Didn’t I do well? – Addressing Hindsight Bias

CFA UK Annual Conference. 21 Jun 2012.

Investment managers should use journals, documenting their key decisions at the time they are made. There are good reasons for doing this, rooted in the findings of behavioural finance. We need to learn as observers of our own performance. And, I believe it can make all of us better investors.

My talk is titled “addressing hindsight bias”, but there are a number of other psychological failings also at work. Behavioural finance is often accused of lacking conclusion or application. Identifying errors and biases is all very well, but what specifically can be done about it? This is one area where psychology and research has delivered a workable solution. I think that working with a contemporaneous record is a simple practical way of addressing a number of the problems, and making better decisions as a result.

Our memory is designed for learning, to help us in the future, but not necessarily with the right understanding of complex events. “Memory is the mother of all wisdom” (Aeschylus), but it fails us. Put simply, “hindsight does not equal foresight” (Fischoff). Salient information and more emotionally significant events distort our memory, often blinding us to other factors.

And, we tend to edit our memories to favour our own particular personal narrative. This is not always obvious to us, though it may be to others.

But, the key issue is the tendency to reframe the past with the knowledge we have today. In investment, this typically means we know whether a decision ended up with a good or bad result. This knowledge of the outcome can strip bare all the nuances and factors involved, till all we see is a stark binary choice. We have a tendency to view events as being more predictable than they really are – hindsight bias, often referred to as “I knew it all along”. Oversimplifying causality – seeing history as being more determinant and the outcome as inevitable – is a coping mechanism, helping us to come to terms with adverse outcomes.

As Kahneman puts it; a general limitation of the human mind is its imperfect ability to reconstruct past states of knowledge, or beliefs that have changed. Once you adopt a new view of the world, you immediately lose much of your ability to recall what you used to believe before your mind changed. And, we tend to attribute our mistakes to circumstance rather than bad calculation. In contrast, the mistakes we make in judging others, tend to be the opposite, where we place too much emphasis on their personal characteristics rather than on context and the environment; fundamental attribution error.

In essence, we judge others by their actions, but we wish that we instead are judged by our intentions.

And, self-serving bias means that we tend to want to claim personal credit for success but blame failure on external factors. Preserving self-esteem and our personal narrative gets in the way of any new contradictory evidence. So, for the same reason that we do not usually let pupils mark their own homework, we are not the best judge of our own decisions, even in retrospect.

Of course, we could ask others to review this, and it may be in team work this sort of cross-appraisal goes on. But we often do not trust it and therefore do not learn the right lessons. Many investment managers have sophisticated performance and attribution analysis systems, putting on their screen detailed analysis of their trading history. But, this too can be reshaped to suit our perspective. There are so many numbers, that it is easy to focus on the ones, or the specific timescale, that looks best. And, these systems typically capture numbers rather than context.

Typically we want to blame the context for bad decisions, but do not actually document it.

In other areas of work, such as experimental science, the concept of contemporaneous records is well established. As a profession we aspire to being scientific, yet typically do not take the same rigorous approach. Certainly, many will write up an investment recommendation or meeting and some will make occasional jottings of useful thoughts. But, I do think it is important that a record can be organised, allowing search, and refined.

The key is that it must be a trusted tool that can beused alongside the other information sources for your work.

The perspective put on tools and cognition by philosophy lecturer and author, Professor Andy Clark of Edinburgh University, is a useful one. In his paper with David Chalmers, The Extended Mind, he argues that humans integrate their tools into their decision-making. If someone asks us if we know a friend’s phone number, we might answer yes before getting out our phone contact list to see exactly what it is. In terms of what goes into a decision, the boundaries between what lies in our existing memory and what is readily available in tools we trust, is blurred. Nearby objects in the environment can function as part of the mind.

The key is to maintain a record that we do trust as a guide to what has happened in the past. That will make it more likely we will integrate it into what we do in future.

As many of us will already have analysis systems, I think the key additional contribution that a notebook can make is in documenting other factors not usually captured. Without documentation, it is hard to recall accurately the influence of the circumstances at the time and context of a decision.

We need also to record at the time the outcome we expected. What was the exit strategy? Traditionally, analysing habits involves looking at time, place, emotional state, presence of others and preceding action/catalysts. Some of this may be relevant. For example, I have found decisions taken outside the office are generally poor and are usually reactive with a negative catalyst. In terms of time of decisions, some may conclude that a rush to make decisions before market opening or in reaction to early trading, might be poor.

For investment decisions, I think it is good to capture some of the contrary information at the time and an impression of where the decision stands against the consensus. Whether a manager feels under pressure at the time, or what the background of overall performance is might also be worth noting.

And, I think one of the key factors to capture is what were the alternatives at the time? Even that a decision might have been delayed to capture further information is worth noting and can allow a pattern to be later identified. We are more willing to accept our decisions might be mistimed rather than plain wrong.

Through this, we should be able to identify the patterns that are supportive of good decision-making against poor. We should not, however, always allow the fact that history gives us the luxury of perfect knowledge of the result, to always give a binary verdict of right or wrong on past decisions. This is a failure of many accident enquiries, which tend to focus on a simplistic category of human error rather than trying to understand why it made sense at the time for people to do what they did. It is important we do not oversimplify past choices, but understand the context.

What might come out of record keeping?

Well, I think we can derive our own personal rules and checklists, which are covered separately by Joachim. We are inclined to regret, but it can reinforce learning and a notebook is the way to learn the right things, particularly when emotion is part of judgement.

I do not believe that emotion need always mean poor decisions. Emotion may capture subtle understandings of past events and decisions and therefore be an inevitable part of our decision making, sometimes helpful. We just need a good tool to appraise this mix of emotion and progression in our past decision-making.

The key constituents of hindsight bias are the impression of the inevitability of outcomes, exacerbated if we can pin things on a salient cause. And, the less surprised we are by the outcome, the more we view the result as foreseeable. And, our memory is distorted as to timing and the actual events themselves. 

Helpfully, research points to the value of spending more time in studying the record. Recall typically improves as we gather facts, and the notebook can present those. The process of making and reviewing the record addresses each of these issues in hindsight. And, for those who can continue to make the effort to learn, there is an additional benefit. The process of gaining new knowledge tends to make us revise down our belief in our past wisdom. So there are practical conclusions from behavioural finance.

No comments yet... Be the first to leave a reply!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: