Don’t stay up and decide: sleep deprivation and the culture of late night summits

Would you trust a minister of finance explaining how he just fixed the latest euro-zone deal if he came out of the summit chambers tipsily waving a glass of wine? No? What about if he gave a press conference after an all-night session? Most likely nobody would even notice.

Yet 24 hours without sleep has (roughly) the same effect on decision-making as a 0.1% blood alcohol content (six glasses of wine in an hour). You would not be allowed to drive at this alcohol level, but you are apparently allowed to make major political decisions.

The example is from a blog essay (in Swedish) by Andreas Cervenka, where he asks the sensible question: can we trust sleep-deprived political leaders?

That sleep deprivation is bad is well-known: it causes lapses in attention and working memory, and sleep-deprived people are bad at evaluating just how impaired they are. It causes thinking to become rigid and less adaptive, changes how risk is perceived, makes people less trusting and more emotional in economic games, lowers emotional intelligence while increasing magical thinking. Just the kind of thing you do not want at a summit that will affect the lives of hundreds of million people, involving complex issues and participants from different cultures.

Yet as Cervenka points out, in politics and big business late night meetings are common and not just accepted, but even revered: they signal commitment. Stories abound about the brief sleep needs of great people: their somewhat iffy veracity does not matter in the mythology of power. In some situations there might be a competition element: you are not going to show weakness by adjourning, right?

Leaving aside direct crises it seems that this pattern of decisionmaking under the influence is bad for everyone. It harms the decisionmakers, and it has a high chance of stalling negotiations and producing policies that are irrationally risk-taking, yet rigid. While the signalling, competitive and cultural aspects of all-nighters might be hard to deal with on an individual basis, it is possible to coordinate rules like EU directive 2003/88/EC on working time that would ban such practices without anybody losing face. However, the fact that the top EU decisionmakers do not follow their own occupational safety rules is a bad sign…

Many safety critical jobs have zero tolerance or strict policies for alcohol or drug influence on the job: the harm from mistakes could be large. Typical examples are pilots, offshore platform workers, nuclear plant workers and soldiers. Similarly truck drivers and some other occupations have rules mandating getting enough sleep and banning the use of heavy equipment when sleepy. Of course, that doesn’t mean they are free from drug use or sleepiness, but there is a general recognition that they better be at their best when on the job.

However, it seems that the demand for peak performance is mainly tied to direct risk. An engineer designing a machine is potentially endangering far more people through an unsafe design than an ever so clumsy pilot. The causal distance seem to make us less likely to see the link between the impairment and the eventual harm. Yet industrial disasters, security breaches, health problems and economic losses due to remote mistakes or stupidities probably outweigh the harm from direct mistakes enormously. I suspect this distance is one reason sleep deprived decisionmaking is not seen as a major problem by the public.

Another reason is that indirect decisions and actions tend to be collective. Bad choices can be corrected by subsequent layers of people. If you are an optimist, you could say that maybe the summits are just for show: the real decisions have already been hashed out by other layers of the bureaucracy. But in that case they are not necessary.

A few years ago I was planning a paper on whether political decisionmakers ought to take cognitive enhancer drugs. The case looked pretty clear: they suffer sleep deprivation, they have high stress jobs with big cognitive demands, and their decisions matter. After a week in the Houses of Parliament I quietly scrapped the paper. I had realized just how little effect enhancing individual MPs would have: while no doubt good for them, the eventual decisions do not come about due to a single or few bright minds deciding, but through a formal collective process where collective decisionmaking happens in layer after layer. Smarter members does not mean the group becomes smarter, especially since groups have their own cognitive biases. If we want to enhance the political decisionmaking process we should fix how people work together: reducing groupthink, in-group bias and polarization - for example by deliberate rules of procedure. That ought to include a ban on late night negotiation.

So my recommendation to decisionmakers would be to  obey 2003/88/EC: you are not above it, it is for your own sake, and it will actually improve your ability to reach your goals. Sleep deprived negotiation is irresponsible and harmful. When some party tries to mess with the agenda in the hope of coming out on top because of exhaustion, refuse to play the game: everybody loses if joint decisions are of low quality. If you cannot avoid the late meeting, at least take some modafinil – and make sure to pass it around the table.

My recommendation to voters, shareholders and pundits is to call out politicians and businessmen on their meeting habits: if you wouldn’t drink and negotiate, why stay up late and negotiate?

  • Facebook
  • Twitter
  • Reddit

5 Responses to Don’t stay up and decide: sleep deprivation and the culture of late night summits

  • gwern says:

    If summits aren’t about deep intellectual thought and analysis (which will have been done beforehand by flunkies and academics), but about clashing power and manipulation and wearing down the other party to secure oneself the best deal possible, then the effects of sleep deprivation on the opposing side are not a bug but a feature.

    • Anders Sandberg says:

      But around half of the delegates are going to be less good at the wearing down bit: unless you are overconfident (which I think is common – these are the people who typically dominate other groups) you should assume you have no certain advantage, and may instead be victimized by the others. So looking for other strategies for winning makes sense. Especially when the situation is not a zero-sum game.

      • gwern says:

        > you should assume you have no certain advantage, and may instead be victimized by the others. So looking for other strategies for winning makes sense.

        Only if you think that the endurance contest is not a Nash equilibrium (and given the commoness of it, there must be *something* going on); perhaps rotating out fresh conference members forfeits more social connections or respect than it gains in intellectual horsepower.

  • “My recommendation to voters, shareholders and pundits is to call out politicians and businessmen on their meeting habits: if you wouldn’t drink and negotiate, why stay up late and negotiate?”

    Because pulling an all-nighter is a symbolic action. It shows that you are working really hard on the problem, even if the opportunities for success are non-present. Observe the mammoth-size various UN conferences of the parties. Most often, the conference would be a slim affair by involving the actors with the opportunity to really influence the issue. Inviting 200+ national delegation with NGOs magnifies the complexity while adding little. In a complex environment effort shows the image of commitment and foresight.

    A comparison with productivity in the workforce might shed further light on the issue. In Sweden, labour productivity has more than doubled since 1970. But still today the workforce is putting in the same amount of hours. New, more complex tasks and entire lines of professions have been added, but partially the explanation is that we have added more “empty work” à la Dilbert.

    • Anders Sandberg says:

      Yes, I think you are right about the symbolic function and its powerful draw on this kind of people.

      So fighting the symbolic/social role is important: the best way of disrupting the behaviour is to make it uncool. That is the essence of my call out proposal: make it embarrassing to do this kind of things. “Hey, finance minister! Can’t your staff do a schedule?!” “Hey, general-director, why are you wasting money and time when you are trying to safe the Earth?!” “Have you heard that the parties did an all-nighter again? It sounds like they are not serious about fixing unemployment.”