Harold Winslow

Choosing Leaders

Proposal: remove congressional district boundaries

Allocating legislative seats by geography makes sense only to the extent that people's interests are tied to where they live. This is much less true now than it was historically, as people move more and the internet facilitates the homogenization of culture. District boundaries are also a contentious topic because they're always subject to gerrymandering. We could improve the way seats are allocated by simply eliminating all boundaries. Winners would then be determined by choosing the 535 candidates that received the most votes nationwide; each winner would be responsive to the people that voted them in, wherever those voters happened to be.

This formulation would effectively allow congress to partition the electorate into groups with common interests, whether those interests were geographically based or not. Some candidates would continue to run local campaigns, but others would find constituents based on issues or demographics. One person might win on the votes of veterans nationwide while another wins on the votes of small business owners in the Southwest. Many minority groups would get more representation under this system since they would be able to aggregate their votes in support of regional or national candidates.

As a practical matter, it would probably be best to allow voters to choose more than one candidate since so many options would be available on the ballot.

Proposal: treat voting like jury duty

First, there is such a thing as a good president or a bad president. This is somewhat independent of your ideological leanings. The vast majority of people will agree that certain extreme outcomes (like sustainable 5% annual GDP growth or major war) are desirable or undesirable. And leaders differ in their capacity to move the country towards or away from outcomes like these.

To put it differently, a person on the far left should generally prefer a competent right-wing leader to an incompetent one. This sounds contradictory, because a sitting duck should be easy to override and outmaneuver. But an incompetent leader can do great damage by failing to act when necessary, by misallocating resources, or by unknowingly taking great risks. Better to have a competent leader who moves the nation orthogonal to your desired direction than an incompetent one who pulls it backward.

Among qualified and competent candidates it may be virtually impossible to select the 'right' candidate. But there are very clearly wrong candidates. Your local gas station attendant, however nice, should probably not be president. Now the candidates we receive on the national stage are not generally so egregiously unqualified, but you may be able to cherry-pick one or two examples who you judge to be worse. It would be nice if our elections could screen out such candidates with some reliability.

Second, individuals differ in their ability to distinguish good candidates from bad ones. Gullibility, background knowledge, and attentiveness during the campaign are all relevant factors and all vary considerably across the population. I will not belabor this point, but if you wonder about the scope of human variation with regards to knowledge and opinions about politics you may find it useful to get out more. Log on to Reddit and browse the front page. Do the same on Youtube while logged out, so that you can see its recommendations for the general public. Now imagine this as your primary media diet for a week, month, or four-year term. It is hard to believe that you would come into the general election in anything resembling the intellectual shape needed to sift through ads designed specifically to mold your opinion in a predetermined way.

On a related note, consider that door-to-door get-out-the-vote efforts work and that states find it necessary to randomize the order of names on their ballots. That is, people's voting behavior can be changed by features of their environment so small and irrelevant as the presence of someone at their door or the order in which they see names in the ballot box. It is hard to see how people who are swayed by such minutiae could be positively contributing to the selection of our public officials.

We presently have no real compensation for voters who do a good job at selecting representatives, which means that the legitimate effort required to sift through policy recommendations and sort fact from fiction is probably massively underdone. Perhaps worse, there is no disincentive for people who get their information from unreliable sources, vote myopically, and generally allow themselves to become irrationally led by dumb campaign ads.

At the same time, we should take seriously the costs of becoming informed, ignoring unwanted advertisements, registering to vote, traveling to the polls, enduring endless political conversation at home and at work, and tallying the votes. The sheer number of hours spent on these activities is substantial. (It is possible that we actually don't spend enough time on these activities, but given the scale it nevertheless behooves us to consider if we could do things more efficiently.) How many people die from car crashes on their way to or from the polls? How many of their votes were decisive? How much money is spent on advertising arms races? It is possible that, after running the numbers, you would find that all of these costs are very worthwhile given whatever benefits you suppose they produce. But if you have not actually attempted to estimate these costs in some way and already 'know' that they are worthwhile, then you are failing to acknowledge them at all.

In sum: we employ millions of volunteers to select our policymakers but do nothing to screen out the people who will do a bad job or encourage those who are likely to do a good job to participate. In no other context would this seem sensible. It is virtually guaranteed to result in a massive tragedy of the commons...indeed, it does, every election cycle (at least as measured by approval ratings).

Now, you may object that none of this matters because everyone has a right to vote and restricting suffrage is abhorrent. I am not convinced. If you think that the results of elections are significant - that some candidates can be dramatically better than others for the well-being of the country overall - then you should be pretty concerned about unqualified people voting too much. If much of an election is driven by the votes or views of people who have not given the issues much thought, elections can quickly devolve into choices over two popular products, like a vote over Pepsi vs Coke. We take the right to vote too seriously and the responsibility to vote in an informed way too lightly.

People like to say that it is your civic duty to vote. This is nonsense. It is your civic duty to exert a positive influence on your community. Whether you are personally capable of doing that via voting is an open question. If you happen to be very uninformed, it may very well be your civic duty to either become quite informed or to abstain from voting. If you have no particular preference between the candidates, it may be your duty to abstain. If you spend a great deal of time reading about politics and vote but never write op-eds or run for office, maybe you are not fulfilling your civic duty. If you are a valuable member of your community and you are more likely to die in a car accident on your way to the polls than you are to be a decisive voter in the election, maybe you have a duty to stay home. Whatever civic duty you personally have, it is not likely to cleanly stop and end with a single trip to the polls.

Of course, there is no case to be made for restricting or unduly weighting suffrage based on irrelevant characteristics - gender, race, location, etc. But that doesn't mean it's impossible to design a system wherein the voting pool is demographically similar to, but more knowledgeable than, the population as a whole. Which brings us to the proposal.

It is pretty difficult to set any well-defined standards for who should be able to vote (if you're looking for some guidelines about how suffrage might profitably be restricted). You could set educational standards or require voters to pass some sort of test, but this feels like a messy solution, ripe for continuous political infighting regarding the contents of the standards/test.

So instead of attempting to filter out the least knowledgeable voters, why don't we try to bring out the best in the voters we do get? Imagine that our elections were structured like jury trials. Each election, 10000 Americans would be randomly selected for voting duty, divided into 1000 groups of 10. Each would report to a local site each day for a month - perhaps 5 hours per day for 20 days, for a total of a 100-hour commitment. Candidates from the five largest parties would each get a chance to present their cases and respond to one another's arguments; this could be done via real-time video feed to each voting center. At the end of each day of presentations, each group would be allowed to submit one question to the candidates, and a small number of these would be randomly selected for the candidates to answer the following day.

10000 people out of 300 million doesn't sound like very many, but the outcome is likely to be very close to the one that the entire population would have come to at a very small fraction of the cost.

There would be ample time for deliberation, and some experimentation would need to be done regarding how much time to allot for this and how to structure that time, if at all. At the end of the commitment, each voter would cast a ballot, perhaps using quadratic voting, and this would determine the outcome of the election.

There would be a lot of details to sort out, like what outside information the voters should have access to during their service and under what circumstances they should be allowed to opt out of service. But one can imagine that the net effect would be a much better-informed population of voters. In practice, it is hard to see this idea getting any traction from the two major parties since it gives equal airtime to parties which presently have far smaller budgets. But it may have a chance of adoption in local politics where party budgets are not as decisive and there may be fewer than five candidates running for office.

Proposal: limit presidential candidates to qualified people

"Qualified people": former or existing governors, generals, ambassadors, and directors of federal agencies.

Additionally limit the total field to 50 candidates, with preference determined by recency of service and length in office. This restriction would screen for seriousness and competence and would provide voters with a field of candidates that were somewhat easier to compare since they would be more similar to one another than the candidates we get today. Losing variance is bad in that it eliminates some of the very best potential candidates, but it should also remove most of the very worst, and that may be much more important.

A nice side effect of this approach is that there would then be a defined pool of candidates to fund in any election year. It would then be easy enough to provide public funds to each candidate in equal amounts for campaign purposes while severely restricting or eliminating private contributions. Since regulatory capture is the root of all evil, this would be a great advantage.

Proposal: give legislators a fixed number of votes instead of a fixed term

Say, for the sake of example, that there are 100 votes in the Senate per year. Then by today's system, every state is allotted 200 votes per year or 1200 votes per session. Imagine that your senators only vote on half of the issues, so that your senators have 600 votes remaining at the end of the session. Then at the next election, the incumbents remain in power (with 600 more votes to use between them over the course of their lives) while the voters choose who will be awarded the next set of 1200 votes. Award the votes proportionally, so that a candidate who receives 10% of the popular vote receives 120 votes in the Senate.

This system allows candidates to market themselves in ways that they can't today (e.g. as single-issue voters), and it would probably allow for a better fit between the populace and their politicians. A potential downside is that it might introduce additional strategizing among senate leaders in an effort to manage vote counts.