Table 2. Governance and fiscal cyclicality - all countries 2000–2020

accountability across countries. The coefficient of interest is β2. It captures how procyclical fiscal expenditures are in response to accountability. The level of accountability determines the cyclicality of fiscal expenditures as follows, and note that a positive β implies more procyclicality.

ġc,t / ẏc,t = β1 + β2LogAccc,t-1 + β3 Logypcc,t-1 (2)

Table 2 shows the association between accountability and the cyclicality of fiscal expenditures in all years, in good (i.e., HP-filtered GDP>=0) and bad (i.e., HP-filtered GDP<0) times for all countries. Table 3 shows the association for Middle East and North Africa countries only.

Table 2. Governance and fiscal cyclicality - all countries 2000–2020

Y = Cyclical Fiscal Expenditure All Years Good Times Bad Times Vertical Diagonal Horizontal Vertical Diagonal Horizontal Vertical Diagonal Horizontal (1) (2) (3) (4) (5) (6) (7) (8) (9)

Cyclical GDP

Cyclical GDP * Log real per capita GDP (t-1) Cyclical GDP * Log vertical acc (t-1) Cyclical GDP * Log diagonal acc (t-1) 1.938** 1.994** 2.014* 1.691** 1.486* 1.578 2.405** 2.512*** 2.535** (0.836) (0.765) (1.040) (0.804) (0.774) (1.002) (0.979) (0.934) (1.136) -0.145 -0.168** -0.146 -0.102 -0.108 -0.106 -0.206** -0.232** -0.195* (0.090) (0.083) (0.105) (0.089) (0.078) (0.095) (0.100) (0.100) (0.116) -0.026 (0.081) 0.209** (0.085) -0.121 (0.116)

-0.171 (0.115) -0.128 (0.155) -0.201* (0.117)

Cyclical GDP * Log horizontal acc (t-1)

0.045 (0.249)

0.023 (0.260)

0.070 (0.222)

Constant

0.021*** 0.021*** 0.021*** 0.028*** 0.029*** 0.029*** 0.017*** 0.016*** 0.018*** (0.005) (0.005) (0.005) (0.009) (0.009) (0.009) (0.006) (0.006) (0.006) Number of observations 3,459 3,459 3,459 1,676 1,676 1,676 1,783 1,783 1,783 Number of countries 171 171 171 171 171 171 171 171 171 Year fixed effects yes yes yes yes yes yes yes yes yes Country fixed effects no no no no no no no no no

Adjusted R2 0.111 0.115 0.111 0.094 0.093 0.091 0.133 0.138 0.132

Source: Fiscal expenditure (as a percentage of gross domestic product) data are obtained from the International Monetary Fund World Economic Outlook dataset, real per capita GDP (constant 2015 US$) and real per capita GDP (constant local currency unit (LCU)) are obtained from the World Bank World Development Indicators dataset, and the accountability indices (vertical, diagonal, horizontal) are obtained from the V-DEM Varieties of Democracy dataset. Note: For level of significance in regressions, *** 1%; ** 5%; * 10%. Cyclical fiscal expenditures are obtained by applying a Hodrick–Prescott (HP) filter to isolate the cyclical component of log of real fiscal expenditures (constant LCU). Real fiscal expenditures are obtained by multiplying share of fiscal expenditures (as a percentage of GDP) by real per capita GDP (constant LCU). Cyclical GDP is obtained by applying an HP filter to isolate the cyclical component to the log of real per capita GDP (constant LCU). All accountability measures are normalized from 0 to 1. Good times are when the cyclical GDP is positive. Bad times are when the cyclical GDP is negative.

For all countries from 2000 to 2020, after controlling for the lagged log of per capita GDP, the associations between the three measures of accountability and fiscal cyclicality β2 are not statistically significant at the 10 percent level. Moreover, as will become clearer, the association between vertical accountability and fiscal expenditures has opposite signs in good and bad times.

In good times, from 2000 to 2020, countries with better lagged vertical accountability have more procyclical fiscal expenditures (column 4 of Table 2). The coefficient of 0.209 is statistically significant at the 5 percent level. The coefficient 0.209 implies that when the vertical accountability score increases by 1 percent, the cyclicality of government