Does the CPI mirror costs-of-living?

There is considerable interest in identifying the magnitude of the difference between increases in CPI and costs-of-living, and this article uses the technique proposed by Hamilton (2001) to measure this discrepancy for Norway for the 90s. The method is extended along several dimensions by introducing a framework in which measurement errors are modelled. A non-parametric approach is then employed to segment households into demographic types while allowing for flexibility in costs-of-living increases for different standards. Hamilton finds that American CPI overstates costs-of-living in the U.S. for the period 1974-1991, Norwegian results for 1990-1999 indicate that CPI sometimes may understate costs-of-living, perhaps because of a credit-financed boom in house prices. The Norwegian CPI rose 22 percent in the period, but the general consumer behaved as if costs-of-living increased more than 35 percent. For some segments of society, for example single-person households, the increase was substantially larger, suggesting potentially important distributional effects.