The European Journal 2021 - Vol 18 No 2 by Singidunum University

Vol. 18 Nº 2

journal.singidunum.ac.rs

Vol. 18 Nº 2 OCTOBER 2021 journal.singidunum.ac.rs

2021

Economic growth of the tourism sector in a Covid-19 pandemic during 2021 pp. 1-14

Corruption impact on east European emerging markets development pp. 39-54

Structural Breaks, Twitter and the Stock Liquidity of Internet Dot-com Company: Evidence from US Companies pp. 15-38

English digital playground: Friend or foe to the children? pp. 55-72

The long-run impact of bank credit growth on social and economic inequalities in Morocco evidence from the Johansen's cointegration analysis pp. 106-125

The impact of control variables on entrepreneurial intentions among employed persons pp. 126-136 Preferred attributes of employer brand attractiveness among potential employees in the hotel industry pp. 137-150

Macroeconomic determinants of tax revenue in economic community of west African states pp. 73-88 Concentration of supply on the chosen markets of Serbian electronic communications sector pp. 89-105

Political instability and informality in Uganda: an empirical analysis pp. 151-172

Vol. 18 No. 2

The Europiean Journal of Applied Economics has been supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia.

w w w. j ou r na l.sing idunum.a c .rs

Vol. 18 No. 2 Publisher: Singidunum University E d it o r ia l B o a r d

Professor Milovan Stanišić, Singidunum University, Serbia mstanisic@singidunum.ac.rs Professor Francesco Frangialli, Hong Kong Polytechnic University, Hong Kong frangialli@gmail.com Professor Gunther Friedl, Technische Universität München, Germany gunther.friedl@wi.tu-muenchen.de Professor Karl Ennsfellner, IMC University of Applied Sciences, Krems, Austria (karl.ennsfellner@fh-krems.ac.at Professor Gyorgy Komaromi, International Business School, Budapest, Hungary gyorgy@komaromi.net Professor Vasile Dinu, University of Economic Studies, Bucharest, Romania dinu_cbz@yahoo.com Professor Ada Mirela Tomescu, University of Oradea, Oradea, Romania ada.mirela.tomescu@gmail.com Professor Radojko Lukić, University of Belgrade, Serbia rlukic@ekof.bg.ac.rs Professor Alexandar Angelus, Lincoln University, USA angelus@lincolnuca.edu Professor Milan Milosavljević, Singidunum University, Serbia mmilosavljevic@singidunum.ac.rs Professor Olivera Nikolić, Singidunum University, Serbia onikolic@singidunum.ac.rs Professor Goranka Knežević, Singidunum University, Serbia gknezevic@singidunum.ac.rs Professor Mladen Veinović, Singidunum University, Serbia mveinovic@singidunum.ac.rs Professor Jovan Popesku, Singidunum University, Serbia jpopesku@singidunum.ac.rs Professor Zoran Jeremić, Singidunum University, Serbia zjeremic@singidunum.ac.rs Professor Vesselin Blagoev, Varna University of Management, Bulgaria blagoev@vum.bg Professor Michael Minkov, Varna University of Management, Bulgaria minkov@iuc.bg Professor Ionel Bostan, Department of Economics, Al. I. Cuza University, Romania ionelbostan@yahoo.com Associate Professor Christine Juen, Austrian Agency for International Mobility and Cooperation in Education, Science and Research, Wien, Austria chrisine.juen@oead.at Associate Professor Anders Steene, Södertörn University, Stockholm/Hudinge, Sweden anders.steene@sh.se Associate Professor Ing. Miriam Jankalová, University of Zilina, Prague, Czech Republic miriam.jankalova@fpedas.uniza.sk Associate Professor Bálint Molnár, Corvinus University of Budapest, Budapest, Hungary molnarba@inf.elte.hu Associate Professor Michael Bukohwo Esiefarienrhe, University of Agriculture, Dept. of Maths/Statistics, Makurdi, Nigeria esiefabukohwo@gmail.com Associate Professor Goh Yen Nee, Graduate School of Business, Universiti Sains Malaysia, Malaysia yngoh@usm.my Associate Professor Blaženka Hadrović Zekić, Faculty of Economics in Osijek, Croatia hadrovic@efos.hr Research Associate Professor Aleksandar Lebl, Research and Development Institute for Telecommunications and Electronics, Belgrade, Serbia lebl@iritel.com Senior Lecturer Nor Yasmin Mhd Bani, Universiti Putra, Malaysia nor_yasmin@upm.edu.my Roberto Micera, PhD, Researcher, National Research Council (CNR), Italy roberto.micera@ismed.cnr.it Assistant Professor Patrick Ulrich, University of Bamberg, Germany patrick.ulrich@uni-bamberg.de Assistant Professor Jerzy Ładysz, Wrocław University of Economics, Poland jerzy.ladysz@ue.wroc.pl Assistant Professor Konstadinos Kutsikos, University of the Aegean, Chios, Greece kutsikos@aegean.gr Assistant Professor Theodoros Stavrinoudis, University of Aegean, Chios, Greece tsta@aegean.gr Assistant Professor Marcin Staniewski, University of Finance and Management, Warsaw, Poland staniewski@vizja.pl Assistant Professor Gresi Sanje, İstanbul Bilgi Üniversitesi, Istanbul, Turkey gresi.sanje@bilgi.edu.tr Assistant Professor Michaeł Biernacki, Wrocław University of Economics, Poland michal.biernacki@ue.wroc.pl Assistant Professor Piotr Luty, Wrocław University of Economics, Poland piotr.luty@ue.wroc.pl Assistant Professor Vânia Costa, Polytechnic Institute of Cávado and Ave, Barcelos, Portugal vcosta@ipca.pt Assistant Professor Tihana Škrinjarić, University of Zagreb, Croatia tskrinjar@net.efzg.hr Luu Tien Dung, PhD, Lecturer - Researcher, Lac Hong University, Dong Nai, Vietnam dunglt@lhu.edu.vn Assistant Professor Dharmendra Singh, Modern College of Business and Science, Oman dharmendra@mcbs.edu.om Associate Professor Slađana Čabrilo, I-Shou University, Kaohsiung City, Taiwan (R.O.C.) sladjana@isu.edu.tw Ed it o r ia l O f f ice

Editor in Chief: Managing Editor: Technical Editor: English Language Editor:

Professor Žaklina Spalević, Singidunum University Gordana Dobrijević, Associate Professor, Singidunum University Jovana Maričić, Singidunum University Radmila Suzić, Associate Professor, Singidunum University

zspalevic@singidunum.ac.rs gdobrijevic@singidunum.ac.rs jmaricic@singidunum.ac.rs rsuzic@singidunum.ac.rs

Prepress: Miloš Višnjić Design: Aleksandar Mihajlović, MA ISSN: 2406-2588 The European Journal of Applied Economics is published twice a year. Contact us: The European Journal of Applied Economics 32 Danijelova Street, 11010 Belgrade, Serbia Phone No. +381 11 3094046, +381 11 3093284 Fax. +381 11 3093294 E-mail: journal@singidunum.ac.rs Web: www.journal.singidunum.ac.rs Printed by: Caligraph, Belgrade

Access to full text articles: The European Journal of Applied Economics (www.journal.singidunum.ac.rs), SCindeks (https://scindeks.ceon.rs/journaldetails.aspx?issn=2406-2588&lang=en). Copyright © 2021 Singidunum University, Belgrade All rights reserved.

This is an open access journal which means that all content is freely available without charge to the user or his/her institution. Users are allowed to read, download, copy, distribute, print, search, or link to the full texts of the articles in this journal without asking for prior permission from the publisher or the author. This is pursuant to the BOAI definition of open access.

CONTENTS

1 - 14 15 - 35 36 - 48 49 - 61 62 - 75 76 - 94

95 - 126

127- 145

Economic Growth of the Tourism Sector in a Covid-19 Pandemic During 2021 Žaklina Spalević, Snježana Stanišić

Structural Breaks, Twitter and the Stock Liquidity of Internet Dot-com Company: Evidence from US Companies

Osarumwense Osabuohien-Irabor

Corruption Impact on East European Emerging Markets Development

Dušan Dobromirov

English Digital Playground: Friend or Foe to the Children?

Nataša Krstić

Macroeconomic Determinants o Tax Revenue in Economic Community of West African States

Gideon I. Ihuarulam, Gbenga Peter Sanusi, L.O. Oderinde

Concentration of Supply on the Chosen Markets of Serbian Electronic Communications Sector

Milan Kostić, Jelena Živković

The Long-Run Impact of Bank Credit Growth on Social and Economic Inequalities in Morocco Evidence from the Johansen's Cointegration Analysis Ahmed Khattab, El Khlifi Imad

The Impact of Control Variables on Entrepreneurial Intentions Among Employed Persons

Predrag Mali, Bogdan Kuzmanović, Milan Nikolić, Siniša Mitić, Edit Terek Stojanović III

146 - 160

161 - 177

Preferred Attributes of Employer Brand Attractiveness Among Potential Employees in the Hotel Industry Jasmina Ognjanović

Political Instability and Informality in Uganda: An Empirical Analysis Stephen Esaku

EJAE 2021, 18(2): 1 - 14 ISSN 2406-2588 UDK: 338.1:338.48(4)"2021" 616.98:578.834]:338.48 DOI: 10.5937/EJAE18-33977 Original paper/Originalni naučni rad

ECONOMIC GROWTH OF THE TOURISM SECTOR IN THE COVID-19 PANDEMIC DURING 2021 Žaklina Spalević*, Snježana Stanišić Singidunum University, Belgrade, Serbia

Abstract:

Article info:

The tourism sector around the world has been hit hard by the Covid-19 virus pandemic. The consequences of the pandemic during 2020 on the entire tourism sector have significantly reduced the income of both individuals and the collection in the state treasury. The introduced measures, as well as the mass vaccination of citizens, enabled the opening of tourist destinations during 2021, which brought long-awaited revenues to this branch of economics. The opening of tourist destinations has started the travel sector, both air, and road, rail and water transport. The aim of this paper is to analyze the current economic growth of the tourism sector and compare the current situation with the situation in 2019 and the situation during the crisis in 2020. The paper also gives an overview of the legal measures adopted in order to overcome the problems caused by the closure. The analysis of the economic growth of the tourism sector during 2021 was performed on the basis of available data and information both in the world and in the countries of the region. Also, the analysis was performed for both international tourism and domestic tourist destinations. Based on the conducted analysis, it is concluded that the number of tourists who visited the observed regions in the first half of 2021 is still lower compared to the same period in 2019.

Received: September 16, 2021 Correction: September 20, 2021 Accepted: September 23, 2021

Keywords: tourism, Covid-19, Europe, economic growth.

INTRODUCTION The economy of many countries around the world is directly dependent on tourism, as one of the main industries. On the other hand, it can be said that there is almost no country in the world where tourism does not bring a certain profit. Practically observed from the angle of both citizens and the state, if it is a question of foreign tourists who visit a tourist destination, this is a form of income by which new amounts of money are introduced into the state budget. *E-mail: zspalevic@singidunum.ac.rs

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

Precisely for these reasons, we have been constantly working on improving tourism and economic growth, which can be achieved by increasing the number of tourist destinations, as well as increasing the number of both domestic and foreign tourists who visit specific tourist destinations. Closing the doors of tourist destinations either voluntarily or forcibly means great economic losses for all people living in that part of the country from tourism. At the same time, the closure of tourist destinations affects the state budget. The Covid-19 virus during the tourist year 2020 caused the closure of a large number of tourist destinations. Practically the way the virus has spread, as well as the unknowns about the dangers it carries, have led many countries to introduce measures globally to close tourist destinations, close hotels, restaurants, and finally close their borders to foreign tourists. The introduced measures led to the collapse of the tourism economy (Sigala, 2020). The improvement in terms of re-normalization of travel and operation of all tourist facilities began in 2021 after the mass vaccination and immunization of the population around the world. Precisely for this reason, the aim of this research was to analyze the current state of tourism in relation to 2019 and in relation to 2020. In addition to mass vaccination, to mitigate the effects of full closure measures, many countries around the world have enacted legal provisions and decisions to fund tourism. The goal is to compare the current state of tourism with the year that preceded the pandemic (2019) and for which data show that it is one of the best years, as well as with the year in which the pandemic was in full swing. This approach was chosen to show the extent to which tourism could recover to its pre-pandemic status. The paper is organized as follows. The second chapter presents the relevant literature. The third section deals with the economic analysis of the tourism sector during the last year. Within the fourth section, key conclusions and ideas for future work are given.

LITERATURE REVIEW The field of tourism industry largely depends on many socio-economic factors. For example, human relations, wars, pandemics, natural disasters can affect the development of tourism in some parts of the world. It is for these reasons that tourism organizations must have a developed mechanism for acting in crises, as well as a strategy for recovery from them (Qiu, Park, Li, & Song, 2020). A large number of papers examine the crisis that may arise in the tourism sector. In one of the papers, a quantitative study of crisis management was published regarding the crisis in tourism that occurred during the Covid-19 pandemic. The current data obtained from the conducted interviews were analyzed using a statistical tool. The results of the research showed that honest communication is the best solution in terms of fighting the pandemic, and that it is the main trump card when it comes to overcoming the crisis in tourism (Yeh, 2020). In one of the papers, the authors defined different scenarios aimed at predicting the future requirements of international tourism. The basic concept is to consider the further development and struggle of tourism organizations in order to restore tourism after the Covid-19 pandemic. The authors applied two completely different methodologies in their research. By applying these methodologies, they managed to predict the development of tourism in the next year. Using appropriate methods, they developed accuracy metrics which showed that their pedicure models have a high degree of accuracy. The results of the survey showed a large decline in the number of tourist arrivals during 2021 (Fotiadis, Polozos & Huan, 2021). 2

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

As a number of tourist destinations are known in casinos, the Covid-19 pandemic has also affected the gambling sector. One of the research is aimed at determining the impact of Covid-19 pandemics on the gambling industry. Using quantitative analysis and relevant data, the research showed that the gambling industry in Macau is significantly dependent on the number of international tourist arrivals (Lim &To, 2021). As the number of tourist arrivals has fallen sharply, the survey showed a significant decline in GDP, as well as a decline in employment (Lapointe, 2020). During the pandemic, one of the ways to help the tourism sector was through non-repayable economic aid. However, these measures do not cover tourism workers working in the gray zone. In one of the surveys, the authors showed that one in 165 citizens of the European Union prefers to be an unregistered worker. Of course, the research refers to workers in the tourism sector. In order to help these workers during the Covid-19 pandemic, they had to be counted and financially supported (Williams, 2020). The tourism of a country can be divided into national and international. The division is of course made according to the tourists who visit specific tourist destinations. Different experiences in both domestic and international tourism can contribute to its development in each of the countries. During the Covid-19 virus pandemic, there was a significant increase in revenue from domestic tourists around the world. In one of the papers, experiences in terms of domestic and international tourism in Uzbekistan are summarized. Based on the available data, some of the key conclusions have been drawn regarding the stimulation of further development (Nurov, Khamroyeva &Kadirova, 2021). In one of the papers, as the goal of the research an assessment of the short-run economic impacts of the inbound tourism industry on the Australian economy during the pandemic. The research has shown that pandemic has had a major impact on the various industries behind tourism. The authors also conclude that financial assistance to tourism is much needed. Financial assistance to tourism will provide further development of various branches of industry that are directly dependent on tourism (Ngo, Su, Dwyer& Pham, 2021).

ECONOMIC ANALYSIS OF THE TOURISM SECTOR The Covid-19 virus pandemic started at the end of 2019 and has changed until this day the tourist picture around the world daily. Favorite tourist destinations, one day full of tourists, are already closed to visitors the next day (Gretzel, et. al., 2020). The problems that have arisen have greatly affected at least the earnings of the people employed in the tourism sector. The consequences of the Covid-19 pandemic on both domestic and international tourism are certainly more significant precisely because of the fact that the tourism sector is largely dependent on the free movement of people (Stankova, Amoiradis, Velissariou & Grigoriadou, 2021). This is especially the case when it comes to international tourism (Andrades & Dimanche, 2019). The economic deficit of international tourism due to the closure of borders directly impairs the survival of a large number of families whose livelihoods are directly based on income from foreign tourists (Bakar & Rosbi, 2020). The closure of borders during 2020 directly caused the impossibility of movement of tourists, and thus reduced the number of arrivals and reduced wages for all those working in the tourism sector (Kourgiantakis, Apostolakis & Dimou, 2020). The comparison of the number of tourist arrivals in certain parts of the world or regions is directly related to the popularity of tourist destinations in a given region. Also, the increase or decrease in economic income is conditioned by the number of tourists who arrived at a destination as well. In order to make a comparison, the data from the World Tourism Organization were used (UNWTO, 2020). 3

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

If we look at the number of arrivals at the global level during 2020 and compare it with the number of arrivals during 2019, we can clearly see the decline in the number of arrivals (Figure 1 and Figure 2, respectively). Also, based on the available data on the number of tourist arrivals worldwide, it can be seen that the percentage during January 2020 was still growing in the number of tourists. In the following months, a significant decline in the number of tourist arrivals is evident, which corresponds to the period in which measures were introduced in the fight against the Covid-19 virus (Bozovic et. al., 2021). Figure 1. International tourist arrivals in 2020

Figure 2. International tourist arrivals comparison of 2019 and 2020

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

As immunization led to the possibility for tourists to move during 2021, the number of tourist arrivals around the world increased (Gallego & Fort, 2020). Figure 3 shows the percentage number of arrivals for the first 5 months of 2021 (Pantic, 2021). As 2020 was considered one of the worst years in terms of tourism, a comparison of the number of arrivals in the first 5 months of 2021 was made with the same period in 2020. The results of the comparison can be seen in Figure 4. If we compare the decline in the number of arrivals during 2020 and the available data on the number of arrivals during 2021, it is clear that the number of arrivals during 2021 has increased compared to the number of arrivals during 2020. The growth in the percentage of tourist arrivals worldwide in 2021 will begin after March. This growth trend continues during April and May. Projections show that the same growth rate can be expected during the other months of 2021. As there is no data available to show the number of tourist arrivals in other months, the interpretation is based solely on the projection. Figure 3. International tourist arrivals in the first 5 mounts of 2021

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

Figure 4. International tourist arrivals comparison of 2020 and 2021

The decline in the number of arrivals at the world level is also transmitted at the level of the European Union. If we compare the available data on the number of arrivals during 2019, 2020, 2021 in the European Union, we can clearly see the decline in the number of arrivals during 2020 and 2021 compared to 2019 (Figure 5, Figure 6, Figure 7 and Figure 8).Regarding the comparison of the number of tourist arrivals in the first months of 2021, the number of tourist arrivals is less than in the first months of 2020 (Cai, et. al. 2021). This trend can be justified by the fact that the number of tourists traveling around the world in the first months of 2020 was significantly higher because there was no presence of the Covid-19 virus or measures that restricted movement. As for 2021, the first trimester is still characterized by a great potential for the spread of the virus, as well as restrictive measures in terms of travel and movement. If we compare the data from April onwards, it can be clearly seen that in the case of 2020, the number of tourist arrivals decreases significantly in these months, while in the case of 2021, the number of tourist arrivals increases.

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

Figure 5. Number of arrivals on the territory of Europe during 2020

Figure 6. Comparison of the number of arrivals on the territory of Europe during 2019 and 2020

Projections for the remaining period of 2021 as available data on the number of tourists and economic earnings show that the trend of growth in the number of tourist arrivals will continue, and thus revenue in the tourism sector (Puska, et. al, 2020). Measures to combat the Covid-19 virus pandemic also contribute to the increase in the number of tourists, both international and domestic. Namely during this period, active immunization of the population as well as the abolition of restrictive measures contributed to the increase in the number of tourist arrivals (Nientied, & Shutina, 2020). 7

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

The adoption of new measures that include the possibility of travel with the possession of appropriate documents has greatly facilitated the movement of tourists and enabled the renewal of the tourism sector (Cvijanovic, Pantovic, Djordjevic, 2021). Depending on the region or specific country, it still differs which of the documents is needed in order to be able to travel and enter the territory of a given country. For example, some countries require that a permit entering their territory must have a vaccination certificate. On the other hand, some of the countries allow tourists to enter if they have a negative PCR test or a negative antigen test. Figure 7. Number of arrivals on the territory of Europe during the first 5 months of 2021

Figure 8. Comparison of the number of arrivals on the territory of Europe during 2019 and 2021

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

If we compare the percentage of tourist arrivals in the Republic of Serbia as well as in neighboring countries with the data available worldwide, we can conclude that the downward trend in the number of tourists is almost identical. All this can be seen in Figure 9. Also, from Figure 9 it can undoubtedly be seen that in the first months there is still an increase in the number of tourists, while after that there is an evident decline (Demirovic, 2021). Such data can be justified by the fact that the number of infected people in the territory of the Republic of Serbia started to grow at the beginning of March. Also in the first months of 2020, a large number of Serbian citizens who were temporarily frightened by the Covid-19 pandemic while working abroad returned to the country, which caused an increase in the number of arrivals (Beraha, Djuricin, 2020). Figure 9. Number of arrivals on the territory of the Republic of Serbia and surrounding countries during 2020

Figure 10. Number of arrivals on the territory of the Republic of Serbia and surrounding countries during the first five months of 2021. year

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

In May 2021, in accommodation facilities in the Republic of Serbia, there were 194 479 registered tourist arrivals, which means it increased by 307.5% when compared to the same period 2020. In relation to May 2020, the number of foreign tourist arrivals increased by 1371.0%, while the number of domestic tourist arrivals increased by 219.7%. In May 2021, the number of 585 798 tourist overnight stays was recorded, 74.7% of which were made by domestic tourists, and 25.3% by foreign tourists. Compared to May 2020, the number of overnight stays increased by 243.2%. In relation to the same period 2020, in May 2021 the number of overnight stays of foreign tourists increased by 689.4%, while the number of overnight stays made by domestic tourists increased by 188.1%. Regarding the structure of foreign tourist overnight stays, in May 2021 the largest share was made up the tourists from Bosnia and Herzegovina (10.8%), Turkey (9.5%), followed by tourists from Russian Federation (9.2%). Observed by regions for the first seven months of 2021, tourist accommodation capacities in the north of Serbia recorded 4 287 116 registered overnight stays. On the other hand, tourist accommodation capacities in the south of the Republic of Serbia recorded 2 918 515 registered overnight stays. As for the north of the Republic of Serbia, out of the total number of 741 435 overnight stays, it refers to foreign tourists, while the remaining number of overnight stays is by domestic tourists. As for the south of the Republic of Serbia, in the first seven months of 2021, 121 618 overnight stays belonged to foreign tourists, while the remaining number of overnight stays was made by domestic tourists. If we look at individual tourist destinations, the total number of nights spent by domestic and foreign tourists in Vrnjacka Banja was 129 192 and it can be said that it is in the lead when it comes to spa tourism. If we look at mountain tourism in the first seven months of 2021 – 111 640. In the category of cities, Belgrade realized the largest number of overnight stays – 271 821. As in other regions, certain European governments helped minimise the decline in domestic spending through stimulus initiatives. Italy, for instance, implemented the ‘Italy Cure’ rescue plan in May 2020, which included a ‘holiday bonus’ of up to EUR 500 that low-income families could spend on tourism accommodation (Mandic, 2021). Further support announced in August included grants for tourist activities open to the public in the historic centres of art cities, and EUR 15 million for tourism promotion. Visit Sicily launched the ‘See Sicily’ voucher scheme, offering tourists of the island a discount on flights, a free night’s stay, a free tour, and entry to a cultural attraction. Travel & Tourism employment fell by 9.3%, equating to 3.6 million jobs; however, the situation could have been far worse if there were not the government’s prompt action, which introduced job retention schemes to save millions of jobs under threat. In fact, job protection schemes were introduced in many European countries, including the largest Travel & Tourism economies such as France, Germany, Italy, Spain, and the United Kingdom, with different levels of support. The UK’s Job Retention Scheme brought significant relief to millions of employees across the UK whose jobs have been sustained. The furlough scheme, as it is also known, has been in place since March 2020 and will end in September 2021 (Payne, Gil-Alana, & Mervar, 2021). For most of this period, the grant covered 80% of wages up to GBP 2,500 (USD 3,500) for employees kept on payroll but with no work, as well as national insurance and pension contributions. Between March 2020 and mid-February 2021, 11.2 million jobs were furloughed across the UK, with GBP 53.8 billion paid out across the country. A similar scheme was also set up for the self-employed. In terms of the global rankings, some European countries improved while others fell in the rankings. For example, Germany had a decline of 46.9% in GDP, while Italy at the same time had a decline of 51% of GDP. If we look at France and the United Kingdom as one of the leading tourist destinations in terms of the number of arrivals, the data show that they experienced a decline of 48.8% and 62.3% respectively (Qiu, et al. (2020). 10

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

A combination of stringent lockdowns, continuing travel restrictions and damaging quarantines caused it to suffer the biggest collapse of the 10 largest Travel & Tourism markets (Prideaux, Thompson, & Pabel,2020). What is more, the Netherlands rose two places from 15th to 13th position following a 36.5% drop in Travel & Tourism GDP (below the regional and global average decline), while Switzerland joined the top 20 largest Travel & Tourism economies following a GDP decline of 35.4%. While Spain was the top country in Europe and third worldwide for international visitor spending in 2019, it fell three places in 2020 and was overtaken by France, Germany, and Italy. In fact, international visitor spending experienced an unprecedented drop of 78.2%, partly due to restrictions in the country’s key source markets such as the UK. Tourism Revenues in Greece increased to 790.71 EUR Million in June from 168.28 EUR Million in May of 2021 (Sigala, 2020).

CONCLUSIONS The Covid-19 pandemic has changed a lot since its inception in the daily lives, work, and habits of people around the world. Various measures to combat this virus came into force and were abolished both locally and globally. The reduction of travel, both domestic and foreign tourists has greatly affected the development of tourism. The decrease in the number of tourists has reduced the potential for the continued growth of income from tourism in relation to 2019 and all the years before it. The year behind us was very unfavorable because of all the above for all people who are directly related to tourism and who live from it. The lifting of travel restrictions allowed an increase in tourism revenues during 2021 compared to 2020. The conducted research showed that the number of tourists both in the Republic of Serbia and in the world has significantly increased compared to 2020. The increased number of tourists directly conditioned the increased number of overnight stays, and thus the increased income. Observed on the example of Greece, earnings in the first half of the year are significantly higher than in the same period in 2020. However, although there has been a significant increase, direct earnings from tourism have not reached the 2019 level. In order for the tourism sector to return to the level before the Covid-19 virus pandemic and for the total earnings to reach the desired level, it is necessary to achieve as many arrivals and overnight stays as possible. As the number of overnight stays is directly conditioned by the reduction in the spread of the Covid-19 virus, it is necessary to build a strategy on a global level. All actors in this strategy must contribute. The key conclusion of this research is that even in a pandemic, with appropriate legal and economic measures, it is possible to improve the tourism sector.

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

REFERENCES Andrades, L., & Dimanche, F. (2019). Destination competitiveness in Russia: tourism professionals’ skills and competencies. International Journal Contemp Hospit Management. 31(2), pp. 910–930. https://doi.org/10.1108/IJCHM-11-2017-0769 Bakar, A.N., & Rosbi, S. (2020). Effect of Coronavirus disease (COVID-19) to tourism industry. International Journal of Advanced Engineering Research and Science (IJAERS).7(4), 189-193. Beraha, I., & Djuricin, S. (2020). The Impact of COVID-19 Crisis on Medium-sized Enterprises in Serbia. Institut ekonomskih nauka, pp. 14-27. Bozovic, T., Blesic, I., Knezevic Nedeljkovic, M., Djeri, L., & Pivac, T. (2021). Resilience of tourism employees to changes caused by the Covid-19 pandemic. Journal of the Geographical Institute “Jovan Cvijić” SASA. 71(2), 181-194. Cai, W., Gebbels, M., & Shukri, W. (2021). Performing authenticity: Independent Chinese travellers’ tourism dining experiences in Europe. Tourism Management. 86, 104339. Cvijanovic, D., Pantovic, D., & Djordjevic, N. (2021). Transformation from urban to rural tourism during the covid-19 pandemic: the case of Serbia. In D. Cvijanovic (Ed.) Sustainable Agriculture and Rural Development (pp. 123-132). Belgrade, Serbia. Fotiadis, A., Polozos, S., & Huan, T.C. (2021). The good, the bad and the ugly on COVID-19 tourism recovery. Annals of Tourism Research. 87, 1-35. Gallego, I., & Fort, X. (2020). Changes in air passenger demand as a result of the COVID-19 crisis: using Big Data to inform tourism policy. Journal of Sustainable Tourism. 29(9), 1470-1489. Demirovic, B., Terzic, A., Petrovic, M. D., Radovanovic, M., Tretiakova, T. N., & Hadoud, A. (2021). Will we have the same employees in hospitality after all? The impact of COVID-19 on employees’ work attitudes and turnover intentions. International Journal of Hospitality Management. 94, 102754. Gretzel, U., Fuchs, M., Baggio, R., Hoepken, W., Law, R., Neidahardt, J., Pesonen, J., Zanker, M. & Xiang, Z. (2020). e‑Tourism beyond COVID‑19: a call for transformative research. Information Technology & Tourism. 22, 187-203. Kourgiantakis, M., Apostolakis, A., & Dimou, I. (2020). COVID-19 and holiday intentions: the case of Crete, Greece. Anatolia An International Journal of Tourism and Hospitality Research. 32(1), 148-151. Lapointe, D., (2020). Reconnecting tourism after COVID-19: the paradox of alterity in tourism areas. Tourism Geographies An International Journal of Tourism Space, Place and Environment. 22(3), 633-638. Lim, W.M., & To. W. M (2021). The economic impact of a global pandemic on the tourism economy: the case of COVID-19 and Macao’s destination- and gambling-dependent economy. Current Issues in Tourism. 1-13. https://doi.org/10.1080/13683500.2021.1910218 Mandic, A. (2021). Protected area management effectiveness and COVID-19: The case of Plitvice Lakes National Park, Croatia. Journal of Outdoor Recreation and Tourism. 100397. https://doi.org/10.1016/j.jort.2021.100397 Phama T. D., Dwyer L., Sua J-J., & Ngo T. (2021). COVID-19 impacts of inbound tourism on Australian economy. Annals of Tourism Research. 88, 103179. https://doi.org/10.1016/j.annals.2021.103179 Nientied, P. & Shutina, D. (2020). Tourism in transition, the postcovid-19 aftermath in the Western Balkans. Co-Plan Resilience Series. 2, 1-20. Nurov, Z .S., Khamroyeva, F. K., & Kadirova. D. R. (2021). Development of domestic tourism as a priority of the economy. In Z. V. Chevychalova (Ed.) Research Innovation In Multidisciplinary Sciences, (pp. 271-275). New York, USA.. Pantic, N. (2021). Tourism Challenges amid Covid-19. Hotel and Tourism Management. 9(1), 145-148. Payne, J., Gil-Alana, L., & Mervar, A. (2021). Persistence in Croatian tourism: The impact of COVID-19. Tourism Economics. 1-7. https://doi.org/10.1177/1354816621999969 12

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

Peric, G., Dramicanin, S., Conic, M. (2021). The impact of Serbian tourists' risk perception on their travel intentions during the COVID-19 pandemic. European Journal of Tourism Research, 27, 1-22. Puska, A., Pamucar, D., Stojanovic, I., Cavallaro, F., Kaklauskas, A., & Mardani, A. (2020). Examination of the Sustainable Rural Tourism Potential of the Brčko District of Bosnia and Herzegovina Using a Fuzzy Approach Based on Group Decision Making. Sustainability, 13(2), 583. Prideaux, B, Thompson, M, &Pabel, A (2020). Lessons from COVID-19 can prepare global tourism for the economic transformation needed to combat climate change. Tourism Geographies, 22(3), 667–678. Qiu, R., Park, J., Li, S. & Song, H. (2020). Social costs of tourism during the COVID-19 pandemic. Annals of Tourism Research. 84. 102994. https://doi.org/10.1016/j.annals.2020.102994 Sigala, M. (2020). Tourism and COVID-19: Impacts and implications for advancing and resetting industry and research. Journal of Business Research. 117, 312-321. Stankova, M., Amoiradis, C., Velissatiou, E., & Grigoriadou, D. (2021). Accessible tourism in Greece: a satisfaction survey on tourists with disabilities. Management Research and Practice. 13(1), 5-16. Sumner, A., Hoy, C. & Ortiz-Juarez, E. (2020). Estimates of the impact of COVID-19 on global poverty. WIDER Working Paper 2020/43. UNU-WIDER. https://doi.org/10.35188/UNU-WIDER/2020/800-9 United Nations World Tourism Organization (UNWTO) (2020). UNWTO World Tourism Barometer and Sta- tistical Annex, UNWTO World Tourism Barometer. 18(6), 1–23. Wachyuni, S. S., & Kusumaningrum, D. A. (2020). The Effect of COVID-19 Pandemic: How are the Future Tourist Behavior? Journal of Education, Society and Behavioural Science. 33(4), 67-76. Williams, C.C. (2020). Impacts of the coronavirus pandemic on Europe's tourism industry: Addressing tourism enterprises and workers in the undeclared economy. International Journal of Tourism Research. 23. 79-88. Yeh, S. S. (2020). Tourism recovery strategy against COVID-19 pandemic. Tourism Recreation Research. 46(2), 188-194.

EJAE 2021  18 (2)  1 - 14

SPALEVIĆ. Ž., STANIŠIĆ. S.  ECONOMIC GROWTH OF THE TOURISM SECTOR IN A COVID-19 PANDEMIC DURING 2021

EKONOMSKI RAST SEKTORA TURIZMA TOKOM PANDEMIJE COVID-19 2021. GODINE Rezime: Turistički sektor širom sveta teško je pogođen pandemijom virusa Covid-19. Posledice pandemije tokom 2020. na čitav turistički sektor značajno su smanjile prihode pojedinaca i naplatu u državnoj kasi. Uvedene mere, kao i masovna vakcinacija građana, omogućile su otvaranje turističkih destinacija tokom 2021. godine, što je ovoj grani ekonomije donelo dugo očekivane prihode. Otvaranjem turističkih destinacija pokrenut je sektor putovanja, vazdušni, drumski, železnički i vodeni saobraćaj. Cilj ovog rada je analiza trenutnog ekonomskog rasta turističkog sektora i upoređivanje trenutnog stanja sa stanjem u 2019. i sa stanjem tokom krize 2020. U radu se daje i pregled zakonskih mera usvojenih u cilju prevazilaženja problema izazvanih zatvaranjem. Analiza ekonomskog rasta turističkog sektora tokom 2021. godine izvršena je na osnovu dostupnih podataka i informacija kako u svetu tako i u zemljama regiona. Takođe, analiza je urađena kako za međunarodni turizam tako i za domaće turističke destinacije. Na osnovu sprovedene analize zaključuje se da je broj turista koji su posetili posmatrane regione u prvoj polovini 2021. godine i dalje manji u odnosu na isti period 2019. godine.

Ključne reči: turizam, Covid-19, Evropa, ekonomski rast.

EJAE 2021, 18(2): 15 - 35 ISSN 2406-2588 UDK: 336.761.5 658.153(73)"2019/2020" DOI: 10.5937/EJAE18-27857 Original paper/Originalni naučni rad

STRUCTURAL BREAKS, TWITTER, AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES Osarumwense Osabuohien-Irabor* Department of International Economics, School of Economics and Management, Ural Federal University, Russia

Abstract:

Article info:

The goal of this paper is to explore relationship between Twitter and stock liquidity of some large US internet Dot-com companies in the presence of unknown structural breaks for the period from September 2019 to April 2020. Using the Andrews-Ploberger and Andrews-Quandt structural break models, we identify the major structural breakpoints in the stock liquidity and find that most of these structural changes are significantly perceived. When we examined the sub periods as well as the full sample, Tweets and likes from most numbers of companies were found not to have links with stock liquidity. These results provide crucial insight into portfolio strategy to both international and local investors.

Received: February 7, 2020 Correction: October 30, 2020 Accepted: September 20, 2021 Keywords: structural breaks, Twitter, stock company, Andrews-Ploberger, liquidity, regime.

INTRODUCTION Dot-com companies are companies that do their businesses strictly on the internet with a known website on the World Wide Web (WWW) with the domain “.com”. The “.com” domain of a website URL usually (but not always) indicates commercial or profit oriented companies (compared to the companies with “.org” domains which are usually used for commercial or non-prof organization), see Wikipedia1. The “.com” companies conduct their businesses, be those products or services, via webbased mechanism, even when tangible goods, products, or services are involved. However, some “.com” companies do not deal with tangible products. Many of these companies respond or communicate with customers and investors through their social media handles, particularly Twitter. Scientific research has shown the existence of a relationship between Twitter’s tweets, “likes”, retweets, etc., and stock market activities, (see Pöppe et al., 2020, Guijarro et al., 2019, Shiva and Singh, 2019, Broadstock and Zhang, 2019). Structural change is common in stock prices relationships, and it can be quite risky to ignore. 1 Wikipedia, Dot-com company https://en.wikipedia.org/wiki/Dot-com_company

*E-mail: oosabuokhien-irabor@urfu.ru

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

This can cause inferences to go astray, inaccurate forecasts, and misleading policy recommendations. Therefore, the goal of this paper is to explore the role of Twitter and stock liquidity of US internet Dotcom companies in the presence of unknown structural breaks. In past decades, investors got information about market situation through watching television, reading newspapers, or by word-of-mouth from friends and families. But with the advent of social media, the ways how information is generated and dispersed on financial markets have fundamentally been transformed (Dugast & Foucault, 2016). Unarguably, social media have significantly influenced daily human lives and changed the way individuals and businesses perform, create awareness, and seek advice (Nisar & Yeung, 2018). For the past decade, Twitter has remained one of the largest social media microblogging service providers that has shown steady growth both in its services and the number of subscribers. Twitter networking service allows users (subscribers) to post and interact with messages referred to as "tweets." Launched in 2006, Twitter currently has 330 million active subscribers and 145 million daily active users (Twitter, 2020). Twitter has grown in popularity and reliability as a means of messaging for individuals, and an official channel of communication by many corporate organizations. Tweets from Twitter usually contain short text messages within 140 characters. Just like individuals, companies also maintain Twitter accounts, so as to create a two-way communication channel where customers can publicly communicate with companies and leave the record public. With abundance of information available online, scholars and practitioners have successfully applied Twitter data to predict and analyze several variables ranging from health, politics, academics, biology, financial markets, and stock markets, etc. Twitter has widespread coverage and is generally accepted both in the financial sector and research community. However, avalanche of studies exists on the relationship between Twitter and financial markets. And some studies (such as those conducted by Saurabh and Dey (2020), Albarrak et al. (2020), Affuso and Lahtinen (2019), Ge et al. (2019), Behrendt and Schmidt, A. (2018), Zhang et al. (2018), etc.) have revealed the existence of a link between Twitter and stock market. Also, numerous studies and events have clearly shown that Twitter influences both the companies and the market. For example, on the 23rd of April 2013, the Associated Press Twitter account was hacked, and the hackers posted “Breaking: Two Explosions in the White House and Barack Obama is injured”. This incident caused a 0.9 percent immediate decline in the S&P 500. Empirical studies have revealed that Tweets and sentiments associated with Twitter have shown to impact return on investments. If so, does it also affect stock liquidity? Also, given that under the same economic conditions, stock liquidity of separate companies is different, the question is: Does company capitalization affect the relationship between Twitter and stock liquidity? We provide an empirical analysis of the role of social media, specifically Twitter and stock liquidity by tracking the history of individual companies tweets and the corresponding likes. Liquidity is a complex concept and one of the most researched area in theory of finance simply because of its role in functioning of the financial markets. Hence, O’Hara (2004) said “liquidity is hard to define, but easy to feel it”. Generally speaking, liquidity refers to the ease with which assets are sold immediately after purchase without incurring any forms of losses. These losses could be results of price changes or various transaction costs. Therefore, whenever investors consider investing an asset, some of the first things to thoroughly consider are: the ability to re-sell it, its future cost, as well as the price to sell at. These various forethoughts relate to asset liquidity and these issues which are to be considered can potentially affect the cash flow of the asset. Future cash flows are known to affect liquidity; hence it is seen as an important factor in asset pricing. However, forced sales with regards to price reduction and cost trading are not pricing factors that are significantly related to a financial asset like stock. Therefore, higher measure of non-liquidity attracts the risk of higher losses for the investors along with higher gains in comparison to the liquid markets due to price volatility. 16

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

On liquidity market, investors remain uncertain when it comes to performing large transaction as it may create significant price change which can cause higher losses. Therefore, stock market development is impeded as higher illiquidity lowers down capital inflows. Apart from financial stock market liquidity which this study focuses on, the concept of liquidity can also be explained in other forms: (i) asset liquidity; (ii) an asset market liquidity; (iii) a financial market liquidity and (iv) the liquidity of a financial institution. Understanding the microstructure of the market is important. Hence, number of studies have proposed liquidity measures as proxies for investors’ liquidity and transaction costs. Datar et al. (1998) proposed liquidity tests based on turnover rate. Specifically, whilst the former uses the turnover rate as a proxy for liquidity that correlate with trading frequency, the latter employs the turnover rate in a cross-sectional regression to perform an experiment. Their study reveals stock returns as a decreasing function of the turnover rates. Amihud (2002) introduced the illiquidity index to measure liquidity with regard to the traded volume. The illiquidity measure is the average of the daily impacts over a particular sample period, thus provides an understanding of the relations between volume and price change. Apart from the volume-type liquidity indices already enumerated, another group of liquidity measure is the price-type liquidity index. This category includes measures that use asset or market liquidity based on price behavior (Gabrielsen et al., 2011). The Market Efficient Coefficient (MEC) is another liquidity measure that assesses the effects of execution costs on price volatility over short period of time. This measure is also known as the variance ration which is a widely used liquidity index in many empirical references. The notion is that more liquid market indicates smaller variance of transaction around equilibrium price. The bid-ask spread, and its variants have also been relied on and used by economists and other market participants as liquidity measure. This is because, it conveys insightful information about market conditions. The bid-ask spread can also be explained as the difference between the lowest ask price and the highest bid price. However, in this study, we apply the Amihud (2002) illiquidity ratio to determine the stock liquidity for the US internet Dot-com companies. It is the most commonly used and generally accepted liquidity proxy by scientists, academia, economists, stocks participants, etc. Nevertheless, numerous studies have examined the relationship between Twitter and stock market activities (Sakhare et al., 2020; Fan et al., 2020; Broadstock & Zhang, 2019; Chahine & Malhotra, 2018). While most of these studies found meaningful relationship, other literature documents offered contrary results. Similarly, it is of common knowledge that economic shocks, global pandemic, political incidents and unrests, policy alterations, etc., greatly affect companies’ revenue. According to Amihud (2002) conventional liquidity ratio, etc., are known market liquidity indices which depend on the traded stock volume and change in prices. And if this is so, do companies’ tweets and “likes” influence liquidity in the presence of structural changes? Again, are these structural breaks in companies’ liquidity connected to or relevant for the global pandemic crisis of COVID-19 even as the companies continue to tweet? First, we attempt to identify the structural break points in the time series of companies’ liquidity with external Twitter variables, using Andrews-Ploberger (1994) and Andrews-Quandt Structural Break Tests which allow for the presence of breaks in a linear model. The rest of this paper is organized as follows. Section 2 reviews the relevant literature related to structural break, Twitter, and stock liquidity. Section 3 discusses the data as well as methodology applied in our empirical analysis. Section 4 reports and discusses the results of our empirical analysis, whereas the section 5 is the summary of our research study.

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

LITERATURE Factors which can cause structural changes in countries, markets, and industries include new economic development, global shifts in labor and capital, changes due to disaster, changes as a result of global pandemic, changes due to demand and supply of resources. That is why academia, policy makers, and other concerned stakeholders conduct ex ante or ex post research to investigate theses structural changes in numerous financial and economic variables. Ruch et al. (2020), Anguyo et al. (2020), Hegerty (2020), Nath and Sarkar (2019), Gil-Alana (2019), Orlowski (2017) examine the break dates and the impact of structural changes on inflation series with regards to other variables. Ruch et al. (2020) predict inflation variables using factor-augmented VARs (FAVAR), time-varying parameter vector autoregressive models (TVP-VARs), and structural break models. They found that models with heteroscedastic errors performed better than models with homoscedastic errors. Their results also showed that structural break did not enhance the predictability of inflation. Anguyo et al. (2020) examine structural changes and measurement of inflation persistence over time using the Uganda data. Using the regression quantile, they find higher levels of persistence after 2006 and during the inflation targeting period. Hegerty (2020) study reveals that no Central and Eastern European members have inflation rate with break point that corresponds to Euro adoption. Results also show that CEE members have multiple breaks with final structural breaks occurring in 2013. Nath and Sarkar (2019) investigate relations between inflation and relative price variability (RPV) using quarterly consumer price index for seventy-four (74) consumption categories in Australia. The results of their empirical research show that a J-shaped non-line relationship between inflation and unexpected inflation exists. Two structural breaks were also identified in inflation-RPV relationship. Gil-Alana et al. (2019) study the behavior of inflation rate in Iran from 1992 to 2017, using fractional integration. He documents extremely large degree of persistence in series with an order 2 integration. Orlowski (2017) uses Bai-Perron multiple breakpoint and two-state Markov regime switching tests to investigate the sensitivities of Poland’s interest rate to inflation and exchange rate for over two decades. His results reveal a major structural break and regime change at the start of 2002. The break timing reflects de facto inflation targeting effective and credible policy. Other references in this category include Gil-Alana and Mudida (2017), Clemente et al. (2017), etc. Sani et al. (2020), Nasir and Vo (2020), Liu et al. (2020), Phiri (2020) examine the structural changes of foreign exchange rates in countries and microeconomics variables. Sani et al. (2020) examine relationship between exchange rate and interest rate differential for the BRICS countries. Their results show that the exchange rate predominantly responds asymmetrically to the interest rate differential in four out of the five countries examined. Nasir and Vo (2020), using the monthly data from October 1976 to September 2017, investigate implications of Inflation Targeting (I.T.) for the exchange Rate PassThrough (ERPT) to inflation and trade balance for New Zealand, UK, and Canada. They employed the TVSVAR framework, and their result show a time-variation in the ERPT to inflation and trade balance in the three countries. Liu et al. (2020) examine the relationship between the US exchange rate and crude oil prices within the structural break detection context. Results reveal that crude oil prices shocks have both immediate and short-term effect on exchange rate movements which are emphasized during the confidence intervals of structural breaks. Phiri (2020) investigates the impacts of two structural events on exchange rate-equity returns nexus for four (4) Johannesburg Stock Exchange (JSE) indices using the nonlinear autoregressive distributive lag (ARDL) cointegration.

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

A monthly sample data was collected from 2000 to 2017 and the empirical analysis shows that sub-periods correspond to breaks caused by crisis and the use of new trading platform. Their results also suggest that prior to the crisis, exchange rates appreciations cause stock returns. But depreciations show to unlikely cause stock returns to decrease. Other references include Jeelani et al. (2019) who study the relationship between India’s macroeconomic factors that affect exchange rate (ER) (INR/ USD). Data such as ER, GDP, inflation, interest rate (IR), FDI, money supply, trade balance (TB) and terms of trade (ToT) were collected from the RBI website. The study investigates whether there is any structural break with the application of the Chow’s Breakpoint Test. Multiple structural breaks were found between 2003 and 2009 which explained the fact that volume of crude oil imported by India is high, and oil price rise led to a deficit in TB which caused a structural change. Relationship between structural breaks and revenue is another growing strand in literature. GilAlana et al. (2019) analyze the structural pattern of Brazilian tourism revenue over the period of 20 years. Results show that benefits obtained from tourism revenue can jeopardize the economic structural problems reflected in currency fluctuation. Kumar et al. (2019) investigate the effects of tourism industry on the economic growth of Fiji over a period from 1975 to 2015, using a neoclassical and autoregressive distributed lag (ARDL) bound framework while accounting for structural breaks. The short-run and long-run results reveal that 1% increase in visitor’s arrival contributes with about 0.20% and 0.13% to the per capita income, respectively. Min et al. (2019) test the structural break of visitor’s arrival to Taiwan from China, Japan, the USA and Hong Kong. The Bootstrap with multiple structural break framework is used to model the Taiwan’s four inbound time series. They document that the severe acute respiratory syndrome (SARS) outbreak impacts the Taiwan’s four major inbound markets. Stauvermann et al. (2018) explore the relationship between tourism, exchange rate and the economic growth in Sri Lanka from 1980 to 2014. They employ the Augmented Solow and ARDL framework for the empirical analysis. Their results reveal a long-run relationship among tourism, exchange rate, capital per worker and output per worker. In addition to investigating the structural breaks in the tourism industry, Amiraslany et al. (2019) examine structural breaks in biased estimation and forecast errors in GDP series of Canada against the USA. Their result suggests a structural break for Canadian gross domestic product (GDP) when there was a switch from the Standard Industrial Classification system (SIC) to the North American Industry Classification (NAIC) System. Their results also reveal that failure to identify in-sample breaks may adversely affect the model’s out-of-sample forecast. One of the major concerns of many scientific researchers and policy makers is whether Twitter’s posts and sentiments influence stock market movements. Thus, huge volume of literature has been documented in this strand. Urlam et al. (2020) employ the Long Short-Term Memory (LSTM) neural networks to predict and analyze stock market data. Their results confirm that longer horizon prediction is more useful than the shorter horizon prediction. Their study also compares sentiments analysis and the predicted stock value, and the results reveal that the two are similar. Emotions also show to affect the future of stock prices. Ajjoub et al. (2020), Brans and Scholtens (2020), Klaus and Koser (2020), Ge et al. (2019) examine the impact of presidential tweets on stock prices. Their results show that President Donald Trump’s tweets impact the stock prices, increase trading volume and volatility. Positive tweets were shown to have a pronounced positive impact on the stock price. Besides, negative and neutral tweets have little or no effect on the stock market. Reboredo and Ugolini, (2018) investigate the effects of Twitter sentiment and sentiment mood on stock returns, trading volumes, and volatility for renewable energy stocks. They document that Twitter sentiment has no impact on stock market activities, but Twitter sentiment divergence is shown to generate feedback effects on volatility and trading volumes. 19

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Chahine and Malhotra (2018), Reed, M. (2016) in their studies also examine the impact of social media on the stock market. While the former finds significant market reaction around Twitter activities for subsample of firms contaminated by corporate announcements but not for full sample. The latter empirical results show that sentiment impacts stock prices, particularly, S&P 500 and the Dow Jones Industrial Average. Guijarro et al. (2019) investigate whether Twitter sentiment impacts the financial market liquidity in S&P500 Index. Their results find that the investors' mood had little influence on the spread of the index. Other literature references include: Groß-Klußmann (2019), Wu (2019), Shelar and Huang, (2018). However, our study seeks to investigate the impact of companies’ Twitter posts and “likes” on its stock liquidity in the presences of structural changes due to global pandemic. And to the best of our knowledge, no previous study exists in this direction of research. We employ Andrews-Ploberger (1994) and Andrews-Quandt Tests, which allow for the presence of structural break in linear model to examine the US internet Dot-com companies structural break date in stock liquidity using Amihud (2002) liquidity proxy, as well as the impact of Twitter posts for the different sub periods. We found that for nearly all the companies examined, the structural breaks captured are significant. Our results also reveal that neither the Twitter’s posts nor “like” appear to influence investors’ ability to exchange an asset for cash.

HYPOTHESES DEVELOPMENT Many previous studies such as Affuso, E., Lahtinen, K.D. (2019), Zhang, X., Fuehres, H., Gloor, P.A. (2012), Vanstone, B.J., Gepp, A., Harris, G. (2019), Makrehchi, M., Shah, S., Liao, W. (2013), Sul, H.K., Dennis, A.R., Yuan, L. (2014), etc., reveal that Twitter sentiment impacts stock price returns. And that there is a strong link between social media and stock returns. There are other studies that document company-initiated news via Twitter leads to the improved liquidity of that company’s stock. That is why one of our aims is to test the following hypotheses on the US internet Dot-com companies stock market: H1: Twitter posts from specific US internet Dot-com companies increases its Stock liquidity H2: Twitter likes from specific US internet Dot-com companies increases its Stock liquidity

METHODOLOGY AND DATA The daily stock data used in this study is obtained primary from Yahoo! Finance (http://finance. yahoo.com/). Then a total of 571,499 and 20,912,993 English language tweets and ‘likes’ from the twelve (12) US internet Dot-com companies examined were collected for the period from September 4th, 2019. to April 1st, 2020. Information on companies’ tweets and ‘likes’ were accessible via https://popsters. com/ website’s application programming interface (API). In 2012, Twitter introduced the addition of a cashtag sign to stock tickers to stress on the stock being referred to. For example, Microsoft’s tweets were query by $MSFT, Facebook Tweet query by $FB, Apple by $AAPL, etc. Therefore, we used the cashtag ($) preceding the internet Dot-com companies’ ticker symbols to search for each companies’ Twitter information, such as tweets, likes, messages, retweets, etc. 20

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

To avoid noisy data, we cleaned all ‘spam’, duplicated tweets, and other irrelevant tweets associated with datasets. ‘Like’ is a function on many social media platforms which indicates engagement or validation with a piece of content, such as message. According to Cabellon and Ahlquist (2016), like is a form of external validation for social media posts. And the more ‘likes’ a post gets, the more positively the user is perceived. Table 1 shows that Amazon.com company, an American multinational technology company based in Seattle which focuses on e-commerce, cloud computing, digital streaming, and artificial intelligence, had the biggest revenue of $280.50 billion, market capitalization of $920.22 with 798000 employees in the US in 2019 fiscal year. The Facebook.com company also has the highest numbers of tweets and the corresponding ‘likes’ as shown in Table 2. While Table 2 also shows Twitter data and the search features, Table 3 describes the companies’ stock data used in our analysis. The statistical property for variables such as liquidity, stock volume, firm size, and stock returns were examined in our pre-empirical analysis. With exception of stock returns, each variable appears to be a standard normal distribution with a mean of zero and standard deviation of 1, see Table 3. The standard normal distribution is centered at zero and the measurement of the degree of deviates from the mean is given by the standard deviation. The interquartile range (IQR) as a measure of dispersion indicates that our stock data are not spread out. This is evidence in the 25th and 75th quartile scores. Table 1. Large International Internet Companies located in US., ranked by total revenues and market capitalization for their respective fiscal years ended on or before March 31, 2019. FY

Nos. of Employee

Market Cap ($B)

2019

798000

$920.22

Seattle, USA

1994

$161.80

2019

118,899

$921.14

Mountain View

1998

Facebook

$70.69

2019

45,000

$585,37

Menio Park, USA

2004

Tesla

$24.58

2019

48,016

$75.72

Palo Alto, USA

2008

Netflix

$20.16

2019

8,600

$141,98

Los Gatos, USA

1997

PayPal

$17.77

2019

23,200

$126.88

San Jose, USA

1998

Salesforce.com

$17.10

2019

49,000

$161.71

San Francisco

1999

Booking Holdings

$15.06

2019

26,400

$85.06

Norwalk, USA

1996

Expedia

$12,07

2019

25,400

$15.42

Bellevue, USA

1996

Adobe

$11.17

2019

22,634

$149.30

San Jose, USA

1982

eBay

$10.80

2019

13,300

$28.74

San Jose, USA

1995

Wayfair

$09.13

2019

16,983

$08.50

Boston, USA

2005

Company Name

Revenue ($B)

ending

Amazon

$280.50

Google

Rank

Headquarters Location

Year Founded

Note: Revenue: Annual revenue of company in USD billion in previous fiscal year FY: Company's fiscal year Employee: Number of employees of company Market Cap.($B): Market capitalization as of March 2019 in USD billion Company: Name of the international company with Headquarter in USA Headquarter: Location of company's headquarters

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Table 2. Companies and Twitter Statistics(Dot.Com) Some Ranked US. Largest Internet Twitter Features Rank

Company Name

Twitter Data

Stocks Tickers

Search Keywords

Total Tweets Collected

Tweets After filtering

Like Tweets

Amazon.com, Inc.

AMZN

$AMZN

121007

92883

172416

Alphabet Inc, Class A.

GOOGL

$GOOGL

96123

61803

889104

Facebook, Inc.

$FB

179023

156283

7557101

Tesla, Inc.

TSLA

$TSLA

156115

124084

483837

Netflix, Inc

NFLX

$NFLX

88324

62239

1976294

PayPal Holdings, Inc.

PYPL

$PYPL

44997

17394

739551

Salesforce.com, Inc.

CRM

$CRM

9768

1115

353498

Booking Holdings

BKNG

$BKNG

18765

11607

576023

Expedia, Inc.

EXPE

$EXPE

16664

10731

1122085

Adobe, Inc.

ADBE

$ADBE

21864

12284

2072187

eBay, Inc.

EBAY

$EBAY

32765

16254

2595924

Wayfair, Inc.

11087

4822

374973

Data Transformation by Standardization Using a sample of the US internet Dot-com companies, we examine whether these companies’ posts are reflected in its stock liquidity for the period observed. And whether structural break affects this relationship. The downloaded data for these companies varies in sizes. Therefore, there is the need to provide equal weight for these variables in our experiments, hence the need for data standardization. Many notable authors, such as Bijl et al. (2016), Nisar and Yeung (2018), Kim et al. (2019), etc., have applied standardization transformation in their research. This is also referred to as the z-score. The formula in equation 1, shows the x and mean dataset at the numerator and standard deviation at the denominator of the equation. Defined as: (1)

Where: SVt indicates the standardized values, Xt indicates the raw datasets and deviation of the dataset

represents the standard

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Companies’ Stock Liquidity Measure Liquidity is seen as an elusive concept. This is because it encompasses many observed transaction properties on the markets. Therefore, there is an array of liquidity measures employed by empiricists and policy makers, etc., that take into account different liquidity aspects. However, in our research paper, we employed Amihud’s (2002) illiquidity ratio, and then obtained the liquidity measure as the inverse of the ratio. Hur and Chung (2018) also applied similar methodology in their study. Thus stated (2) Where: lt denotes the number of daily observations of stock i in day t. Rt denotes the market returns, Vt denotes trading volume, and Pt shows the stock price.

Tests for Single Structural Change with Unknown Change Point We employ Andrews-Ploberger (1994) and Andrews-Quandt structural break model to examine whether regime changes have broken down the stability of the relationship between companies’ tweets and its stock market liquidity. The Quandt (1960) and Andrew (1993), now popularly known as QuandtAndrews breakpoint test tests a specific sample for an unknown structural breakpoint. The Chow’s single breakpoint test which performs at every observation between two break points, or observations, and is the idea behind the Quandt-Andrews test. The k test statistics from the Chow tests are put together into a single test statistic against the null hypothesis of no breakpoints between and . With the assumption that m or is unknown, Quandt employs the LR statistic against This is the maximal

. statistic over range of break dates m0 , ...,m1: (3)

Where: = trimming parameters, i=0,1 QLR is also the Andrews’ sup-F statistic The break data m and break fraction respectively.

are estimated using

Since we have no knowledge of the break data, we set our trimming parameter

and .

Whilst Andrews-Quandt test uses as the maximum of the LM statistics, Andrews-Ploberger uses the average of the Chow breakpoint statistics

to compute the QLR statistic.

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Therefore (4)

(5) k = Number of regressors being tested The critical P-values given in Andrews and Ploberger can be computed using Hansen (1997) techniques or approximations. The distribution of these statistics degenerates at the beginning or when it approaches the end of the equation. To compensate for this behavior, we trim or exclude the first and the last observations by 15%. That is, the same numbers of observations are removed from the beginning and the end of the estimation sample.

RESULTS OF EMPIRICAL ANALYSIS We study the impact of twitter on the US internet Dot-com companies in the presence of unknown structural breaks. A sample of twelve (12) top companies of US internet Dot-com companies were used for our analysis for the period from September 4th, 2019 to April 1st, 2020. As mentioned earlier in section 3.3, we employ the methodological framework of Andrews-Ploberger (1994) and AndrewsQuandt structural break model to analyze whether structural change influences the relationship between company’s tweets and its stock liquidity. Figure 1 in Appendix A illustrates the plots of natural log stock liquidity for the twelve (12) US internet Dot-com companies. The graph shows the various structural breaks for each company. Although the COVID-19 started late, i.e., September - December (Li et al. (2020), Yongchen et al. (2020), Lone and Ahmad (2020)), Table 4 and Figure 1 show that the pandemic caused a major structural break for the top US internet Dot-com companies at different break points. However, most of these breaks occur between November 2019 to March 2020. In our analysis, we applied five (5) variables – stock liquidity, stock returns, volumes, firm size, tweets and likes. The stock liquidity as an internal variable, tweets and likes as external variables, and the stock returns, volumes, and size as the control variables. Apart from revealing the major structural break date, our study also examines the sub-period before these breaks, precisely September 4th, 2019. Table 5 shows the empirical results of the companies’ tweets and likes on their stock liquidity before the major structural break. Our results reveal that most US internet top companies’ tweets and ‘likes’ do not to influence their stock liquidity rate for the period examined. Specifically, only Adobe, Expedia and PayPal Holding companies’ tweets appear to influence their stock liquidity. These influences are very weak, negatively significant at 10% levels and do not boost firm values during pre-structural break periods, see Table 5. This means that these few companies appear to have small numbers of orders to buy and sell on the stock market. It is argued that increase in stock liquidity, increases firm values. This is due to the fact that assets are discounted at a lower cost of capital when there is an improvement in stock liquidity. However, almost all companies’ stock volume is shown to impact their stock liquidity. With five (5) of these companies actually have their stock volume increase companies’ values. 24

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

The values of R squared, and R adjusted are low, but they are reported in HeteroscedasticityConsistent (Eicker-White) Standard Errors. Hence, the results estimates are consistent. Our analysis does not confirm the study hypothesis of H1 and H2 that companies’ tweets and likes increase the companies’ stock liquidity. Some companies’ stock liquidity has also shown to react to firm sizes in different behavior (increase or decrease). Table 4. Andrews Ploberger (1994) and Andrews - Quandt Structural Break Tests Firms

Andrews-Quandt

Resid. Analysis

Andrews-Plober.

Resid. Analysis

Structural Dates

Test

P-Val

Test

P-Val

Test

P-Val

Test

P-Val

Break Date

Resid. Date

ADOBE

74.013

0.000

1.006

0.998

33.170

0.000

0.069

1.000

2019:12:19

2019:12:20

AMZN

67.152

0.000

0,757

1.000

30.043

0.000

-0.237

1.000

2020:02:26

2020:02:27

BKNG

61.763

0.000

0.850

1.000

28.481

0.000

-0.078

1.000

2020:01:16

CRM

34.321

0.222

0.804

1.000

14.005

0.197

-0.216

1.000

2020:12:10

2019:12:10

EBAY

22.059

0.933

0.381

1.000

9.0822

0.848

-0.132

1.000

2019:10:25

2020:01:14

EXPE

50.589

0.004

1.695

0.861

22.326

0.002

0.400

0.529

2019:11:04

2019:11:15

28.147

0.583

1.743

0.850

10.402

0.664

0.566

0.392

2019:11:14

2020:01:22

GOOGL

40.916

0.053

1.878

0.816

17.411

0.037

0.469

0.466

2019:10:29

2019:10:30

NFLX

39.866

0.068

1.407

0.929

17.114

0.043

0.320

0.618

2019:11:19

PYPL

27.284

0.643

0.770

1.000

10.518

0.646

0.309

1.000

2020:01:20

2020:01:17

TSLA

110.002

0.000

0.495

1.000

50.395

0.000

0.549

1.000

2020:03:03

100.131

0.000

2.348

0.700

45.555

0.000

0.589

0.377

2019:10:15

Note: The Andrews-Quandt test uses as its test statistic, the maximum of the LM statistics, while Andrewe-Ploberger uses the geometric mean. These both have highly non-standard distributions. P-values are computed in Hansen (1997) approximations.

However, empirical analysis in Table 6 describes the results using estimation by Least Squares with Heteroscedasticity-Consistent (Eicker-White) Standard Errors, the impact of tweets and likes on stock liquidity after post-structural break. The structural break points are at different dates with different sample sizes. Companies’ tweets and likes are shown to have mix results on companies’ stock liquidity. Preponderance of companies’ tweets and likes also appears to influence their stock liquidity post-structural break. Nevertheless, this influence tends to be negative, which indicates that most of the companies examined appear to have small numbers of orders to buy and sell on the underlying market. Conversely, tweets from PayPal Holdings, Inc; Booking Holdings; eBay, Inc. and Wayfair, Inc. appear to have created large numbers of orders to buy and sell in the underlying market during poststructural break. This large numbers of order to buy and sell increase the probability that the highest prices a buyer is willing to pay and the lowest price a seller is happy to accept converge or move closer together. However, our empirical analysis in Table 6 also shows that Tesla, Salesforce.com, Expedia, and Adobe tweets do not influence the stock liquidity, as the estimation coefficients are not statistically significant. Besides, preponderance of investors validation of companies’ tweets with the use of likes also shown to be related to the stock liquidity, i.e., the Tesla, Inc. and Netflix, Inc. 25

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

With the mix results of the influence of ‘likes’ and tweets on companies, our stated research hypotheses of H1 and H2, that tweets and likes from US internet Dot-com companies increase their stock liquidity, are partly fulfilled. In addition, our estimation results are consistent with the standard errors. Table 7 shows the estimation of a relationship between companies’ stock liquidity and some of the activities, such as tweets and likes, performed at their official twitter accounts. We examine these relations using the full sample size of 151 of our experiment without taking into consideration the possible impact of the structural break. The empirical result reveals that neither the companies tweet nor the ‘likes’ impact on stock liquidity. However, companies’ stock volume shows to influence stock liquidity. This impact does not increase firm values as shown by majority of companies examined. Our results estimates are in robust standard errors.

Robustness Check According to Andrew and Ploberger (1994) when checking the robustness of our estimated test results, Andrew-Quandt breakpoint test is used to estimate the breakpoint of companies’ stock liquidity. The P-values of these results as well as the residual analysis are shown in Table 4 and Figure 2. It is indicated that there are significant break points for TSLA, W, EXPE GOOGL, NFLX, ADOBE, AMZN, and BKNG US internet Dot-com companies. However, our analysis reports that the break points in CRM, EBAY, FB and PYPL companies show no significant impact on stock liquidity. Table 4 also shows the residual break points analysis of the stock liquidity with respect to the twitter data. The reported residual structural break dates show to be exact or very close to dates of Andrews-Ploberger (1994) and Andrews-Quandt estimated structural break dates. The Andrews-Quandt and Andrews-Ploberger residual tests and its p-values confirmed the residual analysis to be statistically not different from zero. This analysis points out that our results on the changes of stock liquidity are robust to the break point selection. Nguyen et al. (2017), Chuen and Gregoriou (2014), etc., have also employed this approach.

CONCLUSION This paper examines the relationship between Twitter posts and stock liquidity for the US internet Dot-com companies in the presence of structural breaks. Our study provides an in-depth analysis of this relations for twelve (12) different US stock index, over a period of seven (7) months. Based on the structural break models of Andrews-Ploberger (1994) and Andrews-Quandt which allow for the presence of breaks in a linear model, we have obtained mix results which reveal that nearly all structural breaks recorded are significant. Out of twelve (12) companies’ stock liquidity examined, only four (4) companies have stock liquidity structural changes that are non-significant. For investment strategy, this implies that companies’ managers and investors have to be company-specific to ascertain whether or not to ignore structural change. In addition to reporting and analyzing the structural change dates for the US internet Dot-com companies, it is crucial to understand the relations between Twitter and companies’ stock liquidity so as to figure out whether the extent to which assets are bought and sold quickly at stable prices has any relationship with Twitter posts and likes. Our empirical analysis reveals results for two sub-samples analysis, as well as for full sample experiment. 26

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

We find that neither Twitter’s posts nor “like” appear to influence investors’ ability to exchange an asset for cash. Our results also provide the necessary insight for further research on this study, particularly the use of sentiment analysis. That is, whether sentiments from US internet Dot-com companies’ Tweets can explain the dynamics of the stock liquidity considering the structural changes. This will be an interesting future direction.

ACKNOWLEDGMENTS My sincere thanks go to the editor and the anonymous reviewers for their very helpful comments and insightful suggestions. For any other error(s), the usual disclaimer applies. Table 5. Estimation by Least Squares with Heteroscedasticity Consistent (Eicker White) Standard Errors for Pre-structural Break Dates Endogenous variable: Stock Liquidity Firm Size

-0.0441*

0.0230

0.201

-2.0668*

-0.1680*

0.0693*

-0.182*

-5.4532

0.0114**

0.0842

-0.029

-0.5139

-0.0794

-0.0030

0.1554

PayPal Holdings, Inc.

-0.0670

-0.048*

Salesforce.com, Inc.

-0.088***

Booking Holdings

Company

Constant

Tweets

Amazon.com, Inc.

-0.0316

0.0772

Alphabet Inc, Class A

-0.136***

Facebook, Inc.

Returns

Volume

0.0554

-0.6187

-0.0110

0.0042

-0.0860

-0.1315

Tesla, Inc.

-0.0576

Netflix, Inc.

Adj.R2

Break Date

Obs.

0.200

2019:9:4-2020:02:26

124

0.216

0.164

2019:9:4-2019:10:29

-0.013

0.140

0.120

2019:9:4-2019:11:14

-0.0816*

0.0978

0.082

0.0803

2019:9:4-2020:03:03

128

0.9023*

-0.0196*

0.000*

0.188

0.175

2019:9:4-2019:11:19

-0.072

0.0495**

-0.0904

0.115

0.203

0.211

2019:9:4-2020:01:20

-0.0308

-0.0028

0.2794

0.2101**

-0.266*

0.334

02049

2019:9:4-2019:12:10

-0.107***

0.0927

-0.0239

-2.1898

0.0175*

-0.1044

0.183

0.169

2019:9:4-2020:01:16

Expedia, Inc.

4.1493*

0.0094*

0.0329

1.2211

12.958*

-3.3456*

0.427

0.399

2019:9:4-2019:11:04

Adobe, Inc.

-0.084***

-0.0257*

-0.0102

-0.484

-0.0337

-0.0075

0.255

0.224

2019:9:4-2019:12:19

eBay, Inc.

0.3879

-0.0276

-0.1109

15.056

1.4297*

-0.563**

0.266

0.105

2019:9:4-2019:10:25

Wayfair, Inc.

-1.2075

-0.200

0.0432

1.2546

-6.8080

3.5096

0.035

0.025

2019:9:4-2019:10:15

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Table 6. Estimation by Least Squares with Heteroscedasticity-Consistent (Eicker White) Standard Errors for Post - structural Break Dates Endogenous variable: Stock Liquidity Constant

Tweets

Returns

Volume

Amazon.com, Inc.

- 0.6304*

-0.3150*

0.0317*

0.6436*

0.0722*

0.4765

0.1200

Alphabet Inc, Class A

0.0363**

-0.1968**

0.1021*

-0.7795*

-0.2202

0.0879*

Facebook, Inc.

-0.0120

-0.0841*

0.0628*

-0.8274*

-0.2103

Tesla, Inc.

4.0087

-0.4854

-0.0035

-2.8326

Netflix, Inc.

-0.0779*

-0.0132*

-0.0794

PayPal Holdings, Inc.

-0.0715

0.1967*

Salesforce.com, Inc.

0.0968**

Booking Holdings

Company

Firm Size

Adj.R2

Break Date

Obs.

0.111

2020:02:26-2020:04;01

0.0769

0.0604

2019:10:29-2020:04:01

111

0.0936

0.366

0.2420

2019:11:14-2020:04:01

12.574*

-13.435

0.5157

0.2463

2020:03:03-2020:04:01

-0.7586

-0.0865

0.000***

0.0903

0.0912

2019:11:19-2020:04:01

0.172**

0.60130

-0.3210

0.4506*

0.4639

0.3687

2020:01:20-2020:04:01

-0.0430

0.0815*

0.0895*

-0.3788*

0.2653

0.2161

0.1681

2019:12:10-2020:04:01

0.4992*

1.7922*

0.0335

0.0907**

0.5959*

-0.7008

0.6268

0.5501

2020:01:16-2020:04:01

Expedia, Inc.

0.0304**

-0.0932

0.1615*

0.7648

-0.0898

0.0355*

0.2031

0.1973

2019:11:04-2020:04:01

107

Adobe, Inc.

0.0904

-0.1131

0.363*

-0.9992

-.04396

0.4894

0.3721

0.2510

2019:12:19-2020:04:01

eBay, Inc.

0.1711*

0.0197*

0.0643

2.9348

-0.0584

-0.2147

0.1708

0.1302

2019:10:25-2020:04:01

Wayfair, Inc.

-0.0627

0.0075**

-0.013*

0.1511

-.0.0895

0.0392

0.1453

0.1017

2019:10:15-2020:04:01

121

Notes: 1. The symbols ***, ** and * indicates significance at the 1%, 5%, and 10% levels, respectively. 2. White heteroscedasticity - consistent (Eicker White) standard errors and covariances are applied to the liner models. Table 7. Least Squares Estimation with Heteroscedasticity-Consistent (Eicker White) Standard Errors for Full Sample (4th September 2019 to 1st April 2019) Endogenous variable: Stock Liquidity Company

Constant

Tweets

Returns

Volume

Firm Size

Adj.R2

Obs.

Amazon.com, Inc.

0.0015**

0.0226

0.0624

-0.9016

0.2097*

-0.1142

0.1176

0.1124

151

Alphabet Inc, Class A

0.0008

-0.0484

0.0779*

-0.8069

-0.2426

0.1352

0.2491

0.2113

151

Facebook, Inc.

0.0040**

-0.0435

-0.0353

-0.4163*

0.0116**

-0.1846

0.2382

0.1020

151

Tesla, Inc.

0.0030

-0.0103

-0.0523

-0.5360

0.0337

0.1269*

0.2771

0.2130

151

Netflix, Inc.

0.0029*

0.0988

-0.0231

-0.6447

-0.0793*

0.0000**

0.3079

0.2161

151

PayPal Holdings, Inc.

0.0001

-0.0361

0.0529

-0.1514

-0.2618

0.3412

0.4360

0.2997

151

Salesforce.com, Inc.

0.0002**

-0.0342

0.0401

-0.0255

-0.1027**

0.0306

0.1561

0.1027

151

Booking Holdings

0.0013

0.3237

0.0679

0.3706*

-0.1204*

0.1237

0.2146

0.1011

151

Expedia, Inc.

-0.0022

0.0309

0.1100

0.6680

-0.1606

0.1576

0.2646

0.1253

151

Adobe, Inc.

-0.0011

-0.0512

0.1581*

-0.6394*

-0.1737*

0.2003

0.3299

0.2183

151

eBay, Inc.

0.0047*

0.0651

0.0425

3.2092

-0.2267

0.1472

0.1924

0.1543

151

Wayfair, Inc.

0.0022**

-0.0284

0.0270

0.1256

-0.1160

0.0318

0.0314

0.0100

151

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

REFERENCE Andrews, D., W. K., & Ploberger W., (1994). Optimal Tests When a Nuisance Parameter is Present Only Under the Alternative, Econometrica, 62 (6), 1383-1414. Affuso, E., & Lahtinen, K.D. (2019). Social media sentiment and market behavior, Empirical Economics, 57(1), 105-127. Amihud, Y. (2002), Illiquidity and stock returns: cross-section and time-series effects, Journal of Financial Markets, 5 (1), 31-56. Amihud, Y., Mendelson, H. & Pedersen, L. H. (2006), Liquidity and asset prices. Boston: Publishers Inc. Albarrak, M.S., Elnahass, M., Papagiannidis, S., & Salama, A. (2020). The effect of twitter dissemination on cost of equity: A big data approach, International Journal of Information Management, 50, 1-16. Ajjoub, C., Walker, T., & Zhao, Y. (2020). Social media posts and stock returns: The Trump factor, International Journal of Managerial Finance, 17(2), 185-213. https://doi.org/10.1108/IJMF-02-2020-0068 Anguyo, F. L., Gupta, R., & Kotzé, K. (2020). Inflation dynamics in Uganda: a quantile regression approach, Macroeconomics and Finance in Emerging Market Economies, 13(2), 161-187. Amiraslany, A., Luitel, H.S., & Mahar, G.J. (2019). Structural Breaks, Biased Estimations, and Forecast Errors in a GDP Series of Canada versus the United States, International Advances in Economic Research, 25(2), 235-244. Andrews, D.W.K. (1993). Tests for Parameter Instability and Structural Change with Unknown Change Point, Econometrica, 59, 817-858. Benson, K., Faff, R. & Smith, T. (2015). Injecting liquidity into liquidity research, Pacific-Basin Finance Journal, 35, 533-540. Broadstock, D.C., & Zhang, D. (2019). Social-media and intraday stock returns: The pricing power of sentiment, Finance Research Letters, 30, 116-123. Behrendt, S., & Schmidt, A. (2018). The Twitter myth revisited: Intraday investor sentiment, Twitter activity and individual-level stock return volatility, Journal of Banking and Finance, 96, 355-367. Brans, H., & Scholtens, B. (2020). Under his thumb the effect of President Donald Trump’s Twitter messages on the US stock market, PLoS ONE, 15(3), e0229931. Chuen, D. L. K., & Gregoriou, G.N. (2014). Handbook of Asian Finance: Financial Markets and Sovereign Wealth Funds. New York, USA: Academic Press. Cabellon, E.T., & Ahlquist, J. (2016). Engaging the Digital Generation: New Directions for Student Services, 155. New York, USA: Jossey-Bass. Chahine, S., & Malhotra, N.K. (2018). Impact of social media strategies on stock price: the case of Twitter, European Journal of Marketing, 52(7-8),1526-1549. Clemente, J., Gadea, M.D., Montañés, A., & Reyes, M. (2017). Structural breaks, inflation and interest rates: Evidence from the G7 countries, Econometrics, 5(1),11. https://doi.org/10.3390/econometrics5010011 Dugast, J., & Foucault, T. (2016) Data Abundance and Asset Price Informativeness, Working paper, Luxembourg: Luxembourg School of Finance and HEC Paris. Datar, V. T., Naik, N. Y. & Radcliffe, R. (1998). Liquidity and Stock Returns: An Alternative Test, Journal of Financial Markets, 1, 203-219. Fan, R., Talavera, O., & Tran, V. (2020). Social media, political uncertainty, and stock markets, Review of Quantitative Finance and Accounting, 55(3), 1137-1153. Florackis, C., Gregoriou, A. & Kostakis, A. (2011). Trading frequency and asset pricing: evidence from a new price impact ratio, Journal of Banking and Finance, 35(12), 3335-3350. Guijarro, F., Moya-Clemente, I., & Saleemi, J. (2019). Liquidity Risk and Investors' Mood: Linking the Financial Market Liquidity to Sentiment Analysis through Twitter in the S&P500 Index, Sustainability, 11(24),7048. 29

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Groß-Klußmann, A., König, S., & Ebner, M. (2019). Buzzwords build momentum: Global financial Twitter sentiment and the aggregate stock market, Expert Systems with Application, 136, 171-186. Gabrielsen, A., Marzo, M., & Zagaglia, P., (2011). Measuring market liquidity: an introductory survey, Munich Personal RePEc Archive. Ge, Q., Kurov, A., & Wolfe, M.H. (2019). Do Investors Care About Presidential Company-Specific Tweets? Journal of Financial Research, 42(2), 213-242. Gil-Alana, L.A., Dadgar, Y., & Nazari, R. (2019). Iranian inflation: peristence and structural breaks, Journal of Economics and Finance, 43(2), 398-408. Gil-Alana, L.A., & Mudida, R. (2017). CPI and inflation in Kenya. Structural breaks, non-linearities and dependence, International Economics, 150, 72-79. Gil-Alana, L.A., Dos Santos Figueiredo, O.H., & Wanke, P. (2019). Structural breaks in Brazilian tourism revenues: Unveiling the impact of exchange rates and sports mega-events, Tourism Management, 74, 207-211. Hansen, B. E. (1997). Approximate Asymptotic P-Values for Structural Change Tests, Journal of Business and Economic Statistics, 15, 60-67. Holden, C. W., Jacobsen, S. E., & Subrahmanyam, A. (2014). The empirical analysis of liquidity. Kelley School of Business Research Paper, 2014-09. Hegerty, S.W. (2020). Structural breaks and regional inflation convergence for five new Euro members, Economic Change and Restructuring, 53(2), 219-239. Hur, S.-K., & Chung, C.Y. (2018). A novel measure of liquidity premium: application to the Korean stock market, Applied Economics Letters, 25(3), 211-215. Kumar, N., Kumar, R.R., Patel, A., & Stauvermann, P.J. (2019). Exploring the effect of tourism and economic growth in Fiji: Accounting for capital, labor, and structural breaks, Tourism Analysis, 24(2), 115-130. Kisling, Whitney, Eric Lam, & Nina Mehta. (2013). Human Beats Machine This Time as Fake Report Roils Stocks. Accessed 24.05.2020. Klaus, J., & Koser, C. (2020). Measuring Trump: The Volfefe Index and its impact on European financial markets, Finance Research Letters, 101447. Lone, S.A., & Ahmad, A. (2020). COVID-19 pandemic - an African perspective, Emerging microbes & infections, 9(1), 1300-1308. Li, L., Li, R., Wu, Z., Liu, J., & Chen, D. (2020). Therapeutic strategies for critically ill patients with COVID-19, Annals of Intensive Care, 10(1), 45. https://doi.org/10.1186/s13613-020-00661-z Makrehchi, M., Shah, S., & Liao, W. (2013). Stock prediction using event-based sentiment analysis, In Web Intelligence, 1, 6690034, (pp. 337-342). New York, USA: IEEE. Min, J. C. H., Kung, H.-H., & Chang, T. (2019). Testing the structural break of Taiwan inbound tourism markets, Romanian Journal of Economic Forecasting, 22(2), 117-130. Nasir, M.A., & Vo, X.V. (2020). A quarter century of inflation targeting & structural change in exchange rate pass-through: Evidence from the first three movers, Structural Change and Econ. Dynamics, 54, 42-61. Nisar, T. M., & Yeung, M., (2018). Twitter as a tool for forecasting stock market movements: A short- window event study, The Journal of Finance and Data Science, 4, 101e119. Nath, H.K., & Sarkar, J. (2019). Inflation and relative price variability: new evidence from survey-based measures of inflation expectations in Australia, Empirical Economics, 56(6), 2001-2024. Nguyen, A.D.M., Dridi, J., Unsal, F.D., & Williams, O.H. (2017). On the drivers of inflation in Sub-Saharan Africa, International Economics, 151, 71-84. Nisar, T.M., & Yeung, M. (2018). Twitter as a tool for forecasting stock market movements: A short-window event study, The Journal of Finance and Data Science, 4, 101e119. Orlowski, L.T. (2017). Sensitivity of Interest Rates to Inflation and Exchange Rate in Poland: Implications for Direct Inflation Targeting, Comparative Economic Studies, 59(4), 545-560. 30

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

O’Hara, M. (2004). Liquidity and Financial Market Stability, National Bank of Belgium, Working paper 55. Quandt, R.E. (1960). Tests of Hypotheses that a Linear System Obeys Two Separate Regimes, Journal of the American Statistical Association, 55, 324-330. Pöppe, T., Kolaric, S., Kiesel, F., & Schiereck, D. (2020). Information or noise: How twitter facilitates stock market information aggregation, In T. Chairs (Ed.) Crowds, Social Media And Digital Collaborations. Munich, Germany: ICIS. Reboredo, J.C., & Ugolini, A. (2018). The impact of Twitter sentiment on renewable energy stocks, Energy Economics, 76, 153-169. Reed, M. (2016). A Study of Social Network Effects on the Stock Market, Journal of Behavioral Finance, 17(4), 342-351. Ruch, F., Balcilar, M., Gupta, R., & Modise, M.P. (2020. Forecasting core inflation: the case of South Africa, Applied Economics, 52(28), 3004-3022. Sakhare, N.N., Imambi, S.S., Kagad, S., Malekar, H., & Dalal, M. (2020). Stock market prediction using sentiment analysis, International Journal of Advanced Science and Technology, 29(4), 1126-1133. Saurabh, S., & Dey, K. (2020). Unraveling the relationship between social moods and the stock market: Evidence from the United Kingdom, Journal of Behavioral and Experimental Finance, 26,100300. Shiva, A., & Singh, M. (2019). Stock hunting or blue-chip investments? Investors’ preferences for stocks in virtual geographies of social networks, Qualitative Research in Financial Markets, 12(1), 1-23. Sul, H.K., Dennis, A.R., Yuan, L. (2014). Trading on twitter: The financial information content of emotion in social media, In System Sciences, 6758703, (pp. 806-815). Hawaii, USA. Stauvermann, P.J., Kumar, R.R., Shahzad, S.J.H., Kumar, N.N. (2018). Effect of tourism on economic growth of Sri Lanka: accounting for capital per worker, exchange rate and structural breaks, Economic Change and Restructuring, 51(1), 49-68. Sani, Z., Salisu, A., Onyia, E., Anih, O., Kanu, L. (2020). Modeling exchange rate-interest rate differential nexus in BRICS: The role asymmetry and structural breaks, Economics and Business Letters, 9(2), 73-83. Twitter. Company: About; (2020). https://about.twitter.com/company. Accessed May 9, 2020. Urlam, S.P.S., Mandal, S., Poornima, S. (2020). Stock prediction using twitter sentiment analysis, International Journal of Psychosocial Rehabilitation, 24(8), 1031-1035. Vanstone, B.J., Gepp, A., Harris, G. (2019). Do news and sentiment play a role in stock price prediction? Applied Intelligence, 49(11), 3815-3820. Vayanos, D. and Wang, J. (2012). Market liquidity -- Theory and empirical evidence, National Bureau of Economic Research, w18251. Wu, C.-F., Wang, C.-M., Chang, T., Yuan, C.-C. (2019). The nexus of electricity and economic growth in major economies: The United States-India-China triangle, Energy, 188, 116006. Wu, D. (2019), Does Social Media Get Your Attention? Journal of Behavioral Finance. 20(2), 213-226, Yongchen, Z., Shen, H., Wang, X., Chen, Y., Gu, B. (2020). Different longitudinal patterns of nucleic acid and serology testing results based on disease severity of COVID-19 patients, Emerging microbes & infections, 9(1), 833-836. Zhang, X., Fuehres, H., Gloor, P.A. (2012). Predicting asset value through twitter buzz, Advances in Intelligent and Soft Computing, 113, 23-34. Zhang, X., Shi, J., Wang, D., Fang, B. (2018). Exploiting investors social network for stock prediction in China's market, Journal of Computational Science, 28, 294-303.

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Table 3. Financial Market Variables Summary Statistics (Full Sample) Table 3: Financial Market Variables Summary Statistics (Full Sample) Panel A Company

Mean

LIQUIDITY P25 P50 P75

Amazon.com, Inc.

-0.000

-0.277

-0.210

0.138

1.000

0.000

-0.830

-0.369

1.353

1.000

151

0.000

-0.241

-0.207

0.198

1.000

-0.000

-0.875

-0.370

1.638

1.000

151

Facebook, Inc.

-0.000

-0.496

-0.355

0.941

1.000

-0.000

-0.872

-0.341

1.584

1.000

151

Tesla, Inc.

-0.000

-0.219

-0.176

0.146

1.000

0.000

-0.968

-0.207

1.318

1.000

151

Netflix, Inc.

-0.000

-0.314

-0.263

0.534

1.000

-0.000

-0.804

-0.247

0.946

1.000

151

PayPal Holdings, Inc.

0.000

-0.425

-0.305

0.571

1.000

0.000

-0.988

-0.244

1.447

1.000

151

Salesforce.com, Inc.

-0.000

-0.214

-0.179

0.051

1.000

-0.000

-0.867

-0.326

1.596

1.000

151

Booking Holdings Inc

0.000

-0.182

-0.154

-0.043

1.000

0.000

-0.900

-0.343

1.550

1.000

151

-0.000

-0.423

-0.303

0.625

1.000

0.000

-0.664

-0.305

0.897

1.000

151

0.000

-0.200

-0.172

0.129

1.000

0.000

-0.957

-0.304

1.764

1.000

151

eBay, Inc.

0.000

-0.541

-0.340

0.798

1.000

0.000

-0.832

-0.417

1.405

1.000

151

Wayfair, Inc.

-0.000

-0.264

-0.184

0.268

1.000

-0.000

-0.821

-0.347

1.420

1.000

151

Alphabet Inc, Class A

Expedia, Inc. Adobe, Inc.

Panel B Company Amazon.com, Inc.

-0.000

-0.068

Alphabet Inc, Class A

-0.000 0.000

Tesla, Inc. Netflix, Inc. PayPal Holdings, Inc.

Trading Volume P25 P50 P75

Market Returns P25 P50 P75

Mean

1.442

1.000

0.000

-0.018

0.0001

0.021

0.019

151

-1.118

0.224

1.607

1.000

-0.000

-0.023

0.000

0.019

0.024

151

-1.093

-0.198

1.581

1.000

-0.000

-0.029

0.001

0.023

0.026

151

-0.000

-1.384

0.066

1.334

1.000

0.005

-0.035

0.004

0.048

0.053

151

-0.000

-0.804

-0.247

0.046

1.000

0.002

-0.030

0.000

0.035

0.028

151

-0.000

-1.234

-0.063

1.426

1.000

-0.001

-0.027

-0.000

0.023

0.086

151

Salesforce.com, Inc.

0.000

-1.031

-0.149

1.525

1.000

-0.000

-0.026

0.001

0.020

0.028

151

Booking Holdings Inc

-0.000

-1.223

-0.153

1.525

1.000

-0.002

-0.027

0.000

0.019

0.054

151

Expedia, Inc.

-0.000

-1.247

0.154

1.341

1.000

-0.006

-0.027

-0.000

0.019

0.047

151

Adobe, Inc.

-0.000

-1.215

-0.118

1.621

1.000

0.000

-0.023

0.001

0.022

0.030

151

0.000

-1.030 -1.186

-0.282 -0.181

1.440 1.520

1.000 1.000

-0.002 -0.005

-0.032 -1.060

-1.000 -0.001

0.020 0.042

0.022 0.068

151 151

eBay, Inc. Wayfair, Inc.

-0.000

Author’s Calculation;

Authors Calculation: Where; SD, P25, P50, and P75 are the standard deviation, 25th, 50th, 75th percentiles respectively. Where; SD, P25,P50, and P75 are the standard deviation 25th, 50th, 75th percentiles respectively.

Obs.

-0.231

Facebook, Inc.

Mean

Firm Size P25 P50 P75

Mean

Obs.

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

APPENDIX A Figure 1. US’ Internet Dot.com Companies Liquidity Series and Structural Break Date for the Period 4th September 2019 to 1st April, 2020

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

Figure 2. Autocorrelation Analysis on a Series of Residuals for the US Internet Dot-com Companies

EJAE 2021  18 (2)  15 - 35

OSABUOHIEN. O.I.  STRUCTURAL BREAKS, TWITTER AND THE STOCK LIQUIDITY OF INTERNET DOT-COM COMPANY: EVIDENCE FROM US COMPANIES

STRUKTURNI PREKIDI, TWITER I LIKVIDNOST AKCIJA INTERNET DOT-COM KOMPANIJE: DOKAZI AMERIČKIH KOMPANIJA Rezime: Cilj ovog rada je istražiti odnos između Twitter-a i likvidnosti akcija nekih velikih američkih internet Dot-com kompanija u prisustvu nepoznatih strukturnih prekida za period od septembra 2019. do aprila 2020. Koristeći Andrews-Ploberger i Andrews-Quandt strukture modela prekida, identifikujemo glavne strukturne tačke prekida u likvidnosti akcija i otkrivamo da se većina ovih strukturnih promena značajno percipira. Kada smo pregledali podperiode, kao i ceo uzorak, otkriveno je da tvitovi i lajkovi većine kompanija nemaju veze sa likvidnošću akcija. Ovi rezultati pružaju ključan uvid u portfeljsku strategiju kako međunarodnim tako i domaćim investitorima.

Ključne reči: strukturni prekidi, Twitter, akcionarsko društvo, Andrews-Ploberger, likvidnost, režim.

EJAE 2021, 18(2): 36 - 48 ISSN 2406-2588 UDK: 336.76(4-11) 005.31:[336.76:343.352 DOI: 10.5937/EJAE18-31151 Original paper/Originalni naučni rad

CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT Dušan Dobromirov* University of Novi Sad, Faculty of Technical Sciences, Novi Sad, Serbia

Abstract:

Article info:

This paper analyzes the impact of corruption level on East European financial markets development. Financial market liquidity for 15 national markets is presented and market volume per capita is used as an indicator for market maturity. Market volume per capita values are compared to corruption perception index values, using classical logic method. Findings of the research are quite interesting and unexpected, as they show modest impact of corruption on financial market development. Results suggest that further research of corruption should be done, in order to develop better, quantitative corruption indicator.

Received: March 4, 2021 Correction: August 30, 2021 Accepted: September 7, 2021

Keywords: corruption, financial markets, market liquidity.

INTRODUCTION The phenomenon of corruption has been present since ancient times and has significantly affected the degree of order in a particular society. The development of human communities brought systems that were based on a series of rules and social norms. Compliance with regulations and rules was imposed and violations of them were sanctioned. Penalties for breaking the rules were often very severe and violators expected material incentives to take the risk of being punished. Material incentives were easiest to achieve in violation of the rules in the field of economy, so corruption most often appeared in this important area. Modern economies are based on a whole series of complicated regulations and laws that are regulated and enforced by state institutions, so the impact of corruption is more significant. Corruption always exists to some extent and there is no state without this phenomenon. Specialized international organizations such as Transparency International deal with the registration and analysis of corruption in individual countries. Their fight against corruption aims to point out the problems that arise in countries affected by high levels of corruption. 36

*E-mail: ddobromirov@gmail.com

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

There is also a tendency to make recommendations based on which to reduce the level of corruption. We can understand the importance of the fight against corruption if we accept the fact that an orderly and well-organized state provides equal economic opportunities to all its citizens. The ideal of modern societies is to provide the same opportunities for all citizens in all spheres of life and especially in the economic domain. Knežević and Dobromirov (2016) show that different factors specific to a particular market have a significantly greater impact on the economy than macroeconomic indicators. Sovbetov and Kaplan (2019) show that tranquility of economic environment is significantly important for the development. A good way to look at the development and democracy of a society is to observe the orderliness of the financial market. A well-developed and active financial market indicates a country where there is fair market competition in which there are many participants. K. Mishra (2018) argues that globalization and democracy are mutually supportive and that together they hinder corruption, support human development and social progress. Investors operate freely in conditions where there is equality between them and when the success of a financial transaction depends solely on proper risk assessment and market skill. In such circumstances, the market is controlled by strong and functional institutions immune to corruption. On the other hand, in conditions of high levels of corruption, success in the financial market is achieved through unpunished violations of the rules. In such a situation, it is important to have special personal contacts, privileged information and the possibility of non-compliance with the rules of trade, which is all achieved through corrupt practices. Such financial markets do not benefit most investors and they avoid such markets, which is manifested in lower turnover and poor development of such a market. Therefore, it is possible to state that the liquidity of the financial market will serve well as an indicator of the level of corruption of an economy. The definition of corruption is described by various terms such as: “abuse of public authority” and “moral setback” and strict legal definitions of corruption as an act of bribery of a public official and exchange of material resources. The phenomenon of corruption indicates the problem of political, economic, cultural and moral underdevelopment (Lučić et al., 2016). Corruption is a phenomenon where there is a deviation from the legal rules that govern the actions of a public servant for personal gain such as money, power or social status. Corruption is also described as an exchange between private and public sector agents where there is an illegitimate conversion of collective goods into payments in the private interest. Another definition of corruption is that it is a process in which the influence of public office is used for private interests in a non-regulatory manner (Jain, 2008), or that government officials can perform corruption for personal gain. Corruption, which could generally be defined as the abuse of public authority to pursue a personal interest, is a complex and ambiguous concept. Although there is a whole range of definitions of corruption, the common view is that corruption negatively affects society. Corruption can be a significant problem in the process of economic development and modernization of public administration, as it affects the weakening of institutions on which economic growth largely depends. Also, corruption is a specific additional unofficial tax on business transactions. Concerning financial markets as a part of an economic environment, it is important for government regulatory principles to provide orderly market conditions through prudential regulation Brezigar-Masten et al. (2011). The aim of this paper is to analyze the impact of corruption level on East European financial markets and their liquidity. The idea is to prove that higher corruption level in a country has a bad influence on financial market institutions and infrastructure, leading to lower financial market turnover. The paper is organized in following manner: the first section is literature review followed by data and methodology definition. Results are presented and discussed in result and discussion part and at the end a short conclusion summarizes the research. 37

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

LITERATURE REVIEW Numerous researchers are involved in understanding the economic phenomenon of corruption and the impact of corruption on the most important macroeconomic indicators. There is an intense public debate about the interdependence between the level of corruption and the speed of economic development. Some earlier studies suggest that corruption can help the most efficient firms bypass bureaucratic hurdles and rigid laws and that this has a positive effect on economic development. On the other hand, some recent works do not find a significant negative relationship between corruption and economic development. High levels of corruption can put a country in a bad position where high levels of corruption accompanied by low growth block economic development, while some other countries may develop by reducing corruption and increasing economic growth (Murshed and Mredula,2018) . Some authors (Popovaa and Podolyakinaa, 2014) argue that corruption is a disorder that causes low economic growth and that the causality is such that corruption affects GDP. They define corruption as a disease that affects the poor and that the disease disappears when countries develop, so the causality is described from the level of GDP to corruption. Bota-Avram et al. (2018) present evidence of interdependence present in both directions. The basic theoretical arguments indicate that there is a connection between the level of economic development and the level of corruption. Assiotis and Sylwester (2014) have shown that corruption adversely affects economic growth only in states that have well-developed independent institutions. At the same time, economic development reduces corruption. On the other hand, corruption does not affect economic growth in countries where independent institutions are weak. Schneider and Buehn (2018) showed that the correlation between levels of corruption and economic development is negative and reciprocal. Igwike and Hussain (2012) only partially succeeded in showing evidence in favor of a two-way relationship. It depends on economic growth towards a lower level of corruption and on increasing levels of corruption towards a fall in GDP per capita. Although it is justified and logical to claim that a decline in economic activity could cause an increase in the level of corruption or that a higher level of corruption could lead to a lower growth rate, the main direction of interdependence between these two variables has not been proven. Some research shows that there is no correlation between economic development and levels of corruption in certain cases. Zaman and Goschin (2015), Stojanovic et al. (2016) examined the degree of political freedoms of a society as a key factor influencing the dependence between corruption and long-term economic growth. Analyzing data from the period from 1960 to 2000, they failed to establish a correlation between the level of corruption and economic growth in countries with a low level of political freedoms. Luminita (2011) found that high corruption has a positive effect on economic growth when political and economic freedoms are limited, but that the positive impact of corruption diminishes when political and economic freedoms increase. Ionescu (2011) show that in countries with weak independent institutions, the level of corruption has no impact on economic growth. Gallego-Alvarez et al. (2014) concluded that to the extent that we can measure corruption in the environments of different countries, it does not affect growth. Littvay and Donica (2006) examined the period from 1986 to 2003 and did not find a link between corruption and economic growth in non-Asian countries, but found a positive correlation in the case of Asian countries. Paiders (2008) analyzed the values and changes in GDP per capita for the period from 1998 to 2005 and the values and changes in the Corruption Perceptions Index (CPI) for the period from 1998 to 2007. The conclusion of such research is that the mutual connection between changes in CPI and GDP per capita cannot be noticed, and that values fluctuate independently of each other when looking at data from countries in the world and European countries. 38

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

The dominant view in the scientific literature is that a change in the level of corruption affects the change in the level of GDP and that there is an interdependence between these two quantities. The prevailing view in most published papers in this area is that the correlation between the change in the level of corruption and the change in the level of GDP is negative. The World Bank and the IMF are of the view that corruption has significant negative effects on economic growth. High levels of corruption negatively affect economic development by undermining the rule of law and weakening independent institutions on whose successful operation economic growth depends. Similarly, the IMF states, "Many causes of corruption are economic in nature, and so are its consequences - poor governance is clearly detrimental to economic activity and well-being." Empirical studies of the impact of corruption on economic growth are gaining in importance due to the social impact of these factors. Corruption adversely affects economic growth by reducing the level of private investment and thus changing the structure of government spending, especially by reducing part of the education budget. Groşanu et al. (2015) showed that countries with high levels of corruption continuously tend to record poor economic results. Corruption further deepens the inequality gap between rich and poor sections of society. In his research, Aidt (2011) finds that corruption has a very specific impact on growth, as it has the greatest negative effect in countries with quality institutions, but little impact in countries with weak institutions. Aidt confirms the negative correlation between growth and corruption. It defines four channels through which corruption negatively affects economic growth: increased values of public investment, lower tax revenues, lower operating costs and poor quality of public infrastructure facilities. The impact of a 1% increase in corruption reduces the growth rate of economic activity by about 0.72%, with the most important channel for the impact of corruption being political instability, accounting for about 53% of the total effect. Some research shows that reducing the level of corruption by one index point affects GDP growth by 0.5 percentage points. Ionescu (2012) reveals significant negative effects of corruption on economic growth due to poor institutions and shows how corruption can have a negative impact on foreign direct investment and net capital inflows, which are important components of economic growth. An analysis of the relationship between the level of economic development measured by GDP and the estimated level of corruption among countries shows a strong interdependence: poor countries tend to be corrupt. Examining the relationship between the recorded level of corruption and the rate of economic growth among countries, we can see that in countries with high levels of corruption there are significantly different growth rates. In other words, most highly corrupt countries had a low rate of economic growth but, there are countries that have achieved rapid economic growth with a very high level of corruption. This shows that certain countries can achieve a high level of economic growth despite a high level of corruption. Huang (2012) explores 10 Asian countries (China, Indonesia, Japan, South Korea, Malaysia, the Philippines, Singapore, Taiwan, Thailand, and Vietnam) from 1995 to 2010 and analyzes the impact of corruption on economic growth. The result is a positive impact of corruption, indicating that corruption boosts economic growth in large East Asian newly industrialized economies. The explanation for this phenomenon, which is specific to the Asian continent, is that a strong centralized government can limit the negative effects of bribery compared to a decentralized corrupt bureaucracy. Corruption makes business processes more efficient and accelerates them, thus boosting economic growth, relaxing the rigid bureaucratic regulations imposed by governments.

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

Absalyamova et al. (2016) argue that in countries with a higher level of social trust, corruption is socially acceptable and thus less detrimental to economic growth. Swaleheen (2011) in his research tested the relationship between corruption and economic growth and the results showed that corruption does not reduce growth at all levels and that there can be a significant increase in economic growth even with a high degree of corruption. Proponents of "efficient corruption" argue that bribery helps companies to be more efficient and to contract jobs faster in an economy characterized by bad and rigid laws. According to this model of increasing efficiency, companies, with the help of "quick money", bypass bad laws and inefficient institutions. There are two views regarding the usefulness or harmfulness of corruption. Studies claiming that corruption is detrimental to economic growth draw attention to the bad implications of corruption on efficiency, especially in the long run. Other studies explain how corruption lubricates business and trade outlets and thus stimulates economic growth and investment. The prevailing view of these two opposing views is the view that corruption is detrimental to economic growth. The usual view in the academic literature is that corruption hinders and hinders economic growth and development, weakens institutions and has a negative impact on society as a whole. A review of the literature indicates that two observations can be made. First, there is a correlation between corruption and GDP levels. Second, empirical evidence on the link between corruption and economic development shows that negative dependence is present in the vast majority of cases. So far, a limited number of papers have been published that analyze the temporal interdependence between the change in the level of corruption and the change in the level of GDP. Sahakyan and Stiegert (2012) show that the effect of corruption on economic growth is negative and statistically significant only in the medium and long term, while it is insignificant in the short run. Thus, the actions of economic policy makers focus more on the medium and long term, considering the consequences of corruption, than on the short-term effects. Ruzek (2015) finds that the causality of GDP and corruption is long-term, as the country becomes richer and thus the motivation for corruption decreases. Based on their research, this long-term interaction works only in one direction. Borlea et al. (2017) show a negative correlation between the level of corruption and the long-term growth rate. Looking at two periods: 1990–2005 and 1980–2005, they find that less corrupt countries achieve more significant economic growth, while countries with high levels of corruption achieve negative growth rates. Hoinaru et al. (2020) showed that there is a strong long-term relationship between the level of corruption and GDP, analyzing the changes in values in the period 1984–2008. All estimates of the effects of long-term business show that corruption has a direct negative impact on GDP per capita. Using comprehensive data on 47 countries from 1996–2007, Chen N. (2010) shows that, when reference variables are controlled, corporate liquidity is lower in countries with more efficient securities laws or greater control of corruption. In addition, cash can increase the value of the company. This positive relation is more pronounced in countries with efficient securities laws or with low levels of corruption. Moreover, excess cash can reduce a company’s value in countries with inefficient securities laws or low corruption controls. This impairment effect of companies can be mitigated or reversed when corruption control is improved or when securities laws are improved. Although most economists argue that corruption can act as a good solution in the short term for market disruptions that can cause wrong government procedures and policies, in the medium and long term corruption reduces market efficiency and in that period negative effects are dominant. 40

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

Gründler and Potrafke (2019) determine a decline in real GDP per capita of 17% when the CPI increases by one standard deviation as a long-term effect of corruption. This fact, which arises from the results of previous research, indicates a significant impact of the level of corruption on the development of the financial market, measured by the liquidity of national stock exchanges.

DATA AND METHODOLOGY The data set used in the research represents two different indicators: national financial markets volumes for year 2019 from 16 East European countries and 2019 Corruption Perception Index for corresponding countries. Radišić and Dobromirov (2017) describe Central and Eastern European (CEE) emerging markets as active, due to their continuing aspiration to position themselves on the global portfolio investment map. Eastern European markets volumes are represented by annual trading volume from representative, national stock exchanges. Data collection was done by visiting official web sites and taking into account annual trading report. However, certain markets were not taken into consideration, such as Albania, Latvia, Estonia, Lithuania and Russian Federation. Due to non-existing trading volume in 2019, Albania could not be included into data analysis. Latvia, Estonia and Lithuania passed through market consolidation by joining NASDAQ Baltic, established form US stock exchange giant NASDAQ. NASDAQ Baltic as an exchange operates in Sweden, Finland, Latvia, Lithuania and Estonia. Such an international market organization would not be appropriate for planed methodology. Russian Federation is omitted from the research due to conflicting market volume data from several different sources. After collecting official market data, population size was taken into consideration for each of observed countries. Financial market volume per capita is calculated as an indicator V, with an idea to present each financial market relative to size of a country. Corruption Perception Index (CPI) for 2019 is obtained from official Transparency International web site and presented as a country corruption indicator. Transparency International in their Corruption Perception Index (CPI) methodology assigns a higher value of CPI index to a country with lower corruption level, representing indicator C. CPI index values are rounded and in some cases two countries have the same index value. In these circumstances, unrounded values were taken into consideration in order to create corruption ranking (Table 2). The process of proving and testing the hypothesis is established on classical logic and the principle of exclusive disjunction. We will compare CPI rank (C) and market volume (V) for each pair of countries and check if the following is true (T) or false (F): (1) where Cm and Vm are CPI rank and volume per capita rank for one country respectively, and Cn and Vn are CPI rank and volume per capita rank for another country. After testing every presented country’s indicator values with each other (105) pairs, we will get certain number of true and false samples. Following specific cases of condition testing, we denote by R a function as: (2)

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

We introduce the following scheme: Strong match, if R ≥ 2.5, Mean match, if 1.5 ≤ R <2.5, Poor match, if 0 ≤ R <1.5. Result interpretation: if a calculated value of R is going to be over 2.5 (strong match) that will indicate that initial hypothesis is confirmed. With R valued between 1.5 and 2.5 (mean match) we will have a conditional hypothesis approval and with R valued under 1.5 we should prove that hypothesis is not valid.

RESULTS AND DISCUSSION The financial market volume data related to population of corresponding country are presented in Table 1 as follows: Table 1. Population, market volumes, volumes per capita and volume ranking Country

Population

Financial market volume per capita

FM volume per capita ranking

(in mil US$)

Poland

37,840,465

50,394.38

1,331.77

Hungary

9,652,821

8,668.02

898.05

Czech Republic

10,722,241

5,569.71

519.46

Bosnia and Hertzegovina

3,269,170

559.89

171.27

Romania

19,523,621

2,044.51

104.92

Croatia

4,105,493

328.51

80.02

Slovenia

2,070,050

149.94

72.43

Montenegro

622,359

43.51

69.91

Slovakia

5,462,617

264.68

48.45

Moldova

4,031,141

190.22

47,18

Northern Macedonia

2,084,162

92.89

44.57

Serbia

6,963,764

240.60

34.55

Bulgaria

7,000,039

170.42

24.35

Belarus

9,452,123

14.97

1.58

Ukraine

43,642,532

1.32

0.03

Source: author’s calculation

2019 financial market volume

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

Financial market volumes per capita in US$ were sorted in descending order. The highest financial market volume per capita is ranked 1, the second is ranked 2 and so on. These ranks represent V indicator value that will be used in further analysis. The CP index score for 2019 is presented in Table 2: Table 2. Corruption ranking Country

CP index score 2019

CP index Ranking C

Slovenia

Poland

Czech Republic

Slovakia

Croatia

Belarus

45a

Montenegro

45b

Hungary

44a

Romania

44b

Bulgaria

Serbia

Bosnia and Hertzegovina

Northern Macedonia

Moldova

Ukraine

Source: author’s calculation

The highest CP index score for 2019 is ranked 1 the second is ranked 2 and so on. These ranks represent C indicator value that will be used in further analysis. Applying the presented methodology (formula 1) to the data shown in Table 1 and Table 2, 105 tests were performed. The test results of all pairs, which were obtained by crossing on a one-to-one basis, are shown in Table 3:

Poland

Source: author’s calculation

Ukraine

Moldova

N Maced

Bosnia&H

Serbia

Bulgaria

Romania

Hungary

Monten

Belarus

Croatia

Slovakia

Czech R.

Poland

Slovenia

Czech R.

Table 3. Condition testing results

Slovakia

Croatia

Belarus

Monten

Hungary

Romania

Bulgaria

Serbia

Bosnia&H

Moldova

N Maced

Ukraine

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

The number of correct logical statements is 70 and the number of incorrect logical statements is 35, so for R we get the value of 2. This value of R represents mean match as defined in the initial research setting. This means that the initial hypothesis has been conditionally confirmed and it cannot be determined that there is a medium dependence of the degree of corruption and the value of turnover on the financial market. This is a somewhat surprising and unexpected result. Analyzing in detail the results shown in Table 3, it must be noted that the tests in which one of the pairs was Belarus differ significantly from the other results. This is not negligible because by omitting the Belarus from the research sample, we would get different results and unconditionally confirm the hypothesis. A logical explanation would be that the Belarus was not adequately assessed in compiling the annual report on corruption in 2019. Another two countries that show irregular and unexpected results are Hungary and Bosnia and Herzegovina. Hungary has well developed and organized financial market but, the result is odd due to lower CPI index score in recent years. On the other hand, Bosnia and Herzegovina did have certain big financial market transactions in connection with privatization process. These one-time transactions influenced the 2019 financial market volume and influenced the results of research. However, thorough result analysis raises a question of CPI index methodology. There is an impression that there is a need for more reliable and better corruption level index.

CONCLUSIONS This research analyzed the impact of corruption level on 15 East European financial markets development. Although a strong correlation of low corruption and well developed financial markets was expected, results show medium impact of corruption on financial market performance. Although initial hypothesis was logical and proof was expected, it was not unconditionally proven. As the methodology in this paper was based on classical logic, it was a good opportunity to record some illogical issues aroused concerning the corruption level assessment. These issues were discussed and it is explained how they influenced the research result. Moreover, it was shown that CPI as an indicator has certain shortcomings and heuristic nature. The major finding is that a better corruption indicator should be found and used in scientific research. Further research should try to discover a model for quantitative based corruption indicator.

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

REFERENCES Absalyamova S., Absalyamov T., Khusnullova A. and Mukhametgaliev C. (2016). The impact of corruption on the sustainable development of human capital, Journal of Physics: Conference Series, 738, 012009. doi:10.1088/1742-6596/738/1/012009 Aidt, T. S. (2011). Corruption and sustainable development, International handbook on the economics of corruption, 2, 3-51. Assiotis, A. , Sylwester, K. (2014). Do the effects of corruption upon growth differ between democracies and autocracies? Review of Development Economics, 18(3), 581-594. Borlea, S.; Achim, M.V.; Miron, M. (2017). Corruption, Shadow Economy and Economic Growth: An An Empirical Survey Across the European Union Countries. Studia Universitatis „Vasile Goldis” Arad – Economics Series, 27, 19–32. Bota-Avram, C.; Grosanu, A.; Răchisan, P.R.; Gavriletea, M.D. (2018). The Bidirectional Causality between Country-Level Governance, Economic Growth and Sustainable Development: A Cross-Country Data Analysis. Sustainability, 10, 502. Brezigar-Masten, A., Coricelli,F., Masten, L. (2011). Financial Integration and Financial Development in Transition Economies. Economic and Business Review, 13(1/2), 119-137. Chen N. (2010). Securities Laws, Control of Corruption, and Corporate Liquidity: International Evidence, Corporate Governance, 19(1), 3-24. https://doi.org/10.1111/j.1467-8683.2010.00823.x Gallego-Alvarez, I., Vicente-Galindo, M., Galindo-Villardón, M., Rodríguez-Rosa, M. (2014). Environmental Performance in Countries Worldwide: Determinant Factors and Multivariate Analysis. Sustainability, 6, 7807–7832. Groşanu A., Boţa-Avram, C., Răchişan, P.R., Vesselinov, R. and Tiron-Tudor, A., (2015). The Influence of Country-Level Governance on Business Environment and Entrepreneurship: a Global Perspective. Amfiteatru Economic,17(38), 60-75. Gründler, K.; Potrafke, N. (2019): Corruption and Economic Growth: New Empirical Evidence, CESifo Working Paper, No. 7816. Munich, Germany: Center for Economic Studies and ifo Institute (CESifo). Hoinaru R., Buda D., Borlea S.N., Văidean L.V. and Achim M.V. (2020). The Impact of Corruption and Shadow Economy on the Economic and Sustainable Development. Do They “Sand the Wheels” or “Grease the Wheels”?, Sustainability, 12, 481. https://doi.org/10.3390/su12020481 Huang, C. J. (2012). Corruption, economic growth, and income inequality: evidence from ten countries in Asia. World Academy of Science, Engineering and Technology, 66, 2012. Ionescu, L. (2011). The Influence of Corruption on Economic Growth. Economics, Management and Financial Markets, 6(1), 453-458. Ionescu, L. (2012). Corruption, Unemployment, and the Global Financial Crisis, Economics, Management, and Financial Markets, 7(3), 127-132. Igwike, R. S., Hussain, M. E., & Noman, A. (2012)..The impact of corruption on economic development: A panel data analysis.SSRN Working Paper Series. Mishra, S. K. (2018). A Simultaneous Equation Model of Globalization, Corruption, Democracy, Human Development and Social Progress. The European Journal of Applied Economics, 15(1), 46-82. https: doi:10.5937/EJAE15-16414 Knežević A., & Dobromirov D. (2016). The determinants of Serbian banking Industry profitability, Economic research, 29(1), 459-474. Littvay, L., & Donica, A. N. P. (2006). Corruption: A Cause of Effect? Canadian Political Science Association Conference, Toronto, Ontario, Canada. Lučić D., Radišić M. & Dobromirov D. (2016). Causality between corruption and the level of GDP, Economic Research-Ekonomska Istraživanja, 29(1), 360-379. doi:10.1080/1331677X.2016.1169701 46

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

Luminita, I. (2011). The Influence of Corruption on Economic Growth, Economics, Management, and Financial Markets, 6(1), 453-458. Murshed, M.,& Mredula, F.A. (2018). Impacts of Corruption on Sustainable Development: A Simultaneous Equations Model Estimation Approach, Journal of Accounting, Finance and Economics, 8, 109–133. Paiders, J. (2008). Relationship between corruption level changes and economic growth in the world end Europe. International scientific conference: Research for Rural Development 2008, 14, Jelgava (Latvia), 21-23., May 2008. Popovaa Y., Podolyakinaa N. (2014). Pervasive impact of corruption on social system and economic growth, Procedia - Social and Behavioral Sciences, 110, 727 – 737. Radišić M. & Dobromirov D. (2017). Statistical analysis of the regional stock market indices price returns, Transformations in Business & Economics, 16(3), 175-186. Ruzek, W. (2015).The Informal Economy as a Catalyst for Sustainability. Sustainability, 7, 23–34. Sahakyan, N., & Stiegert, K. W. (2012). Corruption and firm performance. Eastern European Economics, 50(6), 5–27. Schneider, F., & Buehn, A. (2018). Shadow Economy: Estimation Methods, Problems, Results and Open Questions, Open Economics, 1(1), 1–29. Sovbetov, Y., & Kaplan, M. (2019). Causes of Failure of the Phillips Curve: Does Tranquility of Economic Environment Matter?. The European Journal of Applied Economics, 16(2), 139-154. doi:10.5937/EJAE16-21569 Stojanovic, I.. Ateljevic, J.. Stevic, R. S. (2016), Good Governance as a Tool of Sustainable Development. European Journal of Sustainable Development, 5(4), 558-573. https://doi.org/10.14207/ejsd.2016.v5n4p558 Swaleheen, M. (2011). Economic growth with endogenous corruption: an empirical study. Public Choice, 146, 23–41. Zaman, G.. & Goschin, Z. (2015). Shadow economy and economic growth in Romania. Cons and pros, Procedia Economics and Finance, 22, 80–87.

EJAE 2021  18 (2)  36 - 48

DOBROMIROV. D.  CORRUPTION IMPACT ON EAST EUROPEAN EMERGING MARKETS DEVELOPMENT

UTICAJ KORUPCIJE NA RAZVOJ ISTOČNIH EVROPSKIH TRŽIŠTA

Rezime: U ovom radu se analizira uticaj nivoa korupcije na razvoj istočnoevropskih finansijskih tržišta. Prikazana je likvidnost finansijskog tržišta za 15 nacionalnih tržišta, a tržišni volumen po stanovniku se koristi kao pokazatelj zrelosti tržišta. Vrednosti tržišnog obima po glavi stanovnika upoređuju se sa vrednostima indeksa percepcije korupcije, koristeći klasičnu logičku metodu. Nalazi istraživanja su prilično zanimljivi i neočekivani, jer pokazuju skroman uticaj korupcije na razvoj finansijskog tržišta. Rezultati ukazuju na to da je potrebno dodatno istraživanje korupcije kako bi se razvio bolji, kvantitativni pokazatelj korupcije.

Ključne reči: korupcija, finansijska tržišta, likvidnost tržišta.

EJAE 2021, 18(2): 49 - 61 ISSN 2406-2588 UDK: 338.46:004.42]:34 004.42:79 341.231.14-053.6(497.11) DOI: 10.5937/EJAE18-29481 Original paper/Originalni naučni rad

DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN? Nataša Krstić* Singidunum University, Faculty of Media and Communications, Belgrade, Serbia

Abstract: For the central entertainment industry of the 21st century - online gaming, children are undoubtedly the key consumer group. Although research on the impact of the gaming industry on children mainly deals with adverse effects such as addiction, violent content, inappropriate conduct and monetisation of personal data, there are also many positive effects – family fun, virtual socialising, improving cognitive skills and using games as a teaching tool. Therefore, all participants' task in the gaming industry value chain is to maximise the positive and minimise the negative impacts on children. A survey conducted among 893 young gamers in Serbia exposed their habits in consuming online games and indicated whether their rights are protected during the gameplay. The conclusion provides recommendations for key stakeholders in the gaming industry's ecosystem on making the digital playground inclusive, safe, and responsible for respecting children's rights.

Article info: Received: November 19, 2020 Correction: January 8, 2021 Accepted: March 3, 2021

Keywords: child rights, gaming, gaming industry, Covid-19, Serbia.

INTRODUCTION The available data show that in 2019 games earned $120 billion globally – i.e., $64.4 billion on mobile devices, $29.6 billion on desktop computers (PC) and $15.4 billion on consoles (Superdata, 2020), thus making the gaming industry exceed other entertainment sectors by far, generating three times box office revenues ($43 billion) and twice music industry revenues ($57 billion) (Hall, 2020). The industry's exponential growth is grounded on the massive expansion of information and communications technology (ICT). More precisely, the perfection of gaming hardware and broad accessibility of bandwidth and mobile internet accelerated a shift towards the delivery of online games via mobile and cloud-based platforms (ibid). In that respect, online gaming is defined as “playing any type of single or multiplayer commercial digital game via any internet-connected device” (UNICEF, 2019, p. 5).

*E-mail: krstic.natasa@gmail.com

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

The proliferation of platforms and borderless access to various online games have also changed the way how gamers, or “individuals who play video games regularly” (Granic, Lobel and Engels, 2014, p. 68), play, buy and interact. Consequently, the business model of the gaming industry has evolved. Consumers today buy fewer games than before but spend more time in gaming (Hall, 2020), which is why “finding a balance between providing engaging experiences and generating revenue is key for online gaming companies” (UNICEF, 2020, p. 8). As a result, the industry has become entirely focused on increasing user engagement and monetising the game and the gaming experience through expansions, new features and tools (Pappas, Mikalef, Giannakos, and Kourouthanassis, 2019). The today’s gaming industry ecosystem is home to many stakeholders – designers, publishers, console and application (app) store distributors, streaming services, e-sports organisers and teams, sponsors and professional gamers with an abundance of followers. All of them impact the design and the way of game consumption, where children, as “a key consumer group for the gaming industry” (UNICEF, 2019, p. 5) should receive special consideration. Due to exponential growth and easier access to digital technology, children start to play online games at an ever-younger age, resulting in childhood development (Grimes, 2015; Lissak, 2018). Namely, increasing exposure to games and other gamification experiences has strongly influenced the generation of children and adolescents growing up as “digital natives” (Prensky, 2012). When it comes to online games' behaviour (Arı, Yılmaz and Elmastas Dikec, 2020), children tend to spend more time on free gaming sites than on subscription or pay-to-play games and play online games frequently but for short durations (Grimes, 2015). Online gameplay is second only to social media as the most common digital venue for adolescents to meet new friends (Lenhard, 2015). In terms of gender differences, boys spend substantially more time gaming than girls (Lemmens, Valkenburg & Peter, 2011), and show more severe online game addiction (Pawłowska, Potembska & Szymańska, 2018). Boys also play games with friends to a greater extent (Hastings, Karas, Winsler, Way, Madigan & Tyler, 2009) and are more likely to play age-inappropriate games (Coyne, Padilla-Walker, Stockdale and Day, 2011). When it comes to girls, those with lower levels of life satisfaction, coming mostly from urban areas, are more likely to engage in video gaming to a greater extent (Brooks, Chester, Smeeton & Spencer, 2016, p. 49; Pawłowska et. al, 2018). Consequently, previous academic research findings led to the structuring of the first research question (RQ1) that this article should answer: How do children and youth in Serbia consume online games? The direct and indirect impact of gaming on children can be both positive and negative. Social and collaborative play in online games has been identified as key to learning (Ventura, Shute & Zhao, 2013; Schaaf & Mohan, 2014) and fun (Kaye & Bryce, 2012), boosting creativity (Jackson, Witt, Games, Fitzgerald, von Eye & Zhao, 2012), improving cognitive abilities (Przybylski, 2014), strategic thinking and digital skills (Hygen, Belsky, Stenseng, Skalicka, Kvande, Zahl-Thanem & Wichstrom, 2020), all of which lead to a sense of well-being of children (Amerijckx & Humblet, 2014). However, as with many other online activities involving children, gameplay can have some negative aspects. In addition to concerns about age-inappropriate content (Charmaraman, Richer & Moreno, 2020), online games can lead to cyberbullying (Huang, Yang & Hsieh, 2019), hate speech in multiplayer games (Geraniols, 2010), foster gender and race stereotypes (Lynch, Tompkins, van Driel & Fritz, 2016; Malkowski & Russworm, 2017), depression and anxiety (Mannikko, Billieux & Kaariainen, 2015) and screen addiction (Lissak, 2018). Moreover, through the collection and monetisation of personal data of young gamers, “advergames, specifically created to incorporate and promote advertisers’ products into games and immersive online environments” (UNICEF, 2019, p. 25), can strongly influence children's consumer habits (Folkvord & van ‘t Riet, 2018; Agante & Pascoal, 2019). 50

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

In some cases, the monetisation of games has inspired governments and industry associations to introduce new regulations aimed at protecting the rights of underage players from unwanted content and classified ads, such as the Pan-European Game Information (PEGI), the Entertainment Software Rating Board (ESRB) and International Age Rating Coalition (IARC), which all deal with age and content ratings of online games. Consequently, companies that produce, sell and promote online games “have a great responsibility to shape their platforms in ways that maximise positive and minimise negative impacts on children” (UNICEF, 2020, p. 3), which was especially evident during the Coronavirus Disease 2019 (COVID-19) outbreak (López-Cabarcos, Ribeiro-Soriano & Piñeiro-Chousa, 2020). Lockdown and curfews around the globe have forced millions of people to spend more time at home, where the online entertainment industry, driven by online games and esports, has sprung into action to capture their attention (Hall, 2020). The analysis from the leading website about the global video games industry - GamesIndustry.biz, showed that sales across 50 key markets during the virus outbreak rose by 63% (Dring, 2020). Prolonged periods of social isolation that divert time to technology-based activities pose a risk of perpetuating unhealthy lifestyle patterns, which can lead to difficulties in re-adaptation once the COVID-19 crisis has passed (King, Delfabbro, Billieux and Potenza, 2020), where children are particularly vulnerable (World Health Organisation, 2020). Consequently, during the virus outbreak, the United Nations Children’s Fund brought to the fore the issues of “healthy game time, ensuring inclusion, avoiding toxic environments, considerations around age-limits and verification, combatting grooming and sexual abuse, and managing commercial influence” (2020, p. 3) for all the participants in the gaming ecosystem. With this in mind, the second research question (RQ2) to which the article should respond is formulated as follows: Are the rights of young gamers in Serbia protected during the online gameplay? Accordingly, the paper aims to gain insight into the gaming habits and attitudes of young Serbian gamers, so that based on the survey results, recommendations for the key stakeholders can be suggested. In this regard, the objectives of the research are (a) to examine the habits of young gamers in Serbia in terms of the age at which they started gaming, the frequency of play, motives and social interactions; and (b) scrutinise their exposure to unwanted content, advert-games and negative gaming experiences. Finally, the article is structured as follows. The introductory part outlines the exponential growth and advancement of the gaming industry and its value chain, with a focus on the core target group - children from the earliest age. The positive and negative impact of gaming on children based on previous academic research is highlighted and the recommendations for the gaming industry regarding the opportunities for protection and promotion of children's rights on the digital playground given by United Nations Children Fund (UNICEF) during the COVID-19 outbreak. The section on methodology explains the applied methods of the conducted quantitative research involving children and youth. The obtained survey results are divided into two subsections based on the stipulated research questions; the first part deals with young gamers' habits in Serbia, while the second part deals with the protection of their rights during gameplay. In conclusion, answers are given to the research questions, which provide further recommendations for the gaming industry ecosystem on improving children’s rights in the digital playground.

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

METHODOLOGY The opinion pool entitled “How do online games affect children and youth in Serbia?” was active during August 2020 for one month, via an electronic survey on the U-Report platform [1]. The U-Report represents youth's voices and promotes their participation in creating positive social changes on topics that concern them. The involvement of young U-reporters on the platform is voluntary and anonymous. The data processing was performed by UNICEF, which preserved the concept of ethics in research involving children (Kiili & Moilanen, 2019). In the context of the conducted study, we defined a child, based on the Convention on the Rights of the Child, as every human being below the age of eighteen (United Nations, 1989), and youth according to the National Strategy for Youth as persons from 15 to 30 years of age (Serbian Ministry of Youth and Sports, 2015). The survey consisted of 16 questions with predefined single-choice answers, two of which related to the sample. Of the 7,300 registered U-Reporters, 893 decided to take part in the survey, finding themselves close to the topic of gaming. The survey was slightly dominated by girls (52%) and mostly youth aged 15-19 (71%), reflecting the structure of registered U-Reporters. The rest of the survey questions focused on understanding the gaming patterns of youth in Serbia - the age at which they started gaming, the frequency of play, motives and social interactions, and their potential exposure to inappropriate content, unwanted ads and negative gaming experiences. In some findings, the gender difference in the responses between girls and boys was additionally emphasised. Finally, the young survey respondents made recommendations to the gaming industry to make its digital playgrounds relevant, useful and safe for young gamers. For additional insights, the quantitative data collected through the survey with children and youth have been uploaded on the Mendeley repository [2].

GAMING HABITS OF THE SERBIAN YOUTH The first question in the survey was about the age when youth started to play games. In this regard, half of the respondents pointed out that they had started playing online games before the age of ten (50%), and a quarter of them between the ages of 10-12 (26%). In other words, Generation Z from the survey (15-19 years) mainly played games before the age of ten, while the surveyed Millennials (26-30 years old) started a couple of years later - between 10 and 12 years, which indicates a gradual but inevitable shift in the age limit. When it comes to the frequency of playing, one-third of the young survey respondents confirmed playing games every day or several times a week (34%), which proves that gaming has become an integral part of children and youth's lives in Serbia. Interestingly, almost a third of the surveyed girls play games only when they have free time (30%), while over a fourth of boys do so daily (27%). Despite the global trends of accelerated gaming during the lockdown, the survey results revealed that COVID-19 has not significantly changed Serbian youth's playing habits, as less than half of the respondents confirmed that they play online games the same amount of time as before (42%). However, that this finding should not be particularly encouraging is indicated by the fact that more than a third of the young respondents (36%) have been spending more time on games in recent months compared to the time before the virus outbreak, which requires special attention from the industry and parents. 52

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

Among the main reasons for which children and youth play online games, one-quarter of the respondents highlighted fun (29%), followed by boredom and relaxation (12% each), all belonging to emotional motives for online gameplay (Olson, 2020). Bearing in mind that virtual socialising was chosen as a gaming motive by only 8% of young respondents and that more than half gameplay alone (57%), online gaming in Serbia among children and youth cannot be characterised as a social activity. However, it is encouraging to have found that less than a third of young survey participants gameplay online with friends (29%) or in their presence (14%), and that the vast majority have never replaced seeing friends with gaming (64%), in which respect the girls lead. Table 1. Gaming habits of the Serbian youth (U-Report survey, n=893) Gaming habits

Behavioural patterns

The age at which youth start playing games

Pushing the boundaries of starting gaming at an early age: Generation Z started gaming two years earlier than the Millennials

Frequency of gameplay

One-third of the youth play games daily or several times a week

Gender differences and gaming frequency

Girls play games in their free time; boys daily

COVID-19 and online gameplay

So far, the youth play online games at the same time as before

Main reasons for online gameplay

Emotional motives: fun, boredom, relaxation

Individual or social gaming

Over half of the youth play games alone

Gaming versus seeing friends

Most of the youth do not need gameplay at the expense of seeing friends, especially girls

RESPONSIBLE GAMEPLAY Developers and game creators have vast opportunities to establish rules for responsible gaming that protect children and youth's rights on the digital playground. One of the most common control mechanisms for suppressing screen addiction is limiting the game's duration and encouraging players to take breaks. Coming back to our research, almost half of the young survey participants (42%) confirm the possibility of independent decision-making on the duration of the games they play. Still, nearly a quarter of them point out that the games they play have no time limits (23%). Other types of restrictions that are less frequently offered to young gamers in Serbia relate to limiting the number of sessions or levels (13%), time per session or level (9%), encouraging players to take a break (4%), while only 2% of the young gamers stated the existence of an age limit, the setting of which is required for games sold through app stores and consoles. Exposure to uncontrolled monetisation in games consumed by Serbian children and youth is indicated by the finding that only a third of the respondents do not leave personal data during registration (34%). Namely, one-fifth of the young players who participated in the research leave several pieces of personal data during gameplay (19%), where e-mail (23%) and social media accounts (10%) prevail. 53

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

From the gender perspective, boys tend to leave more personal data to the game distributors, as every third male gamer does so (30%) instead of girls who are more responsible in personal data revealing during gaming (50%). Based on the previous findings, it does not surprise that most of the surveyed young gamers confirmed that they notice advertisements while playing, such as commercials and sponsorships (59%). In this regard, of particular concern is the finding that more than a quarter of them (28%) confirmed seeing unwanted advertisements that could not be disabled, thus endangering young gamers' privacy. Furthermore, it is positive that half of the surveyed young gamers did not encounter inappropriate content (50%). Though over a third of the surveyed youth did notice some type of disturbing or potentially harmful content that upset them, with violence prevailing (25%), and gender stereotypes and sexualised content appearing sporadically (4% each). Moreover, most young, surveyed gamers believe that the games they consume are equally adequate for both genders in terms of their content (69%). Still, almost a third of them consider games they play as primarily intended for boys (30%). A positive finding is that most young respondents feel safe on the digital playground and state that they have not had any unpleasant experiences (63%) in improper contact and conduct. It does happen sporadically, though, as every sixth young gamer noticed other players' inappropriate behaviour in multiplayer games (17%). A small number of them confirmed that they had insomnia as a result of gaming (7%), experienced bullying in an online chat (6%) or witnessed the mockery of other players (5%). Consequently, game developers and creators can safeguard children and youth's rights in the digital playground through options such as clearly defined and child-friendly terms of service, privacy policies, community standards, and codes of conduct (UNICEF, 2019). In practice, conditions and instructions that encourage responsible gameplay are often barely noticeable or written in legal language that children and youth do not understand. Therefore, it is not surprising that the young survey respondents generally do not notice conditions for players (42%). Just over a third of them pay attention only to certain conditions (35%), while less than a quarter pay attention to all the requirements listed for players (23%), where girls aged 20-25 dominate. Finally, Serbia's young gamers believe that the industry could be more responsible for children and protect their rights if a mandatory age limit was set in all games for minors (27%), along with a mechanism to reduce the risk of screen addiction (21%). Other recommendations relate to the game design, so that they are more in line with the interest of young gamers (17%), enhanced policies for the protection of personal data (17%) and the removal of inappropriate content (13%).

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

Table 2. Situation regarding responsible gameplay of young Serbian gamers (U-Report survey, n=893) Mechanisms for responsible gaming

Research findings

Time or age limits

Half of the surveyed gamers independently decide on the duration of the game. Nearly a quarter confirm that games they play have no time limit while having age limits is rare.

Personal data exposure

Only a third of the respondents, predominantly girls, do not leave personal data during registration. Among the provided personal data, e-mail, and social media accounts prevail.

In-games monetisation

Most young gamers spot advergames, and more than a quarter of them cannot disable unwanted ads.

Inappropriate content

Half of the young gamers did not encounter inappropriate content, but more than a third of them noticed potentially harmful content that disturbs them, where violence prevails. Close to a third of the respondents believe that the games are primarily intended for boys.

Inappropriate conduct

Most of the young gamers feel safe on the digital playground. However, every sixth gamer noticed inappropriate behaviour in multiplayer games, while some of them had insomnia or experienced bullying in a chat.

Protection of children’s rights

Young gamers mainly do not notice conditions for players, predominantly boys.

Recommendations to the gaming industry

The setting of the age limit and the mechanism for reducing the risk of screen addiction prevails.

CONCLUSIONS The gaming industry is one of the fastest-growing industries globally, as it is based on the synergy of an exponential development of digital technology, online connectivity and entertainment. Its key consumers and target group are children from the earliest age, so the positive and negative impact of the gaming industry ecosystem on children and their rights is a frequent academic research topic. When it comes to Serbia, this is the first research involving children and youth about their gaming habits and the gaming industry's impact on their rights. The responses related to the first research question showed that the age limit in starting gaming is slowly but surely moving towards the earliest childhood, considering the finding that Generation Z began to gameplay two years earlier than Millennials. Furthermore, gaming has become a part of everyday life of youth in Serbia, especially for boys, as one-third of the young survey respondents confirmed playing games every day or several times a week. The Serbian youth gameplay for fun, relaxation and out of boredom, mainly alone. Despite that, close to half of the survey respondents game online with friends or in their presence. Additionally, girls are at the forefront of not being inclined to miss seeing friends because of gaming. Although gaming in Serbia is not a social activity, it can be concluded that the paradigm of sociability is shifting from the real to the virtual environment in the new generations. The global pandemic has not yet led to an increase in gameplay, as less than half of the respondents confirmed that they play online games the same as before. 55

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

Since the research was conducted during summer, it is questionable how it will affect children and youth in the winter months. Since children and youth start gaming at an early age, parents and educators should look at it as an inalienable part of their childhood and adolescence. Online gameplay is impossible to limit, but parents can encourage children to game with friends, which would motivate their socialisation and reduce addiction, all critical during COVID-19. Game creators should keep in mind parental concern about excessive gameplay and the primary role that online games play in children’s lives as they grow up, treating them as key stakeholders to be informed, educated and listened to. In addition to game creators, gaming industry associations such as the Serbian Gaming Association can play a positive role towards parents, encouraging its members to treat them as stakeholders, and providing them with information and channels through which they will better understand their children's gaming habits. On the other hand, educators, due to the reduced attention span in children (Schaaf and Mohan, 2014) and the receptivity of digital environments to informing digital natives, should actively use gamification in the new context by the pandemic - online teaching. The education of teachers in primary schools, and the creation of stimulating and educational gaming solutions should be in the mission of the domestic gaming industry, but also of the Ministry of Education, especially in subjects that have proven to be more difficult to master in the online environment, as well as for younger pupils attending primary school. When it comes to the second research question concerning the protection of young gamers' rights on the digital playground, the research showed that children and youth in Serbia feel safe on the digital playground stating not having any unpleasant experiences, but are also feeling relatively unprotected. Limits such as time per session, in-game duration or the number of levels, and encouraging players to take a break do not exist in most of the children's games. Setting the age limit is exceedingly rare, which is reflected in young gamers' recommendations for the game developers. The surveyed children and youth are also not fully aware that the personal data they leave during the registration is exploited to monetise free games, so they are exposed to unwanted advertising that cannot be disabled. In this regard, the gaming industry should consider versions of their games adapted to children without collecting data, which at the same time would protect their rights to play and comply with the regulations on the privacy of data of underage users. Although some games, especially those offered through the app stores and consoles, have a mandatory set age limit, it should always be assumed that children will be present in adult game users' database if there are no robust age verification systems. Surveyed young Serbian gamers confirm that violence and stereotypes sometimes permeate games' content and report occasional inappropriate behaviour in multiplayer environments and chats. Game creators believe that setting codes of conduct, rules for players and respecting regulatory requirements are sufficient, not considering that the legalised slang and used forms are often inconspicuous and unclear not only for children but also for adult players on the platform. Educating young gamers on protecting their rights on the digital playground should become an embedded part of the game, which should be passed within the gaming. Furthermore, popular online gamers can play an important role, actively engaging in underage players' education about their rights and calling for the additional responsibility of the gaming industry. Besides, the psychology of gaming and its impact on children should be in the gaming industry's focus, which should commit to having a team psychologist. The person in this working position could perform impact and due diligence analysis of the online game on children, with the pre-set performance indicators to minimise children's adverse effects. To reduce or neutralise violent content, racial and gender stereotypes embedded in the finding that a third of the surveyed young gamers consider games they play as primarily intended for boys, the gaming industry should establish a balanced and diverse work environment, with adequate representation of female game designers and developers, besides different races. 56

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

To this end, the development and popularisation of educational profiles for the gaming industry, not only in technical occupations but also among related ones, such as architecture, design, arts, digital marketing and psychology, along with diversity scholarships, can have an impact on curbing stereotypes and supplying skilled personnel for the gaming industry in the medium term. The limitation of the research is that the sample of the surveyed children and youth is not nationally representative, which is why the findings cannot be generalised to all underage players in Serbia but only indicate the habits of young gamers and their attitudes regarding the protection of their rights on the digital playground. Further, how the U-Report platform, which was used to survey children and youth, is configured, was a limiting factor in structuring the questions that would lead to a more advanced statistical exploration of the findings. Recommendations for future academic research are to reconsider the habits of young gamers after the end of the pandemic, as well as to cross-reference the views of the gaming industry on children-gamers with the views of children and their parents, to identify the gaps and opportunities for improving children's rights on the digital playground.

ENDNOTES [1] https://serbia.ureport.in/ [2] https://data.mendeley.com/datasets/4tfx3gp9sr/1

ACKNOWLEDGEMENTS The research is part of the project funded by UNICEF Serbia: "Implementation of the Business for Results (B4R) Initiative 2020”.

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

REFERENCES Agante, L., & Pascoal, A. (2019). How much is “too much” for a brand to use an advergame with children? Journal of Product Brand Management, 28(2), 287–299. https://doi.org/10.1108/JPBM-08-2017-1554. Amerijckx, G., & Humblet, P. C. (2014). Child Well-Being: What Does It Mean? Children and Society, 5, 404-415. DOI:10.1111/chso.12003. Arı, E., Yılmaz, V., & Elmastas Dikec, B. (2020). An extensive structural model proposal to explain online gaming behaviors. Entertainment Computing, 34. 100340. https://doi.org/10.1016/j.entcom.2020.100340. Brooks, F. M., Chester, K. L., Smeeton, N. C., & Spencer, N. H. (2016). Video gaming in adolescence: factors associated with leisure time use. Journal of Youth Studies, 19(1), 36–54. https://doi.org/10.1080/136762 61.2015.1048200. Charmaraman, L., Richer, A. M., & Moreno, M. A. (2020). Social and Behavioral Health Factors Associated with Violent and Mature Gaming in Early Adolescence. International Journal of Environmental Research & Public Health, 17(14), 4996. https://doi.org/10.3390/ijerph17144996. Coyne, S. M., Padilla-Walker, L. M., Stockdale, L., & Day, R. D. (2011). Game On… Girls: Associations Between Co-playing Video Games and Adolescent Behavioral and Family Outcomes. Journal of Adolescent Health, 49(2), 160–165. https://doi.org/10.1016/j.jadohealth.2010.11.249. Dring, Ch. (2020, March). What is happening with video game sales during coronavirus? GamesIndustry.biz. Retrieved October 8, 2020, https://www.gamesindustry.biz/articles/2020-03-28-what-is-happeningwith-video-game-sales-during-coronavirus Entertainment Software Rating Board (n.d.). ESRB Game Ratings. Retrieved October 15, 2020, from https://www.esrb.org/ Folkvord, F., & Riet, J. V. T. (2018). The persuasive effect of advergames promoting unhealthy foods among children: A meta-analysis. Appetite, 129, 245–251. https://doi.org/10.1016/j.appet.2018.07.020. Geraniols, N. K. (2010). Hate speech corrodes online games. Games on NBCNews.com. Retrieved October 10, 2020, from http://www.nbcnews.com/id/36572021/ns/technology_and_science-games/t/hate-speechcorrodes-online-games/ Granic, I., Lobel, A., & Engels, R. C. M. E. (2014). The benefits of playing video games. American Psychologist, 69(1), 66–78. https://doi.org/10.1037/a0034857 Grimes, S.M. (2015). Online Games and Children. In The International Encyclopedia of Digital Communication and Society (Eds P.H. Ang and R. Mansell). New York, USA: John Wiley & Sons. DOI:10.1002/9781118767771. wbiedcs108 Hall, S. (2020, May). How COVID-19 is taking gaming and esports to the next level. World Economic Forum. Retrieved October 8, from https://www.weforum.org/agenda/2020/05/covid-19-taking-gaming-andesports-next-level/ Hastings, E. C., Karas, T. L., Winsler, A., Way, E., Madigan, A., & Tyler, S. (2009). Young children’s video/ computer game use: Relations with school performance and behaviour. Issues in Mental Health Nursing, 30, 638–649. https://doi.org/10.1080/01612840903050414. Huang, C. L., Yang, S. C., & Hsieh, L. S. (2019). The cyberbullying behavior of Taiwanese adolescents in an online gaming environment, Children and Youth Services Review, 106, 104461. https://doi.org/10.1016/j. childyouth.2019.104461. Hygen, B. W., Belsky, J., Stenseng, F., Skalicka, V., Kvande, M. N., Zahl-Thanem, T., & Wichstrom, L. (2020). Time Spent Gaming and Social Competence in Children: Reciprocal Effects Across Childhood. Child Development, 91(3), 861-875. https://doi.org/10.1111/cdev.13243. International Age Rating Coalition (n.d.). About IARC. Retrieved October 27, 2020, from https://www.globalratings.com/about.aspx

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

Jackson, L. A., Witt, E. A., Games, A. I., Fitzgerald, H. E., von Eye, A., & Zhao, Y. (2012). Information Technology Use and Creativity: Findings from the Children and Technology Project. Computers in Human Behavior, 28, 370–376. DOI:10.1016/j.chb.2011.10.006. Kaye, L. K., & Bryce, J. (2012). Putting The “Fun Factor” Into Gaming: The Influence of Social Contexts on Experiences of Playing Videogames. International Journal of Internet Science, 7(1), 23–36. Kiili, J. & Moilanen, J. (2019). Participation as a Methodological and Ethical Issue in Child Protection Research. Journal of Children's Services. 14(3), 143–161. DOI: 10.1108/JCS-02-2019-0009. King, D. L., Delfabbro, P. H., Billieux, J., & Potenza, M. N. (2020). Problematic online gaming and the COVID-19 pandemic. Journal of Behavioral Addictions, 9(2), 184–186. https://doi.org/10.1556/2006.2020.00016. Lemmens, J. S., Valkenburg, P. M., & Peter, J. (2011). The effects of pathological gaming on aggressive behavior. Journal of Youth and Adolescence, 40(1), 38–47. https://doi.org/10.1007/s10964-010-9558-x Lenhard, A. (2015, August). Teens, Technology and Friendships. PewResearch Center. Retrieved October 12, 2020, from https://www.pewresearch.org/internet/2015/08/06/teens-technology-and-friendships/ Lissak, G. (2018). Adverse physiological and psychological effects of screen time on children and adolescents: literature review and case study. Environmental Research, 164, 149–157. DOI:10.1016/j.envres.2018.01.015. López-Cabarcos, M. Á., Ribeiro-Soriano, D., & Piñeiro-Chousa, J. (2020). All that glitters is not gold. The rise of gaming in the COVID-19 pandemic. Journal of Innovation & Knowledge, 5(4), 289–296. https://doi. org/10.1016/j.jik.2020.10.004. Lynch, T., Tompkins, J. E., van Driel, I. I., & Fritz, N. (2016). Sexy, Strong, and Secondary: A Content Analysis of Female Characters in Video Games across 31 Years. Journal of Communication, 66(4), 564–584. https:// doi.org/10.1111/jcom.12237. Malkowski, j. & Russworm, T.M. (2017). Gaming Representation: Race, Gender, and Sexuality in Video Games. Indiana University Press. Mannikko, N., Billieux, J., & Kaariainen, M. (2015). Problematic digital gaming behaviour and its relation to the psychological, social and physical health of Finnish adolescents and young adults. Journal of Behavioral Addictions, 4, 282-288. https://doi.org/10.1556/2006.4.2015.040. Ministry of Youth and Sports of the Republic of Serbia. (2015). National Youth Strategy for the period from 2015 to 2025. Retrieved October 6, 2020 from https://www.mos.gov.rs/wp-content/uploads/download-managerfiles/nacionalna_strategija_za_mlade0101_cyr.pdf Olson, C. K. (2010). Children’s Motivations for Video Game Play in the Context of Normal Development. Review of General Psychology, 14(2), 180–187. https://doi.org/10.1037/a0018984. Pappas, I.O., Mikalef, P, Giannakos, M.N. & Kourouthanassis, P.E. (2019). Explaining user experience in mobile gaming applications: an fsQCA approach. Internet Research, 29(2), 293–314. https://doi.org/10.1108/ IntR-12-2017-0479. Pawłowska, B., Potembska, E., & Szymańska, J. (2018). Demographic and family-related predictors of online gaming addiction in adolescents, Polish Journal of Public Health, 128(1), 9-13. DOI: https://doi.org/10.2478/ pjph-2018-0002. Pan European Game Information (n.d.). About PEGI. Retrieved October 13, 2020, from https://pegi.info/whatdo-the-labels-mean Prensky, M. (2012). From digital natives to digital wisdom: Hopeful essays for 21st century learning. Thousand Oaks, California, USA: Corwin Press. Przybylski, A. K. (2014). Electronic gaming and psychosocial adjustment. Pediatrics, 134, 716–722. https://doi.org/10.1542/peds.2013-4021. Schaaf, R. & Mohan, N. (2014). Making School a Game Worth Playing: Digital Games in the Classroom. Thousand Oaks,California, USA: Corwin. SuperData. (2020). 2019 Year in Review: Digital games and interactive media, SuperData Research Holdings Inc. Retrieved October 8, 2020, from https://www.superdataresearch.com/reports/p/2019-year-in-review 59

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

United Nations. (1989). The Convention on the Rights of the Child. General Assembly Resolution no. 44/25, c.1. Retrieved October 6, 2020 from https://www.unicef.org/child-rights-convention United Nations Children Fund. (2020). Online Gaming and Children’s Rights: Recommendations for the Online Gaming Industry on Assessing Impact on Children, UNICEF. United Nations Children Fund. (2019). Child Rights and Online Gaming: Opportunities & Challenges for Children and the Industry. Discussion Paper Series: Children’s Rights and Business in a Digital World. UNICEF. Ventura, M., Shute, V., & Zhao, W. (2013). The relationship between video game use and a performance-based measure of persistence. Computers & Education, 60, 52–58. DOI:10.1016/j.compedu.2012.07.003. World Health Organisation (2020). Excessive screen use and gaming considerations during #COVID19. WHO Regional Office for the Eastern Mediterranean. Retrieved October 13, 2020, from http://www.emro.who. int/mnh/news/considerations-for-young-people-on-excessive-screen-use-during-covid19.html

EJAE 2021  18 (2)  49 - 61

KRSTIĆ. N.  DIGITAL PLAYGROUND: FRIEND OR FOE TO THE CHILDREN?

DIGITALNO IGRALIŠTE: PRIJATELJ ILI NEPRIJATELJ DECE?

Rezime: Za centralnu industriju zabave 21. veka - onlajn igrice, deca su nesumnjivo ključna potrošačka grupa. Iako se istraživanje uticaja gejming industrije na decu uglavnom bavi negativnim efektima poput zavisnosti, nasilnog sadržaja, neprimerenog ponašanja i monetizacije ličnih podataka, postoje i mnogi pozitivni efekti - porodična zabava, virtuelno druženje, poboljšanje kognitivnih veština i korišćenje igara kao nastavno sredstvo. Stoga je zadatak svih učesnika u lancu vrednosti gejming industrije da maksimiziraju pozitivne i minimiziraju negativne uticaje na decu. Istraživanje sprovedeno među 893 mladih u Srbiji pokazalo je njihove navike u konzumiranju onlajn igara i ukazalo na to da li su njihova prava zaštićena tokom igranja. Zaključak daje preporuke ključnim zainteresovanim stranama u ekosistemu gejming industrije o tome kako digitalno igralište učiniti inkluzivnim, sigurnim i odgovornim kada je reč o poštovanju dečjih prava.

Ključne reči: prava deteta, gejming, gejming industrija, Kovid-19, Srbija.

EJAE 2021, 18(2): 62 - 75 ISSN 2406-2588 UDK: 336.1/.5(6-15) 336.225(6-15) 338.23:330.101.54 DOI: 10.5937/EJAE18-30727 Original paper/Originalni naučni rad

MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES Gideon I. Ihuarulam, Gbenga Peter Sanusi*, L.O. Oderinde Department of Economics Anchor University, Lagos, Nigeria

Abstract: The determination of the effects of macroeconomic environment on tax revenue is very vital for every country and more so for an economic community aiming for harmonization of macroeconomic environment and ultimately integration. However, the extent to which aggregate output, inflation, and unemployment affect tax revenue in ECOWAS has been less studied in the literature. Therefore, this study empirically investigates how tax revenue is related to selected macroeconomic variables. Panel data analysis is employed on six ECOWAS countries’ data set on tax revenue, gross domestic product, inflation, unemployment, trade openness and exchange rate over 2005-2019. The Wald’s test and Hausman test indicated that the fixed effects regression was appropriate for the study. The results showed that inflation was positively related to tax revenue and statistically significant at 5 percent. A unit increase in inflation led to 0.007 increase in tax revenue measure; economic growth was also positive and statistically significant at 5 percent; a unit rise in GDP resulted in 0.78 rise in governmental tax revenue variable. Finally, Tax revenue variable decreased by 0.10 with a unit increase in unemployment. It is recommended that ECOWAS countries should carefully manage their macroeconomic environment to boost tax revenue.

Article info: Received: February 4, 2021 Correction: Jun 4, 2021 Accepted: September 9, 2021

Keywords: Macroeconomic variables, Tax revenue, ECOWAS, Hausman Test, Optimal tax theory. JEL Classification: H20, E62, C23, B22

INTRODUCTION Tax Revenue as one of the resources needed by the government of any nation of the world, either developing or developed, is very crucial in the discharge of duties and obligations (Michael, 2012 ; Bersley & Persson, 2014 ; Andersson & Lazuka, 2019).There are various sources of the revenue available to the government and its agencies for its efficient and effective functioning. According to Afuberoh and Okoye (2014) these sources include amongst others: revenue from natural resources, rents, royalties, foreign aid, grants, interests from loans, interest from capital investments, and tax revenue, which seems to be the oldest form of revenue. 62

*E-mail: gsanusi@aul.edu.ng

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Furthermore, tariff on international trade could form additional sources of government revenue (Lashkaripour, 2020).

Tax revenue is a revenue generated through taxation. According to Michael (2012), a tax instrument is a means by which a government generates a large amount of its revenue, thus manipulating the economy. Taxation as an instrument plays a key role in the regulation of any economy, as well as its performing, as it serves as a tool of either increasing or decreasing money supply in the economy. The importance of revenue in general and tax revenue in particular cannot be underestimated or toyed with by any government if it will succeed in the discharge of its expected duties and obligations. Given an ever-increasing population in every country especially 15 ECOWAS nations (Benin, Burkina Faso, Cape Verde, Cote d'Ivoire, The Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, and Togo) and around the world, the role of government in these population exploding situations has become increasingly enormous. Thus, there is a need for a larger pocket of revenue, and tax revenue could be a very important and significant source. Furthermore, Terefe and Teera (2018) argued that the drive towards sustainable development, which is a base of the improved welfare and living standard, depends crucially on the availability of massive resources mobilized within the economy via tax revenue. Improvements in macroeconomic variables have been put forth as a boost to government tax revenue achievement (see Saibu & Olatunbosun, 2013; Castro & Camarillo, 2014; Andrejouska & Pulikova, 2018; Mawejje & Munyambonera, 2016; Arnold, Brys, Heady, Johansson, Schwellnus, & Vartia, L, 2011). How does inflation and exchange rate impact on tax revenue in ECOWAS? What is the impact of trade openness and economic growth on government revenue from taxation ECOWAS? These are the questions that this present study intends to answer by providing empirical evidence which could influence policy formation. The broad objective of this research is to examine how tax revenue is related to key selected macroeconomic variables. Specifically, it investigates: 1. The effects of inflation and exchange rate on tax revenue in selected countries in ECOWAS 2. The role of trade openness and economic growth on tax revenue in ECOWAS, and 3. The impacts of unemployment rate on government tax income in selected countries in ECOWAS The rest of the study is as follows: section two focuses on Literature review - in this section, literature related to the subject of consideration was reviewed following a sequence of theoretical review and empirical reviews. Methodology of the study is discussed in Section three – this section underscores the research design employed, the description and sources of data used, and data analysis method as well as specification of the model employed. Empirical analysis and discussion of results are presented in section four. Finally, the summary of major findings and conclusion drawn are discussed in the concluding section.

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

BRIEF LITERATURE REVIEW

First, two theories of tax collection are discussed: the Laffer curve hypothesis and Optimal tax theory. Second, a brief empirics on tax revenue and macroeconomic variables are also examined. Economist Arthur Laffer (and others) popularized the idea of Laffer curve hypothesis, originally proposed in a limited form by North African philosopher and social scientist Khaldun around AD 1377. Rather than implying that the relationship between tax rates and tax revenue is not a simple straight line, the theory suggests that the increasing tax rates beyond some point result in lower tax revenues (Kenneth, 2019). Laffer describes the situation as following: “At a tax rate of zero percent, the government would collect no tax revenues, no matter how large the tax base. Likewise, at a tax rate of 100 percent, the government would also collect no tax revenues because no one would be willing to work for an after-tax wage of zero (that is, there would be no tax base). Between these two extremes there are two tax rates that will collect the same amount of revenue: a high tax rate on a small tax base and a low tax rate on a large tax base” (as cited in Kenneth, 2019). The standard theory of optimal taxation posits that a tax system should be chosen to maximize a social welfare function subject to a set of constraints (Mankiw, Weinzierl & Yagan, 2009). The literature on optimal taxation typically treats the social planner as utilitarian, i.e., the social welfare function is based on the utilities of individuals in the society. In its most general analyses, this literature uses a social welfare function that is a nonlinear function of individual utilities. Nonlinearity allows for a social planner who prefers, for example, more equal distributions of utility (Mankiw et al, 2009; Blundell, 2012). Other studies considered some salient aspect of optimal taxation theory, like its design in an open economy (Ogawa & Hosoe, 2020), marginal social welfare weighting (Saez & Stantcheva, 2016), accounting for the possibility of tax evasion (Brunner, Eckerstorfer & Pech, 2013; Artavanis, Morse & Tsoutsoura, 2016). A study conducted by Ade, Rossouw, & Gwatidzo, (2018) investigates the determinants of tax revenue performance in all 15 Southern African Development Community countries during 1990-2010, using panel data. He considered variables such as foreign direct investment, tax policy harmonization measures, government expenditure, and growth rate of domestic credit, export share of GDP, inflation and VAT harmonization indicator. The empirical findings indicated the importance of foreign direct investment inflows towards tax revenue collected in the SADC and the existence of reverse causality (Ade et al, 2018). Ali and Audi (2018) considered Macroeconomic environment and tax revenue in Pakistan where they examined the impact of macroeconomic situations on tax revenues in the case of Pakistan over the period from 1975 to 2016. They found out that unemployment had a positive and significant impact on tax revenue, money supply was also positive and significant on tax revenue, while inflation was negative and significant on tax revenue. It indicated that the macroeconomic environment was healthy, which further fostered the increase in tax revenue. Shivanda and Obwogi (2018) investigated the effects of macroeconomic variables on tax revenue in Kenya. Their analysis included data set from 1995 to 2016 using ANOVA. Their findings, among other, indicated that there is no statistically significant relationship between interest rate, inflation and exchange rate and tax revenue; however, interest rate and exchange rate were important in predicting tax revenues. The general conclusion was that interest rate and exchange rate are important macroeconomic factors influencing tax revenue collection in Kenya. 64

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Mawejje and Munyambonera (2016) study on Uganda, stressed the importance of sectoral economic performance, while the study of Saibu and Olatunbosun (2013) on Nigeria presented results that confirm the importance of macroeconomic variables. The study conducted by Onakoya, Afintinni, & Oyeyemi (2017) utilized the periods 2005-2014 to determine the significant relationship between tax revenue performance, trade liberalization and macro-economic variables of 22 sub-Saharan African countries. Several tests were conducted, and the Vector error correction model was engaged to check for possible long or short run connection among the variables. The Granger causality test was applied to determine the shock impact of one variable on the other. The findings concluded that inflation, interest rate and trade openness had a short run relationship with tax revenue, unlike exchange rate and unemployment. All variables apart from exchange rate were positively related to the dependent variable coupled with the fact that there existed a one-way causation with the absence of serial correlation and heteroskedasticity among the variables (Onakoya, et al., 2017). However, the case of outliers in the data set of some countries may have adversely affected the results obtained. Rodríguez’s (2018) contribution to this research area is the focus on developing countries. The author examined the role of governance indicators like government effectiveness, democracy, political stability and other additional variables like financial intermediation, internal trade volume, agriculture share or aggregate economy. In this panel data study, over 138 developing countries were covered for the period (1976–2015). The findings confirm the importance of governance variables and Agriculture’s share in gross domestic product in determining the level of tax revenue. Botlhole, Asafu-Adjaye, & Carmignani (2012) and Mawejje (2019) extend the analysis of the determinants of tax revenue performance in sub-Saharan Africa to account for the significant role of institutions and natural resource abundance and its governance. Studies on OECD countries like the work of Castro and Camarillo (2014) showed that key macroeconomic variables, including but not limited to GDP per capita, exert significant positive influence on tax revenue generation. Arachi, Bucci, & Casarico (2015) study on a panel of OECD countries corroborates this. The study by Terefe & Teera (2018) focused on East African countries for the period 1992-2015. Their main findings were that key macroeconomic variables, such as GDP per capita, exert positive impact on tax revenue while inflation and exchange rates impact on tax revenue negatively. Neog & Gaur (2020) investigated the determinants of tax performance in the BRICS (Brazil, Russia, India, China and South Africa) countries for the period 1996-2017, with the focus on economic and political variables. Results obtained showed that economic development, trade openness and control of corruption are revenue-enhancing factors for BRICS, whereas the agriculture sector discourages the tax revenue performance. Having looked at the existing literature, it is evident that there is a gap in the literature, as the number of studies that examine the aggregate impact of macroeconomic variables on the ability of various economies to generate revenue from tax, either indirect or direct tax, with specific reference to ECOWAS is limited. This study thus aims to fill the existing gap in literature by carrying out an empirical analysis of the impact of the selected macroeconomic variables on tax revenue in Economic Community of West Africa States.

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

METHODOLOGY

Theoretical framework Analysts investigating the role of macroeconomic variables on tax revenue performance frequently adopt the Laffer Curve Hypothesis or the optimal tax theory as theoretical framework. These were expounded upon in section 2.1 of this paper. The standard optimal tax theory is adopted in this study as the theoretical basis for the empirical investigation following the framework drawn from the works of Ade et al (2018), Shivanda and Obwogi (2018), Terefe and Teera (2018).

Model Specification The general model indicating the relationship between tax revenue and selected macroeconomics variables is as specified in equation 1. (1) Where: TR = Tax revenue as a percentage of GDP, INF = Inflation Rate, GDP = Gross Domestic Product, EXC = Exchange Rate, TO = Trade Openness, UNM = Unemployment Rate To determine the responsiveness of dependent variable to the independent variables, the general panel econometric model is as specified in equation 2: (2) The variables are as defined earlier, while µ represents the error term. Following the panel method technique, three models were considered and estimated as follows: the Pooled Regression Model (PRM), Fixed Effect Model (FEM) and Random Effect Model (REM). The Hausman test was conducted in order to choose between Fixed Effect Model (FEM) and Random Effect Model.

Data The data were sourced from World Bank Indicators, 2020 and relevant statistical bulletins of selected countries in ECOWAS. The variables were analyzed and their measurements included: tax revenue as a percentage of GDP (%), inflation rate in percentage (%), unemployment rate in percentage (%), exchange rate in percentage (%), gross domestic product in billion USD ($), trade openness as calculated by the world bank. Inflation rates are calculated by the annual percentage change in the CPI; GDP is measured as the real GDP in billion USD ($) generated within the year, unemployment rate is measured as the ratio of unemployed individuals and total work force; tax revenue is measured as the percentage of total GDP; exchange rate is measured as the ratio of home currencies to a unit of foreign currency; and, trade openness is measured by the ratio between the sum of exports and imports and gross domestic product as calculated by the world bank.

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

RESULTS AND DISCUSSION Results Descriptive statistics

Descriptive statistics of the variables used in the model are presented in Table 1. Table 1. Descriptive statistics TR

GDP

INF

UNM

EXC

Mean

20.32

22.31

117.43

5.28

85.42

234.29

Median

20.55

22.46

107.50

4.27

80.75

127.23

Maximum

24.18

26.78

268.36

12.25

311.35

612.00

Minimum

13.87

15.83

52.93

1.98

20.72

0.91

Std. Dev.

2.71

3.17

45.98

3.12

49.51

214.83

Skewness

-0.83

-0.59

1.67

0.98

2.18

0.53

Kurtosis

3.03

2.55

5.51

2.78

10.19

1.59

Jarque-Bera

10.48

6.09

65.85

14.61

265.93

11.76

Probability

0.005

0.04

Sum

0.00

0.0006

0.00

Sum Sq. Dev.

1829.04

2008.61

10568.87

475.37

7688.58

21086.70

Observations

655.63

898.45

188201.2

870.79

218173.5

4107589.

Source: Authors’ Computation, 2020

Table 1 showed that exchange rate and trade openness recorded the highest variation, while tax revenue and unemployment have relatively lower variability. Some variables recorded a positive skewness, which means that the distributions are tailed to the right, while tax revenue and GDP have a negative skewness.

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Stationarity (unit root) Test

The results of Augmented Dickey–Fuller (ADF) and Philips–Perron (PP) Tests, showing the order of integration I (d) of the series as presented in Table 2. It shows that the series are integrated of order zero; I (0) or integrated of order one; I (1). Table 2. Results of Unit Root Test Augmented Dickey Fuller test Variables

Phillip Peron Test

Levels

1st diff.

2nd diff.

Stationarity

Levels

1st diff.

2nd diff.

Stationarity

0.10

0.00

I(1)

0.00

I(0)

GDP

0.23

0.00

I(1)

0.03

0.00

I(0)

INF

0.00

0.01

0.00

I(0)

0.00

I(0)

EXR

0.99

0.00

I(1)

0.99

0.00

I(1)

0.05

0.00

I(1)

0.16

0.00

I(1)

UNMP

0.18

0.00

I(1)

0.63

0.00

I(1)

Source: Authors’ Computation, 2020

Under the Augmented Dickey Fuller test, inflation was stationary at levels, while LGDP, LTAX, unemployment, trade openness and exchange rate were stationary at first differencing.

Variance inflation factor (multicollinearity) Table 3. Multicollinearity results Coefficient

Un-centered

Centered

Variance

VIF

35.69

15727.78

GDP

0.06

15227.89

1.52

INF

2.77E-06

19.05

2.23

UMP

0.003

42.14

1.62

3.21E-06

11.47

1.14

EXC

2.48E-06

61.81

1.89

Variable

Source: Authors’ Computation, 2020

Multicollinearity test is a test to confirm which variables have multicollinearity. When the centered VIF is 1, it is said that no multicollinearity exists. When the centered VIF is between 1 and 5, there is multicollinearity but at a negligible level. When the centered VIF is between 5 and 10, there is multicollinearity at a reasonable level and corrections may be necessary. 68

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Finally, when the centered VIF is above 10, it indicates a high correlation and is cause for concern. From Table 3 multicollinearity results, multicollinearity exists but at a negligible level, hence no correction is required. Model Selection The model selection is between the Pooled OLS Regression Model, Random Effects Model and Fixed Model Effects. Table 4 presents the results for each of the models, without any prior adjustment. Table 4. Analysis results POOLED

FIXED

RANDOM

Coef.

St-error

Prob.

Coef.

St-error

Prob.

Coef.

St-error

Prob.

1.88

0.37

0.00

2.59

2.71

0.34

1.88

0.86

0.00

GDP

0.79

0.01

0.00

0.78

0.11

0.00

0.79

0.02

0.00

INF

0.005670

0.00

0.007

0.0009

0.00

UMP

0.028884

0.00

-0.10

0.017

0.0000

0.02

0.019

0.73

-0.0019

0.00

0.02

-0.0019

0.0012

0.12

-0.0019

0.002

0.34

EXC

0.00045

0.0001

0.00

0.00025

0.0004

0.57

0.000456

0.00

0.13

A-RSQR

0.95

0.96

0.95

F-STAT

2404.26

221.41

378.19

Durbin

0.82

1.45

0.58

Source: Authors’ Computation, 2020

Correlated Random Effects - Hausman Test The Hausman test is employed to determine the appropriate model. The Hausman test is used to check which type of regression analysis model between random effect and fixed effect is appropriate for interpretation. Table 5. Correlated Random Effects - Hausman Test Equation: Hausman Test Test cross-section random effects Test Summary

Chi-Sq. Statistic

Chi-Sq. d.f.

Prob.

Cross-section random

18.53

0.0023

Source: Authors’ Computation, 2020

This formed the basis of our decision to use the fixed effect method. The probability value from Table 5: Correlated Random Effects - Hausman Test is 0.023, which is less than 5%, thereby we reject the null hypothesis and state that the fixed effects model is appropriate. 69

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Wald’s Test

Wald’s test is further employed to determine the best model to employ between fixed effect model, and pooled regression analysis. Table 6. Wald’s test result Wald Test: Equation: Eq01 Test Statistic

Value

Probability

F-statistic

1060.618

(5, 79)

0.0000

Chi-square

5303.091

0.0000

Source: Authors’ Computation, 2020

From Table 6: Wald’s test result, we deduce that the coefficients are neither equal to each other, or equal to zero (0), hence we reject Null hypothesis and conclude that the fixed effect model is the appropriate model as the probability value (0.0000) is less than 5%. This is reasonable, given that different countries are involved. Thus, between the pooled regression and the fixed effect regression, the fixed effect regression is the best model for the interpretation.

Fixed Effects regression Results Table 7. Fixed Panel EGLS (Cross-section SUR) Variable

Coefficient

Std. Error

t-Statistic

Prob.

2.59

2.71

0.95

0.34

GDP

0.78

0.11

6.62

0.00

INF

0.007

0.00098

7.55

0.00

UNF

-0.10

0.017

-5.77

0.00

TRO

-0.0019

0.0012

-1.53

0.12

EXC

0.00025

0.00044

0.56

0.57

R-squared

0.96

Adjusted R-squared

0.94

F-statistic Prob. (F-statistic)

Durbin-Watson stat

1.45

161.91 0.00

Source: Authors’ Computation, 2020

Findings as shown in table 7 reveal that Inflation is positively related to tax revenue, and also significant. This implies that as inflation increases by 1 unit, tax revenue as percentage of GDP will increase by 0.0074 unit. Exchange rate is positively related to tax revenue, but insignificant, implying that as exchange rate increases by 1 unit, tax revenue as percentage of GDP will increase by 0.00025. 70

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

Gross domestic product is positively related to tax revenue, and also significant. This implies that as gross domestic product increases by 1 unit, tax revenue as percentage of GDP will increase by 0.784216 unit. Trade openness is negatively related to tax revenue, but insignificant, implying that a unit increase in trade openness, leads to a fall in tax revenue as percentage of GDP by 0.001916 unit. Further, unemployment is negatively related to tax revenue, and also significant, implying that as unemployment increases by 1 unit, tax revenue will decrease by 0.102872 units. The Adjusted R2 indicates that the independent variables in consideration account for 95% of the changes in the dependent variable.

Residual Cross-Section Dependence Test The Residual Cross-Section Dependence Test is used to check for correlation between the errors of different cross sections. If there is correlation, then the appropriate GLS weights will be used to correct for the correlation. Table 8. Residual Cross-Section Dependence Test Test

Statistic

df.

Prob.

Breusch-Pagan LM

30.76064

0.0095

Pesaran scaled LM

2.877486

0.0040

Bias-corrected scaled LM

2.663200

0.0077

Pesaran CD

2.087956

0.0368

Source: Authors’ Computation, 2020

Due to the span of the period (15), the study employed the Pesaran CD test to determine crosssection dependence in residuals. From the Table 8: Residual Cross-Section Dependence Test, we reject the null hypothesis, because the calculated t-statistics (2.087) is greater than the critical value (1.8) and state that there exists correlation between the cross-sectional residuals. We corrected for this using the Generalized Least Square Method (cross section sur).

Discussion Inflation is positively related to tax revenue, significantly implying that as inflation increases by 1 unit, tax revenue as percentage of GDP will increase by 0.007414 US. This can be a result of an increase in income of workers by employers, to account for the inflation, and hence an increase in purchase of goods and services leading to a greater tax revenue from value added tax. This coheres with the results of Onakoya, et al., 2017; Saibu & Olatunbosun, 2013; Andersson & Lazuka, 2019). Exchange rate is positively related to tax revenue, but insignificant, implying that a unit increase in exchange rate leads to 0.000251 increase in tax revenue as percentage of GDP. This could be a result of the dynamics of exchange rate theories. Even though trade import would be expected to fall, which may invariably reduce tax collected, a close look suggests that inelastic products, such as essential 71

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

goods which are not locally available or raw materials, may not reduce import demand. Thus, a rise in exchange rate could lead to increase in tax revenue from import duties or trade. This result is consistent with the results of Caroline and Joseph (2017), where they found out that the real exchange rate tends to rise by the full amount of any consumption tax increase as well as levies. Gross domestic product is positively related to tax revenue, and also significant. When gross domestic product increases by 1 unit, tax revenue as percentage of GDP increases by 0.784. Theoretically speaking, it is expected for tax revenue to increase in response to a corresponding increase in GDP, due to the fact that the tax base has increased. Trade openness is negatively related to tax revenue, but insignificant, implying that as trade openness increases by 1 unit, tax revenue as percentage of GDP will reduce by 0.001916 units. A reason for this result, which did not follow theoretical foundation, could be the presence of gross mismanagement or porous trade routes, borders, the presence of high corruption and lots of smuggling. In contrast, Micah, Bbaale, & Hisali (2017) established a theoretically correct result that the average tariff rate used as a measure for trade openness positively influences total tax, indirect tax and trade tax while the average tariff rate squared is negative, illustrating the “Laffer effect”. Unemployment rate is negatively related to tax revenue, and also significant. This suggests that as unemployment increases by 1 unit, tax revenue as percentage of GDP will decrease by 0.102 US dollars. This is because, the higher the unemployment rate, the lower the tax base and the lower the tax revenue from wages and salaries, which is a major source of the governmental tax revenue.

CONCLUSION This paper contributes to a growing strand of literature on the macroeconomic determinants of tax revenue performance in Sub-Saharan African countries, particularly in Economic Community of West African States, where there is a dearth of accessible studies. The main contribution of this research is to provide estimates of the quantitative impact on tax revenues of changes in key macroeconomic variables like Inflation, GDP, Trade openness, Unemployment and Exchange rate in recent times. The paper focused on some selected countries that are part of the Economic Community of West African States (ECOWAS) including; Nigeria, Ghana, Liberia, Burkina Faso, Togo, and Cape Verde. Robust panel data analytical approach was adopted for data set of these countries using the standard optimal tax theory as theoretical foundation. The results showed that the key drivers of tax revenue performance in the countries studied and the period covered are economic expansion measured via GDP, Inflation rates and unemployment. GDP and inflation exert a positive effect while increasing unemployment rate indicates underutilization of limited resources and thus depresses tax revenue attainment of government. Furthermore, the study provided evidential basis for the attainment of robust tax revenue by careful management of the macroeconomic environment.

EJAE 2021  18 (2)  62 - 75

IHUARULAM. I. G., SANUSI. G. P., ODERINDE. L. O.  MACROECONOMIC DETERMINANTS OF TAX REVENUE IN ECONOMIC COMMUNITY OF WEST AFRICAN STATES

REFERENCES

Ade, M., Rossouw, J. & Gwatidzo, T. (2018). Determinants of tax revenue performance in the Southern African Development Community (SADC) ERSA working paper 762). Retrieved October 2020 from https://econrsa. org/system/files/publications/working_papers/working_paper_762_final.p Afuberoh, D. & Okoye E. (2014). The Impact of Taxation on Revenue Generation in Nigeria: A Study of Federal Capital Territory and Selected States. International Journal of Public Administration and Management Research, 2(2), 22-47. Ali, A & Audi, M. 2018. Macroeconomic Environment and Taxes Revenues in Pakistan: An Application of ARDL Approach. MPRA Paper No. 88916 Retrieved October 2020 from https://mpra.ub.uni-muenchen. de/88916/1/MPRA_paper_88916.pdf Andersson, J & Lazuka,V.(2019). Long-term drivers of taxation in francophone West Africa 1893–2010, World Development, 114, (294-313), Andrejouska, A. & Pulikova, V. 2018. Tax Revenues in the Context of Economic Determinants Montenegrin Journal of Economics, 14(1), 133-141. Arachi, G., Bucci, V. & Casarico, A. (2015): Tax structure and macroeconomic performance International Tax and Public Finance, 22, 635–662. Arnold, J., Brys, B., Heady, C., Johansson, A., Schwellnus, C., & Vartia, L. (2011). Tax policy for economic recovery and growth. The Economic Journal, 121(550), F59–F80. Artavanis, N., Morse, A., & Tsoutsoura, M. (2016). Measuring income tax evasion using bank credit: Evidence from Greece. The Quarterly Journal of Economics, 131(2), 739-798. Bersley T, & Persson T (2014). Why do developing countries tax so little? Journal of Economic Perspectives, 28(4),99-120. Blundell, R. (2012): Tax policy reform: The role of empirical evidence. Journal of the European Economic Association, 10(1), 43-77. Botlhole,T., Asafu-Adjaye,J., & Carmignani, F. (2012), Natural resource abundance, institutions and tax revenue mobilization in Sub-Saharan Africa. South African Journal of Economics, 80, 135-156. https://doi. org/10.1111/j.1813-6982.2011.01305.x Brunner,J.K., Eckerstorfer, P. & Pech, S. (2013): Optimal taxes on wealth and consumption in the presence of tax evasion Journal of Economics, 110(2),107-124. DOI: 10.1007/s10797-015-9364-1 Caroline F. & Joseph E.G. (2017). Effects of Consumption Taxes on Exchange Rates and Trade Balances. Working Paper. Peterson Institute for International Economics. Castro, G.A. & Camarillo, D.B.R. 2014. Determinants of tax revenue in OECD countries over the period 20012011 Contaduría y Administración, Accounting and Management, 59(3), 35-60. Kenneth J. P. 2019. A Plot of the Laffer Curve Retrieved October 2020 from https://www.researchgate.net/ publication/332719140_A_Plot_of_the_Laffer_Curve Lashkaripour, A. (2020): Can Trade Taxes be a Major Source of Government Revenue?, Journal of the European Economic Association, Early View, jvaa058. https://doi.org/10.1093/jeea/jvaa058 Mankiw, N. Gregory, Matthew Weinzierl & Canny Yagan 2009, Optimal taxation in theory and practice, Journal of Economic Perspectives, 23, 147-174. Mawejje, J. & Munyambonera, E. (2016), Tax Revenue Effects of Sectoral Growth and Public Expenditure in Uganda. South African Journal of Economics, 84, 538-554. https://doi.org/10.1111/saje.12127 Mawejje,J., (2019) Natural resources governance and tax revenue mobilization in sub-Saharan Africa: The role of EITI, Resources Policy, 62,176-183. Micah, S. G., Bbaale, E., & Hisali, E. (2017) Trade Openness and Tax Revenue Performance in East African Countries. College of Business and Management Science. Modern Economy, 8, 690-711. 73

EJAE202118(2)62-75

IHUARULAM.I.G.,SANUSI.G.P.,ODERINDE.L.O.MACROECONOMICDETERMINANTSOFTAXREVENUEINECONOMICCOMMUNITYOFWEST AFRICANSTATES

Michael, K. (2012) Taxation and Development—Again. IMF Working Paper, Fiscal Affairs Department. Serial no. WP/12/220. Neog,Y.& Gaur, A.K.(2020) Shadow economy, corruption, and tax performance: A study of BRICS, Journal of Public Affairs, Early View. https://onlinelibrary.wiley.com/doi/10.1002/pa.2174 Ogawa, Y., Hosoe, N. (2020): Optimal indirect tax design in an open economy. International Tax and Public Finance 27, 1081–1107. https://doi.org/10.1007/s10797-020-09592-8 Onakoya, B.A., Afintinni, O.I., and Oyeyemi, G.O. (2017). Taxation and Revenue Growth in Africa. Journal of Accounting and Taxation. 9, 11-22. DOI: 10.5897/JAT2016.0236 ID.56461882 Rodríguez, V.M.C (2018): Tax determinants revisited. An unbalanced data panel analysis. Journal of Applied Economics, 21(1), 1-24. Saibu, O.M. and Olatunbosun O. S. 2013. Macroeconomic Determinants of Tax Revenue in Nigeria (1970-2011) World Applied Sciences Journal, 28(1), 27-35. DOI: 10.5829/idosi.wasj.2013.28.01.1189 Saez, E. and Stantcheva, S. (2016). Generalized Social Marginal Welfare Weights for Optimal Tax Theory The American Economic Review, 106(1), 24-45. Shivanda, A.R. and Obwogi, J. (2018). Effect of Macroeconomic Variables on Tax Revenue in Kenya International Journal of Social Sciences and Information Technology, 4(11), 33-47. Terefe, K.D., and Tera, J. (2018). Determinants of tax revenue in East African countries: An application of multivariate panel data cointegration analysis Journal of Economics and International Finance, 10(11), 134-155, DOI: 10.5897/JEIF2018.0924 World Bank .2020. World Development Indicators 2020 Retrieved October 2020 from https://databank.worldbank. org/source/world-development-indicators

EJAE202118(2)62-75

IHUARULAM.I.G.,SANUSI.G.P.,ODERINDE.L.O.MACROECONOMICDETERMINANTSOFTAXREVENUEINECONOMICCOMMUNITYOFWEST AFRICANSTATES

MAKROEKONOMSKE DETERMINANTE POREZNIH PRIHODA U EKONOMSKOJ ZAJEDNICI ZAPADNO-AFRIČKIH DRŽAVA Rezime: Određivanje efekata makroekonomskog okruženja na poreske prihode je od vitalnog značaja za svaku zemlju, a posebno za ekonomsku zajednicu koja teži harmonizaciji makroekonomskog okruženja i na kraju integraciji. Međutim, stepen u kojem agregatna proizvodnja, inflacija i nezaposlenost utiču na poreske prihode u ECOWAS -u je manje proučavan u literaturi. Stoga ova studija empirijski istražuje kako su poreski prihodi povezani sa odabranim makroekonomskim varijablama. Panel analiza podataka koristi se na setu podataka šest zemalja ECOWAS-a o poreskim prihodima, bruto domaćem proizvodu, inflaciji, nezaposlenosti, otvorenosti trgovine i kursu tokom 2005-2019. Wald i Hausman test pokazali su da je regresija fiksnih efekata odgovarajuća za studiju. Rezultati su pokazali da je inflacija pozitivno povezana s poreskim prihodima i statistički značajna na 5 posto. Jedinično povećanje inflacije dovelo je do 0,007 povećanja mere poreskih prihoda; ekonomski rast je takođe bio pozitivan i statistički značajan sa 5 odsto; jedinični rast BDP -a rezultirao je povećanjem varijable državnih poreskih prihoda za 0,78. Konačno, varijabla poreskih prihoda smanjena je za 0,10 sa jedinstvenim povećanjem nezaposlenosti. Preporučuje se da zemlje ECOWAS -a pažljivo upravljaju svojim makroekonomskim okruženjem kako bi povećale poreske prihode.

Ključne reči: Makroekonomske varijable, Poreski prihod, ECOWAS, Hausmanov test, Optimalna poreska teorija.

EJAE 2021, 18(2): 76 - 94 ISSN 2406-2588 UDK: 340.137:338.47(4-672EU) 346.545:621.391(4-672EU) 339.137.2(497.11) DOI: 10.5937/EJAE18-32186 Original paper/Originalni naučni rad

CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR Milan Kostić, Jelena Živković* University of Kragujevac, Faculty of Economics, Kragujevac, Serbia

Abstract: The electronic communications sector is very important sector of a national economy. It provides an opportunity for facilitating business activities and it is a good area for investment. A small number of operators and a large number of users operate in most markets of the electronic communications sector. The conditions under which electronic communications services are available to users in different countries diverge, among other things, due to the degree of liberalization of individual markets in these countries. The Republic of Serbia has significantly liberalized this sector of economic activity by adjusting it according to the regulations of the European Union. However, certain entry barriers have remained, primarily related to the licenses that operators need to obtain to operate in the Serbian market. The paper aims to investigate the level of concentration in selected markets of the electronic communications sector, namely: mobile telecommunication networks and services, broadband Internet services and media content distribution. The research showed a high level of concentration measured by the Herfindahl-Hirschman index in all three markets. A further increase in concentration is expected in all analyzed markets in the short term. These research results require more attention from regulatory bodies because they increase the possibility of some form of non-competitive behaviour by undertakings.

Article info: Received: May 11, 2021 Correction: September 17, 2021 Accepted: September 20, 2021

Keywords: electronic communications sector, market concentration, distortion of competition, competition policy.

INTRODUCTION The electronic communications sector is an important segment of a national economy because it is a very propulsive industry. This sector development can represent the level of national economic development but it also provides an excellent support for its accelerated growth. One should not dismiss the fact that the electronic communications sector is intertwined with the other economic sectors and they cannot operate and give the expected results if they are not supported by an adequate telecommunications network. 76

*E-mail: jelena.zivkovicfbb@gmail.com

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

It turns out that this type of service provides support for the permanent development of other economic sectors. Also, wider adoption and more intensive diffusion of modern telecommunication technologies have high influence on the competition in other markets (Jerbashian & Kochanova, 2017, p. 650). This sector is also attractive for investments due to its potential for long-term large profits. It is not surprising that the global telecom services market revenue was USD 1,516,909 million in 2019 and will reach USD 1,975,411 million in 2025. What is important to point out is that this sector is characterized by an oligopolistic market, namely a small number of undertakings of great market power operating. Western Balkans countries have entered the process of liberalization of the electronic services sector following the example of developed European countries due to the importance of this sector and their ambition to join the European Union. The liberalization process implies the introduction of free competition in this sector and free entry of new participants onto the market. The liberalization process of the electronic communications services market in the Republic of Serbia has begun with the adoption of the Law on Telecommunications in 2003, when the Republic Telecommunications Agency was established, which is now the Regulatory Agency for Electronic Communications and Postal Services (RATEL). The paper aims to analyse the degree of concentration in individual markets of the electronic communications sector of the Republic of Serbia and to investigate the trends present in these markets. Additionally, the paper aims to provide recommendations for regulatory authorities as to which actions to perform to provide better conditions for achieving free competition in the analysed markets. In addition to the introduction and concluding remarks, the paper contains three interconnected parts according to the defined goal. The first part of the paper refers to the literature review and previous studies that analysed the level of concentration in certain markets of the electronic services sector. Furthermore, an overview of the situation on the global market of electronic services, mainly mobile telecommunication networks and services is presented in this part of the paper. The second part of the paper presents the research methodology and data sources, while the third part presents the research results of the current market and the level of concentration in the selected ones.

CONDITIONS IN THE GLOBAL ELECTRONIC COMMUNICATIONS MARKET The electronic communications services markets are oligopolistic and are characterized by a small number of undertakings with high market power in most countries. The main characteristic of those markets is the supply of unified products and services by several operators whose business requires a license from the regulatory body. Obtaining the license is one of the most important entry barriers limiting the number of undertakings and keeping them at the low level. Based on this it can be concluded that this industry operates as a natural oligopoly. The most important and dynamic part of the electronic communications sector is the market of mobile telecommunication networks and services. Mobile telecommunications are an important generator of innovation and represents one of the most significant sources of revenue in the electronic communications sector. Furthermore, mobile telecommunications offer higher quality services and thus overcome the inefficiency generated by fixed telephony services thanks to the existence of modern technological solutions (Wellenius, 1993). A large number of studies have dealt with the analysis of the competition in this market. 77

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

The first studies related to the mobile communications market focused on the analysis of competition, mainly in duopoly markets such as the ones in the USA, Great Britain and Sweden for many years. Parker and Roller (1997) investigated the U.S. mobile phone market in the period from 1984 to 1988 and concluded that prices were significantly above those where there was an oligopoly. As one of the reasons why prices are higher, these authors also state the fact that participants perform in more connected markets, which further strengthens their market position and enables them to set higher prices. In his research, Valletti (2003) defined a strategic interaction model between operators in the mobile communications industry. Furthermore, this author addressed the question of how the characteristics of mobile communications and customer preferences affect the nature of competition and found that, in most cases, the mobile telecommunication market takes on the characteristics of a natural oligopoly. An analysis of the market share and market strength of undertakings in the mobile telephony markets in 49 European countries, including Serbia, showed that the most significant undertakings are those with the largest market share in most of the observed countries. Although the market share of undertakings can be determined in different ways, the dominant one is the one based on the number of users. The study showed that most mobile telecommunication markets are highly concentrated and that the two largest operators control more than half of the domestic mobile telecommunications markets (Whalley & Curwen, 2012). Sung (2014) examined the concentration of the mobile telecommunication markets in 24 OECD countries between 1998 and 2011 and recognized a positive relationship between market concentration, prices and profits, thus more concentrated markets provide higher prices and profit. This research is in accordance with the well-known SCP paradigm. This paradigm origins can be traced back to the middle of the twentieth century and it begins with the fact that the market structure, the level of its concentration and entry barriers have an impact on the behaviour of successful companies, namely the performance of those companies. Participants can be expected to behave in such a way as to increase the prices of their services and limit supply to make higher profits in markets characterized by high concentration and where significant entry barriers are. Valaskova et al. (2019) determined the existence of an oligopolistic market structure and a high degree of concentration in the Slovenian telecommunications market in the period from 2013 to 2017 using Concentration ratio, Herfindahl-Hirschman index, Gini coefficient and Lorentz curve. Furthermore, Madleňáková et al. (2018) in their study investigated the issue of determining the degree of concentration in the electronic communications sector using the Herfindahl-Hirschman index. Pejić Bach et al. (2013) analysed the degree of concentration in the telecommunications market in Croatia using Concentration ratio, Herfindahl-Hirschman index and Gini coefficient from 2003 to 2008. The study concluded that the concentration changes differently in diverse segments of the telecommunications market and that the degree of concentration in this market is heavily influenced by entry barriers. Krstić et al. (2016) measured the level of concentration in the mobile services market in the Republic of Serbia by sales revenue from 2009 to 2014 using some concentration indicators (Concentration ratio, Herfindahl-Hirschman index, Lorentz curve, Gini coefficient, Entropy coefficient, etc.). They concluded that there is a high level of market concentration in the mobile telecommunication networks and services and an oligopolistic market structure characterized by the change of two companies in a leading position. Also, there is a declining trend in the level of concentration. Kostić et al. (2016) also measured the level of concentration in the mobile telecommunications market in the Republic of Serbia using various concentration indicators such as Herfindahl-Hirschman index, Lorentz curve and Gini coefficient from 2008 to 2013. 78

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

The results have shown that there is a high degree of concentration in this market and the dominance of few undertakings. The correlation and regression analysis results confirmed a positive relationship between concentration and market performance of undertakings. Furthermore, a high level of market concentration was determined in the media content distribution market of the Republic of Serbia from 2007 to 2009 using Concentration ratio, Herfindahl-Hirschman index, Lorentz curve, Entropy coefficient and Horvat concentration index (Maksimović et al., 2011). Trifunović and Mitrović (2016) studied the externalities of the market of electronic services in Serbia from 2003 to 2014 and noted the high market concentration and relatively low prices. The conclusion is that regulatory policy should not be based solely on the degree of market concentration. More detailed information is needed, like the behaviour of undertakings and switching costs consideration, as well as entry barriers and potential price discrimination by economic entities. Particular attention should be given to the company’s partnership in R&D, which is fairly common in the high-tech industry. Although such cooperation reduces risks and pressures due to high-risk investments, there is a risk of the horizontally integrated companies turning into market cartelization (Stojanović el al., 2019, 149). To assess future responses competition policy, Stojanović and Kostić (2018) analysed the level of concentration in selected markets of electronic communications services of the Republic of Serbia using the Herfindahl-Hirschman Index from 2007 to 2016. They concluded that the concentration level is extremely high but with a declining trend. The analysis of market concentration in the electronic communications sector has gained importance in recent years due to the accelerated growth of this sector globally, especially the part related to mobile telecommunications. It is a very dynamic industry, which is constantly evolving thanks to new technologies and infrastructure. When observing current trends, it is interesting to note that in 2020 there were approximately 7.7 billion active mobile subscribers in the world, which is an increase of 3.3 billion in the last five years. When it comes to fixed telephony, where there are over 1.1 billion subscribers and which has grown by about 830 million subscribers in the last five years, we can say that communications are rapidly moving from fixed to mobile telephony. Graph 1. The number of mobile and fixed telephone subscribers from 2008 to 2018.

Source: Authors based on data ITU https://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx accessed on December 15, 2019.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Graph 1 shows trends in the number of subscribers of fixed and mobile networks globally, especially in developed countries and developing countries. There is a higher growth rate of mobile subscribers compared to fixed telephony subscribers, especially in developed countries. At the beginning of the observation period, there were about 30 subscribers per 100 inhabitants in mobile telephony in developed countries. At the end of the period, we see that this number reaches 115 per 100 inhabitants. The number of mobile subscribers per 100 inhabitants in 2018 is 61 in developing countries. It was an increase compared to the beginning of the period when that number was less than 10 subscribers. It is noticeable that the number of subscribers in developing countries is far below that which exists in developed countries at both the beginning and end of the period.

METHODOLOGY AND DATA SOURCES In accordance with the goal set for this paper, the conducted research relates to the level of concentration and dynamics of changes in selected markets of the electronic communications sector of the Republic of Serbia. The most promising markets of this sector were examined, which also have the largest share in the total revenue of the sector. The paper examines the following markets: mobile telecommunication networks and services, broadband Internet services and media content distribution. Market concentration indicators are very important for analysing competition conditions (Oliveira, & Oliveira 2018; p. 166). There are many market concentration indicators. The researchers opted for the Herfindahl-Hirschman index (HHI) as the most comprehensive and understandable index from the analytical point of view. The OECD recommendation is to use the Concentration Ratio (CR) and the Herfindahl-Hirschman index to determine the level of market concentration (OECD, 1993, p. 24–25). The Concentration (CRn) ratio is the sum of the market shares from several leading undertakings (the number can vary from 3 to 10 and usually ranges between 3 and 5 undertakings). Having in mind that this index represents the sum of individual market shares of the largest companies and does not take into account the dispersion of shares between them, it can be stated that its analytical significance is insufficient. The Herfindahl-Hirschman index (HHI) is taken as an alternative. The HerfindahlHirschman index is calculated by squaring the market shares of all undertakings (Dafny, et. al., 2012, p. 1162). The HHI takes account of the differences in the sizes of market participants (undertakings), as well as their number (Thembalethu, et al., 2019, p. 354). It gives particular importance to companies with a high market share while including in the calculation all companies in the market (Chiang-Ming et al., 2014, p. 147; Kostić, et al., 2016, p. 339; Ivanova et al., p. 2018, 35; Grosche et al., 2020, p. 79, Kruger, et al., 2021, p. 2). The Herfindahl-Hirschman index (HHI) is calculated as follows (Yeong-seok Ha & Jung-soo Seo, 2013, p. 261; Lieshout, et al., 2016, p. 72; Chih-Wen, 2016, p. 292; Lipczynski et al., 2017, p. 271, Barra & Zotti, 2019, p. 109; Kastratović, et al., 2019, 219; Nguyen, et al. 2020, p. 4): (1) namely: Si market share of i company and n total number of companies on the market. Value of the HHI ranges between 0 and 10,000 (alternatively 0 and 1). In the case of an atomized supply, where there is a large number of undertakings and the supply tends to equal 0, the index value also tends to equal 0. In the case of a monopoly, the HH index is 10,000 (or 1) (Yuan, et al., 2019, 476; Bakhtiari, 2021, p. 59). 80

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Table 1 provides reference values for this indicator, which are presented in Horizontal Merger Guidelines by the Department of Justice, and the U.S. Department of Justice and the U.S. Federal Trade Commission in 2010 (Yeong-seok Ha & Jung-soo Seo, 2013, p. 261; Brown, et al., 2020, p. 2). Table 1. HH index reference values HH index

Supply concentration levels

HHI < 1,500

Unconcentrated market

1,500 ≤ HHI ≤ 2,500

Moderately concentrated market

HHI > 2,500

Highly concentrated market

Source: U.S. Department of Justice and Federal Trade Commission, (2010), HorizontalMerger Guidelines, p. 18

In addition, the Horizontal Merger Guidelines set out the rules for interpreting the change in the value of HHI, as follows: (1) any change in HHI below 100 index points is considered a slight change, which does not affect competition and does not require further analysis; (2) if it is an unconcentrated market (index value below 1,500 points), any integration, regardless of the change in the value of the index, does not require further analysis, because it is unlikely to have adverse competitive effects; (3) in moderately concentrated markets (index value between 1,500 and 2,500 points), any increase in the value of HHI above 100 index points is a matter of concern and requires analysis; and (4) in highly concentrated markets (index values above 2,500) any increase in HHI between 100 and 200 index points raises considerable concerns regarding restrictions of competition and requires detailed supervision by regulatory bodies. Otherwise, in the situation of a highly concentrated market, any increase in the index above 200 points, which occurs as a consequence of the integration of undertakings, indicates a significant increase in market power of related undertakings and is subject to restriction by the regulatory body. Due to the simplicity of calculation and interpretation, the Herfindahl-Hirschman index is widely accepted by most developed market economies and regulatory bodies dealing with competition policy. In markets where there is a small number of undertakings with a high market share and a large number of undertakings with a negligible market share, which is difficult to identify, we use a modification of the HHI in the following form (Kostić, 2018, p. 165): (2)

namely: n number of undertakings with identified market share, m number of undertakings with unidentified market share, Si market share of identified i undertaking. In the research, we use the data by Regulatory Agency for Electronic Communications and Postal Services (RATEL) presented in the publication of An Overview of the Telecom and Postal Services Market in the Republic of Serbia at the following link https://www.ratel.rs/cyr/page/cyr-godisnjipregledi-trzista. The study covers the period from 2008 to 2019.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

RESEARCH RESULTS The electronic communications sector revenue ranged from 1.4 to 1.75 billion euros in Serbia during the observed period (2008-2019). The lowest revenue was recorded in 2010, while the highest revenue was achieved in 2019 - 1.75 billion euros, which is 4.1% more than in 2018 (RATEL, 2020, p. 6). Regarding the share of the electronic communications sector in GDP, it was 4.87% at the beginning of the period, and 3.8% in the end (2019). The data showed a decline in the share of this sector's revenue in GDP, although the revenue was measured in absolute terms at a relatively similar level throughout the observed period (Graph 2). Graph 2. The electronic communications sector revenue and share in the GDP of the Republic of Serbia from 2008 to 2019

Source: Authors based on RATEL data

During 2019, the largest share of total revenues in the electronic communications sector was generated from mobile telecommunications services. This market earns 59.8% of the electronic communications sector total revenue. If compared to the share in total revenues in 2008, it is clear that the share remained at a similar level. The most extensive change occurred in the fixed telephony market, where the share significantly decreased. The spillover was realized in favour of broadband Internet services and media content distribution. The share of those markets of the electronic communications sector total revenue in 2019 was 12.6% and 12%, whereas it was 7% and 4%, in 2008, respectively (Graph 3). It can be concluded that these two markets have the highest revenue growth rate in the electronic communications sector.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Graph 3. The revenue structure of the electronic communications sector of the Republic of Serbia in 2008 and 2019

Source: Authors based on RATEL data

When it comes to mobile telephony, as the most significant part of the telecommunications sector in terms of revenue in Serbia, it is a market that has all the features of an oligopolistic market structure. In 2019 there were three operators in the mobile communications market of the Republic of Serbia, namely Telekom Srbija a.d., Telenor d.o.o., and Vip mobile d.o.o. There have been two virtual operators in Serbia since 2016 (Mundio Mobile d.o.o. and Globaltel d.o.o.). In 2019 only Globaltel d.o.o. was represented on the market with a modest market share. As Figure 4 represents, Telenor has been taking the leading position by revenues since 2010, although the market leader was Telekom Srbija at the beginning of the observed period. The individual market share of the two leading operators has ranged from 36% to 43% since 2011. Telenor has the advantage regardless of the downward trend. Vip mobile is the youngest network operator, which has been operating in Serbia since 2006. What distinguishes Vip mobile from other operators is the multiannual accelerated revenue growth, thus managing to multiply its revenues from 2008 to 2019. Graph 4. Mobile operators’ revenues in the Republic of Serbia from 2008 to 2019

* Revenues in millions of euros Source: Authors based on RATEL data 83

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Graph 4 shows slight differences in revenues between Telekom Srbija and Telenor and indicates that Telenor is the market leader during the most part of the observed period. The situation is somewhat different if we analyse the operators market share based on the number of subscribers. Telekom Srbija is the leader by the number of subscribers in the mobile telecommunication market. According to the number of subscribers, Telekom Srbija has had a market share exceeding 40% during the entire observed period. In 2009 it was even 59.7%. The second-largest market share company is Telenor, with average market share of approximately 30%. Vip mobile has the lowest market share by the number of subscribers, but there is a constant growth in the number of users, as well as in revenues. Regarding the number of subscribers in 2019, Telekom Srbija market share was 44.3%, Telenor 31.4%, and Vip mobile 24%. (Globaltel market share was 0.03) in 2019. Graph 5 shows the market share movement in the mobile telecommunication market measured by the number of subscribers. Graph 5. Mobile telecommunications operators market share by the number of subscribers in the Republic of Serbia from 2008 to 2019

Source: Authors based on RATEL data

The difference between the market share by revenue and the market share by the number of subscribers is that Telenor is primarily focused on legal entities, while Telekom is oriented towards the individuals. Table 2 shows the obvious difference in operators market share by total revenues and by the number of subscribers.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Table 2. Operators market share by total revenues and number of users from 2008 to 2019 Telekom Srbija

Telenor

Vip mobile

Globaltel

Year

Total revenue

Users

Total revenue

Users

Total revenue

Users

Total revenue

Users

2008

52.15

58.93

41.83

31.94

6.02

9.13

2009

50.70

59.70

40.50

28.70

8.80

11.60

2010

43.70

56.00

42.40

30.30

13.90

13.70

2011

38.84

53.70

43.58

30.20

17.58

16.10

2012

38.10

45.80

42.40

33.90

19.50

20.30

2013

37.00

44.80

41.00

33.30

22.00

21.90

2014

36.81

44.56

40.29

33.27

22.90

22.17

2015

36.00

46.10

41.00

32.30

23.00

21.60

2016

36.70

46.80

39.90

31.20

23.40

22.00

2017

37.00

45.70

39.10

31.10

23.90

23.20

0.00

2018

37.10

45.00

37.20

31.60

25.68

23.20

0.02

2019

36.81

44.30

36.33

31.40

26.81

24.00

0.05

0.03

Source: Authors’ calculation based on RATEL data

Table 3 shows the calculation of HHI in the mobile telecommunication networks using the market share based on the number of users in the Republic of Serbia from 2008 to 2019. Table 3. Herfindal-Hirschman index on the mobile telecommunication market of the Republic of Serbia from 2008 to 2019 Year

HHI

2008

4,576.26

2009

4,522.34

2010

4,241.78

2011

4,054.94

2012

3,658.94

2013

3,595.54

2014

3,583.99

2015

3,635.06

2016

3,647.68

2017

3,593.94

2018

3,561.84

2019

3,524.54

Source: Authors’ calculation based on RATEL data 85

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Based on the data from Table 3, we can conclude that there is a very high supply concentration in the mobile telecommunication market of the Republic of Serbia. HHI values were high during the whole observed period; namely, when the value of HHI exceeds 2,500, the market is considered to be highly concentrated, which is the case here. Nevertheless, regardless of the extremely high values of the HHI, the encouraging fact is that the value of the HHI index is declining year by year. The trend extrapolation of the HHI index value (Graph 6) represents the trend curve corresponding to the presented data and one can see that HHI value is gradually reversing. A slight increase in the value of HHI can be expected in 2020 and 2021 (Graph 6). Graph 6. The HHI value trend on the mobile telecommunication market with the forecast for 2020 and 2021

Quadratic Trend Model Yt = 4928 - 294,4*t + 15,42*t**2 4750

Variable A ctual Fits Forecasts

4500

A ccuracy Measures MAPE 2,21 MAD 84,48 MSD 9116,87

HHI

4250 4000 3750 3500 08 09 10 11 12 13 14 15 16 17 18 19 20 21 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Year

Source: Authors’ calculation based on RATEL data

The next significant electronic communications market is the market for broadband Internet services, which has rapidly grown in recent years. A large number of households have constant access to the Internet, while the business is literally unimaginable without the use of information technology and the Internet. It can be said that its use has become an integral part of business life over the world and in Serbia respectively (Đorđević Boljanović et al., 2014, p. 17). Revenues from Internet services amounted to 8.76 million dinars at the beginning of the observed period, and they reached the amount of 26.20 billion dinars in the end of the period. The share of revenues from these services in the electronic communications total revenues was 7% in 2008, while it was between 12 and 13% in 2019 (Graph 7).

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Graph 7. Broadband Internet services revenue in billions of dinars and the share of this segment the electronic communications total revenue of the Republic of Serbia from 2008 to 2019

Source: Authors based on RATEL data

The broadband Internet services market is characterized by two large and many small undertakings. Telekom Srbija is the dominant operator in Internet services with a significant market share. The emergence of SBB and its expansion in recent years has led to Telekom gaining a significant competitor. SBB's market share is increasing, but we cannot say that Telekom market share is decreasing. Telekom has taken over some of the Serbian broadband Internet operators with a smaller market share, such as Copernicus Technology and Radius Vector, in the end of 2018. This action prevented the decline in market share. In addition to Telekom and SBB, other operators have significantly lower market shares (Table 4). Table 4. Operators market share in the broadband Internet market based on the number of subscribers from 2015 to 2019 Year

Telekom Srbija

SBB

Other operators

2015

46.00

21.00

33.00

2016

47.50

26.30

26.20

2017

44.17

32.28

23.55

2018

46.00*

33.00

21.00

2019

48.00*

35.00

17.00

Source: Authors’ calculation based on RATEL data

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

The Herfindahl-Hirschman index was applied for a more detailed concentration level examination in broadband Internet services. It can be noticed that HHI is continuously above the value of 2,500 points within the four-year period (available data). It implies a high market concentration with a tendency for further increase (Table 5). Table 5. Herfindal-Hirschman index on the Serbian broadband Internet services market from 2015 to 2019 Year

HHI

2015

2,746.00

2016

3,074.38

2017

3,101.31

2018

3,314.00

2019

3,624.00

Source: Authors’ calculation based on RATEL data

Media content distribution is another important market of the Serbian electronic communications sector regarding the number of subscribers and the realized revenue. This market has shown significant development dynamics in the last few years. The number of subscribers has grown significantly in the last few years, so the number of subscribers was 922,000 at the beginning of the observed period in 2008, while there were 2 million subscribers a decade later. Most subscribers use the media content distribution service within the service package, along with the broadband Internet service, which sufficiently makes the market analysis of this electronic communications services segment difficult. Graph 8 shows media content distribution revenues are constantly growing. There was a significant increase in revenue in the observed period, so revenue increased fivefold in 2019 compared to 2008. Revenue growth rates in certain years were extremely high; in 2008 the growth rate was 50%. Graph 8. Revenues in billions of dinars and the revenue growth rate from media content distribution from 2008 to 2019

Source: Authors based on RATEL data 88

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

There are about 80 undertakings in the media content distribution market in the Republic of Serbia. There were 81 in 2019. However, two undertakings with a high market share stand out, namely SBB and Telekom. SBB has about 50% of the market share during the entire observed period, and Telekom, with all related participants, such as Supernova, Copernicus technology, Radius Vector, etc., has about 38% market share. If Telekom is added to PTT Serbia's market share as another state-owned distributor, market share exceeds 40%. Around 10% of the market share belongs to operators that operate outside the system of SBB (United Group) and Telekom and PTT Serbia, which are in ownership of the Republic of Serbia. Table 6 shows the change in HHI values from 2008 to 2019. HHI index value is above 2,500 points throughout the period, indicating a high market concentration level. It is even higher if we regard all state-owned operators as one market participant (index value, in that case, would be 3,895.82 instead of 3,667.28 in 2019). Table 6. HHI values on the media content distribution market from 2008 to 2019 Year

HHI

2008

3,177.32

2009

2,931.74

2010

2,603.58

2011

2,857.92

2012

3,051.74

2013

3,141.64

2014

2,564.86

2015

2,905.21

2016

2,993.21

2017

3,588.83

2018

3,746.08

2019

3,667.28

Source: Authors’ calculation based on RATEL data

It should be emphasized here that we can expect a further increase in the market concentration and HHI value in the following period, and that may become serious issue. Graph 9 shows the HHI value trend, and the estimates for 2020 and 2021 based on the trend line that best corresponds to the presented data.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Graph 9. The HHI value trend in the media content distribution market with forecasts for 2020 and 2021 Quadratic Trend Model Yt = 3253 - 184,7*t + 19,39*t**2 4500

Variable A ctual Fits Forecasts

HHI

4000

A ccuracy Measures MA PE 6,0 MA D 180,6 MSD 42150,4

3500

3000

2500

08 09 10 11 12 13 14 15 16 17 18 19 20 21 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Year

Source: Authors’ calculation based on RATEL data

The trend curve indicates that a significant jump in the HHI value is expected in the following period (Graph 9). It requires additional caution from regulatory bodies and focusing on regular and detailed market conditions analysis.

CONCLUSION This study estimates the market concentration of individual electronic communications services markets in Republic of Serbia: mobile telecommunication networks and services, broadband Internet services and media content distribution. The research findings, obtained by calculating the values of Herfindahl-Hirschman index, indicate that a high market concentration exists in all analysed electronic communications services in Serbia. It is an oligopolistic market structure with a high market concentration level where there are two or three leading undertakings. This conclusion is in line with the results of previous research (Maksimović et al., 2011; Whalley & Curwen, 2012; Krstić et al., 2016; Kostić et al., 2016; Valaskova et al., 2019). As for the individual markets, mobile telecommunication networks and services have the best opportunity for some kind of concentration reduction. Although the trend line changes, based on data of the HHI value, it indicates a slight increase in market concentration in the following period. It is realistic to expect a continuation of the third mobile operator (Vip mobile) growth, as well as the operators of the new generation (virtual operators). This fact can reduce market concentration. A short-term market concentration increase is possible in this market, while we can expect a decrease in concentration in the long run. We can expect both short and long-term concentration increase in Serbian broadband Internet services and media content distribution. 90

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

The main characteristic of these two overlapping markets is two undertakings that dominate the market, whereas one is in the property of the United group, and the other two (Telekom and PTT Serbia) are the property of Republic of Serbia (these two can be considered as one participant, due to clear ownership structure). In the last few years, there has been intense rivalry between these two market participants, both trying to achieve market dominance. It is of particular concern that the activities against competitors involve deals prohibited by the competition law, like "predatory" pricing policy. It is necessary to regularly use sectoral analysis (such as this one) to prevent any form of competition distortion. It is an open warning for the regulators to pay more attention to these markets, especially regarding concentration increase and deteriorating competitive conditions. The study contributes to a better understanding of competition conditions in electronic communications services in Serbia. Findings provide novel insights into the literature and have implications for competition authorities, also. The concentration of these markets is high and more attention should be paid to the actions of undertakings, especially in case of mergers. Special problem is that undertakings participate in several markets, so their market power is even greater. For more detailed consideration of the competition intensity between undertakings on these markets further research can include this problem in research model.

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

REFERENCES Bakhtiari, S. (2021), Trends in Market Concentration of Australian Industries. The Australian Economic Review, 54(1), 57 –75. https://doi.org/10.1111/1467-8462.12393. Barra, C., Zotti, R. (2019). Bank Performance, Financial Stability and Market Concentration: Evidence from Cooperative and Non-cooperative banks. Annals of Public and Cooperative Economics, 90(1), 103-139. https://doi.org/10.1111/apce.12217 Brown, D. P., Zarnikau, J., Adib, P., Tsai, C. H. & Woo, C. K. (2020). Rising market concentration in Texas’s retail electricity market. The Electricity Journal, 33(8), 106848, https://doi.org/10.1016/j.tej.2020.106848 Chiang-Ming C., Kuo-Liang C., Langelett, G. (2014). How demand uncertainty and market concentration affect long-term price instability. International Journal of Hospitality Management, 37, 146-149. https://doi. org/10.1016/j.ijhm.2013.11.001 Chih-Wen Y. (2016). Entry effect of low-cost carriers on airport-pairs demand model using market concentration approach, Journal of Air Transport Management, 57, 291-297, https://doi.org/10.1016/j.jairtraman.2016.09.001 Dafny, L., Duggan, M., & Ramanarayanan, S. (2012). Paying a Premium on Your Premium?? Consolidation in the US Health Insurance Industry. American Economic Review, 102(2), 1161–1185. https://doi.org/10.1257/ aer.102.2.1161 Đorđević Boljanović, J., Vukašinović, J. & Veinović, M. (2014). Information technologies in knowledge economy. Singidunum Journal of Applied Sciences, 11(2), 11-19. https://doi.org/10.5937/SJAS11-4447 Grosche, T., Klophaus, R., & Seredyński, A., (2020). Market concentration in German air transport before and after the Air Berlin bankruptcy, Transport Policy, 94, 78-88, https://doi.org/10.1016/j.tranpol.2020.05.006 Ivanova, B., Barjaktarović, L., & Ivanov, I. [2018]. Uticaj investicionog potencijala banaka na dugoročno stambeno kreditiranje sistema odbrane. Oditor, 4(2), 27-38. Jerbashian, V. Kochanova, A. (2017), The Impact of Telecommunication Technologies on Competition in Services and Goods Markets: Empirical Evidence The Scandinavian Journal of Economics,119(3), 628-655. https:// doi.org/10.1111/sjoe.12183 Kastratović, R., Lončar, D, & Milošević, S. (2019). Market concentration and profitability: the empirical evidence from Serbian manufacturing industry. Zbornik radova Ekonomskog fakulteta u Rijeci: časopis za ekonomsku teoriju i praksu/Proceedings of Rijeka Faculty of Economics: Journal of Economics and Business, 37(1), 213- 233. https://doi.org/10.18045/zbefri.2019.1.213 Kostić, M., Maksimović, L., & Stojanović, B. (2016). The limitations of competition in the insurance markets of Slovenia, Croatia and Serbia. Economic Research, 29(1), 395-418. https://doi.org/10.1080/133167 7X.2016.1169703 Kostić, M. (2018), Teorijsko-metodološka analiza uslova konkurencije u antimonopolskoj politici, Kragujevac: Ekonomski fakultet Univerzitet u Kragujevcu. Kostić, Z., Stojanović, B. & Radukić, S. (2016). Measuring the Level of Competition on the Serbian Mobile Tel- ecommunications Market. Economic Themes, 54(3), 323–343. Krstić, B., Radivojević, V. & Stanišić, T. (2016). Measuring Market Concentration in Mobile Telecommunications Market in Serbia. Facta Universitatis, Series: Economics and Organization, 13(3). 247–260. Kruger, W., Nygaard, I., & Kitzing, L. (2021). Counteracting market concentration in renewable energy auc- tions: Lessons learned from South Africa. Energy Policy, 148(B), 111995. https://doi.org/10.1016/j. jmoneco.2021.04.005. Lieshout, R., Malighetti, P., Redondi, R., & Burghouwt G. (2016), The competitive landscape of air transport in Europe. Journal of Transport Geography. 50, 68-82, https://doi.org/10.1016/j.jtrangeo.2015.06.001, Lipczynski, J., Wilson, J. O. S. & Goddard, J. (2017). Industrial Organization: Competition, Strategy and PolicyLondon, UK: Pearson. 92

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

Madleňáková, L., Matúšková, M., Madleňák, R. & Droździel, P. (2018). Quantitative Analysis of the Competitive Environment in the Electronic Communications Sector. У I. Kabashkin, I. Yatskiv, & O. Prentkovskis (Eds.), Reliability and Statistics in Transportation and Communication (pp. 413–421). Springer International Publishing. Maksimović, Lj., Radosavljević G. & Borisavljević K. (2011). Concentration in the radio television program distribution market in the Republic of Serbia. Industrija, 39(3), 31–41. Nguyen, P.N., Woo, S. H., Beresford, A., & Pettit, S. (2020), Competition, market concentration, and relative efficiency of major container ports in Southeast Asia, Journal of Transport Geography, 83, 102653, https:// doi.org/10.1016/j.jtrangeo.2020.102653 OECD. (1993). Glossary of Industrial Organization Economics and Competition Law. Organisation for Economic Co-operation and Development. Oliveira, V. R. M. & Oliveira V. M. A. (2018). What drives effective competition in the airline industry? An empirical model of city-pair market concentration. Transport Policy, 63, 165-175, https://doi.org/10.1016/j. tranpol.2017.12.021 Parker, P. M. & Röller, L.-H. (1997) Collusive conduct in duopolies: multimarket contact and cross-ownership in the mobile telephone industry, The RAND Journal of Economics, 28, 304– 22. https://doi.org/10.2307/2555807 Pejić Bach, M., Zoroja, J. & Jirous, Z. (2013). Croatian Telecommunication Market: Concentration Trends in the Period from 2003 to 2008. Interdisciplinary Description of Complex Systems, 11, 131–142. https://doi. org/10.7906/indecs.11.1.11 RATEL. (2020). Pregled tržišta telekomunikacija i poštanskih usluga u Republici Srbiji u 2019. godini. Regulatorna agencija za elektronske komuikacije i poštanske usluge. Stojanović, B. & Kostić, M. (2018). Information and communication technologies product market and protection of competition in the Republic of Serbia. Ekonomika, 64(1), 1–12. https://doi.org/10.5937/ekonomika1801001S Stojanović, B., Ranđelović, M. & Vučić, V. (2019). Challenges of the competition policy in Serbia and other Western Balkan countries. Knowledge International Journal, 34(1), 147 -152. https://ikm. mk/ojs/index. php/KIJ/article/view/2099 Sung, N. (2014). Market concentration and competition in OECD mobile telecommunications markets Applied Economics, 46(25), 3037–3048. https://doi.org/10.1080/00036846.2014.920480 Thembalethu, B., Thando, M. & Liberty M. (2019). The extent of market concentration in South Africa’s product markets, Journal of Antitrust Enforcement, 7(3), 352–364, https://doi.org/10.1093/ jaenfo/jnz014 Trifunovic, D. & Mitrovic, D. (2016). Network externalities in telecommunication industry: An analysis of Serbian market. Industrija, 44, 63–87. https://doi.org/10.5937/industrija1-8886 U.S. Department of Justice and Federal Trade Commission, (2010), Horizontal Merger Guidelines. Valášková, K., Ďurica, M., Mišanková, M., Gregova, E. & Lazaroiu, G. (2019). Oligopolistic Competition among Providers in the Telecommunication Industry: The Case of Slovakia. Administrative Sciences, 9, 1–15. https://doi.org/10.3390/admsci9030049 Valletti, T. M. (2003). Is Mobile Telephony a Natural Oligopoly? Review of Industrial Organization, 22(1), 47–65. https://doi.org/10.1023/A:1022191701357 Wellenius, B. (1993). Telecommunications: World Bank experience and strategy, Washington D.C, USA: The World Bank. Whalley, J. & Curwen, P. (2012). Incumbency and market share within European mobile telecommunication networks. Telecommunications Policy, Elsevier, 36(3), 222–236. https://doi.org/10.1016/j.telpol.2011.11.020 Yeong-seok Ha, Jung-soo Seo, (2013). An Analysis of Market Concentration in the Korean Liner Shipping Industry, Asian Journal of Shipping and Logistics, 29(2), 249-266. Yuan, J., Zhou, Z., Zhou, N. & Zhan, G. (2019). Product market competition, market munificence and firms’ unethi- cal behavior. Chinese Management Studies, 13(2), 468-488. https://doi.org/10.1108/CMS-06-2018-0569 https://www.itu.int/en/ITU-D/Statistics/Pages/stat/default.aspx accessed on December 15, 2019. https://www.ratel.rs/cyr/page/cyr-godisnji-pregledi-trzista accessed on February 23, 2020. 93

EJAE 2021  18 (2)  76 - 94

KOSTIĆ. M., ŽIVKOVIĆ. J.  CONCENTRATION OF SUPPLY ON THE CHOSEN MARKETS OF SERBIAN ELECTRONIC COMMUNICATIONS SECTOR

KONCENTRACIJA PONUDE NA ODABRANIM TRŽIŠTIMA SEKTORA ELEKTRONSKIH KOMUNIKACIJA SRBIJE Rezime: Sektor elektronskih komunikacija je veoma značajan za funkcionisanje nacionalnih ekonomija. On pruža mogućnost za lakše odvijanje poslovnih aktivnosti i predstavlja dobro područje za investiranje. Na većini tržišta sektora elektronskih komunikacija funkcioniše mali broj operatora i veliki broj korisnika. Uslovi pod kojima su dostupne usluge elektronskih komunikacija razlikuju se između ostalog i zbog stepena liberalizacije pojedinačnih tržišta u konkretnim državama. Republika Srbija je značajno liberalizovala ovaj sektor ekonomske aktivnosti usklađujući ga sa regulativom Evropske unije. Ipak ostale su izvesne ulazne barijere vezane, pre svega, za licence koje operateri treba da pribave kako bi poslovali na tržištu Srbije. Rad ima za cilj da istraži nivo koncentracije na odbranim tržištima sektora elektronskih komunikacija i to: mobilne telefonije, širokopojasnog pristupa Internetu i distribucije medijskog sadržaja. Istraživanje je pokazalo visok nivo koncentracije meren Herfindal-Hiršmanovim indeksom na sva tri tržišta. Na svim analiziranim tržišta se očekuje dalje povećanje koncentracije u kratkom roku. Ovakvi rezultati istraživanja zahtevaju veću pozornost regulatornih tela jer povećavaju verovatnoću nekog oblika nekonkurentnog ponašanja učesnika na tržištu.

Ključne reči: sektor elektronskih komunikacija, tržišna koncentracija, narušavanje konkurencije, politika zaštite konkurencije.

EJAE 2021, 18(2): 95 - 126 ISSN 2406-2588 UDK: 005.334:336.71(64) 330.43:336.774 DOI: 10.5937/EJAE18-30597 Original paper/Originalni naučni rad

THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS Ahmed Khattab*, El Khlifi Imad Abdelmamek Essadi University, Morocco

Abstract: The subject of the economic and social inequalities growth has become a major concern for economic researchers around the world. As a result, there is a great attention and concern about the significant costs of rising inequalities on peace and social coherence in society. To counteract the harms of these inequalities, Morocco has undertaken several reforms, including the implementation of a new development model set up in 2018. The main objective of this article is to estimate the impact of bank credit growth on the increase of economic and social inequalities in Morocco. In other words, we have verified whether there is a positive impact on the rise of income inequalities in Morocco. For this reason, in this study, we tested an econometric model, using the cointegration method, particularly, the error correction model. Thus, our results confirms that the degree of trade openness, bank credit and gross domestic product per capita are considered to be determinants of the equilibrium of the Gini index in the long run. We used annual data covering the period 1990-2019 from the Central Bank of Morocco (BAM) database and the World Bank database (WBD). A wide range of studies demonstrate the significant positive impact between bank credits and income inequalities.

Article info: Received: February 16, 2021 Correction: March 31, 2021 Accepted: April 15, 2021

Keywords: Income inequalities, bank credits, economic growth, trade openness, cointegration.

INTRODUCTION In recent years, researchers have given more importance on the issue of global wealth distribution. We are all convinced today that natural laws, as they are described in the capitalist system; where private actors own and control goods in accordance with their interests, and supply and demand freely set prices in self-regulating markets optimally for society. In reality, however, these laws do not work effectively as liberal theorist’s postulate. It is thus understood that these laws exert a fatal influence creating an order of things that is advantageous only to capitalists, and that the distribution of wealth is becoming increasingly unequal in both developed and underdeveloped countries. *E-mail: akhattab@uae.ac.ma

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Indeed, financial globalization through the mobility of capital, logically leads to another sharing between profits and wages. Competition from countries with low wages or social protection has a fatal influence on state policies to create an order of things that is advantageous only to the few rich, either by influencing wage negotiations or by redistributing taxes. Thus, favoring capital responds to the concern to preserve the capacity of firms to invest and puts the vast majority of the population at a disadvantage against the domination of the rich. In a word, this leads to an unequal growth of income, wealth and living conditions. From an empirical point of view, we find that several authors have attempted to verify empirically the existence of a relationship between, on the one hand, the dynamics of income inequality and, on the other hand, the expansion of credit. Income inequality on the higher and continuous level will create a significant social cost. It will weaken the choices of education, healthcare, and occupation. It will also cause other social problems such as corruption, nepotism, criminal, and many others Stiglitz (2012). The increase in income inequality, accompanied by low purchasing power of the affected households, has in fact occurred in parallel with a rapid growth in private indebtedness, which can be explained as a response by households suffering from economic insecurity and seeking to improve their quality of life or simply to maintain it at an acceptable level. Through this study we try to participate in the debate on the link between finance and income inequality by modeling the impact of the bank credit growth in Morocco, especially during the period of real estate growth, on the increase in income inequality and the decrease in household purchasing power, through macro-econometric modeling. In this sense, our problem can be formulated as follows: What is the impact of the bank credit growth on the increase in inequalities in Morocco? To answer this question, we based on existing concepts in the literature, we are expecting four possible scenarios regarding the impact of bank credit on income inequality: i) The bank credit growth has a positive impact on the increase of income inequality in Morocco, ii) The Bank credit growth has a negative impact on the rise of income inequality in Morocco. Methodologically, the method used to analyze the data in this study was the Vector Error Correction Model (VECM) Method. In order to increase the explanatory power of financial development on inequality, we use the following as control variables; income inequality (Gini index) in Morocco, traded openness index, financial development (private credit as a ratio of GDP) and real per capita GDP growth. We find a positive and statistically significant relationship between bank development and the growth of the Gini coefficient, suggesting that improvements in banking sector may increase income inequality in emerging economies. The paper is structured as follows: The next section briefly reviews the pertaining theoretical concepts and relevant empirical studies on the relationship between financial development and inequality. In section 3 discusses the methodology and presents our results, which are interpreted in section 4. Section 5 draws conclusions and offers policy recommendations.

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

LITERATURE REVIEW

Defining inequalities is a very complex operation, given those inequalities are constructed in a relationship between individuals through differences that may exist on several levels, for example biological, social or economic. However, these differences are not sufficient for them to become inequality; they must therefore be accompanied by unequal access between these individuals to certain resources that are valued because of their difference. That is to say that differences in color or gender, for example, are not inequalities only when they are taken as an obstacle to access to socially hierarchical goods and services. Inequalities are therefore defined in the Social Science as a difference judged unfair in access to valued resources including all possibilities of human action: political, economic, cultural, social, etc. For the sociologist Zamora (2019), two opposing conceptions of the response to inequality are opposed, "A conception limited to effects, and therefore focused on strict income disparity, leads to increase equality by reducing the monetary gap between rich and poor. The result would be a world where economic competition would still be ruthless, but where no one would fear material deprivation. A world that none of the socialist thinkers of the nineteenth century could have imagined, so strongly did they associate inequality with the problem of economic liberalism". A second conception seeks to achieve equality through the demarcation and democratization of goods such as health care, education, transport, energy, etc. A world that, by socializing and guaranteeing access for all to the most important elements of our existence, would reduce dependence on the market, and thus on the mechanism that is at the origin of inequalities. For a long time, this project was not considered scandalously utopian, even by the most moderate reformers. For Polymnia Zagefka (2009), slavery is "as an emblematic form of inequality, the negation of all identity to the other". Formally, five main orders of inequality can be distinguished: ◆ Political inequalities; ◆ Legal inequalities; ◆ Economic inequalities; ◆ Social inequalities (in which gender inequalities are often classified; health inequalities which sometimes also have political, religious and legal origins); ◆ Ecological inequalities. Dynamic aspects (development, reproduction, or reduction of inequalities) can also be taken into account. So, it is a multi-dimensional phenomenon that affects several domains, although the question of inequalities is often reduced to income and, more broadly, it extends to education, employment, health, and leisure.

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Figure. Financial development and social inequalities.

Source: Author’s research.

Economic growth and income inequalities The question of the distribution of wealth was not always a subject of debate since the ancient economy was very largely based on slavery and the question did not arise, in which a small part of the population represented by the owners, who had a divine right, could take the largest share of the wealth without any objection from the others. It was in the Middle Ages and with the birth of Islam that we have given more importance for the individual, which is explained by the gradual disappearance of slavery, and that the idea of the value of labor power developed, giving a little balance to society. On the other hand, the application of Islam has been a way of dealing with the systematic discrimination against women, religious minorities and, on the other hand, on a policy of redistribution to the lower social classes "Azzakat" which takes place through compulsory levies on household wealth, whether granted in monetary and then redistributed to those in need. In this regard, Tohirin et al., (2019), Mohamad et al., (2020) find that Islamic banking financing has potential to reduce income inequality. In the 19th century, economists believed that the dynamics of private capital accumulation led to an ever-increasing concentration of wealth and power in the hands of a small social group. For David (1815), during a period when economic activity was based on the cultivation of foodstuffs such as wheat, the main concern was the long-term evolution of land prices and the level of land rent. 98

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Moreover, from the moment when population and production growth is prolonged over time, we notice that because of the law of supply and demand, land tends to become increasingly scarce and expensive, as do the rents paid to landowners. Thus, an increase in the number of workers makes the demand for labor increasingly greater for identical supply, hence the tendency for wages to fall to the limit where the level of the wage allows simple survival. In the end, therefore, landowners will receive an increasingly large share of the national income, and the rest of the population which is an increasingly smaller share, and this situation will threaten the social balance and give rise to inequalities. For Ricardo, the only solution to get out of this situation will be heavier taxes on land rents. For Marx et al.,l (1936) the question was very obvious in his thinking, as he believed that the economic laws specific to the mode of production of the old regime allowed the appearance of a new social class, which was the bourgeoisie represented by the industrial capitalists. The interplay of these economic laws leads to a strengthening of the power of the bourgeoisie, and an urban misery of the industrial proletariat that will be more extreme than the rural misery. Indeed, as a result of industrial growth, a huge rural exodus to industrial zones will make the working days longer for the proletariat and wages increasingly low. It was in this context that the communist and socialist movements developed, which demanded the concentration of industrial capital accumulating without a limit, so that a small social group would appropriate an ever-increasing share of production and income. Thus, capitalism mechanically produced inequalities that continued to worsen. Towards the 20th century, in the Glorious Thirties, a new theory of Kuznets (1955) appeared, according to which income inequalities were in fact spontaneously destined to decrease in the advanced phases of capitalist development, and then to stabilize at an acceptable level. According to Kuznets (1955) growth is a rising wave that carries all the boats, it is in fact a balanced growth path, in which all social classes progress at the same rate, and each social group benefits from growth in the same proportions without considerable differences, unlike the inequalities of Ricardo & Marx. This is an optimistic theory according to which inequalities would follow a bell-shaped curve, that is, increasing and then decreasing, where the first stage of industrialization is characterized by a natural growth of inequalities, after which it will decrease as an increasingly large fraction of the population that did not benefit from the new wealth brought by industrialization in the first stage, will take its share of the wealth, thus reducing inequalities. This theorem influenced the world considerably until the 1990s. Since the 1990s, we notice that inequalities have reached a record level in most northern and southern countries, even in emerging countries that have experienced very strong growth in recent decades, such as China, India, and Brazil. This worsening of inequality now occupies a primordial place in world current affairs and calls into question the character of the balanced growth path described by Kuznets (1955). Empirical studies also provide evidence on whether the linkage between growth and income inequality. For example, Woo (2011) introduced fiscal policy volatility as a new channel to explain the negative link between inequality and growth. Amri (2018), Binatli (2012), Neves et al., (2016), Malinen (2013), Herzer & Vollmer (2012) also find a negative and significant relationship between the economic growth and income inequality in the long-run. In a theoretical paper, Majeed Region (2010) provided a unified theory that combines both contradictory approaches, by analyzing the linkage at the global level for some developing countries in Asia, the authors suggest a positive and significant relationship between growth and inequality. They stated that inequalities and growth are positively correlated at earlier stages, due to the main role played by physical capital accumulation in economic growth. Furthermore, Huang et al., (2015), using the annual state-level panel data of United States, also proves that growth volatility is positively and significantly associated with higher income inequality. 99

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Finally, Stewart & Moslares (2012), Dabla-Norris et al., (2015), and Saez & Zucman (2016) also demonstrate that income inequality has a negative impact on growth. They perform their work analyzing Indian states for the period 1980-2010, and conclude that the regional coefficient of Gini has a negative impact on growth.

Finance and income inequalities The imbalance in access to financing is a major finding that proves that we do not have the same opportunities between different social classes. While we notice that the richest people take on debt at low cost and invest in more lucrative investments, on the contrary, the middle classes take on debt at higher interest rates for the sole purpose of maintaining their standard of living. In addition, it is the rich who still hold these credits through securitization. Thus, either the poor classes who cannot access bank loans or the middle classes who will also be victims, finance plays a role that aggravates income inequality. In theory, financial development should in the first place benefit the most modest, by relaxing this constraint and improving the allocation of capital, thus reducing inequalities. This idea of a linear and positive relationship between financial development and inequality was first widely accepted in the economic literature, particularly in development economics. Neil & Goldstein (2015, Jakob & Jan-Egbert (2016), Olivier Godechot (2016), Allen, Todd & Wallace (2018), Haan & Sturm (2017), Tan & Law (2012) summarizes this apparent consensus, "The results indicate that finance has a disproportionately positive impact on the poor and contributes to rising income inequality. Thus, according to Seven & Coskun (2016) find that neither banks nor stock markets play a significant role in poverty reduction. Several other studies support this result Muhammad, Bhattacharya & Mantu (2017) a 1% increase in financial development increases income inequality by 0.09% in the long run and in short-run estimate of 0.047, Mookerjee & Kalipioni (2010), Jeanneney & Kpodar (2011), Jakob & Jan-Egbert (2017), Jeong & Kim (2018) show that financial development increases inequalities in urban (rural) areas. The relationship between financial development and inequality could be non-linear Kim & Lin (2011), Bayaran & Yilmaz (2017), Thorntona & Tommaso (2020) thus show that financial development reduces inequality only beyond a certain threshold of depth of financial markets, while Tan & Law (2012) from another database of inequality also find non-linearity, but in the opposite direction to Kim & Lin (2011). For them, financial development reduces inequality at low levels of financial development, but tends to increase it at higher levels. As a note of Chen & Kinkyo (2016), Hyde & Wallace (2018), the detrimental effect of financial development on income inequality can be associated with susceptibility to financial crises and the quality of governance. They show that the low frequency of crisis and good governance can soften this detrimental effect. Therefore, it is worth investigating the relationship between financial development and income inequality at different levels of institution. Another recent breakthrough in the belief that financial development is systematically beneficial to inequality reduction is the analysis of Denk & Cournède (2015) for various periods between 1974 and 2011, they conclude that the expansion of the financial sphere (represented by private credit, the share of value added from the financial sector and the size of stock markets) has fueled income inequality in OECD countries. 100

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Also, in a panel of 59 countries for the period 1985–2015, Anthony & Kwon (2017), Yi & Chien (2019) find a positive relationship between financial development and income inequality for low-income countries. Another dimension to be taken into account concerns the effect of high wages on inequality, Philippon & Reshef (2012) show that employees in the financial sector in the United States enjoy wages that are 50% higher than in the non-financial sector, with equal characteristics. This could explain between 15% and 25% of the increase in inequality observed since the early 1990s. Denk & Cournède (2015) shows a similar effect for the European Union, but with a smaller amplitude (28% higher wages in the financial sector on average). In a study of some African countries, Batuo et al., (2010) find a linear negative relationship between financial development and inequality in their study on African countries in the years from 1980-2004. In addition, in a panel of 62 countries for the period 1973 to 2005, Agnello et al., (2012), (Delis et al.,(2014) find that financial reforms, measured by the aggregate financial index due to Abiad et al., (2010), is associated with less income inequality. In addition, their study also suggests that the effect on income distribution varies across liberalization policies. Especially directed credit and removal of excessively high reserve requirements seem important for reducing income inequality. Yet, other financial liberalization policies, such as privatization, reducing entry barriers and increases in international capital flows do not affect income distribution. Moreover, Rewilak (2013), "I examine Whether the revenues of the poor are systematically increasing with an average income, and if financial development improves the poorest Quintile revenues ", using cross-country regression analyzes. Rewilak (2013) concludes that on economic growth benefits universally with rich and poor, Financial development does not necessarily take poverty in all regions. In this regard, Rewilak (2013) attracts particular attention to South Asia and Latin America as two opposite examples. Although the results show positive effects of financial development at the poorest quintile income in South Asia, the opposites the case for Latin America. Lower results suggest a negative effect of low Sahara Africa's income funding and a positive impact in Eastern Europe and Central Asia. Kunieda et al., (2014), suggest that an increase in financial depth improves income distribution only in closed economies. By employing data for the period of 1975-2011. (Abosedra et al., (2016) find that financial development reduces poverty when domestic credits to the private sector is used as proxy for financial development. In this respect, Uddin et al., (2014) Examine short and long-run relationships between financial development, economic growth and poverty reduction in Bangladesh. The authors show that a long-run relationship exists between these variables. More recently, Park & Shin (2015), Chiu (2019) examines the relationship between financial development and income inequality. Their results indicate that financial development helps reduce inequalities to a point, but since financial development goes further, it contributes to greater inequality. In this context, the banking sector seems to exert stronger influence on income inequality Gimet & Lagoarde-Segot (2011). While an analysis of the strong presence of financial sector employees at the top of the income distribution is interested in itself, it provides at the same time one channel behind the negative relationship between finance and income equality established in Denk & Cournède (2015). The inequality-narrowing hypothesis puts forward the idea that when the financial sector grows, the poor, who were previously excluded from obtaining loans, might gain access Mookerjee & Kalipioni (2010), Jalil & Feridun (2011), Hamori & Hashiguchi (2012), Amine & Samir (2015), Manish & Colin Reilly (2019), Ibrahim & Tidjani (2020), Fligstein & Goldstein (2015), Shahbaz & Mahalik (2017). 101

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

In contrast, other studies predict that financial development may fail to reduce income inequality and poverty. Sehrawat & Giri (2015) investigate the finance-inequality nexus in India for the period 1982-2012, and suggest that financial development aggravates the income inequality in both longrun and short-run. Stiglitz (2013) contends that the financial sector has contributed so powerfully to inequality in the US through several channels. The author underlines that while financial firms pursue their own benefits via several rents seeking channels, inefficient regulation/supervision/enforcement framework and regulatory capture have also played roles, with consequences for distribution. Financial crises will have diverse effects on income distribution and inequality. Bourguignon states that the consequences of financial crises depend on the size of the financial sector, which is directly correlated to the severity of the crisis. The 2007-2008 crisis also allowed progress to be made in understanding the distributive effects of crises, which may pass through more indirect channels linked to fiscal policies. Jenkins et al., (2012) study country-by-country effects within the OECD. They find few direct effects on income distribution in the two-year following the crisis (2007-2009). Nevertheless, it seems that social protection systems and counter-cyclical policies implemented at the beginning of the crisis played a major role in this result. Fiscal consolidation policies certainly had a greater anti-distributive effect. This intuition is confirmed by Ball et al., (2013), who show a very clear positive correlation between the level of inequality and these consolidation policies. The latter also reduce the share of labor in value added and increase the level of long-term unemployment. Woo et al., (2013) evaluate the effect of a fiscal contraction of 1 percentage point of GDP at a [0.4%-0.7%] increase in the Gini coefficient in the first two years. 15% to 20% of this increase in inequality would be explained by the rise in unemployment directly attributable to the policies of fiscal austerity. Finally, Jauch & Watzka (2016) in an empirical analysis of 138 countries over the period 1960-2008, find a positive effect of financial development on inequality, according to their estimates, 10% increase in credit would increase the Gini coefficient by 0.23.

Trade openness and income inequalities According to Stiglitz, international trade is a game positive-sum, the absence of full employment and free trade can lead to job destruction at a faster rate than job creation, these limited fruits of trade openness are increasingly unevenly distributed, and the distribution of wealth becomes more unequal as the bargaining power of employees is reduced, Unskilled workers are the ones who benefit the least from it. Socials support measures are therefore necessary to compensate for the negative effects of free trade according to the study by Azevêdo, the opening of trade is solely responsible for the rise in inequalities, 80% of jobs have been destroyed in recent years due to technological progress. According to Romer, trade is not alone responsible for the rise in inequalities. He takes the example of the United States and Denmark, which have been confronted since the 1990s with the same economic changes but have not experienced the same changes in terms of inequalities. Inequalities have decreased in Denmark (the Gini coefficient has gone from 31% to 21%) while they have increased in the United States (the Gini coefficient has risen from 43% to 47%). The failure of our laws and public policies, which instead of mitigating the process of rising inequality in periods when market forces would have led to growing gaps between the rich and the rest, shows that the State has not tempered the excesses of the market. 102

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

However, higher rates of trade between countries, partly activated by improving technology often lead to income inequality Norris et al., (2015). Technology can reduce transportation and communication costs, and the degree of automation is increasing. Trade opening makes way for economic growth in developed and developing countries. Nevertheless, this also increases the income inequality rate. This fact is due to the imbalance between technical proficiency and other aspects of the growth process. Krugman (2005) noted out that owners with a many number of factors of production in a country can benefit from trade, but owners with many factors of production suffer losses due to international trade if a country has a large amount of capital. Trade openness will increase income inequality, and if there is a large amount of labor, trade openness will reduce income inequality Asteriou et al., (2014). Internal forces result in market forces that drive, for example, the balance between supply and demand, and that manages competition, or in general the fundamental rules of the game of social life. It is clear that the contribution of these market forces in the rise of income inequality is approved, because the rich have the power to edict these rules in their favor, while at the same time not giving others any chance to balance the game or make it fair.

DATA AND METHODOLOGY In order to assess the impact of bank credit on inequalities, it is necessary to explain the model to be estimated, the introduced variables, and the estimation method, before presenting the econometric results. In this section, we will first present the explanatory variables and the dependent variable of our econometric model, the sample and the frequency of observation of the data, and finally, we will graphically present the descriptive statistics of the variables. The method used to analyze the data in this study is the Vector Error Correction Model (VECM). It is a restricted VAR model, used for non-stationary variables, but has co-integrated potencies. After testing the model, it is recommended to enter the cointegration equation into the model used. On time series data mostly has stationary on the first difference or I (I). Then, VECM uses that co-integrated restriction information into its specification. Therefore, VECM is often known as VAR design for the non-stationary series that has a co-integrated correlation. Furthermore, in this econometric model, there is a speed of adjustment from short term to long term. The analysis tool provided by VAR/ VECM had done through four kinds of usage, such as forecasting, Impulse Response Function (IRF), Forecast Error Variance Decomposition (FEDV), and Co-integration tests Johansen (1988), Johansen and Juselius (1990) confirmed the presence of potential long run equilibrium relationship between two variables. We used Johansen’s technique in order to establish how many cointegration equations exist between variables. Our test suggests that our set of cointegrated time series have an error-correction representation, which reflects the long run adjustment mechanism. The diagnostic tests are lastly performed to examine serial autocorrelation, normality in error distribution and stability of the model.

103

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

In this study the error correction model as suggested by Hendry (1995) has been used. The general form of the VECM is as follows: (1)

Where is the first difference operator; is the error correction term lagged one period; is the short-run coefficient of the error correction term ; and is the white noise. The error correction coefficient ( ) is very important in this error correction estimation as the greater co-efficient indicates higher speed of adjustment of the model from the short-run to the long-run. The error correction term represents the long-run relationship. A negative and significant coefficient of the error correction term indicates the presence of long-run causal relationship.

The Empirical Model: For this study, we estimated the following model: Referring to the work of Li et al. (1998), Beck et al. (2007), Kpodar and Jeanneney (2008), Bazillier et al. (2017), the basic model is written as follows: (2) Where represents income inequality, designed by the statistician Corrado Gini (1912), it measures the degree of deviation of the income distribution within an economy from a perfectly equal distribution, represents the private credit, is trade openness index, and is real income per capita.

Statistical analysis The purpose of descriptive statistics is to structure and represent the information contained in the data. It includes many techniques: central tendency indicator, dispersion indicator and shape indicator. The following table presents the descriptive statistics for each variable. The descriptive statistics for these variables are given in the following table: Table 1. Summary statistics of variables Variable

Mean

Min

Max

Gini

39.98

0.6187

38.20

40.72

0.1578

0.0974

0.0456

0.288

65.43

14.12

47.1

87.99

GDPP

2440.6

586.13

1704.7

3396.1

Observations

Source: Authors, results obtained from the software (GRETL).

104

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Econometric analysis:

- Stationarity test of the variables According to the graphical representation of the variables (Figure. 3), it can be seen that all the variables studied are not stationary, which is why the Gini index, bank credits, trade openness and gross domestic product per capita are not integrated in level. The analysis of stationary is the first step before estimating our econometric model; it consists of verifying the order of integration of the variables used. The search for stationarity was carried out using Dickey and Fuller's tests on unit root search according to the strategy proposed by Harris (1995). The method applied makes it possible to analyze the level of stationary and the existence of cointegration between all the variables. Table 2. ADF Unit Root Test At Level Variables

At First difference I(1)

Constant and trend

Constant

None

Constant and trend

Constant

None

GINI

0.9999

0.9851

0.2690

0.2450

0.0000

I(1)

0.5986

0.7478

0.0516

0.1988

0.0470

0.0378

I(1)

0.1801

0.9397

0.9996

0.0000

I(1)

GDPP

0.1400

0.8348

0.9485

0.0000

I(1)

Source: Authors, results obtained from the software (GRETL).

According to this table, we notice that the results obtained for the level variables indicate that the series are not stationary at the 5% threshold, we also notice that the variables are integrated of order one I (1). Therefore, this shows that these variables are stationary in first differences, we notice that the trend is not significant for the four time series retained. The test is based on a model containing constants. In all cases, the observed value is lower than the significant value, which means that hypothesis I (1) is retained. Subsequently, we tested the existence of two unit roots to avoid the problems reported by Dickey and Pantula (1987). After the first distinction of each series, the tests are carried out on a model containing neither constants nor trends. In all cases, we reject the hypothesis of a series I (2). Consequently, we can deduce that the four time series retained are integrated of order 1.

The optimal number of delays Selection Criteria: Before proceeding with Johansen's multivariate cointegration test, one step is preliminary. The magnitude of the lags in each series are determined according to the BIC criterion of Schwarz et al. (1978), the calculation of the information criteria, AIC, SC, and HQ for lags ranging from 1 to 5, show that the majority of the criteria indicate that the number of lags to be retained is P=5.

105

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Table 3. Results of the selection criterion Lag

AIC

-9.069678

-8.874658

-9.015588

-13.32622

-12.35112

-13.05577

-14.40164

-12.64646

-13.91483

-13.98841

-11.45315

-13.28524

-14.74114

-11.42580

-13.82161

-18.24595*

-14.15053*

-17.11006*

Source : Authors, results obtained from the software (GRETL).

The delay that minimizes the three criteria AIC, SIC and HQ is -18.24 and corresponds to P = 5. Thus, we obtain the VAR (5).

Eigenvalues of the characteristic polynomial: Figure 2. Inverse roots of characteristic polynomial in the VAR model

Source: Authors, results obtained from the software (GRETL).

The VAR model (5) is stationary because all characteristic roots lie inside the unit circle. As a result, the system is stable and converges towards its long-term equilibrium, which is confirmed by the Student test associated with the calculated "t" of the model parameters are in absolute value greater than 1,96 which represents the value tabulated at the 5% threshold.

106

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Johansen cointegration test:

Johansen's approach consists of a cointegration test based on trace analysis ( ) and maximum eigenvalue analysis (maximum Egenvalue ( )). In addition, this model is estimated in order to determine the existence of a long-term equilibrium relationship between all the variables integrated into the econometric model. tests the hypothesis of a number of cointegrating relationships at most equal to “r” against an alternative hypothesis of “r + 1” of cointegrating relationships while tests the hypothesis of at most “r” cointegrating relationships against an alternative hypothesis of at least “r + 1” vectors. The test works by exclusion of alternative hypotheses, i.e., the null hypothesis is tested first: r = 0 against the alternative hypothesis r > 0 (r being the number of cointegrating relations). If it is accepted, the procedure stops, there are no cointegrating relations, if not, we move on to the next step by testing r = 1 against r > 1. The results of the statistics and the employee to determine the level of cointegration are presented in Tab. 4 and Tab. 5, according to the multivariate approach of Johansen, (1988), with the estimation of an error-correction model (ECM) which should lead us to the final choice of the model.

Estimation by the trace test:

The following statistic is used: (3) Where: “T” is the number of observations, “N” the number of variables, and (when ordered in ascending order).

is the eigenvalue

The critical values of the LRtrace statistic were tabulated by Johansen and Juselius (1990) and then by Osterwald-Lenum et al. (1992). H0 is accepted when the value of the LRtrace statistic is below its critical value. The maximum eigenvalue test: The following statistic is used: (4) The results obtained from this test are presented below:

107

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Table 4. Cointegration test (trace statistics).

Trace test

Max Eigenvalue test

H(r )

Eigenvalue

Statistic

5% p-value

Statistic

5% p-value

r =0

0.945555

144.6714

0.0000

72.76418

0.0000

r ≤1

0.849954

71.90727

0.0000

47.42032

0.0000

r≤2

0.566354

24.48695

0.0017

20.88815

0.0039

r ≤3

0.134071

3.598806

0.0578

3.598806

0.0578

Source: Authors, results obtained from the software (GRETL).

Trace test: The results obtained from this test are as follows: - H0: r = 0: H1: r > 0 = 144.67 > 47.85. We reject H0 so there is at least one cointegrating relation. - H0: r = 1: H1: r > 1 = 71.91 > 29.80. We reject H0 so there is at least one cointegrating relation. - H0: r = 2: H1: r > 2 = 24.49 > 15.49. We reject H0 so there is at least one cointegrating relation. - H0: r = 3: H1: r > 3 = 3.60 < 3.84. The null hypothesis (existence of 3 cointegrating relations) was accepted because we have 3.60 < 3.84. So, we have found three cointegrating relations; the error-correction model can then be estimated. Max Eigenvalue test: The results obtained from this test are as follows: - H0 : r = 0 : H1 :r > 0 = 72.76 > 27.58. We reject H0 so there is at least one cointegrating relation. - H0: r = 1: H1: r > 1 = 47.42 > 21.13. We reject H0 so there is at least one cointegrating relation. - H0: r = 2: H1: r > 2 = 20.88 > 14.26. We reject H0 so there is at least one cointegrating relation. - H0: r = 3: H1: r > 3 = 3.60 < 3.84. The null hypothesis (existence of 3 cointegrating relations) was accepted because we have 3.60 < 3.84. So, there are three cointegrating relations. The error-correction model can then be estimated.

108

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

The results, therefore, indicate the existence of a long-run relationship between the variables. Since a long-run relationship exists between the variables, the vector error correction model is next estimated for short-run and long-run coefficients. The Short-Run and Long-Run Effects: The final outcome is given in Table 5, 6 and 7. The entire Eq 1 can be deduced from Table 5. The target model D (GINI) which is the dependent variable given in Eq 5 (a) (b), (c) between Table 5 & 6. D (GINI) is identified as . (a) (b)(5) (c) Table 5. Long-Run Coefficients CointEq1

CointEq2

CointEq3

1.0000 (0.00000)

GDPP

0.079554 (0.031768)

0.68603 (1.0622)

0.46175 (0.45385)

Constant

-4.2501 (0.24383)

-0.59620 (8.1524)

-6.6783 (3.4835)

GINI

Source: Authors, results obtained from the software (GRETL).

109

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Table 6. Short-Run Coefficients

110

D(GINI)

D(PC)

D(CO)

D(GDPP)

ECt-1

−1.08538 (1.50337) [−0.7220]

3.04774 (2.13836) [1.425]

6.75635 (1.09527) [6.169]

2.54901 (1.12809) [2.260]

ECt-2

0.0805839 (0.240341) [−0.3353]

−0.659333 (0.341855) [−1.929]

0.586695 (0.175099) [3.351]

0.0716799 (0.180345) [0.3975]

ECt-3

0.199190 (0.627394) [0.3175]

1.29052 (0.892390) [1.446]

−1.84936 (0.457084) [−4.046]

−0.274560 (0.470779) [−0.5832]

ΔGINIt-1

0.111682 (1.51870) [0.07354]

−4.53589 (2.16017) [−2.100]

−8.07833 (1.10644) [−7.301]

−2.45044 (1.13959) [−2.150]

ΔGINIt-2

5.16058 (4.09831) [1.259]

16.2508 (5.82933) [2.788]

14.5977 (2.98580) [4.889]

0.296903 (3.07526) [0.09655]

ΔGINIt-3

−2.22632 (2.25869) [−0.9857]

−0.506290 (3.21270) [−0.1576]

−1.47738 (1.64555) [−0.8978]

1.93935 (1.69486) [1.144]

ΔGINIt-4

−1.20713 (2.89783) [−0.4166]

−15.0651 (4.12179) [−3.655]

−11.6548 (2.11119) [−5.520]

−2.50844 (2.17445) [−1.154]

ΔPCt-1

−0.114666 (0.175131) [−0.6547]

0.109303 (0.249102) [0.4388]

−0.0442864 (0.127591) [−0.3471]

0.0194306 (0.131414) [0.1479]

ΔPCt-2

−0.159371 (0.164036) [−0.9716]

−0.303333 (0.233321) [−1.300]

−0.385272 (0.119507) [−3.224]

0.0713866 (0.123088) [0.5800]

ΔPCt-3

0.0419384 (0.134058) [0.3128]

−0.177047 (0.190681) [−0.9285]

−0.653157 (0.0976672) [−6.688]

0.00516872 (0.100594) [0.05138]

ΔPCt-4

0.0136402 (0.206621) [0.06602]

0.0386999 (0.293892) [0.1317]

−0.101676 (0.150532) [−0.6754]

−0.0727191 (0.155043) [−0.4690]

ΔCOt-1

0.0109583 (0.374894) [0.02923]

−1.28411 (0.533239) [−2.408–]

0.388465 (0.273126) [1.422]

−0.0451870 (0.281310) [−0.1606]

ΔCOt-2

0.129525 (0.340923) [0.3799]

−0.450767 (0.484920) [−0.9296]

0.670267 (0.248377) [2.699]

−0.0212034 (0.255819) [−0.08288]

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

D(GINI)

D(PC)

D(CO)

D(GDPP)

ΔCOt-3

0.0989050 (0.244137) [0.4051]

−0.307420 (0.347254) [−0.8853]

0.499004 (0.177864) [2.806]

−0.0304529 (0.183194) [−0.1662]

ΔCOt-4

0.173160 (0.204725) [0.8458]

−0.132674 (0.291195) [−0.4556]

0.252569 (0.149151) [1.693]

−0.00989379 (0.153620) [−0.06440]

ΔGDPPt-1

−0.00989379 (0.153620) [−0.06440]

3.17486 (0.780059) [4.070]

1.05049 (0.399548) [2.629]

−0.312189 (0.411519) [−0.7586]

ΔGDPPt-2

1.17794 (0.753815) [1.563]

4.18467 (1.07221) [3.903]

0.117177 (0.549187) [0.2134]

−0.159584 (0.565642) [−0.2821]

ΔGDPPt-3

1.04999 (0.980975) [1.070]

4.11626 (1.39531) [2.950]

−0.329000 (0.714683) [−0.4603]

−0.478141 (0.736096) [−0.6496]

ΔGDPPt-4

0.294360 (0.672035) [0.4380]

2.28910 (0.955886) [2.395]

0.102849 (0.489607) [0.2101]

−0.425444 (0.504277) [−0.8437]

Source: Authors, results obtained from the software (GRETL).

Table 7. Statistical data of the outcome with the AIC. D(GINI)

D(PC)

D(CO)

D(GDPP)

R-squared

0.653037

0.964277

0.982049

0.926194

Adj.R-Squared

0.565420

0.828530

0.913837

0.645729

Sum sq.resids

0.005412

0.010950

0.002873

0.003047

F-statistic

0.850445

189.3956

81.76158

98.84134

Mean dependent

−0.001399

0.064111

0.022873

0.024201

S.D dependent

0.025454

0.092142

0.078250

0.033321

Log likelihood

305.01899

Akaike information criterion

-18.0015

Bayesian information criterion

-14.1011

Hannan–Quinn information

-16.9197

Determinant resid covariance

2.9693238e-016

Source: Authors, results obtained from the software (GRETL).

111

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Validation of the VECM model

This step consists in checking the absence of autocorrelation and the homoscedasticity at the level of the residues of each model. To do this, we will use the following tests: - The Ljung-Box test for the absence of autocorrelation; - The ARCH test for homoscedasticity. Autocorrelation test: This step consists of verifying that the residues are white noise, using the Q-statistic of Box and Pierce (1970). The Ljung-Box test is based on the analysis of the Q-statistic which is defined by: (6) Where “n” is the number of observations and refers to the empirical autocorrelation of order “k”, the null hypothesis of white noise (p1 = p2 = - - - = pk = 0) is rejected at the threshold of “a” if the statistic “Q” is higher than the critical value read in the table of “X2” to “ℎ” degrees of freedom. - Residuals of the VAR system, GINI: The Ljung-Box Q statistic with delay h = 15 confirms that there is no autocorrelation. Indeed, the test probability of h = 15 is 0.999> 0.05, so the null hypothesis of white noise is accepted (see Table. 10 and Figure. 6 in the Appendix). - Residuals from the VAR system, PC: The Ljung-Box Q statistic with delay h = 15 confirms that there is no autocorrelation. Indeed, the test probability of h = 15 is 0.170 > 0.05, so the null hypothesis of white noise is accepted (see Table. 10 and Figure. 6 in Appendix). - Residuals from the VAR system, CO: The Ljung-Box Q statistic with delay h = 15 confirms that there is no autocorrelation. Indeed, the test probability of h = 15 is 0.083 > 0.582, so the null hypothesis of white noise is accepted (see Table. 10 and Figure. 6 in Appendix). - Residues from the VAR system, GDPP: The Ljung-Box Q statistic with delay h = 15 confirms that there is no autocorrelation. Indeed, the test probability of h = 15 is 0.230 > 0.05, so the null hypothesis of white noise is accepted (see Table. 10 and Figure. 6 in appendix). We conclude that all values are not random and independent in time, so there is no autocorrelation in our model. Homoscedasticity test: We used the ARCH test. The multivariate version is based on the following regression: (7) “ 112

Where: “ pt” is a white noise process and “Vm ” the vectorization operator for symmetrical matrices. ” pour are of dimensions .

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

The assumptions of the ARCH test are as follows: H0: homoscedasticity

H0: heteroskedasticity and there is at least one coefficient “

” significantly different from 0.

The test statistic is defined as follows:

(8)

(9)

with:

where “

” designates the variance-covariance matrix of the regression relationship.

The test statistic follows a law

Using the ARCH test, one finds: 231.72 < 233, 994 = (q) designates the critical value appearing in the Chi-square table, we accept the hypothesis of homoscedasticity at the threshold of 5%. Using the ARCH test, we find: 231.72 < 233, 994 = (q) designates the critical value in the Chi-square table, we accept the hypothesis of homoscedasticity at the 5% threshold.

EMPIRICAL RESULTS AND DISCUSSION In this framework, we will analyze the results obtained from the estimation of the elasticities of our econometric model, through the comparison of the signs and values of the coefficients with the economic theory, and the results of other empirical work. First of all, it should be noted at this level that the estimated parameters of the model represent elasticities because we have worked with logarithmic data. For this reason, we have used the (Log) of the variables. Based on the result of Johansen's cointegration test, we can see that there is a long-term equilibrium relationship between income inequality (GINI), bank credits (PC), trade openness (CO) and average per capita income (GDPP). To estimate this relationship, the Vector Error Correction Model (VECM) was used. The short-run effects of private credit on income inequality are mixed; the effect is not significantly negative at lower lags and significantly positive at higher lags (both the short-run coefficients are significant at 10% significance level). The short-run effect of per-capita real on income inequality is significantly negative at 5% significance level; thus, greater gross domestic product (GDP) per capita will expectedly reduce income inequality in the Morocco in the short-run. The short-run effect of trade openness on income inequality is significantly positive at 5% significance level; thus, greater trade openness will increase income inequality in the short-run (Table .6). With respect to long-term dynamics (LT), the equation shows the existence of a long-term equilibrium relationship (LT). This explains that there is an impact of variations in the explanatory variables on the variation of the dependent variable. In other words, variations in bank credit, trade openness, and gross domestic product per capita have an impact on income inequality (Table. 6). 113

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Eq 5(c) is the long-term relationship as each variable at (t−1) is equal to each variable at t. An outlier can be defined as an observation with a larger residual, which represents the difference (positive or negative) between the actual value and the estimated value from the regression model. When the residual is large, compare with other residuals. Usually, a large residual will attract attention because the vertical distance between it and the estimated regression line is quite large. The relationship of GINI is deduced from (Table. 6) using the error correction model where the values in brackets ‘()’ is the standard error, and the values in square brackets ‘[ ]’ is the t statistic. In fact, an increase of 1% in GDPP, for example, will lead in the long term to a decrease of 0.0608 in the Gini index. Moreover, the impact of trade openness on the dependent variable is positive, because simply an increase of 1% in this variable causes an increase of 0.1991 in the Gini index. In addition to the complexity of the distribution effect, the long-term results are consistent with the theorem of (Hecksher-Ohlin Samuelson and Stolper-Samuelson), that, in developed economies, increased trade openness will lead to increased income inequality, an increase of 1% in bank credits, for example, will lead in the long term to a 0.0806 increase in the Gini index. The estimated elasticities of our econometric model were in accordance with the economic theory. All the coefficients have the expected signs, and they are all statistically significant. Thus, the growth in bank credit has a positive impact on the rise in income inequality in the long run. Finally, based on the signs and values of the estimated elasticities, we find that our result is consistent with results obtained from other empirical work, for example, Arora (2012), Gimet & Lagoarde-Segot (2011), (Kim & Lin, 2011), Bazillier et al., (2015), Jauch & Watzka (2016), Bazillier et al., (2017). These empirical studies have found that the growth of bank credit has a positive impact on income inequality.

Impulse Response Functions The objective in impulse analysis is to show the impact of a shock to one variable in the system on the other variables since there is a dynamic structure in the composition of a VAR, i.e. to represent the effect of a shock to an innovation on the other variables. In the (Figure. 5), we notice that the shock on GINI has an instantaneous impact on itself, which explains why the curve relating to GINI ⇒ GINI starts from a value sufficiently higher than 0. A positive shock on bank credit results in a negative effect on the Gini index during the same period. A positive shock on trade openness translates into a positive effect on the Gini index during the first 10 years. In addition, a shock to gross domestic product per capita has a negative effect on the Gini index during this period. A positive shock on bank credits has an instantaneous effect on itself, which explains why the curve for PC ⇒ PC starts from a value sufficiently above 0, while a positive shock on the Gini index translates into a positive effect on PC throughout the first 10 years, moreover it is an increasing function. Thus, this confirms that the impact of PC volatility on GINI remains positive in the long term. And a positive shock on trade openness creates a positive effect on PC over the same period. A positive shock on GDPP creates a positive effect on PC during this period. A positive shock on trade openness has an instantaneous effect on itself, which explains why the curve for CO ⇒ CO starts from a value sufficiently above 0, a positive shock on the Gini index has a negative effect on CO throughout the first 10 years. And a positive shock on bank loans creates a positive effect on CO over the same period, a positive shock on GDPP creates a positive effect on CO during this period. 114

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

A positive shock to GDP per capita results in a positive impact on GDP per capita during the first 10 years, while a positive shock to PC has a positive effect on OGDI, a positive shock to CO has a positive effect on GDPP over the same period. Finally, a shock on GDPP has an instantaneous impact on itself, which explains why the curve for GDPP ⇒ GDPP starts from a value that is sufficiently higher than 0.

Variance Decomposition Analysis The interest is to know what is the contribution of each innovation to the total variance of the prediction error. The percentage contribution of the residuals of each variable to the variance of the prediction error of the variable under consideration is presented in (Table. 9), from which conclusions can be drawn on the variable that has the greatest influence on the other variables. But to make this study, it is necessary to order the variables from the most exogenous to the most endogenous because in the decomposition of "Cholesky" there will be a change of innovations in a way that the first variable will be according to its innovations; the second according to its innovations as well as the innovations of the first variable etc. Variance decomposition for GINI: According to (Table. 9), we obtain on average an innovation of the GINI index that contributes 99.09% of its variance of the prediction error, bank credits contribute on average 0.2353% of the variance of the GINI error, 0.5779% for trade openness, 0.0936% for the gross domestic product per capita. So, we conclude that the GINI index contributes a good part in the determination of the variance of the prediction error. Variance decomposition for PC: Also, on average an innovation of the Gini index which contributes 0.1276% of the variance of the forecast error, bank credits contribute on average 86.84% of its own variance of the error, 1.012% for trade openness, 12.01% for the gross domestic product per capita. We conclude that the gross domestic product per capita has the largest contribution after bank credits in the determination of the variance of the forecast error. Variance decomposition for CO: A Gini index innovation contributes 4.23% of the variance of the forecasting error, bank credits contribute on average 80.85%, trade openness contributes on average 14.45% of its own variance of the error, 0.452% for the gross domestic product per capita. It is concluded that trade openness has the largest share after bank credits in determining the variance of the forecast error. Variance decomposition for GDPP: On average, an innovation of the Gini index that contributes 0.1272% of the variance of the error of prediction, the bank credits contribute on average 12.83%, 10.73% for the commercial opening, the gross domestic product per capita contributes on average 76.31% its own variance of the error. It is concluded that bank credits have the second largest share after gross domestic product per capita in determining the variance of the forecast error.

115

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

CONCLUSION

While remaining within the framework of the main lines dealt with in this article and while basing ourselves on the explanatory economic theories presented and the explanatory models that allow us to study the problematic posed on the impact of the bank credit on inequalities and also on the reduction of purchasing power in the Moroccan economy, we have demonstrated the theoretical and empirical links between the two variables as it has already been done in several research works of economists on the international level. This paper examines the short-run and long-run effects of bank credits on income inequalities in the Morocco. Due to the limitation of data availability, the study is limited to 1990-2019, we conclude that there is a positive impact of bank credits on income inequalities in the short-run and the longrun, so we find in fact that bank credits (consumer credits and real estate credits) represent a direct consequence of the rise of income inequalities. Reason why the number of poor people in the world is increasing and the income inequality between rich and poor is also increasing. This regression model does not have evolutionary technologies such artificial intelligence (neural networks, etc.). Future work will solve this problem by establishing a hybrid model that combines Johansen's cointegration principle and machine learning technique. Finally, inequality remains a subject of analysis and economic reflection among economists by continuing to carry out theoretical and empirical research in order to develop new economic theories and simplified econometric modeling. It is preferable to propose new avenues of research, given that the problem of rising inequality is multidimensional and that the study of the impact of several other variables on the rise in inequality is still an acute question.

116

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

REFERENCES

Abiad, A., Detragiache, E., & Tressel, T. (2010). A new database of financial reforms. IMF Staff Papers, 57, 281–302. Abosedra, S., Shahbaz, M., & Nawaz, K. (2016). Modeling causality between financial deepening and poverty reduction in Egypt. Social Indicators Research, 126, 955–969. Agnello, L., Mallick, S. K., & Sousa, R. M. (2012). Financial reforms and income inequality. Economics Letters, 116, 583–587. Amine, S., & Ouattara, P. (2015). The effect of incentive return-to-work policies on single-parent families: A comparative approach. The European Journal of Applied Economics, 12, 9–15. Amri, K., & others. (2018). Is there causality relationship between economic growth and income inequality?: Panel data evidence from Indonesia. Eurasian Journal of Economics and Finance, 6, 8–20. Asteriou, D., Dimelis, S., & Moudatsou, A. (2014). Globalization and income inequality: A panel data econometric approach for the EU27 countries. Economic modelling, 36, 592–599. Ball, L. M., Furceri, D., Leigh, M. D., & Loungani, M. P. (2013). The distributional effects of fiscal consolidation. International Monetary Fund. Batuo, M. E., Guidi, F., Mlambo, K., & others. (2010). Financial development and income inequality: Evidence from African Countries. African Development Bank, 1–28. Bayar, Y., & Karamelikli, H. (2017). Financial development and tax revenues in Turkey: A non-linear cointegration analysis. The European Journal of Applied Economics, 14, 31–42. Bazillier, R., & Najman, B. (2017). Labour and financial crises: is labour paying the price of the crisis? Comparative Economic Studies, 59, 55–76. Binatli, A. O. (2012). Growth and Income Inequality: A Comparative Analysis. Economics Research International, 1, 1–7. Chen, W., & Kinkyo, T. (2016). Financial Development and Income Inequality: Long-Run Relationship and Short-Run Heterogeneity. Emerging Markets Finance and Trade, 52, 733-742. Chiu, Y-B., & Lee, C-C. (2019). Financial development, income inequality, and country risk. Journal of International Money and Finance, 93(C), 1-18. https://doi.org/10.1016/j.jimonfin.2019.01.001 Dabla-Norris, M. E., Kochhar, M. K., Suphaphiphat, M. N., Ricka, M. F., & Tsounta, M. E. (2015). Causes and consequences of income inequality: A global perspective. International Monetary Fund. De Haan, J., & Sturm, J-E. (2017). Finance and income inequality: A review and new evidence. European Journal of Political Economy, 50, 171-195. https://doi.org/10.1016/j.ejpoleco.2017.04.007 Delis, M. D., Hasan, I., & Kazakis, P. (2014). Bank regulations and income inequality: Empirical evidence. Review of Finance, 18, 1811–1846. Denk, O., & Cournède, B. (2015). Finance and income inequality in OECD countries. Fligstein, N., & Goldstein, A. (2015). The emergence of a finance culture in American households, 1989-2007. Socio-Economic Review, 13(3), 575-601. https://doi.org/10.1093/ser/mwu035 Gimet, C., & Lagoarde-Segot, T. (2011). A closer look at financial development and income distribution. Journal of Banking & Finance, 35(7), 1698–1713. Godechot, O. (2016). Financialization is marketization! A study of the respective impacts of various dimensions of financialization on the increase in global inequality. Sociological Science, 3, 495-519. https://doi.org/10.15195/v3.a22 Hamori, S., & Hashiguchi, Y. (2012). The effect of financial deepening on inequality: Some international evidence. Journal of Asian Economics, 23(4), 353–359. https://doi.org/10.1016/j.asieco.2011.12.001 Hendry, D. F., & others. (1995). Dynamic econometrics. Oxford University Press on Demand. Herzer, D., & Vollmer, S. (2012). Inequality and growth: evidence from panel cointegration. The Journal of Economic Inequality, 10, 489–503. https://doi.org/10.1007/s10888-011-9171-6 117

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Huang, H-C. R., Fang, W., Miller, S. M., & Yeh, C-C. (2015). The effect of growth volatility on income inequality. Economic Modelling, 45, 212-222. https://doi.org/10.1016/j.econmod.2014.11.020 Hyde, A., Vachon, T., & Wallace, M. (2018). Financialization, Income Inequality, and Redistribution in 18 Affluent Democracies, 1981–2011. Social Currents, 5, 193-211. https://doi.org/10.1177/2329496517704874 Ibrahim, A. T. (2020). An exploratory analysis of financial inclusion in Chad. The European Journal of Applied Economics, 17, 34–53. Jalil, A., & Feridun, M. (2011). Long-run relationship between income inequality and financial development in China. Journal of the Asia Pacific Economy, 16(2), 202–214. https://doi.org/10.1080/13547860.2011.564745 Jauch, S., & Watzka, S. (2016). Financial development and income inequality: a panel data approach. Empirical Economics, 51, 291–314. https://doi.org/10.1007/s00181-015-1008-x Jeanneney, S. G., & Kpodar, K. (2011). Financial development and poverty reduction: can there be a benefit without a cost? The Journal of development studies, 47(1), 143–163. https://doi.org/10.1080/00220388.2010.506918 Jenkins, Brandolini, A., Micklewright, J., & Nolan, B. (2012). The great recession and the distribution of household income. OUP Oxford. Jeong, H., & Kim, S. (2018). Finance, growth, and inequality: New evidence from the panel var perspective. Seoul Journal of Economics, 31(2), 121-143. Johansen, S., & Juselius, K. (1990). Maximum likelihood estimation and inference on cointegration—with appucations to the demand for money. Oxford Bulletin of Economics and statistics, 52(2), 169–210. https://doi.org/10.1111/j.1468-0084.1990.mp52002003.x Kim, D-H., & Lin, S-C. (2011). Non linearity in the financial development–income inequality nexus. Journal of Comparative Economics, 39(3), 310–325. https://doi.org/10.1016/j.jce.2011.07.002 Krugman, P. (2005). For richer. Critical social issues in American education: Democracy and meaning in a globalizing world, 3, New York, USA: Routledge, Taylor & Francais Group. Kunieda, T., Okada, K., & Shibata, A. (2014). Finance and inequality: How does globalization change their relationship? Macroeconomic Dynamics, 18(5), 1091–1128. https://doi.org/10.1017/S1365100512000843 Kuznets, S. (1955). Economic growth and income inequality. The American economic review, 45, 1–28. Malinen, T. (2013). Inequality and growth: Another look with a new measure and method. Journal of International development, 25(1), 122–138. https://doi.org/10.1002/jid.2812 Manish, G. P., & O’Reilly, C. (2019). Banking regulation, regulatory capture and inequality. Public Choice, 180, 145-164. https://doi.org/10.1007/s11127-018-0501-0 Marx, K., Engels, F., Huard, R., & Cornillet, G. (1936). Les luttes de classes en France (1848-1850). Éd. sociales internationales. Mohamad, N. M., Masron, T. A., Wijayanti, R., & Jamil, M. M. (2020). Islamic Banking and Income Inequality: The Role of Corporate Social Responsibility. Jurnal Ekonomi Malaysia, 54(2), 77-90. https://doi.org/10.17576/JEM-2020-5401-7 Mookerjee, R., & Kalipioni, P. (2010). Availability of financial services and income inequality: The evidence from many countries. Emerging Markets Review, 11(4), 404–408. https://doi.org/10.1016/j.ememar.2010.07.001 Neves, P. C., Afonso, Ó., & Silva, S. T. (2016). A Meta-Analytic Reassessment of the Effects of Inequality on Growth. World Development, 78, 386-400. https://doi.org/10.1016/j.worlddev.2015.10.038 Park, C.-Y., & Mercado, R. (2015). Financial inclusion, poverty, and income inequality in developing Asia. Asian Development Bank Economics Working Paper Series. Rewilak, J. (2013). Finance is good for the poor but it depends where you live. Journal of Banking & Finance, 37(5), 1451–1459. https://doi.org/10.1016/j.jbankfin.2012.04.022 Roberts, A., & Kwon, R. (2017). Finance, inequality and the varieties of capitalism in post-industrial democracies. Socio-Economic Review, 15(3), 511-538. https://doi.org/10.1093/ser/mwx021

118

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Saez, E., & Zucman, G. (2016). Wealth in equality in the United States since 1913: Evidence from capitalized income tax data. Quarterly Journal of Economics, 131(2), 519-578. https://doi.org/10.1093/qje/qjw004 Sehrawat, M., & Giri, A. K. (2015). Financial development and income inequality in India: an application of ARDL approach. International Journal of Social Economics, 42(1), 64-81. https://doi.org/10.1108/IJSE-09-2013-0208 Seven, U., & Coskun, Y. (2016). Does financial development reduce income inequality and poverty? Evidence from emerging countries. Emerging Markets Review, 26, 34-63. https://doi.org/10.1016/j.ememar.2016.02.002 Shahbaz, M., Bhattacharya, M., & Mahalik, M. K. (2017). Finance and income inequality in Kazakhstan: evidence since transition with policy suggestions. Applied Economics, 49(52), 5337-5351. https://doi.org/10.1080/00036846.2017.1305095 Stiglitz, J. E. (2012). The price of inequality: How today's divided society endangers our future. WW Norton and Company. Stiglitz, J. E. (2013). Society Endangers Our Future. New York: WW Norton & Company. Tan, H-B., & Law, S-H. (2012). Nonlinear dynamics of the finance-inequality nexus in developing countries. The Journal of Economic Inequality, 10, 551–563. https://doi.org/10.1007/s10888-011-9174-3 Thornton, J., & Tommaso, C. D. (2020). The long-run relationship between finance and income inequality: Evidence from panel data. Finance Research Letters, 32. 101180. https://doi.org/10.1016/j.frl.2019.04.036 Tohirin, A., & Husaini, F. (2019). Does Islamic Banking Financing Help the Poor? Proceeding of International Conference on Accounting, Business & Economics, 1, 41–50. Uddin, G. S., Shahbaz, M., Arouri, M., & Teulon, F. (2014). Financial development and poverty reduction nexus: A cointegration and causality analysis in Bangladesh. Economic Modelling, 36, 405–412. https://doi. org/10.1016/j.econmod.2013.09.049 Woo, J. (2011). Growth, income distribution, and fiscal policy volatility. Journal of Development Economics, 96(2), 289–313. https://doi.org/10.1016/j.jdeveco.2010.10.002 Woo, J., Bova, M. E., Kinda, M. T., & Zhang, M. Y. (2013). Distributional consequences of fiscal consolidation and the role of fiscal policy: What do the data say? International Monetary Fund. Zamora, D. (2019). Déplorer les inégalités, ignorer leurs causes. Le Monde Diplomatique. Retrieved May 13, 2019, from: https://www.monde-diplomatique.fr/2019/01/ZAMORA/59419.

119

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

APPENDIX

Table 8. Variable information. Variable name

Code

Definition

Gini Coefficient

GINI

Household Income Inequality

GDP per capita

GDPP

GDP per capita is gross domestic product divided by mid-year World Bank population. (2020)

Trade openness

Exports and Imports added together as a ratio of GDP

Private credit

Calculated as private credit as a ratio of GDP, measured by credit to the private sector (consumer credit and real estate credit)

Figure 3.Graphical representation of the variable

120

Source: Authors, results obtained from the software (Jupyter Notebook).

Source World Bank (2020)

World Bank (2020) Bank AlMaghrib

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Figure 4. Relationships Between Variables

Source: Authors, results obtained from the software (Jupyter Notebook).

121

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Figure 5. Impulse Response Function (IRF)

Source: Authors, results obtained from the software (Jupyter Notebook). Note: The numbers on the axis suggest the percentile, the model is fixed to see the response in a period of 10 years

122

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Figure 6. Test For Serial Correlation

Source: Authors, results obtained from the software (Jupyter Notebook).

123

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Table 9. Variance decomposition analysis

Variance decomposition of GINI

Variance decomposition of PC

Period

Erreur Std

GINI

GDPP

Period

Erreur Std

GINI

GDPP

0.0282

100.0 0

0.000 0

0.0000

0.0796

0.23 28

99.76 7

0.00 00

0.0000

0.0300

99.09 4

0.23 35

0.577 9

0.0936

0.1482

0.12 76

86.84 6

1.01 21

12.013

0.0400

97.67 3

0.15 18

1.301 0

0.8735

0.2419

18.5 21

59.52 2

2.98 76

18.968

0.0404

95.52 7

2.26 25

1.342 3

0.8674

0.2976

13.8 96

55.09 6

4.16 88

26.838

0.0497

95.01 2

1.60 46

1.994 1

1.3866

0.3449

11.6 62

55.30 5

4.97 26

28.058

0.0503

94.02 7

2.49 46

2.122 5

1.3550

0.3850

10.3 17

53.47 0

6.72 67

29.484

0.0650

95.89 2

1.50 99

1.588 7

1.0089

0.4169

8.81 14

54.96 8

7.17 40

29.046

0.0656

95.48 8

1.71 34

1.615 3

1.1829

0.4517

11.5 27

52.51 6

7.65 61

28.300

0.0816

96.38 9

1.35 55

1.390 0

0.8646

0.4742

11.0 06

52.83 9

8.28 97

27.864

0.0835

95.61 2

1.66 32

1.357 1

1.3674

0.5008

13.9 14

50.78 3

8.42 61

26.875

Variance decomposition of CO

Variance decomposition of GDPP

0.0547

0.311 6

72.7 09

26.97 8

0.0000

0.0246

0.00 15

13.12 4

12.9 56

73.917

0.0794

4.236 4

80.8 58

14.45 2

0.4520

0.0272

0.12 72

12.83 5

10.7 28

76.308

0.1118

27.83 4

55.9 13

8.520 9

7.7308

0.0568

51.1 03

6.072 6

5.42 89

37.394

0.1203

24.48 7

49.3 79

11.98 4

14.149

0.0811

67.9 93

3.744 0

3.15 54

25.107

0.1315

24.44 5

48.5 86

11.51 7

15.449

0.1121

76.3 60

4.678 6

2.23 04

16.730

0.1661

42.38 7

39.6 32

7.664 9

10.315

0.1443

79.1 08

2.916 0

2.48 02

15.495

0.1833

38.87 9

40.2 93

9.366 9

11.459

0.1782

83.2 48

4.220 1

1.64 27

10.888

0.2069

44.09 6

36.1 08

8.115 8

11.679

0.2199

84.5 39

2.778 7

2.13 29

10.548

0.2155

41.60 0

38.9 04

7.871 5

11.623

0.2638

87.7 86

3.283 0

1.48 24

7.4477

0.2445

48.96 7

31.0 62

8.545 2

11.424

0.3208

89.4 54

2.276 9

1.49 49

6.7736

Source: Authors, results obtained from the software (GRETL).

124

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

Table 10. Autocorrelation Test Delay

Autocorrelation function of GINI ACF

PACF

Q-stat

P.crit

0.039

0.0437

0.834

0.002

0.000

0.0438

0.014

-0.027

Delay

Autocorrelation function of PC ACF

PACF

Q-stat

P.crit

-0.060

0.1056

0.745

0.978

-0.206

-0.211

1.3952

0.498

0.0503

0.997

-0.037

-0.068

1.4377

0.697

-0.029

0.0753

0.999

-0.206

-0.272

2.8484

0.584

-0.026

-0.024

0.0980

1.000

-0.014

-0.093

2.8555

0.722

-0.142

-0.140

0.8297

0.991

0.324

0.225

6.6824

0.351

0.026

0.038

0.8557

0.997

0.242

0.300

8.9230

0.258

0.089

0.088

1.1740

0.997

-0.374

-0.312

14.575

0.068

-0.109

-0.116

1.6844

0.996

-0.157

-0.173

15.630

0.075

0.088

0.091

2.0347

0.996

-0.066

-0.104

15.830

0.105

0.054

0.042

2.1775

0.998

-0.019

0.047

15.848

0.147

-0.028

-0.049

2.2172

0.999

0.144

-0.121

16.919

0.153

0.073

0.088

2.5177

0.999

0.116

-0.136

17.673

0.170

-0.046

-0.033

2.6454

1.000

-0.159

-0.125

19.201

0.157

0.096

0.071

3.2597

0.999

-0.114

0.166

20.062

0.170

Autocorrelation function of CO

Autocorrelation function of GDPP

0.052

0.0800

0.777

0.181

0.9535

0.329

-0.035

-0.037

0.1163

0.944

-0.313

-0.357

3.9209

0.141

-0.442

-0.440

6.3056

0.098

-0.119

0.027

4.3684

0.224

-0.142

-0.127

6.9702

0.137

0.117

0.034

4.8204

0.306

0.091

0.093

7.2598

0.202

0.011

-0.073

4.8245

0.438

0.384

0.238

12.617

0.050

-0.293

-0.270

7.9468

0.242

0.092

-0.022

12.940

0.074

-0.120

-0.004

8.5016

0.290

-0.229

-0.253

15.062

0.058

0.112

-0.047

9.0087

0.342

-0.249

-0.009

17.709

0.039

0.095

0.003

9.3981

0.401

-0.113

0.021

18.295

0.050

-0.102

-0.104

9.8705

0.452

0.097

-0.108

18.747

0.066

-0.168

-0.131

11.239

0.423

0.066

-0.251

18.975

0.089

0.080

0.013

11.572

0.481

0.143

0.113

20.122

0.092

0.038

-0.150

11.655

0.556

-0.206

-0.057

22.702

0.065

-0.098

-0.076

12.240

0.587

-0.075

-0.101

23.072

0.083

-0.124

-0.133

13.261

0.582

Source: Authors, results obtained from the software (GRETL)

125

EJAE 2021  18 (2)  95 - 126

KHATTAB. A., EL. KHLIFI. I.  THE LONG-RUN IMPACT OF BANK CREDIT GROWTH ON SOCIAL AND ECONOMIC INEQUALITIES IN MOROCCO EVIDENCE FROM THE JOHANSEN'S COINTEGRATION ANALYSIS

DUGOROČNI UTICAJ RASTA BANKARSKIH KREDITA NA SOCIJALNE I EKONOMSKE NEJEDNAKOSTI U MAROKU: DOKAZI IZ JOHANSENOVE ANALIZE KOINTEGRACIJE Rezime: Predmet rasta ekonomskih i društvenih nejednakosti postao je glavna briga za ekonomskih istraživača širom svijeta. Kao rezultat toga, postoji velika pažnja i zabrinutost zbog značajnih troškova rastućih nejednakosti u pogledu mira i socijalne koherentnosti u društvu. Da bi se suzbila šteta ovih nejednakosti, Maroko je preduzeo nekoliko reformi, uključujući i implementaciju novog razvojnog modela postavljenog 2018. Glavni cilj ovog članka je proceniti uticaj rasta bankarskih kredita na povećanje ekonomskih i društvenih nejednakosti. u Maroku. Drugim rečima, proverili smo da li postoji pozitivan uticaj na porast nejednakosti u prihodima u Maroku. Iz tog razloga smo u ovoj studiji testirali ekonometrijski model, koristeći metodu kointegracije, posebno model ispravljanja grešaka. Dakle, naši rezultati potvrđuju da se stepen otvorenosti trgovine, bankarski kredit i bruto domaći proizvod po glavi stanovnika dugoročno smatraju determinantama ravnoteže Gini indeksa. Koristili smo godišnje podatke za period 1990-2019 iz baze podataka Centralne banke Maroka (BAM) i Svetske banke (VBD). Širok spektar studija pokazuje značajan pozitivan uticaj između bankarskih kredita i nejednakosti u prihodima.

126

Ključne reči: Nejednakosti u prihodima, bankarski krediti, ekonomski rast, otvorenost trgovine, kointegracija.

EJAE 2021, 18(2): 127 - 145 ISSN 2406-2588 UDK: 005.321 005.322:316.46 DOI: 10.5937/EJAE18-30186 Original paper/Originalni naučni rad

THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS Predrag Mali, Bogdan Kuzmanović, Milan Nikolić, Edit Tarek Stojanović* University of Novi Sad, Tehnical faculty ‘Mihajlo Pupin’, Novi Sad,Serbia

Abstract: The paper presents the results of the study of individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, depending on seven variables: the respondents' gender, their age, the educational level of the respondents, the ownership structure of the enterprises, the respondents' previous experience in entrepreneurship, their perceived job performance, and their perceived finances. The specificity of the research is that the respondents are employed persons. The sample included 540 respondents from 72 organizations in Serbia. Data analysis was performed via a t-test. A statistically significant difference in the influence of the observed variables exists in most cases, except for the variable - the respondents' level of education. Thus, four of the seven hypotheses were fully confirmed, two were partially confirmed, while one hypothesis was rejected. The profile of an employed person who, potentially, has the greatest chances of becoming an entrepreneur is the following: a younger man with a high school diploma (a degree does not have such a significant impact), who is employed in a private company, has previous entrepreneurial experience, is successful at work and has adequate finances.

Article info: Received: January 5, 2021 Correction: Jul 31, 2021 Accepted: August 25, 2021

Keywords: Individual entrepreneurial orientation, the theory of planned behavior, entrepreneurial intentions, employed persons, Serbia.

INTRODUCTION There is a significant number of studies addressing entrepreneurial intentions among students, for example (Kwong & Thompson, 2016; Espiritu-Olmos & Sastre-Castillo, 2015; Altinay, Madanoglu, Daniele, & Lashley, 2012; Shinnar, Giacomin, & Janssen, 2012; Siu & Lo, 2013). Likewise, there is a large body of research indicating the importance of corporate entrepreneurship and its role in contemporary business (Mohedano-Suanes & Benitez, 2018; Hornsby, Peña-Legazque, & Guerrero, 2013; Gawke, Gorgievski, & Bakker, 2017; Wei & Ling, 2015). *E-mail: terekedita@gmail.com

127

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

However, to date, the external entrepreneurial intentions of employed persons (the intentions of employed persons to leave their current job, undertake an entrepreneurial venture and start their own business) have largely remained outside the focus of researchers. Some researchers (Hormiga, Hancock, & Valls-Pasola, 2013; Marshall & Gigliotti, 2018), point out this problem, as well as the need for its deeper study. According to some authors (Hormiga et al., 2013; Miralles, Giones, & Riverola, 2016; Saraf, 2015), employed persons are actually more likely to succeed in an entrepreneurial venture. The reasons for this are logical: employed persons have much more experience, practical knowledge, better knowledge of the situation in their industry, and especially knowledge about any deficiencies in the field in which they work, which leaves room for defining the idea of the type of future enterprise. Finally, employees often have more financial opportunities to start their own business (they have been making money for a while). Here, the question of the expediency of researching entrepreneurial intentions among people who already have a job may be raised. However, the benefits of an employed person starting a new business are obvious: jobs will be created in that new business as well as within the organization left by the (new) entrepreneur. This paper examines individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions of employed persons. In doing so, these dimensions are observed depending on seven variables: the respondents’ gender, their age, the educational level of the respondents, the ownership structure of enterprises, the respondents’ previous experience in entrepreneurship, their perceived success at work, and their perceived finances. The research was conducted in organizations in Serbia. Detecting and understanding the significant differences of the observed dimensions, depending on the given variables, has both theoretical and practical significance. The theoretical significance is reflected in the extension of knowledge in an area that has not been sufficiently researched. The practical importance derives from the identification of the entrepreneurial orientation, motivation, attitudes and intentions of employed persons, depending on the control variables, thus creating a theoretical basis for the proper functioning of state structures in the sphere of promoting and developing entrepreneurship.

THEORY AND HYPOTHESIS In research dealing with detecting and explaining the effects on entrepreneurial intentions, variables such as gender, age, type of education, type and characteristics of previous work experience, the existence of entrepreneur parents, etc. are often considered. Such studies most often seek to identify the categories and groups in the general population that are most prone to starting their own businesses. As mentioned, unemployed persons or employed persons in the sense of internal entrepreneurship are usually considered.

Gender When it comes to the gender of entrepreneurs and potential entrepreneurs, there are also some typical research questions here which capture the attention of researchers in the field of entrepreneurship. One group of references deals with the level of the participation of women and men in entrepreneurial activities. 128

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

According to Hisrich, Peters, and Shepherd (2008), there has been a recent increase in self-employment among women, who are now more likely to start new entrepreneurial ventures than men, and in the US women start up twice as many businesses as men and stay in business longer. The researchers (Poggesi, Mari, & De Vita, 2016) highlight the very significant role and representation of women in entrepreneurship globally. It is difficult to make a general, objective assessment of the distribution of entrepreneurial activities among men and women. Nevertheless, there seems to be a balance in that men no longer dominate entrepreneurship, and women are increasingly involved in this type of occupation. Furthermore, some studies address the strength of entrepreneurial intentions among women and men. One group of references shows a uniformity of intensity of entrepreneurial intentions in both genders, while another group shows that entrepreneurial intentions are higher for men. Thus, according to Saraf (2015), entrepreneurial intentions are similar for men and women, as is motivation for starting a business (Knorr, Garzón, & Martinez, 2011). However, Leppel (2016) found that men nevertheless have a greater tendency to become entrepreneurs than women. According to Santos, Roomi, and Liñán (2016), there are similarities in entrepreneurial intentions among men and women, although they are somewhat greater for men. This is due to the fact that in society men are more encouraged towards self-employment, which makes women feel that entrepreneurship is not an acceptable career choice for them. Male students in Turkey and Norway show significantly higher levels of entrepreneurial intentions than female students (Shneor, Metin Camgöz, & Bayhan Karapinar, 2013). The next group of studies deals with the success rate of male and female entrepreneurs, as well as certain differences between them (degree of satisfaction, goals, career orientation). There are numerous similarities between male and female entrepreneurs (Jayawarna, Jones, & Marlow, 2015), thus eliminating prejudices about women's disadvantages in the role of entrepreneurs. Also, according to Conroy and Weiler (2016), there are no differences in the economic performance of firms and the growth of business volume depending on whether the business owners are men or women. Differences exist in some “soft” parameters. Thus, female entrepreneurs show higher levels of job satisfaction than their male counterparts (Bender & Roche, 2016). Another study (Hechavarria, Terjesen, Ingram, Renko, Justo, & Elam, 2017) shows that female entrepreneurs pay more attention than men to goals that are focused on social values, and less attention to those that are fundamentally economic.

Age Other researchers have tried to answer the question as to what motivates older people to start a private business later in life. Research in Australia (Perenyi, Zolin, & Maritz, 2018), has shown that older entrepreneurs start their own businesses, first and foremost, because they are given the right opportunity, and not out of necessity. Also, internal motives are primary. According to Kean, Van Zandt, and Maupin (2008), entrepreneurship is gaining increasing importance among elderly persons primarily because older people often have low incomes. The success of older entrepreneurs is linked to the desire for autonomy and independence, confidence, personal effectiveness and intergenerational support. Other studies found that entrepreneurial intentions decline with age. Thus, the desire for entrepreneurial endeavours, as well as perceived feasibility, culminates in early adulthood, and declines considerably with age (Minola, Criaco, & Obschonka, 2016). Similarly, researchers (Tsai et al., 2016) found that age negatively affected perceived chances and entrepreneurial intentions. Research in France (Sahut, Gharbi, & Mili, 2015) found that there was a negative relationship between the respondents' age and entrepreneurial intentions. 129

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Also, it is very important for older respondents to have the competencies and resources they need. This indicates greater caution for older respondents and lower risk preference. Among employed people in Australia (Hatak, Harms, & Fink, 2015), entrepreneurial intentions are found to decline with age. The survey, in which the respondents were entrepreneurs and internal entrepreneurs (Jain & Ali, 2012) found that younger entrepreneurs were more proactive than others, while middle-aged entrepreneurs exhibited a higher locus of control and risk preference. According to Miralles, Giones, and Gozun (2017), individuals currently engaged in entrepreneurial behaviors and activities logically have stronger intentions to start a new venture, whereby this connection is reinforced in older individuals.

Education One of the common questions that arises when it comes to the link between education and entrepreneurship is how educated entrepreneurs are. Entrepreneurs are often believed to be less educated than the rest of the general population, however, according to Hisrich et al. (2008), research shows that this is not true. In fact, education is very important for the development of entrepreneurs as it helps them to cope with the problems they encounter. On the other hand, formal education is not so important in the start-up phase. Also, there are cases where successful entrepreneurs have only a high school diploma. The next question is whether there is a link between entrepreneurship education, some form of training and entrepreneurship education, and the subsequent development of entrepreneurial intentions. According to Hoppe (2016), the introduction of entrepreneurship education in the Swedish education system has significant effects. The researchers (Bergmann, Hundt, & Sternberg, 2016), state that when students attend some form of entrepreneurial education, this may encourage actions to open their own businesses. However, in the study of do Paço, et al. (2015), a comparison of entrepreneurial intentions was made between girls attending business school and boys attending sports school. It was shown that boys have stronger intentions to start their own businesses even though they do not actually receive any entrepreneurial education at school. There does not seem to be a clear answer to this question either: special entrepreneurship education is important, but in some cases it may not have a decisive influence. An interesting study conducted among Romanian students (Luca, Cazan, & Tomulescu, 2013) found that students who are aware of their entrepreneurial potential are more likely to engage in entrepreneurial training. Also, such training will have greater effects on such students. Perhaps these results are an indicator that entrepreneurial training and education should be selectively applied to students who show certain preferences and the desire for entrepreneurship, and thus the effects of these actions will be greater.

Previous experience (work and life) For most entrepreneurs their most significant entrepreneurial venture was not their first (Hisrich et al., 2008). Namely, entrepreneurs constantly find new ideas and opportunities, and thus start more ventures throughout their careers. Previous experience helps them to gain a better understanding of the circumstances, to improve their assessment of risks and opportunities, to better anticipate certain possible situations, and make comparisons with previous jobs. In these circumstances, it is clear that previous entrepreneurial experience helps the entrepreneur and increases the chances of the venture’s success. 130

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

According to Hatak et al. (2015), previous entrepreneurial experience has an influence on entrepreneurial intentions. Employees who were previously entrepreneurs have higher entrepreneurial intentions than other employees (Hsu et al., 2017). That desire is greater if they have been entrepreneurs for the longer part of their careers. In the study of Miralles et al. (2016) it was also found that individuals with entrepreneurial knowledge and previous entrepreneurial experience have stronger entrepreneurial intentions. Previous experience with social problems predicts entrepreneurial intentions in the field of social entrepreneurship (Hockerts, 2017). In terms of the impact of work experience and, in particular, previous entrepreneurial experience on entrepreneurial intentions and success, there appears to be a consistency in the research findings as well as the authors' agreement that this impact is positive. Based on the previous considerations, as well as the fact that the analysis is performed according to seven control variables, in this paper, seven hypotheses are posed: H1: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to gender of employed persons. H2: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to age of employed persons. H3: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to level of education of employed persons. H4: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to ownership structure of the respondents’ enterprise. H5: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to previous experience in entrepreneurship of employed persons. H6: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to perceived job performance of employed persons. H7: There are significant differences in the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions, according to perceived finances of employed persons.

131

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

METHOD

Survey instruments (measures) The Individual Entrepreneurial Orientation (IEO) instrument was used to measure individual entrepreneurial orientation (Bolton & Lane 2012). The questionnaire contains 3 dimensions, consisting of 10 items: 1. Risk-taking, 2. Innovativeness and 3. Proactiveness. The respondents evaluated them on a seven-point Likert scale. The Cronbach's alpha values are: a=0,798 (risk-taking), a=0,848 (innovativeness) and a=0,820 (proactiveness). The need for achievement was measured by means of the achievement dimension of the Attitude Toward Enterprise (ATE) Test (Athayde, 2009). This dimension consists of 4 items. The respondents rated them using a seven-point Likert scale. The Cronbach's alpha value for achievement is a=0,866. The Entrepreneurial Intention Questionnaire (EIQ) instrument was used to measure the dimensions of the Theory of Planned Behavior (Liñán & Chen 2009). The questionnaire includes 4 dimensions which are made up of 20 items: 1. Personal Attitude, 2. Subjective Norm, 3. Perceived Behavioral Control and 4. Entrepreneurial Intention. The respondents evaluated them on a seven-point Likert scale. The Cronbach's alpha values are: a=0,906 (personal attitude), a=0,807 (subjective norm), a=0,898 (perceived behavioral control) and a=0,954 (entrepreneurial intention).

Participants and data collection The research was conducted in companies in Serbia. The survey covered medium and large enterprises, and according to the type of activity, production, service and public companies were included. The respondents were employed in these enterprises, and were of mixed gender, age, level of education (minimum secondary school) and position in the organization. In most of the companies surveyed, several questionnaires were distributed. The respondents then filled in the questionnaires. The sample includes 540 people and the survey included 72 companies.

A sample structure according to the observed variables According to the observed variables (gender of respondents, years of respondents, level of education, ownership structure of the company, previous experience in entrepreneurship, perceived success at work, their perceived finances), the structure of the 540 respondents sample is as follows: 1. There were NM = 285 men and NF = 255 women in the sample. 2. The sample was represented by employees between 20 and 64 years of age. By age, respondents are divided into those between 20 and 42 years of age (younger respondents) and those between 43 and 64 years of age (older respondents). Thus, the younger respondents were NY = 273 and the older respondents were NO = 267. 3. In the sample there were 90 respondents with secondary education, 59 respondents with tertiary education and 391 respondents with higher education. The respondents with tertiary education are added to the respondents with secondary education. Thus, in the sample there were NHS = 149 respondents with secondary and tertiary education and NFAC = 391 respondents with higher education. 132

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

4. According to the ownership structure, there were NST = 350 state-owned enterprises and NPR = 190 private enterprises in the sample.

5. According to the previous entrepreneurial experience of the respondents, the sample included NYEX = 141 respondents with entrepreneurial experience and NNEX = 399 respondents with no entrepreneurial experience. 6. Perceptions of their own performance at work were rated by the respondents with a score of 1 to 7. In group 1 - 1 was very unsuccessful, in group 2 - 7 were unsuccessful, in group 3 - slightly unsuccessful (3), 4 - averagely successful (74), 5 - mildly successful (49), 6 - successful (266) and 7 - very successful (140). Because the respondents perceived their performance as relatively high, according to this variable, the sample was divided into those who expressed their performance at work with grades 1 to 5 (low performance) and those who showed their performance at work with grades 6 and 7 (high performance). Thus, in the sample, there were NLSUC = 134 low performance at work and NHSUC = 406 high performance. 7. The perceptions of owning finances for starting a private business were rated by the respondents from 1 to 7. In group 1 - very low was 203 respondents, in group 2 - low was 80 respondents, in group 3 - slightly low (32) , 4 - average (131), 5 - slightly high (38), 6 - high (35) and 7 - very high (21). Given that respondents perceived their own finances to start their own business as relatively low, according to this variable, the sample was divided into those who expressed their financial possession with grades 1 to 3 (low financial ownership) and those who expressed their financial ownership with grades 4 to 7 (high financial ownership). Thus, the sample had NLFIN = 315 respondents with low financial ownership to start an own business and NHFIN = 225 respondents with high financial ownership to start their own business.

RESULTS Above average scores of individual entrepreneurial orientation dimensions, achievement dimension and the theory of planned behavior dimensions, t-test was performed. In doing so, the variables were: 1. Gender of the respondents (GEN). 2. Age of the respondents (YEA). 3. Level of education of the respondents (level of education - EDU). 4. Ownership structure of the respondents’ organization (OWN). 5. Previous experience in entrepreneurship (EEX). 6. Perceived performance at work (SUC). 7. Perceived possession of finances of the respondent (FIN). The results of the t-test are presented in Tables 1 to 7. The aim of this analysis is to determine if there are statistically significant differences in the average ratings of individual entrepreneurial orientation dimensions, need for achievement, and theory of planned behavior, depending on the variables indicated. Also, the goal is to determine the directions of these differences, that is, in which conditions the observed dimensions have higher. In Tables 1 to 7, the average scores with a statistically significant difference are shown in bold font and shaded fields. 133

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Table 1. T-test with the average scores for men (M) and women (F)

GEN

RT IN PR ACH PA SN PBC EI

Mean

Std. Deviation

Std. Error Mean

285

4.646784

1.4663660

0.0868600

255

4.538562

1.5189582

0.0951209

285

4.868421

1.2668801

0.0750435

255

4.868627

1.3575380

0.0850123

285

5.804678

1.1460526

0.0678863

255

5.730719

1.1254598

0.0704790

285

5.275439

1.1633761

0.0689125

255

5.150980

1.2200140

0.0764002

285

4.760702

1.3832497

0.0819366

255

4.328627

1.4288658

0.0894790

285

5.097076

1.3013818

0.0770872

255

5.048366

1.3190892

0.0826046

285

4.347368

1.3072560

0.0774352

255

4.107190

1.3389206

0.0838465

285

3.477778

1.5593041

0.0923652

255

3.150980

1.6700013

0.1045795

Levene's Test for Equality of Variances F

Sig.

0.146

0.702

0.834

0.362

0.086

0.769

1.877

0.171

0.230 0.632 0.020

0.888

0.000 0.999 5.514 0.019

t-test for Equality of Means t

Sig. (2-tailed)

Mean Difference

0.842

538

0.400

0.1082215

0.840

526.682

0401

-0.002

538

0.999

-0.002 521.093

0.999

0.755

538

0.451

0.756

533.350

0.450

1.213

538

0.226

1.210

524.768

0.227

3.568

538

0.000

3.561

527.098

0.000

0.431

538

0.666

0.431

529.726

0.667

2.107

538

0.036

2.104

528.321

0.036

2.351

538

0.019

2.342

521.188

0.020

-0.0002064 0.0739594 0.1244582 0.4320743 0.0487100 0.2401789 0.3267974

Table 2. T-test with the average scores for younger (Y) and older (O) respondents

YEA

RT IN PR ACH PA SN PBC EI

134

Mean

Std. Deviation

Std. Error Mean

273

4.859585

1.3377007

0.0809613

267

4.325843

1.5907346

0.0973514

273

5.090659

1.1904085

0.0720468

267

4.641386

1.3863258

0.0848418

273

5.766789

1.0773534

0.0652044

267

5.772784

1.1948686

0.0731248

273

5.276557

1.1280514

0.0682728

267

5.155431

1.2512019

0.0765723

273

4.738462

1.3274342

0.0803400

267

4.370787

1.4887702

0.0911113

273

5.347985

1.1299316

0.0683866

267

4.794007

1.4176112

0.0867564

273

4.368742

1.2346971

0.0747273

267

4.096130

1.4032118

0.0858752

273

3.600122

1.6175192

0.0978967

267

3.040574

1.5744267

0.0963534

Levene's Test for Equality of Variances F

Sig.

10.660 0.001 4.596

0.032

5.063

0.025

3.540

0.060

2.052

0.153

12.523 0.000 1.979 0.073

0.160 0.787

t-test for Equality of Means df

Sig. (2-tailed)

Mean Difference

4.223

538

0.000

0.5337422

4.215

518.591

0.000

4.043

538

0.000

4.036

522.332

0.000

-0.061

538

0.951

-0.061 529.684

0.951

1.182

538

0.238

1.181

529.673

0.238

3.031

538

0.003

3.027

528.185

0.003

5.027

538

0.000

5.015

507.604

0.000

2.398

538

0.017

2.395

526.266

0.017

4.072

538

0.000

4.074

537.988

0.000

0.4492736 -0.0059953 0.1211261 0.3676750 0.5539779 0.2726125 0.5595478

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Table 3. T-test with the average scores for respondents with secondary (HS) and higher (FAC) education Std. Deviation

Std. Error Mean

149 4.655481

1.4474873

0.1185828

FAC

391 4.572890

1.5084807

0.0762871

149 4.894295

1.2758083

0.1045183

FAC

391 4.858696

1.3232443

0.0669193

149 5.677852

1.2504750

0.1024429

FAC

391 5.804774

1.0887772

0.0550618

149 5.087248

1.2476079

0.1022080

FAC

391 5.265985

1.1665340

0.0589942

149 4.719463

1.4753688

0.1208669

FAC

391 4.494629

1.3955044

0.0705737

149 5.174497

1.3006027

0.1065495

FAC

391 5.035806

1.3115190

0.0663264

149 4.209172

1.3595611

0.1113796

FAC

391 4.243393

1.3153470

0.0665200

149 3.418345

1.6555505

0.1356280

FAC

391 3.150980

1.6700013

0.1045795

EDU

RT IN PR ACH PA SN PBC EI

Mean

Levene's Test for Equality of Variances F

Sig.

0.207

0.649

0.412

0.521

2.895

0.089

0.002

0.965

1.826

0.177

0.067

0.795

0.677

0.411

0.014

0.907

t-test for Equality of Means t

Sig. (2-tailed)

Mean Difference

0.575

538

0.566

0.0825910

0.586

277.797

0.559

0.282

538

0.778

0.287

276.575

0.774

-1.161

538

0.246

-1.091

238.317

0.276

-1.561

538

0.119

-1.515

252.410

0.131

1.647

538

0.100

1.606

254.877

0.109

1.101

538

0.271

1.105

269.562

0.270

-0.268

538

0.789

-0.264

259.864

0.792

0.840

538

0.401

2.342

521.188

0.020

0.0355996 -0.1269217 -0.1787363 0.2248339 0.1386910 -0.0342207 0.1310470

Table 4. T-test with the average scores for respondents employed in state (ST) and private (PR) enterprises

OWN

RT IN PR ACH PA SN PBC EI

Mean

Std. Deviation

Std. Error Mean

Levene's Test for Equality of Variances F

350 4.393333

1.5011043

0.0802374

190 4.968421

1.4013294

0.1016631

350 4.700000

1.3026862

0.0696315

190 5.178947

1.2670338

0.0919203

350 5.701905

1.1556773

0.0617736

190 5.894737

1.0905647

0.0791179

350 5.148571

1.1856947

0.0633780

190 5.342105

1.1935966

0.0865926

350 4.353143

1.3691144

0.0731822

190 4.931579

1.4396328

0.1044419

350 4.824762

1.2547609

0.0670698

190 5.533333

1.2847406

0.0932049

350 4.149048

1.3117004

0.0701133

190 4.390351

1.3428214

0.0974185

350 3.091905

1.4788269

0.0790466

190 3.750000

1.7768475

0.1289061

Sig.

1.109 0.293 0.001 0.980 3.502

0.062

1.130

0.288

2.329 0.128 0.028 0.868 0.948 0.331 5.741 0.017

t-test for Equality of Means t

Sig. (2-tailed)

Mean Difference -0.5750877

-4.351

538

0.000

-4.440

411.358

0.000

-4.119

538

0.000

-4.153

397.298

0.000

-1.888

538

0.060

-1.921

407.634

0.055

-1.807

538

0.071

-1.804

385.777

0.072

-4.604

538

0.000

-4.536

371.635

0.000

-6.214

538

0.000

-6.171

380.204

0.000

-2.024

538

0.043

-2.010

380.257

0.045

-4.593

538

0.000

-4.352

332.409

0.000

-0.4789474 -0.1928321 -0.1935338 -0.5784361 -0.7085714 -0.2413033 -0.6580952

135

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Table 5. T-test with the average scores for respondents with prior entrepreneurial experience (YEX) and respondents with no previous entrepreneurial experience (NEX)

EEX

RT IN PR ACH PA SN PBC EI

Mean

Std. Deviation

Std. Error Mean

YEX

141 4.758865

1.4075231

0.1185348

NEX

399 4.538012

1.5169397

0.0759420

YEX

141 4.902482

1.2623579

0.1063097

NEX

399 4.856516

1.3267617

0.0664212

YEX

141 5.825059

1.1605838

0.0977388

NEX

399 5.750209

1.1279082

0.0564660

YEX

141 5.363475

1.1867856

0.0999454

NEX

399 5.164787

1.1895987

0.0595544

YEX

141 4.870922

1.3380444

0.1126837

NEX

399 4.445614

1.4332816

0.0717538

YEX

141 5.234043

1.2766180

0.1075106

NEX

399 5.017544

1.3168893

0.0659269

YEX

141 4.723404

1.2425699

0.1046433

NEX

399 4.060986

1.3134132

0.0657529

YEX

141 3.693853

1.5940893

0.1342465

NEX

399 3.192565

1.6097563

0.0805886

Levene's Test for Equality of Variances F

Sig.

0.675

0.412

0.217

0.642

0.253

0.616

0.195

0.659

0.138 0.710 0.143

0.705

0.005 0.946 0.665 0.415

t-test for Equality of Means df

Sig. (2-tailed)

Mean Difference

1.514

538

0.131

0.2208536

1.569

262.934

0.118

0.358

538

0.720

0.367

256.866

0.714

0.672

538

0.502

0.663

239.659

0.508

1.706

538

0.089

1.708

246.151

0.089

3.081

538

0.002

3.184

261.432

0.002

1.691

538

0.091

1.717

252.524

0.087

5.220

538

0.000

5.360

258.217

0.000

3.187

538

0.002

3.202

247.765

0.002

0.0459660 0.0748502 0.1986882 0.4253080 0.2164987 0.6624185 0.5012887

Table 6. T-test with the average scores for respondents who perceive their job performance as low (LSUC) and high (HSUC)

SUC

RT IN PR ACH PA SN PBC EI

136

Mean

Std. Deviation

Std. Error Mean

134 4.176617

1.3913987

0.1201985

HSUC 406 4.733990

1.4985533

0.0743720

134 4.526119

1.2263816

0.1059432

HSUC 406 4.981527

1.3174859

0.0653857

LSUC LSUC

134 5.291045

1.2380906

0.1069547

HSUC 406 5.927750

1.0549935

0.0523584

134 4.565299

1.1495051

0.0993021

HSUC 406 5.431650

1.1256447

0.0558648

LSUC LSUC

134 4.092537

1.3946529

0.1204797

HSUC 406 4.709852

1.3966632

0.0693152

134 4.644279

1.2500277

0.1079859

HSUC 406 5.215928

1.2981609

0.0644266

134 3.723881

1.1945477

0.1031932

HSUC 406 4.402299

1.3261763

0.0658170

134 2.941542

1.4061058

0.1214690

HSUC 406 3.449507

1.6661673

0.0826905

LSUC LSUC LSUC LSUC

Levene's Test for Equality of Variances Sig.

0.727 0.394 0.330 0.566 8.776 0.003 0.051 0.822 0.322 0.571 0.252 0.616 2.331 0.127 3.233 0.073

t-test for Equality of Means t

-3.799

538

Sig. (2-tailed)

Mean Difference

0.000

-0.5573732

-3.943 242.650

0.000

-3.528

0.000

538

-3.658 242.084 -5.794

538

0.000 0.000

-5.347 200.601

0.000

-7.685

0.000

538

-7.604 223.167 -4.438

538

0.000 0.000

-4.460

0.000

-4.546 234.768

0.000

-5.259

0.000

538

-5.543 249.649

0.000

-3.175

0.002

538

-3.457 266.070

-0.6367056 -0.8663517

0.000

-4.441 227.435 538

-0.4554077

0.001

-0.6173149 -0.5716491 -0.6784183 -0.5079651

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Table 7. T-test with the average scores for respondents who perceive their finances as low (LFIN) and high (HFIN)

FIN

RT IN PR ACH PA SN PBC EI

Mean

Std. Deviation

Std. Error Mean

LFIN

315 4.431746

1.5111689

0.0851447

HFIN

225 4.825185

1.4344326

0.0956288

LFIN

315 4.752381

1.3484312

0.0759755

HFIN

225 5.031111

1.2370440

0.0824696

LFIN

315 5.761905

1.1180001

0.0629922

HFIN

225 5.780741

1.1629722

0.0775315

LFIN

315 5.166667

1.1642567

0.0655984

HFIN

225 5.286667

1.2265843

0.0817723

LFIN

315 4.401270

1.4663090

0.0826172

HFIN

225 4.774222

1.3257924

0.0883862

LFIN

315 4.987302

1.3295518

0.0749118

HFIN

225 5.195556

1.2721373

0.0848092

LFIN

315 3.907407

1.3027876

0.0734038

HFIN

225 4.691111

1.2227263

0.0815151

LFIN

315 3.004762

1.6588122

0.0934635

HFIN

225 3.769630

1.4523316

0.0968221

Levene's Test for Equality of Variances F

Sig.

1.068

0.302

3.084

0.080

0.000

0.988

0.610

0.435

1.733

0.189

0.693

0.406

0.251

0.617

14.297 0.000

t-test for Equality of Means t

Sig. (2-tailed)

Mean Difference

-3.046

538

0.002

-0.3934392

-3.073

497.076

0.002

-2.450

538

0.015

-2.486

505.711

0.013

-0.190

538

0.850

-0.189

470.944

0.851

-1.155

538

0.249

-1.145

467.080

0.253

-3.031

538

0.003

-3.083 509.151

0.002

-1.827

538

0.068

-1.840

494.955

0.066

-7.069

538

0.000

-7.144 500.021

0.000

-5.560

0.000

538

-5.684 516.207

-0.2787302 -0.0188360 -0.1200000 -0.3729524 -0.2082540 -0.7837037 -0.7648677

0.000

DISCUSSION The respondents’ gender (GEN) According to Table 1, statistically significant differences in the average scores occur in three dimensions: PA - Attitude Towards Entrepreneurship, PBC - Perceived Behavioral Control, and EI - Entrepreneurial Intentions. These dimensions have higher average scores (statistically significantly higher) among men. Thus, men have a more pronounced attitude toward entrepreneurship, perceived behavioral control, and entrepreneurial intentions than women. This is most prominent for the PA dimension - Attitude Towards Entrepreneurship. It is yet to be noted that the difference in the mean scores between attitude and intention is greater for men than for women. Therefore, women have more specific entrepreneurial thinking: if they have a positive attitude towards entrepreneurship, they are more likely to have entrepreneurial intentions. Based on these results, it can be said that hypothesis H1 is partially confirmed. These results can be interpreted by the fact that in society, men are often more encouraged to start an entrepreneurial venture, and women are less likely to think that owning a business is an acceptable career option. These results are in line with previous research that shows that entrepreneurial intentions are higher among men than women (Leppel, 2016; Santos et al., 2016; Shneor et al., 2013).

137

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

The respondents’ age (YEA)

According to Table 2, a statistically significant difference in the average scores occurs across most dimensions: RT - Risk Taking, IN - Innovation, PA - Attitude Towards Entrepreneurship, SN - Subjective Norm, PBC – Perceived Behavioral Control and EI - Entrepreneurial Intentions. These dimensions have higher average scores (statistically significantly higher) among the younger subjects. Thus, the younger respondents have a more pronounced preference for risk, innovation, attitude toward entrepreneurship, support from the environment, perceived behavioral control and entrepreneurial intentions than their older counterparts. This is strongly expressed in all dimensions where there is a statistically significant difference. The older respondents can match the younger ones only when it comes to PR - Proactivity and ACH - Need for Achievement. It should also be noted that the difference in the mean scores between attitude and intentions is smaller in the younger respondents. Therefore, the younger respondents are more specific and realistic in their entrepreneurial thinking: if they have a positive attitude towards entrepreneurship, they are more likely to have entrepreneurial intentions. Based on these results, it can be said that hypothesis H2 is confirmed. Such results can be interpreted in light of the fact that the younger respondents have more energy, willpower, courage and confidence to start an entrepreneurial venture and are readier to take risks. It is possible that in the process of starting a business, younger people are more able to count on the help of their parents, especially for financial support. In addition, younger people always have time to try again. Based on the above, this research can be categorized as a reference that shows that entrepreneurial intentions are higher among younger people than older people, for example (Minola et al., 2016; Tsai et al., 2016; Sahut et al., 2015; Hatak et al., 2015). The respondents’ education level (EDU) Table 3 shows no statistically significant differences in the mean ratings for the observed dimensions. However, there are certain tendencies that are not statistically significant but may deserve to be listed. Thus, the respondents who have completed higher education have a slightly higher average score for the dimensions PR - Proactivity and ACH - Need for Achievement, while those with secondary education have a slightly higher average score for the PA dimensions - Attitude Toward Entrepreneurship and EI - Entrepreneurial Intentions. The respondents with higher education are somewhat more proactive and in greater need of achievement, as confirmed through education. Those with secondary education see a greater chance to succeed in entrepreneurship, probably because without a higher education diploma it is difficult to achieve some of their ambitions in another way. However, it should be concluded here that the degree of the observed dimensions of entrepreneurial performance does not depend significantly on the level of education. Based on these results, it can be concluded that hypothesis H3 has not been confirmed. The issues related to entrepreneurship education are discussed in the theoretical part of the paper. However, the existing literature does not provide a clear answer to the following questions: what is the level of entrepreneurship education, how important is education for entrepreneurship, and does education increase the chances of starting an entrepreneurial venture and the subsequent success of that business? There are practical examples that give different answers to these questions. It can be concluded here that education is not so important in the initial stage of starting an entrepreneurial 138

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

venture, but it is important in the stage of running and maintaining a business. The results obtained here somehow confirm the previous findings and the rather hesitant results of previous research on this topic (Hoppe, 2016; Bergmann et al., 2016; do Paço et al., 2015; Luca et al., 2013). The ownership structure of the respondents’ organizations (OWN) According to Table 4, a statistically significant difference in the average scores occurs across most dimensions: RT - Risk Taking, IN - Innovation, PA - Attitude Towards Entrepreneurship, SN - Subjective Norm, PBC - Perceived Behavioral Control and EI - Entrepreneurial Intentions. These dimensions have higher average scores (statistically significantly higher) among the respondents employed in private companies. Thus, the respondents employed in private enterprises have a greater preference for risk, innovation, attitude toward entrepreneurship, support from the environment, perceived behavioral control, and entrepreneurial intentions than their counterparts who work in state-owned enterprises. This is very pronounced for all the dimensions where there is a statistically significant difference, with the exception of the PBC dimension - Perceived Behavioral Control. Although there is no statistically significant difference in the average grades for the dimensions PR - Proactivity and ACH - Need For Achievement, employees in private companies also have higher scores in these two cases. Based on these results, it can be said that hypothesis H4 is confirmed. Such results can be interpreted (generally speaking) as a job in a private company is not as secure as one in a state-owned company. Private enterprise employees are aware that any mistakes and / or poor performance can place them in the position where they may lose their job. Consequently, they take on more risk (accustomed to risk), and are forced to be innovative and proactive. In addition, employees in private companies have a director who is an entrepreneur, they are accustomed to such an environment, and may be convinced from the immediate vicinity of the benefits of an entrepreneurial call and have the knowledge and skills to start and run their own private business. As a result, the observed dimensions have a higher average rating for those respondents employed in private companies. These results cannot be compared with other studies, because there is very little research which addresses the entrepreneurial attitudes and intentions of employees, especially as a function of the ownership structure of an organization. The previous experience in entrepreneurship (EEX) According to Table 5, a statistically significant difference in the average scores occurs for three dimensions: PA - Attitude Towards Entrepreneurship, PBC – Perceived Behavioral Control, and EI - Entrepreneurial Intentions. These dimensions have higher average scores (statistically significantly higher) among those respondents with previous entrepreneurial experience. Thus, the respondents with previous entrepreneurial experience have a more pronounced attitude towards entrepreneurship, perceived behavioral control and entrepreneurial intentions than the respondents without previous entrepreneurial experience. This is very strongly expressed in all dimensions where there is a statistically significant difference. Based on these results, it can be said that hypothesis H5 is partially confirmed. Such results are easy to understand: respondents who have previous experience in entrepreneurship are far more experienced in this regard, know what to expect, are more able to recognize their strengths and weaknesses for the job, know where they have gone wrong and what they are able to resolve. 139

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

In addition, they have already shown that they have an entrepreneurial spirit and are prone to starting a private business. Existing references (Hatak et al., 2015; Hsu, Shinnar, Powell, & Coffey, 2017; Miralles, Giones, & Riverola, 2016), confirm that previous experience in entrepreneurship has a positive impact on entrepreneurial intentions. The respondents’ perceived performance (SUC) According to Table 6, a statistically significant difference in the average scores occurs across all dimensions: RT - Risk Taking, IN - Innovation, PR - Proactivity, ACH - Need for Achievement, PA Attitude Toward Entrepreneurship, SN - Subjective Norm, PBC - Perceived Behavioral Control and EI - Entrepreneurial Intentions. All of these dimensions have higher average scores (statistically significantly higher) among those respondents who perceive their job performance as high. This is very strongly expressed for all dimensions where there is a statistically significant difference. Based on these results, it can be said that hypothesis H6 is confirmed. It should be borne in mind that most of the respondents rated their performance as very high (on a scale of 1 to 7, as many as 406 respondents rated this variable with grades 6 and 7). The respondents who rated their performance low are obviously people who do not have high ambitions, do not make particularly great efforts at work, and who seem to accept this situation relatively easily. Such people certainly have low ratings for the observed dimensions of individual entrepreneurial performance. A similar result can be seen in the previous research (Dechawatanapaisal, 2018), where it was shown that respondents who have a high degree of self-efficacy show greater intentions to leave the organization in the event of adverse circumstances within the organization. The respondents’ perceived finances (FIN) According to Table 7, a statistically significant difference in the average scores occurs across most of the dimensions: RT - Risk Aversion, IN - Innovation, PA - Attitude Toward Entrepreneurship, PBC - Perceived Behavioral Control and EI - Entrepreneurial Intentions. The dimensions mentioned above have higher average scores (statistically significantly higher) among the respondents who perceive their financial situation as high. Thus, respondents who perceive their finances as high have a greater risk appetite, greater innovation, a better attitude toward entrepreneurship, and stronger perceived behavioral control and entrepreneurial intentions than those respondents who perceive their financial situation as low. This is very strongly expressed for all of the dimensions, but especially for PBC - Perceived Behavioral Control and EI - Entrepreneurial Intentions. Also, the difference in the average scores between attitude and intentions is smaller for the respondents who perceive their financial standing as high. Accordingly, respondents who perceive their own finances as high are significantly more prepared and determined to concretize their positive attitudes towards entrepreneurship through entrepreneurial intentions. Based on these results, it can be said that hypothesis H7 is confirmed. This situation is understandable, since the perception of a good financial position certainly gives an individual more security, greater confidence, and more room for risk taking, and, finally, having finances makes it easier to cope with the costs at the very beginning of an entrepreneurial venture. Some existing research has shown that having finances can be important for starting an entrepreneurial venture (Iakovleva, Kolvereid, Gorgievski, &, Sørhaug, 2014; Kim, Longest, & Aldrich, 2013; Rajković, 140

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Nikolić, Ćoćkalo, Terek, & Božić, 2020), and that potential financial difficulties significantly prevent individuals from leaving their current jobs (Virick, Basu, & Rogers, 2015).

CONCLUSION Based on the results of the t-test and previous analyses of all seven observed variables, the profile of the employee who, potentially, has the greatest chances of becoming an entrepreneur emerges. The profile is as follows: a younger man with a high school degree (education does not have such a significant impact), who is employed in a private company, has previous entrepreneurial experience, is successful in his job and has the appropriate finances. Theoretically speaking, such a defined profile may serve to quickly check one's entrepreneurial attitudes and intentions: the greater the number of characteristics which coincide with the characteristics of an "ideal" profile a person has, the more likely that person is to have entrepreneurial intentions. Note that this applies to employees. From the seven hypotheses posed, four were confirmed, two were partially confirmed, while one hypothesis was rejected. Taking into account the previous presentations, a general conclusion can be drawn: a statistically significant difference in the influence of the observed control variables (the respondents’ gender, their age, the respondents’ level of education, the ownership structure of the company, the respondents’ previous experience in entrepreneurship, their perceived success at work, and their perceived finances) on the level of the individual entrepreneurial orientation dimensions, the achievement dimension and the theory of planned behavior dimensions among employed persons, exists in most cases, with the exception of the respondents’ level of education. As the research was carried out in organizations in Serbia, this fact can be considered as a limitation of this research. Specifically, the results obtained apply primarily to organizations in this country. However, it is assumed with a high degree of certainty that similar results could be obtained in some other countries. This is especially true of countries in the region, as well as other countries in transition. The theoretical significance of this research stems from the fact that there is an insufficient number of papers dealing with the entrepreneurial intentions of employed persons, and especially the effects of the control variables observed here under such conditions. The practical significance of this research is that it shows that employees should be considered as potential entrepreneurs. Also, the survey indicated which groups, among employed persons, have the highest chances of becoming entrepreneurs. This should certainly be kept in mind when defining future state-level entrepreneurship promotion strategies and programs, as well as when directing entrepreneurial training and deciding on financial support for employed persons with entrepreneurial intentions.

ACKNOWLEDGEMENT This paper was supported by the Provincial Secretariat for Science and Technological Development, Autonomous Province of Vojvodina, project number: 142-451-2139/2019-01.

141

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

REFERENCES

Altinay, L., Madanoglu, M., Daniele, R., & Lashley, C. (2012). The influence of family tradition and psychological traits on entrepreneurial intention. International Journal of Hospitality Management, 31(2), 489-499. Athayde, R. (2009). Measuring Enterprise Potential in Young People. Entrepreneurship Theory and Practice, 33(2), 481-500. Bender, K.A., & Roche, K. (2016). Self-employment and the paradox of the contented female worker. Small Business Economics, 47(2), 421-435. Bergmann, H., Hundt, C., & Sternberg, R. (2016). What makes student entrepreneurs? On the relevance (and irrelevance) of the university and the regional context for student start-ups. Small Business Economics, 47(1), 53-76. Bolton, D.L., & Lane, M.D. (2012). Individual Entrepreneurial Orientation: development of a measurement instrument. Education + Training, 54(2/3), 219-233. Conroy, T., & Weiler, S. (2016). Does gender matter for job creation? Business ownership and employment growth. Small Business Economics, 47(2), 397-419. Dechawatanapaisal, D. (2018). The moderating effects of demographic characteristics and certain psychological factors on the job embeddedness – turnover relationship among Thai health-care employees. International Journal of Organizational Analysis, 26(1), 43-62. do Paço, A., Matos Ferreira, J., Raposo, M., Gouveia Rodrigues, R., & Dinis, A. (2015). Entrepreneurial intentions: is education enough? International Entrepreneurship and Management Journal, 11(1), 57-75. Espíritu-Olmos, R., & Sastre-Castillo, M.A. (2015). Personality traits versus work values: Comparing psychological theories on entrepreneurial intention. Journal of Business Research, 68(7), 1595-1598. Gawke, J.C., Gorgievski, M.J., & Bakker, A.B. (2017). Employee intrapreneurship and work engagement: A latent change score approach. Journal of Vocational Behavior, 100, 88-100. Hatak, I., Harms, R., & Fink, M. (2015). Age, job identification, and entrepreneurial intention. Journal of Managerial Psychology, 30(1), 38-53. Hechavarria, D.M., Terjesen, S.A., Ingram, A.E., Renko, M., Justo, R., & Elam, A. (2017). Taking care of business: the impact of culture and gender on entrepreneurs’ blended value creation goals. Small Business Economics, 48(1), 225-157. Hisrich, R.D., Peters, M.P., & Shepherd, D.A. (2008). Entrepreneurship (7th ed.). Zagreb: Mate (in Croatian). Hockerts, K. (2017). Determinants of Social Entrepreneurial Intentions. Entrepreneurship: Theory and Practice, 41(1), 105-130. Hoppe, M. (2016). Policy and entrepreneurship education. Small Business Economics, 46(1), 13-29. Hormiga, E., Hancock, C., & Valls-Pasola, J. (2013). The relationship between employee propensity to innovate and their decision to create a company. Management Decision, 51(5), 938-953. Hornsby, J., Peña-Legazque, I., & Guerrero, M. (2013). Guest editorial: the role of corporate entrepreneurship in the current organizational and economic landscape. International Entrepreneurship and Management Journal, 9(3), 295-305. Hsu, D.K., Shinnar, R.S., Powell, B.C., & Coffey, B.S. (2017). Intentions to reenter venture creation: The effect of entrepreneurial experience and organizational climate. International Small Business Journal: Researching Entrepreneurship, 35(8), 928–948. Iakovleva, T.A., Kolvereid, L., Gorgievski, M.J., & Sørhaug, Ø. (2014). Comparison of perceived barriers to entrepreneurship in Eastern and Western European countries. International Journal of Entrepreneurship & Innovation Management, 18(2/3), 115-133. Jain, R., & Ali, S.W. (2012). Personal Characteristics of Indian Entrepreneurs and Intrapreneurs An Empirical Study. Management and Labour Studies, 37(4), 295-322. 142

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Jayawarna, D., Jones, O., & Marlow, S. (2015). The influence of gender upon social networks and bootstrapping behaviours. Scandinavian Journal of Management, 31(3), 316-329. Kean, R.C., Van Zandt, S., & Maupin, W. (2008). Successful Aging: The Older Entrepreneur. Journal of Women & Aging, 5(1), 25-42.

Kim, P.H., Longest, K.C., & Aldrich, H.E. (2013). Can You Lend Me a Hand? Task-Role Alignment of Social Support for Aspiring Business Owners. Work and Occupations, 40(3), 213-249. Knorr, H., Garzón, D., & Martinez, D. (2011). Motivations and differences upon reconciling professional and personal life: an empirical study of businesswomen and businessmen in the Valencian Community. International Entrepreneurship and Management Journal, 7(3), 391-412. Kwong, C., & Thompson, P. (2016). The When and Why: Student Entrepreneurial Aspirations. Journal of Small Business Management, 54(1), 299-318. Leppel, K. (2016). The incidence of self-employment by sexual orientation. Small Business Economics, 46(3), 347-363. Liñán, F., & Chen, Y.W. (2009). Development and Cross-Cultural Application of a Specific Instrument to Measure Entrepreneurial Intention. Entrepreneurship Theory and Practice, 33(3), 593-617. Luca, M.R., Cazan, A.M., & Tomulescu, D. (2013). Entrepreneurial personality in higher education. Procedia Social and Behavioral Sciences, 84, 1045-1049. Marshall, D.R., & Gigliotti, R. (2018). Bound for entrepreneurship? A career-theoretical perspective on entrepreneurial intentions. International Entrepreneurship and Management Journal, 16(1), 287-303. Minola, T., Criaco, G., & Obschonka, M. (2016). Age, culture, and self-employment motivation. Small Business Economics, 46(2), 187-213. Miralles, F., Giones, F., & Riverola, C. (2016). Evaluating the impact of prior experience in entrepreneurial intention. International Entrepreneurship and Management Journal, 12(3), 791-813. Miralles, F., Giones, F., & Gozun, B. (2017). Does direct experience matter? Examining the consequences of current entrepreneurial behavior on entrepreneurial intention. International Entrepreneurship and Management Journal, 13(3), 881-903. Mohedano-Suanes, A., & Benitez, D.G. (2018). Intrapreneurs: Characteristics and Behavior. In Tur Porcar A., Ribeiro Soriano D. (Eds.), Inside the Mind of the Entrepreneur. (pp. 109-119). Cham: Springer. Perenyi, A., Zolin, R., & Maritz, A. (2018). The perceptions of Australian senior entrepreneurs on the drivers of their entrepreneurial activity. International Journal of Entrepreneurial Behavior & Research, 24(1), 81-103. Poggesi, S., Mari, M., & De Vita, L. (2016). What’s new in female entrepreneurship research? Answers from the literature. International Entrepreneurship and Management Journal, 12(3), 735-764. Rajković, J., Nikolić, M., Ćoćkalo, D., Terek, E., Božić, S. (2020). National culture and the entrepreneurial intentions of students in Serbia. Journal for East European Management Studies, 25(1), 105-141. Sahut, J.M., Gharbi, S., & Mili, M. (2015). Identifying factors key to encouraging entrepreneurial intentions among seniors. Canadian Journal of Administrative Sciences, 32(4), 252-264. Santos, F.J., Roomi, M.A., & Liñán, F. (2016). About Gender Differences and the Social Environment in the Development of Entrepreneurial Intentions. Journal of Small Business Management, 54(1), 49-66. Saraf, N. (2015). What Determines Entrepreneurial Intention in India? Journal of Entrepreneurship and Innovation in Emerging Economies, 1(1), 39-55. Shinnar, R.S., Giacomin, O., & Janssen, F. (2012). Entrepreneurial Perceptions and Intentions: The Role of Gender and Culture. Entrepreneurship Theory and Practice, 36(3), 465-493. Shneor, R., Metin Camgöz, S., & Bayhan Karapinar, P. (2013). The interaction between culture and sex in the formation of entrepreneurial intentions. Entrepreneurship & Regional Development An International Journal, 25(9-10), 781-803. Siu, W., & Lo, E.S. (2013). Cultural Contingency in the Cognitive Model of Entrepreneurial Intention. Entrepreneurship Theory and Practice, 37(2), 147-173. 143

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

Tsai, K.H., Chang, H.C., & Peng, C.Y. (2016). Refining the linkage between perceived capability and entrepreneurial intention: roles of perceived opportunity, fear of failure, and gender. International Entrepreneurship and Management Journal, 12(4), 1127-1145. Virick, M., Basu, A., & Rogers, A. (2015). Antecedents of Entrepreneurial Intention among Laid-Off Individuals: A Cognitive Appraisal Approach. Journal of Small Business Management, 53(2), 450-468. Wei, L.Q., & Ling, Y. (2015). CEO characteristics and corporate entrepreneurship in transition economies: Evidence from China. Journal of Business Research, 68(6), 1157-1165.

144

EJAE 2021  18 (2)  127 - 145

MALI. P., KUZMANOVIĆ. B., NIKOLIĆ. M., TAREK STOJANOVIĆ. E.  THE IMPACT OF CONTROL VARIABLES ON ENTREPRENEURIAL INTENTIONS AMONG EMPLOYED PERSONS

UTICAJ KONTROLNIH VARIJABLI NA PREDUZETNIČKE NAMERE MEĐU ZAPOSLENIM LICIMA Rezime: U radu su prikazani rezultati proučavanja dimenzija individualne preduzetničke orijentacije, dimenzije postignuća i teorije dimenzija planiranog ponašanja, u zavisnosti od sedam varijabli: pola ispitanika, njihove starosti, obrazovnog nivoa ispitanika, vlasničke strukture preduzeća, prethodnog iskustva ispitanika u preduzetništvu, njihove pretpostavljene performanse posla i njihove percepcije finansija. Specifičnost istraživanja je da su ispitanici zaposlene osobe. Uzorak je obuhvatio 540 ispitanika iz 72 organizacije u Srbiji. Analiza podataka je izvršena putem t-testa. Statistički značajna razlika u uticaju posmatranih varijabli postoji u većini slučajeva, osim varijable - nivo obrazovanja ispitanika. Tako su četiri od sedam hipoteza u potpunosti potvrđene, dve delimično potvrđene, dok je jedna hipoteza odbačena. Profil zaposlene osobe koja potencijalno ima najveće šanse da postane preduzetnik je sledeći: mlađi muškarac sa srednjom stručnom spremom (diploma nema tako značajan uticaj), koji je zaposlen u privatnoj kompaniji, ima prethodno preduzetničko iskustvo, uspešan je na poslu i ima odgovarajuće finansije.

Ključne reči: individualna preduzetnička orijentacija, teorija planiranog ponašanja, preduzetničke namere, zaposlene osobe, Srbija.

145

EJAE 2021, 18(2): 146 - 160 ISSN 2406-2588 UDK: 640.4:659 005.952:[005.966:640.4 DOI: 10.5937/EJAE18-32929 Original paper/Originalni naučni rad

PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY Jasmina Ognjanović* University of Kragujevac, Faculty of Hotel Management and Tourism in Vrnjačka Banja Kragujevac, Serbia

Abstract: Successful management of preferred attributes of employer brand attractiveness provides appropriate benefits for hotels. The research aim is to examine the preferred attributes of employer brand attractiveness among potential employees in the hotel industry. The research was conducted at the Faculty of Hospitality and Tourism and includes 148 fourth-year undergraduate students and master students studying in Hospitality Management. Descriptive statistics were used in the paper. By analyzing the results, author concluded that the preferred attributes of employer brand attractiveness are career advancement and employee training and development. Employees’ desire to work for an employer that provides employee training and development opportunities brings numerous benefits to hotels, since employee development contributes to the creation of high-quality hotel services, more efficient employees, and, thus, better organizational performance. The least preferred attributes of employer brand attractiveness among potential hotel employees are “corporate reputation” and “corporate culture”.

Article info: Received: Jun 29, 2021 Correction: Avgust 20, 2021 Accepted: September 21, 2021

Keywords: employer brand attractiveness, human resource, attributes of employer brand attractiveness, hotel

INTRODUCTION The hospitality industry has a skilled and committed workforce that is seen as vital to the firm’s success (Scott, 2009). The hotels can achieve the desired results, among other things, by investing in a highly qualified workforce with the appropriate knowledge, experience, and skills. The hotel industry is justifiably referred to as the “people industry,” as it depends on the ability and enthusiasm of front-line staff (Baum, & Nickson, 1998). The role of human resources becomes crucial in every business segment, bearing in mind that no activity in the value chain can be fully automated and function without human resources (Perić, Dramićanin, & Pavlović, 2021). On the other hand, the hotel industry does not have the image of a desirable employment industry. Employees have very low salaries while working hours and conditions are unacceptable compared to other industries (Baum, & Nickson, 1998). 146

*E-mail: jasmina.lukic@kg.ac.rs

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Employing low-skilled workers justifies low salaries as well as high employee turnover rates (Baum, & Nickson, 1998). Given the disproportion present between the need for qualified employees, on the one hand, and inadequate employment conditions, on the other hand, hotel management must develop tools to enhance the benefits that employees derive from performing their business tasks. Modern business conditions, as well as the consequences of the pandemic crisis, have increasingly highlighted the role and importance of employees and conditioned the need to build sustainable and stable relations between employers and employees, as the most important internal stakeholders (Ognjanović, 2019). Employer brand attractiveness appears as an important tool in the process of creating sustainable and stable relationships between companies and employees. As some studies point out, in the struggle to build an image of an attractive employer, companies must possess a variety of attractive attributes to attract talented and hard-working applicants (Tanwar, & Kumar, 2019). The research aim is to examine the preferred attributes of employer brand attractiveness among potential employees in the hotel industry. The basic concept of employer branding relies on the fact that the company’s attractiveness depends on the perception of attributes by potential applicants (Jain, & Bhatt, 2015). Scientists and organizations today are becoming increasingly interested in identifying the attributes of employer attractiveness that distinguish companies in the labor market (Younis, & Hammad, 2021). For this reason, hotels need to know which attributes of employer brand attractiveness potential employees prefer to develop and actively participate in the “war for talent”. Companies use various marketing channels to promote the development of employer brand attributes in the labor market. Promotion helps potential employees to perceive the employer, so they observe company as a desirable or undesirable workplace. Research shows that potential employees associate employer brand attractiveness with their own needs, personality character (Backhaus, & Tikoo, 2004; Sivertzen, Nielsen, & Olafsen, 2013), and primary values (Tkalac Verčić, 2021). When the needs, personality and values of potential employees match the employer brand, the company becomes attractive to that person (Backhaus, & Tikoo, 2004; Sivertzen et al., 2013). The hotel industry, as a labor-intensive activity, must especially research and include in its HRM activities the preferred attributes of potential and current employees. Developing preferred attributes and matching them with the personality of employees, bring greater work efficiency and results. When it comes to the analysis of preferred attributes of employer brand attractiveness, several research gaps have been observed in the literature. First, scholars and organizations at the present time are interested in identifying the factors/attributes that differentiate companies and make them an attractive employer besides identifying how this process occurs (Younis, & Hammad, 2021). Employees within the company and potential employees may perceive these factors differently. For these reasons, there is a need to clearly separate potential and current employees and clearly differentiate the attributes of the employer brand attractiveness that are relevant to both groups of participants (Alshathry, Clarke, & Goodman, 2017). Reis and Braga (2016) and Biswas and Suar (2014) consider that research about employer attractiveness attributes, used in employer brand strategies, is still scarce. For potential employees to perceive the company in the right way, it is necessary, first of all, to know in which direction to develop the company from the aspect of the development of benefits for potential employees. Second, researchers emphasize the need to focus on personality traits that influence preferences relating to employment characteristics (Barrick, Mount, & Li, 2013; Horng, Tsai, Yang, & Liu, 2016; Bellou, Stylos, & Rahimi, 2018). A small number of studies have attempted to investigate how personal characteristics influence applicants’ individual decisions to apply for a job (Bellou et al., 2018). 147

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

By researching the preferred attributes of the employer among potential employees, the company has the opportunity to, through the ranking of preferred attributes, get to know what potential employees want and what their focus in career development will be. Potential employees can assess whether such benefits match their value system, culture, and personality. Third, most studies in the field of employer brand attractiveness have been conducted in the field of social sciences, business, and law (Oliveira, Proença, & Ferreira, 2021). The importance of this concept in the hotel industry has yet to be studied (Bellou et al., 2018). The study focuses on researching attributes of employer brand in the hotel industry, which is in line with Santiago’s (2019) recommendations that future research cover relevant sectors. This is further justified by the fact that the hotel industry is labor-intensive and that its business is largely based on the work of human resources. On the other hand, the hotel industry does not offer attractive working conditions for employees, which only deepens the gap between the need for human resources, on the one hand, and the benefits offered by hotels, on the other hand. In addition, Baum and Nickson (1998) see a special problem in the failure to recognize the connection between what is studied in the hotel industry in the hospitality context and the wider theoretical context. Due to the difficulties that employers face in attracting and retaining talent with certain profiles, there is a growing interest of both researchers and practitioners in researching the employer brand attractiveness (Benraïss-Noailles, & Viot, 2021). This topic becomes attractive having in mind several observed problems / issues in the research process. First, Brusch, Brusch and Kozlowski (2018) state that in larger cities many well-trained professionals are available to the companies, which is not the case in smaller cities and rural areas. In the process of researching the preferred attributes of employer brand attractiveness, one should keep in mind the quality of the supply and characteristic of potential employees at the labor market in developing countries, such as Serbia. As many hotels operate in the international market, an important question is whether, in today’s globalized business world, management can use standardized strategies to attract talent or they must adapt employer brand to cultural differences between countries (Alnıaçık, Alnıaçık, Erat, & Akçin, 2014). Cultural environment is an important element of the general environment that affects the perception of a good / bad employer among potential employees. Based on the above, the results of the study should provide an answer to 2 research questions: 1. What are the preferred attributes of employer brand attractiveness among potential employees in the hotel industry? 2. Is there a difference in preferred attributes among potential employees with work experience and no work experience? The paper, in addition to the introduction and conclusion, contains three parts. The first part includes a review of the literature describing the concept of an employer brand as well as the benefits that the construction of this intangible asset brings. The attributes of employer brand attractiveness are described in more detail, as well as the importance of developing these attributes in the process of attracting and recruiting applicants. The second part of the paper describes the used research instrument with an overview of the observed sample according to the appropriate characteristics. The third part of the paper describes the obtained results with discussion and connection with the results of previous research.

148

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

LITERATURE REVIEW Employer brand concept Employer branding assumes that human resources contribute to value creation and through skillful investment in human capital, the company’s performance can be improved (Backhaus, & Tikoo, 2004). Human resources are viewed as a critical asset of the company and a key factor of sustainable efficiency (Ognjanović, 2020). For this reason, for labor-intensive hotel companies it is necessary to improve management of human resources. Some authors (Kucherov, & Zavyalova, 2012; Santiago, 2019) view employer brand as a new field of study within HRM and a progressive approach based on general theory, which uses specific tools and principles of branding to make the people management process effective. The theoretical foundation of the employer brand concept can be found in the resource-based view, according to which human resource characteristics can contribute to competitive advantage and value creation (Backhaus, & Tikoo, 2004; Sivertzen et al., 2013). Sivertzen et al. (2013) believe that employer branding is also based on human resource theory, with a focus on developing the company’s image as an attractive employer (Backhaus, & Tikoo, 2004). To further foster reputation and increase attractiveness, employers seek to strengthen the company name as a brand, which is referred to as employer branding (Sivertzen et al., 2013). Backhaus and Tikoo (2004, p. 502) view employer branding “as the process of building an identifiable and unique employer identity, and the employer brand as a concept of the firm that differentiates it from its competitors”. The employer brand is a part of intangible assets that is the result of the successful implementation of the employer branding strategy that promotes the company as an exceptional employer by providing applicants with realistic expectations and fulfillment of promises given to current employees (Ognjanović, 2020). Santiago (2019) views the employer brand as benefits that the company offers to employees as to build a unique identity in the eyes of employees and applicants, thus encouraging them to join the company. Employer attractiveness is defined as the degree to which a company is attractive as an employer (Jain, & Bhatt, 2015). Employer brand techniques take two ways: attracting potential employees and retaining current employees (Sivertzen et al., 2013; Chhabra, & Sharma, 2014; Alshathry et al., 2017; Ognjanović, 2019). When talking about the importance of the employer brand for attracting potential employees, some authors (Sivertzen et al., 2013; Rampl, & Kenning, 2014; Xie, Bagozzi, & Meland, 2015; Ognjanović, 2019; Santiago, 2019) consider the term “employer brand attractiveness” appropriate. Other authors (Berthon, Ewing, & Hah, 2005) also discuss the appropriateness of the term “employer attractiveness. The essence of the employer branding concept refers to a proactive approach to managing the company’s image as an employer (Rampl, & Kenning, 2014) by current employees. Thus, current employees build the image of an attractive employer on the labor market, which, through the development of attractive job offers, aims to attract potential talent. Chhabra and Sharma (2014) prove a significant and positive correlation between a strong brand image and the likelihood of applicants applying. Berthon et al. (2005, p. 156) define employer attractiveness as “the envisioned benefits that a potential employee sees in working for a specific organization”. In addition to employer brand attractiveness, similar terms that denote the company’s activities to attract talented employees are “employer of choice”, “job pursuit intentions”, “acceptance intentions” (Tanwar, & Kumar, 2019). That is why it is necessary to distinguish between them. Employer of choice is a company that employees strive to work for and that they prefer over other companies (Tanwar, & Kumar, 2019). 149

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Job pursuit intention is reflected in the potential applicant’s desire to work for a particular company (Tanwar, & Kumar, 2019). Acceptance intentions measure the probability of an applicant accepting a job offer or the likelihood of an applicant accepting a job offer from the company (Tanwar, & Kumar, 2019).

Benefits of building employer brand attractiveness The ultimate goal of employer brand attractiveness is to attract potential employees to join the company while the goal of existing employees is to ensure that they continue to experience the value associated with belonging to the organization (Alshathry et al., 2017). By effectively branding the company, the employer can give the company an advantage in the “war for talent”, attracting and retaining a talented workforce, which is crucial for the quality of services delivered to the guest (Santiago, 2019). Attracting and retaining highly talented employees contributes to competitive advantage and is an important issue for all companies (Alnıaçık et al., 2014). Backhaus and Tikoo (2004) believe that a more attractive employer brand internalizes company value and helps retain employees. Some of the advantages of companies with an attractive employer brand are reduced employee recruitment costs, improved employee relations, higher employee retention rates, and lower wages for the same categories of employees compared to wages paid by companies with a less attractive employer brand (Chhabra and Sharma, 2014). Kucherov and Zavyalova (2012) conclude, based on research results, that companies with employer brands gain numerous economic benefits, such as lower rates of staff turnover and higher rates of HR investment in training and development activities of employees. Sivertzen et al. (2013) consider that employer branding attractiveness is used to enhance corporate reputation.

Attributes of employer brand attractiveness and potential employees The attributes of employer brand attractiveness represent the expected benefits that employees and potential employees can gain by working for a particular company. Ronda and Valor (2018, p.574) observed attributes of employer brand as “employer-extrinsic traits set by companies that constitute organization’s offering to employees”. By developing the attributes of an attractive employer, the company gains appropriate benefits by building an appropriate workforce structure (Ognjanović, 2019). The identification of preferred attributes is especially important when it comes to front-line employees who represent their company in terms of beliefs, values, mission in the minds of guests (Bellou et al., 2018). Many studies link applicants’ characteristics to job descriptions (Bellou et al., 2018), while other studies focus on the job and environment perceptions among potential employees (such as Berthon et al., 2005; Backhaus, & Tikoo, 2004; Kaur, & Shah, 2021). The theoretical basis of the analysis of the preferred attributes of employer brand attractiveness is based on signaling theory which explains the attractiveness of the employer among applicants (Younis, & Hammad, 2021). According to this theory, applicants are affected by any information that refers to the attributes of the organization and is called “signals” (Younis, & Hammad, 2021). Applicants are motivated to seek information about the company as to create an accurate perception of the employer (Younis, & Hammad, 2021). The more potential employees find an employer attractive, the more they will show interest in working for a particular company (Jain, & Bhatt, 2015). The literature emphasizes that potential employees can observe the direct characteristics of the employer, such as location, salary (Jain, & Bhatt, 2015), size, number of employees, working hours. 150

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Other information, also known in the literature as symbolic attributes, will not be noticed by applicants (employee orientation, work culture, career development, and innovativeness) (Jain, & Bhatt, 2015; Younis, & Hammad, 2021). Research shows that potential employees spend less effort on gathering information about potential employers, while, on the other hand, most of these employers may share too little information about employment and working conditions (Younis, & Hammad, 2021). Attributes of employer attractiveness are observed through functional, psychological, economic (Ambler, & Barrow, 1996), and organizational attributes (Kucherov, & Zavyalova, 2012). Functional attributes refer to the specifics of the job, work content, training, and development opportunities for employees as well as the perspective of career development of the employee. Psychological attributes refer to the feeling of belonging and membership of the hotel employee. Economic attributes are understood as a system of financial and non-financial compensations that the employee acquires in the company by the invested work. Organizational attributes refer to the company’s image among various stakeholders in the external market. The research has addressed desirable employer attributes that attract potential employees to apply. Chhabra and Sharma (2014) conclude that among students, the preferred employer attributes are organizational culture, brand name, and compensation. The results of the study (Tanwar, & Kumar, 2019) show that two dimensions, work culture, and work content, drive employer of choice. Jain and Bhatt (2015) believe that there are some factors, such as company stability, work-life balance, and job security, that potential employees consider important when applying, both in the case of the private and in the case of the public sector. Santiago (2019) concludes that economic factors, such as an aboveaverage salary or opportunities for promotion, are less important for applicants when considering future jobs. Reis, Braga, and Trullen (2017) conclude that companies become more competitive in attracting talent if their employment strategies emphasize psychological benefit. Kaur and Shah (2021) analyze the instrumental-symbolic attributes of an employer when applying for a job. They conclude that potential employees prefer, within instrumental attributes, job security and task diversity, and within symbolic attributes – competence. Lee, Kao and Lin (2018) also analyze instrumental-symbolic attributes and conclude that companies use functional attributes to build an emotional relationship with employees and satisfy psychological sustenance, which attracts young Taiwanese applicants. Younis and Hammad’s (2021) results show that the employer image and corporate image have a positive and significant impact on employer attractiveness. Xie et al. (2015) conclude that corporate reputation influences applicants’ intention to apply for a job. Sivertzen et al. (2013) prove that the development of employer value (innovation value, psychological value, application value, and the use of social media) can improve corporate reputation, which in turn contributes to positive intentions for potential employees to apply for a job.

151

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

RESEARCH INSTRUMENTS AND SAMPLE DESCRIPTION The sample includes potential employees in the hotel industry who are fourth-year undergraduate students and master students in the Hotel Management department. Information on the preferred attributes of employer brand attractiveness among potential employees was obtained through a questionnaire. Respondents were tasked with opting for only one, the most desirable attribute of employer brand attractiveness among the attributes offered. The selection of the offered attributes of employer brand was made based on the following papers: Functional attributes: 1. The employer provides opportunities for career advancement (Bellou et al., 2018; Lee et al., 2018; Deepa, & Baral, 2019); 2. The employer provides training and development opportunities for employees (Schlager, Bodderas, Maas, & Cachelin, 2011; Tanwar, & Prasad, 2016; Kashyap, & Verma, 2018; Sharma, & Prasad, 2018; Lee et al., 2018; Santiago, 2019; Ognjanović, 2020; Kaur, & Shah, 2021); 3. The employer offers a meaningful and interesting job (Bellou et al., 2018; Lee et al., 2018). Psychological attributes: 4. The employer has a developed corporate (organizational) culture that provides fair and collegial cooperation with colleagues (Chhabra, & Sharma, 2014; Tanwar, & Prasad, 2016; Sharma, & Prasad, 2018; Tanwar, & Kumar, 2019; Deepa, & Baral, 2019; Ognjanović, 2020). Economic attributes: 5. The employer pays adequate salaries according to the work invested (Schlager et al., 2011; Sivertzen et al., 2013; Chhabra, & Sharma, 2014; Bellou et al., 2018; Santiago, 2019; Deepa, & Baral, 2019; Kaur, & Shah, 2021). Organizational attributes: 6. The employer is known on the market, i.e. has a built-in corporate reputation (Schlager et al., 2011; Tanwar, & Prasad, 2016; Alshathry et al., 2017; Sharma, & Prasad, 2018; Deepa, & Baral, 2019). The research was conducted in May 2020. The questionnaire was sent to 251 e-mail addresses of students in the fourth year of basic academic studies and master students. A total of 148 questionnaires were returned, so the response rate was 59%. The sample is observed based on the following characteristics: gender, age, work experience (Table 1). The sample is dominated by female respondents (75%), respondents aged between 21 and 30 (88.5%), and respondents with no work experience in the hotel industry (53%).

152

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Table 1. Description of the sample Criterion

Number of respondents

Male

25%

Female

111

75%

148

100%

21 - 30

131

88.5%

31 and above

11.5%

148

100%

I have no experience

53%

Up to 5 years

45%

From 6 to 10 years

1.5%

11 year and above

0.5%

148

100%

Gender

∑ Age up to 20 years

∑ Work experience

∑ Source: Author

RESEARCH RESULTS AND DISCUSSION Data processing is performed using the statistical program for social science, SPSS. Based on the listed frequencies, the preferred attributes of employer brand attractiveness, ranked according to importance for potential employees, are: 1. “The employer provides opportunities for career advancement” (38%) 2. “The employer offers training and development opportunities for employees” (23%) 3. “The employer pays adequate salaries in accordance with the invested work” (19%) 4. “The employer offers a meaningful and interesting job” (10%) 5. “The employer has a corporate reputation” (6%) 6. “The employer has a developed corporate (organizational) culture that ensures correct and collegial cooperation with colleagues” (4%)

153

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Figure 1. Preferred attributes of employer brand attractiveness The employer pays adequate salaries in accordance with the invested work 0%

23%

The employer provides opportunities for career advancement and development

19%

The employer has a corporate reputation 4% 10% 6%

38%

The employer offers a meaningful and interesting job

The employer has a developed corporate (organizational) cultur that ensures correct and collegial cooperation with colleagues The employer offers training and development opportunities for employees Slice 7

By identifying preferred attributes among potential employees, attributes significant in the process of recruitment and attraction employees are determined. The results show that potential employees in the hotel industry prefer those employers who support career advancement Slice 8 as well as training and employee development. These are the functional attributes of the employer that describe the desirable elements of employment in the company (Backhaus, & Tikoo, 2004). Promoting these attributes of an employer brand attractiveness can help potential employees to understand how company operates Slice 9 (Tkalac Verčić, 2021). Authors (Backhaus, & Tikoo, 2004; Bellou et al., 2018; Lee et al., 2018; Santiago, 2019) also concluded that career advancement, training and development are crucial for potential employees. Kaur and Shah (2021) point out the connection between these two attributes by the fact that improvement of knowledge and skills, due to training and development, provides greater opportunities for career advancement, which leads to the greater organizational commitment of employees. Ognjanović (2019) connects working conditions that are a consequence, among other things, of available training and development programs as well as opportunities for career advancement, with employee satisfaction. Backhaus and Tikoo (2004) state that thanks to the “career advancement” attribute, the number of qualified candidates applying for professional jobs has increased by 30%. Baum and Nickson (1998) believe that training and development are the main motivators of employee work that can lead to a reduction in attrition rates. Learning and development provide benefits such as stimulating engagement and desired behavior in the workplace as well as strengthening the identity of employees in the organization (Kaur, & Shah, 2021). Kucherov and Zavyalova (2012) conclude that investing in training and development also contributes to the company through the provision of appropriate economic benefits. 154

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Jain and Bhatt (2015) believe that training and learning can help potential employees in honing their employability. It should also be noted that the attributes of career advancement and training and development are symbolic attributes of employer brand attractiveness, whose information cannot be obtained completely accurately by potential employees (Younis, & Hammad, 2021). What is interesting is that salary, a financial benefit that the employees get in exchange for their invested work and effort, ranks third in importance. The fact that the research was conducted in a developing country makes this fact even more surprising. These results are in line with the conclusion of Santiago (2019) that economic attributes are not crucial for applicants when considering future jobs. This is a positive signal for all managers because it is more important to potential employees how the management treats the knowledge and development of employees than the amount of salary they are paid. If the employees are ready to improve, that will bring additional benefits to the hotels because they will work more efficiently, which will also affect the satisfaction of the guests. In the last place of the preferred attributes is the development of “corporate reputation” and “corporate culture” in the hotel. Potential employees do not link corporate reputation to the reputation of the employer, which can only further motivate the management of companies to invest and develop the attractiveness of the employer brand. The results also show that respondents still do not have a developed awareness of the importance of corporate culture in the business of hotel companies. Finally, it should be noted that the development of the attractiveness of the employer is not the task of one department but of the entire company where the employer plays a central and important role (Brusch et al., 2018). The second research question relates to checking the difference in preferred attributes among potential employees with work experience and no work experience. The sample was observed according to three criteria: gender of respondent, age, and level of education. The results are shown in Table 2 and Table 3. Table 2. Preferred attributes of employer brand attractiveness among potential employees without work experience according to the criteria (gender, age, level of education) Criterion

Preferred attributes of employer brand

Percentage response rate

Gender Male

„The employer offers training and development opportunities for employees“

Female

„The employer provides opportunities for career advancement“

up to 20 years

21 - 30

„The employer offers training and development opportunities for employee“

31 and above

„The employer offers training and development opportunities for employee“ and

„The employer offers a meaningful and interesting job“

Bachelor study

„The employer offers training and development opportunities for employee“

Master study

„The employer provides opportunities for career advancement“

Age

Level of Education

Source: Author 155

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

The results in Table 2 show that potential employees without work experience prefer attributes training and development and career advancement, which is in line with the results obtained for the whole sample. An additional preferred attribute by older potential employees without work experience is a meaningful and interesting job. Employers, in addition to training and developing employees and their career advancement, must insist on introducing creativity into regular work tasks. Table 3. Preferred attributes of employer brand attractiveness among potential employees with work experience according to the criteria (gender, age, level of education) Criterion

Preferred attributes of employer brand

Percentage response rate

Gender Male

„The employer provides opportunities for career advancement“

Female

„The employer provides opportunities for career advancement“

up to 20 years

21 - 30

„The employer provides opportunities for career advancement“

31 and above

„The employer offers training and development opportunities for employee“

Bachelor study

„The employer provides opportunities for career advancement“

Master study

„The employer provides opportunities for career advancement“

Age

Level of Education

Source: Author

The results in Table 3 show that employees with work experience prefer the career advancement attribute. These respondents prioritize career development over the financial benefits they can gain by working for an employer. This means that with the increase of work experience, employees become more and more interested in the non-financial benefits that the employer provides them.

CONCLUSION Based on the obtained research results, answers to research questions were given. Potential employees in the hotel industry prefer the following attributes of employer brand attractiveness: “career advancement” and “employee training and development”. Employee motivation to work for an employer that provides opportunities for training and advancement of employees brings numerous benefits to hotels since employee development contributes to the development of quality hotel services, more efficient work of employees, and, thus, better organizational performance. These results show that potential employees are willing to invest in their knowledge and improve it, which increases the value of human capital for hotels. The least preferred attributes are “corporate reputation” and “corporate culture”. 156

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

The results show that potential employees without work experience prefer the attributes of training and development and career advancement. In this situation, potential employees, by improving their knowledge, contribute to the growth of the value of human capital. The preferred attribute of employer brand attractiveness among potential employees with work experience is career advancement. Potential employees expect a non-financial benefit in terms of promotion and a better position in the hotel because they will become satisfied with their job. Information about preferred attributes of employer brand attractiveness can be useful for hotel management because they know the character of potential employees, recognize the values and principles that potential employees value. Recognizing the character of potential employees and the employer attributes which they prefer, facilitates the process of selecting employees for hotel management. By choosing the preferred attributes of the employer, business culture is formed among the employees with a clear attitude what the priorities of work are and whether they choose to stay with the employer. Managerial implication. The results of the research can help hotel management by pointing them to the attributes of employer brand attractiveness that they need to specifically analyze and develop. The hotel industry is not known as an attractive option when it comes to employment, and therefore does the management would have to deal with the employees and develop appropriate benefits, to attract and keep the best candidates in the hotel. The situation in the labor market of the hotel industry in the Republic of Serbia is quite favorable if we consider that the preferred attributes of employer attractiveness are “career advancement” and “employee training and development”. Potential employees are aware that by developing and improving their knowledge, skills, and abilities, they can achieve personal satisfaction, through career advancement, as well as hotel management satisfaction, by improving the quality of hotel services and work efficiency. This should be supported by the hotel management in terms of developing and providing preferred attributes to potential employees. Limitation. Research has several limitations. First, the respondents are students of the Hotel Management department who are expected to work in the hotel industry in the future. The research does not include potential employees who are active job seekers in the labor market in the hotel industry. Second, the research does not link the preferred attributes of employer brand attractiveness to any of the hotel business factors. The research concluded in a broader context would consider the implications of employer brand attractiveness for the hotel business while analyzing possible connections with some of the key business factors. For these reasons, no additional analyzes and research methods have been performed that would give a more comprehensive overview of the factors that affect the employer brand attractiveness. Third, the research does not provide an answer to the question of whether the surveyed potential employees would apply for a job in a company with developed preferred attributes. We cannot claim with certainty that these are the attributes of the employer that would stimulate potential applicants to apply for open job position. Fourth, certain attributes of employer brand attractiveness are observed through only one question (psychological, economic, and organizational attributes). More detailed results could be obtained in the case of observing attributes through multiple questions as well as by including some additional attributes (job security, work-life balance, etc.). Future research could be based on observing many attributes of employer brand attractiveness. Such attributes could be related primarily to the applicants’ desire to apply for a job as well as to some other factors of the hotel business. Future research could also be based on a comparative analysis of the preferred attributes of employer brand attractiveness between the hotel industry and some other service industries. 157

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

REFERENCES Alnıaçık, E., Alnıaçık, U., Erat, S. & Akçin, K. (2014). Attracting Talented Employees to the Company: Do We Need Different Employer Branding Strategies in Different Cultures? Procedia - Social and Behavioral Sciences, 150, 336 – 344. Alshathry, S., Clarke, M. & Goodman, S. (2017). The role of employer brand equity in employee attraction and retention: a unified framework. International Journal of Organizational Analysis, 25(3), 413-431. DOI: 10.1108/IJOA-05-2016-1025 Ambler, T. & Barrow, S. (1996). The employer brand. The Journal of Brand Management, 4(3), 185-205. Backhaus, K. & Tikoo, S. (2004). Conceptualizing and researching employer branding. Career Development International, 9(5), 501-517. DOI: 10.1108/13620430410550754 Barrick, M.R., Mount, M.K. & Li, N. (2013). The theory of purposeful work behavior: the role of personality, higher-order goals, and job characteristics. Academy of Management Review, 38(1), 132-153. Baum, T. & Nickson, D. (1998).Teaching human resource management in hospitality and tourism: a critique. International Journal of Contempory Hospitality Management, 10(2), 75–79. Bellou, V., Stylos, N. & Rahimi, R. (2018). Predicting hotel attractiveness via personality traits of applicants: The moderating role of self-esteem and work experience. International Journal of Contemporary Hospitality Management, 30(10), 3135-3155. DOI: 10.1108/IJCHM-06-2017-0369 Benraïss-Noailles, L. & Viot, C. (2021). Employer brand equity effects on employees well-being and loyalty. Journal of Business Research, 126, 605–613. Berthon, P., Ewing, M. & Hah, L.L. (2005). Captivating company: dimensions of attractiveness in employer branding. International Journal of Advertising, 24(2), 151-172. Biswas, M. & Suar, D. (2014). Antecedents and consequences of employer branding. Journal of Business Ethics, 136(1), 57-72. DOI:10.1007/s10551-014-2502-3 Brusch, I., Brusch, M. & Kozlowski, T. (2018). Factors influencing employer branding Investigations of student perceptions outside metropolitan regions. International Journal of Quality and Service Sciences, 10(2), 149-162. DOI:10.1108/IJQSS-09-2017-0085 Chhabra, N.L. & Sharma, S. (2014). Employer branding: strategy for improving employer attractiveness. International Journal of Organizational Analysis, 22(1), 48-60. DOI: 10.1108/IJOA-09-2011-0513 Deepa, R. & Baral, R. (2019). Importance-performance analysis as a tool to guide employer branding strategies in the IT-BPM industry. Journal of Organizational Effectiveness: People and Performance, 6(1), 77-95. Horng, J.S., Tsai, C.Y., Yang, T.C. & Liu, C.H. (2016). Exploring the relationship between proactive personality, work environment and employee creativity among tourism and hospitality employees. International Journal of Hospitality Management, 54, 25-34. Jain, N. & Bhatt, P. (2015). Employment preferences of job applicants: unfolding employer branding determinants. Journal of Management Development, 34(6), 634-652. DOI: 10.1108/JMD-09-2013-0106 Kashyap, V. & Verma, N. (2018). Linking dimensions of employer branding and turnover intentions. International Journal of Organizational Analysis, 26(2), 282-295. Kaur, R. & Shah, R. (2021). Employer branding at armed forces: current and potential employees’ perspective. European Journal of Training and Development, Article in press. DOI: 10.1108/EJTD-03-2019-0043 Kucherov, D. & Zavyalova, E. (2012). HRD practices and talent management in the companies with the employer brand. European Journal of Training and Development, 36(1), 86-104. Lee, C., Kao, R. & Lin, C. (2018). A study on the factors to measure employer brand: the case of undergraduate senior students. Chinese Management Studies, 12(4), 812-832. Ognjanović, J. (2019). Building employer attractiveness as part of the employer branding strategy. Ekonomija, teorija i praksa, XII(3), 53–68. 158

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

Ognjanović, J. (2020). Employer brand and workforce performance in hotel companies. Hotel and Tourism Management, 8(2), 65-78. Oliveira, M., Proença, T. & Ferreira, F. (2021). Do corporate volunteering programs and perceptions of corporate morality impact perceived employer attractiveness? Social Responsibility Journal. DOI: 10.1108/SRJ-032021-0109. Perić, G., Dramićanin, S., &, Pavlović, N. (2021). The influence of internal service quality and employee satisfaction on organizational commitment in travel agencies: The case of Serbia. Hotel and Tourism Management, 9(1), 43-60. Rampl, L.V. & Kenning, P. (2014). Employer brand trust and affect: linking brand personality to employer brand attractiveness. European Journal of Marketing, 48(1/2), 218-236. DOI: 10.1108/EJM-02-2012-0113 Rampl, L.V. (2014). How to become an employer of choice: transforming employer brand associations into employer first-choice brands. Journal of Marketing Management, 30(13-14), 1486-1504. Reis, G.G., Braga, B.M. & Trullen, J. (2017). Workplace authenticity as an attribute of employer attractiveness. Personnel Review, 46(8), 1962-1976. DOI:10.1108/PR-07-2016-0156 Reis,G.G. & Braga, B.M. (2016). Employer attractiveness from a generational perspective: Implications for employer branding. R.Adm., São Paulo, 51(1), 103-116. DOI: 10.5700/rausp1226 Ronda, L. & Valor, C. (2018). Are they willing to work for you? An employee-centric view to employer brand attractiveness. Journal of Product & Brand Management, 27(5), 573–596. DOI: 10.1108/JPBM-07-2017-1522 Santiago, J. (2019). The relationship between brand attractiveness and the intent to apply for a job: A millennials’ perspective. European Journal of Management and Business Economics, 28(2), 142-157. DOI: 10.1108/ EJMBE-12-2018-0136 Schlager, T., Bodderas, M., Maas, P. & Cachelin, J.L. (2011). The influence of the employer brand on employee attitudes relevant for service branding: an empirical investigation. Journal of Services Marketing, 25(7), 497 – 508. Scott, R. (2009). Undergraduates' Perceptions of Tourism and Hospitality as a Career Choice. International Journal of Hospitality Management, 28(3), 382-388. DOI: 10.1016/j.ijhm.2008.10.006 Sharma, R. & Prasad, A. (2018). Employer brand and its unexplored impact on intent to join. International Journal of Organizational Analysis, 26(3), 536-566. Sivertzen, A., Nilsen, E.R. & Olafsen, A. (2013). Employer branding: employer attractiveness and the use of social media. Journal of Product & Brand Management, 22(7), 473–483. DOI: 10.1108/JPBM-09-2013-0393 Tanwar, K. & Kumar, A. (2019). Employer brand, person-organisation fit and employer of choice Investigating the moderating effect of social media. Personnel Review, 48(3), 799-823. DOI: 10.1108/PR-10-2017-0299 Tanwar, K. & Prasad, A. (2016). The effect of employer brand dimensions on job satisfaction: gender as a moderator. Management Decision, 54(4), 854 – 886. Tkalac Verčić, A. (2021). The impact of employee engagement, organisational support and employer branding on internal communication satisfaction. Public Relations Review, 47, 102009. DOI:10.1016/j.pubrev.2021.102009. Xie, C., Bagozzi, R. & Meland, K. (2015).The impact of reputation and identity congruence on employer brand attractiveness. Marketing Intelligence & Planning, 33(2), 124-146. Younis R.А. & Hammad, R. (2021). Employer image, corporate image and organizational attractiveness: the moderating role of social identity consciousness. Personnel Review, 50(1), 244-263. DOI: 10.1108/PR02-2019-0058 Younis, R.A.A. & Hammad, R. (2021). Employer image, corporate image and organizational attractiveness: the moderating role of social identity consciousness. Personnel Review, 50(1), 244-263. DOI: 10.1108/PR-022019-0058

159

EJAE 2021  18 (2)  146 - 160

OGNJANOVIĆ. J.  PREFERRED ATTRIBUTES OF EMPLOYER BRAND ATTRACTIVENESS AMONG POTENTIAL EMPLOYEES IN THE HOTEL INDUSTRY

PREFERIRANI ATRIBUTI ATRAKTIVNOSTI BRENDA POSLODAVCA MEĐU POTENCIJALNO ZAPOSLENIMA U HOTELSKOJ INDUSTRIJI Rezime: Uspešno upravljanje preferiranim atributima atraktivnosti brenda poslodavca pruža odgovarajuće prednosti hotelima. Cilj istraživanja u radu jeste da ispita preferirane atribute atraktivnosti brenda poslodavca među potencijalno zaposlenima u hotelskoj industriji. Istraživanje je sprovedeno na Fakultetu za hotelijerstvo i turizam i obuhvata 148 studenata završne godine osnovnih akademskih studija i studenata master studija, smera Menadžment u hotelijerstvu. U radu je korišćena deskriptivna statistika. Analizom rezultata može se zaključiti da su preferirani atributi atraktivnosti brenda poslodavca razvoj karijere, razvoj karijere, obuka i razvoj zaposlenih. Želja zaposlenih da rade kod poslodavca koji pruža mogućnost usavršavanja i napredovanja zaposlenih donosi brojne koristi i hotelima, budući da razvoj zaposlenih doprinosi razvoju kvaliteta hotelskih usluga, efikasnijem radu zaposlenih, pa samim tim i boljim organizacionim performansama. Najmanje preferirani atributi atraktivnosti brenda poslodavca među potencijalno zaposlenima u hotelijerstvu su „korporativna reputacija“ i „korporativna kultura“.

160

Ključne reči: atraktivnost brenda poslodavca, potencijalno zaposleni, atributi atraktivnosti brenda poslodavca, hotel.

EJAE 2021, 18(2): 161 - 177 ISSN 2406-2588 UDK: 330.341:323(676.1) 339.194 DOI: 10.5937/EJAE18-31159 Original paper/Originalni naučni rad

POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS Stephen Esaku* Department of Business and Management at Cavendih University Uganda, Kampala, Uganda.

Abstract: In this paper, we analyzed the long-run relationship between political instability and the shadow economy in Uganda using the autoregressive distributed lag bounds testing approach to cointegration. We found a negative and statistically significant relationship between political instability and the shadow economy, in both the long-run and short-run. This implies that an improvement in political processes that creates stability in the incumbent regime significantly reduces the shadow economy, consistent with the view that political institutions play a crucial role in facilitating political processes, which in turn reinforce the allocation of economic resources and the provision of public goods and services that improve the welfare of the citizens. This makes it less attractive for the citizens to operate in the shadow economy as the formal economy can now provide much of the needed goods and services. The practical implication of these results is that any attempts by policy makers to reduce activities in the shadow economy should also involve reforming the political system and encouraging civic engagement between the political elites and the citizenry or voters. Additionally, policy makers should formulate policies that reinforce the functioning of political institutions independent of any interference from political elites with rent-seeking behavior.

Article info: Received: March 4, 2021 Correction: April 12, 2021 Accepted: April 20, 2021

Keywords: shadow economy, governance, democracy, informal sector, political economy.

INTRODUCTION Over the last two decades or so, there has been increasing evidence that activities in the shadow economy1 are expanding in many economies of the world (Medina & Schneider, 2018). Across many countries, the growth of the shadow economy has persisted and many policy makers are acknowledging its impact on the size and growth of the formal economy (Medina & Schneider, 2018; Berdiev, Pasquesi1 Also known as shadow economy, or informal economy/ sector, or underground economy. In this paper, we use these words to denote the same thing.

*E-mail: esaku_stephen@yahoo.com

161

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Hill, & Saunoris, 2015; Orsi, Raggi & Turino, 2014). During the early 1950s through 1970s, activities in the shadow economy were considered activities that happened on the fringes of the formal economy because of low development and backwardness. Thus, as economies start to record substantial economic gains resulting from growth and development, shadow economy activities start to decline and eventually shrink as the formal sector can provide the much needed goods and services ( Kanbur, 2017). However, there is mounting empirical evidence indicating that activities in the shadow economy are increasing, implying that this sector plays a key role in the production and distribution of goods and services and should not be considered a temporary or momentary sector (Fourie, 2018; Esaku, 2020). For example, Medina & Schneider (2018) provide estimates of the size of shadow economy around the world and suggest that shadow economy activities have been expanding in both the developing and developed economies. Similarly, Alm & Embaye (2013) provide estimates of the shadow economy for 111 economies that show that the informal economy is a substantial part of production. This provides evidence that substantial level of productive activities does take place outside the radar of government regulators and tax bodies (Elgin, 2020). Some of the reasons suggested by the literature for the increase in shadow economy activities include overregulation, higher taxes, underdevelopment, income inequality and financial development among others (Buehn & Schneider, 2012a; Esaku, 2021a; Goel & Nelson, 2016; Hajilee, Stringer & Metghalchi, 2017; Luong, Nguyen & Nguyen, 2020). The increase of activities in the shadow economy causes serious distortions and undermines governments’ ability to collect taxes which could facilitate the provision of public goods and services. Allowing shadow economy activities to crowd out production and distribution of goods and services is detrimental for economic planning because of biases introduced by these activities into the economy wide indicators like unemployment, inflation, income, consumption expenditure, among others, which in turn distorts the real economic situation of the country (Capasso & Jappeli, 2013). For instance, Blackburn, Bose & Capasso (2012) suggest that because of the economic distortions caused by the shadow economy, relying on statistics produced by economies with large shadow economies could be misleading since not all economic activities in such countries have been fully accounted for in the national accounts. Thus, relying on data with the above distortions affects governments’ planning processes (Capasso & Jappeli, 2013). Consequently, there is a growing interest among researchers to investigate the drivers of the shadow economy so as to ameliorate its adverse impacts. In line with the above mentioned, a number of studies have examined the size, causes and impact of activities in shadow economy (Bucek, 2017; Buehn & Schneider, 2012a; Mugoda, Esaku, Nakimu & Bbaale, 2020). The bulk of studies that examine shadow economy activities have mainly focused on investigating how public finance and public administration such as the burden of taxes, overregulation, government spending and others affect the size of the shadow economy (Goel & Nelson, 2016). Recently, a new line of research has emerged that analyses how the shadow economy interacts with economy wide variables. One strand of literature investigates the importance of the financial sector in shaping shadow economy. For example, Berdiev & Saunoris (2016) analyzing the dynamic link between the shadow economy and financial development. Their results indicate that a developed financial sector is important in reducing the size of shadow economy. Correspondingly, some studies have emphasized the role of institutions in influencing the growth of shadow economy (Elgin & Oztunali, 2014; Huynh, Nguyen & Nguyen, 2020). This literature suggests that institutions impose constraints that regulate activities in the shadow economy and determine how economic agents participate or organize social, political and economic engagement. The importance of institutions cannot be overemphasized. 162

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Another line of research analyzes the relationship between corruption and the shadow economy. For instance, Buehn & Schneider (2012b), in a sample of 51 economies around the world investigate the relationship between corruption and the size of shadow economy and find evidence that corruption increases the size of shadow economy. González-Fernández & González-Velasco (2014) provide evidence of a positive relationship between the size of shadow economy and corruption. Recently, some studies have started to investigate the role of governance in shaping the size of shadow economy. For instance, Teobaldelli & Schneider (2013) examine the effect of direct democracy on the shadow economy across 57 democracies around the world and find that democracy reduces the size of shadow economy. The rise in the level of democratization significantly reduces the size of shadow economy since democracy strengthens fiscal policy formulation and implementation leading to efficient collection of revenue and provision of public goods and services, such as security and infrastructure. The implication of this finding is that some components of governance, such as democracy, are important in hindering the increase in the shadow economy. Whilst previous studies have examined shadow economies around the world, much of their focus has been on the underlying proximate causes of shadow economy with less attention paid to its interrelationships with other phenomena like the business and political environment. Not much is known about how political processes in a given country drive shadow production of goods and services. This paper investigates the relationship between political instability and the shadow economy in Uganda. Specifically, how does political instability shape the size of the shadow economy in a less developed country like Uganda? Since its independence from British colonial rule in 1962, this East African country has experienced political instability which has slowed economic growth and development for nearly four decades. Starting with the military coup of 1971 that brought in the government of Idi Amin, the country introduced many radical changes that caused political instability. The Amin regime declared ‘economic war’ which resulted into the expulsion of (within ninety days), British-Asians and confiscation of their property and businesses. This was followed by the clamping down of any dissenting voices and ‘perceived enemies’ of the state, which further complicated the economic situation in the country. Consequently, businesses began to operate underground to avoid being labeled by the government as mafuta mingi, a kiswahili slang at the time, which meant ‘rich and fat’ businessmen/ women who were profiteering by not paying taxes. For nearly 10 years, the country was plunged into political chaos where different factions were fighting for the control of government. The situation was compounded by Uganda’s attack on neighboring Tanzania, in which Tanzania retaliated by invading Uganda in October 1978. Amin’s government was overthrown in June 1979 and a new government came into power after the general elections of 1980. However, the results of the general elections were disputed resulting into another war, this time, a protracted bush war that brought in the regime of Museveni into power in January 1986. Throughout the 1980s and 1990s, the new government faced a lot of resistance from elements of previous regimes, resulting into nearly two decades of conflict and political uncertainty. However, when multiparty system of governance was reinstated in the 2000s, political party activities brought in competition and political engagement amongst the population. Thus, the study of how political instability shapes the size of shadow economy is important especially in the geographic contexts, like Uganda, that have experienced a lot of political uncertainty for some extended periods of time. One can argue that government policies that influence the size of shadow economy are a result of political processes taking place in the country. Theoretical models have previously suggested that political environment in which businesses operate in, are a result of government policies that reinforce the growth of shadow economy (Elbahnasawy, Ellis, & Adom, 2016). 163

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Furthermore, the relationship between political instability and the size of shadow economy is possible given that the incumbent government may find difficulty in tax administration in situations of polarization and political instability in the country. A reduction in the levels of taxes or efficiency in tax administration will impact negatively on the provision of essential goods and services. This leads to the growth of shadow economy as the population struggles to meet the daily needs as a result of low provision of public goods and services. Given that tax evasion and failure to collect taxes by the government are some of the proximate causes of the shadow economy, we investigate how the political system shapes the shadow economy in Uganda. This paper makes three main contributions to the literature. First, we explore an important relationship between political instability and the size of shadow economy. To the best of our knowledge, except for the study by Elbahnasawy et al. (2016), this might be the first attempt to explore this relationship outside of a developed economy. Second, this paper provides evidence of the relationship between political instability and the shadow economy in Africa, a context that has experienced turbulent political periods for decades. We believe that Africa offers the best testing ground for our empirical ideas on the relationship between the shadow economy and political instability. Third, this paper uses a robust econometric technique (autoregressive distributed lag, ARDL- bounds testing approach) for testing the long-run and short-run relationship between our variables and analyzing time series data where the sample size is small (Tang, 2010). Furthermore, this technique can be used regardless of the order of integration, that is, whether the variables are integrated of order zero and/ or one, that is, I(0)s and/ or I(1)s. Moreover, this technique is recommended because of its ability to correct for any possible endogeneities among the explanatory variables (Ali, Law & Zannah, 2016). The remainder of the paper is sequenced as follows; section 2 presents the data and descriptive statistics, section 3 is the methodology. Section 4 reports the findings and discussion, while section 5 presents the conclusion.

DATA AND DESCRIPTIVE STATISTICS This paper uses annual time series data from internationally recognized sources that cover the period from 1991 to 2015. Data on our dependent variable, the shadow economy (Se), comes from Medina and Schneider (2018). The above authors use a variety of estimation methods to derive new estimates of the shadow economy that cover a span of 20 years. The data for our core independent variable - political instability, which is proxied by regime durability index, (dur), comes from the Center for Systemic Peace (Polity5 Version) which reports a measure of authority trends in a given period of time. In this case, regime durability is the duration since it is the last regime alteration that changes authority characteristics of a given state (Elbahnasawy et al. 2016). Further, we also control for a number of other proximate measures of the shadow economy in line with the extant literature. We include the following variables: financial development (dob), which is proxied by domestic credit by banks as share of gross domestic product (GDP); institutionalized democracy (demo); fractionalization (Frac); annual growth rate of gross domestic product (GDP) which we denote as (growth); and the ratio of government spending to GDP (gov/gdp). Institutionalized democracy measures the existence of political institutions and processes that facilitate citizens with a means to express preferences about the type of leadership and leaders they want, which puts administrative constraints on the exercise of power by the executive arm of government. 164

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Our measure of fractionalization index (frac) is from the World Bank’s Database of Political Institutions (Cesi, Keefer & Scartascini, 2015) while institutionalized democracy data comes from the Center for Systemic Peace (Polity5 version data, 2019). The value of this index ranges from -10 denoting strongly autocratic to +10, denoting strongly democratic. The remaining variables, domestic credit by banks to private sector (dob), growth (gw) and government spending (gov/gdp), are from World development indicators of the World Bank (2020). In Table 1, panels (a) and (b), we report the summary statistics and correlation matrix, respectively. From Table 1, we observe that the average values of the key variables are: shadow economy (Se) is 38.743, regime durability (dur) is 5.280, domestic credit (dob) is 7.985, institutionalized democracy (Demo) is 0.440, fractionalization (frac) is 0.393, growth (gw) is 3.344, and government spending (gov/gdp) is 11.775. Besides summarizing statistics, we also present the correlation matrix in panel (b). Panel (b) shows a negative correlation between the size of the shadow economy and all the three indicators that measure political processes, that is, dur, demo and frac. However, negative correlation does not necessarily indicate that political instability does reduce the size of shadow economy. We need to formally test this relationship using an empirical model, which this paper attempts to implement in the next section. Table 1. Summary statistics and correlation matrix Se

Dur

Dob

Demo

Frac

Gov/gdp

Panel (a): Summary statistics Mean

38.743

5.280

7.985

0.440

0.393

3.344

11.775

Median

40.720

5.000

7.702

0.000

0.550

3.080

11.757

Maximum

43.250

11.00

13.786

1.000

0.687

8.140

16.792

Minimum

31.880

0.000

3.529

0.000

0.030

6.636

Std.Dev.

4.009

3.260

3.191

0.507

2.245

3.005

Skewness

-0.521

0.003

0.276

0.242

0.360

-0.178

Kurtosis

1.679

1.956

1.770

1.058

2.280

1.878

#Obs.

Panel (b): Correlation matrix Se

1.000

Dur

-0.202

1.000

Dob

-0.893

0.289

1.000

Demo

-0.905

-0.078

0.866

1.000

Frac

-0.496

0.056

0.369

0.354

1.000

0.143

-0.458

-0.028

-0.007

-0.009

1.000

Gov/gdp

0.584

0.012

-0.234

-0.411

-0.273

0.374

1.000

Source: Author’s calculation

165

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

METHODOLOGY Model specification This section presents the empirical model for testing the long-run and short-run relationship between political instability and the size of shadow economy. Thus, we suggest that the shadow economy is a function of variables that can be expressed as: (1) Where Se is the size of shadow economy, dur is a measure of regime durability, dob is domestic credit by banks to private sector, demo is institutionalized democracy index, frac is the index for fractionalization, growth is annual rate of GDP per capita growth, and gov/gdp is the share of government spending to GDP. The choice of these explanatory variables stems from a number of factors as reviewed in the literature. Recently, studies have shown that the political processes determine what policies come into place, or are enacted, which in turn influences public administration. Political process can either create instability and conflict or generate consensus which results into weak (strong) political and democratic institutions (Elbahnasawy et al. 2016). Thus, the political system can therefore affect the welfare of citizens through its impact on the size of shadow economy activities. As shown in the introduction section of this paper, financial development is important for the proper functioning of the financial sector (Berdiev & Saunoris, 2016) and as such it’s a key indicator of the level of informality in the economy (Cahn & Thanh, 2020). Additionally, governance is important for the successful allocation of productive resources and provision of public goods and services hence we include (demo) variable to proxy for governance in the country (Teobaldelli & Schneider, 2013). Furthermore, the level of a country’s development is also important in the provision of public goods and services that have the potential to improve the quality of life of the citizenry, thereby reducing their desire to operate in the shadow economy (see, Baklouti & Boujelbene, 2020; Esaku, 2021b). Some studies use the log of GDP per capita, and in some cases they use GDP growth rate per capita. This study uses GDP growth rate per capita to proxy for the level of economic development. Correspondingly, recent studies have shown that government spending is a proximate cause of the shadow economy (see Buehn & Schneider, 2012b). Accordingly, we include this variable in the main estimation equation. In what follows, we present the econometric methodology in the next section.

Econometric methodology We follow three main steps in the estimation process. The unit root tests are done in the first step, while ARDL bounds tests of the existence of both long-run and short-run relationships between the variables are carried out in the second step. In the third step, we conduct empirical estimation of the ARDL model for relationship level among variables, and diagnostics tests to ensure that the results are reliable, stable and not driven by biases. Thus, we follow the ARDL bounds testing approach to cointegration of Pesaran, Shin & Smith (2001). This approach has a number of benefits over traditional cointegration techniques. Firstly, it is a robust econometric method for analyzing time series data even in cases where the sample size is small (Tang, 2010). 166

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Secondly, it can be applied regardless of the order of integration of the variables, whether the variables are integrated of order zero and/ or one. This means that variables can either be I(0)s and/ or I(1)s but not I(2)s. Thirdly, this method allows for the correction of outliers with impulse dummies (Marques, Fuinhas & Marques, 2017). Fourthly, the interpretation of the ARDL bounds testing method and its implementation can be done in a straightforward manner (Bayer and Hanck, 2013). Given the above benefits of this method, we can specify the ARDL for the empirical model in equation (1) as follows:

(2)

Where, denotes the constant term while run coefficients, respectively, and denotes the error term.

denote the short-run and long-

We implement ARDL bounds testing by testing for cointegration among variables using the F-statistic and t-statistic. The null hypothesis of no cointegration, , is then tested against the alternative hypothesis that there is cointegration among variables, . The outcome of the F-statistic or t-statistic is then compared to the critical values specified in Pesaran et al. (2001). The decision rule is as follows: if the calculated values of F-statistic are above the upper critical bound values, the (the null hypothesis) is rejected and vice versa. However, if the F-statistic values fall within the bounds, then it means that the test result is inconclusive. Before conducting ARDL bounds testing procedure, we first determine the optimal lag length for the ARDL model, which we choose according to the appropriate lag selection criteria based on the Schwartz-Bayesian criterion (SBC). If the results of the cointegration test on equation (2) show cointegration among variables, we can proceed to express the error correction model (ECM) as follows:

(3)

Where denotes the coefficient of the ECT (error correction term) capturing the long-run adjustment to the equilibrium after deviations, while is the residual error term. The importance of the ECT coefficient is in its size and sign, which denotes the speed of adjustment and validity of the results. Thus, the coefficient of the error correction term ( ) should be negative, less than 1 and statistically significant.

167

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

RESULTS AND DISCUSSION This section reports the empirical investigation of the relationship between the shadow economy and political instability. We report the findings of stationarity tests, ARDL bounds tests, and empirical analysis of the relationship between political instability and the size of shadow economy in the subsequent sections.

Stationarity tests Before implementing the empirical analysis, this paper first tested the variables to determine whether they are integrated of order zero, I(0), or order one, I(1) to enable a valid analysis. Accordingly, we used two different tests for unit root. We used Augmented-Dickey-Fuller test (ADF), and Phillip-Perron (PP) tests, with intercepts and with intercepts and trend. The results of these tests are reported in Table (2). The test results in Table (2) confirm that variables are either stationary in levels or after first differencing, and vary according to the type of test used. We then proceeded to conduct the ARDL bounds testing procedure after the stationarity test was implemented. In Table (3) panel (a), we formally express the equation to be tested, while panel (b) reports the results of the ARDL bounds test for cointegration. We can note that the calculated F-statistic is higher than the asymptotic critical value bounds given in Pesaran et al. (2001). This leads to the rejection of the null hypothesis of no cointegration and the conclusion that the variables are cointegrated. Additionally, our diagnostic test shows that these results are reliable and not driven by any biases. After the test, this study proceeded to estimate the long-run and short-run coefficients for the model by first determining the optimal lag length according to the Schwartz information criterion (SIC), which selected ARDL (1,0,0,0,0,0,0) model. Table 2. Results of stationarity tests for all variables In Levels ADF

First difference: PP

ADF

Intercept

Trend & Intercept.

Intercept

Trend & Intercept

Intercept

Trend & Intercept

Intercept

Trend & Intercept

0.087

-2.128

0.087

-2.047

-5.818***

-6.081***

-5.818***

-6.084***

Dur

-1.579

-2.209

-1.637

-2.255

-5.201***

-5.252***

-5.199***

-5.246***

Dob

-0.874

-5.172***

-1.020

-5.175***

-8.460***

-8.289***

-21.746*** -22.550**

Demo

-0.923

-2.036

-0.923

-2.098

-5.000***

-4.888***

-5.000***

-4.888***

Frac

-2.132

-2.118

-2.132

-2.199

-4.632***

-4.559***

-4.632***

-4.559***

Growth

-3.740***

-3.999**

-3.740***

-3.989**

-5.496***

-5.512***

-10.481*** -17.476***

Gov/gdp

-2.165

-2.739

-2.058

-2.652

-5.588***

-3.650*

-7.822***

-11.472***

Source: Author’s calculation. *,**,***, denote statistical significance at 10%, 5% and 1% levels, respectively.

168

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Table 3. Results of the ARDL Bounds test Panel (a): The equation to be tested Equation

Dependent variable

Function

Eq. (1)

Shadow economy

F (Se | dur, dob, demo, frac, growth, gov/gdp)

Panel (b): Results of ARDL bounds test Model

ARDL

F-Stat.

Diagnostics X2(Normality)

Eq. (1)

(1,0,0,0,0,0,0)

7.848

X2 (Heteroscedasticity) X2 (Correlation)

0.646

0.986

0.641

Actual sample size (T=24) Critical values Lower Bound I (0)

Upper Bound I (1)

10%

2.12

3.23

2.45

3.61

2.5

2.75

3.99

3.15

4.43

Source: Author’s calculation

The long-run relationship between political instability and the size of shadow economy The results of the long-run relationship between political instability and the shadow economy are reported in Table 4, column 2. The results indicate that political instability and the shadow economy are directly related in the long-run. Thus, the coefficient on regime durability is negative and statistically significant at 1% level, implying that regime durability (our measure of political instability) has a negative impact on the size of shadow economy in the long-run. Specifically, these results confirm that an improvement in regime durability significantly reduces the size of shadow economy by 0.194 units. This implies that more stable political processes or regime durability is crucial for addressing the factors that drive the increase and expansion of the shadow economy in the long-run. Further, our findings seem to bode well with the suggestion that political processes seem to shape the landscape where businesses operate. A more politically stable business environment provides the impetus for the growth and survival of both domestic and foreign businesses. If investors view the business environment as politically unstable, this discourages investment and could even drive out existing businesses from operations. Given that political processes are important, regimes that last long periods seem to guarantee political stability and economic prosperity, and reduce uncertainty (see Elbahnasawy et al. 2016). This might be the case for Uganda considering that the country has been ruled by one regime for nearly 36 years. Regime durability is important because it reduces a country’s vulnerability to macroeconomic uncertainty, given that no investor is willing to invest in such uncertain business environments. Indeed, political environments in which businesses operate result from government policies that either reinforce or slow down the growth of shadow economy. These results confirm the expectations of this study and have important implications. 169

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

One main implication of this long-run relationship is that political processes that reinforce regime durability provide enabling environment for businesses to operate and produce goods and services that improve the welfare of general population. This implies that citizens with better livelihood opportunities will be less inclined to operate in the shadow economy since there will be better chances of success in the formal economy as shown in Teobaldelli & Schneider (2013). Furthermore, these findings also imply that as political institutions develop, the politicians and bureaucrats are induced to enact and implement fiscal policies that closely offer better chances of utility maximization for citizens as these policies nearly mirror the citizens’ choices and preferences. This makes it less attractive for the country’s citizens to operate in the shadow economy since the formal economy can now provide most of the needed goods and services. We also discuss the remainder of the results in Table (4). We observe that improvements in financial development (dob) and the level of institutionalized democracy (demo) reduce the size of shadow economy by 0.384 units and 3.399 units, respectively; all statistically significant at 1%. This confirms the view that a well-developed financial structure and functioning democracy is crucial in limiting the expansion of shadow activities (see Bayar & Ozturk, 2016; Berdiev & Saunoris, 2016; Teobaldelli & Schneider, 2013). Similarly, we find evidence that a rise in fractionalization and economic growth do limit the increase and expansion of shadow activities. The results indicate that an improvement in fractionalization and economic growth significantly reduce shadow activities by 1.418 units and 0.121 units; statistically significant at 1% and 10% levels, respectively. These findings are consistent with previous studies that suggest that the level of economic growth is crucial in reducing shadow activities in the long-run (Baklouti & Boujelbene, 2020; Esaku, 2021b). Government expenditure has the expected effect on the shadow economy. The results indicate that an increase in government expenditure increases the size of shadow economy by 0.379 units; statistically significant at 1% level. This could be due to the fact that more government spending requires that businesses pay up their tax liability, which in turn affects the operations of these businesses, especially if they are start-ups. Indeed, the literature shows that high taxes imposed by governments on businesses could be one of the key drivers of informality around the world (Elgin, 2015; Goel & Nelson, 2016). As a robustness check, Menegaki (2019) advises the usage of DOLS (dynamic OLS) and FMOLS (fully modified OLS) in order to validate the robustness of the results. This is because the above two econometric techniques generate asymptotically efficient coefficients, since they can address the issue of endogeneity and serial autocorrelation (Menegaki, 2019). We report these results in Table (4) columns 5 and 8. These results are qualitatively and quantitatively similar to the ARDL results. We can observe that an increase in regime durability significantly reduces the size of shadow economy in the long-run, in the case of Uganda. Furthermore, we conduct residual diagnostic to ascertain that the results are driven by bias. Specifically, we conducted Breusch-Godfrey Serial Correlation LM Test, Heteroskedasticity Test (ARCH), and Normality test. The results of these tests are shown on the lower panel of Table 4. These results show that the findings are reliable and not biased. In sum, it can be concluded that an improvement in political institutions, financial development, institutionalized democracy, fractionalization and economic growth could be viewed as long-run effective remedies to addressing the expansion of the shadow activities in Uganda.

170

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Table 4. Long run relationship between political instability and the size of shadow economy Dependent variable: The shadow economy Explanatory

ARDL

Fully Mod. OLS (FMOLS)

Dynamic OLS (DOLS)

Coeff.

t-stat.

Prob.

Coeff

t-stat

Prob.

Coeff.

t-stat.

Prob.

Dur

-0.194***

-3.967

0.001

-0.188**

-2.866

0.012

-0.194***

-3.967

0.001

Dob

-0.384***

-6.618

0.000

-0.390***

-3.383

0.004

-0.384***

-6.618

0.000

Demo

-3.399***

-4.357

0.001

-3.122***

-3.618

0.003

-3.399***

-4.357

0.000

Frac

-1.418***

-3.594

0.002

-1.198**

-2.814

0.013

-1.418***

-3.594

0.002

Growth

-0.121*

-2.002

0.063

-0.128*

-2.065

0.057

-0.121*

-2.002

0.063

Gov/gdp

0.379***

3.678

0.002

0.045***

6.532

0.000

0.379***

3.678

0.002

Constant

37.516

6.852

0.000

35.104***

8.351

0.000

37.516***

6.852

0.000

R-sq.

0.979

0.976

0.979

R-bar-sq.

0.969

0.966

0.969

Durb W.

2.104

Residual diagnostics X2 (Correlation-BG LM Test) X2 (Heteroskedasticity-ARCH Test) X2 (Normality Test)

Source: Author’s calculation. Note: *,**,***, indicate statistical significance at 10%, 5% and 1% levels, respectively. HAC standard errors and covariance (Bartlett kernel, Newey-West fixed bandwidth=3.000) used

The short-run relationship between political instability and the size of shadow economy The findings of the short-run relationship are shown in Table 5, column 2. We find evidence of a short-run relationship between political instability and the size of shadow economy. We observe that, in the short-run, an improvement in regime durability significantly reduces the size of shadow economy by 0.182 units, statistically significant at 10% level, all else equal. The implication of these results is that political instability could be effective in reducing the shadow economy in the short-run, as well as in the long-run. This is probable given the fact that political processes involve negotiations and building alliances, a process which if done well could create policies that are important in allocation of productive resources and provision of public goods and services. Relatedly, this paper also finds evidence confirming that the control variables in the estimation equation are significant determinants of the shadow economy in the short-run. First, this paper establishes that an improvement in financial development is a significant determinant of shadow activities in both the short-run and long-run. We can observe that the coefficient on (dob) is negative and statistically significant at 1% level. This implies that an improvement in the financial structure of the overall economy significantly improves access to financing for businesses, thus reducing their (businesses) incentive to operate informally (see Berdiev & Saunoris, 2016). 171

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Similarly, improvement in institutionalized democracy, growth, and fractionalization significantly limit the expansion of the shadow economy in the short-run as well as in the long-run. We observe that an improvement in democracy, economic growth and fractionalization significantly reduces the size of shadow economy by 3.663 units, 1.938 units and 0.096 units, respectively. This is consistent with previous findings in the literature indicating that improvements in the above variables do improve the welfare of citizens resulting from political processes that rationalize the allocation of resources (see Elbahnasawy et al. 2016; Teobaldelli & Schneider, 2013) Furthermore, these results show that the lagged coefficient of the error correction term (ECT) is negative and statistically significant at 1% level. The ECM results indicate that the shadow economy adjusts to any deviations from long-run equilibrium at a speed of adjustment that is shown by the coefficient of the lagged error correction term being 94.1% and is statistically significant at 1% level. As a robustness check, we carried out residual diagnostic as before to ensure that the results are not biased. Breusch-Godfrey Serial Correlation LM Test, Heteroskedasticity Test (ARCH), and Normality test were conducted and results reported on the lower panel of Table 5. The results indicate that the ARDL model results are reliable and not biased. In sum, these results seem to suggest that improvement in political institutions in Uganda could be an effective weapon for addressing the expansion of the shadow economy in Uganda, in both the short-run and long-run, since strong political institutions seem to strengthen citizens’ engagement in political processes. The practical implication of these findings is that any attempts by policy makers to reduce activities of the shadow economy should also involve reforming the political system by encouraging political reforms and civic engagement between the political elites and the voters. Table 5. Short-run relationship between political instability and the size of shadow economy Explanatory variable

Outcome variable: Shadow economy Coefficient

t-statistic

Probability

-0.182*

-2.090

0.055

∆Dob

-0.375***

-4.443

0.000

∆Democracy

-3.663***

-3.107

0.008

∆Government fractionalization

-1.938***

-3.767

0.002

∆Growth

-0.096*

-2.042

0.060

∆Government spending/gdp

0.342***

6.670

0.000

ECM(-1)

-0.941***

-3.759

0.002

Constant

-0.034

-0.239

0.815

R-squared

0.799

R-bar-squared

0.685

Durbin Watson

2.193

∆Regime durability

Residual diagnostics: X2 (Correlation BG LM Test)

0.203

X (Heteroskedasticity-ARCH Test)

0.858

X (Normality Test)

0.603

2 2

Source: Author’s calculation. Note: **, ***, indicate statistical significance at 10%, and 5% levels respectively. Ordinary covariance method is used.

172

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Diagnostic tests To confirm the reliability of these results we conduct stability diagnostics and report the plots of cumulative sum of recursive residuals (CUSUM) and cumulative sum of squares of recursive residuals in Figures (1) and (2). As shown in these figures, the residual plots are tightly bound within the boundaries at a significance level of 5% and provide additional evidence on the stability of the estimated model. In what follows, we sum up by arguing that these results show evidence of a negative and statistically significant relationship between political instability and the size of shadow economy in both the long-run and short-run in the case of Uganda. Figure 1. Long-run Plot of Cumulative sum (CUSUM) and Cumulative Sum of Squares (CUSUM) of Recursive Residuals for ARDL model 10.0

1.6

Long-run: Plot of CUSUM of Recursive Residuals

Long-run: Plot of CUSUMQ of Recursive Residuals

7.5

1.2

5.0 2.5

0.8

0.0

0.4

-2.5 -5.0

0.0

-7.5 -10.0 16

CUSUM

-0.4 16

5% Significance

CUSUM of Squares

5% Significance

Figure 2. Short-run Plot of Cumulative sum (CUSUM) and Cumulative Sum of Squares (CUSUM) of Recursive Residuals for ARDL model 10.0

1.6

Short-run: Plot of CUSUM of Recursive Residuals

Short-run: Plot of CUSUMQ of Recursive Residuals

7.5 1.2

5.0 2.5

0.8

0.0 0.4

-2.5 -5.0

0.0

-7.5 -0.4

-10.0 16

CUSUM

5% Significance

CUSUM of Squares

5% Significance

173

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

CONCLUSIONS This paper analyses the relationship between political instability and the size of shadow economy in a country-level analysis using ARDL modeling technique. We use annual time series data drawn from a variety of data sources, covering the period from 1991 to 2015. The results indicate a negative and statistically significant relationship between political instability and the size of shadow economy, in both the short-run and the long-run. These findings suggest that, in the case of Uganda, an improvement in political processes that creates stability in the regime significantly reduces the size of shadow economy. These results are consistent with the view that political institutions play a crucial role in facilitating political processes, which in turn reinforce the allocation of economic resources and provision of public goods and services that improve the welfare of the citizens. Improvement in citizens’ welfare reduces their incentive to operate in the shadow economy and increases citizens’ engagement in the political processes. In sum, this paper provides evidence that improvement in political instability significantly reduces the size of shadow economy in Uganda. These results seem to suggest that improvement in political institutions could be an effective weapon in the long-run, for addressing the expansion of the shadow economy in Uganda since strong political institutions strengthen citizens’ engagement in political processes. The practical implication of these findings is that any attempts by policy makers to reduce activities of the shadow economy should also involve reforming the political system by encouraging political reforms and civic engagement between political elites and voters. Furthermore, policy makers should formulate and enact policies that reinforce the operation of political institutions, independent of any interference from the political elites. As is the case with some studies, one main limitation of this study is that it uses data on the size of the shadow economy that covers 25 years. We suggest that future studies use historical data on the size of shadow economy that covers more years.

ACKNOWLEDGEMENT The author is thankful to the editor and anonymous reviewers of this journal for their helpful comments that improved the quality of this paper. The remaining errors and omissions are solely the responsibly of the author.

174

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

REFERENCES Ali, H.S., Law, H., & and Zannah, T.I. (2016). Dynamic impact of urbanization, economic growth , energy consumption, and trade openness on CO2 emissions in Nigeria. Environmental Science and Pollution Research, 23, 12435–1243. DOI:10.1007/s11356-016-6437-3 Alm, J. & Embaye, A. (2013). Using Dynamic Panel Methods to Estimate Shadow Economies Around the World, 1984-2006. Public Finance Review, 41(5), 510-543 DOI:10.1177/1091142113482353 Baklouti, N., & Boujelbene, Y. (2020). A simultaneous equation model of economic growth and shadow economy: Is there a difference between the developed nd developing countries? Economic Change and Restructuring, 53, 151-170. DOI: 10.1007/s10644-018-9235-8 Bayar, Y. & Ozturk, O.F. (2016). Financial Development and Shadow Economy in European Union Transition Economies. Managing Global Transitions, 14(2), 157-173 Bayer, C., & Hanck, C. (2013). Combining non-cointegration tests. Journal of Time Series Analysis, 34(1),83–95. DOI: 10.1111/j.1467-9892.2012.00814.x Berdiev, A.N., & Saunoris, J. (2016). Financial development and the shadow economy: A panel VAR analysis. Economic Modelling, 57, 197-207. DOI:10.1016/j.econmod.2016.03.028 Berdiev, A.N., Pasquesi-Hill, C. & Saunoris, J.W. (2015). Exploring the dynamics of the shadow economy across US states. Applied Economics, 47(56), 6136-6147: DOI:10.1080/00036846.2015.1064081 Blackburn, K., Bose, N., & Capasso, S. (2012). Tax evasion, the underground economy and financial development. Journal of Economic Behaviour and Organization, 83, 243-253. Bucek, J. (2017). Determinants of the Shadow Economy in the Czech Regions: A Region-Level Study. Review of Economic Perspectives-Narodohospodarsky Obzor, 17(3), 315-329. DOI:10.1515/revecp-2017-0016. Buehn, A., & Schneider, F. (2012а). Corruption and the shadow economy: like oil and vinegar, like water and fire? International Tax and Public Finance, 19(1), 172-194. DOI:10.1007/s10797-011-9175-y Buehn, A., & Schneider, F. (2012b). Shadow economies around the world: novel insights, accepted knowledge, and new estimates. International Tax and Public Finance, 19, 139-171. DOI:10.1007/s10797-011-9187-7. Canh, N.P., & Thanh, S.D. (2020). Financial development and the shadow economy: A 0multi-dimensional analysis. Economic Analysis and Policy, 67, 37-54. DOI:10.1016/j.eap.2020.05.002 Capasso, S., & Jappelli, T. (2013). Financial development and the underground economy. Journal of Development Economics, 101, 167-178. DOI:10.1016/j.deveco.2012.10.005 Cesi, C., Keefer, P., & Scartascini, C. (2015). Database of Political Institutions (DPI2015)', Inter-American Development Bank. Updated version of Beck,T., Clarke, G., Groff, A., Keefer, P., and Walsh, P, 2001. 'New tools in comparative political economy: The Database of Political Institutions. World Bank Economic Review, 15(1), 165-176 Elbahnasawy, N.G., Ellis, M.A., and Adom, A. D. (2016). Political Instability and the Shadow Economy. World Development, 85, 31-42. DOI:10.1016/j.worlddev.2016.04.009 Elgin, C. (2015). Informal Economy in a Dynamic Political framework. Macroeconomic Dynamics, 19, 578-617. DOI:10.1017/S1365100513000497 Elgin, C. (2020). Shadow Economies Around the World: Evidence from Metropolitan Areas. Eastern Economic Journal, 46, 301-322. DOI:10.1057/s41302-019-00161-4 Elgin, C., & Oztunali, O. (2014). Institutions, Informal Economy, and Economic Development. Emerging Markets Finance and Trade, 50(4), 145-162. DOI:10.2753/REE1540-496X500409 Esaku, S. (2020). Job creation, job destruction and reallocation in Sub-Saharan Africa: Firm-level evidence from Kenyan manufacturing sector. Cogent Economics & Finance, 8(1), 1-22. DOI:10.1080/23322039.2020.17 82113 Esaku, S. (2021).Does the shadow economy increase income inequality in the short- and long-run? Empirical evidence from Uganda. Cogent Economics & Finance, 9(1), 1-21. DOI: 10.1080/233220392021.1912896 175

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

Fourie, F. (2018). Enabling the forgotten sector: Informal sector realities, policy approaches and formalization in South Africa. In Frederick C.v.N.Fourie (Ed.), The South African informal sector: Creating jobs, reducing poverty. Human Sciences Research Council, pp. 439-476 Goel, R. K., & Nelson, M. A. (2016). Shining a light on the shadows: Identifying robust determinants of the shadow economy. Economic Modelling, 58, 351-364. DOI:10.1016/j.econmod.2016.06.009 González-Fernández & González-Velasco, C. (2014). Shadow economy, corruption and public debt in Spain. Journal of Policy Modeling, 36(6), 1101-1117. DOI: 10.1016/j.polmod.2014.10.001. Hajilee, M., Stringer, D.Y., & Metghalchi, M. (2017). Financial market inclusion, shadow economy and economic growth: New evidence from emerging economies. The Quarterly Review of Economics and Finance, 66, 149-158. DOI: 10.1016/j.qref.2017.07.015 Huynh, C.M., Nguyen, V.H.T, & Nguyen, H.B. (2020). One –way effect or multiple-way causality: foreign direct investment, institutional quality and shadow econom? International Economics and Economic Policy, 17, 219-239. DOI: 10.1007/s10368-019-00454-1 Kanbur, R. (2017). Informality: Causes, consequences and policy responses. Review of Development Economics, 21(4), 939-961. DOI:10.1111/rode.12321 Luong, T.T.H., Nguyen, T.M., & Nguyen, T.A.N. (2020). Rule of Law, Economic Growth and Shadow Economy in Transition Countries. The Journal of Asian Finance, Economics and Business, 7(4), 145-154. DOI:10.13106/ jafeb.2020.vol7.no4.145 Marques, L.M., Fuinhas, J.A., & Marques, A.A. (2017). Augmented energy-growth nexus: Economic, political and social globalization impacts. Energy Procedia, 136: 97–101. DOI:10.1016/j.egypro.2017.10.293 Medina, L., & Schneider, F. (2018). Shadow Economies Around the World: What Did We Learn over the Last 20 Year?. IMF Working Paper No. WP/18/17 Menegaki, A. N. (2019). The ARDL Method in the Energy-Growth Nexus Field; Best Implementation Strategies. Economies, 7(4), 1-16. DOI:10.3390/economies7040105 Mugoda, S., Esaku, S., Nakimu, R. K., & Bbaale, E. (2020). The portrait of Uganda's shadow sector: what main obstacles do the sector face? Cogent Economics & Finance, 8(1), 1-28. DOI:10.1080/23322039.2020.1843255 Orsi, R., Raggi, D., & Turino, F. (2014). Size, trend, and policy implications of the underground economy. Review of Economic Dynamics, 17(3), 417-436. DOI:10.1016/j.red.2013.11.001 Pesaran, H., Shin, Y., & Smith, R.J. (2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16, 289-326. DOI:10.1002/(ISSN)1099-1255. Polity IV Database (2018). Polity IV project: Political regime characters and transitions, 1800–2018. Available from http://www.systemicpeace.org/polity/polity4.htm. (Retrieved May 11, 2020). Tang, T.C. (2010). A reassessment of aggregate import demand function in the Asean-5: A cointegration analysis. The International Trade Journal, 18(3), 239-268 Teobaldelli, D., & Schneider, F. (2013). The influence of direct democracy on the shadow economy. Public Choice, 157, 543-567. DOI:10.1007/s11127-013-0098-2 World Bank, Africa Development indicators 2019/20 computer file https://datacatalog.worldbank.org/dataset/ africa-development-indicators. Accessed on 13 January, 2020.

176

EJAE 2021  18 (2)  161 - 177

ESAKU. S.  POLITICAL INSTABILITY AND INFORMALITY IN UGANDA: AN EMPIRICAL ANALYSIS

POLITIČKA NESTABILNOST I NEFORMALNOST U UGANDI: EMPIRIJSKA ANALIZA Rezime: U ovom radu smo analizirali dugoročni odnos između političke nestabilnosti i sive ekonomije u Ugandi koristeći autoregresivni pristup testiranja granica zaostajanja za kointegraciju. Utvrdili smo negativan i statistički značajan odnos između političke nestabilnosti i sive ekonomije, kako dugoročno tako i kratkoročno. To implicira da poboljšanje političkih procesa stvara stabilnost u aktuelnom režimu i značajno smanjuje sivu ekonomiju, u skladu sa stavom da političke institucije igraju ključnu ulogu u omogućavanju političkih procesa, što zauzvrat pojačava raspodelu ekonomskih resursa i obezbeđivanje javnih dobara i usluga koje poboljšavaju blagostanje građana. Zbog toga je građanima manje privlačno da rade u sivoj ekonomiji, jer formalna ekonomija sada može da obezbedi veći deo potrebne robe i usluga. Praktična implikacija ovih rezultata je da bi svaki pokušaj kreatora politike da smanje aktivnosti u sivoj ekonomiji trebao uključivati i reformu političkog sistema i podsticanje građanskog angažmana između političkih elita i građana ili birača. Osim toga, kreatori politike treba da formulišu politiku koja jača funkcionisanje političkih institucija nezavisno od bilo kakvog uplitanja političkih elita u ponašanje koje traži rentu.

Ključne reči: siva ekonomija, upravljanje, demokratija, neformalni sektor, politička ekonomija.

177

CIP - Каталогизација у публикацији Народна библиотека Србије, Београд 33

The EUROPEAN Journal of Applied Economics / editor-in-chief Žaklina Spalević. - Vol. 12, No. 1 (2015)- . - Belgrade : Singidunum University, 2015- (Belgrade : Caligraph). - 28 cm Polugodišnje. - Je nastavak: Singidunum Journal of Applied Sciences = ISSN 2217-8090. Drugo izdanje na drugom medijumu: The European Journal of Applied Economics (Online) = ISSN 2406-3215 ISSN 2406-2588 = The European Journal of Applied Economics COBISS.SR-ID 214758924

Vol. 18 Nº 2

journal.singidunum.ac.rs

Vol. 18 Nº 2 OCTOBER 2021 journal.singidunum.ac.rs

2021

Economic growth of the tourism sector in a Covid-19 pandemic during 2021 pp. 1-14

Corruption impact on east European emerging markets development pp. 39-54

Structural Breaks, Twitter and the Stock Liquidity of Internet Dot-com Company: Evidence from US Companies pp. 15-38

English digital playground: Friend or foe to the children? pp. 55-72

The long-run impact of bank credit growth on social and economic inequalities in Morocco evidence from the Johansen's cointegration analysis pp. 106-125

Macroeconomic determinants of tax revenue in economic community of west African states pp. 73-88 Concentration of supply on the chosen markets of Serbian electronic communications sector pp. 89-105

Political instability and informality in Uganda: an empirical analysis pp. 151-172