Krugman Followup

What I like about people in academics is that when we disagree, we actually care about figuring out who’s right — and therefore we have a tendency to reach consensus, though it can take a while.

Anybody who blogs often enough (very much not excluding yours truly) is occasionally going to post something that, at least as written if not as intended, is objectively plain flat out wrong. Paul Krugman did that a couple of days ago, I responded, he’s responded to my response, and at least 4/5 of our disagreement is now resolved. That’s exactly as it should be.

To review the bidding:

1) Eric Cantor demands that any spending on disaster relief should be paid for through cuts in current spending.

2) Krugman makes a blog post arguing that as a matter of theory, the cost of any new spending at all should be spread over as many categories of taxes and cuts as possible. He makes a correct argument, but it’s based on the premise that the current structure of spending and taxes is optimal.

3) I respond with a major point and a minor point. The major point is that it’s nuts to assume that the current structure of spending and taxes is optimal; therefore Krugman’s argument proves nothing. The minor point is that in the nature of politics, sometimes it’s wise to hold even a good policy as a hostage while negotiating for other policies that you also think are good.

4) Krugman concedes the major point, which of course he must, because he’s objectively flat out wrong. (Again — I’ve been that myself from time to time on other issues.):

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. If you think the government’s priorities are all wrong, then theory doesn’t tell you much about what should happen.

5) We now agree that there is no principle requiring new spending to be paid for from a broad base of spending cuts and tax increases.

6) Krugman now goes on to say that Eric Cantor is asserting a principle that all new spending on disaster relief must be offset with current spending cuts. I do not know why Krugman says this; my impression is that Cantor defends this as a policy preference, not as a principle. In any event, when Krugman writes “I should have been clearer about that” (that being the question of what point he was trying to make), I think he’s being a bit disingenuous; his original post was very explicitly not a critique of the take-it-all-from-one-place principle but a detailed careful correct argument (though based on a faulty premise) in favor of the spread-it-out principle. But this is an argument over what is now (in Internet time) ancient history; Krugman and I are in total agreement on substance going forward.

7) Let me stress that: We are in total agreement on substance. Namely: There is no general principle that supports Cantor’s policy preference, and (given reasonable assumptions) there is also no general principle that supports the opposite preference.

8) Krugman then turns to the minor point: Is it ever a good idea to hold good policies hostage in order to get other good policies? I think it sometimes is; he seems to think it’s not. Or he thinks it’s not in this particular instance; I’m having trouble telling how general a statement he’s trying to make here.

9) I’d like to be clear that I never took a stand on whether in this instance it was a good idea to hold one policy hostage to another. I said only that there are surely some such instances and therefore we can’t settle this issue by appeal to general all-encompassing principles.

10) Krugman objects to the analogy in which I compare disaster relief to letting your kid go to the prom. Here he has quite entirely misunderstood the point of the analogy (whether the fault lies more in his reading or in my writing, I’ll let you decide). I was not saying “disaster relief is as frivolous as going to the prom”. I was saying: “Look, there is such a thing as relatively unimportant spending, and therefore if you’re going to argue in favor this particular spending, you’ve got to argue for it on its merits, not on the basis of some general principle that `if it’s good we should do it’ “.

11) Krugman closes by repeating the charge that Cantor has invented a nonsense principle. Here I think Krugman is being mildly libelous, because, as I said above, I’ve seen Cantor claim no such principle; all I’ve seen is a statement of a policy preference. But the language of politicians is sufficiently vague that it’s often hard to be sure what they’ve really meant.

12) The language of academics, by contrast, is refreshingly precise, which is why we’re able to reach something pretty close to consensus even when we start off in opposite corners.

Share/Save

31 Responses to “Krugman Followup”


  1. 1 1 John

    Say it three times:
    Reductio ad absurdum
    Reductio ad absurdum
    Reductio ad absurdum

    Now you own it! :)

  2. 2 2 Ken B

    It will be fun to see how those on this board who stuck up for Krugman’s now-confessed error will react.

    Let’s not forget the obvious here. Krugman, in a fit of annoyance, shot off an unthinking and pretentious blast. Had it not been for Steve’s acute catch and mocking post there would have been no correction.

  3. 3 3 David Welker

    I feel though you are seriously confused. It is as if you did not understand economics at all.

    We don’t proceed from economic models to absolute proofs about policy. We proceed from economic models to insights and subjective probabilities (as opposed to frequentist probabilities) about policy.

    That you have shown that the Ricardian public finance model is not an absolute proof that Rep. Cantor is wrong is not any more of an accomplishment than me showing you that sometimes price decreases result in a decrease in the quantity demanded, contrary to the “law” of supply and demand. We are dealing with models which hopefully reflect some aspects of reality, not models that describe reality.

    As far as it goes, Krugman is no more “flat out wrong” than someone who would predict, based on the laws of supplies and demand, that an increase in price is likely to result in a decrease in the quantity demanded. Yes, the law of supply and demand does not prove that price increases result in decreased quantities demanded. But the model usually is applicable.

    Krugman never conceded that he was flat out wrong. He merely said you were misinterpreting him, and that perhaps he could have been more clear. So it appears that you are in major disagreement still.

  4. 4 4 Steve Landsburg

    David Welker:

    That you have shown that the Ricardian public finance model is not an absolute proof that Rep. Cantor is wrong is not any more of an accomplishment….

    Surely it’s not much of an accomplishment, but it does seem appropriate in a response to a post by Krugman claiming that the Ricardian public finance model is an absolute proof that Rep Cantor is wrong.

  5. 5 5 David Welker

    But Krugman said in his reply that this was not a correct interpretation of what he said.

    I can see how he can possibly be read that way. But when someone says that is not what they meant even when they wrote it (as Krugman says in his follow up post), I usually take that as the honest truth rather than an attempt to cover tracks.

    I will say I liked your post on this topic, though. Regardless, we can agree that there was something to be desired by the original Krugman post, whether that is an issue of clarity or an issue of correctness or both.

  6. 6 6 Ken B

    @David Welker:
    “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. If you think the government’s priorities are all wrong, then theory doesn’t tell you much about what should happen.”

    Sounds like an admission of error to me. Of exactly the key point at issue for exactly the reason SL gave. I must add: “It is as if you did not understand” English at all.

  7. 7 7 PrometheeFeu

    @David Welker:

    The big difference is that demand curves tend to slope downwards and supply curves tend to slope upwards in real life. (Or at least so do most people think) So using that model to make predictions often does help. But government spending does not tend to be anywhere close to MC=MB. As a result, what Ricardo tells us is: take the money where it hurts the least. If as Cantor most likely believes there are many large government programs which are hugely wasteful, (He probably thinks some have negative net total benefit) it hurts less to take money from a big wasteful program than from taxes or by borrowing. In other words: what Cantor is saying makes perfect sense within the Ricardian framework.

    Also, the point Krugman made was not that the insights of Ricardian theory generally point us away from Cantor’s ideas. What he said was: “So the bottom line is that basic, regular economics says that Cantor isn’t making sense.” That is flat out wrong. The theory says that Cantor may or may not be right depending upon the various levels of spending and MC and MB curves for all programs.

    If Krugman wants to attack Cantor’s estimated MC and MB curves, that’s fine. But then he has to say Cantor and himself have disagreements on the numbers to put on a graph which is not quite as good sounding as insinuating Cantor is an idiot who doesn’t even understand his own beliefs.

  8. 8 8 Ken B

    I’m going to snark a bit (quite a bit actually) because I am struck by how many people
    seem just unwilling to even *try* to read.
    The lesson for today is that when an authour lays out in advance what he intends to argue, this might guide one in reading.
    Steve, alluding to Candide, writes in a mocking tone that
    “Krugman has implicitly declared in his latest blog post that we are now living under the best of all policy regimes.”
    There are several points a neophyte reader might well attend.
    There is the rubric ‘Panglossian’. Google will help.
    There is the “implictly”. This is a hint, subtle to be sure, that Landsburg will be looking at what Krugman says implicitly not explicitly.
    There is the “best of all possible”. Your ‘panglossian’ google should help with this.

    I’d say Steve is likening Krugman to Dr Pangloss.
    I’d say he’s going to base this comparison on implicit assumptions he alleges Krugman to be making.
    Armed with this insight the neophyte reader can now attend to Steve’s logic and to its goal.

  9. 9 9 John

    You’ve missed the point. Of course Krugman’s premise was wrong—-he was engaging in a proof by contradiction (or more accurately, a reductio ad absurdum).

    He’s just as wrong as you are in thinking that affirming the consequent is a valid inference form, though.

  10. 10 10 John

    “…post by Krugman claiming that the Ricardian public finance model is an absolute proof that Rep Cantor is wrong.”

    Case in point. You’ve misread him and have disregarded his clarification in his followup. If this is intentional on your part, it’s excellent debating technique but piss-poor reasoning and shows bad faith.

    Krugman set out to evaluate the economics of Cantor’s proposal under the best possible circumstances for Cantor — that is, supposing that Cantor had an economic model in mind, specifically, Ricardo’s public finance theory. (It’s perhaps debatable whether Ricardo actually does provide Cantor with the strongest possible model, but that’s another story.)

    He then reasoned that if Cantor supports offsets from spending cuts, he is also bound, by the aforementioned model, to support offsets from tax hikes.

    Krugman implicitly assumes (justifiably) that Cantor would not support such a corollary to his policy proposal.

    Therefore, Cantor “has no principle”—that is, he has no coherent economic basis for his proposal.

  11. 11 11 John

    “We don’t proceed from economic models to absolute proofs about policy. We proceed from economic models to insights and subjective probabilities (as opposed to frequentist probabilities) about policy.”

    +1

  12. 12 12 Steve Landsburg

    John:

    Krugman set out to evaluate the economics of Cantor’s proposal under the best possible circumstances for Cantor

    The best possible circumstances for Cantor would presumably mean under assumptions that Cantor would buy. Where has Cantor ever indicated that he believes all current policies are optimal?

    Let me put this another way: We all agree that there is a perfectly reasonable model under which Cantor’s proposal makes sense (i.e. a model in which there’s a whole lot of wasteful spending that ought to get cut). Instead of assuming the model that Cantor obviously has in mind, and under which his proposal makes sense, Krugman pulled out another model — one in which all policies are currently optimal — that nobody, including Cantor, has ever claimed to be applicable here.

    It’s as if you had said “My mother painted her house pink”, and I said, “Well, I’m going to evaluate that statement under the best possible circumstances for you, by assuming your mother painted her house blue. And under those circumstances, you are wrong.”

  13. 13 13 Bob Murphy

    At first I was baffled by your post, Steve. But now I get it: Just as Krugman took your devastating refutation and somehow turned it into evidence of your moral bankruptcy, here you take Krugman’s weasely post and somehow turn it into evidence of Krugman’s academic honesty.

    For an atheist, you make a good Christian.

  14. 14 14 Ken B

    Actually Bob he’d make a good old testament character: Job, whose patience he has shown here. That aside there is no reason to gratuitously insult the man; has he threatened to burn anyone at the stake, declared his adversary damned, or embraced a fairy tale?

  15. 15 15 Scott H.

    Krugman is grasping at straws here. Cantor is not citing any principle. It is very obvious that he feels the government has entirely too much wasteful spending and that needs to be cut before we go further into debt. I’m pretty sure if debt was at zero we wouldn’t be having this conversation (to make the extreme point).

    The blackmail is in the eye of the beholder. I could argue Krugman is blackmailing the Tea Party into taking on more debt than agreed to — using this disaster as the blackmail. Shouldn’t he or someone propose immediate cuts so that disaster relief can go forward? I mean, if he really cared about the disaster relief? How disgusting to use this tragedy to expand the national debt. (I’m making fun of Prof. K’s rhetoric here)

  16. 16 16 GBurtless

    “Krugman concedes the major point, which of course he must, because he’s objectively flat out wrong.” I have re-read Paul Krugman’s original blog post on Cantor’s policy preference or “principle” three times. I am hard pressed to understand where an economist as smart as Steve Landsburg finds an error (“flat out wrong”) in that post. Krugman states a general proposition about how a rational government would respond to an unanticipated disaster:

    “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 makes the further observation that, for governments that can borrow, this basic idea also applies to considering spending reductions today versus in the future and increased taxes today versus some point in the future. For temporary events or emergencies this probably means the additional costs should not all be offset by higher taxes or spending reductions today. (In fact, this pretty much describes how all America’s war have been financed; a sizeable percentage of the costs were deferred through public borrowing.)

    Steve Landsburg offers the correct, but unhelpful, observation that this reasoning only applies if all the original public spending amounts and tax rates were optimal, that is, “set all the marginals equal.” This is surely correct, but it is also fatuous and useless for helping us think about how we should pay for a war or disaster. Why? Because in the entire history of organized states, when can we confidently claim that a government has got “all the marginals equal” in the judgement of every citizen in the state? It should take readers about a nanosecond to give the right answer: Never. Therefore, what Landsburg is telling readers is that Ricardo’s (and Krugman’s) proposition has never been helpful for resolving the questions: “How should we pay for a war?” or “How should we finance aid to victims of a disaster?”

    My reading of the Krugman column is that he was trying to make the point that there is no particular reason, based on a sound understanding of public economics, that governments should offset 100% of the cost of disaster relief or wars with corresponding cuts in other current public spending. (I cannot see why it makes the least difference whether we classify Cantor’s proposed constraint as a “principle” or a “policy preference”; Krugman was arguing that Cantor’s conclusion doesn’t make much sense.) He was offering a public finance economist’s short-hand way of thinking about the problem. Steve Landsburg is right to say that if the original amount and distribution of public spending and taxes are not optimal, then who knows how we should finance wars and emergencies?

    We don’t know. But our analysis of what a sensible government would do if it “set all the marginals equal” both before *and* after the disaster is helpful in showing us one thing: There is no compelling reason to think the optimal response to a disaster is to reduce spending on other government obligations dollar-for-dollar to offset our added spending for the victims of the disaster. If Landsburg can offer a convincing argument to the contrary, I’d like to see it. Krugman offers a reasonably persuasive argument that sensible governments, faced with a disaster and pursuing policies likely to produce the greatest good for citizens, would spread the pain between tax increases now and in the future and spending cuts (on non-disaster items) now and in the future. No economist, least of all Krugman, would argue that governments are necessarily sensible.

  17. 17 17 Scott H.

    Dang. I think I made a mistake above. Even if debt is at zero we could still be having this debate as long as Cantor feels that there is wasteful spending and we would be better off eliminating it instead of increasing debt. Its still not a principle. Oh… great, now I read the same thing in your post. Double dang. Look away everyone! Look away!

  18. 18 18 Dave

    Ken B – ha! love it!

  19. 19 19 Steve Landsburg

    GBurtless:

    There is no compelling reason to think the optimal response to a disaster is to reduce spending on other government obligations dollar-for-dollar to offset our added spending for the victims of the disaster.

    Agreed, of course. Nobody ever disputed this as a general proposition.

    Krugman offers a reasonably persuasive argument that sensible governments, faced with a disaster and pursuing policies likely to produce the greatest good for citizens, would spread the pain between tax increases now and in the future and spending cuts (on non-disaster items) now and in the future. No economist, least of all Krugman, would argue that governments are necessarily sensible.

    Nor would Eric Cantor. But Krugman wants to conclude that Cantor is “not making sense” solely because Cantor has failed to embrace the proposition that governments are necessarily sensible. That’s a false conclusion in the absence of additional assumptions, which, as you point out, neither Krugman nor anybody else would be likely to want to make.

  20. 20 20 Steve Landsburg

    GBurtless: Let me follow up.

    1) Are we agreed that Krugman’s original argument would apply equally well to “prove” that anyone who wants to raise taxes on the rich without raising taxes on the poor is not making sense?

    2) Are we agreed that Krugman would probably balk from this application of his argument?

  21. 21 21 Mike H

    You seem to make the assumption (perhaps Krugman does too) that Cantor claims to have any knowledge at all of economic principles. Does he in fact make any such claim, or would he be happy to acknowledge (perhaps in different words) that he’s just making things up as he goes along, with respect to economic theory and its applications to policy?

    If so, that puts both your arguments in a different light.

  22. 22 22 Steve Landsburg

    Mike H:

    You seem to make the assumption (perhaps Krugman does too) that Cantor claims to have any knowledge at all of economic principles.

    No, I don’t think so. I think it’s more like this:

    1) Krugman attributes to Cantor a) a knowledge of economic principles together with b) a belief that everything is currently optimal. From this he concludes that Cantor’s policy prescription makes no sense.

    2) I attribute to Cantor a belief that a lot of current spending is very wasteful. From this I conclude that Cantor’s policy prescription makes (by Cantor’s lights) perfect sense.

    3) Various commenters have made the bizarre claim that Krugman is being charitable to Cantor by attributing to him a belief that is both a) ludicrous and b) clearly at odds with everything Cantor has ever said. The charitable interpretation, and I think the correct one, is that Krugman and Cantor disagree about the value of a lot of current spending, and each makes policy prescriptions that are consistent with his own beliefs.

  23. 23 23 Mike H

    If Cantor believed that a lot of current spending was wasteful, he would no doubt have recommended spending cuts irrespective of disaster relief. If so, tying the two together is not economic sense, but political sense.

    Is there any conceivable economic case for tying the two together?

    NB – I have no idea whether or not Cantor in fact recommended that. I know nothing about him except what I’ve read in your blog and Krugman’s.

  24. 24 24 Greg O

    “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. If you think the government’s priorities are all wrong, then theory doesn’t tell you much about what should happen.”

    I think Krugman believes the idea that the government’s priorities are all wrong is too absurd to include in the original argument. It’s like saying with the analogy of the child going to the prom, if the parents and children are all schizophrenics then who knows what their best strategy is. So there is a a flaw in Krugman’s argument in that he leaves out (to him) absurd situations.

    The current situation is close to equal margins but part of the way towards Cantor’s views. Taxes should be raised to some degree as well as some spending cuts and deficit spending which is Krugman’s view. Cantor’s view that taxes should not be raised in this case is an absurd view of the economy, though the temporary nature of the disaster probably makes tax raising too much trouble in the short term. I recall for example California raised its sales tax temporarily after the LA earthquake. Cantor is just rationalizing no new taxes no matter what the situation.

  25. 25 25 John

    “The best possible circumstances for Cantor would presumably mean under assumptions that Cantor would buy. Where has Cantor ever indicated that he believes all current policies are optimal?”

    Right. I completely agree with your here. But this is a different line of attack against Krugman than the one you’ve outlined.

  26. 26 26 Steve Landsburg

    John:

    Right. I completely agree with your here. But this is a different line of attack against Krugman than the one you’ve outlined.

    I think you’ve misunderstood me then; this was essentially my point all along. (More precisely: To make Krugman’s argument, one must either believe that everything is currently optimal or attribute that belief to Cantor; neither is tenable.)

  27. 27 27 bobkamath

    Well, let’s see now what kind of play you are playing. I am a mathematician myself and have three question for you actually :
    -How great an importance do you give math to actually understand things in economics ? Do you really think math has the adequate INHERENT logical consistency to be of any more help than just clarifying thoughts and giving compact representation of arguments in a science like economics ? Do you really mean, that we can prove policy proposals just like we do with the infinity of primes ?
    -Krugman asserts his point in this Cantor debate actually mainly on ethical grounds in my view, that is he says, bundling the two proposals, spending cuts and government relief of disasters, is for him morally unjustifiable, and (I think just as even your own logical digressions about whether Krugman is “right” or not) sound actually pretty cynical if confronted with the pain of the afflicted people and the duty the government has in these cases to its citizens. Social contract anybody ?
    Therefore let we put aside for some moment mathematics of equity and all that stuff which is being studied even now and that as we all know form basic mathematical logics, just as everything in math, will never boil down to sth. perfect and without flaws. What do you think, as a PERSON, as a CITIZEN, as an american, should be the way out of the debt crisis ? Do you also think, as all republicans do, that not taxing the rich is desirable/ a viable solution ?

    Thank you.

  28. 28 28 iceman

    Greg O: “The current situation is close to equal margins”

    It’s great that you have a personal, normative view of the status quo based on your own philosophical preferences…it’s not so great that you’re so willing to declare it “absurd” that others might have a different view about the near-optimality of the government’s current priorities. And again, the main point here was that Ricardo (PK’s appeal to “regular, basic economics”) had no knowledge whatsoever of what those current priorities might actually be.

  29. 29 29 Steve Landsburg

    bobkamath:

    Krugman asserts his point in this Cantor debate actually mainly on ethical grounds in my view…that is he says, bundling the two proposals, spending cuts and government relief of disasters, is for him morally unjustifiable

    I do not believe you read Krugman very carefully. The bulk of his post concerns equating marginals.

  30. 30 30 Jacob

    Steve:

    I’m Krugman fan and happen to agree with him politically, but your point is well taken. It seems to me Krugman let his politics get in the way of his economic analysis.

  31. 31 31 Steve Landsburg

    Jacob:

    It seems to me Krugman let his politics get in the way of his economic analysis.

    Either that or he just wasn’t thinking clearly. It happens to the best of us.

  1. 1 Recap at Steven Landsburg | The Big Questions: Tackling the Problems of Philosophy with Ideas from Mathematics, Economics, and Physics
  2. 2 Awesome Arguments | Shaun Miller’s Weblog
Comments are currently closed.