BSM Targeting Case Study

This section examines the issue of BSM targeting. The study begins by exam-ing the household data and providing a descriptive statistical analysis. Then, estimates based on a probit model are used to investigate the issues.
In order to determine eligibility households, I order households in the data set from worst to best based on expenditure per capita. Then, I estimate the con-sumption quartile in which a household lies (see table 5-6).
Table 5-6: Quartile of expenditure per capita/month in rupiahs ( Author’s Calculation using Susenas Core 2013)
Table 5-6 describes the range of expenditure per capita for different quartiles. There are 284,603 observations in the household data and those who are eligi-ble for BSM program are those with a maximum expenditure per capita of IDR. 365,161.6.
While the coverage of the BSM program is increasing every year, unfortunately effective implementation of the program is quite low due to exclusion and in-clusion errors. As simple illustration designed by Grosh (1994) and Hoddinot (1999), table 5-7 capture detailed implementation of BSM regarding the exist-ence of the

The inclusion error also determines the ratio of leakage rate that is the number of inclusion error divided by the total number of BSM re-ceiver. This result shows the leakage rate column in the table. The leakage rate within the household with children of 7-18 years old is 0.50 (0.59 and 0.71) for BSM primary (junior and senior). Meanwhile, the under coverage rate ratio in each BSM program within this sample is 0.94, 0.98 and 0.99 for primary, junior and senior high school. This number also explains that the program implementation reaches the target around 6, 2, and 1 percent for each type of program. It is higher than World Bank (2012b) reported using Susenas data 2009. In 2009, the program covered the poorest 20 percent of households (4.0 percent of primary