A Simple Poverty Scorecard for the Philippines

How accurate are estimates of households’ poverty likelihoods? To measure, thescorecard is applied to 1,000 bootstrap samples of size n = 16,384 from the validationsub-sample. Bootstrapping entails: 17• Score each household in the validation sample• Draw a new bootstrap sample with replacement from the validation sample• For each score, compute the true poverty likelihood in the bootstrap sample, that is,the share of households with the score and income below a poverty line• For each score, record the difference between the estimated poverty likelihood(Figure 4) and the true poverty likelihood in the bootstrap sample• Repeat the previous three steps 1,000 times• For each score, report the average difference between estimated and true povertylikelihoods across the 1,000 bootstrap samples• For each score, report the two-sided interval containing the central 900, 950, or 990differences between estimated and true poverty likelihoodsFor each score range and for n = 16,384, Figure 7 shows the average differencebetween estimated and true poverty likelihoods as well as confidence intervals for thedifferences.For the national line in the validation sample, the average poverty likelihoodacross bootstrap samples for scores of 20–24 in the validation sample is too high by 1.4percentage points (Figure 7). For scores of 25–29, the estimate is too low by 1.3percentage points. 1817Efron and Tibshirani, 1993.18These differences are not zero, in spite of the estimator’s unbiasedness, because thescorecard comes from a single sample. The average difference by score would be zero ifsamples were repeatedly drawn from the population and split into sub-samples beforerepeating the entire scorecard-building process.25