eThesis - School of Business electronic theses

I this paper I study the effectiveness of financial statement data based stock picking on the Finnish stock market. More specifically I construct portfolios based on F-Score, aggregate of seven individual signals, mean to capture both current state and trend in profitability, operating efficiency, leverage, liquidity and source of funds. Main research question of the study is whether financially sound companies, as defined by having high F-Score, would yield higher risk-adjusted returns. I also combine the scores individually to company size and book-to-market valuation to determine if the F-Score selection works when taking into account well-known value and size effects. I have studied the performance with two holding periods, three and 12 months. My data set consists of eight different income statement and balance sheet based metrics, annual and quarterly share return data for 112 companies listed in the Finnish stock exchange between 1.1.2004 and 31.12.2015.

A literature review on theoretical asset pricing is carried out, starting from basic concepts of market efficiency and Capital Asset Pricing Model (CAPM), moving to main CAPM anomalies and more sophisticated multifactor models. In the end I present Joseph Piotroski's F-Score, originally developed to separate undervalued companies form financially unhealthy ones in order to improve performance of traditional value strategies.

My results show that portfolios formed from companies that rank high on their F-Scores show higher risk-adjusted returns than portfolios of low F-Scores, for the entire data set and individually when controlling for size and valuation and for both three and 12 months holding periods. However the results are mostly not statistically significant. F-Score based portfolio selection worked best when used to small sized companies.