I think the strongest reply to this doesn’t apply to your readers: voters usually have systematically biased and incorrect beliefs on policy, which means voting selflessly rather than selfishly causes even more harm. This is because errors about the diverse interests of voters are more likely to be idiosyncratic in ways that cancel each other out, allowing the aggregate outcome to be biased in the direction of more-informed voters, and because voters are a bit less likely to be systematically wrong about their own lives (e.g., if the economy has improved under a candidate, they might falsely believe the economy is bad but they’re still more likely to be doing financially better). When voters are selfless and systematically biased about the collective interest, the different mistakes move in the same direction rather than diluting each other.
And I'd further argue that it still does apply to the readers (myself included), because even philosophically inclined folks or Rationalists or Effective Altruists are often unable to meaningfully escape their biases when voting.
I disagree. I’d much rather a rationalist/EA with tons of biases vote altruistically than selfishly, because as long as they’re just modestly better at reasoning about the issues they tend to prioritize (e.g., reducing AI risk or helping mitigate factory farming) than chance, the expected value is still much better than if they voted selfishly because the importance of issues follows a heavy-tailed distribution and hence is way more important than issues they would vote on if voting based on self interest.
Great points. I’m not sure about the local analysis here.
I live in San Francisco where elections have been determined by a few dozen votes. Local issues can be very impactful on people’s lives, like whether the highway next to someone’s house turns into a public park. (How much would you pay to keep a favorite park next to your home of 2 decades from closing down?)
There can also often be many issues on the ballot. Might’ve been 15+ candidates, ballot measures, etc. for my recent June election. So there’s multiple opportunities to shape outcomes for each voting effort.
Not saying this invalidates your overall thesis. I mainly think that local voting may be higher EV (personally and publicly) than implied here.
"(This article was edited. I’m thinking of hiring the person who edited it as an editor long-term, so let me know if you found the writing noticeably cleaner than usual)."
This question is hard to answer because of confounds. I might find another post less clear because the concepts it discusses are more difficult or because I spent less time thinking about it. (You said "clean," not "clear," but I imagine that above all you want your articles to be understood.)
Why don't you post the original version of the article, so that your readers can compare the two and decide to what extent the edited version is an improvement?
I'm generally okay with hand-waved numbers in arguments like these, but the first round of number wrangling feels way too casual to me. Are there really no scenarios where a candidate would (or you've convinced yourself they would) benefit you a significant amount? First thing that comes to my mind is job creation. If it's reasonable to believe that you'd be likely to secure higher-paying employment (given your resume) then you could easily stand to directly benefit tens of thousands over multiple years. I don't disagree that *the vast majority* of voters would only benefit at the level you've described, but the edge cases are big enough that they need to be addressed.
Second, I'd argue there are *very* few cases where a voter really knows that one candidate will be severely detrimental to the country, so that's a largely useless form of hypothetical weight. Instead, people generally believe that, at most, there will be a very mild detrimental effect (and even that is usually seen merely as "possible", not necessarily "likely").
Also if I believe that everyone else is only going to get 10¢ if I choose $1000, but 0¢ if I choose $2000, the moral math is less striking, and I don't think it's anywhere near as persuasive as the example of depriving everyone else of a 1k payout. You can make arguments about how 10¢ for ten million people is ultimately going to benefit me more than an extra 1k, because it means one million dollars going into the economy around you - but that argument wasn't made, merely that it would be immoral to deprive so many people of a direct payout.
But note that I didn't say you should never vote for the candidate that's better for you. I just said you shouldn't if it conflicts with the one who you expect to be best overall.
I think the strongest reply to this doesn’t apply to your readers: voters usually have systematically biased and incorrect beliefs on policy, which means voting selflessly rather than selfishly causes even more harm. This is because errors about the diverse interests of voters are more likely to be idiosyncratic in ways that cancel each other out, allowing the aggregate outcome to be biased in the direction of more-informed voters, and because voters are a bit less likely to be systematically wrong about their own lives (e.g., if the economy has improved under a candidate, they might falsely believe the economy is bad but they’re still more likely to be doing financially better). When voters are selfless and systematically biased about the collective interest, the different mistakes move in the same direction rather than diluting each other.
And I'd further argue that it still does apply to the readers (myself included), because even philosophically inclined folks or Rationalists or Effective Altruists are often unable to meaningfully escape their biases when voting.
I disagree. I’d much rather a rationalist/EA with tons of biases vote altruistically than selfishly, because as long as they’re just modestly better at reasoning about the issues they tend to prioritize (e.g., reducing AI risk or helping mitigate factory farming) than chance, the expected value is still much better than if they voted selfishly because the importance of issues follows a heavy-tailed distribution and hence is way more important than issues they would vote on if voting based on self interest.
Great points. I’m not sure about the local analysis here.
I live in San Francisco where elections have been determined by a few dozen votes. Local issues can be very impactful on people’s lives, like whether the highway next to someone’s house turns into a public park. (How much would you pay to keep a favorite park next to your home of 2 decades from closing down?)
There can also often be many issues on the ballot. Might’ve been 15+ candidates, ballot measures, etc. for my recent June election. So there’s multiple opportunities to shape outcomes for each voting effort.
Not saying this invalidates your overall thesis. I mainly think that local voting may be higher EV (personally and publicly) than implied here.
Yeah, I guess it will depend on the details.
"(This article was edited. I’m thinking of hiring the person who edited it as an editor long-term, so let me know if you found the writing noticeably cleaner than usual)."
This question is hard to answer because of confounds. I might find another post less clear because the concepts it discusses are more difficult or because I spent less time thinking about it. (You said "clean," not "clear," but I imagine that above all you want your articles to be understood.)
Why don't you post the original version of the article, so that your readers can compare the two and decide to what extent the edited version is an improvement?
I'm generally okay with hand-waved numbers in arguments like these, but the first round of number wrangling feels way too casual to me. Are there really no scenarios where a candidate would (or you've convinced yourself they would) benefit you a significant amount? First thing that comes to my mind is job creation. If it's reasonable to believe that you'd be likely to secure higher-paying employment (given your resume) then you could easily stand to directly benefit tens of thousands over multiple years. I don't disagree that *the vast majority* of voters would only benefit at the level you've described, but the edge cases are big enough that they need to be addressed.
Second, I'd argue there are *very* few cases where a voter really knows that one candidate will be severely detrimental to the country, so that's a largely useless form of hypothetical weight. Instead, people generally believe that, at most, there will be a very mild detrimental effect (and even that is usually seen merely as "possible", not necessarily "likely").
Also if I believe that everyone else is only going to get 10¢ if I choose $1000, but 0¢ if I choose $2000, the moral math is less striking, and I don't think it's anywhere near as persuasive as the example of depriving everyone else of a 1k payout. You can make arguments about how 10¢ for ten million people is ultimately going to benefit me more than an extra 1k, because it means one million dollars going into the economy around you - but that argument wasn't made, merely that it would be immoral to deprive so many people of a direct payout.
I showed it held even under outrageous assumptions.
It's implausible politicians will benefit you by 10,000x more than average.
The benefit is that the one that benefits me more also benefits the country more (parties that support free markets and deregulation)
But note that I didn't say you should never vote for the candidate that's better for you. I just said you shouldn't if it conflicts with the one who you expect to be best overall.
Sure, I'm just calling myself better informed than the typical person ;)