ZPR has constructed several proprietary quantitative models over its many years of research.
Earnings Quality and True Profitability (EQTP) Model
The Earnings Quality and True Profitability Model (EQTP) was created by ZPR's quantitative research department as a way to separate the good from the bad in corporate earnings reports, a critical determinant of stock prices.
Many investors and Wall Street analysts simply use the earnings as given by the companies, an error which can lead to a completely false understanding of a company's true economic state. This works in both directions, as at any given time some companies attempt to raise their earnings in the short term to meet analyst expectations, while others attempt to delay earnings to the future in order to show consistent growth.
ZPR's research is built on a seminal paper by Professor Richard Sloan (Sloan, Richard G., 1996, Do Stock prices Fully reflect Information in Accruals and cash Flows About Future earnings?, Accounting Review 71, 289-315) which demonstrated that all earnings are not created equal, and that companies whose earnings have a high accrual component subsequently have lower stock market returns. We have taken this research to a higher level by separating the accrual analysis into two components, Earnings Quality and True Profitability. The two concepts are related but distinct: Earnings Quality examines the portion of accruals in a company's earnings (using ZPR's proprietary formula), while True Profitability strips away the accruals and re-evaluates a company's profitability on a non-accrual basis.
Each quarter ZPR rank orders all of the companies in a given universe of stocks (the S&P 500, for example) by quintile (1-5), with quintile 1 representing the best companies according to the model and quintile 5 representing the worst.
Byproducts of ZPR’S proprietary database to produce this model are an excellent cash flow to market cap model that outperforms the Holt measures, an accruals ratio model, a cash flow ratio model, and a reported profitability model. These are all insightful looks at how the company is keeping its books.
EQTP Improvement Model
In 2005 ZPR made a further enhancement to EQTP designed for its investment management clients. Instead of investing in the entire top quintile of stocks, we created a strategy which uses only the new qunitile 1 stocks (those having a worse rank in the prior quarter). This is the EQTP Improvement Strategy.
Our EQTP Improvement Strategy invests in S&P 500 companies only. New arrivals to EQTP quintile (rank) one are bought each quarter. This investment strategy is only for accounts which can tolerate high turnover, as we have 100% turnover each quarter. Results have been superior and in accordance with our research. From inception on 6/1/08 through June 2011 the annualized return was 6.06% versus the S&P 500 Index return of 0.33%.
This actually is an entire product system for loss avoidance. [More...]
APT Factor Model
Are growth stocks going to outperform value stocks? [More...]
Modified Graham & Dodd Model (MODGD)
''Value'' companies tend to be ''distressed'' companies and cyclicals. This model finds value companies that are not distressed. IC's of risk adjusted returns are spectacular. Despite its low beta, the model's total returns are also very good. It also works well on all universes.
Relative Strength Models
Investors have long been fascinated with relative strength. For almost 20 years, ZPR has had excellent relative strength models that had outperformed all others. Enhancement were made to adjust for mean reversion. This has resulted in even better performance. Our stable of important relative strength variables consists of:
ZPR Momentum Model (ZPRMM)
- Relative Strength 12M (price appreciation only)
- Relative Strength 3M (price appreciation only)
- Relative Strength 11M (total return, price+dividends)
- Standardized Relative Strength 11M
- Beta Adjusted Relative Strength 11M
- Relative Strength Modified (RSTRMOD)
- Relative Strength MAP (RSMAP)
- Relative Strength Potpourri (RSPOT)
At particular times, momentum investing can dominate the equity markets and many renowned managers have made a career from such an investment style. This ZPR model combines relative price strength, earnings revision and earnings surprise for solid performance results on all universes.
MOCON and SuperMo Models
Momentum confirmation or MOCON and SuperMO are our super momentum models. Both are founded on quantitative computer-based models that focus on the upward momentum of a stock’s price and earnings. They track and measure improvements in these and the above mentioned factors. Not surprisingly, these models only work when applied to the under covered stocks found in the micro cap universe.
MOCON model is based on price and earnings momentum as well as estimate revisions and quarterly surprise. We have derived our own proprietary estimate revisions and quarterly surprise models called Prime EREV and QSUR Prime which are very strong by themselves and much more powerful than the basic estimate revisions and quarterly surprise models.
SuperMO is a turbo-charged subset of MOCON. It is a very concentrated portfolio of super momentum stocks. More conditions are added to the model further reducing the number of stocks (normally, no more than 45 stocks and as few as 15 make the list). The model is season dependent and has a double-weight on stocks that meet all criteria – largest estimate revisions, largest quarterly surprise, largest EPS a price momentum. In comparison to MOCON, SuperMO has a much greater upside as well as downside potential because it contains a smaller number of stocks.
ZPR has conducted extensive research of various analyst effects’ impact on stock returns. We measured the performance of analysts’ buy and sell recommendations, we measured how stocks perform without analysts or few analysts, cases were there are analysts covering a company but they give no forecast, only one analyst with a rating (Lone Wolf). We tested analyst overexposure and analyst neglect and also impact of changes in analyst recommendations on stock returns. What happens when a stock gets its first analyst, what if it also gets a strong buy rating? What about performance of companies that lose the only analyst? After conducting extensive research we have the answers to all the questions. All our tests are done separately on each universe because small caps, being less efficient, usually behave very differently from large caps.
For many years, we have known that low volume stocks outperform high volume stocks. And they do that with lower standard deviation. We can define volume as either the 20-day average trading volume relative to shares outstanding or the 3-month volume as a percent of shares outstanding. The original discovery was made by Professors Lee and Swaminathan in their study "Price Momentum and Trading Volume". Their research documented a life cycle theory for momentum, where stocks with good performance and low volume outperform stocks with poor performance and high volume. The theory is that phases of overexcitement follow good stock performance, and phases of neglect follow poor stock performance. Large institutional investors likely perpetuate this cycle, since their trading desks are typically demanders of liquidity, weighting their buys and sells and therefore their portfolios upon a percentage of recent daily trading volume. This is a major source of underperformance by institutional investors, and a source of outperformance for our strategies.
The Volume Winners strategy has the lowest risk profile of our three volume strategies. In fact, it is designed to minimize risk while still providing market beating returns. It accomplishes this by combining three components: low volume, a price momentum measure, and a proprietary short interest ratio. All of these three components produce excess returns in their own right, and they all do so partly by avoiding potential losers. These stocks generally have improving fundamentals (as they are near their 52 week high) while avoiding investor overexcitement (volume) and are less subject to negative catalysts (lack of short sellers). In our study the combination proves to be excellent at avoiding big declines. Volume Winners strategy provides significant downside protection during some of the worst periods in recent market history.
The Volume Value strategy combines low volume and low valuation (cheaper) stocks. This combination is designed to deliver market beating returns with substantially lower volatility. The addition of the value component improved the performance of volume by itself – not only in greater return for quintile one, but also for identifying losers. In our backtests the strategy succeeds in delivering higher returns while reducing volatility by a meaningful amount.
The Volume Momentum strategy has the highest theoretical return but also the highest risk profile of our three volume based strategies. This strategy is a combination of two of our other strategies: SuperMo, a proprietary momentum model, and the Volume Winners strategy (see following description). This combination is designed to deliver market beating returns with modestly lower volatility.
The SuperMo component is a multifactor momentum screen, comprised of earnings revision, earnings surprise, price and earnings momentum, and momentum confirmation measures. Not surprisingly, this model works best on undercovered, thinly traded stocks found in the micro cap universe.
Anti-Gravity Loss Avoidance
Anti-Gravity Loss Avoidance or ''AG'' is a proprietary stock screening process used to identify companies with the potential to disappoint. It acts to enhance rather than change the core strategy. AG uses earnings expectation dynamics and theories of investor behavior, drawn from academic research, to identify companies that are likely to perform poorly. It works in conjunction with any model or technique already used by the investment manager. This was developed in 1991 from the Antigravity theory of Prof. Zavanelli
APT Factor Models
Are growth stocks going to outperform value stocks? Are small cap stocks going to beat large cap stocks? ZPR can answer these questions with statistical confidence using its APT (Arbitrage Pricing Theory) model. We may be the only firm which has a functioning APT model with real time results. Our model focuses on macro economic forces and risk. We use such factors as bond term structure and credit risk to make accurate predictions of investor behavior. The APT model predicts which variables and methodologies will be the most effective predictors in the next quarter.
ICX Composite Modeling Service
ZPR's primary statistic for measuring investment methodology performance is the IC or ''information coefficient.'' The IC statistic is the correlation between a methodology and its excess return over a quarterly holding period. An optimization system uses historical IC's to create an ICX composite of several methodologies that has lower variability and higher expected excess returns than any investment methodology taken alone. ''January'' models contain a ''January Effect'' component to take advantage of the seasonal behavior of markets.
Integrated Effects System
This is ZPR’s system for adjusting quantitative models for special company events which have been identified as causing abnormal returns. ZPR then researches these anomalies on the ICX 2000 stock database to determine our own specifications and applications.
ZPR has standardized the expected abnormal performance of these special events by creating scoring system called Integrated Effects. This system can help the portfolio manager identify stocks whose performance will be affected by buyback, secondary offering, dividend dynamic and spinoff events in the future. The portfolio manager can also have ZPR incorporate his quantitative model into the Integrated Effects system and calibrate Integrated Effects score adjustments to the model. This offers the portfolio manager the means to determine if any special events will be significant enough to override the basic model’s result.
ZPR's EQTP Model is an excellent predictor by itself as well as when used in a muti-variable model. EQTP successfully picks companies with good or poor earnings quality. We found that companies that move to EQTP rank 1 from other quintiles during quarter (improving earnings quality) have superb performance. This strategy is successfully used by us in practice with excellent results since the third quarter of 2001.
In our constant search for statistical patters within the marketplace, we have discovered the super momentum stocks amongst the smallest of public companies. These are companies with significantly increasing earnings, increasing analyst coverage and forecasts, ratings upgrades, and positive earnings surprises. Not surprisingly, this model works best on under covered, thinly traded stocks found in the micro cap universe. Super Mocon strategy is rebalanced quarterly. Normally, no more than 45 stocks and as few as 15 make the list. This strategy has higher than average standard deviation but delivers spectacular results.
For many years, we have known that low volume stocks outperform high volume stocks. And they do that with lower standard deviation. We have three Volume models and we provide quarterly prediction reports on all three: Volume Winners, Volume Value, and Volume Momentum.
Model Performance Report (IC Reports - Quarterly)
To provide timely information on the performance of investment methodologies ZPR publishes its quarterly IC reports. ZPR’s Quarterly IC reports provide the subscriber with an independent and unbiased evaluation of over 40 vendor and academic methodologies. Subscribers receive their IC reports on the 2nd day of the following quarter. The IC report connects the present with the past, providing historical information as well as current. Methodology performance statistics are provided back to 1986 with special subsections detailing performance in January and non-January quarters to control for seasonality.
All historical performance statistics are calculated using ZPR’s ICX 3000 "Live" research database. All survivor and data availability biases have been eliminated to assure accuracy of results. New methodologies are constantly being added to reflect new research.
The IC Statistic: The primary performance statistic provided to subscribers by the IC report is the IC. IC’s or "information coefficients" are a measure of the correlation between a methodology and its excess return. ZPR uses the IC because it is more stable than other popular statistics such as total and excess return. The IC captures the true performance of the methodology without the volatility and market correlations that tend to cloud the true performance result. IC’s also identify the "wrong way" predictors. While some methodologies exhibit high positive correlations with returns, other methodologies are highly negative predictors. IC’s allow the creation of composite models that take advantage of each methodology’s unique properties.
In model predictions, all variables and companies are sorted and then broken to deciles (rank 1 to rank 10) where rank 1 companies are predicted to do the best and rank 10 companies are predicted to do the worst.
Events Performance Anomalies
We have a number of specific databases that will track companies for the following events:
We then use our own research and that of academic studies to provide performance predictions each quarter.
- Identification (Name and Ticker Change)
- Splits (since 1980)
- True Mergers
- Secondary Offerings
- Dividend Dynamics (Initials, Resumptions, Cuts, Omissions)
- Wall Effect
- Investor Behavior Effect
- Anticipated Surprise (are analysts too optimistic in EPS estimates?)
- Analyst Effect
- Volume Effect
- EQTP Improvement Effect
- SUPERMO Effect
ZPR has many years of experience testing special company events and analyzing specific investment strategies. As a by-product of ZPR's years of testing specific investment strategies, we have produced the Event Analysis Database. Then ZPR diligently determined the validity of the data for investment decision-making.
Because of the synergy between the ZPR ICX 3000 Database and the databases that make up these special effects, we have the ability to analyze the excess return performance on a quarterly and monthly basis. We can also analyze the size adjusted excess return performance on quarterly performance. Furthermore, we are able to further break down these databases into smaller groups and even subgroups that are specific to certain characteristics of the companies. This allows us to see if performance significantly differs for any of these groups and/or subgroups as opposed to the entire database of companies. We determined whether the event for the company is an immediate intersection or lagged (delayed) intersection.
ZPR has many years of experience testing these special events and quantitative strategies. If you have an investment strategy or quantitative model that you would like tested (either for historical information or for current prediction use), ZPR can professionally and thoroughly complete the task. The event analysis database can be acquired separately or as part of the Integrated Effects Data Browser system.