Background

A common goal for many investors is to beat the market. However, only a few are able to do so consistently over a long time. The random walk theory and the efficient market hypothesis are two widely accepted theories that state that it should not be possible to consistently generate abnormal returns in an efficient market. There are though some contradicting results that argue against market efficiency and a lot of those studies have value investment in common. Joseph Piotroski was in 2000 able to generate a value investment model that consistently beat the market between the years 1976-1996.

Purpose

The purpose of this paper is to test Piotroski’s model on stock markets with different size and maturity to evaluate if the model, as an investment strategy, can generate a better risk adjusted rate of return than a comparable market index. Unlike recent studies done on Piotroski’s value investing model, we will add a number of additional comparison portfolios and use two different valuation models to determine the source of return variation.

Method

This thesis employed a quantitative research method with a deductive approach. With data from four markets with different characteristics regarding efficiency and development, we performed an ex-ante test from 1995 to 2009. By employing Piotroski’s model, each stock on the four markets was given a score from 0-9; a portfolio for each market was created by the stocks that received a score of 8-9. They were then compared with portfolios from the same market based on the small firm- and book-to-market anomaly. We also performed a test between the markets to see if Piotroski’s model worked better in low efficiency or developed countries. All portfolios in this thesis were risk-adjusted with two different models, CAPM and the Fama & French three-factor model. Since these models use various factors to risk-adjust we have tested if they generate a different valuation of the same portfolio.

Results

Our study has shown that Piotroski’s model is not able to generate significant abnormal returns compared to our portfolios based on anomalies, our results also give an indication that by removing the anomaly premium the model might be destroying value instead of creating it. An explanation to why the model works in Piotroski’s study and not in ours could be the different method of risk adjustment. Piotroski uses a simple method by deducting the market return while we use two models that are taking additional factors into account. Our results are also able to show that choice of the valuation model does have a significant effect on the risk-adjusted return and could therefore affect the end-results of a study. Last of all our results do not give any support for the hypothesis that Piotroski’s model works better in countries with low efficiency compared to high efficiency or in countries that are developed compared to emerging.