This is a "perfect is the enemy of the good" fallacy.
Yes, you COULD do more by working/donating than voting. But you could apply that logic to just about any good thing someone does -- there's almost always something better. The question is, ARE you going to do that ideal thing? Last election day, did you, Chris Freiman, "work some overtime and earn, say, $50 to donate to the Seva Foundation"? Or did you just not vote, thus contributing less to society that day than an informed person who just did the boring thing and voted?
Voting is a collective action problem, because the individual voter gains essentially no benefit, but added up, informed voting makes society richer.
As a result, informed voting should be seen as the contribution to society that it is.
Regarding whether an "example can be set" I think it's quite clear that culture is contagious. Many social groups / cultural tribes have a norm where you must vote -- and as a result, their tribe's say in the policy process is higher.
To the extent that a tribe of people who support freedom or rationality exists, it would be good for society if they maximize their say in politics by having a cultural norm of voting.
Informed voting is good, but there is lots of evidence that almost none of the actual voting that happens is informed. If one is choosing between uninformed voting and nothing, nothing comes out ahead as uninformed voting is negative value. The problem with the modern USA set up is that being informed is a massive amount of work, due to the number of issues and decisions made by government. So switching between “voting” and “informed voting” will lead you astray.
What does "informed" even mean? There are lots of people filled with misinformation, or biased, one sided information, or those who are highly opinionated in ways that are inconsistent, hypocritical, or are against their own self interest. I agree that many or most people should probably not be voting and are just adding noise (possibly biased noise at that). At the least, we should stop encouraging people to vote who don't even understand the issues or have a strong opinion.
Indeed. I think I am using a slightly different version of informed than Maxim. I would say "high value" informed votes are cast by people who know the issues at hand, know the politicians, and know the state of the world/relevant fields of expertise well enough to make educated decisions, decisions where their vote is instrumental in a "I am voting for X because he will vote for A, B and C, while voting against Y, Z, and I want those positions because I am really quite sure those are the best positions based on my learnings."
So in my case I know a lot about economics, but if I don't know the voting history of local politicians on economic matters, I probably shouldn't vote because I don't know enough to use that knowledge. Voting for a president who promises to defund the police to solve crime is a bad idea, partly because the president can't really do that, and partly because it probably wouldn't solve crime in the way people hope. Net neutrality was never going to be an issue government should make decisions on, because it isn't clear that anyone really understood what the hell the question even was, let alone what the answer should be.
Anyway, short version of "informed" is "has knowledge about all relevant issues, and candidates such that votes cast for a candidate that leads to the candidates election are likely to lead to the desired legal outcomes, which are likely to lead to the desired outcomes in reality." Informed voters must know "Do we have problem X? If so, should we do Y to solve X? If we should do Y, will candidate Z reliably vote for Y should he be elected?"
I think that is the strongest argument against the governmentalizing of all of our affairs: who the hell has time to learn all that such that they can be effective voters? Government by the people requires a lot of the people, and that increases exponentially with the number of decisions made by government.
That is the ideal, but I think the value drops away pretty quickly once you get away from it.
Even considering the median voter is terribly uninformed, effectively voting at random for representatives, you have to consider how many potential issues a representative could decide upon should you vote for them. How many things can you be knowledgeable about? Single issue voters are not good, but 2-3 issue voters are not obviously a lot better. Plus, how do you know that potential candidates will vote the way your knowledge of the issues suggests they should?
Plus, what if you are well informed and absolutely objectively correct on 2-3 issues, but dead wrong on a few others? Are you sure you are adding value? You had better hope the 3 issues you are right on come up and the others don't, or at least that the benefit of your representative making the correct vote on those issues outweigh the harm from the others.
It is really important to grasp the multidimensionality of the issue space when considering voting for representatives. The problem with much of political science theory is that it tries to simplify things really hard. Take, for example, the median voter theorem. It assumes everyone votes and just a single dimension of decision. Those make it much more tractable, but pretty much useless in predicting real life where "getting the vote out" is more important than finding that sweet spot of dividing the electorate.
I should say that I do generally vote when I know a bit about the politicians and the relevant issues, but I don't assign a high likelihood of anything valuable coming out of it, even a high expected value, and just try my best. The only way to fix the problem is to lower the range of things to know, and until that happens at best the outcome of anyone voting is "perhaps not net negative." If someone is interested in maximizing the good done by their time, they definitely should focus on something else.
I don't think being informed is as hard as you make it out to be, for the simple fact that your options are limited. If all the candidates running are 80% similar, you only need to focus on the (relatively) small number of differences.
For instance, take the conflict in Ukraine. You could try to become informed, and learn all about the state of the war, and foreign policy in general. Or you could realize that the vast majority of federal politicians will happily spend billions of dollars in military aid to any foreign country no matter what. Consequently, this issue is irrelevant to most people, because they don't have an option of voting for someone who won't do that.
If you look at things this way, it narrows down the issues that you have to worry about. It's not a good situation, but it does make it easier to know how to vote.
You can assert that … but other smart people (as measured by any objective standard) disagree with you, and we actually get a say in which leader is less bad
Voting is meaningless when there are only two choices and there's little difference between them. If that's your system, voting is harmful because it endorses a bad system. In the case of Trump, Biden, Hillary, and Obama, all of them believe that the secret of prosperity is to print money rather than set spending priorities to balance the federal budget; so it makes no difference who is elected.
Maybe that's the point. Cut out the public service ads guilting people into voting. Give them "permission" to admit that they aren't engaged and should stay out of it.
I rather disagree on that; see my reply to Chuck above. I think the range of decisions getting made by government officials and representatives is so vast that none of us can really claim to be fully informed, and even being better than the median doesn't mean much against the ocean of ignorance we face when considering the actions of government. The only real solution is to drastically limit what government can do, and thus what we need to learn about in order to be fairly close to fully informed.
Of course, the chance your vote makes a difference in the outcome is pretty much nil anyway, so meh. Then again, if all low information voters stopped voting the probabilities improve. But very few people admit to being low information, so it would be hard to get enough people to opt out. Ceasing the endless "VOTE OR BE DAMNED" propaganda might help though :)
You've got it right, you should weigh the goodness of voting against whatever else you would actually do. I think some people do in fact just work more when more of their time is free, and donate more when they have more money. For such a person, the earning to give baseline is actually a good thing to make decisions against.
Nearly all voting is uninformed and voting is the greatest evil we suffer today. We can buy quality food and cars and trucks because grocery stores and car dealers aren't government employees. If the government managed those businesses, we'd eat bad food and drive in dangerous cars that get poor gas mileage. And, if you choose to buy bad food in the grocery store, it's available for you but I don't have to eat the same food that you eat.
As Winston Churchill put it, “democracy is the worst form of Government except for all those other forms that have been tried.”
Until prediction markets mature, or Putin becomes all-wise (sarcasm), voting combined with institutional guardrails is the best mechanism which territorial security providers (countries) have found
You didn't attack his argument. The question is this: All else being equal, how should an individual attempting to maximally benefit society allocate their time? Vote, or do something else? You're ignoring the "all else being equal" part.
Issue is, humans don’t care only about benefitting society — so we are more likely to use the prefect as an excuse not to do something good, rather than actually do the perfect, harder thing instead of the good thing.
Case in point, did the author actually work overtime, or donate extra, on election day instead of voting? I’d bet he didn’t. Nothing against him — that’s what most people will do if they buy his line of reasoning.
If you are actually going to do the overtime/donate thing though, good for you!
I think the best argument against this view is that even though the chance of your vote changing the election is tiny, the expected value of voting still seems to be high. Suppose there's a one in one million chance of my vote changing the outcome of the election in the UK - if having a different government in power changes the way that £100bn is spent in a way that I think is preferable to how some alternative government would have spent it, then the EV of voting is 0.000001 * 100,000,000,000 = £100,000. So voting seems like an immensely high EV decision, even though the raw chance your vote makes a difference is extremely low.
Although my numbers may be some way off the true numbers, the fact that David Shor (who is very good with numbers and thinks about the impact of voting/elections a lot) has made the same argument is reassuring: https://twitter.com/davidshor/status/1425506809788477440
In the most decisive states, the probability of being the pivotal vote was around 1 in 10 million. If you believe the benefit of your preferred candidate winning is $1000 per American, then the value of your vote can be $30,000 if you live in the right state.
I can see why you would apply a median valuation model to elections, but I think this specific application is inappropriate. We're not talking about the median population making decisions we're talking about a small percentage of people being swayed either way, so the average consequence would be the same regardless. What we are focusing on than is the cost to an individual vs what should be expected to be gained (In a true sense so say no to billion-dollar trades where we could earn an infinite amount, but the percent chance to earn that is near infinitely small). You will almost certainly no gain anything from doing this other than to say you voted.
Beyond that in most cases people have horrible ideas on how to run an economic system in general. So, for most people the gain isn't positive, it's negative (Left with things like price controls, and the right with things like anti-Immagration). It would be better that the median voter didn't vote so that the non-median voter held more sway (The third largest party in America, I think, is Libertarians; they tend to be more knowledgeable about economics). Bryan's book "Myth of the Rational Voter" was made to point this out.
I always find this type of EV analysis to be an abuse of math. You don't actually do anything unless you are the reason and almost no one will be in their lifetime. You can argue that a few bucks spent in the lottery will be better than investing or anything else, because of the one-off chance you win Hundreds of Millions, but you would be a fool to try to win and in almost every universe you try you would fail and just lose a couple of bucks. Same for Cosinos where the games are rigged against you.
Under a risk-neutral consequentialist value system, I think my math is reasonable. If the goal is to maximize the expected utility of the universe for example, then the probability of having in impact is irrelevant, only the mean impact is, which is what I calculated. Perhaps you have a value system that incorporates risk aversion, if so I'm curious to hear about it.
I don't think the lottery example is a fair comparison because unlike with voting, the mean payout is les than what you put in.
Regarding "Risk aversion" I set up this scenario: Assume that any quantity gained is better than 0 quantity gained. If you had a choice to gain an infinite amount, but risk an infinite amount what would be the best choice? Assume a probability of near infinitesimally small amount and the ability to guarantee a gain of some kind outside of this bet.
Of course, the right answer would be to avoid the bet and pick the guarantee. The reason is because "any quantity gained is better than 0 quantity gained". This is a world closer to our reality; if I bet the house on a small percentage chance to triple my net worth, I would be foolish to do so, because I would almost always end up better off not doing that and the lost would be great if I did lose.
I don’t think this is right. Most people intuitively vote because they think it’s valuable, the people arguing not to vote are the ones who say it’s irrational because the probability is so low. But it isn’t irrational under a fairly standard decision-making framework! So, neither the people who act on their intuitions nor the people who are fully rational reject the idea that voting is good, meaning your ‘you don’t actually do anything unless you are the reason’ claim doesn’t work - most people do vote!
First, you don’t multiple odds of changing outcome by government budget. You need to multiply by the delta between the two in terms of “doing good things.” My guess is it is a lot less than 100b.
Second and more important it is probably true the right government could be a massive difference in value for the populace. But the problem is an ex ante problem of who is the right government.
If the election came down to your vote that means a great many people rejected your view. Indeed, we effectively have a tie.
But we are to believe you possess knowledge this great multitude lacked such that your vote is awesome. It just as easily could be that you pick the wrong candidate making your vote a terrible thing for society.
No, the right way to think about the benefit is (1) what is the delta in benefit between admin 1 and 2, and (2) odds I’m doing better than the crowds which are tied 50-50. We then multiply that by the odds of your vote impacting the outcome.
My intuition is that when you do that math with a little humility the answer is your vote is probably worth less than ten dollars (more likely cents on the dollar).
Yeah, I mean the assumption isn’t that the reader’s preferences are objectively correct in some sense but simply that they’re better off in some way than if their preferences weren’t enacted. It’s taken as given that the reader benefits from money being spent in the way that they prefer.
Are you saying how the election effects the specific individual? In that case, I imagine the impact one way or the other is relatively small especially when discounted for odds of the election being decided by the voter.
As a recent convert to voting who still doesn't think of it as a rational act the signalling value wins out for me. Friends knowing I don't vote will lower their opinion of me and make them less likely to listen to me in more consequential matters, such as donating precious charity dollars to the most effective operators.
I also live in California, where ballots are now mailed to my house automatically and can be returned virtually anywhere, so voting is very low-effort and not doing so would allow me very little time to earn money that might go to more productive ends. There are modest norms against discussing who one votes for, so the choices I make don't even matter all that much to my ends and I mostly follow the suggestions of trusted elites.
1. Wearing a false/fake "I voted" sticker is practically sociopathic.
2. You voting may not matter but "Bryan Caplan's audience" or "people who can follow decision theoretic arguments about opportunity costs and effective altruism"... those are pretty bad subsets of the electorate to disenfranchise!
3. You're at least correct about the "what if *no one* voted??" rebuttal being dumb. A better retort is "ah well in that case I *would* vote, since I could single-handedly decide the election". People say this not because it's a coherent argument for voting but because they're enforcing the norm that voting is a civic duty. Which brings me back to point 2. We in the econ nerd community, or whatever we are, need to also have that norm. Because "econ nerds tend not to vote" would be a pretty bad outcome for our democracy. Which means you deserve scorn and ridicule for writing this essay!
4. I do agree about reading the news though. I was first convinced of this by a classic essay by Aaron Swartz: http://www.aaronsw.com/weblog/hatethenews . Relevant excerpt:
> To become an informed voter all one needs to do is read a short guide about the candidates and issues before the election. There’s no need to have to suffer through the daily back-and-forth of allegations and counter-allegations, of scurrilous lies and their refutations. Indeed, reading a voter’s guide is much better: there’s no recency bias (where you only remember the crimes reported in the past couple months), you get to hear both sides of the story after the investigation has died down, you can actually think about the issues instead of worrying about the politics.
"Wearing a false/fake "I voted" sticker is practically sociopathic."
What about going in, casting a blank ballot, then wearing the sticker?
Or what I did in 2016...I cut out the e and o and reversed them, wearing a "I vetod" sticker. Yes, it was spelled wrong, but the few people who noticed laughed.
"...which actually suggests that at least I am not influential" Love it - worth the read for that alone!
I agree with the post, really. Still, it's ironic that if there could be any outcome beyond rhetoric from making this argument very persuasively, it would likely be to marginally discourage the most reasonable voters from voting. Further irony: any such effect would likely be negligible. I doubt you would actually convince people to do something beyong just not voting.
But after reading this and commenting, I'm feeling unusually civic - I'm going to go order some "I didn't vote" stickers.
So clearly written by a conservative propagandist targeting young voters who are likely to sway elections in more liberal directions. Sick that McCombs approves sending this out to students but not surprising. Please always look at "advice" such as this with a healthy amount of skepticism. They do not care about the "energy you're wasting" being involved in politics, they just don't want your vote interfering with their agenda.
I don't have any great arguments to counter the one's that Chris makes, but his first example is quite a head scratcher. Let's put it this way: how many times, reader, have you ever found yourself even remotely in a position where voting kept you from doing some immediately beneficial task? Maybe I'm weird since I live in a city and voting has always been a matter of stopping off on my way to the subway before work - usually a total of 3-5 minutes, but what other important activity did I miss? I can still do whatever effective altruism research I was planning to do while waiting in line, right?
At any rate, I thought, why not take a look at some evidence that voting is actually important? What would demonstrate that? The OECD library has tables that give voting participation and trust in government, so I took the tables, cross indexed against country and found a generally upward sloping line that indicates that where voting participation is highest, trust in government is highest. Maybe you think trust in government isn't important, but just looking at the list, the countries where trust in government was high tended to be countries that I found would be attractive places to live if I were inclined to move.
The one outlier on the far bottom, right is Belgium. High voting participation, low trust in government. My guess is that's because of the roughly even split between French and Dutch speakers. Neither group trusts the government when so many of the other group are in it so they go to the polls to ensure their own group is on top. Maybe, but that's just a guess.
I'd say this argument supports Frieman's position. People shouldn't trust government, so if lower voter participation = lower trust, then that's a preferred outcome.
I would agree with you if trust in government meant "blind faith irrespective of the actions of your government", but I don't think that's the correct interpretation of the statistic. I think a better interpretation is "a government that has proven through its actions to be one in which you can have trust". A cursory glance at the most trusted governments (Switzerland, Norway, Germany, New Zealand) vs those with least trust (Chile, Poland, Columbia, Latvia - and recall that this list doesn't cover the real basket cases of the world like Somalia, North Korea, etc) seems to bear this notion out.
Truthfully, there are several statistics I could have used to illustrate the point. For example, Health Care outcomes, Education, etc.
> imagine that you’re on your way to the polls to cast a vote in the next presidential election.
A presidetial election is where an individual's vote has the least effect possible.
A more iinterestingcade would be,say, a single local ordinance referendum in a primary during an odd (in the US) year, where your vitehas a chance of being the one vote that makes the difference)
I'm also interested in Effective Altruism and this article by 80,000 hours seems to disagree with your opinion that voting does not matter.
"Again using US to illustrate, over the next four years6 the US federal government will spend about $17.5 trillion.
Written out as a number it looks like $17,500,000,000,000. That’s $53,000 for each American, or $129,000 for each vote cast in 2016.
If you multiply all that spending through a 1 in 10 million chance of changing the outcome, in a swing state like New Hampshire, it comes to $1.75 million. That’s the fraction of the budget you might ‘expect’ to influence by voting in a swing state, in the statistical sense of expectation."
I haven't been engaged academic philosophy since college, but I thought that Jonathan Glover's Baked Bean Bandits was a classic counter example to claims that relied on "actions are too insignificant to "make a difference"" type arguments. Was I misinformed, or is this no longer (never was?) the dominant take? Whats the counter-counter argument these days?
Altruism suffers from an “information problem,” like that under which planners in a socialist system labor: the information needed for reliably effective altruism is too hard to obtain. Instead, for the most part each person should focus on matters close to home—private affairs, where he is more likely to know what is really going on. As for influencing the larger society: putting one’s efforts to a market test is the best way to get some information about how much good he is doing. For example, someone who can earn $75,000/year as an electrician but only $50,000 as a schoolteacher should take that as a sign that his contribution to society as an electrician is the greater.
Thanks for the post and nice to read about your work and book. I will try to read it at some point, partly because I am working with ideas of liquid and decentralised democracy, politics without politicians and political parties, uberisation of politics etc. The future of democracy is very much about humans having more free time for meaningful things and being able to delegate their votes to different individuals and communities https://vladanlausevic.medium.com/abc-of-liquid-democracy-82c4ea85bc1
Even if 100% of voting were attributable to influence-by-example, it would still be impossible for the *average* voter to have influenced more than one other person to vote. So even if the influence argument were true it could only multiply the utility of voting by <2 unless you are exceptional in some way.
I'm unconvinced. I always felt that if I were very wealthy, one of the best things I could do with my money would be to donate to political causes that will keep (or move) the US as free as possible. Maintaining free markets, free speech, gun rights and all the things that make progress and prosperity possible is huge. One could coldly argue that the long term utility of keeping our institutions working properly, especially in the face of the seeming one way ratchet away from freedom, far outweighs the utility of saving some people from malaria right now.
This is a "perfect is the enemy of the good" fallacy.
Yes, you COULD do more by working/donating than voting. But you could apply that logic to just about any good thing someone does -- there's almost always something better. The question is, ARE you going to do that ideal thing? Last election day, did you, Chris Freiman, "work some overtime and earn, say, $50 to donate to the Seva Foundation"? Or did you just not vote, thus contributing less to society that day than an informed person who just did the boring thing and voted?
Voting is a collective action problem, because the individual voter gains essentially no benefit, but added up, informed voting makes society richer.
As a result, informed voting should be seen as the contribution to society that it is.
Regarding whether an "example can be set" I think it's quite clear that culture is contagious. Many social groups / cultural tribes have a norm where you must vote -- and as a result, their tribe's say in the policy process is higher.
To the extent that a tribe of people who support freedom or rationality exists, it would be good for society if they maximize their say in politics by having a cultural norm of voting.
Informed voting is good, but there is lots of evidence that almost none of the actual voting that happens is informed. If one is choosing between uninformed voting and nothing, nothing comes out ahead as uninformed voting is negative value. The problem with the modern USA set up is that being informed is a massive amount of work, due to the number of issues and decisions made by government. So switching between “voting” and “informed voting” will lead you astray.
What does "informed" even mean? There are lots of people filled with misinformation, or biased, one sided information, or those who are highly opinionated in ways that are inconsistent, hypocritical, or are against their own self interest. I agree that many or most people should probably not be voting and are just adding noise (possibly biased noise at that). At the least, we should stop encouraging people to vote who don't even understand the issues or have a strong opinion.
Indeed. I think I am using a slightly different version of informed than Maxim. I would say "high value" informed votes are cast by people who know the issues at hand, know the politicians, and know the state of the world/relevant fields of expertise well enough to make educated decisions, decisions where their vote is instrumental in a "I am voting for X because he will vote for A, B and C, while voting against Y, Z, and I want those positions because I am really quite sure those are the best positions based on my learnings."
So in my case I know a lot about economics, but if I don't know the voting history of local politicians on economic matters, I probably shouldn't vote because I don't know enough to use that knowledge. Voting for a president who promises to defund the police to solve crime is a bad idea, partly because the president can't really do that, and partly because it probably wouldn't solve crime in the way people hope. Net neutrality was never going to be an issue government should make decisions on, because it isn't clear that anyone really understood what the hell the question even was, let alone what the answer should be.
Anyway, short version of "informed" is "has knowledge about all relevant issues, and candidates such that votes cast for a candidate that leads to the candidates election are likely to lead to the desired legal outcomes, which are likely to lead to the desired outcomes in reality." Informed voters must know "Do we have problem X? If so, should we do Y to solve X? If we should do Y, will candidate Z reliably vote for Y should he be elected?"
I think that is the strongest argument against the governmentalizing of all of our affairs: who the hell has time to learn all that such that they can be effective voters? Government by the people requires a lot of the people, and that increases exponentially with the number of decisions made by government.
what you describe seems to me more a platonic ideal than a realistic bar.
If you are closer to the ideal you lay out than the median voter is, then your vote adds value, so that should be the bar IMO.
That is the ideal, but I think the value drops away pretty quickly once you get away from it.
Even considering the median voter is terribly uninformed, effectively voting at random for representatives, you have to consider how many potential issues a representative could decide upon should you vote for them. How many things can you be knowledgeable about? Single issue voters are not good, but 2-3 issue voters are not obviously a lot better. Plus, how do you know that potential candidates will vote the way your knowledge of the issues suggests they should?
Plus, what if you are well informed and absolutely objectively correct on 2-3 issues, but dead wrong on a few others? Are you sure you are adding value? You had better hope the 3 issues you are right on come up and the others don't, or at least that the benefit of your representative making the correct vote on those issues outweigh the harm from the others.
It is really important to grasp the multidimensionality of the issue space when considering voting for representatives. The problem with much of political science theory is that it tries to simplify things really hard. Take, for example, the median voter theorem. It assumes everyone votes and just a single dimension of decision. Those make it much more tractable, but pretty much useless in predicting real life where "getting the vote out" is more important than finding that sweet spot of dividing the electorate.
I should say that I do generally vote when I know a bit about the politicians and the relevant issues, but I don't assign a high likelihood of anything valuable coming out of it, even a high expected value, and just try my best. The only way to fix the problem is to lower the range of things to know, and until that happens at best the outcome of anyone voting is "perhaps not net negative." If someone is interested in maximizing the good done by their time, they definitely should focus on something else.
I don't think being informed is as hard as you make it out to be, for the simple fact that your options are limited. If all the candidates running are 80% similar, you only need to focus on the (relatively) small number of differences.
For instance, take the conflict in Ukraine. You could try to become informed, and learn all about the state of the war, and foreign policy in general. Or you could realize that the vast majority of federal politicians will happily spend billions of dollars in military aid to any foreign country no matter what. Consequently, this issue is irrelevant to most people, because they don't have an option of voting for someone who won't do that.
If you look at things this way, it narrows down the issues that you have to worry about. It's not a good situation, but it does make it easier to know how to vote.
Anyone who votes for Trump or Biden is uninformed, or about 98% of voters in the last election.
You can assert that … but other smart people (as measured by any objective standard) disagree with you, and we actually get a say in which leader is less bad
Voting is meaningless when there are only two choices and there's little difference between them. If that's your system, voting is harmful because it endorses a bad system. In the case of Trump, Biden, Hillary, and Obama, all of them believe that the secret of prosperity is to print money rather than set spending priorities to balance the federal budget; so it makes no difference who is elected.
I disagree — I think there are major differences between them in policy, sanity, and competence.
But yes, if they are equivalent to you, then it doesn’t make much sense for you to vote.
Everyone must judge for themselves whether their opinion is really better than the median voter. Ideally subtract a bit for overconfidence bias.
Maybe that's the point. Cut out the public service ads guilting people into voting. Give them "permission" to admit that they aren't engaged and should stay out of it.
I’m all in favor of that. I also believe the point in this post was quite different.
I agree!
I'll note that we should define "informed" or not, as compared to the median voter.
I'll also note that most (though of course not all) blog-reader types already meet that bar by virtue of their intellectual hobbies.
I rather disagree on that; see my reply to Chuck above. I think the range of decisions getting made by government officials and representatives is so vast that none of us can really claim to be fully informed, and even being better than the median doesn't mean much against the ocean of ignorance we face when considering the actions of government. The only real solution is to drastically limit what government can do, and thus what we need to learn about in order to be fairly close to fully informed.
Of course, the chance your vote makes a difference in the outcome is pretty much nil anyway, so meh. Then again, if all low information voters stopped voting the probabilities improve. But very few people admit to being low information, so it would be hard to get enough people to opt out. Ceasing the endless "VOTE OR BE DAMNED" propaganda might help though :)
You've got it right, you should weigh the goodness of voting against whatever else you would actually do. I think some people do in fact just work more when more of their time is free, and donate more when they have more money. For such a person, the earning to give baseline is actually a good thing to make decisions against.
Nearly all voting is uninformed and voting is the greatest evil we suffer today. We can buy quality food and cars and trucks because grocery stores and car dealers aren't government employees. If the government managed those businesses, we'd eat bad food and drive in dangerous cars that get poor gas mileage. And, if you choose to buy bad food in the grocery store, it's available for you but I don't have to eat the same food that you eat.
As Winston Churchill put it, “democracy is the worst form of Government except for all those other forms that have been tried.”
Until prediction markets mature, or Putin becomes all-wise (sarcasm), voting combined with institutional guardrails is the best mechanism which territorial security providers (countries) have found
You didn't attack his argument. The question is this: All else being equal, how should an individual attempting to maximally benefit society allocate their time? Vote, or do something else? You're ignoring the "all else being equal" part.
I think that's a bad question, which likely leads to *worse* outcomes, as I lay out in my comment.
As another commenter rightly says: "you should weigh the goodness of voting against whatever else you would actually do."
Why is the question bad? For a resource-constrained actor who cares about benefiting society it is the ONLY relevant question.
Again, you and Jalex Stark are ignoring the "all else being equal" part.
Google “perfect is the enemy of the good fallacy”
Issue is, humans don’t care only about benefitting society — so we are more likely to use the prefect as an excuse not to do something good, rather than actually do the perfect, harder thing instead of the good thing.
Case in point, did the author actually work overtime, or donate extra, on election day instead of voting? I’d bet he didn’t. Nothing against him — that’s what most people will do if they buy his line of reasoning.
If you are actually going to do the overtime/donate thing though, good for you!
I'm aware of the saying. Every MBA Business Bill uses it twice a day. It's not a logical fallacy. Quit attempting to apply it as such.
"Issue is, humans don’t care only about benefitting society..." - Ok, so what are we even arguing here? What is the motivation for voting?
It’s pretty simple… if you can get yourself to informed-vote, do that, because it’s good.
If you can get yourself to work overtime and donate $50, do that, because it’s even better.
If you can manage both, even better still.
If you can only do the first thing, that’s fine too.
Maybe we agree.
I think the best argument against this view is that even though the chance of your vote changing the election is tiny, the expected value of voting still seems to be high. Suppose there's a one in one million chance of my vote changing the outcome of the election in the UK - if having a different government in power changes the way that £100bn is spent in a way that I think is preferable to how some alternative government would have spent it, then the EV of voting is 0.000001 * 100,000,000,000 = £100,000. So voting seems like an immensely high EV decision, even though the raw chance your vote makes a difference is extremely low.
Although my numbers may be some way off the true numbers, the fact that David Shor (who is very good with numbers and thinks about the impact of voting/elections a lot) has made the same argument is reassuring: https://twitter.com/davidshor/status/1425506809788477440
To reinforce your point, here's a paper by Andrew Gelman and Nate Silver about the probability of a vote being pivotal in the 2008 election: http://www.stat.columbia.edu/~gelman/research/published/probdecisive2.pdf
In the most decisive states, the probability of being the pivotal vote was around 1 in 10 million. If you believe the benefit of your preferred candidate winning is $1000 per American, then the value of your vote can be $30,000 if you live in the right state.
I can see why you would apply a median valuation model to elections, but I think this specific application is inappropriate. We're not talking about the median population making decisions we're talking about a small percentage of people being swayed either way, so the average consequence would be the same regardless. What we are focusing on than is the cost to an individual vs what should be expected to be gained (In a true sense so say no to billion-dollar trades where we could earn an infinite amount, but the percent chance to earn that is near infinitely small). You will almost certainly no gain anything from doing this other than to say you voted.
Beyond that in most cases people have horrible ideas on how to run an economic system in general. So, for most people the gain isn't positive, it's negative (Left with things like price controls, and the right with things like anti-Immagration). It would be better that the median voter didn't vote so that the non-median voter held more sway (The third largest party in America, I think, is Libertarians; they tend to be more knowledgeable about economics). Bryan's book "Myth of the Rational Voter" was made to point this out.
I always find this type of EV analysis to be an abuse of math. You don't actually do anything unless you are the reason and almost no one will be in their lifetime. You can argue that a few bucks spent in the lottery will be better than investing or anything else, because of the one-off chance you win Hundreds of Millions, but you would be a fool to try to win and in almost every universe you try you would fail and just lose a couple of bucks. Same for Cosinos where the games are rigged against you.
Under a risk-neutral consequentialist value system, I think my math is reasonable. If the goal is to maximize the expected utility of the universe for example, then the probability of having in impact is irrelevant, only the mean impact is, which is what I calculated. Perhaps you have a value system that incorporates risk aversion, if so I'm curious to hear about it.
I don't think the lottery example is a fair comparison because unlike with voting, the mean payout is les than what you put in.
Regarding "Risk aversion" I set up this scenario: Assume that any quantity gained is better than 0 quantity gained. If you had a choice to gain an infinite amount, but risk an infinite amount what would be the best choice? Assume a probability of near infinitesimally small amount and the ability to guarantee a gain of some kind outside of this bet.
Of course, the right answer would be to avoid the bet and pick the guarantee. The reason is because "any quantity gained is better than 0 quantity gained". This is a world closer to our reality; if I bet the house on a small percentage chance to triple my net worth, I would be foolish to do so, because I would almost always end up better off not doing that and the lost would be great if I did lose.
I don’t think this is right. Most people intuitively vote because they think it’s valuable, the people arguing not to vote are the ones who say it’s irrational because the probability is so low. But it isn’t irrational under a fairly standard decision-making framework! So, neither the people who act on their intuitions nor the people who are fully rational reject the idea that voting is good, meaning your ‘you don’t actually do anything unless you are the reason’ claim doesn’t work - most people do vote!
I don’t think this is close to correct.
First, you don’t multiple odds of changing outcome by government budget. You need to multiply by the delta between the two in terms of “doing good things.” My guess is it is a lot less than 100b.
Second and more important it is probably true the right government could be a massive difference in value for the populace. But the problem is an ex ante problem of who is the right government.
If the election came down to your vote that means a great many people rejected your view. Indeed, we effectively have a tie.
But we are to believe you possess knowledge this great multitude lacked such that your vote is awesome. It just as easily could be that you pick the wrong candidate making your vote a terrible thing for society.
No, the right way to think about the benefit is (1) what is the delta in benefit between admin 1 and 2, and (2) odds I’m doing better than the crowds which are tied 50-50. We then multiply that by the odds of your vote impacting the outcome.
My intuition is that when you do that math with a little humility the answer is your vote is probably worth less than ten dollars (more likely cents on the dollar).
Yeah, I mean the assumption isn’t that the reader’s preferences are objectively correct in some sense but simply that they’re better off in some way than if their preferences weren’t enacted. It’s taken as given that the reader benefits from money being spent in the way that they prefer.
Are you saying how the election effects the specific individual? In that case, I imagine the impact one way or the other is relatively small especially when discounted for odds of the election being decided by the voter.
As a recent convert to voting who still doesn't think of it as a rational act the signalling value wins out for me. Friends knowing I don't vote will lower their opinion of me and make them less likely to listen to me in more consequential matters, such as donating precious charity dollars to the most effective operators.
I also live in California, where ballots are now mailed to my house automatically and can be returned virtually anywhere, so voting is very low-effort and not doing so would allow me very little time to earn money that might go to more productive ends. There are modest norms against discussing who one votes for, so the choices I make don't even matter all that much to my ends and I mostly follow the suggestions of trusted elites.
1. Wearing a false/fake "I voted" sticker is practically sociopathic.
2. You voting may not matter but "Bryan Caplan's audience" or "people who can follow decision theoretic arguments about opportunity costs and effective altruism"... those are pretty bad subsets of the electorate to disenfranchise!
3. You're at least correct about the "what if *no one* voted??" rebuttal being dumb. A better retort is "ah well in that case I *would* vote, since I could single-handedly decide the election". People say this not because it's a coherent argument for voting but because they're enforcing the norm that voting is a civic duty. Which brings me back to point 2. We in the econ nerd community, or whatever we are, need to also have that norm. Because "econ nerds tend not to vote" would be a pretty bad outcome for our democracy. Which means you deserve scorn and ridicule for writing this essay!
4. I do agree about reading the news though. I was first convinced of this by a classic essay by Aaron Swartz: http://www.aaronsw.com/weblog/hatethenews . Relevant excerpt:
> To become an informed voter all one needs to do is read a short guide about the candidates and issues before the election. There’s no need to have to suffer through the daily back-and-forth of allegations and counter-allegations, of scurrilous lies and their refutations. Indeed, reading a voter’s guide is much better: there’s no recency bias (where you only remember the crimes reported in the past couple months), you get to hear both sides of the story after the investigation has died down, you can actually think about the issues instead of worrying about the politics.
"Wearing a false/fake "I voted" sticker is practically sociopathic."
What about going in, casting a blank ballot, then wearing the sticker?
Or what I did in 2016...I cut out the e and o and reversed them, wearing a "I vetod" sticker. Yes, it was spelled wrong, but the few people who noticed laughed.
"...which actually suggests that at least I am not influential" Love it - worth the read for that alone!
I agree with the post, really. Still, it's ironic that if there could be any outcome beyond rhetoric from making this argument very persuasively, it would likely be to marginally discourage the most reasonable voters from voting. Further irony: any such effect would likely be negligible. I doubt you would actually convince people to do something beyong just not voting.
But after reading this and commenting, I'm feeling unusually civic - I'm going to go order some "I didn't vote" stickers.
Subjective value, anyone?
So clearly written by a conservative propagandist targeting young voters who are likely to sway elections in more liberal directions. Sick that McCombs approves sending this out to students but not surprising. Please always look at "advice" such as this with a healthy amount of skepticism. They do not care about the "energy you're wasting" being involved in politics, they just don't want your vote interfering with their agenda.
I don't have any great arguments to counter the one's that Chris makes, but his first example is quite a head scratcher. Let's put it this way: how many times, reader, have you ever found yourself even remotely in a position where voting kept you from doing some immediately beneficial task? Maybe I'm weird since I live in a city and voting has always been a matter of stopping off on my way to the subway before work - usually a total of 3-5 minutes, but what other important activity did I miss? I can still do whatever effective altruism research I was planning to do while waiting in line, right?
At any rate, I thought, why not take a look at some evidence that voting is actually important? What would demonstrate that? The OECD library has tables that give voting participation and trust in government, so I took the tables, cross indexed against country and found a generally upward sloping line that indicates that where voting participation is highest, trust in government is highest. Maybe you think trust in government isn't important, but just looking at the list, the countries where trust in government was high tended to be countries that I found would be attractive places to live if I were inclined to move.
A graph of the data is here: https://lh6.googleusercontent.com/_-eBRT-oE2gNscB6JNIrzPhA2neuYESWGmMG7SIQJyWgpRonLuMaPP2DnaIc8NBE4DCwmXzzJQB3zz-K3qIQHys8z0GUpkWB5T8-cUSDC4Oh8nvXlHZdlT_K3Hc9tKYu3w=w1280.
The one outlier on the far bottom, right is Belgium. High voting participation, low trust in government. My guess is that's because of the roughly even split between French and Dutch speakers. Neither group trusts the government when so many of the other group are in it so they go to the polls to ensure their own group is on top. Maybe, but that's just a guess.
I'd say this argument supports Frieman's position. People shouldn't trust government, so if lower voter participation = lower trust, then that's a preferred outcome.
I would agree with you if trust in government meant "blind faith irrespective of the actions of your government", but I don't think that's the correct interpretation of the statistic. I think a better interpretation is "a government that has proven through its actions to be one in which you can have trust". A cursory glance at the most trusted governments (Switzerland, Norway, Germany, New Zealand) vs those with least trust (Chile, Poland, Columbia, Latvia - and recall that this list doesn't cover the real basket cases of the world like Somalia, North Korea, etc) seems to bear this notion out.
Truthfully, there are several statistics I could have used to illustrate the point. For example, Health Care outcomes, Education, etc.
> imagine that you’re on your way to the polls to cast a vote in the next presidential election.
A presidetial election is where an individual's vote has the least effect possible.
A more iinterestingcade would be,say, a single local ordinance referendum in a primary during an odd (in the US) year, where your vitehas a chance of being the one vote that makes the difference)
Typical CDT nonsense: https://handsandcities.com/2021/08/27/can-you-control-the-past/
https://80000hours.org/articles/is-voting-important/
Hi Chris,
I'm also interested in Effective Altruism and this article by 80,000 hours seems to disagree with your opinion that voting does not matter.
"Again using US to illustrate, over the next four years6 the US federal government will spend about $17.5 trillion.
Written out as a number it looks like $17,500,000,000,000. That’s $53,000 for each American, or $129,000 for each vote cast in 2016.
If you multiply all that spending through a 1 in 10 million chance of changing the outcome, in a swing state like New Hampshire, it comes to $1.75 million. That’s the fraction of the budget you might ‘expect’ to influence by voting in a swing state, in the statistical sense of expectation."
Keen to hear your thoughts on this.
I haven't been engaged academic philosophy since college, but I thought that Jonathan Glover's Baked Bean Bandits was a classic counter example to claims that relied on "actions are too insignificant to "make a difference"" type arguments. Was I misinformed, or is this no longer (never was?) the dominant take? Whats the counter-counter argument these days?
Original Paper:
https://www.jstor.org/stable/4106873
Less politics—fine. More altruism—maybe not.
Altruism suffers from an “information problem,” like that under which planners in a socialist system labor: the information needed for reliably effective altruism is too hard to obtain. Instead, for the most part each person should focus on matters close to home—private affairs, where he is more likely to know what is really going on. As for influencing the larger society: putting one’s efforts to a market test is the best way to get some information about how much good he is doing. For example, someone who can earn $75,000/year as an electrician but only $50,000 as a schoolteacher should take that as a sign that his contribution to society as an electrician is the greater.
Thanks for the post and nice to read about your work and book. I will try to read it at some point, partly because I am working with ideas of liquid and decentralised democracy, politics without politicians and political parties, uberisation of politics etc. The future of democracy is very much about humans having more free time for meaningful things and being able to delegate their votes to different individuals and communities https://vladanlausevic.medium.com/abc-of-liquid-democracy-82c4ea85bc1
Even if 100% of voting were attributable to influence-by-example, it would still be impossible for the *average* voter to have influenced more than one other person to vote. So even if the influence argument were true it could only multiply the utility of voting by <2 unless you are exceptional in some way.
I'm unconvinced. I always felt that if I were very wealthy, one of the best things I could do with my money would be to donate to political causes that will keep (or move) the US as free as possible. Maintaining free markets, free speech, gun rights and all the things that make progress and prosperity possible is huge. One could coldly argue that the long term utility of keeping our institutions working properly, especially in the face of the seeming one way ratchet away from freedom, far outweighs the utility of saving some people from malaria right now.