In Democracies, good public policy is largely a public good which is indirectly provided (or not provided) by voters. Because it is a public good, voters do not have strong incentives to be informed or vote intelligently. Good public policy is underprovided because the electorate has biases and because without research and discussion, even great policies may at first appear bad, so it is difficult for many good policies to gain support.
In terms of human computing power, it is not really necessary to have large elections to elect good representatives and officials. A small, but well informed, fraction of the current electorate could conceivably elect good representatives and officials. The reason we do have large electorates is that a small electorate will not necessarily be representative. Having large electorates is a way to ensure that elected bodies represent the interests and perspectives of everyone in the country.
I want to propose a system of professional voting. The system would aim to ensure that voters are informed and thoughtful as well as aim to ensure that everyone’s interests are represented. Essentially, Professional Voting would pay a portion of the electorate to be informed, model what the results of the elections would be if the whole electorate was informed and voted and then use that result as the result of the election.
The purpose of Professional Voting is not to make the electorate formulate good policy by itself, but to create an electoral process that can identify competent representatives and officials and provide the opportunity for policies which are initially unpopular to gain support through evidence and debate. Professional Voting would improve the process of government by creating much more room for debate and evidence to change the outcomes of elections.
In Professional Voting everyone wishing to be a voter takes a political knowledge test and a demographic survey before voting. The information from the political knowledge test and the demographic survey is used to model how a fully informed electorate that was representative of the whole country would vote. The model is used to select the actual representatives and officials.
The political knowledge test would ask relatively detailed questions about various issues that had come up in the media since the last election as well as questions about the common arguments surrounding issues (for example, “(True/False) A common argument against raising the minimum wage is that it would increase unemployment.”), questions about the structure of government relevant to current issues (for example, “(True/False) Parliament does not have the sole power to declare war.”), as well as questions about the positions and history of candidates in elections (for example, “(multichoice) Which of the following is a reason the New Liberal Party has claimed it voted against raising the minimum wage?”).
The demographic survey would ask questions about variables which might conceivably affect a voter’s perspective (race, religion, education level, sexual orientation, occupation, etc.).
Based on the results of the test and the survey, a model would be constructed that predicts what the results of the election would be if everyone had gotten all (or nearly all) of the questions right on the test, as well as if the electorate were completely representative of the population at large.
To assure that a sufficient number of voters do well on the political knowledge test, voters would receive a payment related (though not necessarily proportionally) to how well they did on the test. The payment would be relatively large (at least a few thousand dollars) to give voters the much needed incentive to become well informed. Such a voting subsidy would also reduce the ratio of radical voters to moderate voters because radical voters get more psychological benefits from voting. The voting subsidy would decrease the difference between the private benefits that moderate voters and radical voters experience from voting.
Assuring that the political knowledge test and the demographic survey reflect the important issues and demographics relatively evenhandedly is probably the most difficult part of Professional Voting. The only method I can see of accomplishing this task is to have a system of local and independent election commissions charged with creating the test and survey, as well as conducting elections. The ideal members of such a commission would be interested in accurately and evenhandedly measuring the political knowledge of voters and not especially interested in any particular policies or candidates. There are two ways that could be used to select good members for the commissions. The first way is to have classic (non-professional) local elections to select the members. The second way is to have a selector appoint the members subject to approval by a super-majority of a local level legislative body (I have discussed high approval appointments here).
The benefits of Professional Voting would be that voters would be much more aware of proposals and debates so it would be much easier to change their minds. Quality representatives would be much more common than under classic voting rules because voters would be more aware of corruption and incompetence. Reform, even institutional reform, would be much simpler to accomplish than under classic voting rules because voters would have incentives to learn the details of proposals and understand the arguments for and against them which would create ample room to seriously debate even initially unpopular proposals.
Although, I admit to being somewhat uneasy about this proposal, the more I consider it, the more workable it seems to me.