Stock Mispricing: Empirical Investigation, Development of Mispricing Measure and Trading Strategy

Petar Stiliyanov Petrov

Student thesis: Master thesis


The thesis investigates stock mispricing and its connection with behavioral finance concepts. Various ways of measuring stock mispricing have been used in the literature, none of which is definitive and commonly accepted. Some of the measures even include accounting variables, effectively providing a noisy measure of mispricing. Hence, we apply the framework in Shiller (1980) to construct a new measure of ex-post mispricing (EPM). We adapt it to a real-time mispricing measure (RTM) by introducing new methodology of computation. Then both measures are tested if they move together with investor sentiment proxied by Baker and Wurgler (2006)’s index. The concept behind the link is based on studies that have shown that behavioral biases such as overconfidence sometimes cause divergence of opinion between market participants about the future dividend growth of companies. This extreme sentiment increases arbitrage risk and drives prices away from fundamental values. We have found that both the ex-post and the real-time mispricing measure for our market proxy have statistically significant positive correlation with the investor sentiment index. For the whole period between July 1965 and September 2015 the former had a correlation of 12.5% while the latter had almost 46%. This supports the idea of a connection between the concepts of mispricing and sentiment. EPM and RTM also show strong positive correlation between each other and with the Cyclically Adjusted Price-Earnings Ratio of the S&P 500. The development of the real-time mispricing measure solves some issue with EPM and enables us to test for return predictability. We construct a set of trading strategies based on trading the most overpriced and underpriced stocks according to their RTM measure. Results indicate outperformance of the undervalued companies over the aggregate market and underperformance of the overvalued. The underpricing portfolio provides better excess market-risk adjusted returns which are between 2.5% and 6.4% annually. However, after controlling for size and value premium based on the Fama-French three-factor model, abnormal returns disappear or become insignificant. This is because the underpricing portfolio is tilted toward small and value stocks while the overvalued portfolio on average includes more growth stocks.

EducationsMSc in Finance and Investments, (Graduate Programme) Final Thesis
Publication date2017
Number of pages67