PRIMER – BUILDING A BEHAVIOURAL FINANCE STRATEGY

In this paper, we share our journey of developing investment strategies based on behavioural finance – why we went down this road, how we developed specific strategies and why we believe this approach may be a long-term source of alpha for investors.

 

BUT FIRST, WHY BEHAVIOURAL FINANCE?

Behavioural Finance, which combines finance and psychology, is really the study of how we as investors make decisions, and more poignantly, how we make poor investment decisions.

Psychology has uncovered many heuristics that our brains use to help make decisions quickly. These are rules of thumb or mental shortcuts that enable us to navigate the enormous number of decisions we must make every day. However, many of these heuristics or shortcuts can lead to predictable decision-making errors in certain situations. These biases often result in poor or wrong decisions, and of equal importance, these poor decisions appear to be predictable.

In a low-stakes, calm environment, our biases or heuristics usually help us. For example, “It is cloudy outside; best to bring an umbrella to work.” There’s no need to do a thorough analysis of the weather, barometric pressure, temperature changes, etc. Even if wrong, carrying an umbrella around today is no big deal. However, as the stakes rise and / or the environment becomes more emotionally charged, our biases may hurt us more often. Most would agree investing your money is a higher stakes endeavour than being caught in the rain, since your success or failure may impact your lifestyle at some point. Plus, it can be very emotionally charged, especially at certain times when markets are moving. Both of these are prerequisites for our behavioural biases to get the better of us.

Much research has shown behavioural biases to be systematic. Or, put another way, there are certain situations or events in the market that will trigger the same biases in many investors, such as a big market move, share price reaction to earnings, or a big news-flow day.

If certain circumstances cause investors to act irrationally due to a behavioural bias, does this irrational behaviour impact the price of an asset? And if so, can we profit from the misbehavior of these investors?

We believe the answers are yes and yes.

 

ECONOMICS VS. REAL LIFE

Most economic theories are predicated on the view that the majority of investors behave rationally and in their own self-interest. While stipulating that some investors behave irrationally, for whatever reason, these irrational investors tend to cancel one another out. As a result, the price of an asset is generally assumed to be efficiently priced, incorporating all the available public information.

This may well be the case most of the time. However, sometimes behavioural biases and emotions get the best of many investors, causing them to act irrationally and often, more in one direction over another.

 

We believe that under certain circumstances, there is an increased likelihood that irrational behaviour will be systematically skewed in one direction, causing an asset to be mispriced. The goal, then, is to determine when and under what circumstances this is most likely to occur – and how to profit from the mispriced asset.

All investors make mistakes. If under certain circumstances these mistakes are systematic and largely caused by behavioural biases, then we can develop systematic strategies to profit from them. This is the essence of active portfolio management – finding mispriced assets and profiting from them. We believe a good source of mispricing is investors’ emotional or behavioural mistakes.

 

BEHAVIOURAL STRATEGY DEVELOPMENT
1. Find instances of behavioural bias
There are instances in the marketplace that certainly elicit more biased behaviour than others. Often, these are emotionally charged times caused by significant new information or changing sentiment. One example we have found is the tendency for share prices to overreact to earnings: If a company is truly worth its future discounted cash flow for the next 10-20 years, why would one quarter carry such a big impact if they miss or beat consensus estimates? This is the availability or recency bias in action, which involves placing too much weight on information that is readily or newly available.

When sentiment shifts, it can often trigger behavioural biases. Investors, in aggregate, tend to be susceptible to herd-like behaviour. If it seems like everyone is doing something, there is a strong behavioural desire to join them. There are numerous biases behind this, including confirmation, cognitive dissonance and fear of missing out, and they help fuel bubbles and bear markets.

2. Find evidence of market impact
This is the hard part: discerning whether the behavioural bias is causing an asset to be mispriced. We have researched a great number of scenarios over the past decades where we had a strong belief that investor biases were impacting their behaviour.

During this quantitative stage of our strategy development, we were looking for patterns of consistent mispricing of an asset and how the mispricing corrected over time. Once a reliable pricing anomaly was found, our quantitative approach dug deeper, analyzing how the anomaly acted in different market environments, across different sectors and for companies with different characteristics.

3. Develop trading strategy
Once a reliable mispricing anomaly was discovered, we then developed optimal trading strategies designed to profit from it. The strategy optimization process is rather quant-heavy, and helped refine the strategy to determine in what kind of market or for what kind of company the strategy has had the best results.

For example, our Earnings Overreaction strategy, which is triggered by a price drop following an earnings miss, identified that higher-quality companies (lower beta) tended to partially recover in the subsequent days and weeks with good reliability. The analysis also highlighted which sectors are better for this strategy. Lower-quality companies, conversely, did not show similar recovery patterns with the same consistency.

The trading strategy is further refined to include stop-loss levels, trade duration and profit-taking parameters. This helps further remove our emotions from the strategy implementation.

4. Monitor, test, repeat
While our strategies were developed and tested with up to 20 years of data, the process never ends. Each strategy is continuously tested for efficacy and profitability. We also continue to refine and analyze characteristics and other parameters that can further optimize results.

Given these strategies are predicated on investors’ mistakes, we believe the longevity could prove very long….as long as investors keep making mistakes.

 

BEHAVIOURAL STRATEGIES

On the following pages, we provide a brief synopsis of some of the above-listed strategies. Some target specific behavioural biases, while others target broader market inefficiencies. In each case, the strategy is designed to follow these steps:

  1. Quantitative model – Each strategy employs a proprietary quantitative model designed to find the instances where we believe the biased behaviour is causing an asset to be mispriced. These models generate signals that can become trades for each strategy.
  2. Computer + Human – Quantitative signals are vetted by the investment team, who look for key characteristics that in the past have improved strategy success probabilities. How the trade is incorporated into the overall portfolio is factored in.
  3. Trade parameters – Each trade is sized based on risk and the probability of success. Stop-loss and profit-taking levels are set to avoid our own loss aversion or anchoring biases.
  4. Monitoring and measuring – Markets are always evolving and so are our strategies. We learn as much from trades that didn’t work as from those that did.

 

Note: The following trade examples are for strategy illustrative purposes only and do not denote any actual trade or actual companies. Each trade example has firm stop-loss and profit-taking levels.

Strategy: Emotional Cascade

Targeted behaviour: availability bias, hyper discounting 

Availability bias causes investors to focus more on recent or widely available information and lose sight of the longer-term picture and fundamentals. This becomes especially acute when the quantity of new information spikes due to some news or critical event. This cascading new information can cause an asset price to dramatically overreact, thus presenting an opportunity to be a short-term contrarian.

There are a number of key factors that must be considered when implementing this strategy. As the market is emotional, often ignoring fundamentals, the biggest question is how long this opportunity will last. We have developed a number of tools that help measure when the emotional overreaction is beginning to fade. In addition, we use strict stop-loss levels to help limit downside risk.

 

Trade example: Disparage Co.

Disparage Co., one of the largest operators in its industry, missed second-quarter earnings estimates, triggering a 20% drop, then a further 13% drop the next day on the back of news that two of its competitors were joining forces. Down 33% in two days appeared to be overreaction to both events. As heightened trading volumes slow along with additional factors, this would be a good candidate for the emotional cascade strategy – with a stop-loss in case there are more shoes to drop.

 

Strategy: Earnings Overreaction

Targeted behaviour: availability bias, recency bias 

Availability bias causes investors to overweight the impact of the most readily available news. While earnings are important, investors will often overreact to the release, causing the share price to move more than is justified. We have found an asymmetric regression back to the mean is a function of company quality. That is, higher-quality companies that suffer a significant price drop on bad earnings tend to recover the loss, while lower-quality companies that enjoy a price jump on good earnings don’t tend to hold onto these gains for long.

Trade example: Plummet Inc.

Plummet Inc. reported a big earnings miss in October that sent shares down almost 40%. As Plummet Inc. qualifies as a higher-quality company given historical volatility, this hit our screens as a potential long position. Upon further research into the source of the earnings miss and a review of their outstanding bonds, which were holding in just fine, we determined this was not a solvency issue. This would be a candidate for our earnings overreaction strategy with a stop-loss and profit revisit levels.

 

Strategy: Indexing Bias

Targeted behaviour: market inefficiency, mean reversion 

While counterintuitive, companies added to a major market-capitalization-based index tend to underperform, while those removed tend to outperform. This index bias is really a form of mean reversion: Companies that are added to an index

have already enjoyed a strong price advance as they neared the bottom end of the index’s market cap threshold. They are subsequently bid up more by investors anticipating the potential inclusion into the index, and then once announced, passive investors / ETFs pile on even more buying. By the time they are added, shares have usually been aggressively bid up, and there is often a regression back down to earth. The same process, in the opposite direction, occurs for those being removed.

Trade example: Volte-face Ltd.

In June, Volte-face was added to the S&P/TSX Composite Index. Shares were up 45% over the preceding six months. This would be a short candidate for the indexing bias strategy. A stop-loss level would be set in case the price of the shares continued to climb, as well as profit-taking levels and a time limit. If the shares don’t mean revert in the weeks following the Index change, the trade would be exited.

 

Strategy: Unloved to Less Unloved

Targeted behaviour: confirmation bias, herd behaviour 

Herd mentality is a behavioural bias that is hard-coded in our DNA: we tend to feel more comfortable with consensus. When it comes to analyst ratings, investors generally like to buy, own or add to companies that have a higher percentage of BUY ratings. Conversely, companies with very few BUY ratings over an extended period of time are often forgotten or neglected. The Unloved to Less Unloved strategy attempts to capture unloved companies that start to see analyst upgrades. If the early upgrades are from more forward-looking analysts, more upgrades could follow and create a recovery in the share price.

Trade example: Stellar Ltd.

Between March 2016 and April 2017, Stellar Ltd.’s share price fell more than 60% and the company spent a considerable amount of time being disliked by analysts. However, in May 2017, Stellar received an analyst upgrade that pushed the recommendation ratio (# buys to total analyst recommendations) above our 30% threshold. As a result, it became a candidate for the unloved to less unloved strategy – with stops and profit taking levels, of course.

DOWNLOAD PDF VERSION

Commissions, trailing commissions, management fees and expenses all may be associated with investment funds. Investment funds are not guaranteed, their values change frequently and past performance may not be repeated. The indicated rates of return are the historical annual compounded total returns including changes in unit value and reinvestment of all distributions and do not take into account sales, redemption, distribution or operational charges or income taxes payable by any unitholder that would have reduced returns. Nothing in any commentary should be considered a recommendation to buy or sell a particular security. The Fund may sell these securities at any time, or purchase securities that have previously been sold. The securities may increase or decrease in value after the date hereof, and the Fund may accordingly gain or lose money on the investment in the securities. Please read the prospectus before investing. The opinions expressed are provided by the portfolio manager responsible for the management of the Fund’s investment portfolio, as specified in the Fund’s prospectus. Unless otherwise stated, the source for data cited in any commentary is the portfolio manager. The statements by the portfolio managers in their commentaries are intended to illustrate their approach in managing the funds, do not necessarily reflect the views of Redwood Asset Management Inc.

 

 

 

Return to Richardson GMP