The economics of the politics of the arts

The standard approach to economic analysis of arts and culture is built on the theory of market failure and the idea that arts and culture are a public good. In essence, if left to the market, too little art and culture will be produced from the perspective of aggregate social welfare. Accordingly, optimal arts and cultural policy will seek public support to correct this market failure.

But the market failure/public good approach is, in economic theory, met with critique from what is called public choice theory. They point out that the above argument contains a realistic model of markets, but an idealised model of politics. In essence, they argue that while markets may fail, governments usually fail even worse, and for reasons largely associated with the information and incentive mechanisms of democratic institutions.

Public choice economists study three interrelated problems with modern democracy. First, majorities can exploit minorities (the “tyranny of the majority”). Second, organised minorities can exploit disorganised majorities (“rational ignorance” and “collective action” problems). And three, voting is irrational (costless “expressive voting”, weak incentives to information gathering).

You can see this in arts policy when organised minorities (such as any arts lobby) exploit disorganised majorities (e.g. taxpayers). Or when weak majority preferences defeat strong minority preferences (e.g. suppression of queer art). Or when voting becomes virtue signalling rather than honest preference revelation (e.g. inner city support for the ABC).

But a particularly brilliant young economist called Glen Weyl has come up with a way to fix all of this. It’s called quadratic voting, and you can read about it here, here, here in the context of corporate democracy, and in a particularly good essay discussing it in the context of ancient and modern democracy here. Some scepticism here.

Quadratic voting gets at the idea that intensity of preferences are not counted in normal voting. An indifferent majority can outvote a passionate minority (same-sex marriage is a good example) causing relatively small gains to the winners but imposing very large costs on the losers.

In normal voting, one person gets one vote (1p1v), and the price of the vote is zero. The problem is that someone who cares a little has the same voting power as someone who cares a lot.

Quadratic voting is the idea to modify a simple majoritarian framework to allow vote buying according to a quadratic rule. The total pool is then redistributed pro rata. There are two radical ideas in there: vote buying (and redistribution) and the quadratic rule.

A quadratic rule is that the price of a vote (under quadratic voting, you pay to vote) increases as the square of the number of votes. So if one vote costs A$1, 10 votes costs A$100, and 100 votes costs A$10,000. If I care a lot about an issue, I can influence a vote by paying a lot. The quadratic rule ensures that I can’t go too far with that, and that the size of the pool for redistribution also grows proportionately.

Vote buying is perhaps the one that people (who are not economists) get stuck on the most. Giving everyone only one vote, and making it free sounds fair, but it is economically inefficient. Any good that is free suffers tragedy of the commons, and voting is no exception. Because you have almost no power over what you end up with, you have little incentive to gather useful information or to vote your true preferences.

The great virtue of quadratic voting is that it incentivises honesty about intensity of preference. It was actually discovered not in retail politics, but in honey bees (bees ‘vote’ to make collective decisions following a quadratic rule on the waggle dance). The mathematics of the idea are well worked out and show that it is an efficient and equitable collective decision mechanism.

What would quadratic voting look like if applied in the arts? Consider two hypothetical referenda: one to privatise the ABC, another to expand the funding and scope of the Australia Council for the Arts.

At the core of the politics of arts funding is the critical claim that it is just middle-class welfare (a majority exploiting a minority), or an elite imposing their tastes on the masses (an organised minority exploiting a disorganised majority). Think the ABC in the first instance, and the Australia Council in the second.

The Australia Council’s clients are a relatively small number of people and supporters, with elite tastes, who care a lot about an outcome. Against them are a majority of voters for whom this is a minor issue, yet they can be whipped into outrage. However, on the back of this play out Australia’s culture wars.

What we want to know is whether a mostly indifferent majority is actually harming a passionate minority by withholding further support in a way that in aggregate harms society, or whether it is the organised minority that is doing the exploiting, making this a pure and therefore socially costly transfer from the poor to the rich. It’s hard to tell who is actually harming who here.

Quadratic voting might fix this by subjecting an intensity-scaled measure of these minority preferences, which would ideally draw upon wealthy patrons, to a straight majority vote. A quadratic vote mechanism then redistributes the total vote sum pro rata, potentially compensating those harmed by a win in such outcomes (e.g. taxpayers who are non-supporters of the arts).

That would probably be a better mechanism than the enormously politically charged and divisive process we currently have, as built on a one-person-one-vote mechanism.

A referendum on privatising the ABC could also benefit from a quadratic vote. What we seem to have is a smallish number who passionately support the ABC, a small number who passionately oppose it, and a large number for whom it is really hard to tell whether they actually support it in a willingness-to-pay sense, or because it is a nice thing to say to someone doing a survey, or whether they mildly oppose it but don’t really care enough to get upset about it.

No 1p1v majority vote will never answer that question, but a quadratic vote would. And whatever way the result broke, it would reflect a genuine social welfare maximising outcome. And that would be a lot better than where we are now.

Disclosure

Jason Potts receives funding from the Australian Research Council. He is affiliated with the Institute of Public Affairs.