Working Paper: CEPR ID: DP7344
Authors: Rachel Campbell; Kathryn Graddy; Jonathan Hamilton
Abstract: This paper proposes an alternative specification for the second stage of the Case-Shiller repeat sales method. This specification is based on serial correlation in the deviations from the mean one-period returns on the underlying individual assets, whereas the original Case-Shiller method assumes that the deviations from mean returns by the underlying individual assets are i.i.d. The methodology proposed in this paper is easy to implement and provides more accurate estimates of the standard errors of returns under serial correlation. The repeat sales methodology is generally used to construct an index of prices or returns for unique, infrequently traded assets such as houses, art, and musical instruments which are likely to be prone to exhibit serial correlation in returns. We demonstrate our methodology on a dataset of art prices and on a dataset of real estate prices from the city of Amsterdam.
Keywords: art; index; real estate; repeat sales
JEL Codes: C13; C29; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| traditional methods assuming iid errors (C51) | biased estimates of variance and mean returns (C51) |
| nonparametric approach (C14) | accuracy of estimating standard errors (C51) |
| nonparametric approach (C14) | more reliable return estimates (C51) |
| non-iid errors (Y50) | non-linear relationships in coefficients on time between sales (C32) |
| accounting for non-iid errors (C20) | more accurate estimations of asset returns (G17) |
| nonparametric method (C14) | significant reduction of errors in coefficient estimates (C20) |
| nonparametric method (C14) | improvements in R-squared values (C29) |