How much time to do you spend figuring out how you should vote? A few hours? A few minutes? Any time at all? I could not find a study that measured the average time that voters spend researching their candidates, but I did find that three quarters of Americans do not know how long a senator’s term is. Perhaps an average voter reads the blurbs in the local newspaper, listens to some soundbites on the news about each candidate, and watches a myriad of television ads. Does he read each candidate’s website? If not, this is like owning a bagel shop and hiring a manager without reading his resume. Imagine a bagel shop where all employees are hired with the same consideration that the average voter gives to political candidates. Such a shop would operate with the legendary efficiency and effectiveness of a DMV.

You may think I am writing to encourage you to put more effort into deciding for whom to vote. I am not. You are doing it right. Unlike choosing a bagel shop manager, choosing a governor culminates in the counting of millions of votes. The odds that your vote either way will change anything at all is infinitesimal. Your vote, frankly, doesn’t matter. It is irrational to vote at all. Enough people do it out of altruism which keeps the system running, but not enough people do the research necessary to keep it running well.

How much research would be appropriate to choose the governor? I just spent a week on a jury deciding a workplace injury lawsuit. Companies and universities spend months looking for a president. Would we be better off if everyone took a week off of work to research the candidates for governor? Well, no. We would undoubtedly end up with a better governor, but the cost would astronomical. And that is just for the governor. Imagine having to do this for the mayor, the city council, the state representative, the state senator, the federal representative, the federal senator, and the Presidentnot to mention lots of other positions that probably should not even be elected positions, like dogcatchers and judges. With the cost of research so high and payoff so small we collectively settle for a million snap judgements and hope for the best.

Can we do better, and if so, how? The reason that an individual vote is worthless is simply because there are so many voters. The decision, while valuable and important by itself, once spread over millions of voters is irrelevant. The huge number of voters per election is simultaneously the reason that it would be prohibitively expensive to have all voters be politically educated. A week of one person’s time is not that much; a week of million persons’ time is overwhelming. If we want to have votes count, we need to have fewer voters. A solution to this problem is something I call the electoral jury (technically, a form of sortition). In the electoral jury, a random subset of the eligible voting population is chosen and this subset, the jury, is tasked with electing a particular office. Rather than having the entire adult population of a state vote for the governor, a random few are told that they and they alone get to choose the next governor. I present here a small subset of the details of this plan.

Jury Size

Choosing the size of the jury is the most important variable. If the jury is too small, then stochastic effects become important. With an 11-person jury and 60%/40% split between two candidates, the 60% candidate will win 75% of the time. With a 101-person jury, the 60% candidate will win 98% of the time. With a 1001-person jury, the 60% candidate will win 100% of the time. The closer the race, the more likely it will be for the majority winner to lose simply by chance when jury is small. With a 1001-person jury and a 51%/49% split, the 51% candidate will win 72% of the time. At a 10001-person jury, the odds are 98% in the favor of the 51% candidate. So it takes about 10000 people to completely eliminate chance from the election. I am not convinced that one needs to eliminate chance entirely. Is a 51% candidate  clearly better than a 49% candidate such that there needs to be no chance that the 49% candidate is elected? The main benefit of increasing the jury size is to get rid of chance and prevent the government from constantly changing hands even when there is no change in public opinion. If a governor is elected with 60% of the vote and, going into his reelection, 60% of the people still support him, then he should keep his seat. With a jury of only a dozen people, there is a good chance it will not happen.

If the jury is too small, then a successful bribe carries a lot of weight. We do not think much about secret bribes in today’s politics. The number of voters is so large that secret bribery is infeasible. In the end, this should not be a large problem. We manage to prevent bribery in criminal and civil cases, where the juries are small and the stakes are high. Nevertheless, it is one effect to keep in mind.

On the flipside, if the jury is too big, then the original problem returns. Each vote begins to not matter much. The original US House of Representatives had one representative for every 30,000 people. Each representative was elected by much fewer than 30,000 people, considering that suffrage was fairly restricted at the time.

Keep all this in mind, I think a number of jurors in the 1000-10000 range is about right.

Jury Location

The easiest place to have the voters vote is at home. It’s cheap, easy, and exactly the same as the current system. But with so fewer voters, we can do so much better. Here is where the big payoff of the electoral jury comes. With a small number of people (1000-10000 range), it is feasible to sequester them at a convention center for a week. Here, they will have the opportunity to listen to speeches made by the candidates, read material provided by the candidates, and discuss among each other their ideas and opinions. Importantly, they will be isolated from all forms of external media. Why is this important? Because this removes money from politics. TV ads, newspaper ads, street criers, door-to-door solicitorsthey all swing elections and they all cost money. This requires all candidates to be friendlier toward people with lots of money. With the jury sequestered and every candidate given equal and large access to the voters, campaign contributions are now meaningless. And on top of it all, no one’s freedom of speech is impinged. Anyone can say what he wants, spending as much money as he likes to say it; it just will not matter. The voters will not see the ads, and if they do in the months leading up to the election, the message of the ads will be swamped by a week’s worth of speeches. A 30 second TV ad is only effective because each voter spend so little time considering to the candidates.

Would sequestering the jury eliminate the value of journalism? If a journalist uncovered the fact that one of the candidates was a secret communist, would the jury never see this? No, the jury would see it, because the candidates are not sequestered. If something important is found outside, one of them can bring it in and present it to the jury.

Another advantage of bringing everyone together is that people are typically good at judging the character of those they meet in person. Judging character is one of the most important aspects of voting. It is not what a politician says he will do that matters; it is what he actually does. That people are pretty good at judging people they actually get to know is more of an impression I have. I have been unable to find research to support this, so maybe this is wishful thinking.


Sequestration for a week will place many jurors under hardship. It is important that they be accommodated as best as possible. The juror’s wage should be paid to his family for the week he is absent, so that the electoral jury is not biased against those with dependents. All travel expenses should be paid for by the state so that the jury is not biased against those who live farther away. Childcare either at home or at the convention center should be provided for those who care for small children.

It is impossible to remove all inconvenience. But choosing 10 thousand out of 10 million every four years means that being selected is a once-in-a-lifetime opportunity. Almost everyone will never be chosen to serve on a jury in his lifetime. I expect that most would be willing to tolerate inconvenience to participate.

Final Thoughts

The largest barrier to implementing a plan like this is the status quo. If the politician is chosen by a tiny fraction of the population, how could that be fair? The best point of reference is probably the current court system. The jury is fair, not because everyone in the county votes on it, but because the jurors are randomly selected from everyone in the county. Imagine that you were falsely accused of murder and there are two possible juries. The first possibility is twelve people forced to listen to all the evidence given every day by opposing sides. The second possibility is a million people otherwise going about their day who can choose to watch the case or just catch the news about it (with ads from the prosecutor). The potential jurors may vote on your conviction regardless of how much attention they actually paid. It can be fairer to have a smaller sample represent the whole, especially if we can force that sample to sit through fairly presented evidence before making a decision.

All it takes is one brave state to take up this plan for electing their governor. That state and others watching shall be pleasantly surprised.

  • Zeph

    Very interesting.

    It has also been suggested that at least one legislative chamber be chosen at random (the other chamber might be elected in order to retain acquired competence, eg: in foreign policy). Not sure I agree with this concept, but it’s interesting.

    Your suggestion sounds more feasible, as you are replacing superficial examination by millions with in depth consideration by thousands. Many of the problems one might anticipate are answered by “well, yes it’s not perfect, but is it any worse than the current system, having millions do the choosing?”.

    The biggest problem is cultural, not logical. Voting isn’t just about choosing the best candidate – it’s about getting buy-in from the voters, even those who lost. In the West, we have created a culture where for the most part, those losing in an election don’t just take up arms to get their way. They accept that their side has lost for now (but will get another chance). The question is whether this “buy-in” would work when 1000-10000 folks do the voting on behalf of all of us. The ritual of having YOUR ballot counted, of having had YOUR say is pretty important in psychological or symbolic terms. Accepting that a random but statistically tiny fraction of the voters has represented you, may be a lot more abstract and less intuitive. I can imagine a culture where it has become normal and people trust it, but it’s harder to imagine how to get from here to there.

    • David Hagen

      In my opinion, the mandate is the best argument against an electoral jury. The psychological component may be important to stability. That being said, I think we can agree that many people think they don’t have a voice in the current system, partly because of the huge number of people involved, but also that some people have bigger megaphones than others. If they saw people like them engaged in a fair convention (like if there were bios of the people selected), I think people could feel they have even more say.

  • Kaleberg

    You didn’t address the problem of jury selection. You say “random”, but that doesn’t have a precise legal meaning. How are you establishing the jury pool? Do you use the registered voter list? Do you allow people to opt out as we do for jury duty? Do we disqualify people serving time for felonies or suffering from advanced dementia? Perhaps we should allow the general public to vote to choose the jury, sort of the way senators were formerly and the president is now chosen. There is also the problem that random lotteries are surprisingly easy to rig. Look at the Massachusetts lottery. They also have surprising biases. Look at the draft lottery.

    More seriously, there is the issue of whether this method provides a suitable sense of democracy. One’s vote might not count for much, but with others it can count for a lot. In a sense, we already have a jury consisting of party insiders and regulars who vet and choose candidates for nomination. Even in states and for positions that are supposedly non-partisan, there are people who have spent weeks or longer learning about the issues and choosing candidates. The average voter can even stick an oar into this by going to party meetings and voting in primaries. Most don’t bother, but they do vote the party line.

    • David Hagen

      I envision the electoral jurors being pulled from a list of registered voters. I don’t think any further changes to the qualifications are necessary. (Incarcerated individuals are excluded, people can choose to not participate at will, etc.)

      Perhaps we should allow the general public to vote to choose the jury

      We technically have this for US President right now. If the only job a lower representative has is to select a higher representative, such a system immediately devolves into a direct vote for the higher representative because that is the only thing the lower representative can campaign on. How long were electors for president on the ballot? They are completely absent now, and can’t imagine they survived even a couple elections. It worked fine for US Senators because that was not the main job of the state legislators.

      There is also the problem that random lotteries are surprisingly easy to rig. Look at the Massachusetts lottery. They also have surprising biases. Look at the draft lottery.

      To me, this seems like the lowest hurdle.

      More seriously, there is the issue of whether this method provides a suitable sense of democracy.

      As I said to Zeph, I think this is the strongest argument against sortition. That people accept the byzantine electoral college system makes me think that citizens will accept any system as long as it works well enough. The only way to know for sure is to try it on a smaller scale first.