Hacker News new | ask | show | jobs
by desdiv 4204 days ago
Let's say I like party A 10 times more than I like party B. In a preferential voting system, there's no way to precisely express my preferences using the ticket. I'm forced to choose between "1. A" or "1. A, 2. B". Someone else who like part A 2 times more than party B will end up voting identically to me, despite the huge difference between our preferences. Any preferential voting system is necessarily flawed because how limited its input is.

That's basically what Arrow's impossibility theorem says. Due to its limited inputs, a preferential voting system will necessarily fail one of the three fairness criteria.

There are far better voting schemes out there, none of which are affected by Arrow's impossibility theorem.

1 comments

What does it mean to like party A 10x more than party B? What I'm questioning here is the existence of cardinal preferences, which are necessary for a "will of the people" to be defined. The only way I can make sense of cardinal preferences is to treat them as dollars spent on private goods [1], but I doubt that's what the OP meant.

As it applies to this situation, it's moot - Portland did not express any cardinal preferences.

[1] Non-private goods introduce other incentives that prevent spending from tracking desire.

Let's say I value thing A ten times more than I value things B through K, and I value them all equally whether I get them together or apart because they apply to different spheres of my life. Then I would be indifferent between getting A or getting all of B-K.
Why fo you think cardinal preferences are necessary!?
Arrow's Impossibility Theorem says that ranked preferences are insufficient. One stronger assumption you can make is cardinal preferences, which is what desdiv appealed to.
The point is that invoking an impossibility theorem oftentimes - and also in this case - demonstrates that the formalization one has chosen to work with is not a desirable one.

For example, if a group of people by some social process comes to a consensus then arguably this represents the "will of the people". Thus it makes sense to reason about this concept without requiring the existence of ranked preferences.

So for you, "will of the people" represents a consensus preference? And following this idea, if there is no consensus (i.e., at least one person in Portland wants to ride an Uber), there is no "will of the people"?

The whole point of Arrow is that you need some very strong assumptions (e.g., cardinal preferences) to define a "will of the people". The only real world expression of cardinal preferences is a set of supply&demand curves, however - based on this the "will of the people" says Uber should exist.

You misread, I made the converse claim.