On How to Measure Poverty

The meaningfulness of a poverty measure depends on both the social acceptability of the poverty line and the accuracy of the measured distribution of purchasing power to which the line is applied. In particular, the use of linear programming to construct a minimum food budget and the uncritical use of the 1975 Family Income and Expenditures Survey of the National Census and Statistics Office has led Tan and Holazo(1978) to some very questionable conclusions. As early as 1975, the Social Indicators Project, making the first effort to construct poverty thresholds for application to a national cross-section of purchasing power (namely the FIES series of the NCSO), showed that absolute and relative poverty worsened during 1961-1971. In 1977 the PREPF project found that the 1975 FIES had very serious defects, and rejected using it to update the poverty trend; on the basis of independent PREPF surveys, adjusted for consistency with the National Income Accounts, it felt that about three-fifths of all households in 1975 should be rated as poor (below P10,000 annual income per households, somewhat more generous than the average SIP Total Thresholds), but could not draw any conclusion about the 1971-1975 trend.