An individual decides to drive home after an evening out despite being knowingly over the legal alcohol limit; before completing their journey they are stopped by the police and charged with driving under the influence. In a parallel universe, the same scenario occurs but with one key difference – prior to being pulled over by the police the individual’s intoxication leads to an accident that causes serious injuries for passengers in another car. How should we view the perpetrator in these two incidents?
Our reaction from an ethical[i], and legal, standpoint is often to judge the second version more harshly – because the consequences were far more severe, but should this be the case? In both cases the main failing is identical – the initial flawed decision to drive after excessive alcohol intake. The relative results, however, are due to luck; the individual in the first instance experienced good luck (comparatively), and the other bad. Such judgements are often heavily influenced by the results, even if they are reliant on chance; an example of outcome bias[ii].
Our tendency to judge the quality of a decision by the ultimate consequence is a simple concept. In many instances it is also a prudent one; results often provide a useful gauge of the value of the actions that led to them. However, as with many things, once you add a healthy dose of randomness things start to become problematic.
“A good decision cannot guarantee a good outcome. All real decisions are made under uncertainty. A decision is therefore a bet, and evaluating it as good or not must depend on the stake and the odds, not on the outcome”[iii] (Ward Edwards)
Financial markets are the perfect breeding ground for outcome bias – results are obvious and easy to obtain, whilst judging process and decision quality is incredibly difficult, which means we rely heavily on the former. We also grossly understate the sheer level of unpredictability, largely due to the wonders of hindsight bias and our susceptibility to a compelling narrative.
In reality, our faith in the information provided by any outcome, should be scaled by the amount of luck there is involved. In certain endeavours results provide a good measure of decision quality; in others we hugely exaggerate the importance of outcomes.
Take chess as an example; it is a heavily structured game, dominated by skill, not chance, and with limited luck or randomness in its results. If I played 100 chess matches with Magnus Carlsen, I would lose each one and these outcomes would prove an excellent indication of our relative abilities. You wouldn’t need to watch each match to know this. Outcome bias is rarely a problem in such activities.
Now imagine that I had to enter a portfolio management competition against my seven year old son, where we each had to pick a portfolio of 30 stocks. As much as I might like to believe that I would hold a significant advantage, I know the probability of my selections outperforming his over a single year are not much greater than 50%. Whilst the odds may tilt in my favour as the time horizon extends there are no guarantees – maybe he has picked some stocks that go on to enjoy dramatic growth, or given his portfolio a factor tilt that is in vogue for a number of years. Not only am I faced with prospect of my diligent investment decision making being improved upon by the haphazard selections of a child, but outcome bias means that my son’s investment success may see him appearing on Bloomberg and asked to give his opinion on the Fed’s next move.
Despite the problems of using results as a barometer of decision quality, it remains endemic in investment. We use outcomes as a simple indicator and then weave narratives around these views. We take a difficult problem, simplify it (are results good or bad?) and then create a story to justify the outcome. This pattern of behaviour is evident in a range of poor investment decisions, such as: susceptibility to financial frauds, participation in investment bubbles, performance chasing and excessive short-term trading.
There is an increasing drive by financial regulators to assess the value for money provided by investment professionals by using simple comparative performance metrics, whilst this is an understandable approach to dealing with a fiendishly difficult problem; it creates a situation where a fluky dart thrower is often perceived to have offered a superior service to someone diligent yet unfortunate. These issues are also why performance fees for actively managed funds are so problematic – they egregiously reward the lucky and pay little heed to process or conscientiousness. There are no easy solutions here but being a beholden to outcomes alone is by no means a panacea.
In an investment context it actually seems wrong to refer to outcome bias; rather we should talk about the outcome heuristic. That is we use outcomes as a mental shortcut to simplify a highly complex and inherently unpredictable task. The use of rules of thumb is smart and effective in some domains, using outcomes as a proxy for sound decision making in investment is anything but.
[i] Gino, F., Moore, D. A., & Bazerman, M. H. (2009). No harm, no foul: The outcome bias in ethical judgments.
[ii] Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of personality and social psychology, 54(4), 569.
[iii] Vlek, C., Edwards, W., Kiss, I., Majone, G., & Toda, M. (1984). What constitutes” a good decision”?. Acta Psychologica.