When Is It Wrong To Vote?

Jason Brennan writes:

"On my view, voters are not obligated to vote, but if they do vote, they owe it to others and themselves to be adequately rational, unbiased, just, and informed about their political beliefs."

Right. That would be great. But depending on how you define such terms, the consequences could range from everyone abstaining from voting to no one abstaining from voting.

If the concern is about redistribution of incomes, a better answer might be to restrict the vote to citizens who pay taxes. Whether this includes government employees or not, I will leave as an open question.

Here are a few questions:

How many people who would consider themselves responsible voters are, in fact, not?

How many people abstain from voting just because they don't consider themselves responsible voters? ( I didn't vote in the last 3 elections because of this.)

What percentage of the population could in fact be called "responsible voters," and thus be allowed to vote?

Good questions, both of you. (Thanks, Bryan, for the post.) If you'd like, send me your email and I'll send you the paper, where I try to tackle these kinds of objections.

(Nick: In particular, look for the section on "self-effacingness".)

It seems to me that any sentence of the form "Irresponsible people ought to do X" is largely futile. If they did what they ought to do, they would not be irresponsible.

In a modern democracy, the vote of an individual is unlikely to affect the outcome of an election. The chance is so unlikely as to be zero, for all intents and purposes. Why should anyone agonize over a decision which is consequence-free?

I agree that individual bad votes have next to no expected disutility. I used Geoff Brennan and Loren Lomasky's formulae from their Democracy and Decision. Suppose that the difference between two candidates D and E and easily quantified: E will cost us $33 billion compared to D. Suppose that the number of voters in the election will be the same as in the last US presidential election. And suppose E commands a slight lead in the polls, such that the probability a random voter will voter for E is 50.5%. You vote for E. On Brennan and Lomasky's formulae, the expected disutility of your vote is a mere $ 4.77 x 10^-2650.

So, yeah, individual votes aren't worth much and don't cost much. Bad voting is collectively, not individually, harmful.

That said, it's not obviously decisive.

The basic outline of the argument in the paper is: 1. Individuals have an obligation to refrain from participating in collectively harmful premises provided, at least, that such restraint does not impose any significant personal costs. 2. To vote badly is to participate in a collectively harmful activity, whereas abstaining from voting badly imposes low personal costs. 3. Therefore, individuals ought not vote badly. The premises are each defended at greater length, and various objections are considered.

Cheers,
J

Oops. That should say "activities", not "premises" in premise 1 above.

The problem that I see is with the definitions of "responsible" and "good".

After all, if I'm going to drive a car, how good do I really have to be at it? Do I need the skills of a stunt driver or professional racer? Clearly, I don't. Do I merely need the requisite skills to pass a driver's test? That seems to set the bar rather low; I ought to aspire to be a better driver than the bare minimum. But none of these aspirations suggest that I need to be the very best parent, driver, or surgeon that I can possibly be - merely that I be a "good, responsible" one.

Finally, and this goes to what I found to be the missing element in MRV, what defines "responsible" in a world where my voting choices are strictly limited? A person cannot vote for either McCain or Obama without embracing a number of terrible economic policies. I'm almost certainly a well-above average potential voter. But I don't have an option which not only fails to significantly improve on current harms, but which will not likely result in additional future harms.

Is there a good formal model of this problem, where each voter has a private clue about which candidate is best, voters vary in the quality of their clues, and voters must choose whether to abstain based on their relative quality?

This is what voters in Louisiana faced some years ago. Edwin Edwards vs. David Duke. There were pumper stickers that said, "Vote for the crook. It's important." I was much like the lab rats we tortured in behavioral psychology; frozen in place from fear, damned if they moved and damned if they didn't. Quite a few people I knew did not vote that year.

An Australian visitor just told me that in Australia, citizens are legally required to vote in elections; there is a steep fine for failing to do so.

I wish that topic to be discussed.

Would such a requirement raise or lower the quality of government in the US if it were imposed? (Ralph Nader has advocated such a requirement, but allowing voters to mark the ballot "abstain".)

Has Australia been hurt by the requirement?

Would such a requirement change citizen incentives, so that they might become better informed?

Would such a requirement bias US politics? For example, would the people who are forced to vote who would have otherwise stayed home be more likely to vote, say, Democrat?

The 'harm' is in the not voting, not in the voting badly.

Perhaps you'll agree that whether the election is for POTUS or the mayor of a small town in Alaska it is nearly impossible to be an expert in all the major issues and policy topics bundled into an electoral decision (e.g. foreign policy, zoning, school district parcel tax law).

Perhaps further you'll agree that it is nearly impossible to predict how an elected official will influence actual policy-related outcomes. Impossible to predict in large part because candidates are intentionally vague about most of their positions. Candidates are intentionally vague because the bundle of issues present in their race is large and staking out specific positions is likely to drive down their support. For example, imagine that there were only 3 issues bundled into an election and each issue had only a yea and nay position associated with it. This means that there are 8 possible choice combinations. If each issue is fairly closely divided then a candidate who expressed a definite preference for each of the 3 issues is likely to get only 1/8th of the total vote. You may object that issues choices are clustered in a coherent way among the voters, but, I think the percentage of 'independent' voters is roughly equal to the percentage of voters whose preferences do not fall into the defined party clusters.

It is therefore impossible to make a 'good' (where 'good' means closely aligned with your desired outcome) choice with your vote as you cannot possibly have enough information with sufficient predictive power.

However, choosing the politician is NOT the point of an election. Elections are intended to MANUFACTURE CONSENT of the governed. It is by elections that we convince ourselves that even if we individually get outcomes that we don't want that we are nevertheless willing to continue to play this game rather than burn down the house.

Consenting to be governed is the key benefit here! Therefore the larger the proportion that participate the more convincing are the results, even if the results are narrowly divided or even meaningless in terms of policy choices.

The individual voter who does not participate is harming every one within society by eroding the legitimacy of the 'consenting-to-be-governed' process. For, even those who do not vote themselves are nevertheless impressed by the results of the election if the percentage who participate is large.

As further evidence of manufacturing consent as the primary goal of elections, I refer you to the high voter turn-out generated by totalitarian and despotic regimes as "proof" of their legitimacy.

But, hey, what do I know?

Another bob, you're right.

The advantage of democracy is that it allows regular and peaceful change of the government. It is possible only if people actually vote. If they do not vote, democracy is undermined from various sides and treat to collapse in some kind of dictatorship. Every political force - especially extremists - start to interpret high abstinence as an argument for their case. "People do not vote because they know politicians are communists, homosexuals - and Jews." And without actual votes to show, you cannot easily show that these claims are invalid. Institutions traditionally in search for power, like army, church, secret police etc. can join new game. It can be enough that democracy collapse into dictatorship, especially in young and weak democracies in Africa, Russia etc, especially if accompanied with some kind of economical or foreign policy crisis.

The quality of political decision in democracy is not achieved by system of voting itself, but through education of voters, improvement of political culture, public discussion. Betters voter you have, better decisions you get. And that's it - you cannot get more than that out of democracy.

Instead, I think it is better to appeal on voters to think twice, inform themselves better, not to accept beliefs of the people in their environment for granted - and then to vote the best they can.

It is, of course, valid argument only if Brennan believes in democracy. If he believes that there are better alternatives - that is different thing.

I don't see much difference between a so-called intelligent voter and a voter who is essentially an ignoramus, as long as the choices offered are so meager. Thus, we can vote for a Republican or a Democrat, and any intelligent and semi-objective person can see plenty of defects (and some advantages)in both.

Yes, the choices are meagre - as in differences with no distinction - because, as Dr. Caplan points out both parties must converge toward the median voter.

My personal rules for deciding how to vote apply regardless of the specifics of the candidate, the specifics of the issues:

1. Throw the new rascals in. Vote against the incumbent, it helps prevent corruption.

2. Split government is best - preferably a Democrat executive and a Republican legislature. Vote accordingly. (Single party government is the handbasket we make the trip in.)

3. When 1 and 2 are in conflict (as in POTUS 2008) - vote for a third party.

So, there! Did I mention anything about issues?
;-) What an ignoramous.