Like me, Yale Law School Professor Stephen Carter rejects proposals for mandatory voting. As an alternative strategy for increasing voter turnout, he suggests we should pay people to vote instead:
[I]f increasing turnout is really so important, why signal that fact by punishing people who don’t go to the polls? Maybe we’d do better to reward them instead. If voting is such an unadorned good, let’s pay people to show up. Surely paying people to do a good thing isn’t a bad thing.
Social scientists have understood for some time that cash payments alter people’s incentives, sometimes drastically. For example, paying people money to quit smoking greatly enhances the chances of success. Paying students to keep a certain grade-point average seems to make a difference. Paying teenage girls not to get pregnant greatly decreases the chance that they’ll get pregnant.
These are all behavioral changes we want to encourage. Why not treat voting the same way? It’s likely to work. An experiment conducted by Fordham political scientist Costas Panagopoulos found that paying cash rewards of $25 raised turnout in a municipal election from 14.9 percent to 19.2 percent — no small increase….
I’ve long been mystified by our bipartisan national determination to achieve what we think is best by punishing people who won’t go along. Rewards for good behavior are better than punishments for bad. They force us to discover how much we really value what we claim to want. If increasing turnout is really as important as supporters say, let’s give nonvoters a real incentive to go the polls: cash.
Carter’s proposal has several advantages. Unlike mandatory voting, it does not infringe individual liberty, or unjustly penalize people who have legitimate moral objections to voting – including Carter himself. And, if you believe (as I do not) that increasing voter turnout is an important goal, then paying people to vote is likely to be an effective way to accomplish it. As Carter points out, even small financial incentives often have a big impact on behavior, especially when it comes to incentivizing activities like voting, which are generally not very difficult or time-consuming.
However, there is still an important downside to paying people to go to the polls. One of the biggest objections to mandatory voting is the danger that it will exacerbate the already severe problem of voter ignorance. People who, like Marshawn Lynch at the Super Bowl press conference, only show up to the polls so they won’t get fined, are likely to be more ignorant about the issues they vote on than those who vote voluntarily. The same thing is likely to be true of those who only show up to the polls because they are getting paid. The end result would be an even more ignorant electorate than we have now, and party platforms and government policies that cater to that ignorance. If you are a liberal who believes that increasing turnout would be an unadulterated benefit for the left, keep in mind that relatively ignorant people also tend to be more xenophobic and socially conservative than the current median voter.
One possible solution to this problem, is to couple Carter’s proposal with economist Bryan Caplan’s proposal for a “Voter Achievement Test,” which would pay people to increase their knowledge of politics. I tentatively advanced a similar idea in Chapter 7 of my book on political ignorance. In principle, people who scored well on the VAT could also be paid to turn out to vote (as well as for doing well on the test itself). That way, we can increase voter turnout and diminish voter ignorance at the same time.
For reasons I noted in this 2013 post, Caplan’s idea has a lot of merit:
His plan has some virtues beyond the ones he mentions.
First, unlike with the literacy tests of old, it doesn’t actually deprive anyone of the vote. No one is even required to take the test, much less forfeit the franchise if they don’t do so. Second, liberal egalitarians should like the fact that this plan would probably increase knowledge among the poor more than the affluent. The prospect of winning $500 or $1000 would be a more significant enticement to the former than the latter. Thus, Bryan’s plan could help close the large political knowledge gap between high and low-income voters.
Sadly, however, there is a fly in the ointment, also outlined in my 2013 post:
As Bryan recognizes, the major objection to his idea is the likelihood of partisan bias. Incumbent office-holders will do all they can to bias the test in favor of their party’s positions. I think Bryan dismisses this danger too easily by noting that traditional civics education is often biased as well. There is plenty of political bias in traditional public education. But precisely because under Bryan’s people would have much stronger incentives to actually learn and remember the information in question, the impact of the bias will be enormously magnified. This danger is the reason why I stopped short of actually endorsing the idea of paying voters to learn political information in my book.
In sum, I would be open to supporting Stephen Carter’s idea of paying people to vote if it can be effectively combined with some version of Caplan’s or my scheme for paying voters to increase their knowledge of the issues they are voting on. So far, I don’t see any effective way to forestall the danger of bias that bedevils Caplan’s VAT proposal. But I am not ready to dismiss the idea entirely. Perhaps someone else will come up with a solution to the problem of bias that has eluded me.
Of course, any such solution would have to be feasible under real-world conditions, which include enactment by real-world politicians and administration by real-world government bureaucrats (or perhaps private contractors hired by those same politicians and bureaucrats). I don’t doubt that an ideologically diverse team of economists or political scientists could come up with an excellent and well-balanced test. But I am skeptical that any such objective test would actually be adopted by the government. And even if it is, I worry that it would become corrupted by partisan bias over time.