Timothy Beatty and Thomas Crossley
Abstract:
In two influential papers, Hamilton (2001) and Costa (2001) proposed a simple, straightforward and intuitively appealing method to estimate bias in the CPI as a cost-of-living index. In this paper, we show that while the Hamilton/Costa approach and the CPI each capture changes in the cost-of-living for a representative household, they are unlikely to capture changes in the cost of living for the same household. Thus the Hamilton/Costa approach conflates CPI bias with measuring the cost-of-living at different points of the expenditure distribution.