Careful readers of the conceptual framework in Benjamin Graham and David Dodd’s classic volume Security Analysis (1934) would argue that a stock’sreturn is ultimately dependent on fundamental traits of the business, such as its ability to generate earnings off its assets, and the extent to which the business can efficiently invest the earnings retained by the company.
With this motivation in mind, we recently conducted an empirical investigation into the fundamental factors that separate winning stocks from losing ones, as measured by absolute market returns. By analyzing data for years 1980-2010, we were able to assess the success of publicly traded corporations in the US, and the factors leading to the companies’ ability to create value for shareholders, at least for the last 30 years. Our thesis was that stock performance could be measured through indicators and predictors of value creation, or key business metrics which we refer to as “genes” (analogous to trait-determining genes for organisms in the field of genetics).
Methodology
The selected sample is composed of 70 corporations with 7 firms in each of 10 industry groups. Recognizing that this group is not completely homogeneous (due to industry and size differences), I attempted to make a careful selection of the companies. Ideally, this careful selection allows for a high-quality examination of ratios in time period T to forecast parameters at time T+1 to make predictions about the following annual period (T+1). This 1-year period is not only chosen to simplify the calculations. But also to allow a long-enough time horizon for the chosen fundamental traits (genes) to be relevant.
After the initial industry groups are defined and firms selected, balance sheet and income statement data are collected. Because of the large number of variables found to be significant indicators of corporate problems in past studies, a list of 38 potentially helpful variables and ratios were compiled for evaluation. The variables span several standard ratio categories, including liquidity, profitability, leverage, and solvency. The ratios are chosen on the basis of their popularity in the literature, their potential relevancy to the study, and conversations with experts on the field. Additionally, there are a few “new” ratios in this analysis which we consider relevant traits of business performance. Multicollinearity analysis then reduces the number of variables from 38 to 23.
Results
The analysis indicates that the expected return of a stock at time T+1 is statistically dependent on two key factors: valuation and profitability. The two factors are represented by the proper ratios, which remain unnamed. These factors have the ability to estimate what the company’s expected return (Er) should be. The results indicate that the greater the predicted return based on fundamental factor, the greater the Sharpe ratio.
Additionally, the greater Er the greater the probability of picking a winner.
Back-testing results from the subgroup Er > 25 shows that the model does incredibly well at picking winners: stocks expected to rise significantly by the following year. On average, the model obtained an average return of 29% from years 2000-2009 with no negative years, while the S&P500 returned an average of 2% with 4 negative years.
Interestingly, the model delivers most of its outperformance relative to the S&P500 (alpha) during times of greatest economic distress such as years 2002 and 2008. This, along with cumulative returns, is shown below.
Although the model does not account for company market capitalization, the identification of winning stocks based on the developed Er parameter is independent across a wide market capitalization range. As the results shows, the more negative this parameter is, the more likely it is to underperform the S&P500 benchmark. On the contrary, the more positive the more the absolute returns. Therefore, the model is robust in forecasting stock over-performance regardless of company size or holding period.
Conclusion
Our research shows, through statistical analysis, that a winning stock’s “genetic” characteristics are remarkably simple: bargain prices combined with unusually high profitability. Stocks with factors representing these characteristics tend to outperform their peers in the long-term and produce high returns even during bear markets.
Through back-testing analysis, we have confirmed that stocks with expected “genetic” returns greater than 25 have outperformed the S&P500 benchmark by over 27% over the last 10 years, with no negative year. While 72% of stocks with an expected return parameter of less than -50 lose money the following year, only 14% of those with a parameter greater than 50 do. As shown, the model also serves to increase the likelihood of picking “winners” over “losers”.
Value Huntr
http://valuehuntr.com/
With this motivation in mind, we recently conducted an empirical investigation into the fundamental factors that separate winning stocks from losing ones, as measured by absolute market returns. By analyzing data for years 1980-2010, we were able to assess the success of publicly traded corporations in the US, and the factors leading to the companies’ ability to create value for shareholders, at least for the last 30 years. Our thesis was that stock performance could be measured through indicators and predictors of value creation, or key business metrics which we refer to as “genes” (analogous to trait-determining genes for organisms in the field of genetics).
Methodology
The selected sample is composed of 70 corporations with 7 firms in each of 10 industry groups. Recognizing that this group is not completely homogeneous (due to industry and size differences), I attempted to make a careful selection of the companies. Ideally, this careful selection allows for a high-quality examination of ratios in time period T to forecast parameters at time T+1 to make predictions about the following annual period (T+1). This 1-year period is not only chosen to simplify the calculations. But also to allow a long-enough time horizon for the chosen fundamental traits (genes) to be relevant.
After the initial industry groups are defined and firms selected, balance sheet and income statement data are collected. Because of the large number of variables found to be significant indicators of corporate problems in past studies, a list of 38 potentially helpful variables and ratios were compiled for evaluation. The variables span several standard ratio categories, including liquidity, profitability, leverage, and solvency. The ratios are chosen on the basis of their popularity in the literature, their potential relevancy to the study, and conversations with experts on the field. Additionally, there are a few “new” ratios in this analysis which we consider relevant traits of business performance. Multicollinearity analysis then reduces the number of variables from 38 to 23.
Results
The analysis indicates that the expected return of a stock at time T+1 is statistically dependent on two key factors: valuation and profitability. The two factors are represented by the proper ratios, which remain unnamed. These factors have the ability to estimate what the company’s expected return (Er) should be. The results indicate that the greater the predicted return based on fundamental factor, the greater the Sharpe ratio.
Additionally, the greater Er the greater the probability of picking a winner.
Back-testing results from the subgroup Er > 25 shows that the model does incredibly well at picking winners: stocks expected to rise significantly by the following year. On average, the model obtained an average return of 29% from years 2000-2009 with no negative years, while the S&P500 returned an average of 2% with 4 negative years.
Interestingly, the model delivers most of its outperformance relative to the S&P500 (alpha) during times of greatest economic distress such as years 2002 and 2008. This, along with cumulative returns, is shown below.
Although the model does not account for company market capitalization, the identification of winning stocks based on the developed Er parameter is independent across a wide market capitalization range. As the results shows, the more negative this parameter is, the more likely it is to underperform the S&P500 benchmark. On the contrary, the more positive the more the absolute returns. Therefore, the model is robust in forecasting stock over-performance regardless of company size or holding period.
Conclusion
Our research shows, through statistical analysis, that a winning stock’s “genetic” characteristics are remarkably simple: bargain prices combined with unusually high profitability. Stocks with factors representing these characteristics tend to outperform their peers in the long-term and produce high returns even during bear markets.
Through back-testing analysis, we have confirmed that stocks with expected “genetic” returns greater than 25 have outperformed the S&P500 benchmark by over 27% over the last 10 years, with no negative year. While 72% of stocks with an expected return parameter of less than -50 lose money the following year, only 14% of those with a parameter greater than 50 do. As shown, the model also serves to increase the likelihood of picking “winners” over “losers”.
Value Huntr
http://valuehuntr.com/
