In this article, I will briefly present the Piotroski F-Score and show the output from screening the Russell 3000. Joseph Piotroski received a B.S. in accounting from my alma mater, the University of Illinois in 1989. He worked as a senior tax associate at the firm Coopers & Lybrand before receiving an MBA from Indiana University in 1994, and a Ph.D. in accounting from the University of Michigan in 1999. He taught at the University of Chicago and is currently at Stanford. Piotroski is known for a paper he wrote in 2000 entitled “Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers.”
In the paper, Piotroski presented a model which outperformed the market. He calculated a score with a maximum of 9 points based on 9 questions he asked from each company. The highest returns came from companies that were among the cheapest 20% of stocks based on Price to book and that had a score of 8 or 9. This model returned 23% over the backtest period. The nine questions are:
- Is net income positive? (1 point if positive).
- Is operating cash flow positive? (1 point if positive).
- Net income growth > Total asset growth (1 point if true)
- Operating cash flow > Net income (1pt if true).
- Total asset growth > Total liabilities growth (1pt if true)
- Current ratio is higher this year vs last year. (1pt if true).
- Shares outstanding decrease (1pt if true).
- Gross margin is higher this year vs last year. (1pt if true).
- Sales growth > Asset growth (1pt if true).
To see the entire paper please go to the FatAlpha website, click on Articles and scroll down to “Priceless Academic Papers”. On February 5th, I ran a screen of the cheapest 20% in the Russell 3000 based on price to book and which had a Piotroski F-Score of 8 or 9. The result from this run are the following stocks: ACCO, BCOR, CHK, CPE, LRN, NFX, OC, OMG, PEIX, TEX, TRK.
ACCO is actually a holding in my portfolio.
Please note once again that the results from this or any other screens are not recommendations and I have not checked the integrity of the data or the inputs. This page is meant to provide some food for thought.