Many different approaches have been proposed to determine who is poor and who is not. The gold standard is a household expenditure survey; that is very time-consuming and expensive, so it cannot be done by small organizations or on a massive scale. But poverty scorecards—a relatively new approach based on data from country’s own national survey—have proven to be effective at estimating expenditure-based poverty quickly and easily. And they work for any expenditure-based poverty line, including $1/day and national poverty lines.
During a recent presentation at IDB headquarters, Mark Schreiner, Director of Microfinance Risk Management, L.L.C. and a Senior Scholar at the Center for Social Development at Washington University in Saint Louis, explained how “scoring produces a single, simple, inexpensive estimate of poverty status that is transparent, verifiable, and based on an objective poverty line.”
Unlike the “direct approach” that uses 2–8 hour household expenditure surveys, the poverty scoring “indirect approach” uses 10 simple and verifiable indicators, such as the type of roof on the house or whether the household cooks with wood. The process is simple enough for field workers to compute on paper by hand, in 5 minutes, without the need for software.
One of the scorecard’s best features is its simplicity. Scores range from 0 (most likely poor) to 100 (least-likely poor). For the example of Peru, a score of 15-19 corresponds to a poverty likelihood of 92%, meaning that among households in the national expenditure survey, 92 percent who scored 15–19 had expenditure below the national poverty line.
Indicators and their weights are selected based not on the analyst’s subjective judgments but rather on statistics (logical regression) and a set of other practical criteria that relate indicators to the known poverty status for surveyed households while ensuring that the selected indicators make sense to users. Indicators are limited to those that a field worker can check in a visit to the homestead.
Poverty scorecards are designed to fit local needs. Although they were designed in the context of microfinance, they are not microfinance-specific and work just as well for any pro-poor program.
The three basic uses of poverty scoring are to measure poverty rates (% poor), track changes in poverty rates over time, and to help target services. Accuracy of the scorecards for estimates of poverty rates has been documented to 90-percent confidence level, +/– 2 percentages points if the sample size is at least 5,000.
Scorecards have been piloted in 10 countries and have been adopted for regular, on-going use by BRAC and ASA in Bangladesh, two of the world’s titans in microfinance with 7 million clients each. In Latin America and the Caribbean, scorecards have been developed for Bolivia, El Salvador, Guatemala, Haiti, Mexico, Nicaragua, and Peru.
Sergio Navajas of the IDB’s Multilateral Investment Fund said that scorecards are a good way of determining who the Bank’s programs are reaching, since they are so inexpensive and effective.