Response to IRIS Questions on Poverty Scorecards

This memo responds to Grameen Foundation, USA’s (GFUSA’s) request for comments on IRIS’ questions on approaches to measuring poverty status. In the memo, Mark Schreiner states that he has been in communication with IRIS since January 2004, when Chris Dunford of Freedom from Hunger shared his concept note outlining basic ideas of poverty scoring with national survey data.

Mark Schreiner had communicated with Freedom from Hunger that their proposed approach for dealing with measurement error in expenditure ended up counting some people above the poverty line as poor. Schreiner hopes that the memo will improve poverty measurement and help the poor. Key points include:

• Accuracy matters, but practicality also matters (and probably matters more);
• Small technical differences probably do not matter;
• Logit is standard, in theory and practice;
• Quantile and two-step methods can improve classification of the poor.

The memo addresses IRIS’ specific questions after discussing these points. It states that GFUSA has produced scorecards that are quantitative, data-based and empirically validated, fulfilling the Congressional mandate and worthy of USAID’s imprimatur, in its search for simple tools useful to its affiliates.