Optimism Bias and Market Recoveries
By John M. Burkhardt
Risks & Rewards, February 2022
While the major market crash subsequent to the widespread contagion of the COVID-19 virus was readily predictable, the rapid recovery and ongoing growth subsequent to it was not. Market dynamics around the recovery resist characterization by most traditional analytical approaches. In examining the circumstances around the emergence of a large retail class of investors, one sees that ingrained human tendencies, particularly optimism bias, have fueled much investor behavior responsible for the recovery. Understanding this makes it possible to predict potential future crashes and develop behavioral interventions for subsequent market recoveries.
Pernicious Optimism
As human beings, investors are subject to a wide range of cognitive biases. These include heuristics such as loss and risk aversion, which drive individuals toward safer options even in the face of potentially greater gains. Intertemporal discounting changes the perceived value of outcomes depending on how far in the future they are expected to occur. Whether an outcome is positioned as a gain or a loss (“reflection effect”) changes the decision-making architecture. Confirmation bias causes individuals to ignore evidence that does not support existing beliefs. These and dozens of other biases all influence every decision an individual makes.
Among the many biases that humans experience, optimism bias is especially important for understanding investor and market behavior. Optimism bias refers to individuals’ tendencies to believe that they are less likely to encounter a negative event than historical experience would warrant. This is an intrinsic and not learned bias, and is observed across ages, sexes, ethnicities, and nationalities.
Optimism bias interacts with several other well-established biases. It is partially facilitated by the representativeness heuristic, which leads to misestimation of baseline event and outcome probabilities. It is likewise fueled by the availability heuristic, wherein information that is more easily drawn into working memory is given greater weight in decision-making. As people generally prefer to imagine the results of good investments (and can even be said to self-reinforce by doing so), such thoughts are more easily pulled into attention. Optimism also fuels overconfidence bias, wherein individuals tend to believe they have greater agency and competence than they genuinely do.
It also tends to occur more strongly around negative events. One tends to focus on the benefits of positive outcomes and downplay the consequences of problematic outcomes. This is enormously consequential in investment decisions, where narratives and informational content tend to emphasize negative outcomes in order to mitigate potential loss. It also means that optimism bias interacts with risk aversion, in ways that will be discussed later.
Unlike many biases, there are few precipitating circumstances or factors for optimism bias. It is constitutively active rather than deployed discretely in certain decision-making contexts. This means that optimism factors into virtually every decision people make, including investments.
There is a general evolutionary imperative for optimism bias to occur. In Darwinian terms, an individual always needs to be planning or expecting to reproduce, and a belief in things turning out well is a necessary antecedent to taking such actions. Likewise, optimism bias is an essential factor for any retail investor. The central thesis of securities investing—that it yields a better return than bonds or savings accounts—is predicated entirely on optimism.
An Irrational Crash
Insofar as any market crash can be described as having some degree of rationality, the February 2020 crash of the U.S. stock market was particularly irrational in every objective dimension. Using the Dow Jones Industrial Average (DJIA) as a proxy for overall domestic market activity, a very slight decline began on Feb. 12, 2020, which accelerated into a full-blown crash less than two weeks later on Feb. 21. The DJIA lost just over 10,000 points in the span of a month, declining 37 percent to levels not seen since November 2016.
Curiously, there seems to be no precipitating event or policy action tied to the onset of the crash. The first documented cases of COVID-19 in the United States occurred on Jan. 21, 2020. The World Health Organization (WHO) declared a global health emergency 10 days later on Jan. 31, and the first travel restrictions were put in place on Feb. 2. The United States did not declare a national emergency until March 13, concurrent with large-scale travel restrictions. Large-scale shelter-at-home guidance was not issued until March 19, when the crash was nearly over. In hindsight, the crash was most likely a consequence of behavioral contagion. In a period of great uncertainty with a generally negative outlook, a small drop in market value was seized upon as an indicator of what to do, triggering a widespread sell-off.
When Is a Recovery Not a Recovery?
The COVID crash ended on March 23, 2020. While normally determining the conclusion of a crash and the beginning of a recovery would involve substantial examination of historical financial data and social and policy impacts, no such research is required in this case.
The post-crash recovery was not so much a recovery as a rebound. There was no gradual change in market direction and no malingering at or near the nadir. On March 23 the decline in DJIA value ceased, and the vector wholly reversed. Full recovery to pre-crash market levels was achieved on Nov. 16, 2020, and at that point the recovery transitioned into a bull market. As of this writing, the peak occurred on Nov. 16, 2021, with the DJIA at 36,142—a 23 percent increase over the pre-crash peak and a full 90 percent increase from the crash nadir.
As with the crash itself, no typical news or events appear to be tied to the recovery or subsequent current market growth. The CARES Act was passed on March 26, 2020, experimental use of hydroxychloroquine was authorized on March 30, and the first guidelines for a general reopening of the economy were proposed on April 16.
Moreover, the recovery was apparently immune to negative news events as well. The initially promising studies of hydroxychloroquine studies were retracted on June 4, 2020, and states began reversing their reopening plans a month later on July 2. Neither development had any noticeable impact on the market, and the recovery continued unabated.
A closer examination of the recovery reveals that it has occurred in distinct phases. During the initial phase (roughly March 2020 to July 2020), investor behavior was primarily opportunistic. This could be characterized by both optimism bias—an unfounded belief that things will turn out well; and risk positivity—a preferential weighting of the gain side of a decision over the loss side. Enough investors determined that there were sufficient upside opportunities during the crash to invest heavily, resulting in a complete reversal in market direction. Given the timeline of professional sports league closures, there is some speculation that this phase may have been in part fueled by sports gamblers, a notably risk-positive population, who were bereft of their normal risk-taking outlets.
The second recovery phase spanned roughly from July 2020 to February 2021. This period can be characterized not so much by optimism as by savvy—experienced investors identifying an opportunity and seizing it. At this point the recovery had demonstrated its short-term sustainability, while markets were still viewed as massively undervalued. It is most readily characterized and predicted through classical fundamental and trending tools.
The third phase of the recovery began in February 2021 and continues to the present day; it aligns with the ascent of the retail investor. Flush with CARES Act cash, convinced of their own agency and ability in the wake of the January short squeeze of GameStop, an enormous number of previously passive investors became active participants in securities markets. Already optimistic, this population’s bias was further reinforced as the DJIA surpassed pre-crash levels and kept climbing. At this point, subsequent investing became a feed-forward cycle. Investors became emboldened by market growth during prior months, fueled by their belief that they could do even better. This led to significant new cash infusions into equities, which in turn fueled higher market levels, confirming the optimistic outlook and drawing even more new investors to the asset class.
Notably, with this increase in the number of retail stock participants, markets are now even more strongly subject to the “irrational” behavior of investors. Behavioral analysis of investment trends and market activity will become even more central for any predictive analytics.
What Comes Next?
Currently optimism bias interacts strongly with investors’ available liquidity. Individuals presently have the means to invest more, and their optimism will compel them to do so. When enough retail investors do so simultaneously, it can sustain a market rebound longer than any fundamentals-based model would suggest.
The major question then becomes, how long is such optimism sustainable? Recent events have shown no signs of tapering optimism and, indeed, have emphasized how powerful a motivator optimism bias is. With the emergence of the omicron variant of the COVID virus, markets predictably dropped sharply with the first announcement of the variant. Dips were again recorded when its spread to multiple continents was documented. In both cases, however, there were sudden rebounds only days after the dips that were not associated with any news, public health response, or government policy. Investor optimism continues to fuel what can no longer reasonably be called a recovery and can only be characterized as a full-blown bull market.
The continuous rebounds in turn provide confirmatory evidence for optimism and fuel investor confidence, driving overconfidence and status quo biases. These come from a deep evolutionary belief: The fact that an individual has survived means that their behavior to date is correct. In this context, the fact that an investor is still observing gains and is net positive means that their behavior is correct, so redirecting to different behaviors (e.g., asset classes, investment strategies, etc.) is not necessary.
Absent some genuine infrastructure-altering event, such optimism bias can persist indefinitely provided there is sufficient liquidity available. As long as a critical mass of investors can continue to double down in response to dips, rebounds will drive optimism, which will feed forward into subsequent dip responses.
This is where the inflation becomes especially pernicious, as it functionally reduces liquidity. Once an investor no longer has sufficient liquidity to fuel their optimism, loss aversion rapidly and inevitably sets in. At this point, even modest market dips can prompt an investor to take profits or even close positions entirely. In this scenario, dips are followed by malingering or even further drops rather than an immediate rebound. These in turn make a negative contagion event more likely, leading to a market correction or even a crash. For purposes of forecasting and risk mitigation, understanding what is likely to precipitate this switch from optimism to loss aversion and when it will occur is critical.
Summary
Retail investor behavior is heavily influenced by cognitive biases, and these biases have been key factors in both the post-COVID market crash of 2020 and the unprecedented recovery and growth that immediately followed. Uncertainty and loss aversion drove contagion around the crash, and optimism bias supported sustained active market participation and investment by individuals. These biases can function as an inflection point in market activity when one strongly predominates over the other, turning recession into recovery or vice versa. Understanding the dynamics around these factors is essential for long-term planning and risk mitigation around market behavior.
Statements of fact and opinions expressed herein are those of the individual authors and are not necessarily those of the Society of Actuaries, the editors, or the respective authors’ employers.
John M. Burkhardt, Ph.D., is the CEO of Capita Neuro Solutions and Adjunct Professor in the Enterprise Risk Management program at Columbia University. He can be reached at jburkhardt@capitaneuro.com.