Sustaining improved estimates of poverty in Armenia: An application of multivariate raking ratio adjustment

Abstract

The objective of this paper is to present an application of multivariate raking ratio adjustment with some refinements. Specifically, we applied raking to help improve the measurement of poverty in Armenia. This work was undertaken in conjunction with a Millennium Challenge Corporation (MCC) 2006 grant to reduce rural poverty in Armenia through economic growth. The Armenian Integrated Living Conditions Survey (ILCS) was used to measure poverty, so that the impact of the MCC grant could be assessed. We used raking techniques to adjust survey sampling weights so that totals match that of Armenian census projections. This adjustment required the incorporation of a refinement to the standard raking procedure to adjust multiple demographic dimensions: by person-level variables age and gender, and household-level geography variables. This paper describes the application of raking in this context, and provides background on the MCC work in Armenia. We contend that the work done here was a success because of its collaborative nature, which is leading to a fully sustainable effort long-term. Collaborative work of this type, across cultures and, indeed, across the world is rewarding, both professionally and personally. It never ceases to amaze how much official statisticians have in common, wherever they are found.