July 18, 2008 - The World Bank was recently requested by the Government of Pakistan to examine the validity of the Government’s official poverty estimates for 2005-06.
Using the Pakistan Social and Living Standards Measurement Survey (PSLM), the Centre for Research on Poverty Reduction and Income Distribution (CRPRID) – a research institute responsible for producing the official poverty estimates – had estimated that 22.3 percent of Pakistan’s population was poor in 2005-06.
By performing various validation exercises, the World Bank concluded that CRPRID followed the official methodology accurately. The estimate for 2005-06 (22.3 percent) was also robust to small changes in data cleaning protocol and selection of price index. The validation exercise also found that poverty estimates declined between 2004-05 and 2005-06 but the reduction is not statistically significant.
A caveat for poverty trend analysis: A poverty trend should be measured with poverty estimates derived from the same methodologies/assumptions over the period of interest. This is because poverty estimates often vary substantially depending on what methodologies/assumptions are adopted. For example, one of the previous validation exercises showed an alternative poverty estimate of 29 percent for 2004-05, which is not directly comparable with any estimate from this validation exercise because the poverty line of 2004-05 is estimated differently.
The World Bank is currently preparing a new poverty assessment, which will further deepen the analysis conducted in the context of this validation exercise and include more detailed poverty diagnostics to understand the status of the poor in a comprehensive manner.
Access validation exercise paper
|