Handbook on Poverty and Inequality

Page 246

Haughton and Khandker

11

lower. This introduces a considerable degree of arbitrariness into the determination of poverty rates. To solve this problem Pradhan et al. (2000) propose an iterative procedure. First choose the cost of food for a reference population and then proceed to compute the poverty line. Now take, as the reference group, those households close to this poverty line. Recompute the poverty rate. Repeat the procedure until the poverty line stabilizes. It is to be hoped that this algorithm will converge, but it appears to do so in practice. Using the iterative method, and applying an Engel curve procedure to compute nonfood expenditures (as done by Said and Widyanti 2001), Pradhan et al. (2000) find a headcount poverty rate of 27.1 percent on the basis of the February 1999 Susenas. This is somewhat higher than the official BPS figure of 23.6 percent. The numbers summarized in table 11.7 (from Pradhan et al. 2000, table 3) also show that the iterative method finds a wide difference in poverty rates between urban and rural areas, in contrast to the modest gap reported by the official BPS numbers.

The Problem of Deflation: Poverty Rates over Time. To compare poverty rates over time, it is necessary to deflate the poverty lines to account for inflation. If prices rise modestly, and especially if all prices rise at about the same rate, then deflating poverty lines is relatively straightforward and the results are typically satisfactory. In the Indonesian case, however, the price of food in the consumer price index (CPI) rose by 160 percent between February 1996 and February 1999, while the rise in the price of the nonfood components of the CPI was just 81 percent. Not only was there high overall inflation, but also a massive change in relative prices over this

Table 11.7 Alternative Measures of Poverty in Indonesia Reference population (Rp/month) Iterative method Urban Rural Ratio BPS method Urban Rural Ratio

Lower limit

Upper limit

Poverty line (Rp/month)

Poverty incidence (%)

72,392 64,947 1.11

108,588 97,421 1.11

90,490 81,184 1.11

16.3 34.1 overall: 27.1

80,000 60,000 1.33

100,000 80,000 1.25

93,869 73,8998 1.27

20.0 25.9 overall: 23.6

Source: Pradhan et al. 2000, table 3.

222

Note: BPS = Indonesian Statistics Office (official) approach.


Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.