ACKNOWLEDGING EMOTION WITHIN THE INVESTMENT PROCESS

“STAYING THE COURSE” AMIDST CHOPPY WATERS

The traditional approach to investment management assumes that investors have equal levels of knowledge, and that they make decisions based only on facts and logic. Emotions and external influences play no role in their decision making. In this context, investors maintain their strategies throughout the ups and downs of the markets, providing them a reasonable chance of achieving the returns they expect. Is this realistic? Do emotions play any role in the investment process? If they do, how do we incorporate them into the process of developing strategies that investors can truly maintain for the long term?

Behavioral finance challenges the assumption that investors are not thrown off course by their emotions. Research over the past few decades has identified several “cognitive biases” that affect investors’ decisions, usually causing negative results (as compared to sticking with a well-formulated plan.) Some of these biases include overconfidence (trading too much and making concentrated bets,) herding (following the crowd by buying on the way up and selling on the way down,) and loss aversion (avoiding losses instead of capturing gains.) These are just three of the “top ten” bad behaviors driven by emotion.

Regret is often the driver of emotion-based decisions that result in unnecessary and ill-timed trading that erodes portfolio value throughout the active process. We experience regret in decisions we have made, and in decisions we failed to make.

One example of the first source of regret is found in unrealized losses on an owned asset, after it falls from its high point, causing despair and the regret of not having sold the asset. The second type of regret comes from reading headlines about an unowned asset that increased in value substantially, evoking envy from not having purchased that asset. Often found at the front end of the “despair-inducing” asset is the “fear of missing out” as an asset begins its extraordinary rise in value, only to reverse course and turn envy into despair. This “buy high, sell low” scenario illustrates the “whipsaw” effect that often comes from emotion-based trading. We see that both reactions are emotionally based and unrealistic, since they assume that future events could be anticipated. Sadly, this regret-induced trading incorporates all three of the behavioral biases we identified as the most harmful. Sadder still is the high likelihood of repeating this behavior.

Does reminding ourselves of this tendency toward emotion-based trading help us behave better? Unfortunately, that is not usually the case – these are emotional reactions that are not typically resolved through logic. Behavioral research suggests that we are highly likely to repeat these mistakes, even after being forewarned.

Is there a better way to handle this? We believe so.

In this paper, we focus on the key sources of regret and develop an approach to minimizing its occurrence in the asset allocation process. First, we diversify the portfolio adequately, minimizing the occurrence and the severity of extreme returns (both positive and negative) that drive these emotional, knee-jerk reactions. We then provide a method for quantifying the regret that is inherent in the strategy. This is a crucial step toward understanding whether the investor can realistically maintain the strategy over the long term. (Behavioral finance research has not yet addressed this part, as it simply enumerates the types of emotional responses that investors experience. The research is interesting, but not particularly useful.)

Our first paper demonstrates the relationship between diversification and regret, and then measures the inherent regret of several investment strategies. Our second paper will present a methodology for matching clients to strategies, based on their tendencies toward regret.

THE ROLE OF KNOWLEDGE

We acknowledge that we cannot predict the future with any certainty. Despite having data, analysis, and expert opinions, our forward-looking estimates of market behavior are essentially educated guesses. To manage the uncertainty of this knowledge gap, we plan for the outcomes that may occur, and we hold investments that capitalize on favorable outcomes, combining these with other investments that mitigate the unfavorable ones. Investors can reasonably expect more stable returns from this simple, intuitive, and broadly-based diversification approach.

We outlined this approach to effective diversification in our recent paper: Asset Allocation: No One Knows the Future

Smart Asset Allocation Strategies for an Unpredictable Future | First Rate, Inc.

We evaluated our results using nearly a century of market data for US investments, following the market cycles as the economy moved through times of peace and extreme geopolitical stress. This analysis includes the types of “regret-inducing” events investors are likely to encounter, and builds a portfolio to minimize these extreme returns while capturing a long-term market return that helps clients meet their financial goals.

Our paper focused on a strategy that invested half of the portfolio in equity, with the other half in “reserves” that included a 50-50 mix of Treasury bonds and gold. This provided an asset for good markets (stocks,) bad markets (bonds,) and periods of extreme geopolitical and economic stress (gold.) That strategy earned about 8-3/4% annually with about 12% volatility over the prior 95 years. We also observed 10-year annualized returns between 7-1/4% and 15% for the decades over that time horizon, while containing a small number of “regretful” returns.

EXPANDING THE SET OF INVESTMENT STRATEGIES

In this paper, we increase the number of strategies while also increasing the degree of diversification within each strategy. We continue to allocate assets based on their general behavior in each economic circumstance; we make no claims of forward-looking predictive skills. We simply increase the number of investment options.

• Our equity remains a mix of 70% large cap and 30% small cap stocks.
• Our Treasury bonds are equally-weighted across 10-year and 20-year maturities, with a 20% allocation to cash.
• We combine gold with real estate in equal proportions to create a “real assets” category.
• Our “Reserves” portfolio is equally weighted between bonds and real assets.

Our initial portfolio is the “No Knowledge” allocation of 50% equity and 50% reserves. The difference from the equivalent allocation in our earlier paper is the inclusion of cash in the Treasury allocation and the mix of real estate and gold to form the “real assets” component of our reserves. That strategy now has 50% in equity, 25% in Treasuries and 25% in our equal mix of real estate and gold.

We add two more strategies – one more conservative and the other more aggressive. These are determined by the equity-vs-reserves mix. The conservative strategy has an equal allocation between equity, Treasury bonds and real assets. This is sometimes called a “Talmudic” portfolio because it follows ancient guidance to split wealth equally between companies, land, and liquidity (i.e. equity, bonds/cash, and real estate/gold.) By putting 1/3 of the assets into equity, we achieve this equal split. The aggressive strategy allocates 2/3 of the assets to equity. 

A TRADITIONAL LOOK AT OUR INVESTMENT STRATEGIES

This first chart shows the 95-year annualized compound returns and the volatility risk of our strategies, along with the returns on its component assets. Diversification within the reserves grouping produces a higher return and a lower volatility than either of its assets (i.e. lower risk than Treasuries and higher return than real assets.) We also see diversification benefits from combining equity with our reserves, as the efficient frontier of our strategies is higher than the line connecting equity and reserves. From a mean-variance perspective, our strategies are reasonable and efficient, providing clients with an expected 100 bps of additional long-term return for each increase in equity exposure.

QUANTIFYING REGRET: IT STARTS WITH DEFINITION

So far, we have described regret in terms of its sources and its effects on behavior. But if we are to measure and evaluate regret, we must first define it in terms of the degree of gains or losses that produce these regretful reactions.

We start with the behavioral finance insights which indicate that investors feel the pain of losses much more than they experience the joy of unexpected gains. We use a 2:1 ratio in evaluating the perception of significant losses relative to extraordinary gains. Our regretful return thresholds exceed the range of expected gains and losses for a diversified strategy, and are reasonable estimates of “serious” deviations from an investor’s expectations. This approach also considers the fact that disproportionately large gains are needed to recover from significant losses. Bringing these ideas together, we establish a -12% threshold for regret-inducing losses and +25% for gains the investor may have missed. Returns beyond these limits evoke feelings of regret.

To calculate the “regret penalty,” we took the average of all returns above 25% and all losses greater than -12 percent. We also counted the number of occurrences, and expressed these as a percent of our 95 sample returns. By multiplying these values, we created an “upside” regret (envy or “FOMO”) and a downside regret (despair.) Since both losses and gains produce regret, we treat the regretful losses as positive values in deriving the total regret penalty. This penalty is subtracted from the mean return to derive our “regret-adjusted return.” We focus on mean return because we experience regret in the short term.

The upside regret of “missing out” is the greater source of overall regret, given its higher average return and greater frequency of occurrence. Surprisingly, while overall regret doubles between the 1/3 equity portfolio and the ½ equity portfolio, the regret triples when moving from the ½ equity portfolio to the 2/3 equity portfolio. As a result, the regret-adjusted return drops most dramatically from the expected return in the aggressive portfolio. As a result, the regret adjusted return of the aggressive portfolio is the lowest of the three. This is because we passed the “tipping point” where the negative regret-adjusted return of equity dominates all other effects. This is illustrated in our second chart. The sources of regret (amount vs likelihood) are shown in the third chart.

For investors prone to feelings of regret, either the 1^st or 2^nd strategies would be reasonable choices, but the benefits from the higher expected return in the 3^rd portfolio are more than offset by the significantly higher regret penalty. In practical terms, this strategy is more likely to induce regret that might cause the client to abandon the strategy at the worst time, effectively increasing the likelihood of the whipsaw effect of selling low and eventually re-entering the market at higher levels.

ATTRIBUTION OF REGRET: ENVY VS DESPAIR

The calculation of the regret penalty provides an attribution of the regret effect. Since regret is the expected value of returns that exceed our regret targets, we simply multiply the average of the returns in each regret range (gains or losses) by their likelihood of occurrence. In our example, we set our regret triggers at -12% and +25% and we found the average for the returns that exceed these limits. We then found the probability for those returns and applied these inputs to our regret formula:

Regret = [ Average upside regret return * % Likelihood ] + [ (Average loss return * -1) * % Likelihood ]

We treat both upside (“envy”) and downside (“despair”) regret equally in our calculation. By setting a downside regret trigger at one-half the upside trigger, we incorporate the behavioral principle that losses affect the investor’s psyche (and emotions) twice as much as extraordinary gains.

Our initial attribution analysis of regret compares the proportions of downside vs upside regret for each strategy. Our chart shows that the first two strategies are relatively balanced between the sources of regret, but as these strategies become more aggressive (i.e., higher equity allocation) the proportion shifts from downside regret to the “fear of missing out” (upside regret.)

Our attribution of regret also examines the factors of severity and likelihood as we move across the strategies. We begin with downside regret: a relatively linear increase across the diversified strategies.

The upside regret tells a different story: relatively stable upside regret severity but a dramatic increase in the likelihood of experiencing this regret. Clearly, the driver of overall regret is the likelihood of the upside “envy” which encourages momentum-chasing behavior that begins the “whipsaw cycle” of ill-timed tactical trades.

USING REGRET ANALYSIS TO GUIDE ASSET ALLOCATION

This regret analysis provides an additional aspect of relevance to the asset allocation process. Our initial mean-variance approach provided us with expected long-term returns from 7.4% to 8.4% to 9.3% with allocations to equity ranging from one-third to two-thirds of our assets (with the remaining assets allocated to the “reserves” portfolio.) Given the client’s return goals, we can decide which strategy is most appropriate.

Are we done? For many practitioners, this is the end of the process – but we are just getting started.

This analysis is predicated on the assumption that the client can tolerate the inherent regret of each strategy. But we can only assume that each strategy earns its expected long-term return IF the client maintains that strategy across market cycles. The looming impediment to success is exiting and then re-entering the strategy because the client found the regret of unexpected short-term gains and losses intolerable. It is essential that clients select strategies they can stick with and “stay the course” throughout the inevitable periods of high market volatility. The adage “It’s time IN the market that matters, not TIMING the market” rings true, and it is better for a client to select a lower-returning strategy with tolerable regret, rather than a higher-returning strategy with an intolerable level of regret that likely leads to exiting the market and re-entering at the worst times.

This is what true risk management looks like. It starts with an honest acknowledgement of the human emotions that we all experience. Our regret process provides a way to identify, quantify and control these emotion-provoking conditions. This increases our client’s likelihood of success, as we secure the first part of the portfolio management process.

CUSTOMIZING REGRET ANALYSIS FOR INDIVIDUAL CLIENTS

We proposed a framework for identifying and quantifying regret using “triggers” of 12% losses and 25% gains in evaluating the regret inherent in each strategy. These are reasonable estimates of the regret factor, but our model can be tailored to each client’s individual sensibilities about risk, return and regret. Just as clients have an aversion to volatility risk, they also have an aversion to regret. In our next paper, we will evaluate our strategies in the context of clients along a “regret spectrum.”

Our goal is to establish a credible, behaviorally-based method of matching clients to the most suitable investment strategy, given their “regret aversion.” We believe that this approach is superior to the qualitatively-oriented questionnaires currently in use by investment firms, where clients “self-identify” with a general description of their appetite for volatility risk.