Government is largely about providing public goods, defense and secure property rights for example. Democratic governments have been particularly successful in providing public goods. But democracy is not perfect; democratic governments make frequently poor decisions through inaction or poorly thought out action. For example, despite widespread agreement that farm supports are bad public policy, they persist in many democratic countries. Part of the reason democracies often make suboptimal judgments is that democracy itself relies good decision making by voters, which is itself a public good. Each individual voter has a very small impact on public policy, and public policy generally has a very dispersed effect, so an individual’s vote has a negligible impact, positive or negative, on that individual. Because informed voting is a public good, voters do not have strong incentives to be informed or vote intelligently, and this can lead to bad public policy.

Economists have long noted that because the individual costs of voting intelligently are negligible, voters rationally do not invest time and effort into determining what is good policy and what is not. This can lead to bad policies when voters make errors which do not cancel out, that is, when their errors are not random but systematic. Economist Bryan Caplan has argued, in his book The Myth of the Rational Voter and elsewhere, that this is often the case; voter’s errors do not cancel out, so that voter ignorance has serious effects. One reason voters make systematic errors, is that some policies sound good initially even though they are bad, and other policies appear bad even though they are good. For example, price controls, may sound good initially, but consideration will reveal that price controls are almost always bad policy.

The problem of voter ignorance is not intractable, however. Our democratic institutions do not need to be static. It is possible to make novel changes our democratic institutions to drastically reduce this problem. Here, I will propose a new democratic institution which I call Professional Voting that would aim to provide incentives for voters to become informed about politics.

Basic mechanism in Professional Voting
A political system with Professional Voting would deal with the problem of voter ignorance by paying a relatively small but representative subset of the population to be informed and to vote, thereby encouraging good public policy. Voting in most elections would be restricted to Professional Voters, but anyone would be eligible to become a Professional Voter. The aim of Professional Voting would not be to make voters decide exactly what is good policy themselves, but to create an electoral process that selects competent representatives and administrators and to provide the opportunity for policies which are initially unpopular to gain support through evidence and debate.

Directly paying voters to be informed would give voters an incentive to be informed only if voters who become more informed are paid more, and this requires a way to measure either how much effort an individual has put into becoming informed or how well informed an individual is. The Professional Voting scheme would monitor how well informed individual voters are by giving voters a test right before they vote, which would measure their knowledge of political facts and arguments. To ensure that Professional Voters would be representative of the population at large, they would also be given a demographic survey, which would ask about their income, race, gender geographic location and any other factors which would be likely to shape a voter’s perspective and values. Then, after they vote, voters would be given a relatively large payment (perhaps about $10,000) based on how well they did on the political test and based on how representative they were of underrepresented demographics. The initial voting result would then be statistically corrected so that the final election outcome would be representative of the population at large and to model how voters would have voted if everyone had done very well on the test. This corrected voting result would be used as the final outcome of the election. The body responsible for creating and administering the test and demographic survey would be a proportionally elected Electoral Council, the subject of the next section, which would be elected in a normal election by non-professional voters.

The Electoral Council
The Electoral Council would be responsible for creating and administering the political test and demographic survey given to Professional Voters. and for conducting the elections. To make sure that the Electoral Council always represents the interests of the population at large, it would be proportionally elected by in a normal, non-professional election and voters in the election would receive a small payment, perhaps $10. Independent elections would help avoid self reinforcing differences between the values of Professional Voters and the values of the rest of the population.

The Council would also be almost completely separated from other government institutions. It would not have any power beyond conducting elections, and the other branches of government would have very limited power to interfere with the Council or the elections it would be responsible for, although, other branches would be allowed to police election fraud in both types of elections.

The Electoral Council would put together the political test based on arguments made by political parties, interest groups, and the media on issues relevant to the races decided by Professional Voters. The Council would put together the survey based on what it’s members thought were the most important demographic factors which shape the perspectives and values of the population at large. The proceedings of the Electoral Council and the content of the political test and the demographic survey would only be revealed publicly after the elections have taken place.

Political Knowledge Test
The purpose of the political test would be to measure Professional Voter’s knowledge about political facts and arguments relevant to the issues in the races decided by Professional Voters. The test would likely consist largely of questions about arguments made in politics, for example, “(multi-choice) Which of the following is a reason the New Liberal Party has claimed it voted against raising the minimum wage?”.

Vote trading within the Electoral Council between different factions, would ensure that the test was composed of questions about those arguments and facts which different factions within the electoral council expected to be the most persuasive to Professional Voters. Every individual faction would be willing to trade support for including questions about arguments other factions wanted included on the test in exchange for support for including questions about the arguments arguments they expected to be the most persuasive to Professional Voters included on the test.

Demographic Survey
The purpose of the demographic test would be to make sure that the perspectives and values of Professional Voters would be representative of the population at large. To make the outcome of elections representative of the population at large, the Electoral Council would compare the demographic survey with census data and weight the raw voting results so that the final vote was representative of the population. Vote trading within the Electoral Council would ensure that the demographic questions included in the survey were considered the most relevant questions by the Council, in the same way that vote trading would ensure that the questions included on the political test were considered the most relevant questions.

Benefits of adopting a Professional Voting system
The biggest benefit of adopting a Professional Voting system would be an improvement the process of government by creating much more room for debate and evidence to change the outcomes of elections. Improved policy decision making would also be compounded by an improvement in discourse quality by forcing pundits who want to change policy to acknowledge and respond to the arguments of the opposing side. It would also increase the private benefits to pundits of making novel political arguments. Good pundits would be more famous because more people would pay close attention to them and good arguments would have much larger impacts on policy than they do now. There would also be more potential for good pundits and information aggregators to be rewarded financially because since Professional Voters would be paid to know about politics, they would be willing to spend money to become informed.

A Professional Voting system would also make government more responsive to the needs and desires of the population. This is because informed voters would do a better job pressuring public officials to be competent and honest and to be good agents for the people than conventional voting systems. This pressure would affect even the unelected parts of the government because elected officials would face pressure to improve the behavior of the parts of government responsible to them.

Potential objections to Professional Voting
A Professional Voting system would be a radically different from any form of democracy that governments use now, so there will of course be objections and criticisms. One major objection to Professional Voting systems is that it is hard to tell whether the factions in the Electoral Council, elected by the population at large would have the appropriate incentives to create a test which would accurately gauge the political knowledge of Professional Voters. One worry is that the Council would end up simply making arguments in its test, instead of asking questions to gauge the knowledge of Professional Voters. However, factions within the Electoral Council, if they are motivated to influence policy, would have an incentive to make sure that voters understand whatever arguments they find persuasive. Without conducting some sort of experiment, however, it is difficult to say how an Electoral Council would behave. One might also question whether creating such a test is feasible for anyone. Again, only an experiment could answer this question conclusively.

The fact that Professional Voters would self select is also a potential problem. Professional Voters would inevitably be different from non-voters in ways which the demographic survey does not reveal, and so could not be statistically corrected for. The demographic survey cannot completely solve the problem of representative representation because Professional Voting itself creates a new demographic, that of the Professional Voter. This issue would persist even though Professional Voters will be paid; for example, among professionals, those who become Professional Voters will be those with time on their hands. These people will be distinct from other professionals in ways that are difficult or impossible to discover using a survey; they may be unusually lazy or unusually motivated or have other uncommon characteristics. It is difficult to say whether these differences would have significant effects on election outcomes.

One effect which would mitigate the effect of the invisible differences between voters and non-voters is that voters vote mostly altruistically, for reasons which economist BryanCaplan has explained and supported empirically. This effect should be expected to be greater with Professional Voters because they are informed, so they will have relatively more knowledge about circumstances which do not directly affect them and because altruism will be explicitly socially expected from them. Altruistic voting means that Professional Voters will not purposefully favor their own interests over those of others, especially when those people are very similar to them.

Paying Professional Voters would also reduce these invisible differences by encouraging people engage in professional voting for the monetary rewards instead of for the psychological or ideological rewards, which would help reduce the proportion of strongly ideological voters (which would be impossible to incorporate into the demographic survey) to levels more like those of the general population.

One might also be concerned about the cost of a Professional Voting system. Paying people to be informed and to vote could be quite expensive, but order of magnitude estimates suggest that it would not be. If every state in the U.S. “employed” 1,000 Professional Voters and paid them each $10,000 per year, the system as a whole would cost $500 million per year, which is not very expensive relative to the costs of many government programs. For example, federal agricultural subsidies cost $15 billion per year on average between 1995 and 2005 (source).

Even though I think there are good reasons to think that a Professional Voting system would improve democratic outcomes, I do not advocate adopting a Professional Voting system on a large scale without first testing it thoroughly on a small scale, because the costs of wholesale institutional failure would be very large. A Professional Voting system could be tried on a small scale in a lot of different places. For example, a Professional Voting system could be used to elect school board positions or other local offices. A professional voting system could also be used to elect the board of directors for a corporation. If Professional Voting can work on small scales, then it will be worth considering in larger arenas.