In almost all voting systems, there’s some amount of incentive to vote strategically, by misrepresenting one’s true desires to obtain a more favourable result (which provides a worse result when everybody votes strategically; there’s a prisoner’s dilemma-type situation here). However, an important lens for analyzing systems, is whether a system rewards strategic votes, or punishes non-strategic votes.
In FPTP, the system widely used in the US, a person who chooses not to vote strategically will thereby greatly increase the probability of a candidate they strongly disagree with winning, which makes non-strategic voting become unpalatable.
However, in some other systems, such as approval or score voting, while a non-strategic vote may not result in a person’s favorite candidate being elected, it will still give an advantage to a candidate that they find genuinely likeable, rather than to someone they despise. This means that the incentive to vote strategically is much weaker, and people are more likely to honestly represent their desires, which leads to better outcomes when everybody does so.
There’s a critical modeling question to ask before this one—what is your ACTUAL preferece-aggregation function, in a mind-reading world where strategy doesn’t come into play? https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem is about aggregation of choice, not just voting errors.
In almost all voting systems, there’s some amount of incentive to vote strategically, by misrepresenting one’s true desires to obtain a more favourable result (which provides a worse result when everybody votes strategically; there’s a prisoner’s dilemma-type situation here). However, an important lens for analyzing systems, is whether a system rewards strategic votes, or punishes non-strategic votes.
In FPTP, the system widely used in the US, a person who chooses not to vote strategically will thereby greatly increase the probability of a candidate they strongly disagree with winning, which makes non-strategic voting become unpalatable.
However, in some other systems, such as approval or score voting, while a non-strategic vote may not result in a person’s favorite candidate being elected, it will still give an advantage to a candidate that they find genuinely likeable, rather than to someone they despise. This means that the incentive to vote strategically is much weaker, and people are more likely to honestly represent their desires, which leads to better outcomes when everybody does so.
There’s a critical modeling question to ask before this one—what is your ACTUAL preferece-aggregation function, in a mind-reading world where strategy doesn’t come into play? https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem is about aggregation of choice, not just voting errors.