At the Cornell Capital group we put investment strategies into three general buckets. The first bucket is easy – accept market efficiency as a working hypothesis and buy highly diversified index funds. Note that accepting market efficiency as a working hypothesis is not the same as concluding that the market it is efficient. It means that as an investor you feel you lack skill or resources necessary to exploit inefficiencies that might exist at a cost that makes the effort worthwhile. It makes sense, therefore, to act as if the market is efficient. By highly diversified, Bill Sharpe would say that ideally the investment of choice is a global market fund. If your focus is more domestic, then the choice would be a U.S. market index fund such as the one offered by Vanguard.
[activistinvesting]
Q2 hedge fund letters, conference, scoops etc
The second bucket of investment strategies involves taking account of factor anomalies and factor premiums. The best way to explain this approach is with an example – the small firm effect. Dating back to the work of Rolf Banz in 1981, researchers have found evidence that small firms (measured by market capitalization) offer higher risk adjusted returns than predicted by the capital asset pricing model (CAPM). In addition to size, two of the other classic factors found to be associated with significant premiums are value (stocks with a high book-to-market ratio) and momentum (stocks that have run-up in the last twelve months). These findings set off a debate, unresolved to the current day, as to whether the premiums were an anomalous “free lunch” or whether they were risk premiums that the CAPM failed to incorporate. No matter how the premiums are interpreted, the goal of the second strategy is to harvest them by overweighting in stocks related to the factor (such as value stocks). The drive to uncover new factors and their associated premiums led to an explosion of research. Harvey, Lui and Zhu report that, as of 2016, published research articles had found 316 statistically significant factors![1][1]
At Cornell Capital, we are skeptical of efforts to harvest factor premiums for two reasons: data mining and nonstationarity. The data mining problem arises from the fact that there is only one stock market history. If that history is studied enough times, researchers are sure to find “patterns in returns” that are not true patterns but random idiosyncrasies – that is data mining. In addition, researchers may also find patterns that held in the past but are no longer applicable today – that is nonstationarity. Either way, investors who attempt to exploit factor premiums found in the historical record are going to be disappointed going forward. While we pay attention to factor premiums, particularly classic ones like value and momentum, they are not a centerpiece of the investment strategy at Cornell Capital.
The third bucket is good old stock picking based on fundamental discounted cash flow valuation analysis. This approach amounts to looking for stocks that, according to the investor, have a fundamental value significantly greater or less than the market price and then taking the appropriate position to exploit the mispricing. Such an approach is not easy as emphasized in the previous post. In a highly competitive market, most valuation analyses will indicate that the stock in question is fairly priced. In those instances where mispricing is found, it is almost invariably because the investor and the market disagree regarding the cash flow forecasts for the company. But that begs the question as to whether the market or the investor has made a mistake. And even if the investor is correct, he or she will not reap the benefits of the analysis until the market comes to recognize the error of its ways.
Every investor must choose for himself or herself what combination of the three general strategies to employ. At Cornell Capital, we use a bit of all three with emphasis on the first and the third. That is, we hold a market index fund to give us general exposure to the US equities. We then add specific positions, long or short, often using options, to take account of what we believe are significant mispricing of specific securities. For instance, as readers of this blog know, we have believed that at any price greater than $300 Tesla is overvalued. Finally, we take account of the factor premiums when placing the firm specific bets. For example, we are prone to favor value stocks over growth stocks. Finally, we also take account of the general level of stock prices as measured by indicators like Prof. Shiller’s CAPE ratio, but that is a topic for another day.
[1][1] Harvey, Campbell, Yan Liu and Hequing Zhu, 2016, And the cross-section of expected returns, Review of Financial Studies, 29: 5-68.
Article by Brad Cornell’s Economics Blog
