Working Paper: NBER ID: w21698
Authors: Francisco Barillas; Jay Shanken
Abstract: A common approach to comparing asset pricing models with traded factors involves a competition between models in pricing test-asset returns. We find that such practice, while seemingly reasonable, cannot be relied on to determine which is the superior model for several widely accepted criteria including statistical likelihood, Sharpe ratios and a modified HJ distance. All that matters for model comparison is the extent to which each model is able to price the factors in the other model. Given this information, test assets are actually irrelevant, whether the models are nested or non-nested.
Keywords: No keywords provided
JEL Codes: G11; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| pricing performance of test assets (G19) | superiority of a model (C52) |
| ability of each model to price the factors in the other model (C59) | superiority of a model (C52) |
| test asset performance (C52) | identifying the better model (C52) |
| likelihood ratio for model comparison (C52) | test asset returns (G12) |
| nested model performance (C52) | excluded factors pricing (D49) |