Forensic Value Stock Selections
The following stocks selections using three forensic algorithms from the financial literature rely exclusively on fundamental data and year over year operational performance measures. The combination of all three bankruptcy and financial irregularity algorithms use a total of 22 fundamental financial variables for assessment. The three different forensic algorithms are detailed in the methodology section at the end of this article.
Significant new changes in the trend of performance between positive and negative forensic portfolios are detailed below.
Adverse Scoring Forensic Selections for October:
A sample of 2 stocks is listed below from among the 10 highest adverse scoring stocks screened in the database across all three algorithms. The list of qualifying stocks for October are available exclusively to members of the Value & Momentum Breakout community. Starting this past August 2018 the descending results will be sorted on the Beneish M-Score instead of the Ohlson O-score probability percentage. The magnitude of the Beneish M-Score may also be a factor in future returns that will be tested as well.
Charts of the 2 sampled stocks with high adverse scoring across all three forensic algorithms are listed below:
Performance for each of the Adverse Scoring Forensic Portfolios to Date:
Over the past one year buy/hold period from the Forensic Negative Portfolios (July - Sep) the best returns of well over 100% were from: XOMA Corporation (XOMA), Immunomedics (IMMU), Nutanix (NTNX) and ViewRay (VRAY). However a much higher percentage of stocks in the adverse portfolios were negative compared to the positive forensic portfolios that have had no stocks negative for the year among the first three portfolios.
My working theory is that stocks that achieve high adverse scores on all three forensic algorithms may be much more dependent on momentum and investor sentiment for price performance than on fundamental financial performance. This dependency on weaker fundamental financial data makes the firms more susceptible to larger price fluctuations and potentially lower annual returns.
Analysis of the first 2 Forensic Negative Portfolios (July / August)
An analysis of the 17 stocks from both the July and August completed forensic negative portfolios provides some fascinating results:
An analysis of the stock composition from both the July and August 2017 portfolios after one year shows some unique results relative to the Positive Forensic portfolio returns detailed below. These negative forensic selections had a large standard deviation with 2 stocks gaining over 200% and one losing over 90%. The average return of the negative portfolio stocks was +60.46% with 5 of 17 (29.4%) of the stocks declining over the year.
Analysis of the latest Forensic Negative Portfolio (September)
The September Negative Forensic portfolio results are the most recent and analysis of these nine stocks produced another large sample standard deviation consistent with prior negative forensic portfolios:
Positive Scoring Forensic Selections for October:
Only six stocks qualified as the highest positive scoring stocks screened in the database with positive scores across all three forensic algorithms. A sample of two of the highest positive scoring stocks are listed below. This is a measure of low probabilities of bankruptcy or earnings manipulation that are being tested to find correlations to price returns over a one-year buy/hold.
To further enhance the test of the published evidence, the Forensic Portfolio selection has been sorted on the Beneish M-Score starting in August.
Corbus Pharmaceuticals (CRBP)
Unum Therapeutics (UMRX)
According to research conducted by Beneish, Lee, & Nichols (2013),
"[The evidence] indicates that [the Beneish] M-score has significant ability to predict one-year-ahead cross-sectional returns. Our results show that this predictive power does not come from its correlation with value, momentum, size, accruals, or short interest" (p. 65).
The Forensic Portfolios are generated for subscribers every other month for 6 annual buy/hold portfolios each year. The August 2018 portfolio is the first to sort on the Beneish M-score values. Results so far sorting on the Ohlson O-score probability percentage in this informal forensic study and using the most positive scoring stocks across the Beneish M-score, Altman Z-score and the Olson O-score are listed in the portfolios below:
Performance for each of the Positive Scoring Forensic Portfolios to Date:
The positive forensic portfolios show a much more stable and consistent return performance with every full one-year portfolio outperforming the S&P 500 in their respective time periods. October appears to be well on its way to another strong annual result. The analyzed results for July and August portfolios are shown below, followed by the latest September Positive Forensic portfolio analysis.
Over the past one year buy/hold period from the Forensic Positive Portfolios (July - Sep) the best returns over 50% were from: R1 RCM Inc (RCM), Friedman Industries (FRD), Pure Cycle Corporation (PCYO), and CRISPR Therapeutics AG (CRSP).
Analysis of the first 2 Forensic Positive Portfolios (July / August)
The positive forensic selections from the 2017 July and August portfolios over a one year period had a much smaller standard deviation with 1 stock gaining over 100% and none of the stocks in the two portfolios declining. In fact, the worst performing stock from the positive portfolios was up 11% for the year. The average return of the positive portfolio stocks was +47.9% compared to the average negative portfolio return of +60.46%. This price performance behavior across the monthly portfolios supports my working theory that strong positive fundamental characteristics and undervalued stocks become more attractive to investors during higher periods of risk and negative momentum.
Analysis of the latest Forensic Positive Portfolio (September)
The positive forensic selections from the 2017 September portfolio over a one year period also had a lower standard deviation than the negative portfolios and generated much better returns. One stock gained over 115% and none of the stocks in either the September positive portfolio or the August and July portfolios were negative. In fact, the worst performing stock from the positive portfolios was up 7.86% for the year.
These two forensic categories of positive and negative portfolios that rely upon advanced fundamental analysis to assess bankruptcy risk and financial irregularities may be showing how strong positive forensic scores produce more reliable positive future outcomes. While the negative scoring stocks with higher potential risks of bankruptcy or earnings manipulation may deliver higher returns as compensation to investors for the risks they may be undertaking. The true explanation is not known, but as these portfolios are tested throughout the coming year from each monthly formation period, we may increase the reliability of profitable selections or generate higher risk/return opportunity for higher gains based on the nature of these algorithms.
Ultimately, I intend to add the best algorithms as additional enhancements to my momentum breakout evaluation process to further develop larger long term results.
The purpose of this monthly value selection list is to provide investors with additional tools to evaluate financial irregularities according to three different detection models from academic research. Circumstances surrounding firms are always subject to change, open to extenuating circumstances, and models by their very nature always contain a degree of error.
I learned of these financial detection algorithms and others, during my doctoral research on financial anomalies as well as through texts on financial statement irregularities in my certification programs. Again, it is important to stress that firms identified by these academic models may not be in actual distress or suffer from any adverse irregularities whatsoever. These models are certainly not foolproof and were designed by academic researchers to improve the chance of detection of irregularities leading to bankruptcy, earnings manipulation, or flag the presence of financial distress.
At the same time, these models are among the best peer-reviewed forensic models in the financial literature and have some significant documented value. The Beneish model for example has "correctly identified, in advance of public disclosure, a large majority (71%) of the most famous accounting fraud cases that surfaced after the model’s estimation period" (Beneish, Lee, & Nichols, 2013, p. 57). Further, in a survey of 169 chief financial officers of public companies, Dichev, Graham, and Rajgopal (2012) reported that respondents estimated that approximately 20% of all companies manage earnings to misrepresent economic performance. While three different financial forensic models are applied in the selection of these portfolios, researchers associated with testing the M-score described their approach this way:
Our main hypothesis was that companies that share traits with past earnings manipulators (i.e., those that “look like manipulators”) represent a particularly vulnerable type of growth stock. Because of their strong recent growth trajectory, these companies are likely to be more richly priced. At the same time, they exhibit a number of potentially problematic characteristics, indicative of either lower earnings quality or a more challenging economic environment. Although the accounting games such companies engage in might not be serious enough to warrant legal action, we posited that their earnings trajectory is more likely to disappoint investors (i.e., they have lower earnings quality)(Beneish, Lee, & Nichols, 2013, p. 57).
To my knowledge no similar longitudinal study of positive and adverse forensic scoring using all three models simultaneously has ever been conducted before. The significant benefits of these portfolios are already emerging in two short months. It is also important to constructively consider why such anomalies may exist in these stock selection at this moment in time. The resulting data which varies from month to month may prompt firms and investors to consider further due diligence of publicly available financial characteristics to mitigate any risk or error present in the marketplace.
So far, five of the six portfolios (3 negative / 3 positive) that have completed a one-year buy/hold period from their original date of formation have all greatly exceeded the benchmark S&P 500 one-year returns. Some portfolios by more than double. Though the sample sizes are small, I do think we are beginning to see the emergence of statistically significant performance results. The three positive forensic portfolios (July / Aug / Sep) each had 100% of the stocks gain over 7% and as much as 131% in the year long period. More than 30% of the stocks in the three negative forensic portfolios declined (and one stock by more than 90%). The positive forensic portfolios are now averaging 52.83% and have overtaken the negative forensic portfolios average 1-year return of 45.48%.
Prior tests in the literature of the Beneish M-score have shown the algorithm to generate excellent results on an annual basis for positive scores. What we may be seeing now are the early results that safer fundamental scores perform well, but riskier adverse scores have the potential to generate much higher returns as much as 25% higher on average than the positive scoring stocks. The tests continue and more explanations may develop over time.
I trust these selections will be a significantly positive addition to your investment goals and returns in 2018!
JD Henning, PhD, MBA, CFE, CAMS
Altman, E. I. (1968). The Prediction of Corporate Bankruptcy: A Discriminant Analysis. The Journal of Finance, 23(1), 193–194. doi:10.1111/j.1540-6261.1968.tb03007.x
Beneish, M. D. (1999). The Detection of Earnings Manipulation. Financial Analysts Journal, 55(5), 24–36. doi:10.2469/faj.v55.n5.2296
Beneish, M. D., Lee, C. M. C., and Nichols, D. C. (2013). Earnings Manipulation and Expected Returns. Financial Analysts Journal, 69.2, 57-82.
Ohlson, J. A. (1980). Financial Ratios and the Probabilistic Prediction of Bankruptcy. Journal of Accounting Research, 18(1), 109. doi:10.2307/2490395
Disclosure: I am/we are long MEIP.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.