The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
A Pareto gain: always the best choice?
itylus
Registered User regular
Registered User regular
This topic came up in another thread and I thought it might be interesting to talk about it a bit.
For those who haven't endured an undergrad education in economics, here's a quick definition of Pareto efficiency cribbed from wikipedia:
"Given a set of alternative allocations and a set of individuals, a movement from one allocation to another that can make at least one individual better off, without making any other individual worse off, is called a Pareto improvement or Pareto optimization. An allocation of resources is Pareto efficient or Pareto optimal when no further Pareto improvements can be made."
http://en.wikipedia.org/wiki/Pareto_efficiency
What happens, though, in a situation where you can make lots of people a little bit better off, by making a few people a lot worse off? (Or some similair example). The answer is that if the winners out of the change are able to compensate the losers, and still be better off than they were, you have achieved a Pareto gain. For example, someone wants to build a tower that will block the view from my house. He gains and I lose. But if he's willing to pay me $X, where at $X I'm willing to lose the view, and he still is better off than if he hadn't built the tower, then building the tower is a Pareto gain.
But it's a Pareto gain even if he doesn't actually end up paying me the compensatory sum.
Which leads to the question which I'm leaving you with for discussion: since it's always the case that the average level of wellbeing or happiness or utility or whatever is increased in situations where a Pareto gain is made, does it follow that it's *always* best to go for the Pareto gain? Or are there some circumstances where, for reasons of fairness or... some other reason, it's better to allow a Pareto inefficient situation remain inefficient?
For those who haven't endured an undergrad education in economics, here's a quick definition of Pareto efficiency cribbed from wikipedia:
"Given a set of alternative allocations and a set of individuals, a movement from one allocation to another that can make at least one individual better off, without making any other individual worse off, is called a Pareto improvement or Pareto optimization. An allocation of resources is Pareto efficient or Pareto optimal when no further Pareto improvements can be made."
http://en.wikipedia.org/wiki/Pareto_efficiency
What happens, though, in a situation where you can make lots of people a little bit better off, by making a few people a lot worse off? (Or some similair example). The answer is that if the winners out of the change are able to compensate the losers, and still be better off than they were, you have achieved a Pareto gain. For example, someone wants to build a tower that will block the view from my house. He gains and I lose. But if he's willing to pay me $X, where at $X I'm willing to lose the view, and he still is better off than if he hadn't built the tower, then building the tower is a Pareto gain.
But it's a Pareto gain even if he doesn't actually end up paying me the compensatory sum.
Which leads to the question which I'm leaving you with for discussion: since it's always the case that the average level of wellbeing or happiness or utility or whatever is increased in situations where a Pareto gain is made, does it follow that it's *always* best to go for the Pareto gain? Or are there some circumstances where, for reasons of fairness or... some other reason, it's better to allow a Pareto inefficient situation remain inefficient?
itylus on
0
Posts
That isn't to say that each case shouldn't be considered individually, or that there aren't possible cases were the negative is extreme and localized, while the gain is minuscule and wide spread where the Pareto gain should be disregarded. I'm thinking of cases like extreme versions of Nazi Germany where killing a relatively small number of people made everyone else a bit happier. Summing things up might still result in a net positive, but there are some negatives that aren't worth any amount of positive.
I know I've covered both sides of the issue, but since I was part of the thread that started this and am generally pretty pro-net gain, I figured a bit of ambiguity is good for a discussion thread.
The big problem with only looking at efficiency is that you completely ignore any concept of equity or fairness.
Also, werehippy, would you like to explain why you are opposed to any tax ever that is not correcting for market failures, almost all government, and any form of welfare or income redistribution? All of these things result in a net loss of efficiency.
I thought I was being clear enough, but I was referring both to the fact that certain attributes (like happiness or death) are nearly impossible to quantify in absolute terms (killing person X is worth $Y, being happy for five minutes is worth $Z, etc), and that Pareto gains tend to work well in moderation, but can break down when taken to extremes, as in the case where killing someone makes a lot of people happy for a little while.
Define equity or fairness. How exactly should those play into societies decisions, especially given there is no universal measure of either?
How do ANY of those fail the Pareto test? Besides the fact I'm not at all willing to grant that government, taxes, or social safety nets are automatically inefficient, I don't even need to touch on that to disagree that they don't each represent a net increase in benefit for society.
Taxes represent a little lose to everyone, but a larger gain in terms of public services, safety, and so on. Government represents a lost of some freedoms and autonomy in exchange for a guarantee of the greater number of freedoms and social stability. Welfare and income redistribution (the social safety net) is absolutely critical in terms of a stable society and general welfare of the population, the benefits of which far outweigh the costs in supplying them.
This criticism from the wiki you linked applies also:
So the correct thing to do is to ignore any concept of fairness? How about instead we do what we do with every other opinion-based policy decision: let everybody have their input and see what happens? Because, yes, the concept of economic fairness is opinion-based, but I'm pretty confident that we as a society can reach a better compromise than "oh well, I guess we just won't have fairness!"
Any tax that is not a head tax is inherently inefficient, have you ever taken economics?
Governments have to run off of taxes, therefore they are inefficient to begin with, and anybody with a degree in economics will tell you that, because they don't operate within the restraints of a market, the government uses their already inefficiently acquired funds in an inefficient way, failing Pareto efficiency even harder.
Income redistribution fails Pareto efficiency as long as there is any nonzero transaction cost in redistributing the income (which of course there is) because the very definition of Pareto efficiency implies that there can be no benefit from income redistribution.
A more useful perspective that doesn't abandon economic theory is Rawlsianism (if I'm not remembering incorrectly), which says that income redistribution can be efficient because the marginal benefit of a dollar to a rich person is less than the marginal benefit of a dollar to a poor person.
In pure economics terms, that's not true. For example, the free market could provide police services more efficiently, but we don't do that because it would be pretty fucked up if only people that could afford the police got their protection. Pareto efficiency completely ignores that. The government is only economically useful for providing public goods or regulating common resources. see above One could argue that's true, but the dollar value of stability in society is subjective and therefore beyond the scope of economics. Also see above.
That is it say, some things don't break down into raw numbers, and are impossible to really quantify. Even under Nazi Germany, they couldn't justify the holocaust without claiming that the Jews were less than human. When governments use the death penalty, they basically proclaim that the actions of the criminal have removed them from the human race.
Oh yeah, getting rid of the Jews probably isn't a very smart thing to do for your economy at all. Way to go Spain and Germany.
Obviously its always better to go for the efficient system.
Now, please tell me in all situations what is and what isnt efficient. That is the problem, not figuring out that moving towards efficiency is best.
Also Pareto defined efficiency in a slightly different way, which has applications in attempting to measure specific instances. However, all* other definitions of efficiency, if followed in the ideal produce the same result as the Pareto model.
And as such, we come to the conclusion that the question is not "Should we build if his view is obstructed and it hurts him more than the gain from building". The question is "how much is your view really worth?"
No, this is untrue. Any tax that does not resolve the quantity suplied and demanded to the equilibrium quantity when factoring in externalities into costs or benefits is inherently inefficient.[you might know taxes intended to do this as Pigovian Taxes, the opposite of which are subsidies]
This of course, ignores any benefits that the tax may bring in, which may move the efficiency to correct for any type of tax depending on the situation as subsidies are needed to correct for positive externalities and the amount of negative and positive externality in a system does not nessesarily balance out.
Its still wrong for income taxes you dolt.[in fact, probably more wrong due to decreasing value on monetary gains past a certian point. I.E. if you make 10 million dollars per year, 1 extra dollar is not worth as much as if you make 10,000 dollars per year, which makes head tax type income taxes more inefficient than progressive taxes due to the nessesity to maintain the same monetary flow.]
Do you want to calm down? We all know you know more about us than economics, relax.
Because any tax that does not correct externalities is inefficient. If you are going by Pareto then no income tax is inefficient due to government spending doing more benefit than the tax[and therefor net must be positive in all situations that a head tax must be positive], or similarly they are all inefficient due to government spending doing less than the net benefit of the tax.
If you arent going by praeto and instead going by a similar net efficiency based on net welfare then there is still a reduction in net welfare when a head tax is implemented as opposed to any other type of tax, as a head tax acts as a cost increase to all prodcuts. The fact that its not percentage of income based and that the cost increase is not similar for all individuals doesnt make it any less of an effect as a tax as supply and demand are aggregate values[as in aggregate data, not as in aggregate demand].
ed: and of course if you go by relative value head taxes are less efficient due to decreasing marginal value on wealth increases[though both are still inefficient by classical definitions]
Because taxes drive a wedge between the buyer and seller creating lower net welfare due to higher prices for consumers and lower returns for producers. This lowers the quantity of goods produced. Thus the tax+net producer surplus+net consumer surplus with a tax is less than the < no tax producer surplus+no tax consumer surplus.
I mean, how is the tax not going to produce inefficiency if its non-pigovian when only pigovian taxes dont create inefficiency.
That is the best way i can answer the question because seriously it doesnt make sense.
To give a simple numerical example: suppose I'm crafting gizmods by hand. The cost to me in materials and labour is $4 per gizmod. There are two potential customers, one of who is willing to pay $8 for a gizmod, and the other of whom is willing to pay $5 for a gizmod. Without a tax, I'm going to produce 2 gizmods, and there will be a total surplus of $5 (which will somehow be distributed between profit for me and consumer surplus for them, depending on various things that aren't important for this example.)
Now the gvt. imposes a 50% sales tax. There's no way I can sell the second gizmod at a profit if I make it (the most dude 2 will pay is $5, which after tax would leave me with $3.33, which is less than my production cost) so I won't make it, which reduces the total amount of surplus by $1. (And relocates some proportion of the remaining $4 of surplus into the hands of the gvt, but that's by the by.) The non-making of the second gizmod is what economists call a deadweight loss. Usually they show deadweight loss with weird looking graphs, but (unless I've messed up somewhere in there, entirely possible) this conveys the basics of it.
I thought this was pretty straightforward - the view is worth what I am willing to part with it for. Unless... you think that is a bad definition of what it's value is? Actually, I believe it is a bad definition - or at least an incomplete one - but it's fundamental to the idea of Pareto efficiency, isn't it?
Won't the $5 just get spent somewhere else, probably on something that is more cost-effective to produce?
No, for a variety of reasons. One is that Pareto maximums are just local maximums. Think of a situation - any situation. Striving for and attaining a particular nearby Pareto maximum for that situation could help establish a precedent for how to handle the situation in the future, and that can be damaging in the long run since it can remove focus from finding the optimal solution. The globally optimal solution is always a pareto maximum, but not every pareto maximum is globally optimal.
Yes, but then we have the issue that whether or not something is worth it hedges on the personal evaluation of any individual harmed in the situation.
Without a barganing process to determine what that view is worth to you when you have no means of contesting the building of the project there is no means to determine if the project is worth it or not.
No. the work doesnt get done. The example above is fairly bad though.
O.K. Imagine there are two people willing to buy gizmos one will do so at 6 dollars and one will so at 8 dollars.
Now I produce gizmos and can make the first for 2 dollars and the second for 4 dollars. I will sell these somewhere between 4 and 6 dollars, where it doesnt really matter, but lets say at 5 dollars to assume equal barganing conditions and skills.
The customers profit is 8-5+6-5 or 4. Because they would have payed 8 and 6, but only payed 5. And my profit is 4 as well because i would have sold for 2 or 4, but sold for 5. The total net benefit in the system is 8.
Now the government imposes a 100% tax. This burden shared between consumer and producer equally. This means that its not longer worth it for me to make the second gizmo. As it costs me 4 dollars to make that gizmo, and the customer only wants to buy it for a total of 6 dollars. If i sold for 6 dollars i would lose money, so i wont make the gizmo.
Now this means i sell 1 gizmo and i sell it somewhere between 2[my cost] and 4 dollars[4 and 8 after the tax]. Well lets say i sell it for 3 dollars. I get 3 dollars and the consumer spends 6 dollars.
The consumer has a benefit of 2[8-6], I have a benefit of 1[3-2] and the government has a benefit of 3[6-3]. The total net benefit is 6! This is less than the total net benefit in the previous situation by 2 dollars. The two dollars is called dead weight loss and is work that never gets done.
Now, the guy who didnt spend his money could spend it elsewhere, but if he does, he must nessesarily get less value out of that transaction[or he would have spend the money on my gizmo's] than he would have absent the tax.
Ummm, no it's not. You're worse off. A pareto improvement is where no one is worse off for the trade.
From the wiki:
"A (weakly) Pareto optimal (WPO) allocation is one where there is no feasible reallocation that would be strictly preferred by all agents."
If we assume that I have no legal ownership over the view, and the decision to allow the tower rests in the hands of the local council, whose only options are to allow the building of the tower or to disallow it, (ie, they can't order the tower builder to pay me anything), isn't allowing him to build the tower *weakly* Pareto optimal if the benefit to him exceeds the loss to me?
Goum I don't understand what was "fairly bad" about the other guy's example. The only difference I see between yours and his is that yours incorporates increasing marginal costs. Of course, neither of those examples were actually head taxes, but I can see why a head tax would have a similar effect on a market.
Huh? The $5 surplus is a result of the trade.
No. You're reading your quantifiers wrong. In fact, *both* situations could be pareto optimal. Say the only options were "view" and "no view," where the "preference point" allocations are as follows:
You: 1 for view, 0 for no view
Them: 0 for view, 5 for no view (no view = build).
It's impossible to move from one situation to the other without causing one party to be unhappy, and since they're the only two situations possible, they're both pareto optimal.
http://en.wikipedia.org/wiki/Kaldor-Hicks_efficiency
""Given a set of alternative allocations and a set of individuals, a movement from one allocation to another that can make at least one individual better off, without making any other individual worse off, is called a Pareto improvement or Pareto optimization. An allocation of resources is Pareto efficient or Pareto optimal when no further Pareto improvements can be made." "
Is this wrong?