Irrational voters

Sevenscore and ten years ago today, Abraham Lincoln coined the phrase "government of the people, by the people, for the people." He omitted to add that the people can be systematically wrong, as a new paper neatly shows.

Michela Redoano and colleagues estimate that, in the UK, women whose husbands died in the previous two years are 10-12% less likely to vote for the government than other women. There's a simple reason for this. The less happy people are, the more likely they are to vote for opposition parties; this is a variant of the affect heuristic. But people don't distinguish fully between being unhappy and being unhappy because of bad government. As a result, they - in effect - blame the government for things it is not responsible for.

This is not an isolated finding. Christopher Achen and Larry Bartels have found (pdf) that voters, in effect, blame governments for things they cannot control such as natural disasters. And Neil Malhotra and colleagues show that surprise victories for the local team in US college football games increase support for incumbents in gubernatorial, presidential and senatorial elections. This is consistent with the fact that Harold Wilson blamed Peter Bonetti for losing him the 1970 general election.

I suspect that UK governments have long tried to exploit this misattribution effect.One reason why they prefer to hold elections in the spring is that lighter nights improve our mood, which makes us better inclined towards incumbents. It's no accident that the preferred month to hold general elections - May - is also the month in which share prices often peak.

You might object that the bias from this source is small. Maybe. But if voters are irrational in this respect, isn't it likely they'll be irrational in others. It's insufficiently appreciated that the problem with democracy is not just the shortcomings of our politicians, but those of the voters too.

Comments

[FromArseToElbow]

"The problem with democracy". Hmmm. The trouble with this line of thinking is that it assumes people can make "wrong" choices, whether simply ill-informed or arbitrary. But who is to decide what "wrong" is? Perhaps we should have the people vote on it.

As for the shortcomings of the voters being a characteristic of democracy, I rather thought that was a feature, not a bug.

[Left&rational]

I would have thought that spouse death would retard turnout. Depression depresses activity. If less turnout, then less vote for the incumbent.

The authors should also report the effect on %vote for the opposition. If spouse death has a positive affect then I buy their story. My bet is that it would be equally negative.

[Rob]

"As a result, they - in effect - blame the government for things it is not responsible for."

Presumably it follows that they also credit the government for things it is not responsible for, and such erroneous judgements are likely to cancel each other out, wisdom-of-crowds-style. Only if a government presided over a significant run of bad luck affecting a substantial number of people would the effect be large enough to swing an election, and at this point you might be justified in doubting the government's claim that it's all down to bad luck.

Part of this comes from the sheer scope and ambition of government too - governments seem to want to be judged on the "what have you done for me lately?" principle, which is why they try to bribe us with tax cuts or spending increases. Governments which think that they should be responsible for "happiness" can hardly complain when they're voted out because people aren't happy.

One interesting example is local council elections, where it's a commonplace that hard-working local councillors are routinely thrown out of office once their party achieves power at Westminster, their local role to be taken by a donkey in a correctly-coloured rosette, an outcome that seems highly irrational. In this case it's not so much that voters are irrational as that they value "sending a message" more than they value having competent local councillors. In fact, given the centralised nature of British government, it might actually be a rational strategy to engage in ritual sacrifice of one's local government officers in the hope of attracting the attention of those at the real levers of power. We should be wary of calling something irrational unless we've considered all of the ways it might turn out to be rational after all.

[Robertnielsen21.wordpress.com]

I think you're right about the timing of elections. Governments tend to prefer to hold elections in summer and close to the weekend too.

[Phil Beesley]

Apologies for length.

Sometimes a banking collapse, or of a different sort of complex system, occurs. Individuals who contributed to it (eg Fred the Shred) may be identified alongside a class of people ("bankers"). In other words, blame is placed on specific people *and* on an ill defined group. But it is casual attribution of responsibility.

Owing to complexity of the system, survivors and bystanders cannot accept responsibility for failure. No single person can understand the system, so nobody feels culpable.

In a system where everyone knows how it works, passing the buck up and down the chain is more difficult. The point where the system went wrong can be identified; it might be the point where a rational person cuts the head from the body, necking off a lot of senior management.

But that isn't going to work most* of the time. Nobody knows how the system works -- some processes are like magic -- and everyone prays that the failure point occurred in something obscure or in something common that happened at a competitor.

* Apply perceptual bias algorithm at this point.

Construction of a space shuttle is beyond the comprehension of one person and there aren't many Isambard Kingdom Brunels to work on more modest projects. Every engineer wishes to work with a Brunel, which is a damned good thing because it means that they wish to be responsible for something great. Failure/inability to work with a great mentor partially explains why so many engineers are grumpy.

People do not trust systems because they know that there aren't many Brunels. They'd trust more if the people who knew stuff -- the sub-Brunels -- appeared to make more decisions. Most of the time, people trust sub-Brunels to get them to work or to live 17 levels above the ground. Even if they don't know it.

And we don't/can't simplicate systems enough for one or two people to be in control. Even if we simplicated everything, top bods could not prevent a train crash; train drivers, and track and signal operators, do that.

Professional engineers, Network Rail op techs, rail gangs: they set the rails on which you to travel to London and upon which you are unlikely to die. Engineering removes a lot of randomness.

[Phil Beesley]

Robert Nielsen: "I think you're right about the timing of elections. Governments tend to prefer to hold elections in summer and close to the weekend too."

Elections on a Thursday? It is a boring day, I agree, for us UK folks. Wednesday is in the middle of the week, which is when we might conduct inconsequential affairs. Like voting.

I do not know much about sweaty summers, other than avoidance. And everything that I know about politicians is that they do not suffer UK warmth.

UK elections for local government are conducted in the UK spring. It is by decree.

[Metatone]

Worth noting of course that up to the present day, many women who are widowed are suddenly confronted with the fact that many of our governmental systems are callous and still biased against women. Perhaps the various ways in which widows get a raw deal out of the legal, governmental and benefits systems affects their voting patterns?