Confirmation bias as a feature, not a bug: The facts don’t really matter
Reasoning was not designed to pursue the truth. Reasoning was designed by evolution to help us win arguments. That’s why they call it The Argumentative Theory of Reasoning. So, as they put it, "The evidence reviewed here shows not only that reasoning falls quite short of reliably delivering rational beliefs and rational decisions. It may even be, in a variety of cases, detrimental to rationality. Reasoning can lead to poor outcomes, not because humans are bad at it, but because they systematically strive for arguments that justify their beliefs or their actions. This explains the confirmation bias, motivated reasoning, and reason-based choice, among other things.
–Why Do Humans Reason? Arguments for an Argumentative Theory – Behavioral and Brain Sciences, Hugo Mercier and Dan Sperber.
I found this quote via a great article by Jonah Lehrer on the reason we reason. Read the whole thing. It’s a great piece. Here’s the thing, though.
Psychologists have shown that people have a very, very strong, robust confirmation bias. What this means is that when they have an idea, and they start to reason about that idea, they are going to mostly find arguments for their own idea. They’re going to come up with reasons why they’re right, they’re going to come up with justifications for their decisions. They’re not going to challenge themselves…
The idea here is that the confirmation bias is not a flaw of reasoning, it’s actually a feature. It is something that is built into reasoning; not because reasoning is flawed or because people are stupid, but because actually people are very good at reasoning — but they’re very good at reasoning for arguing.
On some level, this is quite a disturbing revelation. What Mercier is saying is that we use logic and reason not to prove anything but rather to argue our case better than others. Confirmation bias, where we look for data points to bolster our argument, is a feature not a bug. It helps us to win arguments irrespective of the facts on the ground. In essence, the facts don’t really matter.
Now, think about this in the context of finance, economics or politics. Is it really relevant whether your beliefs actually fit the data? This morning we got a pretty good establishment survey that recorded 268,000 private sector jobs and 244,000 jobs overall. Those are pretty good numbers. 244,000 was the most in five years. The market rallied strongly on the back of this apparently bullish report. Nevertheless, the unemployment rate rose to 9.0%, which is disconcerting. Having looked at the data, I wrote:
Note the items highlighted on the household survey. Labor participation is NOT increasing. Rather, there is a divergence between the household and establishment surveys. The household survey data are weak. The establishment data are much better. The increase in unemployment rate is being driven by new entrants to labour market according to household survey. Although not highlighted, note that the number of people counted as unemployed because of re-entry into the labour market has not increased. You can see this near the end of the graphic above the highlighted line for new entrants. This factoid tells me the labour market is still relatively weak because those who have dropped out of the labour force are not re-entering en masse yet. [emphasis added]
If you are of a bullish mindset, you ignore the household survey and run with the establishment survey. And that’s exactly what has happened. Thais doesn’t mean that the establishment survey is right and the household survey is wrong. It’s just that the prevailing wisdom is that the jobs picture is improving so confirmation bias tells us to ignore the household survey data.
On the politics front, there is even a philosophy of hypocrisy for why elites don’t practice what they preach. It rhymes with the confirmation-bias-as-a-feature theory. I wrote about this last month saying:
Let’s call it the hypocrisy of arrogance. As Newt Gingrich said to his wife when caught cheating, ""It doesn’t matter what I do. People need to hear what I have to say. There’s no one else who can say what I can say. It doesn’t matter what I live."…Humans prefer cockiness to expertise. They would rather someone be consistent and wrong than inconsistent but right.
My takeaway for the economics profession is that in the face of cognitive dissonance, you should expect economists to ignore negative data that disproves their pet theories. You should expect them to continue to argue using data they selectively cherry-pick using confirmation bias because the reason we reason is to win arguments. The facts don’t really matter. That is certainly what happens in politics. So I expect no different in economics.
My takeaway here for markets and finance goes back to why markets fail:
The psychology of prior price movements can dominate price activity because humans are social animals. When prices go up 5%, you could consider the actions of market participants random because the motivations and psychology of buyers and sellers cancel each other out. However, when prices go up 50%, this is no longer true. It is very important whether my neighbour is making a killing in the housing market or in Internet stocks. That is exactly how bubbles form. And the same is also true on the downside, the principal reason markets that crash tend to overshoot to the downside.
Great piece.
Number one thing I tell new folks to the capital markets: Watch out for two things: Confirmation bias, and the standardization of deviations. If you can avoid falling victim to either … you’ll do well. But be warned that it is literally … fighting the way that we are wired.
Great piece.
Number one thing I tell new folks to the capital markets: Watch out for two things: Confirmation bias, and the standardization of deviations. If you can avoid falling victim to either … you’ll do well. But be warned that it is literally … fighting the way that we are wired.
But remember this in the long run … the markets are not efficient … but ignoring data for too long that conflicts with your theory?
That’s how “2008’s” are born …
People say “stocks are risky.” Yes and no. They are risky if you suffer from confirmation bias. Because, if you suffer from confirmation bias, you are always doing the wrong thing.
But what if we openly talked about confirmation bias, as you do in this post? Then we would over time all become aware of its influence and become able to overcome it. The risk of stocks would disappear. We would be a far richer people.
Overcoming the economic crisis is precisely that easy.
And that difficult.
Super post.
Rob
People say “stocks are risky.” Yes and no. They are risky if you suffer from confirmation bias. Because, if you suffer from confirmation bias, you are always doing the wrong thing.
But what if we openly talked about confirmation bias, as you do in this post? Then we would over time all become aware of its influence and become able to overcome it. The risk of stocks would disappear. We would be a far richer people.
Overcoming the economic crisis is precisely that easy.
And that difficult.
Super post.
Rob
A companion “feature” to confirmation bias is what has been described as the backfire effect — a sort of disconfirmation bias. When presented with evidence that contradicts our beliefs, we expend disproportionate efforts at attacking it, discrediting its source, and often end up more convinced that our beliefs are correct. This type of argumentation was quite common in the midst of credit bubbles.There’s a whole deck of strategies that can be applied to dismiss any type of corrective facts.
The broader issue of winning arguments has surprisingly little to do with having facts on your side. Robert Cialdini’s, Psychology of Persausion, categorises the techniques used into six groups:
Reciprocation – make a small gesture that the other party feels they are obligated to return in kind. The “free” gift leads to all sorts of irrational decision making.
Consistency & commitment – small commitments can often be built upon toward a larger decision. 409 scams depend on this. People hold onto a shrinking investment far too long because of this.
Social proof – in an ambiguous situation, one looks to see what everyone else is doing. This can lead to highly correlated market behaviour.
Liking – we tend to agree with people we like. Charismatic people are highly persuasive.
Authority – appealing to authority is a common technique in debates. Having a Nobel Prize winner on your side is always a good thing. Each competing economic ideology has their own favoured authority figure.
Scarcity – I’ve lost count how many times the finite size of the British Isles has been used to justify why house prices must always go up.
You have really nailed it here. What surprises me is the degree to which economists are blind to these things. The magnitude of the crisis should have caused some major cognitive dissonance which caused them to decanter some previously held beliefs. But these economists speak with the same cerainty that they did before the crisis.
That the public doesn’t realize this is all confirmation and disconfirmation bias is astounding. I am continually Mazedonien at our collective ability to delude ourselves.
Since we weren’t alive during the Depression it’s hard to know if similar things happened. But my sense is it did despite the depression. It was only the magnitude of the downturn which caused the public to look elsewhere in desperation. So I would expect the same ideological battle lines to persist today until the public forces policy out of desperation. And that will take an economy which is much worse than today’s.
A companion “feature” to confirmation bias is what has been described as the backfire effect — a sort of disconfirmation bias. When presented with evidence that contradicts our beliefs, we expend disproportionate efforts at attacking it, discrediting its source, and often end up more convinced that our beliefs are correct. This type of argumentation was quite common in the midst of credit bubbles.There’s a whole deck of strategies that can be applied to dismiss any type of corrective facts.
The broader issue of winning arguments has surprisingly little to do with having facts on your side. Robert Cialdini’s, Psychology of Persausion, categorises the techniques used into six groups:
Reciprocation – make a small gesture that the other party feels they are obligated to return in kind. The “free” gift leads to all sorts of irrational decision making.
Consistency & commitment – small commitments can often be built upon toward a larger decision. 409 scams depend on this. People hold onto a shrinking investment far too long because of this.
Social proof – in an ambiguous situation, one looks to see what everyone else is doing. This can lead to highly correlated market behaviour.
Liking – we tend to agree with people we like. Charismatic people are highly persuasive.
Authority – appealing to authority is a common technique in debates. Having a Nobel Prize winner on your side is always a good thing. Each competing economic ideology has their own favoured authority figure.
Scarcity – I’ve lost count how many times the finite size of the British Isles has been used to justify why house prices must always go up.
You have really nailed it here. What surprises me is the degree to which economists are blind to these things. The magnitude of the crisis should have caused some major cognitive dissonance which caused them to decanter some previously held beliefs. But these economists speak with the same cerainty that they did before the crisis.
That the public doesn’t realize this is all confirmation and disconfirmation bias is astounding. I am continually Mazedonien at our collective ability to delude ourselves.
Since we weren’t alive during the Depression it’s hard to know if similar things happened. But my sense is it did despite the depression. It was only the magnitude of the downturn which caused the public to look elsewhere in desperation. So I would expect the same ideological battle lines to persist today until the public forces policy out of desperation. And that will take an economy which is much worse than today’s.