Swedroe: The Perils Of Bargain Hunting

As I have been discussing in a series of articles (which you can find here, here and here), we now have a substantial body of evidence demonstrating that individual investors possess a preference for low-priced equities. This is anomalous behavior, because the level of a company’s stock price is arbitrary—firms can manipulate it by adjusting the number of shares they have outstanding.

The research from a trio of studies on the U.S. stock market has found that this irrational preference is explained by individual investors who are searching for a cheap bet, as with lottery tickets. Therefore, such investors find lower-priced stocks attractive. The research also confirms the view that individual investors may see low-priced stocks as being closer to zero and farther from infinity. Thus, they are perceived to have more upside potential and less to lose.

Summarizing the findings from these three recent papers, investor preferences and beliefs lead to the pricing anomaly in which low-priced stocks—especially those with higher betas, high idiosyncratic volatility and high skewness—show poor raw returns and very poor risk-adjusted returns.

The Out-Of-Sample Evidence
As noted, while the cumulative weight of the evidence is compelling, it’s all based on U.S. data. We would have more confidence in the authors’ findings if we had an out-of-sample test that produced similar results. Thanks to Ulrich Hammerich, Christian Fieberg and Thorsten Poddig, authors of the October 2016 study “Nominal Stock Price Investing,” we now have that out-of-sample evidence.

In their study, which covered the period 1990 through 2013, the authors investigated the relevance of nominal prices when applied as an investment style on the German stock market. They found that nominal prices explain portfolios’ excess returns (built via price deciles) when applying the CAPM (market beta), the three-factor model (market beta, size and value) and a four-factor model (adding momentum). Following is a summary of their findings:

• When moving from decile P1 to decile P10, returns increase virtually monotonically, while standard deviations basically decrease monotonically and Sharpe ratios basically increase monotonically.
• There is a significant return difference between the cheapest 10% of stocks (P1) and the most expensive 10% of stocks (P10).
• Although there only is a barely/marginally statistically significant difference in the returns in favor of high-priced stocks, the Sharpe ratios and the standard deviations show a very robust, significant difference concerning P1 and P10 (P10 has a lower standard deviation of returns than P1 and at the same time higher mean returns; thus, a much higher Sharpe ratio).
• If an investor scales up the volatility of P10 via the use of leverage to the same level as P1, the return difference in favor of P10 would be dramatically higher.
• Results from the CAPM and four-factor model show that the main differences between high-priced and low-priced stocks are that high-priced stocks are, on average, low-beta stocks and have positive and very significant coefficients on the momentum factor, while low-priced stocks likely are high-beta stocks and have a very significant negative factor loading on momentum.

The authors noted “the skewness of the returns of P10 is less favourable as to investors’ preferences, being skewed to the left” (that is, negatively skewed). As we observed previously, investors prefer positive skewness, as with lottery tickets. The authors also point out that their results held up to several tests of robustness.

Consumers Are Bargain Hunters

Hammerich, Fieberg and Poddig offered another interesting explanation for individual investors’ preference for low-priced stocks, providing this analogy: “Everyday consumers are trained to identify bargain buys in the (local) store as well as (and even more) on the internet.”

Thus, it seems unlikely that, due to the daily training in bargain hunting, consumers would be able to erase their inherent dispositions completely when deciding which stocks to buy.

Contributing to this bias is that stocks’ past prices serve as reference prices. They write: “Since low-priced stocks generally suffer from a bad/negative (past) momentum, past prices of low-priced stocks tend to be higher than current prices, aggravating the bargain buy illusion. On the other hand, high-priced stocks typically exhibit good/positive (past) momentum, leading to higher present prices in relation to past prices, seemingly rendering high-priced stocks a bad buy.”

Unfortunately for investors, as the authors demonstrate, expression of their preferences leads to the lower-priced stocks being relatively expensive and the higher-priced stocks being relative bargains. In addition, well-known limits to arbitrage prevent sophisticated institutional investors from correcting the mispricing.

In summary, the out-of-sample test by Hammerich, Fieberg and Poddig confirms the relevance of the nominal stock price insofar as investors’ future expectations concerning stock (performance) characteristics and investment behavior.

It also points the way toward a strategy that avoids these low-priced stocks, something that investment firms such as Dimensional Fund Advisors and Bridgeway have been doing for a long time. (In the interest of full disclosure, my firm, Buckingham, recommends Dimensional and Bridgeway funds in constructing client portfolios.)

Larry Swedroe is the director of research for The BAM Alliance, a community of more than 140 independent registered investment advisors throughout the country.