Sampling Frequency and the Comparison between Conformable Matched Model and Hedonic Regression Price Indexes

Matched Model price indexes are generally thought to over-estimate the quality-adjusted price level. This bias stems from the fact that only a fraction of the models are sold in consecutive sampling periods and that the price/performance ratio of these models is worse than that of new models. The "unrepresentativeness" of the sample used to construct such indexes could potentially be reduced by obtaining higher frequency data, thus increasing the fraction of models that are matched. We propose a set of matched model price indexes that is conformable with regression based price indexes in order to test this hypothesis. Using computer prices from the Buy Direct press we find that the bias in the matched model price index increases with the sampling frequency. This result suggests that models that last for only a brief time are models for which the price/performance ratio has deteriorated very rapidly. Thus, increasing the sampling frequency actually increases the selection bias