Introduction


Empirical researchers have long sought to determine the most accurate model for explaining differences in expected asset returns. Model comparison matrices include alpha based tests, the most widely used on is the GRS test. Recently, Barillas and Shanken (2017) proposed the use of  squared Sharpe ratio as a measure of model. Models that accurately price factors from other models are considered to have less mispricing and therefore exhibit higher squared Sharpe ratios, indicating their superior performance in capturing risk-return dynamics. By focusing on the “excluded-factor” alphas, which indicate the extent to which each model prices factors from other models, the comparison becomes driven by the models’ ability to accurately capture pricing information from various sources. Interestingly, the test assets themselves become irrelevant in the model comparison process when using the squared Sharpe ratio improvement metric. Instead, the analysis centers on how well each model prices the factors from other models. This highlights the crucial role of factor pricing in determining the relative performance of different models and reinforces the notion that the ultimate preference lies with the model that permits the highest squared Sharpe ratio to be achieved.

Our team has developed a Stata program called asgrs which consolidates the most popular tests for asset pricing models. The program is user-friendly and only requires a basic understanding of Stata. The built-in tests within the program include:

  1. Gibbons, Ross and Shanken (1989) GRS test
  2. GRS p-value
  3. Mean alpha of the test assets
  4. Mean absolute alpha of test assets
  5. Mean adjusted R²
  6. Mean standard error (SE)
  7. Sharpe ratio (SR) of alphas – test assets
  8. Maximum squared sharp ratio of risk factors
  9. Ratio of absolute mean alpha to average cross-sectional deviations (Fama and French, 2015)
  10. [Mean squared adjusted alpha] / [Mean squared adjusted cross-sectional deviations], See Fama and French (2015), Table 5, page 9

Pricing


  • The asgrs program is available for USD 99 with some example data.
  • asgrs plus nested models: USD 149
  • asgrs plus non-nested models: USD 179
  • asgrs plus nested and non-nested models USD USD 279

You can pay with PayPal or a credit card by clicking on the following button.

For further details, please contact us at:

attaullah.shah@imsciences.edu.pk
Stata.Professor@gmail.com

 

How to get the program


Once you have paid the fee, just send email to the above address. We shall send the program installation link and the example dataset through email.

 

References

Barillas, F., Kan, R., Robotti, C., & Shanken, J. (2020). Model comparison with Sharpe ratios. Journal of Financial and Quantitative Analysis55(6), 1840-1874.

Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of finance52(1), 57-82.

Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of financial economics33(1), 3-56.

Fama, E. F., & French, K. R. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116(1), 1-22.