What happens when we pay MPs more?

[Confidence: Fairly low, and given the amount of literature, also sceptical of anyone who is particularly confident on this question]

If you follow British politics, you already know about the details of MPs earning vast sums of money from second jobs that has been dominating the headlines in the past week (few weeks?). I won’t bore you by repeating the details: Owen Paterson, Geoffrey Cox, you know the score. The second jobs scandal has prompted a second debate (one that seems to happen maybe once a year or once every two years on Twitter): should MPs be paid more? There are two distinct arguments here: the first argument is that MPs who are paid more are less likely to take second jobs that distract from their work as an MP, or that lead them to behave badly as an MP (promoting the interests of the company they work for a la Owen Paterson). The second argument is that MPs who are paid more will generally be higher quality MPs - both through selection effects (you get better candidates if you pay more money) and because better paid MPs might work harder. To cover both of these arguments is probably too much for one post, so I’m just going to focus on the argument about higher quality MPs for now, rather than MPs who take second jobs (even though it is less relevant to the news cycle at the moment).

There are a few decent studies on the effect of increasing wages on the quality and performance of politicians - the first one worth mentioning is a difference-in-differences analysis looking at candidates in Finland that exploits a 35% increase in the salaries of parliamentarians, using municipal candidates as a control group. We might quibble with the measure of MP quality (they use education levels as a simple metric to gauge the quality of a parliamentarian), but the research design is solid. The finding here is pretty interesting: bumping up the salaries of MPs increased the quality of female parliamentary candidates, but not male parliamentary candidates. The primary explanation offered is the differing position of men and women in the labour market - women have relatively lower paid positions to men and so the increase in salary for an MP is more attractive to women than it is to men. The other point they make is that highly educated women are more likely to work in the public sector, and so the wage increase for MPs would have made it competitive with other jobs they would be likely to be doing, whereas highly educated men are disproportionately in the private sector, and the pay increase for MPs is somewhat irrelevant for men in the private sector aspiring to earning much higher wages than are possible for MPs to ever earn.

There are other studies examining how candidate quality changes in response to pay increases - Fisman et al (2015) find that increasing the salary of MEPs actually leads to less well-educated candidates standing for election (as well as showing that MEPs who attended a top university are more productive in parliament). Their study suggests that doubling the salaries of MEPs would lead to a 15% reduction in the number of MEPs who attended a top university (decline from 30% to 25.5%). They write:

This pattern is consistent with the view that higher-quality MEPs (who have better outside options in the market sector) are motivated to hold office due to nonpecuniary returns whereas lower-quality MEPs (whose salaries in the market sector would be lower) respond more to salary when deciding whether to run for political office.

This is pretty interesting to me - I’ve heard people make the argument before that the sorts of people we want in parliament aren’t those who are motivated by money, they’re people who are really committed to public service. I’m pretty sure I said the same thing myself when I was first becoming interested in politics and was instinctively opposed to the idea of MPs being paid more in the midst of a very lacklustre recovery. In the past few years though, I’ve been unfairly dismissive of this argument - my assumption was a (seemingly pretty naïve) rational-choice model wherein candidates respond to incentives, yada yada yada. I’m now more sceptical of the crude incentive model of MP quality.

The other side of the literature, studying how pay increases affect the performance of MPs (as opposed to the quality of the candidates standing to become MPs) is possibly more interesting. It seems like there is some evidence that raising the pay of legislators makes their performance worse - or at the very least does nothing to improve their performance. Let’s start by looking at another difference-in-differences analysis, using a change in legislation in Turkey that led to some MPs getting a pay rise while others received the same pay as before. The effect of the pay increase on performance (measured using a composite of metrics - number of speeches in parliament, number of reports written, and so on) was negative, the effect size being 12.3% of a standard deviation. The graph above shows the difference, with the dotted line showing MPs who did not receive the pay increase and the solid line showing MPs who did, where 2012m1 is the point at which the increase came into effect. Fisman et al (2015) also write about how pay increases affect MEP performance, using the same dataset of MEPs we looked at earlier - their finding is that MEP salary basically has no effect on MEP performance - they don’t attend more/less after a change in their salary, they aren’t more likely to shirk (meaning attending parliament just to sign the attendance register, and then immediately leaving), and so on.

I want to pull back a minute here - while I think the evidence so far is that the effects of increases in MP pay are probably either sort of neutral (on both candidate quality and MP performance) or slightly negative, I think the number of good studies we have on this is low enough that we should be unsure about making any substantive judgement. To add, I’m not really convinced that the external validity of these studies is particularly high (i.e. I’m not sure studies about MPs in Finland, MPs in Turkey, or MEPs allow us to make really meaningful inferences about MPs in the UK or elsewhere), even if the internal validity and research designs are solid. Similarly, the pecuniary cost of increasing MP pay is really low (increasing the pay of MPs by 50% would cost about £26m), so I think we shouldn’t think of that as a barrier to increasing MP pay if new research shows that it might make a difference. But for now, I don’t think the evidence in favour of increasing MP pay is particularly good.