PsyFi Search

Wednesday 26 October 2011

One Long Argument: A Big List of Behavioural Biases

Over the hundreds of articles published on the Psy-Fi Blog we’ve generated a long list of examples of behavioral biases, which is now collated in The Big List of Behavioral Biases.  It’s an extraordinary compendium of irrationality, an imperfect register of the various strange ways in which rational economics goes astray.  Ultimately, though, what we have here is really just one long argument.

It’s a formidably long list already, and we haven't got to the bottom of the ever-increasing range of peculiar behaviors that we exhibit when placed in the appropriate situation.  Unfortunately, the longer the list becomes the more perplexing becomes the underlying issue: after all, many of these biases could easily conflict with each other, dependent on the situation we find ourselves in.  Some researchers have started trying to understand this by looking at neurology and so called heuristics and biases models, but it's clear that we're still a long way from any real understanding of the mechanisms involved.

Certainly some of these issues seem  to be implicated in lots of others.  Anchoring is a prime example which appears to colour people’s perceptions in all sorts of different contexts: and context is another, through the mechanism of framing, which creates the context within which we make a decision.  Some biases appear to be deeply rooted in prior beliefs, such as confirmation bias, and others are almost impervious to any attempt to change them, such as hindsight bias

Other errors are straightforward logic problems – such as the infamous Linda problem, otherwise known as the conjunction fallacy in which the combination of two events is rated as more likely than a single event.  It’s statistically impossible for a person to be more likely to be a librarian and a feminist activist than just a librarian, no matter how carefully framed is the question: but many studies have show that’s what many people believe.  There’s a clue, however, that part of the problem with all of these biases is the experimental framework within which they’re conducted.  Whether we like it or not we make inferences from context, and if that inference appears to point in the direction of Linda the librarian being a raging feminist then we’ll make that logical leap, no matter how actually illogical it is.

Of all the biases listed here, though, the one that seems to have the most power to cause enduring problems for investors is overconfidence.  It may be simply that we need to be optimistic about the future in order to keep going, but whatever the reason we seem to operate on the basis that things will eventually turn out OK, even if we can’t foresee how that will happen.  It’s not just pure overconfidence that’s a problem: we have an illusion of control of situations when we don’t and we believe that our successes are down to our own talents and our reverses to the quirks of luck: the fundamental attribution error.  A few of us never learn anything, no matter how much the evidence mounts that we’re incompetent: the Dunning-Kruger effect.

We’re also gullible: we believe gurus who tell tales using specific examples and ignoring vast oceans of contradictory evidence – the so-called Texan sharpshooter effect – and we’re fooled by randomness by skilled con-artists who can “prove” anything they want by data mining: such that the movements of the S&P500 are predicted by the price of Bangladeshi butter or the likelihood of finding buried treasure in the UK.  In addition data alone can fool us: we think that random sequences of good or bad results must mean revert (the Gambler's  Fallacy) and we anchor ourselves on round numbers (round number effects), and let ourselves be gulled by market professionals exploiting our weaknesses. Oh, and simply exposing us to something can make us more inclined to like it: the mere exposure effect.

Sometimes we fool ourselves – there are whole class of self-fulfilling biases which cause things to happen because we expect them to: that's almost a definition of markets, in fact.  So there’s the Titanic Effect, where we build systems to fail in the ways we expect and not in the ways we don’t expect, and the Clever Hans Effect, where we can convince ourselves a horse can mind-read, and the associated Placebo Effect, where we can make ourselves well – or ill – through the power of thought.  Or there are biases caused because we’re attracted to certain things – there's the Beauty Effect where we think attractive people are more intelligent and we like to associate ourselves with success, even if that aura of success is simply created by someone renaming their corporation, an issue known as the Halo Effect.

Then there are are straightforwardly social effects – like the problem of bystander apathy where everyone leaves helping someone in trouble to each other, and commitment bias, where publically putting ourselves on the lines makes us extremely resistant to changing our minds, even when the evidence is overwhelming.  Associated with this is herding, where people group together and take their cues from each other, rather than trying to think for themselves.  This is particularly common in times of economic uncertainty when no-one really knows what’s going to happen next, so everyone just follows everyone else, in a dizzying spiral of inconsistent decision making.

We also suffer from cognitive overload, where we simply can’t process the information given to us. Choice overload is one such issue, where we’re more likely to make a decision when presented with a smaller number of options.  Procrastination is a related problem – where we simply fail to make any decision at all because it’s all a bit too difficult.  These issues are related to self-control, where we actually are unable to control aspects of our behaviour – such as spending now on the credit cards for some immediate burst of pleasure, because the pain of paying can be put off to the future. 

Loss aversion is a critical issue for investors, because the overwhelming desire to avoid booking losses means we sell winners and keep losers in order to avoid regret with the general result that we’re less successful investors than would otherwise be the case.  Mental accounting is a key concept here: this sees us assigning money to separate pots and then juggling them frantically to make it look as though we’re not actually losing money.  If an investor needs to learn one thing about this mess of mental manipulation it’s this: learn to cut your losses.

Perhaps, though, the biggest issue of all is that we will forget the past just as soon as it’s not in our face reinforcing its lessons.  This is disaster myopia, and our failure to remember the past is the reason why historians are amongst the most important people in our societies.  Those amongst them capable of showing us that we’re no different from our ancestors, that we’re still the same cognitively challenged, behaviorally biased, emotionally compromised humans as ever we were do us all a favour. 

In the end, though, this e-rag is just one long argument in favour of people learning self-control.  And, as ever, the first step to achieving that is to realise it's necessary in the first place ... See: The Big List of Behavioral Biases


  1. The link to Big List of Behavioral Biases is not working. ( Please check. Thank you :)

  2. Seems to have moved:

  3. This comment has been removed by the author.

  4. It was very useful for my report. Thank you for it.