You can’t address macro questions with micro tools

Suppose the recent fiscal stimulus had no impact on NGDP.  Also suppose that fiscal stimulus was aimed at some states but not others, some job categories but not others, some industry categories but not others.  Would a cross-sectional test show that fiscal stimulus impacted the relative share of jobs in certain states, industries and professions?  Almost certainly yes.  And if total employment and output was unchanged, would those relative shifts also impact the absolute employment levels in various sectors?  Again, the answer is clearly yes.  I thought this was pretty obvious, but people keep citing these cross-sectional studies as if that have something interesting to tell us about the efficacy of fiscal stimulus.  Here’s how Brad DeLong titles a new post discussing a recent cross-sectional study by James Feyrer and Bruce Sacerdote at the NBER:

Did the Stimulus Stimulate? Real Time Estimates of the Effects of the American Readjustment and Recovery Act

The post should have been entitled:  “Did the stimulus redistribute output?”


Tags:

 
 
 

34 Responses to “You can’t address macro questions with micro tools”

  1. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 08:11

    These aren’t micro tools.

    Who taught you micro.

  2. Gravatar of Scott Sumner Scott Sumner
    8. February 2011 at 08:22

    Greg, Gary Becker and Donald McCloskey. Who taught you?

  3. Gravatar of W. Peden W. Peden
    8. February 2011 at 08:28

    As far as I can tell, in a situation of tight/tightening monetary conditions, all the old Treasury View arguments apply. That seems to me to be the only consistent, non-question-begging way of explaining the history of fiscal stimuluses.

  4. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 08:31

    Paul Heyne.

    Ask Peter Boettke …

  5. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 08:34

    Well, I’m not a fan at all of Becker’s attempt to make choice theory “scientific” — “De Gustibus Non Est Disputandum” is important as part of showing which way is a dead end in making economics a science.

  6. Gravatar of Morgan Warstler Morgan Warstler
    8. February 2011 at 08:35

    This is why we say, “we’re not even sure Macro exists.”

    Anything that disappears when there is only one currency, is likely smoke and mirrors.

  7. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 08:37

    Scott — the point is these are statistical tools.

    These are not the tools of the logic of marginal valuation, i.e. _micro_.

    It’s just a linguistic point, a point of clarification.

  8. Gravatar of Bogdan Bogdan
    8. February 2011 at 08:40

    I agree that agregate demande matters; I’m skeptical about fiscal stimulus; I’m uncertain about the long run consequnces of monetary stimulus. Right now, different stories are not exclusive.

  9. Gravatar of Mark A. Sadowski Mark A. Sadowski
    8. February 2011 at 09:16

    Scott,
    This is about the shortest and funniest smackdown I’ve ever seen. However, just to play Devil’s Advocate, how about cross sectional country studies of the effects of fiscal stimulus? (Where there is no common currency.)

  10. Gravatar of Charles R. Williams Charles R. Williams
    8. February 2011 at 09:25

    This discussion begs the question of whether the recent stimulus had an effect on NGDP at the aggregate level. There is no way to rigorously answer that question. It is even harder to answer the question of whether the stimulus leaves us better off. If you treat the economy as self-contained sectors – dismissing any effects of the totality of the stimulus on the total economy – then you can try to estimate the differential impact on one sector vs another and extrapolate the benefits to the stimulus package as a whole. Their approach, assuming it does make sense, can be looked at as applying macro tools to open economies.

    The real issue here is their presuppositions.

    Suppose the federal government bribed the people of South Dakota to build a $10 billion subway system in Bismark, the state capital. Probably every unemployed construction worker in the state would sign up to work overtime at premium wages. You could certainly use macroeconomic tools and concepts to measure what happens to GDP. Can you then extrapolate that to a $4 T boondoggle program throughout the US?

    Common sense tells us that if we bribe people to do irrational or destructive things on a grand scale and fund this with debt which their children will have to pay off, nothing good will come of it. It doesn’t change anything if the money spent on boondoggles gets counted as GDP.

  11. Gravatar of Morgan Warstler Morgan Warstler
    8. February 2011 at 09:30

    http://research.stlouisfed.org/fred2/series/M2?rid=21

  12. Gravatar of Scott Sumner Scott Sumner
    8. February 2011 at 09:46

    W. Peden, I think you’ll like my new post.

    Greg Ransom, I do like Becker.

    Micro is concerned with individual sectors and professions, macro with the entire economy.

    Morgan, Never heard of closed economy macro? Do you think deflationary shocks would no longer affect unemployment with a single money?

    Bogdan, Check out my new post.

    Mark, Thanks. Yes, cross country comparisons are fair game.

    Charles, You said;

    “then you can try to estimate the differential impact on one sector vs another and extrapolate the benefits to the stimulus package as a whole. Their approach, assuming it does make sense, can be looked at as applying macro tools to open economies.”

    But does it make sense? I don’t see how.

    Morgan, If I was Milton Friedman, I would have found that graph to be interesting.

  13. Gravatar of johnleemk johnleemk
    8. February 2011 at 09:58

    Well if we’re talking about old teachers, I took macro with James Feyrer… I’ve also had dinner with Bruce Sacerdote.

    This post is a bit opaque, but if I understand it correctly, it implies that the whole methodology of studying fiscal policy impacts based on variations across states/counties is flawed?

  14. Gravatar of Morgan Warstler Morgan Warstler
    8. February 2011 at 10:34

    Scott, I don’t believe in unemployment, I just think we haven’t adopted Guaranteed Income yet.

    And with only one currency printed only for population growth.
    With everyone working.

    It’s very hard for me to see Macroeconomics as a science.

    Physics and math do not disappear when currency and bad policy go away. Philosophy doesn’t ceases to exist when the government ends UI and minimum wage and does GI instead.

    Scott, you get judged on whether and why you think that is an interesting graph.

  15. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 12:06

    My point about the dead end which is “De Gustibus Non Est Disputandum” has nothing to do with “not liking Becker”.

    Becker should be a hero to every economist who values the expansion of the “publish or perish” iron rice bowl of academic economics.

  16. Gravatar of Greg Ransom Greg Ransom
    8. February 2011 at 12:12

    Scott, this use of a (non) persuasive definition of an essentially contested term is called “begging the question” in the science of philosophy.

    One of the pathologies of the “science” of economics is that economists use the word “micro” for all sorts of over-lapping and non-compatible things — including things that some economists call “macro”.

    You can’t compel assent to a bogus conclusion by the use of a word with multiple and overlapping uses — many of them contrary to your own specified use.

    Scott writes,

    “Micro is concerned with individual sectors and professions, macro with the entire economy.”

  17. Gravatar of Lorenzo from Oz Lorenzo from Oz
    8. February 2011 at 13:15

    Mark: Shortest perhaps. Not sure about funniest: try this (and then try to take Freud seriously ever again).

  18. Gravatar of Mark A. Sadowski Mark A. Sadowski
    8. February 2011 at 14:16

    Lorenzo,
    Scott’s not known for his brevity and this particular post hit close to home for me on more than one front so it could be that the joke is just private to me.

    After reading your link on Freud’s theory concerning fires I immediately searched for my copy of “Civilization and its Discontents” and was astonished to find that the footnote is indeed there. I gather you don’t share Freud’s complusion either (and since I have a firepit in my backyard some of my neighbors would probably have complained if I had such an unrestrained one). Yes, I think this raises many more questions about Freud’s personal neuroses than it says anything about why women have been traditionally placed in charge of the family hearth. Sometimes a fire is just a fire.

  19. Gravatar of Rien Huizer Rien Huizer
    8. February 2011 at 19:58

    Fiscal policy and microecomic intervention are the domain of politicians who use economists for certain tasks, including PR. Monetary policy is the domain of economists (not in Europe though, look at the current Chair of the ECB a Christmas tree of degrees but non in the dismal science. His predecessor was nominally a macroeconomist, but not really a scholar) who use politicians. I think is is logical that Keynesian approaches (new or second hand) are unlikely to work since their tools and techniques have redistributive effects that politicians will always exploit for their own purposes. This apart from inherent flaws in Keynesian doctrine, as you point out so regularly.

    I do have a question though regarding NGDP targeting. What would be the most effective tools and support/detection/control systems to do that and is the current Fed (including its statutory mandate) equipped to conduct such a policy with reliable results? Is there a chance that politicians (if there would be some money left fro them to play paternalism and pork to keep them away from important things) might actually like a central bank doing this? Looking at the backlash against the much more politically remote ECB from pro-democracy advocates (usually in countries (and their allies) that have just consumed their lunch for the next 20 years), I doubt we will ever see US politicians leaving economic management to experts.

  20. Gravatar of Mark A. Sadowski Mark A. Sadowski
    8. February 2011 at 20:13

    Rein Huizer,

    You wrote:

    “Monetary policy is the domain of economists (not in Europe though, look at the current Chair of the ECB a Christmas tree of degrees but not in the dismal science).”

    I didn’t believe you at first. (How could this be true. Europe is an ancient continent.) But it’s true, I researched it, Trichet only has degrees in political science and state administration.

    WTF! Are you kidding? The future of the second largest monetary zone in the entire world is in the hands of a complete novice?

  21. Gravatar of Mark A. Sadowski Mark A. Sadowski
    8. February 2011 at 20:20

    Actually what irritates me the most is the fact that I actually have more degrees than him, regardless of their relevance. Egad!!!

  22. Gravatar of Nick Rowe Nick Rowe
    9. February 2011 at 02:37

    Yep. There are 3 reasons why a cross-country study would work but a cross-state study would not work. Because fiscal policy might fail for 3 reasons:

    1. The monetary policy reaction to fiscal policy (monetary policy is country-wide not state-specific).

    2. Ricardian effects (federal taxes are country-wide not state-specific).

    3. The AS curve is vertical (labour can flow freely across state lines but much less freely across international borders).

  23. Gravatar of Doc Merlin Doc Merlin
    9. February 2011 at 05:19

    @Nick Rowe
    “Ricardian effects (federal taxes are country-wide not state-specific).”

    This isn’t actually true due to PPP differences between states: the real effect of taxes varies quite a bit between states. US federal taxes are also /very/ progressive compared to most countries so they affect states with high median incomes (and higher cost of living) much more than those with lower median incomes (and lower cost of living).

  24. Gravatar of Doc Merlin Doc Merlin
    9. February 2011 at 05:24

    @Mark:

    Since when is monetary policy the domain of economists? Baulderdash, it is and always will be the domain of lawyers, financeers, and politicians… economists will just talk about it endlessly.

    @Morgan:
    I agree about not being sure macro actually exists. /Most/ (not all, though) macro effects seem to vanish when pressed hard or seem to be spurious correlations that vanish a few decades later.

  25. Gravatar of Rien Huizer Rien Huizer
    9. February 2011 at 05:52

    Mark,

    Your research in the esteemed mr Trichet (no Dr alas), was too shallow. His first degree was in mining (the verb derived from “mine!”). No novice though, he was a pretty good minster of finance in France. Once you start delving into the backgrounds of CP presidents (Holland has a one with a law degree, but he does not do monetary policy; he is one of the driving forces behind the Basle Committee).

    As to lawyers: I once worked for a brilliant guy (who was also a talented politician), who had a law degree, a dad with almost a Nobel in astrophysics and who went to the University of Chicago got himself an economics PhD (not too mathematical in those days) became an economics professor and went on to become a top civil servant, top banker and important tax reformer. You see it is not degrees that count, it is talent. And talent is genetic, as my example shows.

  26. Gravatar of Scott Sumner Scott Sumner
    9. February 2011 at 14:51

    Johnleemk, Yes. No one, and I mean no one, denies that if the federal government spends a billion dollars building a base in Fargo ND, or a bridge to nowhere, funded by money collected from all over the country, that employment will rise in the locality where all the money is being spent. That’s a given. The question is whether the indirect effects of fiscal stimulus result in lower output in the rest of the country.

    See Nick Rowe’s comment.

    Morgan, I think macro disappears with 5% NGDP targeting. But I do prefer wage subsidies to minimum wage and UI.

    Greg, Arguments about definitions aren’t very fruitful. The question is what propositions are at stake, and what definitions are useful in evaluating those propositions. I was always taught that macro is the economics of aggregates, and micro is the economics of individual markets. That seems like a useful dichotomy to me.

    Rien, The single most powerful tool is level targeting. If the Fed isn’t willing to do level targeting, it’s job will be much more difficult. Another great tool is creating and subsidizing NGDP futures markets. But even without those two powerful tools, the Fed could have done much better in late 2008–they knew NGDP expectations were crashing.

    Nick, Thanks. I had the first two in mind, but forgot about the third.

  27. Gravatar of Mark A. Sadowski Mark A. Sadowski
    9. February 2011 at 16:36

    Rien,
    Well, I agree, talent is usually inherited. And it is talent that often enables one to acquire the credentials that make one a professional. The question in my mind is whether we want our monetary policy to be run by a bunch of talented amateurs. Many nations did similar things in the exercise of foreign policy in the previous century and we all know where that led us (and more than once).

    Many of the Fed presidents and Board of Governor members are bankers. How many of the FOMC members are qualified to be making decisions about monetary policy, on the basis of their education? Far too few for my comfort. (Kevin Warsh’s main qualification is that he’s married Jane Lauder, a granddaughter and heiress of Estée Lauder for example.) Having bankers serve on the FOMC would be similar to having accountants, or mining engineers serve on the Supreme Court. Except in my opinion the FOMC, given that it’s deciding the fate of the world’s largest economy, is actually a little more important.

    Bernanke at least is a professional monetary economist (and he’s also an extremely talented one as well). Even so he’s not handled things so well, has he? But I suspect most of his failures have stemmed from a greater desire to reach consensus in the FOMC than to exercise sound policy. (I sometimes think of him as the “FOMC meeting coffee and donut meister.”)

    But having a person who has no formal bankground in monetary economics placed in charge of making monetary policy for what is perhaps the second largest monetary zone in the world is just plain nonsense in my opinion. Trichet is the worst kind of monetary policy hawk: a talented amateur. No wonder the Eurozone is in such a mess.

    P.S. Trichet is probably much more talented than me. But I’m still on the verge of having one more degree than him.

    P.P.S. Having a rules rather than a discretionary monetary policy (such as NGDP level targeting) would solve a lot of these problems.

  28. Gravatar of Rien Huizer Rien Huizer
    9. February 2011 at 18:52

    Mark,

    Just one more on the subject of education as relevant to a person’s suitability for determining monetary policy (I like the Supreme Court one though: you would be surprised what would happen if (smart, sensible and unbiased) people but not formally trained as law professionals) from outside the profession (but with good expert staff) would sit on that Bench).

    Would the abundance of “non experts” in CB roles have effects on outcomes? Can that be measured?

    Or, given the fact that many who follow this blog, believe that by and large CBs tend to have inappropriate mandates (I certainly believe that), maybe it does not really matter what kind of people preside over these institutions?

  29. Gravatar of Mark A. Sadowski Mark A. Sadowski
    9. February 2011 at 20:07

    Rien,
    You wrote:
    “Would the abundance of “non experts” in CB roles have effects on outcomes? Can that be measured?”

    I think the best measurement is the current outcome. Do you think very many Americans (or Europeans) are satisfied by the results? Has it ever crossed your mind that having policy made by husbands of rich heiresses and talented mining engineers rather than people who have studied the subject might be partially responsible?

    And you wrote:
    “Or, given the fact that many who follow this blog, believe that by and large CBs tend to have inappropriate mandates (I certainly believe that), maybe it does not really matter what kind of people preside over these institutions?”

    I believe the only appropriate mandate is an NGDP level target mandate. In that case we only need monetary economists to monitor the results, not to make policy (and amybe once and for all we get monetary policy out of the hands of the talented amateurs who keep !@#$%^& it up).

  30. Gravatar of Rien Huizer Rien Huizer
    10. February 2011 at 05:27

    Mark,

    If a central bank with would commit to maintain “money” (as defined, let’s say a comprehensive definition) at a fixed percentage of ndgp (this was in the sixties/early seventies, not much forecasting sophistication available but also not much innovation, Bretton Woods was still alive and the country in question has some residual FX controls. But an extremely open real sector. Plus the CB could also control bank credit (the only one available, everything else was forbidden) and conduct an active reserves policy (hey, this looks a bit like China now, no CP, no CDs bonds required CB permission). Also, while doing all these things they would (but not publicly state) expect (hence come close to “targeting”) that one of the outcomes would be a stable growth path for NDGP. Would you like that?

    And what would you have done as a CB official if the country (with no output gap) would be hit fist by the US walking away from BW, then the oil crisis and then the Dutch disease. How would NDGP targeting have looked in those days and with the tools available (and without creating martkets for which the technology did not exist).

    It is not too hard to figure out which country. And what happened next.

  31. Gravatar of DanC DanC
    12. February 2011 at 04:57

    Political markets are about rewarding friends and punishing enemies thus politicians often prefer fiscal stimulus.

  32. Gravatar of Scott Sumner Scott Sumner
    12. February 2011 at 13:41

    DanC, Yes, but I still say it’s partly ignorance. If it was just corruption, they’d also do it in boom times.

  33. Gravatar of TheMoneyIllusion » Small is irrelevant (in macro) TheMoneyIllusion » Small is irrelevant (in macro)
    25. February 2011 at 09:53

    […] few weeks ago I criticized a Brad DeLong post on similar grounds.  If Paul Krugman would have read that post he would have avoided falling into the fallacy of […]

  34. Gravatar of TheMoneyIllusion » Faith in fiscal stimulus TheMoneyIllusion » Faith in fiscal stimulus
    8. October 2011 at 18:05

    […] past February I criticized (here and here) a paper by Feyrer and Sacerdote that used regional data to estimate the fiscal policy […]

Leave a Reply