The Risk in Stocks of Fast Growing Companies
A fast-growing stock – what clients believe is the next Google – is likely to be a disappointing investment. New research, which validates the theory of behavioral economics, shows that “representativeness” explains why clients overweight stocks with high asset growth.
Investing is more about uncertainty (where at best we can estimate the odds of something occurring) than risk (where the odds are known, like at the craps table). When faced with uncertainty while trying to make a decision, people often rely on a mental shortcut known as the “representativeness heuristic” – a simplified approach to problem-solving using judgments arrived at by comparing things to concepts we already have in mind.
In the 1970s, psychologists Amos Tversky and Daniel Kahneman proposed that the representativeness heuristic explains behaviors that lead to mistakes – when people rely on representativeness to make judgments, they are likely to judge wrongly because the fact that something is more representative does not actually make it more likely. In addition, research demonstrates that people overestimate their ability to accurately predict the likelihood of an event.
Miao He, Nishad Kapadia and Sheri Tice contribute to the literature with their August 2020 study, “Can the Representativeness Heuristic Explain the Asset Growth Anomaly?” They began by noting that the low average returns of firms with high asset growth are now well documented. The investment factor (CMA), the return on stocks with low/conservative investment minus the return on stocks with high/aggressive investment, is now incorporated into both the Fama-French five-factor and the Hou, Xue and Zhang q-factor models. The logic for the premium is that asset growth reflects investment and firms that can invest a lot, all else equal, should have low discount rates. However, they noted that “asset growth is not a particularly good measure of investment and better measures do not share its effects on returns. A natural question that arises is what causes high asset growth stocks to have low returns?”
Their study sought to determine if the low returns to firms with high asset growth could be explained by the representativeness heuristic. They explained: “The stereotype that we investigate is that firms with high asset growth are ‘the next Google.’ If investors believe that high asset growth is representative of the ‘next Google’, they will overweight the likelihood of future Googles among high asset growth firms, leading to over-valuation and low future returns. It is intuitively reasonable for investors to believe that high past growth is a representative of Googles and it is also consistent with prior research. Because ‘Google-type’ firms typically grow rapidly for several years, they are likely to be more frequently observed in the sample of firms with high past asset growth relative to the general population.” They also noted that the sales growth-based measure is highly correlated with its asset growth counterpart and results in similar predictive ability for CMA returns. Their data sample covered NYSE, AMEX and Nasdaq stocks. For the distress factor, the data spanned the period 1972-2018, and for analyst forecasts, the period 1984-2018. Following is a summary of their findings:
- There are times when the news about high-asset-growth firms is good – more firms with high asset growth have high returns. In such times, the representativeness bias is more pronounced and investors overestimate the probability of high returns for high asset growth firms to a greater extent. This leads to a greater probability of crashes and lower average future returns.
- When representativeness was high (above its median value), conservative growth firms earned returns that were 18% more than aggressive growth firms over the next three years. When it was low (below its median value), average returns of conservative firms beat aggressive firms by only 5%. The difference was statistically significant.
- Representativeness predicted crashes for the aggressive growth portfolio, but it did not predict booms for that portfolio. Nor did it predict crashes or booms for the conservative growth portfolio. The magnitude of predictability was meaningful, as being in a high representativeness state increased the probability of a crash by almost four times, from 4.4% to 15.6%. And because CMA is short aggressive growth firms, high representativeness predicted booms in CMA returns over the next three years – the probability of a boom in CMA returns rose from 2.2% to 18% when going from a low state to a high state.
He, Kapadia and Tice also noted that representativeness is related to investor sentiment. “Investor sentiment” refers to the general mood investors exhibit toward a particular market or asset. Sentiment can be an important determinant of investment performance because investors who exhibit relatively high sentiment tend to be overconfident and engage in excessive trading, resulting in subpar investment performance because they are trading on “noise” and emotions. Such activity can lead to mispricing. Eventually, any mispricing would be corrected when the fundamentals are revealed, making investor sentiment a contrarian predictor of stock market returns. Studies such as “Global, Local, and Contagious Investor Sentiment,” “The Short of It: Investor Sentiment and Anomalies,” “Investor Sentiment and the Mean-Variance Relation” and “Investor Sentiment: Predicting the Overvalued Stock Market” have found that high-sentiment periods result in overpricing being more prevalent than underpricing because short-sale restrictions limit the ability of rational investors to exploit overpricing (but not underpricing). Given these findings, it is no surprise that He, Kapadia and Tice found that “representativeness is correlated with sentiment, because sentiment is typically high in market booms, when speculative high growth stocks are doing well. … Representativeness predicts sentiment 12 months in the future.” They also found that “representativeness beats sentiment in a ‘horse race’ in predicting future CMA returns.”
He, Kapadia and Tice also examined the connection between high asset growth and default risk. Their hypothesis was that the effect of asset growth on returns is most pronounced in stocks with high default probability because these stocks have the largest amount of risk, which investors might neglect due to the representativeness heuristic. They found that the interaction between asset growth and default probability significantly predicts lower future returns after controlling for the direct effects of asset growth and default probability. Supporting this thesis, they found that the four-factor alpha of the high-minus-low asset growth portfolio was actually slightly positive (0.1% per month) in the low default risk portfolio but fell to 0.64% per month in the high default risk portfolio. They also found that the distress effect was smaller in the low asset growth tercile, with the high-minus-low distress alpha at -0.18% per month, while alpha was -0.92% in the high asset growth tercile.
They noted that the above findings are consistent with those of the authors of the study, “Anomalies and Financial Distress,” who found that “the asset growth effect (along with other anomalies) is more pronounced in stocks with low credit ratings.” That led them to conclude: “The neglect of risk due to the representativeness heuristic provides an economic explanation for why measures of credit risk based on publicly available information are important for the asset growth effect.”
He, Kapadia and Tice next examined whether investors neglect the risk of adverse macroeconomic shocks for firms that have high asset growth and high distress risk. They found that the asset-growth effect is especially bad during recessions, raising the bar for rational explanations, because stocks that do badly in economic downturns should have high, not low, expected returns. They also tested whether analysts and investors overestimate the earnings of distressed stocks with high asset growth. Consistent with their other findings, distress and asset growth predict more negative forecast errors and earnings-announcement-day returns – both analysts and investors are guilty. Representativeness leads to overconfidence and neglect of risk. Also consistent with their other findings is that overconfidence and neglect of risk are higher when representativeness is high.
In a test of the robustness, He, Kapadia and Tice found that a sales growth-based measure is highly correlated with its asset growth counterpart and results in similar predictive ability for CMA returns.
The research findings provide powerful evidence that both the representativeness heuristic and sentiment lead investors to make the behavioral error of overconfidence, which leads to underestimating risks and overestimating the odds of hitting the jackpot – finding the next Google.
Behavioral economists will also be pleased to learn that the evidence presented explains the relatively recently discovered empirical anomaly of high-asset-growth firms having poor returns with the behavioral theories proposed in the early 1970s work of Amos Tversky and Daniel Kahneman.
Larry Swedroe is the chief research officer for Buckingham Strategic Wealth and Buckingham Strategic Partners.
Important Disclosure: The information contained in this article is for educational purposes only and should not be construed as specific investment, accounting, legal or tax advice. The analysis contained in this article is based upon third party information available at the time which may become outdated or otherwise superseded at any time without notice. Certain third-party information is based upon is deemed to be reliable, but its accuracy and completeness cannot be guaranteed. By clicking on any of the links above, you acknowledge that they are solely for your convenience, and do not necessarily imply any affiliations, sponsorships, endorsements or representations whatsoever by us regarding third-party websites. We are not responsible for the content, availability or privacy policies of these sites, and shall not be responsible or liable for any information, opinions, advice, products or services available on or through them. The opinions expressed by featured authors are their own and may not accurately reflect those of the Buckingham Strategic Wealth®, Buckingham Strategic Partners® (collectively Buckingham Wealth Partners). LSR-21-27