Interesting! But I think you’re using a standard term in a not quite standard way. Right now people use “Pareto” in two ways:
1) A deal can be a Pareto improvement compared to another deal, or to never making any deal.
2) A deal can be Pareto optimal (which means no other possible deal is a Pareto improvement compared to this one).
You seem to be proposing a third usage, where a deal must be a Pareto improvement compared to all other possible deals. But as you rightly say, such deals would be rare. I’d prefer to stay with standard usage, which acknowledges that there’s a whole Pareto frontier for people to bargain over.
I think philh is using it in the first way you described, just while honoring the fact that potential future deals factor into how desirable a deal is for each party. We do this implicitly all the time when money is involved: coming away from a deal with more money is desirable only because that money makes the expected outcomes of future deals more desirable. That’s intuitive because it’s baked into the concept of money, but the same consideration can apply in different ways.
Acknowledging this, we have to consider the strategic advantages that each party has as assets at play in the deal. These are usually left implicit and not obvious. So in the case of re-opening Platform 3, the party in favor of making the platform accessible has a strategic advantage if no deal is made, but loses that advantage if the proposed deal is made. The proposed deal, therefore, is not a Pareto improvement compared to not making a deal.
I don’t really think of it in terms of deals, but in terms of world states. A world state might or might not be a Pareto improvement over another.
But then I guess I’m equivocating between deals and outcomes of deals . So when I say it wouldn’t be a Pareto improvement for Adam to give Beth his card, I mean that the world state “Beth has the card” isn’t a Pareto improvement over the world state “Adam has the card”.
(I don’t mean that Adam wouldn’t agree to the deal—that’s just not the question I’m pointing at. Though in this toy world, where taking the deal makes him worse off, I’m not sure why he would.)
I think that most takes would say that that is a Pareto improvement, and I think that when I disagree, I’m not just using the term differently.
If you judge the Pareto goodness of the state “Adam has the card” by all its possible futures, then no future can be a strict Pareto improvement, right? That seems like a drawback of using the term in this way.
I’m not sure why you think I’m doing that. I’m judging based on the assets people have and the affordances they grant.
In another thought experiment, we would say that money has no intrinsic value beyond what it can buy people. But if Adam gives all his money to Beth, who uses it to buy something, we probably wouldn’t say this is a Pareto improvement on the basis that Adam’s only lost something he doesn’t intrinsically value.
Interesting! But I think you’re using a standard term in a not quite standard way. Right now people use “Pareto” in two ways:
1) A deal can be a Pareto improvement compared to another deal, or to never making any deal.
2) A deal can be Pareto optimal (which means no other possible deal is a Pareto improvement compared to this one).
You seem to be proposing a third usage, where a deal must be a Pareto improvement compared to all other possible deals. But as you rightly say, such deals would be rare. I’d prefer to stay with standard usage, which acknowledges that there’s a whole Pareto frontier for people to bargain over.
I think philh is using it in the first way you described, just while honoring the fact that potential future deals factor into how desirable a deal is for each party. We do this implicitly all the time when money is involved: coming away from a deal with more money is desirable only because that money makes the expected outcomes of future deals more desirable. That’s intuitive because it’s baked into the concept of money, but the same consideration can apply in different ways.
Acknowledging this, we have to consider the strategic advantages that each party has as assets at play in the deal. These are usually left implicit and not obvious. So in the case of re-opening Platform 3, the party in favor of making the platform accessible has a strategic advantage if no deal is made, but loses that advantage if the proposed deal is made. The proposed deal, therefore, is not a Pareto improvement compared to not making a deal.
I think I’m using it in the standard way.
I don’t really think of it in terms of deals, but in terms of world states. A world state might or might not be a Pareto improvement over another.
But then I guess I’m equivocating between deals and outcomes of deals . So when I say it wouldn’t be a Pareto improvement for Adam to give Beth his card, I mean that the world state “Beth has the card” isn’t a Pareto improvement over the world state “Adam has the card”.
(I don’t mean that Adam wouldn’t agree to the deal—that’s just not the question I’m pointing at. Though in this toy world, where taking the deal makes him worse off, I’m not sure why he would.)
I think that most takes would say that that is a Pareto improvement, and I think that when I disagree, I’m not just using the term differently.
If you judge the Pareto goodness of the state “Adam has the card” by all its possible futures, then no future can be a strict Pareto improvement, right? That seems like a drawback of using the term in this way.
I’m not sure why you think I’m doing that. I’m judging based on the assets people have and the affordances they grant.
In another thought experiment, we would say that money has no intrinsic value beyond what it can buy people. But if Adam gives all his money to Beth, who uses it to buy something, we probably wouldn’t say this is a Pareto improvement on the basis that Adam’s only lost something he doesn’t intrinsically value.