Koen Smets has published a blog post where he uses the Brexit referendum as an example and behavioral economics as a toolbox to question the rationality of referenda. He uses the standard catalogue of decision-making biases to make his case, and prima facie, he does so quite convincingly. But I think there are some problems in his line of argument.
First of all, he is obviously against Brexit. That is not a problem in itself; it is a perfectly legitimate political position. But some of Smets’ arguments, and arguments by Daniel Kahneman he cites approvingly, seem to start from the implicit premise that a vote for stay is the rational option, while a vote for leave is probably the result of arguments that “look short-term” and are “based on irritation and anger.” That is a pattern of debate which is also often observable when Swiss referenda are discussed in the German media: There is a choice that the foreign observer considers rational, but the stubborn Swiss will simply not comply. And with these seemingly irrational direct-democratic decisions we repeatedly observe, we should better protect our hyper-rational German democracy from the threat of direct-democratic participation.
The problem is, though, that a few years later we often (if we bother to have a second look) find out that this strange and incomprehensible Swiss referendum decision was actually far-sighted and led to perfectly reasonable economic and political outcomes. In these instances, it is the elitist bias that leads us to bad judgments. Thus, we should not judge direct democracy based on whether we like the result in one concrete case, but based on the process in general. Will decision-making biases be at work within the electorate? Obviously. The relevant question, however, is whether these biases will be worse in a referendum process than in a purely representative democracy.
Look, for example, at the argument that members of parliament are better-informed than the ordinary voter. It looks like a straightforward point, because our elected representatives are people who specialize on political decision-making. But in fact, they are often forced to specialize on very narrow fields of policies, and know virtually nothing about other fields. During the worst period of the Euro crisis in 2011, the very reputable political magazine Panorama asked federal MPs in Germany some simple questions on facts related to the crisis. The vast majority commanded an embarrassingly low level of knowledge about the (huge and important!) issue they were just about to cast a historic vote on, and most of them would simply go on to vote the way suggested by their party’s whip.
Such a high degree of specialization is not necessarily a good thing. It implies that compromises are found and politically enforced by few individuals who are highly invested in an issue, in a political sense. And that sets the stage for all the Public Choice problems of representative decision-making which Koen Smets so conveniently ignores in his line of argument. Log-rolling, rent-seeking, unchecked bureaucrats unduly extending their power – you name it. In fact, are these not exactly the problems that triggered the Brexit campaign in the first place? And how do you plan to deal with them in the absence of a referendum?
On the other hand, there is good empirical evidence showing that the degree of voter information is endogeneous to political institutions. In particular, people get better-informed if they live in an institutional framework that allows for referenda regularly. Mind you, this is not about a one-off referendum every now and then, but this argument is about a political culture where direct-democratic participation is a normal part of the political process. Let people get used to have a say, and they will rise to the occasion.
Representative politics without direct-democratic checks, on the other hand, is incomplete. Citizens need an instrument that allows them to signal their discontent with representative policy-making on single, specific issues. This, rather than the opportunity to oust a government every four or five years, is the instrument that allows citizens to break up coalitions of special interests, and to increase the efficiency of representative policy-making itself.
What about some of the other biases introduced into the discussion by Koen Smets? He mentions that voters may suffer from hyperbolic discounting, and therefore neglect long-term costs of their political choices. They certainly might, but it is a very bold, and frankly unsubstantiated claim that representatives whose personal time-horizon is always reset at the next election are less myopic. Smets also mentions the fallacy of appeal to authority: Some people may have voted for a Brexit, because they trust the figureheads of the Brexit campaign, rather than because they checked their arguments and found them convincing. Yes, that may be the case. But is this not what representative democracy is all about? Finding some candidate who seems trustworthy enough to represent a voter on a whole range of utterly diverse issues seems to be the essence of representative democracy. How that is supposed to redeem the appeal to authority fallacy is left in the dark.
And then there is the talk about the moderating effect of representative democracy, about long searches for compromise behind closed doors in parliamentry committees. Indeed, this is a part of representative policy-making, although the usual Public Choice caveats again apply. But, once more, we should have a look at the Swiss experience with direct democracy. The political discourse preceding important referenda usually is of a quality that, viewed from north of the border, is reason enough to envy the Swiss. Even in the UK, where referenda are not part of the everyday political culture, the Brexit issue has spawned debates of a quality that is, to be honest, unheard of in the rather dismal and shallow political culture of today’s, purely representative, federal-level system in Germany.
If you want to use behavioral political economy to denounce direct democracy, something more is required than throwing a bunch of biases at the issue and seeing if something sticks. All policy-making is subject to biases. So where is the sound comparative analysis that supports the claim that purely representative systems actually work better?
The above part of this blog post was written on the evening of the British referendum, before the result was published. Does my line of argument change after getting to know that Britons voted for leaving the EU? Not at all. The short-term economic outlook for Britain is now going to be uncertain, and the country is facing a bumpy road ahead economically. But that is not all. If your political preferences are such that you favor national-level sovereignty over distant power in Brussels, that is a price you might rationally find worth paying.
As far as the long-term prospects are concerned – let’s simply wait and see how the United Kingdom fares economically ten years from now. In good British tradition, I’m willing to offer Koen Smets a wager: Unemployment will not be more than one percentage point higher in ten years in the UK than it is today, and real wages will not be lower in ten years than they are now, and Britain will be neither politically nor economically isolated from the EU or the rest of the world. So, I’m willing to bet than none of the apocalyptic predictions of the remain campaign will come true, and I’m even willing to bet that in ten years, a clear majority of Britons will in fact be pleased with yesterday’s result.
Apart from that, I suppose even the rest of the EU will be grateful. Not because they got rid of the UK, but because now it is clear that the EU might in fact dissolve if it does not thoroughly reform itself, and become much more modest in its claims to power and competencies.