Sources of Bias and Solutions to Bias in the CPI

Four sources of bias in the Consumer Prices Index (CPI) have been identified. The most discussed is substitution bias, which creates a second order bias in the CPI. Three other changes besides prices changes create first order effects on a correctly measured cost of living index (COLI). (1) Introduction of new goods creates a first order effect of new good bias' (2) Quality changes in existing goods will lead to quality' bias, which has first order effects (3) Shifts in shopping patterns to lower priced stores can create first order outlet bias'. I explain in this paper that a pure price' based approach of surveying prices to estimate a COLI cannot succeed in solving the 3 problems of first order bias. Neither the BLS nor the recent report C. Schultze and C. Mackie, eds., At What Price (AWP, 2002), recognizes that to solve these problems, which have been long known, both quantity and price data are necessary. I discuss economic and econometric approaches to measuring the first order bias effects as well as the availability of scanner data that would permit implementation of the techniques. Lastly, I review recent research that demonstrates that these sources of bias are large in relation to measured inflation in the CPI.