International income poverty measurement: Which way now?

In this paper, we critically review conceptual and empirical issues surrounding the derivation of the international poverty line, expressed in PPP-adjusted dollars and linked to various rounds of the International Comparison of Prices (ICP). We find that there are some limitations in the current estimation of these lines, but show that statistically superior methods lead to lines that are relatively robust and confirm the $ 1.25 using 2005PPPs and suggest $ 1.67-1.71 using 2011PPPs; they also roughly confirm the current shape of the proposed "weakly relative" poverty line. Using the new absolute line using 2011 PPPs would lead to substantially lower poverty in our estimation. The extent of the decline depends on whether and how one treats China, India, and Indonesia differently from other countries in the 2005 and 2011 PPPs. More seriously, we note that the dependence on the conceptual and empirical problems associated with the link to successive ICP rounds creates problems that have gotten worse over time so that we suggest that it would be best to consider alternatives to the current reliance on ICP rounds and the resulting PPPs. As a quick fix we propose to fix the international poverty line in national currencies using either the 2005 or 2011 level; in the medium term, we argue for global poverty measurement based on internationally coordinated national poverty measurement.