Panglossian Economics

PanglossIn a radical departure from his previous expressions of dissillusionment, Paul Krugman has implicitly declared in his latest blog post that we are now living under the best of all policy regimes. I presume he will now be able to retire with satisfaction from his career as a gadfly.

The context is Eric Cantor’s demand that any federal disaster relief in the wake of Irene be offset by spending cuts elsewhere. Krugman thinks this is silly, and proves his point with an appeal to the standard Ricardian theory of public finance. According to that theory, which all economists understand and accept, if you’ve got to bear a cost, it’s best to spread that cost out over as many activities as possible. So ideally, you’d pay for disaster relief partly through spending cuts, partly through (current) tax increases, and partly through an increase in the deficit. Therefore says Krugman, “the bottom line is that basic, regular economics says that Cantor isn’t making sense.”

Since Krugman has carelessly neglected to spell out an important detail of his argument, let me fill in the gap for him: The Ricardian conclusion does not come from thin air; instead it follows logically from certain premises, key among which is that you’re starting from an ideal policy regime.

Rather than write down the relevant equations, let me give you an example: Suppose you’ve got a perfect kid, who divides his time optimally among schoolwork, sports, chores and socializing. Now he wants to go to the prom, which is going to cost him some time. Rather than take all that time away from any one activity, he’s best advised to cut back a little bit on schoolwork, a little bit on sports, a little bit on chores, and a little bit on non-prom socializing. (It takes some hidden assumptions to reach that conclusion, but they’re entirely reasonable.)

But suppose, on the other hand, that your kid has not done his chores in a month. Then even if you agree that he really ought to go to the prom, it makes perfect sense to say “You’re not going to the prom until you finish your chores.”

And if your government is as far behind on spending restraint as your kid is on his chores, then it can make perfect sense to say “You’re not providing any disaster relief until you catch up on your spending restraint”.

To get a little wonkier, there are actually two separate points here. First there’s the point that comes from public finance: Unless you believe that everything is perfect to begin with, the Ricardian argument fails, leaving you with no reason to believe that the cost of new spending should be spread widely. Instead, you should start by cutting back on your least wise activities. Cantor and Krugman probably have some legitimate disagreement about what those least wise activities are, but neither of them has any reason that I know of to believe that all activities are currently equally wise at the margin. I’m guessing that if Cantor had proposed paying for disaster relief entirely through an tax increase on the very rich, that Krugman would not have been so quick to dissent with an appeal to Ricardian public finance.

(Come to think of it, if Krugman took his own point seriously, he’d be forced to conclude that every tax increase should be spread across all income groups, and accompanied by cuts in all expenditure categories. This would mark yet another radical change in several of his previous positions.)

Then there’s the separate point that comes from public choice: Sometimes it’s a good idea to constrain people from doing things they want to do until they improve their behavior elsewhere. Even if you approve of the Congress’s intention to provide disaster relief, it can still make sense to hamstring that intention until they get some other stuff right — just like with the kid whose prom plans you approve, but who is still required to finish his chores first.

(Incidentally, all of this completely ignores the issue of whether disaster relief is a proper federal function in the first place — I am granting Krugman that for the sake of argument.)

Notice that Krugman goes beyond implicit insistence that everything today is perfectly hunky-dory; he says that it makes no sense to believe otherwise. In other words, he’s really really sure of this. I await his salute to the Tea Party for its role in bringing us to this state of Nirvana.

P.S. I commented on Krugman’s blog here.

Print Friendly, PDF & Email
Share

102 Responses to “Panglossian Economics”


  1. 1 1 Jonatan

    I think you are slightly misreading Krugman, again.

    He says: “What the government should do, in this case, is set all the marginals equal: (…)”

    He is not saying that the government is doing this, just that it should do that. Then he continues under the assumption that we have such a government, and from here it follows that Cantor doesn’t make sense. (This argument seems sound to me.) So what he is saying is that instead of following Cantor, it’s better to set all the marginals equal. And if you do that, it’s not good to follow Cantor.

    What he should have stated clearly, is that of course the marginals are not equal in his government, so therefore this argument against Cantor does not hold.

    But this sin of omission seems smaller than what you’re accusing him of. Also, the conclusion still seems relevant to me. If it is better to set the marginals equal than to follow the Cantor principle, shouldn’t we discard the Cantor principle and follow the marginal principle instead? It could be better than nothing to follow the Cantor principle, but that doesn’t mean it is the best way to go about it. At least this is up for discussion.

  2. 2 2 Stephan

    >I presume he will now be able to retire with satisfaction from his career as a gadfly.

    Aha. Where can I return your books and get my money back? I don’t like gadflies on my book shelf!

  3. 3 3 Jonathan Campbell

    A quibble with the analogy: assuming perfect policy, the kid should *cut back* on chores, just like all other of his activities if he wants to add a new activity (prom), thereby making time for the prom. In the real world, the kid should *increase* his chore activity.

    Surely, except for the effect described in the 2nd half of the post (“it can still make sense to hamstring that intention until they get some other stuff right”), the prom should make it marginally less urgent to increase the chore level — since marginal cost of prom goes up, marginal benefit is unchanged (unless there is some interaction between the prom and chores other than the simple fact that they both compete for the kid’s same limited time), as compared to the circumstance where there is no prom .

    This is in contrast with the case being analogized (hurricane), where the disaster makes it more urgent to cut back on wasteful government activities (as you suggest).

  4. 4 4 Jonathan Campbell

    correction: “marginal cost of prom goes up” should be “marginal cost of chores goes up”

  5. 5 5 Steve Landsburg

    Jonatan:

    shouldn’t we discard the Cantor principle and follow the marginal principle instead?

    To follow the marginal principle, you have to first determine what the marginal costs and benefits of various activities are. Krugman and Cantor assess those costs and benefits differently. So “follow the marginal principle” is not enough to determine what policy should be. Cantor is of course attempting to follow the marginal principle, applied in tandem with his assessments of costs and benefits. Krugman surely would do the same, in tandem with his assessments. The only way I can make any sense out of Krugman’s argument is that it denies those legitimate differences of opinion and starts from the presumption that we can all agree that the status quo is perfect.

  6. 6 6 Jonatan

    Krugman’s argument, as I understand it, is that assuming we are already in a world where all marginals are equal, then the Cantor principle is not good. (Cantor principle: That spending increase should be offset by cuts elsewhere.)

    This makes sense right?

    There is a discussion topic, which he conveniently ignores, about whether the Cantor principle is good in a world where the marginals are not equal. (Eg our world.)

  7. 7 7 Steve Landsburg

    Jonatan:

    There is a discussion topic, which he conveniently ignores, about whether the Cantor principle is good in a world where the marginals are not equal. (Eg our world.)

    One could, I suppose, discuss this, but it is self-evidently false. I think everyone realizes that if you’re doing one thing that’s smart and another thing that’s dumb, it’s better to stop doing the dumb thing than to stop doing the smart thing.

  8. 8 8 Ken B

    @Steve: Premises? Krugman reaches the same conclusion from all premises.

  9. 9 9 Steven L

    Krugman’s argument about equating marginal utilities is basic economics. But what happens when you start in a world where marginal utilities are not equated? The same argument implies that you should cut spending wherever the marginal value is low, but you shouldn’t have to wait for something like a disaster to happen to start cutting. Cantor’s position is that current government spending delivers lower marginal value than any of the other possible use of funds, so it should be cut, and if you have to shift resources to disaster relief, it should be cut even more. Why he thinks the marginal value of current government services is so low I don’t understand, but he’s been consistent on this point. And given this position, his remarks are completely in line with the economic theory presented by Krugman.

  10. 10 10 Steve Landsburg

    Steven L : Yes, exactly.

    (Note to readers: Steven L is not me!!)

  11. 11 11 ObliviousScout

    I don’t understand how a famous economics professor such as yourself can so badly mis-state the idea of marginal benefit and marginal cost. The relative marginal costs and benefits of every activity are different, but we should balance things where marginal cost equals marginal benefit for each activity.

    Krugman here is saying that the marginal benefit of disaster assistance is HUGE, and the relative marginal cost is small (because disaster assistance is temporary and can be debt financed). Because the relative net benefit is so large, we must re-balance other priorities. It doesn’t matter whether current priorities are a “perfect world” or not: the benefits are so much larger than any other priorities that disaster assistance would always take precedence.

    To take your contrived example with the not-so-perfect kid. Let’s say he’s not doing his chores and he falls out of a tree and breaks his arm. The marginal benefit of getting to the hospital right away is huge. Even though he did not have an optimal allocation of his time before the fall, the fall changes all the marginal benefits and costs. No mother would say, “You’re not going to the hospital until you finish your chores.”

    So has it really been that long since you taught an EC101 course? Or are you just deliberately trying to confuse the issue of how marginal costs and benefits work?

  12. 12 12 Steve Landsburg

    Oblivious Scout: Apparently you didn’t read Krugman’s post. He’s not arguing about the value of disaster relief; he’s arguing about how it should be financed.

  13. 13 13 ObliviousScout

    “Now suppose a disaster strikes. What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line…”

    Krugman then talks about the marginal costs of financing this in various methods, including tax increases and borrowing. His argument rests on balancing marginal costs against marginal benefits. You simply ignored his point about marginal benefits.

    The marginal cost of borrowing is at historic lows, so deficit spending should be a larger part of the equation.

    A perfect world is not a necessary condition for this argument.

  14. 14 14 Steve Landsburg

    Oblivious Scout: To balance MCs and MBs, you have to know which ones are out of whack in the first place. Theory tells you that if the world is perfect, then all the MCs and MBs are initially equal, from which Krugman’s conclusions follow.

    But if you don’t believe the world is perfect, then you are forced to assess the current values of the MCs and MBs. Krugman and Cantor assess them differently. Pure economic theory cannot tell you which of those assessments is correct.

  15. 15 15 Paul

    “Apparently you didn’t read Krugman’s post. He’s not arguing about the value of disaster relief; he’s arguing about how it should be financed.”

    PK wrote:
    “Now suppose a disaster strikes. What this does is raise the marginal benefit of spending on disaster relief.” Looks like he is in fact arguing about the value of disaster relief. Also you need to work on your reading comprehension if you think that PK actually believes in the Ricardian world. He is simply trying to argue that even in the intellectual world he doesn’t agree with that Cantor doesn’t make any sense. Therefore pointing out that the Ricardian world does not match up to his stated policy preferences is a pointless exercise. Finally I find it amusing that you actually believe Cantor is doing real cost benefit analysis to justify his position.

  16. 16 16 Steve Landsburg

    Paul:

    “Now suppose a disaster strikes. What this does is raise the marginal benefit of spending on disaster relief.” Looks like he is in fact arguing about the value of disaster relief.

    No, he’s taking this as a premise (and I am granting him this premise). His bone of contention is with Cantor’s position on how the relief is to be financed.

    He is simply trying to argue that even in the intellectual world he doesn’t agree with that Cantor doesn’t make any sense.

    That is indeed what he’s trying to argue, but of course it’s wrong. If you believe that programs A, B and C are vastly overfunded, then it does make sense to cut them.

  17. 17 17 Michael

    “What the government should do, in this case, is set all the marginals equal”

    As if doing so were as easy as flipping a switch. Businesses, the economy as a whole, even, are far more complex machines than many economists give them credit for.

  18. 18 18 Steve Landsburg

    Michael: Krugman is correct that ideally, you want to set all the marginals equal; I don’t think he’d disagree that this can be difficult in practice.

    The issue here is not about practicalities; it’s about theory. Here you’ve got a whole bunch of numbers, A, B, C, D, and E (the marginal costs and benefits of various govt activities). You, as a policymaker, can, at least in an ideal world, control these numbers subject to some constraints. Your goal is to equalize them.

    Krugman says: If A goes up (A being the marginal benefit of disaster relief), you should aim to raise B, C, D and E also. The mistake is failing to recognize that B, C, D and E might already be much higher than even the new value of A.

  19. 19 19 Paul

    “That is indeed what he’s trying to argue, but of course it’s wrong. If you believe that programs A, B and C are vastly overfunded, then it does make sense to cut them.”

    Except he’s not proposing to cut them. He’s only proposing to cut them if we increase spending on another program. There is no cost benefit analysis going on here. To claim otherwise is silly, but for the sake of argument let’s assume Cantor is doing some kind of crude CBA. The real rate of return on long term government debt is negative and tax revenues are at historic lows. If Cantor is going to argue that the government in the future will be unable to cut spending by a sufficient amount to pay for disaster relief now then he is assuming that we must have slightly higher taxes in the future if we don’t have offsetting cuts now. You can and you do argue that Cantor and Krugman might have different views on how to evaluate the costs of certain programs and policy choices, but in order to come to Cantor’s conclusion that we must cut spending now versus the tax alternative I just mentioned you would have to come to radical conclusions about the costs of a miniscule tax increase that are not supported by historical data. So yes Cantor might have a different opinion, but that doesn’t mean it makes sense.

  20. 20 20 Michael

    Professor,

    I won’t dispute theory–what you say jives with what I remember from my schooling. However, I didn’t read Krugman’s piece as a strictly theory based discussion. Perhaps I’ll read him again and see where my mistake lies, but he treats the conversation far too cavalierly for my taste.

  21. 21 21 Michael

    *lays

    Sheesh, I should proof read.

  22. 22 22 Henry

    Great post. I’d like to see more posts along these lines, because I think status quo bias is among the common fallacies in political discussion, particular as far as taxation is concerned. Consider the framing of the debate surrounding the renewal of the Bush tax cuts. Democrats were trying to frame it is as “letting taxe cuts for the rich expire” whereas Republicans were trying to frame it as a “tax hike”. Basically, Democrats were trying to make people think that the Bush tax cuts were an aberration from the longer-term status quo, whereas Republicans were trying to make people think the new tax brackets were now the status quo.

  23. 23 23 Ken B

    Many commenters are missing Steve’s point. Krugman cited a result. That result relies on premises. So Krugman IMPLICITLY cited those premises. One such premise is that the marginal utility of all the various bits of gov’t spending is equal. That must always be a debatable proposition and not a result of economic analysis (as it depends on actual utulity curves). It is not and never will be “senseless” to question that — unless as SL wickedly observes, we live in the best of all possible worlds.

    I think Steve the Pangloss reference is just lost on some.

  24. 24 24 nobody.really

    I regard most policy arguments as involving an “all else being equal” assumption. So, whether or not you think that spending is happening in an optimal fashion, disaster relief did not cause the sub-optimality, and there’s no special reason to treat disaster relief as the occasion for resolving it.

    I see three scenarios:

    1. Both parties might well prefer to make incremental changes to the status quo in order to fund disaster relief, and could resume their fight over the budget thereafter.

    2. One or more party may sincerely conclude that disaster relief is not worth the cost, and block the expenditure (unless his consent can be bought via changing something he finds even more objectionable).

    3. Both parties might agree that they should fund disaster relief, but one or more might nevertheless engage in holdout behavior – that is, engage in conduct that he dislikes, simply to extort concessions from others who also dislike the conduct. This arises in eminent domain contexts, when a property owner that would otherwise be willing to sell at market prices opts to demand a premium price because he knows his property is necessary for the larger project. The owner doesn’t really want to kill the project, but he has to act as if he’s willing to do so. Even Posner acknowledges that eminent domain laws are appropriate to deal with this scenario.

    So, imagine that Cantor doesn’t like the way that his household is spending money, and his wife doesn’t like the way their household is spending money, but they haven’t come to any resolution about what changes to make. And then their daughter is hit by a truck. In Scenario 1, they might both agree that they need to make an off-budget expenditure for medical services to save their daughter’s life, even though they each preserve their previously-stated concerns about their budget. In Scenario 2, Cantor might conclude that the value of saving his daughter’s life is not worth the cost — although he would not be averse to considering eliminating expenditures that he thinks are even more wasteful than the cost of saving his daughter’s life. In Scenario 3, Cantor might agree that the cost of saving his daughter’s life is worth it, but conclude that this emergency provides a great opportunity to induce unrelated concessions from his wife.

    I sense that Krugman is assuming Scenario 1: Both parties really believe in disaster relief, and could simply choose to make various adjustments to the status quo in order to fund that relief – all the while preserving their budgetary arguments that existed prior to the disaster.

    Landsburg correctly observes that Cantor may be in Scenario 2: In good faith, he really doesn’t think that we should fund disaster relief. He’s not adverse to being bought off, but he’s also ok if we simply maintain the status quo and withhold the relief.

    I suspect Scenario 3. But hey, that’s just me.

  25. 25 25 Paul

    @KenB

    By arguing that we need offsetting spending cuts immediately Cantor is implicitly arguing that he doesn’t believe the government is capable of spreading the cuts out over time because of commitment issues. That means that if offsetting cuts aren’t made now he is assuming we will have to add to our debt and will have pay higher taxes later in order to pay for this spending. The real return on long term U.S. debt is negative and taxes are at historic lows. Furthermore the amount of additional taxation that would be required to pay for disaster relief spending is trivial. The only way that a cost benefit analysis could yield Cantor’s proposal is that the cost of this trivial amount of increased taxation (I mean how much additional funding are we talking for FEMA like 10 billion?) is astronomical and/or the benefits of current government spending is almost nil. The former is not supported by any historical data or analysis with regards to taxation and the latter is contradicted by Cantor’s position that spending need only to be cut should we choose to offset allocate additional funds for disaster relief to say nothing of the drag that austerity is already placing on the fragile economy. People can have different opinions about costs and benefits, but I think it can be argued that the conclusions one would have to draw to reach Cantor’s position are bordering on the absurd.

  26. 26 26 Sbastonvilla

    I agree with your fundamental point, but Cantor shouldn’t be given a pass either. If Cantor is confident that he can find cost savings in current spending that both cover the cost of disaster relief and better equalize marginal costs to marginal benefits, he ought to say where these savings are specifically.

    While governments spend inefficiently, they often cut spending inefficiently as well

  27. 27 27 Steve Landsburg

    Paul:

    in order to come to Cantor’s conclusion that we must cut spending now versus the tax alternative I just mentioned you would have to come to radical conclusions about the costs of a miniscule tax increase that are not supported by historical data. So yes Cantor might have a different opinion, but that doesn’t mean it makes sense.

    It seems crystal clear to me that Krugman is making an argument not about historical data but about economic theory. He explicitly invokes Ricardo, which makes sense only if you’re starting from an ideal position.

    To put this another way, Krugman is saying that Ricardo’s analysis alone tells you that Cantor is wrong. But Ricardo, of course, had no access to the specifics of the 2011 US Federal budget, so any argument that refers to those specifics is not responsive to Krugman’s point, or to mine.

  28. 28 28 Steve Landsburg

    Ken B:

    Krugman cited a result. That result relies on premises. So Krugman IMPLICITLY cited those premises. One such premise is that the marginal utility of all the various bits of gov’t spending is equal. That must always be a debatable proposition and not a result of economic analysis (as it depends on actual utulity curves). It is not and never will be “senseless” to question that — unless as SL wickedly observes, we live in the best of all possible worlds.

    Yes, that’s the point exactly.

  29. 29 29 Hasdrubal

    Isn’t there another way to remain at the same, or even increase total marginal benefit? Instead of reducing all activities so their marginal costs and marginal benefits remain in balance, can’t you substitute directly? Wouldn’t substituting a new thing for an old thing of equal (marginal?) cost and equal or lower marginal benefit also be utility maximizing?

    To borrow your example, wouldn’t it be equally utility maximizing for your prom-going teenager to sell his Ozzfest ticket and go to the prom instead of cutting back a little bit on everything?

  30. 30 30 Addison

    Krugman: “What the government should do, in this case, is set all the marginals equal”

    keywords: SHOULD, SET

    He implies they are NOT currently equal. They need to be made (SET) equal for this basic economic theory to hold. He does not imply all government spending is CURRENTLY perfectly equated with marginal costs and benefits. He explains the basic theory, assuming his readers can figure this out.

    “Now suppose a disaster strikes”.

    Thats right. He very explicitly stated that the government should achieve optimal allocation of funds, FIRST.

    You need to read a bit more carefully.

  31. 31 31 Steve Landsburg

    Addison:

    “What the government should do, in this case, is set all the marginals equal”

    Right. The question then becomes: In order to set them equal, which ones need to be raised and which to be lowered? Theory can answer this only if you assume something about their initial values.

    Or to put this another way: Krugman says, and we all agree, that (A) all the marginals should be equalized. Krugman concludes that (B) when one of the marginals increases, all of the marginals should be raised. If you try writing out a careful argument that will get you from (A) to (B), you’re going to discover that you need another assumption.

  32. 32 32 Addison

    He does not go into detail how funds are currently inappropriately allocated. I was just pointing out your error:

    “Paul Krugman has implicitly declared in his latest blog post that we are now living under the best of all policy regimes.”

  33. 33 33 Ken B

    As an aside, since we discussed books in another thread, Candide is one funny book (especially en francais).

  34. 34 34 Steve Landsburg

    Addison:

    Here we have statement (A): All marginals should be equalized.

    Here we have statement (B): When one marginal is increased, all others should be raised.

    Krugman asserts that (A) implies (B). To get from (A) to (B), you’ve got to assume (C): All marginals are currently equal. (If you think otherwise, I’ll be glad to see your alternative argument for getting from (B) to (A).

    Therefore saying “(A) implies (B)” is as good as saying (C), which is of course equivalent to saying that we currently live under the best of all policy regimes.

    If that’s still not making sense, please point to the specific line above you’re not following, and I’ll try to make it clearer.

  35. 35 35 Ken B

    @Addison: It was for those like you that I emphasized IMPLICITLY in my remarks above. Steve is teasing out — and mocking — an assertion that Krugman’s post LOGICALLY RELIES UPON. If you ignore what Krugman MUST assume you will continue to miss Steve’s jibe.

  36. 36 36 Addison

    Yes, Steve, he says this is what should be done AFTER the disaster. FIRST saying that they are NOT CURRENTLY EQUAL, AND NEED TO BE SET EQUAL. He does not imply “we are CURRENTLY living under the best of all policy regimes.”

    I’m sorry I cannot make it much clearer than that. It seems you are seeing his explanation of the theory, which obviously assumes optimal conditions, but completely forgetting what he said immediately beforehand, that they are not. I would not criticize him too much for not elaborating in this article where he thinks spending should be changed for it to be optimal, that could be a very long article, and is something I assume he has talked about before.

  37. 37 37 Richard

    There is nothing wrong with what Krugman posted. At all. At $#@*&@! all. Why is it difficult for you to understand?

    I’ve believe other commentators here have already beaten you to death with the whole MC vs MB aspect of the post.

    Steve: Krugman says: If A goes up (A being the marginal benefit of disaster relief), you should aim to raise B, C, D and E also. The mistake is failing to recognize that B, C, D and E might already be much higher than even the new value of A.

    Umm…no he doesn’t. He wrote the following…

    Now suppose a disaster strikes. What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes. So even there it shouldn’t be all offsetting spending cuts.

    But wait: even more important, the government can borrow (or, in principle, lend, if it pays off all its debt). So it should balance its budget in present discounted value terms, not year by year. This means that the tradeoffs should include future spending and taxes as well as this year’s spending and taxes. And a natural disaster, like a war, is a temporary event; it should be met largely through higher taxes and lower spending in the future rather than right away, which is another way of saying that it should be paid for in large part by a temporary increase in the deficit.

    There is nothing in his post that dictates you have to raise spending elsewhere, or even that the marginal benefits and or costs of everything else increases. Nothing goddammit!

    At worst, you could stretch the truth by claiming he ‘sort-of-said’ that due to raising taxes, you don’t necessarily have to offset spending increases with purely spending cuts. All he’s doing is reiterating the basic Keynesian principle that in a time of war, recession, or in this case a natural disaster, deficits don’t matter in the short run.

    In the words of Krugman (when he referred to you): “Discussions like this really disturb me; they indicate that there are a lot of people with Ph.D.s in economics who can throw around a lot of jargon, but when push comes to shove, have no coherent picture whatsoever of how the pieces fit together.”

  38. 38 38 Mike

    As Steve Landsburg states, this has absolutely nothing to do with weather or not the MB of disaster relief is higher. He grants, for the purpose of this argument, that it is.

    The issue at hand is weather or not the MC of all government activities is equal. So, the issue is definitely one of funding.

    The only reasonable way to determine that Cantor’s request that the new spending be offset by cuts is absolutely wrong, is to believe that all current spending is of equal value.

    But, as we all know, Cantor sees some (most?) of government spending as the metaphorical stacking of money in a furnace and burning it. You may disagree. Krugman certainly disagrees. But if he actually does view some spending as having little to no value in the first place, he is certainly correct in suggesting it be the first thing to be cut.

    When the parents of the little boy are presented with the hospital bill, should they finance the required spending by increasing the long term balance on a credit card (at a low rate), or by getting rid of some wasteful spending behavior?

  39. 39 39 Paul

    Steve. Cantor is SAYING that he thinks we should increase disaster relief spending and then offset it with cuts. He is already assuming that the marginal benefit on disaster relief spending is higher (whether you think or believe otherwise is irrelevant since it is Cantor’s position that is being debated). Once you accept this point I don’t see how you could reach Cantor’s position without the assumptions I stated previously which I don’t think make sense.

  40. 40 40 Thomas Bayes

    This is the second time in the last few weeks that people have defended Krugman in this way.

    In the current example, Krugman says “the bottom line is that basic, regular economics says that Cantor isn’t making sense.” Professor Landsburg points out that a person could allocate marginal costs and benefits in a way that basic, regular economics would say that Cantor is making sense. Krugman clearly disagrees with the way Cantor allocates marginal costs and benefits, but he is writing in a way that persuades his readers to believe that Cantor is wrong about basic economics. I believe he is deliberately ambiguous for this reason.

    Here is another example:
    http://econlog.econlib.org/archives/2011/08/paul_krugmans_s.html

    In this case, Krugman said in a 2002 column:

    “To fight this recession the Fed needs more than a snapback; it needs soaring household spending to offset moribund business investment. And to do that, as Paul McCulley of Pimco put it, Alan Greenspan needs to create a housing bubble to replace the Nasdaq bubble.”

    Krugman and his defenders have subsequently claimed that Krugman didn’t believe a housing bubble was a good thing, and that if you interpret his words in a different way you should see that.

    Now I’m starting to wonder . . . if you look over much of Krugman’s “public intellectual” writings, is it common for him to say things in a way that causes his readers to take an extreme position, but that can later be analyzed in a way that gets him off the hook from having actually said that was the correct position?

  41. 41 41 PrometheeFeu

    I’m re-posting here because the NYT comment section is more of a soap box than an opportunity for dicussion:

    Even the MC/MB framework presented by Paul Krugman above is useless. Let’s see how we do this. Well, we draw a graph with spending on one axis and value on the other axis. Then you plot your Marginal Benefit curve and your Marginal Cost curves. You see where they intersect and that is your spending level. Sounds easy right? It’s not.

    You need to draw meaningful curves to start with. You can’t. In order to draw meaningful MB and MC curves for the government you need to aggregate in some way the individual MB and MC curves for everyone in the country. That means drawing an aggregation method.

    When you aggregate the utility functions of everyone, do you weight them at all? Should the utility of a citizen count for the same as that of a resident? What about a tourist? What about someone living abroad? How do you pick the weights? Do you consider diminishing social returns on utility? If a person has a lot of utility, can you take away half their utility to give it to someone with very little utility? What about equality issues? How do you aggregate the utility of people who wish each other harm? Or who wish each other good? There are no clear answers to any of those questions.

    Even if there were. In order to compute such aggregate MB and MC curves, you will need in any but the most trivial of cases to assume cardinal utility. In other words, there is a happiness number in your head. How likely is that to be true? Even if you assume cardinal utility, you need to assume meaningful inter-personal utility comparisons which means you and I can always compare our happiness to each other. Think about that outside of the very easy: he’s crying, I’m not situation.

    In other words you can’t build the MC/MB framework Paul Krugman is selling without assuming away questions which are unresolved after centuries of debate. What are the chances that all these questions have been solved by elected people who can’t even pass a budget without trying to kill each other? No MC/MB model.

  42. 42 42 Richard

    ^Said the poster who linked econlog. *snicker*

  43. 43 43 PrometheeFeu

    @Richard:

    Krugman tells us a story. He starts out his story with:

    “What the government should do, in this case, is set all the marginals equal: the marginal benefit of an additional dollar spent on bombs, dental work, national parks, soup kitchens, etc, should all be equal, and this common marginal benefit should equal the marginal cost of raising an additional dollar of revenue.”

    So at the beginning of the story the government sets perfect policy.

    “Now suppose a disaster strikes.”

    Then, he shows how in his story, the proper response is for the government to reduce spending, raise taxes and borrow some money. OK fine. That’s quite a sensible story.

    However, Krugman then jumps out of the pretend world and says: “So the bottom line is that basic, regular economics says that Cantor isn’t making sense.” Of course, this is only true in Krugman’s story world where the government started before the disaster with MC=MB everywhere. Cantor made a statement about what should happen in the real world. And in the real world, MC was not set to equal MB before the disaster struck. If MC was not equal to MB, Krugman’s story is wrong and so is his conclusion about Cantor being nonsensical.

  44. 44 44 Steve Landsburg

    Richard:

    You quote Krugman to this effect: “The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes

    Do you realize that this is equivalent to “raise B,C,D and E”?

  45. 45 45 Dave

    Richard,

    It looks like you are the one who lacks understanding here.

    You write:
    “All he’s doing is reiterating the basic Keynesian principle that in a time of war, recession, or in this case a natural disaster, deficits don’t matter in the short run.”

    Yet Krugman writes:
    “Still, it may be worth talking about just how bad an idea this is in terms of basic economics — and in this case, regular economics, not fancy-schmancy macro.”

    Keynesian economics is macro, so you’ve demonstrated that you do not understand Krugman’s post. But you’ve also managed to demonstrate that you don’t understand Landsburg’s post. And it’s actually very simple. I’ll see if I can breakdown in a clearer manner.

    Krugman criticizes Cantor for wanting to *completely* offset disaster spending with cuts in other areas:

    “In a way, I may be wasting my time doing any kind of rational analysis of Eric Cantor’s demand that any disaster aid in the wake of Irene be offset by spending cuts elsewhere”

    He then states that ideally the government should set the marginal benefits of each of its activities equally, and do so to the point at which the marginal benefit equal the marginal cost of raising additional revenue:

    “What the government should do, in this case, is set all the marginals equal: the marginal benefit of an additional dollar spent on bombs, dental work, national parks, soup kitchens, etc, should all be equal, and this common marginal benefit should equal the marginal cost of raising an additional dollar of revenue.”

    He then introduces a natural disaster, which raises the marginal benefit of disaster relief spending, and states that the appropriate response is to raise spending on disaster relief while offsetting some, but *not all*, with cuts on all other spending, while also raising tax revenue. But this is dependent on the premise he had set in the prior paragraph about the optimally run government which sets its marginal benefits and marginal costs equally. This is the paragraph that prompts Landsburg’s critcism. See for yourself:

    “Now suppose a disaster strikes. What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes. So even there it shouldn’t be all offsetting spending cuts.”

    Note, this is where he differs from Cantor. Cantor wants to *fully offset* the spending. Krugman makes the above argument to argue that even in a world without borrowing, this isn’t the case, provided that the government’s taxing and spending was optimal to begin with. But this reasoning only holds under that premise. Yet he’s using it as his argument against Cantor’s suggestion for what to do in reality, therefore he implies that he believes the “ideal government” premise *is* reality.

    Next, he makes a case that intertemporal trade-offs can be made as well, so that optimal offests can be made in the future, given that the government can borrow. This continues to rely on the premise that the government was optimally run in the same way that the previous paragraph did. I’ll won’t bother quoting it.

    Finally, toward the end, Krugman states:

    “This isn’t some novel idea, by the way — it’s the standard theory of public finance during war, going all the way back to Ricardo. And the logic of wartime finance applies equally to natural disasters.”

    Ricardo is not Keynesian, so again Krugman is not making a Keynesian point as you’ve suggested.

    Is that clear enough?

  46. 46 46 Richard

    @Prometheefeu

    Of course, this is only true in Krugman’s story world where the government started before the disaster with MC=MB everywhere.

    Umm…*cough*cough

    PK:But wait: even more important, the government can borrow (or, in principle, lend, if it pays off all its debt). So it should balance its budget in present discounted value terms, not year by year. This means that the tradeoffs should include future spending and taxes as well as this year’s spending and taxes. And a natural disaster, like a war, is a temporary event; it should be met largely through higher taxes and lower spending in the future rather than right away, which is another way of saying that it should be paid for in large part by a temporary increase in the deficit.

    Cantor was explicitly talking about cutting elsewhere to finance the disaster relief. Taxes and borrowing had no relevance! If anything, PK’s setup has more merit even if he hadn’t included the borrowing component. That’s why Krugman included the whole PV vs FV of cutting spending and taxes, emphasizing why it should be done later rather than now or on a year by year basis.

    @Steve:Do you realize that this is equivalent to “raise B,C,D and E”?

    No, actually, in Steve-o-land I’m not quite sure how your logic works. That in itself is a “Big Question.”

    You posted “If A goes up (A being the marginal benefit of disaster relief), you should aim to raise B, C, D and E also. The mistake is failing to recognize that B, C, D and E might already be much higher than even the new value of A.”

    From this it seems that B, C, D and E are MB’s relating to something else the government spends money on. Am I to assume now you’re going to fall back and inform me that B, C, D and E are tax rates? That seems the only way it could work. But then, again, why can’t B, C, D and E consist of merely borrowing instead of tax increases?

  47. 47 47 Richard

    Furthermore

    Prom: However, Krugman then jumps out of the pretend world and says: “So the bottom line is that basic, regular economics says that Cantor isn’t making sense.” Of course, this is only true in Krugman’s story world where the government started before the disaster with MC=MB everywhere. Cantor made a statement about what should happen in the real world. And in the real world, MC was not set to equal MB before the disaster struck. If MC was not equal to MB, Krugman’s story is wrong and so is his conclusion about Cantor being nonsensical.

    ^Do you realize that this ‘story’ that PK is using (if it were fully the case) is exactly identical to how Landsburg attacks/critiques other people’s positions? Anyone remember “The Man Who Cannot Be Taxed” blog entry? We assumed a rich guy did nothing but push around cars all day so taxing him wouldn’t affect his consumption…lol!

  48. 48 48 steve

    How do you get from

    ““What the government should do, in this case, is set all the marginals equal”

    to

    “this state of Nirvana.”

    You are using a present tense for Nirvana. Should implies a future, not yet achieved event.

    Steve

  49. 49 49 Ricardo Cruz

    Richard: B, C, D and E refer to courses of action being faced by the government agent. If you want to put it in terms of spending, you can think of them as a decision on how resources get spend in the open economy.

  50. 50 50 Richard

    LOL @ Dave

    Honestly, you’re quite sad. Did you read the link that PK posted when he wrote “Still, it may be worth talking about just how bad an idea this is in terms of basic economics — and in this case, regular economics, not fancy-schmancy macro.”

    Seriously, did you even click on the link?! PK was being ironic. The fancy-schmancy macro was refering only to Keynesian economic models not being like ‘regular macroeconomics.’ That whole post that was linked was in reference to Robert Barro’s Op-Ed in the WSJ. He was critiquing Keynesian macroeconomics in relation to other macroeconomic models (I’m guessing RBC.)

    Pot calling the kettle black. Seriously, the sheer stupidity of some on this blog is outstanding. I could critique the rest of your post, but why bother after the first couple sentences. You, like everyone else, missed to ‘borrowing’ paragraph in PK’s post as well.

  51. 51 51 Ricardo Cruz

    Richard: Dave was replying to this point of yours:
    All he’s doing is reiterating the basic Keynesian principle that in a time of war, recession, or in this case a natural disaster, deficits don’t matter in the short run.

    Krugman’s post has nothing to do with fiscal stimuluses. It is not macro-economics.

  52. 52 52 PrometheeFeu

    @Richard:

    You may have noticed my username is PrometheeFeu. Not Landsburg. Why not answer the argument I made instead of saying some other guy unrelated to me (even though I do like his blog) happens to have done something similar to PK?

    As for your response to the issue of for PK to be right, MC=MB has to be true, FV has nothing to do with it. Yes you can include FV in your MC and MB calculations. All that does is move the numbers around. Are you saying PK thinks MC=MB if we include FV calculations? Obviously he doesn’t and neither does Cantor. It doesn’t matter how you decide to calculate MB and MC if they are not equal prior to the disaster, PK’s argument does not hold period. Borrowing makes sense if it is cheaper than other alternatives. If as Cantor you believe that some government programs are a negative value, cutting those programs to pay for a positive value program such as disaster relief is cheaper than debt.

  53. 53 53 Steve Landsburg

    Richard: You’re clearly not following the A’s, B’s, C’s, D’s and E’s; I’ll make one more attempt to clarify.

    A is the marginal benefit of disaster relief, B and C are the marginal costs of various forms of taxation, D and E are the marginal benefits of various government spending programs. (There are also of course F, G, H, etc., all marginal benefits and/or costs of one thing or another that the govt does.

    1) Krugman says, and no economist will find this controversial, that in an ideal world, A=B=C=D=E(=F=G=….).

    2) Krugman now posits that A increases.

    3) Krugman concludes that appropriate policy is to raise B and C (by raising various forms of taxes) and to raise D and E (by cutting various forms of expenditure). Specifically, he says What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes.

    4) If in fact A=B=C=D=E to begin with, then this is right. But if they are not equal to begin with, it is wrong. For example, suppose that initially, A=2, B=1, and C=4. Now A goes up to 3. Then to optimize, you want to raise B, but you want to lower C.

    5) Therefore Krugman has implicitly assumed that A=B=C=D=E to begin with, i.e. he has assumed that we are starting (pre-disaster) from an optimal set of policies.

    This is pretty simple logic. If you can point to the step that you’re having trouble following, or the step that you think misrepresents Krugman, I’ll try to explain it a little further.

  54. 54 54 Keshav Srinivasan

    I have a question about Step 3. You say “raise D and E (by cutting various forms of expenditures)”. Maybe it’s just me, but I find it a bit strange that raising the marginal benefit of a government program is equivalent to cutting that government program. That just doesn’t make sense to me.

    Are you saying that if you cut a lot of education funding, then people will really need education, so they each dollar you spend in education will be much more valuable to them compared to if they already had a lot of education funding? In other words, are you making an argument about diminishing marginal utility?

  55. 55 55 Andres

    Good arguments, Steve, and the writing is even better.

  56. 56 56 Steve Landsburg

    Keshav Srinivasan: The standard assumption is that there are diminishing marginal benefits, so that the first dollar spent on education is the most valuable, the second is the second most valuable, etc. Thus the less you spend, the higher is the marginal benefit. (Of course, the less you spend, the lower is the total benefit, but that’s not what’s at issue here.)

    This, as Krugman correctly says, is very basic economics, and neither I nor any economist I can imagine would object to it.

  57. 57 57 Charlie

    So if your roads are blocked and your power lines are down and you haven’t had working power or water for days, you make a convenient hostage to force congress to get its fiscal house in order?

  58. 58 58 colin

    If we were spending $0 on defense, obviously we wouldn’t cut defense in order to spend on disaster relief (arguably, this is an actual possibility….we could ask al-qaeda if they are willing to pay us money for some help on their most important projects, thereby spending negative money on defense). But the realization that we were only spending $0 on defense might prompt some people to decide that we should increase defense spend before we get around to disaster relief.

  59. 59 59 Charlie

    I can make my point via analogy. You equate disaster relief with “going to the prom” as if it’s just a crazy lark that brings joy to congress. Let’s change it, instead to, “taking your pet to the vet.”

    Yeah, the kid should have done all his chores and you as a parent are also at fault for letting it get this far, but when the time comes is it really right to sentence Fido to death, while your kid tries to do years of backlogged chores.

  60. 60 60 Steve Landsburg

    Charlie: If your roads are blocked and your power lines are down, the case for fixing them rests on the importance of your roads and your power lines, not on economic theory. Krugman’s claim is that *pure theory* tells you what to do in this case. Your example — where you are forced to refer to the importance of power and water — shows exactly what is wrong with that claim.

    PS — your followup about taking Fido to the vet is in exactly the same vein. The whole point here is that to judge what should be cut and what should be expanded, you need to know some specifics. Krugman claims otherwise. You and I are on the same side.

    (Note also that regarding disaster relief, I explicitly granted the premise that it’s worthwhile.)

  61. 61 61 Steve Landsburg

    Steve:
    How do you get from
    ““What the government should do, in this case, is set all the marginals equal”
    to
    “this state of Nirvana.”

    I get there via “And when one marginal goes up then setting them all equal requires raising all the others”, which implies “All the marginals are equal to begin with”.

  62. 62 62 Scott H.

    Nice job Steve! I am good old comment #35 on Krugman’s list for that post.

    Sadly, my comment does not have the recommendations that yours does — even though my analogy was more apropos!

  63. 63 63 Thomas

    I am bothered by Krugman’s suggestion that government should “set all the marginals equal: the marginal benefit of an additional dollar spent on bombs, dental work, national parks, soup kitchens, etc, should all be equal, and this common marginal benefit should equal the marginal cost of raising an additional dollar of revenue.”

    The cost of raising an additional dollar of income is likely to fall upon a different set of persons than the benefits that flow from the government programs financed by that dollar. If, for example, you take a dollar from me to make X happier, you have made me less happy, and X’s greater happiness doesn’t compensate for my greater unhappiness.

    I don’t even have to be a selfish curmudgeon to object to the transfer of my dollar to X. It could be that I wanted to give the dollar to one of my grandchildren, which would have made both me and my grandchild happy.

    As for X, I couldn’t care less. And it is arrogant of Krugman (or anyone else) to suggest (even by implication) that it is okay to take a dollar from me just to make X (or a government bureaucrat) happier.

  64. 64 64 David Welker

    I definitely enjoyed this post. However, I think it is flawed in certain respects.

    Clearly, Paul Krugman is not so silly as to think that we live in an ideal world where all funds are allocated in an optimal manner.

    However, for his argument to work, we really do not have to live in that world. We merely have to live in a world where an extreme solution is not optimal. Extreme solutions include funding disaster relief fully through current spending cuts or current tax increases. The point of bringing up Ricardian finance theory is not to prove, in some absolute sense, the the extreme solution of financing disaster relief funding solely through spending cuts is optimal. Only that it using standard economic theory you can see that it is almost certainly so. All Landsburg has done is shown that Krugman has not provided an absolute proof; but in the real world, 99.9% is more than good enough.

    Another point. If your child was not allocating enough energy to chores, it follows that he should do so, WHETHER OR NOT there is a prom. The LINKAGE between the prom and increased effort on chores is artificial, as the child should be made to increase efforts on chores whether or not there is a prom.

    While as a parent, it might make sense for a parent to hold a prom “hostage” to force the child to increase effort on chores (assuming your poor parenting skills makes such hostage-taking necessary) it is another thing entirely to say that one should hold disaster relief hostage. After all, denying disaster relief is a whole different level of seriousness than denying a child access to the prom.

    In my opinion, if the Republican Party is seriously threatening to hold disaster relief hostage, they should be renamed the Terrorist Party. And I believe their bluff should be called.

  65. 65 65 Steve Landsburg

    David Welker:

    The point of bringing up Ricardian finance theory is not to prove, in some absolute sense, the the extreme solution of financing disaster relief funding solely through spending cuts is optimal. Only that it using standard economic theory you can see that it is almost certainly so.

    First, I think you meant to say “suboptimal”, not “optimal”. Be that as it may, it is not true that “using standard economic theory you can see that it is almost certainly so”. If there is such an argument from standard economic theory, I would like to see it.

    Standard economic theory tells me nothing at all about the probability that there exists some program sufficiently wasteful, or some set of programs that are collectively sufficiently wasteful, that it would be optimal to fund disaster relief by cutting them. If you think I’ve overlooked something, please by all means provide the argument.

  66. 66 66 John

    I’m afraid you’ve made a logical error. I’m taking your economics on board as suppositions and just dealing with the logic here:

    (1) Your claim is that Krugman has made a claim Q, which proceeds from Ricardo’s “If P then Q” (or just “P”). That Krugman assents to Q does not imply Krugman’s assent to P. I’m sure you’re aware the mistake you’ve made is known as affirming the consequent.

    As an intuition pump, consider an analogous case: Any introductory-economics claim about a market clearing when P = MC proceeds from a set of assumptions that includes perfect competition and perfect rationality. These assumptions, however, are admittedly purely heuristic and do not hold in reality. But I’m sure you’ll agree that does not negate their ability to generate insights and predictions.

    (2) A question. You state:

    “Here we have statement (A): All marginals should be equalized.
    Here we have statement (B): When one marginal is increased, all others should be raised.
    Krugman asserts that (A) implies (B). To get from (A) to (B), you’ve got to assume (C): All marginals are currently equal.”

    Strictly speaking, this is not true. x ! y.

    Countermodel:
    Suppose all marginals are not equal: one is 5, the rest are 3. Increase the first to 8. To equalize the marginals, all others must be raised.

    Or: Suppose all marginals are not equal: one is 3, the rest are 5. Increase the first to 5. Now all marginals are equalized. No others need be raised.

    Being charitable, I’m assuming there’s something in the specifics of marginal this-and-that that obviates these cases, yes? Or…no?

  67. 67 67 John

    Something got mangled in transmission. Trying again:

    “Strictly speaking, this is not true — “x is-not-less-then y” does not imply “x is greater than y”.”

  68. 68 68 David Welker

    Steve Landsburg,

    Thank you for the correction. Yes, I meant suboptimal in that particular sentence.

    What you have missed is that any spending cuts (and spending increases) that should occur because we have subomptimal allocations should occur regardless. Disaster relief has nothing to do with it.

    If cutting these programs that are wasteful (or otherwise changing suboptimal allocations) was HARD before a disaster came along, they will be HARD after the disaster. What is the theory? That because there was a disaster, the political and calculation problems that prevented an optimal allocation before will somehow go away? Of course they wont.

    Clearly, Krugman’s argument works to illustrate the point that there are various tradeoffs. It illustrates (but doesn’t prove) that proceeding through spending cuts alone is unlikely to be correct. To the extent that some programs are overfunded, likely some other programs are underfunded. This is true even if you think the overall level of government funding should be decreased.

    We can think of optimal in two ways. First, we can think of it in terms of what we can best achieve in a perfect world. Second, we can think of “optimal” as doing our best in this world, given the constraints of politics and imperfect calculation. We know that in the “perfect” world (which is a model) that disaster relief should result in a combination of spending cuts, tax increases, and deficit spending. In the real world, which is “optimal” rather than optimal, it is more likely than not going to have the same implications. After all, whatever structural problems preventing our “optimal” world from becoming an optimal world are unlikely to change due to a disaster.

    Overall, I think your response to Krugman is very interesting, as it illustrates the general weakness of economic models when translated into the real world. Krugman, who I feel appreciates the downsides of economic models, but probably not to the extent I think he should, might be wise to think more carefully about such translation problems. That said, in this case, I think the model is probably “good enough” so that we can say that with a high degree of confidence that the best solution for us in our less than perfect world is not an extreme one that involves solely spending cuts. (Do you really believe that such spending cuts would be allocated optimally? Either do I. Indeed, even if we did live in a world where the best course of action was 100% spending cuts, it would be perfectly possible and even likely for such cuts to be implemented so badly – on the basis of politics – that we would be thoroughly worse off and even farther from the optimal mix than before. After all, it wouldn’t be exactly shocking to see programs that are underfunded relative to what is optimal cut and those that are overfunded spared in a plan that was 100% spending cuts.)

    Keep in mind, that all economic models have weaknesses when translated into reality. If I recall correctly, Krugman won his Nobel Prize precisely by hypothesizing certain conditions where “strategic trade” might be superior to “free trade.” Of course, neither Ricardo’s model of comparative advantage nor Krugman’s modified model exactly fits reality. And that IS a real problem and a major weakness in the field of economics generally.

    So, while I think you have identified a real problem with Krugman’s argument, I think it is the very same problem that exists with any economic argument that relies on economic models. Namely, a tendency for models not to map into reality perfectly, and in a way where it is hard to determine, whether quantitatively or qualitatively, the precise significance of that failure.

    Basically, you have not made a criticism of Paul Krugman’s argument that could not be made about any argument that proceeds from simplified economic models. That is, it is a criticism that could be made about pretty much all of standard mainstream economics. Given that Krugman’s stated aim was to make a standard economic argument (where all standard economic arguments depend on imperfect models and thus are subject to similar criticisms as you have aimed at Krugman) rather than a perfect argument, I don’t think you can really say he has failed.

  69. 69 69 Jonatan


    3) Krugman concludes that appropriate policy is to raise B and C (by raising various forms of taxes) and to raise D and E (by cutting various forms of expenditure). Specifically, he says What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes.

    4) If in fact A=B=C=D=E to begin with, then this is right. But if they are not equal to begin with, it is wrong. For example, suppose that initially, A=2, B=1, and C=4. Now A goes up to 3. Then to optimize, you want to raise B, but you want to lower C.

    This is exactly right, and noone disagrees with this (I think). The point where I, and others, are in disagreement with you is about what Krugman said in his post.

    To me it is quite clear that all he is saying is that ASSUMING 4) THEN 3). In other words he is saying exactly what you are saying above. In the real world, the marginals are not equal, so 4) is not true. Krugman does not deny this in his article. He is arguing under the assumption that the marginals are equal. We agree with your overall sentiment, the problem is you are accusing Krugman of saying things he does not say.

  70. 70 70 Rob

    Rob:

    Of course economists are going to assume a starting point of equillibrium!

    But Krugman does not start from an equilibrium; he starts from an optimum.

  71. 71 71 Keshav Srinivasan

    Wait, if the assumption for cutting spending is diminishing marginal benefit, then are you allowed to *increase* spending on things with *increasing* marginal benefits?

  72. 72 72 mark

    but Cantor never considers cutting it out of defense pork, or applying a surtax to the billionaires in the Irene-affected areas

    which put the lie to any such ‘good parenting’ or any principle at all, for that matter

  73. 73 73 Dilip

    I don’t believe you would’ve missed it but Krugman has a response today here:
    http://krugman.blogs.nytimes.com/2011/09/01/cantor-counters/

  74. 74 74 Joe Gould

    Oh, my.

    Why have you reduced the word ‘gadfly’ to an adverb? And just what do you mean by it?

    Now, where did I put my poetic teeth?

  75. 75 75 Roger Schlafly

    Krugman reiterates his faulty reasoning. The govt spends money on disaster relief every year, and hurricane Irene was not particularly unusual. The death toll was much less than many previous years, and most of the property damage was insurable. Spending money on appropriate disaster relief should not increase the deficit, and there are good reasons to say that any increase should be paid for somehow.

  76. 76 76 Steve Landsburg

    John:

    Countermodel: Suppose all marginals are not equal: one is 5, the rest are 3. Increase the first to 8. To equalize the marginals, all others must be raised.

    If you read Krugman, you will see that what he is saying is that for any possible initial values, if any one marginal goes up, then the others must be raised. The only way this can be true is if the initial values are constrained to be equal.

    In your example, if one of the marginals rises from 3 to 4, the first must be lowered, not raised.

  77. 77 77 David Welker

    Interesting.

    Paul Krugman concedes that since a critical assumption of his model is not met, namely, a previous allocation of resources that was optimal, that the model has few implications for policy.

    But, don’t all economic models make critical assumptions (such as rationality, maximizing behavior, etc.) that are not true? Yet, we still think that those models might clarify thought about the real world, just so long as we keep in mind the significance of those assumptions.

    This brings up the question. What makes this model any different?

    We might think that the unrealistic assumption of “optimal” allocation is especially bad. But even then, is it really? After all, to the extent that we cannot in practice fix such suboptimal allocations, doesn’t this suggest that there might be “costs” (whether political or calculation costs) that are preventing us from reaching “optimal” policy? But wait, to say policy is truly optimal, you must take into consideration all costs. I would think the same is true if you say that policy is suboptimal. To the extent that you are ignoring costs and assuming infinitely malleable policy without taking into consideration political constraints, you are in fact making a major calculation error.

    An example might be in order. It is as if we are at the production possibilities frontier at the point that maximizes utility, doing the best we can do. And someone says, well, I bet we can make a lot more widgets if we use magic! So this production mix is actually suboptimal! It would only be optimal if we used magic!

    Well, I don’t think much of more of the argument that we can magically change our political institutions to make them infinitely malleable than that we can use magic to increase the production of widgets. For example, our political system makes change costly, which makes our policies much less malleable. It does so in order to preserve liberty, the thought being that a government that can change policy too easily is not really optimal, because of the costs that such a government would impose on the liberty and also on the planning and reliance interests of private parties. That such liberty interests, planning interests, and reliance interests are not easily quantifiable does not mean they are not important. While I am personally inclined to believe that our system makes change too difficult, I can’t deny that it makes change difficult for a good reason.

    Anyway, all I am saying is that, contrary to both Krugman and Landsburg who seem too quick to say that the assumption of optimality undermines the usefulness of Ricardian public finance model as a possible guide to public policy is, not so fast! I agree that the assumption of optimality is unrealistic. But that is just par for the course when it comes to assumptions used by economic models. Explain to me why the assumption of optimality is not “good enough” for purposes of analysis.

    At the very least, the Ricardian public finance model works well at isolating the impact of a temporary need to increase funding for disaster relief. In the model such funding should come from multiple sources, namely current spending, current taxation, future spending, and future taxation. I would suggest that this is probably a good model for how such financing should ideally take place in the real world as well. While no one denies that public policy is far from optimal, maybe it is “optimal” given the concerns about liberty, planning, and reliance interests that caused us to design a government that makes policy change difficult in the first place. The need to fund a sudden disaster does not change these structural impediments to change that were purposely introduced into our system.

    That said, my general position is that economic models of any kind are not “all that” anyway. I think the Ricardian public finance model may be just as acceptable as an economic model as the Ricardian theory of comparative advantage. (Ricardo was obviously a genius, by the way.) In both cases, I think there are serious and important objections to applying these models to the real world. Indeed, in both cases I think there is a possibility, no, a probability, of either or both models being misused. But in both cases, if used cautiously, I think these models can usefully help us to clarify our thoughts about policy.

    So, I happen to think that the Ricardian public finance model has more applicability to our current situation than either Landsburg or apparently Krugman. And I think Krugman is probably being inconsistent in conceding that the problems with the assumptions underlying the Ricardian public finance model are “really bad” but then overlooking, to some degree, similar problems that exist in other economic models. I think this may very well be an instance of selective reasoning or, shall we say, selective conceding.

  78. 78 78 Fritz Mills

    I think your teenage metaphor is completely wrong. This is a situation where your teenager has been hit by a car and is injured, and you’re demanding that he do his chores before you take him to the emergency room. That’s not an appropriate response. Neither is Cantor’s demand. You don’t need any theories to figure that out. You only need to be a human being.

  79. 79 79 Seth

    “That is, he’s [Cantor] holding suffering Americans hostage to his goal of smaller government.”-PK

    Or, he’s just trying to be responsible and sensible. If I had to pay for a major unexpected repair, I would probably look to pay for it first by cutting my spending in other places, even if someone were telling me that I’m missing good buying opportunities that I should go into debt or ask my boss for a raise to pay for.

  80. 80 80 Steve Landsburg

    Fritz Mills:

    You don’t need any theories to figure that out.

    The substance of my disagreement with Krugman was on whether you can settle these things by pure theory. You and I say you can’t; Krugman says you can. Welcome to my side.

  81. 81 81 kan li
  82. 82 82 John

    Ah, I think I missed this. The whole post was a reductio:
    “So even there it shouldn’t be all offsetting spending cuts.”
    “So [even there] the bottom line is that basic, regular economics says that Cantor isn’t making sense.”

    But this doesn’t make sense to me at all:
    “Landsburg points out, correctly, that the proposition that a spending increase should be offset with a little bit of pain everywhere and everywhen — that is, with higher current and future taxes and lower current and future spending on many things — follows from assuming that the government starts from a position of doing the right thing.” (—Krugman)

    But “P” still does not follow from “If P then Q” and “Q”.

    Even supposing it did, I find it very hard to believe that Cantor has Richardian public finance theory in mind in his call for offsetting hurricane relief with spending cuts.

    Krugman’s argument seems to be: Assume Cantor’s policy prescription derives from a Ricardian model of public finance, then:

    (1) Suppose Cantor is correct and we should offset relief with spending cuts.
    (2) Cantor’s Richardian model implies that higher taxes should be included in the offsets.
    (3) Cantor would surely be unwilling to abide a tax hike.
    (4) Therefore, Cantor’s prescription is self-defeating.

    But the hidden assumption about Cantor appealing to a Ricardian model is false, so this argument seems to do very little work.

  83. 83 83 Awenrain

    This thread is fascinating.

    I’m not an economist and don’t play one on TV. Neither have I spent the night in a Holiday Innm Express, so there’s that too. (Apologies for the Madison Avenue references…)

    Putting theoretical economics aside, is the essential point the correct basis of government policy? If American citizens are suffering, the government can choose to pursue a policy of action (response) or a policy of inaction (no response). This choice can be based upon any number of principles, including those economic, political or moral. (There are certainly more possibilities, these will suffice for this illustration)

    If our government is going to base its policies on purely economic principles, what sort of country is this? The political perspective is subject to endless and ultimately futile debate, so I’m choosing to ignore it.

    So we’re left with the idea of guiding policy based upon moral principles. It is proper for government to repond to disaster simply because it is the right thing to do? No matter the economic or political consequences?

    This seems patently obvious point, but also one too little made in political discourse these days. Analogous to the idea of hurricane response, consider the response of making war. Any reasonable politician during WWII knew that responding to the threat of Fascism was bad fiscal policy, but they pursued it anyway because it was a right and necessary thing. Do you think any of them were performing marginal benefit calculations?

    So, to close this long-winded harangue, it’s a wonderful thing to debate whether Krugman or Cantor are correct on the economic merits. But it’s all beside the point on the moral merits. If we live in a country where the decision of whether or not to ease the suffering of our citizens becomes nothing more than an opportunity to score intellectual points or to club the opposition, then we’ve become a country devoid of soul. And that is a very, very sad thing…

  84. 84 84 David Welker

    The substance of my disagreement with Krugman was on whether you can settle these things by pure theory. You and I say you can’t; Krugman says you can. Welcome to my side.

    Now you are just making Krugman into a strawman. He never said you could resolve the issue with pure theory. Not once. Read the blog post again. Instead all he did was assert that economic theory did not support Cantor’s position.

    That said, I don’t believe I am in perfect agreement with Krugman either. But for different reasons. At least I don’t make a strawman out of his actual argument, though.

  85. 85 85 Steve Landsburg

    David Welker:

    He never said you could resolve the issue with pure theory.

    Actually, the entire content of his post was that you could resolve the issue with pure theory. Note:

    What this does is raise the marginal benefit of spending on disaster relief. The appropriate response is to move all the marginals to get them in line: spend less on everything else, and also raise more in taxes. So even there it shouldn’t be all offsetting spending cuts.

    This is a purely theoretical argument; it invokes absolutely no specifics of the case.

  86. 86 86 David Welker

    Okay, in my view, this comes down to the definition of resolve. Here are two ways to look at it:

    (1)
    One way is to resolve something is to do so deductively. Because x + y = 5, it follows that x + y + 6 = 11. Alright, we have arrived at an absolute answer here about the value of a + b + 6 through deductive logic.

    (2)
    Another way to “resolve” something is to give “persuasive” reasons why you think the “right” answer is “probably” A rather than B.

    I originally thought that Krugman was doing the latter. In general, I think that most of economics is based on the latter sort of reasoning, rather than former.

    Based on his response, it doesn’t seem he was engaged in the latter (namely creating an admittedly unrealistic and imperfect model to use as a guide in heuristic decision-making), although it isn’t entirely clear. Instead, he says he was trying to refute Cantor. But it isn’t exactly clear that Cantor has a theory so much as a preference. I suppose that Krugman’s warning that doing “rational analysis” of Cantor’s demand might suggest that Krugman was aware that he might be imposing a more sophisticated theory, where there was in fact none.

    I will say this, both Krugman’s initial post and his follow up definitely leave something to be desired regarding clarity, at the very least. Maybe there are some problems regarding correctness too, though this hasn’t been definitely resolved in my mind. Surely if he was trying to resolve the question deductively using approach 1, he was way off. The thing is, I just have a hard time thinking that is what he was doing, because that really is not what economic modeling is generally about.

    Regardless, of whether the problem with his post is clarity or correctness or both, I think you have contributed something interesting in your critique of Krugman’s post.

  87. 87 87 Ben

    Let’s follow Cantor’s view of morality and disallow states who are net takers from the federal government from receiving emergency assistance without offsets. But on the other hand, we also should not hesitate to give assistance to states that are net contributors. New Jersey, New York, New Hampshire, Massachusetts, Connecticut, and Delaware are all net contributors, but Vermont is a net taker. So even though Vermont was more damaged than the other states, we should withhold emergency relief from them and give extra to the states with the biggest net contribution (NJ and Delaware.) Anyway, this is the outcome if you treat government spending the same as family spending. It’s the outcome when you can’t tell the difference between an analogy and reality. It’s the outcome when you think that economics is about morality, like Marx or Ayn Rand (ironically similar), rather than about logic, like Keynes or Friedman (less though still ironically similar).

  88. 88 88 Henry

    There is one way in which Ricardian public finance theory is very useful: in cases where politicians have previous revealed preferences over their beliefs in optimal allocations of spending.

    This is most applicable in situations where a party has or had effectively one party rule. For example, after spending 13 years in majority government, the UK Labour Party now opposes many spending cuts despite the economic downturn. But if Labour had allocated government spending optimally according to its own preferences, it should have done them in such a way that it would be optimal to cut spending in the event of a shrinking pie. They could try to come up with a story for why this is consistent, but it would be difficult. Unfortunately, since few voters understand Ricardian finance theory, they don’t need and can simply rely on status-quo bias to win votes.

  89. 89 89 Rob

    You are right: I incorrectly said, ‘Of course economists are going to assume a starting point of equillibrium’, when I should have said, ‘optimum’. You also took this quote out of its context, removing my argument entirely. How manipulative and misleading!

    Assuming maximizing behaviour leading to an optimum is just like consumers maximizing utility, producers maximizing profits, and social planners maximizing social welfare. Standard practice in economics. Or is all of economics Panglossian?

    Krugman is simply pointing out that there is no general case to be made for matching buget cuts in the event of disaster relief spending, despite Cantor implying otherwise. Krugman answers the question in the correct way, and comes up with the right answer.

    Perhaps you do not agree that Cantor was trying to make a general point about budget cuts being necessitated by disaster relief spending. But that does not make Krugman wrong. You just think he is answering the wrong question. And if Cantor had no general point to make, then he is just being mean by holding disaster relief hostage to his political agenda.

  90. 90 90 Steve Landsburg

    Rob:

    Assuming maximizing behaviour leading to an optimum is just like consumers maximizing utility, producers maximizing profits, and social planners maximizing social welfare. Standard practice in economics.

    This is not true. We do not assume that maximizing behavior leads to an optimal outcome; we prove it, via arguments that have certain premises which clearly do not apply to the making of government policy. (For example, to get these results we need to assume that there is a set of prevailing prices at which trade takes place; I don’t see any analogue of that set of prevailing prices in the legislative process.)

    Krugman is simply pointing out that there is no general case to be made for matching buget cuts in the event of disaster relief spending, despite Cantor implying otherwise.

    I don’t see any evidence that Cantor has tried to imply anything about a general principle; thanks for acknowledging this viewpoint in the first line of your last paragraph.

    And if Cantor had no general point to make, then he is just being mean by holding disaster relief hostage to his political agenda.

    How does this follow? It seems to me that one can a) believe that a particular policy is a good policy, b) believe so on the basis of particulars, not general principles, and c) therefore support that policy for good reasons. I think, for example, that Krugman’s support for raising taxes on high-income earners is in this category. He supports this policy because he thinks that in current circumstances it’s a good idea, not because he thinks it follows from some universally applicable principle. Does it follow that he’s just being mean?

  91. 91 91 Floris

    I think the point made by Krugman is still valid even outside of the optimum. Suppose that the US government spends too much. Clearly the marginal benefits of spending are too low in such a situation and the government should cut spending irrespective on any natural disaster. Now suppose that a hurricane hits the nation. Marginal benefits of disaster relief are clearly very large, but it seems to me like the marginal cost and benefits of other expenditure are largely unaffected by this event. Hence, the occurence of a hurricane does not at all increase the urgency for spending cuts. Therefore, even if we start of at non-optimal policy still the choice whether or not to provide disaster relief should be taken independently of the choice whether or not to cut spending under first-best policy.

    Once you include second-best arguments such as your public choice argument, things might be different. But the direction in which this effect goes is not entirely clear. Suppose that GOP only wants to provide disaster relief if spending is cut while democrats only want to provide relief if taxes for the rich are raised. This is a game of chicken and in a worst case scenario there is no disaster relief and no spending cuts or tax increases. We have seen this game of chicken before when the debt ceiling needed to be raised. The effect was that it took to long to raise the debt ceiling and the US lost its triple A credit rating. I personally think that US politicians play very dangerous games when they combine very urgent matters such as disaster relief with much less urgent matters such as spending cuts. From a public choice perspective I would not recommend playing these games.

    Summarizing, from first-best policy perspective spending cuts and disaster relief are completely unrelated. From second-best policy perspective, the combination of urgent with non-urgent matters could lead to dangerous games of chicken!

  92. 92 92 Fritz Mills

    Steve,

    That’s just ingenuous to say I agree with you. In your reply to me, you say, “The substance of my disagreement with Krugman was on whether you can settle these things by pure theory. You and I say you can’t; Krugman says you can.”

    However in the third paragraph of your piece you say, “Since Krugman has carelessly neglected to spell out an important detail of his argument, let me fill in the gap for him: The Ricardian conclusion does not come from thin air; instead it follows logically from certain premises, key among which is that you’re starting from an ideal policy regime.”

    You then proceed to back up your defense of Ricardian theory, and attempt to settle the question, not by getting into the weeds of Ricardian theory (“Rather than write down the relevant equations, let me give you an example:”) – instead with an analogy of a teenager who doesn’t clean his room, and suggest that that’s somehow analogous to victims of a hurricane. My point is that victims of a hurricane are not anything like irresponsible teenagers. They are like teenagers who have just been hit by a car. When someone is injured, you don’t make them clean their room before you take them to the emergency room. That’s an inappropriate response.

    So we do not agree at all. I maintain that your metaphor is inappropriate. I don’t think we need any theories because the question is a simple, human one: are hurricane victims analogous to irresponsible teenagers who must be punished, or are they like a teenager who’s been hit by a car and needs immediate help? You maintain they’re irresponsible teenagers that deserve punishment, and I believe they’ve been injured and need immediate help.

  93. 93 93 Steve Landsburg

    Fritz Mills:

    I maintain that your metaphor is inappropriate.

    Fine, so make it a teenager who’s been hit by a car and needs immediate help. Nothing changes. He still has to give up some activities, and there’s still the issue of which activities to give up.

  94. 94 94 iceman

    Floris – “Suppose the US government spends too much…now suppose that a hurricane hits…the marginal…benefits of other expenditure are largely unaffected by this event.”

    Yes but now the new target for those MBs has been raised, from the MC of the taxes used to fund them to the new “very large” MB of disaster relief, so you will have to cut more to equate them. The bottom line is that we have a new spending priority in a world of finite resources.

    Ben — “It’s the outcome when you think that economics is about morality.”

    Now now, nobody has suggested selectively disbursing federal relief to politically favored localities…that’s what the stimulus was for! (Cue rimshot)
    It sounds like you are conflating an economic argument – how best to fund disaster relief (which was explicitly assumed to be desirable) – with a moral one – that demanding spending offsets represents an unseemly threat to withhold such relief, but demanding higher current or future taxes does not. But hey, at least that was PK’s fallback position too. As others have pointed out here, the “hostage” rhetoric cuts both ways. Can anyone really argue with a straight face that we couldn’t find a few billion dollars in lower MB programs?

  95. 95 95 Eric Nilsson

    Perhaps I’m wrong, but David Ricardo developed his theories in the early 19th Century, when gold was the standard for the British Pound. I may be incorrect, but I don’t believe deficits came into play at that time. Most spending by the British went for the military, especially early in the 19th Century.

    I can understand the call to match spending with both cuts and taxes, but deficits really didn’t come into play until Woodrow Wilson’s bonding for World War I; even that was repaid within a few years, something that appears difficult today.

  96. 96 96 Steve Landsburg

    Eric Nilsson: What step in the Ricardian argument do you think is invalidated by a deficit?

  97. 97 97 Eric Nilsson

    Steven,

    Forgive me, but my interest is more in the history of the times than its economics. I have read some of Ricardo, just as I have read some of Smith, Malthus, Marx, and others, but my main interest is the history. Please correct me if I am wrong regarding England’s use of wealth in the early 19th Century or deficit spending prior to the Great Depression.

  98. 98 98 Steve Landsburg

    Eric Nilsson: I’m sure you know the history better than I do. But I’m equally sure that the equimarginal principle is applicable in virtually any historical circumstances.

  99. 99 99 Eric Nilsson

    Please call me Eric, most people do. If two Erics are involved, the ERic N or the full name will be necessary. The “Name Required” implies full name, so that’s what I’ve given.

    I’m not afraid to say I will have to read Ricardo for a better understanding. As for the equimarginal principle, I’ve also some research. It seems this how how one learns, so thank you.

  100. 100 100 Eric Morey

    Steve Landsburg: “Fine, so make it a teenager who’s been hit by a car and needs immediate help. Nothing changes. He still has to give up some activities, and there’s still the issue of which activities to give up.”

    It would seem that the discussion over the the issue between which activities is only appropriate after the emergency care is committed to. The fact that even prominent economists need time to work through the basics of economic theory and how it applies to the effectively infinite possible spending allocations of the government is a clear sign that brinksmanship is unlikely to lead to a directionally optimal outcome while delaying or denying the clearly ethically proper response. For this reason Cantor is reprehensible. So much so that it is hard to separate this from your critique of Krugman.

    Given that A B C etc. are not equal, what might be an optimal allocation of government activities going forward. Clearly, given favorable debt costs it seems obvious to me that additional debt is the best initial change and possibly a significant portion of of the change between the pre-hurricane budget and medium term post-hurricane budget. In other words, pay for the emergency spending with debt now and decide how to allocate spending going forward using MB/MC analysis.

  101. 101 101 Steve Landsburg

    Eric Morey:

    Clearly, given favorable debt costs it seems obvious to me that additional debt is the best initial change and possibly a significant portion of of the change between the pre-hurricane budget and medium term post-hurricane budget.

    This seems to me to be clearly wrong. Surely if there are some currently funded projects that one believes to be very ill-advised, one must believe it wiser to cut those projects than to incur additional debt.

  102. 102 102 Rob

    You seem to be intentionally misinterpreting my words. When I wrote, “Assuming maximizing behaviour leading to an optimum…”, I meant that the starting point is assuming maximizing behaviour, which leads to an optimum. I did not mean that the optimum was also assumed!

    It strikes me as a standard approach to assume that the government is operating to maximize social welfare by purchasing goods and services. This results in the starting point assumed by Krugman. Of course you can poke holes in this model. You always can in economics. It just seems like people always choose to attack the model assumptions when it suits their point of view.

  1. 1 Know Your Premises | Daily Libertarian
  2. 2 Panglossian Economics « Daniel J. Smith
  3. 3 Quando as premissas são jogadas no lixo… « De Gustibus Non Est Disputandum
  4. 4 Equivalência Ricardiana só é boa quando eu quero « Microeconomia, Econometria e muito mais
  5. 5 Why the “Stimulus” Failed to Stimulate « Politics & Prosperity
  6. 6 Reboot Congress: Paul Krugman Schooled on Ricardian Theory
  7. 7 Cantor Counters - NYTimes.com
  8. 8 Talk About Chutzpah
  9. 9 Krugman Followup at Steven Landsburg | The Big Questions: Tackling the Problems of Philosophy with Ideas from Mathematics, Economics, and Physics
  10. 10 Today’s Amazing Jobs Report « Rhymes With Cars & Girls

Leave a Reply