Academic journal article Real Estate Economics

Revisiting the Past and Settling the Score: Index Revision for House Price Derivatives

Academic journal article Real Estate Economics

Revisiting the Past and Settling the Score: Index Revision for House Price Derivatives

Article excerpt

This article examines index revision in measuring the prices for owner-occupied housing. We consider revision in the context of equity insurance and the settlement of futures contracts. The usefulness of aggregate housing price indexes in these contexts requires stability as they are extended. Methods that are subject to substantial revision raise questions about the viability of derivatives markets. We find that the most widely used house price indexes are not equally exposed to volatility in revision. Hedonic indexes appear to be substantially more stable than repeat-sales indexes and are not prone to the systematic downward revision found in the repeat-sales indexes.


Most of the statistical series used to describe the workings of the economy are subject to periodic reexamination and reestimation. Indexes are revised when either the method used to incorporate data or the data themselves are updated. Index revision of the former type is illustrated by the changes in the Consumer Price Index (CPI) that arose from implementing the findings of the Advisory Commission to Study the CPI (the so-called "Boskin Commission"). In this case, the CPI was rebenchmarked through the adoption of new practices based on economic theory. (1) Index revision of the latter type results from the reestimation of the index after the arrival of new information. (2) We focus on index revision of this latter type. Revision from either source has direct implications for public policy and private investment; these indexes are relied upon in policy formulation, investment decisions and economic modeling, and they may form the basis for the development of markets for products such as housing price futures and home equity insurance.

Our interest is the dynamic performance of commonly used methods of housing price index construction: those based on repeat-sale and hedonic models. While there is an extensive literature on the asymptotic characteristics of these indexes, little is known about their stability as they are reestimated with the arrival of new information about housing prices. At issue is not the influence of observations that are added to the data due to informational lags. Rather it is the extent of index revision as additional sales are included in the set of observations used in their construction.

Index stability is often overlooked as a desirable characteristic of price indexes. This is especially relevant for house price indexes, given their wide use. Economic models of mortgage prepayment and default, measures of household wealth and cost-of-living calculations, among many other applications, are all informed by indexes based on the "latest" data. If, in fact, initial estimates of aggregate prices are subject to substantial revision, results from these models and measures may be misleading. Furthermore, the perception of instability in measured prices may preclude the development of markets for financial assets based on housing price indexes.

In the United States the only widely available set of quality-controlled housing price indexes are based on so-called repeat-sale models. We present extensive evidence on the extent to which revision for this class of indexes is "large." Our benchmark for making this assessment is a chained Fisher ideal index derived from a series of cross-sectional hedonic regressions. Repeat-sale indexes rely on strong assumptions regarding the time-invariance of both dwelling characteristics and their implicit prices in order to recover aggregate housing prices from a sample of dwellings that sell two or more times. In contrast, the chained Fisher ideal index is based on a series of cross-sectional regressions that employ all available information about housing sales, while allowing both housing characteristics and their implicit prices to change over time. We compare this benchmark index to two indexes based on repeat-sale models and one index based on a longitudinal hedonic model, where implicit prices of housing characteristics are assumed to remain constant over time. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.