“Data Driven Investing”

The investment strategy employed on behalf of the clients of Cognition Capital is, broadly speaking, based on the “data driven investing” approach described by Bill Matson and Mitch Hardy in their book, Data Driven Investing – Professional Edition.

Data driven investing rests on three key principles:

• All investment decisions must be supported by objectively verifiable information (i.e. data), not driven by intuition, rumor, emotion, or a desire to be fashionable. Therefore, we analyze the available data to identify persistent market inefficiencies, develop decision rules to exploit those inefficiencies, test and refine our decision rules, and then apply them swiftly and surely to new information as it becomes available.
  
  
• Investment strategies based on identified inefficiencies are likely to achieve long-term success only to the extent to which such patterns can be explained as manifestations of the psychological biases (of investors, individually and collectively) inherent in decision-making under conditions of uncertainty, or of perverse incentives acting upon key market participants.
  
  
• Investment performance can be improved over time by identifying and incorporating new strategies for exploiting behaviorally-driven market inefficiencies through research and the analysis of historical market data, and through an investment decision-making process driven by the correlation of new information with previously identified patterns.

Contrary to the efficient market hypothesis espoused by many academicians, the U.S. equities market is rife with inefficiencies arising from the enduring foibles of human nature – cognitive dissonance, illusory and invisible correlations, anchoring, and the endowment effect, to name a few – and from perverse incentives that influence institutions, analysts, accountants, and other market participants and information providers. Skillful, active management can exploit these inefficiencies to consistently produce alpha.

The application of behavioral science - particularly the psychology of decision-making under conditions of uncertainty - to the development of investment strategies resolves the principal weakness of purely quantitative-, technical-, or momentum-driven strategies, which is: the lack of any good way to assess or predict the likelihood that observed patterns will continue into the future. The various forms of bias that enter into investors’ decision-making processes – loss aversion, conservatism, trend, and cognitive biases – are persistent and universal, or nearly so, and, therefore, market inefficiencies arising from these behavioral patterns are also likely to persist and remain widely available for savvy investors to exploit.

Market Inefficiencies Arising from Behavioral Factors

Many of the most enduring market inefficiencies arise when investors:

• allow their investment decisions to be affected by any of a wide range of psychological biases that cause them to enter orders based on emotion, ignorance, or motivations other than “buying low and selling high,” or
  
  
• fail to appreciate trading patterns caused by the predictable behavior brought about by perverse incentives acting upon politicians, corporate managers, brokerage analysts, and other entities whose activities affect stock prices

Some examples of investment decisions arising from emotion, ignorance, or motivations other than “buying low and selling high” are:

• Monday morning panic selling of low relative strength stocks, due to anxiety induced by the inability to trade during the weekend
  
  
• Naïve market orders to buy or sell inappropriately large quantities of thinly traded stocks
  
  
• Tax-motivated selling of low relative strength stocks in late December
  
  
• Institutional selling of spin-off stocks due to arbitrary market cap limitations

Larger scale inefficiencies arise from investors’ failure to fully appreciate patterns of consistently strong performance of certain groups of stocks, for example, small cap value stocks with high relative strength:

• during the last two years of presidential terms (due to politically motivated fiscal stimulus applied by incumbents);
  
  
• when the Federal Reserve Bank is lowering interest rates; and
  
  
• during quarters following positive earnings surprises (due to the asymmetric nature of corporate managers’ and brokerage analysts’ incentives).

Examples of well-documented psychological biases that the investment strategy employed by Cognition Capital is designed to exploit include:

• Anchoring – the tendency of investors’ valuation estimates to be inappropriately influenced by a stock’s price history
  
  
• Endowment Effect – the tendency of investors to be irrationally protective of their gains (which leads them to sell their winners prematurely)
  
  
• Loss Aversion – the tendency of investors to become obsessed with breaking even on losing positions (which leads them to hold their losers too long)
  
  
• Halo Effect – the tendency of investors to extrapolate companies’ strengths or weaknesses in one area into unrelated areas (which leads to overvaluation of exciting, fast-growing companies’ stocks and undervaluation of stodgy or socially stigmatized companies)

Some of these psychological biases are best exploited through quantitative stock screening strategies. For example, the Halo Effect can be exploited by avoiding stocks with exciting, well-publicized stories and high price ratios (price to earnings, price to sales, etc.) and, instead, targeting the stocks of socially stigmatized (e.g. casket manufacturers) or boring (e.g. basic industries) companies with low price ratios.

Most psychological biases, however, tend to cause delayed investor responses to new information and are therefore best exploited by swift, sure reaction to breaking news.

These are just a few of the numerous examples of market inefficiencies arising from behavioral factors that can be profitably exploited.

Investment Philosophy

Broadly speaking, then, our investment philosophy is to select stocks to own (or avoid) according to data-driven strategies that have worked consistently over long periods in the past because they are based on well-established patterns of behavior, and to trade them according to decision rules that enable a reaction to new information that is swift, appropriate, and automatic.

Investment Style

Our investment style emphasizes the following elements:

• A strong “value” orientation
  
  
• An orientation towards high relative strength stocks, except at year-end
  
  
• Very active trading of individual U.S. equities and ADR’s (clearing through a broker with very low commission rates)
  
  
• Use of Exchange-Traded Funds (“ETF’s”) whenever the availability of good opportunities in individual stocks is insufficient
  
  
• A highly diversified portfolio of individual stocks and ETF’s

The Fed-Election-Calendar Cycle and the Yield Curve Slope

The Fed-Election-Calendar Cycle influences various aspects of our strategy, in particular our appetite for risk. Depending on client-imposed guidelines and restrictions, the Cycle may affect our appetite for small cap stocks, use of leverage, and net market exposure.

The yield curve slope affects the availability of good opportunities in individual value stocks.

The Fed-Election Cycle

The Fed-Election Cycle consists of the combined 4-year presidential election cycle and Federal Reserve interest rate policy cycle, as follows:

• Each four-year presidential term is divided into an early phase (i.e. the first two years of the term) and a late phase (the last two years of the term, year four being the election year).
  
  
• The Federal Reserve policy cycle is also divided into two phases: the expansive phase, when the Fed is cutting short-term interest rates, and the restrictive phase, when short-term rates are rising.
  
  
• Combining the election cycle and the Fed policy cycle produces four possible phases, which, generally speaking, affect our view of the market and strategy as follows:

The Calendar Cycle

The calendar portion of the Fed-Election-Calendar Cycle reflects well-established seasonal patterns of stock market returns. Aspects of the calendar cycle include:

• January tends to be the best month of the year for the market.
  
  
• The market tends to perform better during the November to April period and worse during May through October.
  
  
• Certain tax-related patterns occur at year-end. For example, stocks that have done poorly over the year tend to get sold in December (so investors can realize their tax losses) and then bounce back in January.
  

The Yield Curve

The interest rate yield curve affects our investment process in that the slope of the curve tends to correlate with the relative availability of good opportunities in value stocks.

A normal, positively sloped yield curve places a higher relative value on near-term cash flows than a flat or inverted yield curve. (When short-term rates are lower than long-term rates, near-term cash flows are discounted at a lower rate than cash flows in the distant future.) Therefore, a positively sloped yield curve tends to favor value stocks over growth stocks because a value stock’s price is based more on its near-term earnings prospects (and associated cash flows) than its long-term earnings prospects.

A flat or inverted yield curve, on the other hand, places a higher relative value on long-term cash flows than a normally shaped yield curve. (When long-term rates are the same or lower than short-term rates, long-term cash flows are discounted at a relatively low rate compared to short-term cash flows.) Since a growth company’s stock price is based more on its long-term earnings prospects (and associated cash flows) than its short-term earnings prospects, a flattening yield curve tends to favor growth stocks over value stocks

When the yield curve is growing steeper (i.e. long-term rates are rising faster, or falling slower, than short-term rates), we tend to make money on a higher percentage of our trades and trade less actively because we have fewer losing positions to replace.

When the yield curve is flattening (short-term rates are rising faster, or falling slower, than long-term rates), we tend to make money on a lower percentage of our trades and trade more actively because we have more losing positions to replace.

Diversification

We are typically invested in a well-diversified portfolio of individual stocks, ADR’s, and ETF’s. Our standard is generally to have no more than 5% of the portfolio invested in any one individual stock or at any time.

Investment Process

Our investment process is the means for implementing our investment philosophy and investment style and is designed to maximize risk-adjusted returns in a tax-efficient manner by:

• Maximizing exposure to individual stocks and ADR’s to the extent that good opportunities exist;
  
  
• Maintaining our desired Net Market Exposure through investments in ETF’s when insufficient opportunities exist in individual stocks and ADR’s; and,
  
  
• Holding positions at least one year when doing so would, in our judgment, maximize the net after tax return for the position.

We use information technology to the maximum extent practicable to support all aspects of our research and trading in order to develop data-driven strategies and implement them through pre-determined decision rules that enable us to monitor new information and react to it swiftly, surely, and automatically.

All orders to buy or sell securities are, nevertheless, entered by our Portfolio Manager personally.

We minimize trading costs through judicious trading, low cost electronic trading (currently about $8.00 - $17.95 per trade) and by obtaining good execution. We avoid, insofar as practicable, fees* and other charges that would unnecessarily burden our performance.

In day-to-day operation, our process involves:

• Quantitative Screening of U.S.-traded equities to maintain watch lists of individual stocks and ADR’s, and
  
  
• Event-driven trading.
  

Quantitative Screening and Event-Driven Trading

Investing in Individual Stocks

Our methodology for investing in individual stocks involves a 6-step process that combines quantitative screening and event-driven trading.

1. First, we screen the universe of available U.S. equities and ADR’s, using quantitative filters appropriate to the current market phase, to develop a “watch list” of individual stocks. (The screen is periodically updated.)
   
   
2. Then, we continuously monitor the stocks on our watch list for breaking news that could affect the stock price and for significant, unexplained price movements on high volume that might indicate buying or selling on unreleased news.
   
   
3. We buy an individual stock when breaking news indicates the occurrence of an objectively verifiable positive development likely to affect the stock price, or when a large, unexplained upward price movement on high volume indicates a strong likelihood of buying on unreleased news.
   
   
4. If the market reacts favorably to the positive event, we hold the stock with the goal of achieving long-term capital gain tax treatment for the position. In addition to the positive effect on performance of favorable tax treatment, experience has taught that “good news tends to beget more good news” – in other words, once a company develops a positive operating trajectory, it tends to maintain positive momentum over an extended period of time. Behavioral factors influence investors, management, and the analysts who follow the company to react in ways that tend to reinforce the upward trajectory of its stock price.
   
   
5. If the market does not react favorably, we typically conclude that the event must not have been positive after all, and we close out the position when a better opportunity arises. We close out the position promptly, in most cases, if the market reaction to the news is negative.
   
   
6. We continuously monitor the stocks we own for significant unexplained price movements and breaking news that might indicate a negative development, and then promptly sell, in most cases, when a negative event or a significant decline on high volume occurs.

Short-Selling Individual Stocks

The process of short-selling individual stocks is similar to the process used in buying stocks.

1. First, we screen the universe of available U.S. equities and ADR’s, using quantitative filters to develop a watch list of stocks with characteristics that have historically been associated with poor returns. (The screen is periodically updated.)
   
   
2. Then, we continuously monitor the stocks on our watch list for breaking news that could affect the stock price and for significant, unexplained price movements on high volume that might indicate buying or selling on unreleased news.
   
   
3. We short-sell an individual stock when breaking news indicates the occurrence of an objectively verifiable negative development likely to affect the stock price, or when a large, unexplained downward price movement on high volume indicates a strong likelihood of selling on unreleased news.
   
   
4. If the market reacts as expected to the negative event (i.e. if the stock price declines) we maintain the short position. Experience has taught that that “bad news tends to beget more bad news” – in other words, once a company develops a negative operating trajectory, it tends to continue experiencing problems over time. Behavioral factors typically exacerbate the downward trajectory of its stock price, as management and the analysts who follow the company tend to be unable to accurately perceive or unwilling to fully admit the scope of the difficulties, leading them to report the bad news gradually over time.
   
   
5. If the market does not react as expected (i.e. if the stock price does not decline), we typically cover the short position and replace it by short-selling either another individual stock or an Exchange-Traded Fund to maintain our desired Net Market Exposure.
   
   
6. We continuously monitor the stocks in which we have short positions for significant unexplained price movements and breaking news that might indicate a positive development, and then, in most cases, promptly cover the short upon the occurrence of a positive event or a significant price increase accompanied by high volume.

Results

Historically, our process has tended to produce:

• A relatively large number of small losing bets
  
  
• A relatively small number of large winning bets
  
  
• Tax-wise, relatively high ratios of long-term gains to long-term losses and short-term losses to short-term gains
  
  
• Overall positive returns that exceed those for the S&P 500.