The authors aim to identify the misvaluation of stocks as a result of behavioral weaknesses shown by traders. They propose a model, based on an extension of the market model, that finds cases in which the asset valuation differs from fundamentals because of trader biases. Their model also allows the magnitude of these misvaluations to be measured. A trading strategy based on the model is proposed and backtested. The authors conclude that following the strategy can lead to superior returns; however, because the model does not predict turning points in the valuation of assets, it is more suited to a long-term strategy.

The proposed model is a tractable model that practitioners can use as an input for the investment decision-making processes. The model helps identify situations in which stocks are misvalued because of cognitive biases exhibited by traders. It is an extension of the market model, augmented by a composite error term, that captures systematic errors in asset valuation.

The results obtained from applying the model suggest that it can identify and measure the magnitude of misvaluation in stocks. In addition, the authors suggest a practical trading strategy that can be applied using the model. Results from back testing the strategy show that there is potential for earning above-average returns even after adjustments are made for risk and transaction costs.

Practitioners will be able to leverage the model for identifying stock misvaluation in their own analysis. The model can also estimate the magnitude of the misvaluation, which allows stocks to be compared.

The market model describes an individual stock’s return as a function of the market return, sensitivity to the market return (beta), unsystematic return (alpha), and a single error term to account for the “white noise” element of stock returns. Under the assumptions of the market model, this error term has an expected mean of zero.

The model proposed by the authors extends the market model to allow for systematic errors caused by behavioral biases exhibited by traders. This extension is achieved by including a dual error term: one traditional error term with an expected mean of zero and one term that is driven by behavioral bias. Using a maximum likelihood method, the authors estimate the coefficient of the error term, by which undervaluation or overvaluation can be identified.

Once defined, the authors apply the model to Toyota Motor Corporation stock, which is stock that is known to have been misvalued and subsequently corrected. Between 2002 and early 2010, Toyota grew to become the largest automobile manufacturer in the world, leading to a pronounced rise in its stock price. But after the accelerator pedal recall in 2010, the stock price suffered. The authors confirm that their model could identify the initial misvaluation followed by a return to the fundamental valuation. There is evidence of overvaluation leading up to the product recall; after the recall, there is a significant correction, after which the model shows no significant bias. The coefficients of the error term give an estimate of the magnitude of the bias.

Using the companies in the Dow Jones Industrial Average between 2005 and 2010, the authors further test the robustness of the model by back testing a trading strategy based on the model. The strategy is to equally invest in all of the companies that are identified as significantly undervalued on an annual basis. The authors find that even after adjusting for transaction costs and risk, the strategy can earn returns superior to a passive investment in the market over the same period.

The authors’ model provides a method for traders to identify stocks that deviate from their fundamental valuations, leading to an opportunity for superior returns. The model itself is tractable and practical to apply, which increases the likelihood of its use by practitioners. As noted by the authors, the limitation of such a trading strategy is that no model can predict a market turning point.