Article Detail

Are Investor Sentiments Priced by the CAPM?

Journal 35: Zicklin-Capco Institute Paper Series in Applied Finance

Rahul Verma, Gökçe Soydemir

In this paper, we investigate whether the international version of CAPM can price rational and irrational sentiments of U.S. individual and institutional investor sentiments. The results show that the CAPM prices rational sentiments driven by fundamentals and irrational sentiments not driven by rational risk factors in the case of the U.S. institutional investors. We check our results by additionally estimating a non-equilibrium-based model, such as a fundamental market model, and find similar results.

This indicates that irrational sentiments of institutional investors are significantly positively related to returns predicted, as well as those not predicted by the CAPM. We also compare these findings with the response of actual DJIA and S&P 500 returns to individual and institutional investor sentiments and find that CAPM does a partial job by capturing only institutional investor sentiments. These results are consistent with the view that the studies that use an equilibrium-based asset-pricing model such as CAPM need to factor in institutional investors’ cognitive factors to better capture the return generating processes.

There is now general agreement among academics and practitioners that cognitive factors such as investor sentiments play an important role in determining stock prices. However, studies in the past have merely conjectured that such sentiments have this role. It is yet to be known empirically whether asset-pricing models such as the Capital Asset Pricing Model (CAPM) [Sharpe and Lintner (1964)] are able to capture such sentiments when pricing equities. Previous studies that have accepted the role played by cognitive factors mainly choose to adopt a non-equilibrium-based model to study these effects. Examples used by studies in the past constitute either simple or fundamental market models, or merely non-structural econometric models, such as multiple regressions, GARCH or vector autoregression models. Other studies that examined such behavioral aspects simply assume that CAPM could be directly applied without testing whether or not it actually constitutes a viable equilibrium framework to be used in their study.

We follow two separate approaches to test whether these rational and so-called “noise” components of investors’ cognitive behaviors are captured by CAPM. In the first approach we obtain the predicted and residual values of DJIA and S&P 500 returns by estimating an international version of CAPM with a world market return/risk premium on the right side of the equation. We then test if these predicted and residual values from the CAPM estimate are significantly related to rational and irrational components of individual and institutional investor sentiments. In the second approach, we separately capture the portion of DJIA and S&P 500 returns which could be explained by the irrational and rational components of investor sentiments. Thereafter, we test if these returns driven by irrational and rational sentiments are significantly explained by CAPM. Specifically, we twice examine the magnitude and significance of the coefficients (world beta) of the CAPM equations, which include DJIA and S&P 500 returns induced by irrational and rational sentiments (left side of the equation) and world returns (right side of the equation).

In both these approaches we split the overall investor sentiment variable into rational and irrational components by regressing them to a set of risk factors and calculating both predicted and residual values. The predicted values represent investor sentiments induced by rational risk factors, while the unexplained portions represent those components of sentiment which are totally irrational in nature.


Leave a comment

Comments are moderated and will be posted if they are on-topic and not abusive. For more information, please see our Comments FAQ
This question is for testing whether you are a human visitor and to prevent automated spam submissions.