Working Paper: CEPR ID: DP7943
Authors: Francisco Pearanda; Enrique Sentana
Abstract: Two main approaches are commonly used to empirically evaluate linear factor pricing models: regression and SDF methods, with centred and uncentred versions of the latter. We show that unlike standard two-step or iterated GMM procedures, single-step estimators such as continuously updated GMM yield numerically identical values for prices of risk, pricing errors, Jensen's alphas and overidentifying restrictions tests irrespective of the model validity. Therefore, there is arguably a single approach regardless of the factors being traded or not, or the use of excess or gross returns. We illustrate our results by revisiting Lustig and Verdelhan's (2007) empirical analysis of currency returns.
Keywords: CUGMM; Factor Pricing Models; Forward Premium Puzzle; Generalised Empirical Likelihood; Stochastic Discount Factor
JEL Codes: C12; C13; G11; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Single-step GMM methods (C20) | identical estimates of pricing errors (C13) |
| Single-step GMM methods (C20) | identical estimates of prices of risk (C13) |
| Single-step GMM methods (C20) | identical estimates of overidentifying restrictions tests (C20) |
| CUGMM (C68) | identical estimates of pricing errors (C13) |
| CUGMM (C68) | identical estimates of prices of risk (C13) |
| CUGMM (C68) | identical estimates of overidentifying restrictions tests (C20) |
| Regression method and SDF method (C29) | identical estimates (C13) |
| Empirical findings from Lustig and Verdelhan (2007) (G15) | cautious interpretation (Y20) |
| Different GMM methodologies (C51) | significant discrepancies between models (C52) |
| Causal relationships (C22) | robust to various model specifications (C51) |