« Paul Dirac on economics | Main | Assorted links »
Criticisms of the recalculation argument
Matt and Ryan Avent comment, and Matt Steinglass, I'll put some points under the fold...
1. Sectoral shock theories of unemployment have a long lineage, including search theory, David Lilien (1982), Fischer Black, work by Steve Davis and John Haltiwanger, Mortensen and Pissarides, plus some recent writing by Michael Mandel and much earlier Franklin Fisher's work on disequilibrium adjustment. Avent and Yglesias suggest that Kling is making up his own macro but the innovation is simply to call the adjustment process "recalculation," to give it a more Austrian gloss. Mortensen and Pissarides are sometimes mentioned in the context of future Nobel Prize winners.
Or try Brainard and Cutler (nowadays both Obama-linked), who note sectoral shifts are especially likely to account for unemployment episodes of long duration. Here is a list of some of the relevant real shocks.
2. Ryan's summary of the argument involves several strawmen. Various polemic phrases are used throughout his post, including "makes no sense" and "nuts." When you read language like that, it often indicates the writer has not worked hard enough to imagine a sensible version of the idea he is criticizing.
3. Here are a few claims I do believe and in most cases I am on the record:
a. The AD shock today is very real, albeit overemphasized by many relative to sectoral shocks.
b. There is an optimum delay on the recalculation process. An economy can't always do all the recalculation all at once and that is one way of thinking about why some bailouts have been necessary, plus automatic stabilizers.
c. Reemployment does not in general proceed in accordance with an optimum, especially during major shocks. This follows from many (not all) of the models cited above.
d. The new, on-its-way optimum may well involve new government expenditures in various areas on a permanent basis; pick port security if you want one non-controversial example.
You can believe all those propositions, as I do, and still think that the recalculation argument means that, in absolute terms, significant parts of the current stimulus won't be very effective. As James Hamilton has pointed out, a big chunk of the problem is something other than insufficient aggregate demand and so more stimulus doesn't translate necessarily into better outcomes. In other words, we're spending lots of money for smaller "bang" than was advertised.
You might disagree with those conclusions, combined with propositions a-d, but they're not "nuts." There's a disconnect between the emotional content of the polemic Avent wants to level and the information content in his post.
3. Matt suggests that some of the critiques do not apply to most of the stimulus. He notes that aid to the states is a big chunk of the stimulus, as is tax cuts and increased transfer payments. On the transfer payments, see my point 2d; you may or may not like them but most analysts conclude, following the Bush experience, that such programs aren't very good stimulus. So the recalculation idea doesn't much apply there but the stimulus idea doesn't much apply either.
On aid to the states, the recalculation problem applies very directly (Matt says it doesn't but I don't see where in his post he gives a reason for that view). You can think that some form of state-level aid is necessary, as I do, and still see the recalculation idea as explaining why a big state-level ouch is coming in about two years' time. When (if?) the stimulus is not renewed, a painful sectoral reallocation will have to take place and right now we are only postponing that pain. By the way, it would be nice if state governments played along by having a coherent long-run fiscal plan but right now at least half of them are not doing this, thereby worsening the forthcoming ouch. Wait until you see what happens with state universities in two years' time. Ouch, ouch, and triple ouch.
Overall the recalculation idea does apply to large chunks of the stimulus, albeit not all of it.
4. We should be especially skeptical of gdp measures when: a) governments care about those measures especially much, and b) we face trade-offs between temporary and permanent gains and we are choosing the temporary gains. Fiscal theories of cyclical movements, as outlined by Rogoff and the like, deserve to make a comeback and I predict they will. In fact you can add those theories to the list above at #1.
The bottom line is this: if you're trying to use the recalculation idea to explain why the fiscal stimulus should be zero, that in my view will fail. If you're using the recalculation idea to explain why the stimulus has a lower rate of return than many people think, it hasn't much been dented by the recent criticisms. After all, if the problem were just insufficient AD, a solution would be ready at hand. But it isn't and it's not just because Obama isn't "tough enough" to propose a bigger stimulus. It's a genuinely difficult problem to solve.
I may soon consider Scott Sumner's very good recent posts on real shocks and the business cycle.
Posted by Tyler Cowen on October 4, 2009 at 07:47 AM in Economics | Permalink
Comments
"I may soon consider Scott Sumner's recent posts on real shocks and the business cycle."
Sounds arrogant. Interesting post, though.
Posted by: IWantCookieNow at Oct 4, 2009 7:55:38 AM
Let's play pretend....
Suppose there's 20% of workers involved in building and selling homes and 20% of workers involved in healthcare. Suppose that 50% of income is spent at Wal-Mart buying assorted goods, 20% is spent on homes and 20% on healthcare with 10% random other stuff.
Now suppose, all in the sudden, the economy wakes up and discovers homes are not so cool anymore. After discovering the bubble, half the workforce in homes are laid off. Spending on homes cuts in half while spending on health care booms.
To me 'recalculation' seems to imply that we will have a depression because its going to take a while to get that 10% of workers who were in the home business to get trained to work in healthcare. OK fine, so what's the argument against stimulus?
It seems to me like keeping income relatively stable so that 50% can still be spent at Wal-Mart is not so bad. While there's some exceptions, most of the bailout/stimulus policies ARE NOT centered on trying to force the 20% home 20% healthcare level to remain fixed. Most of the stimulus seems perfectly fine with seeing levels of in some sectors decline gracefully while other sectors add jobs.
Imagine without stimulus what would happen in 'recalculation'. Since the 10% of home workers are 'stuck' until they can get training their income drops to zero. In the meantime, the expanded consumption of healthcare causes healthcare inflation because the sector cannot expand supply (remember the 10% of unemployed workers will be unavailable until their skills can be upgraded). The result would be $0 income for the unemployed workers but a nice big bonus for those already with healthcare skills. Wal-Mart, in theory, should still see the same sales but maybe a change in mix.....fewer CD sales for whatever type of music home workers like and more CD sales for whatever type of music is popular with the healthcare workers. Over time, though, this rent will disappear as unemployed workers get into the health industry and bid down wages and prices.
But with stimulus what harm is done? The pain of the home worker remains but is made a bit less harsh. After all, the home workers were essentially just in the wrong place at the wrong time. Efficient markets theory would say that it is impossible to predict ahead of time what sector would have experienced a bubble and which one would boom after the bubble. (Who knows, perhaps ten years from now healthcare innovation will peak and we will be facing a glut of doctors, nurses and medical assistants who bet their careers on an ever expanding cost curve in healthcare).
To me the problems with this are pretty obvious:
1. It doesn't explain the business cycle very well. Why do we seem to run into miscalculations every 8-10 years?
2. It doesn't address liquidity very well. De Long has done a great job exploring the Keynesian approach to liquidity and financial panic and the implications it has for a modern monetary economy and how those implications differ from older goods and barter based economies. This framework seems rather primitive.
Posted by: Boonton at Oct 4, 2009 9:59:15 AM
To me 'recalculation' seems to imply that we will have a depression because its going to take a while to get that 10% of workers who were in the home business to get trained to work in healthcare. OK fine, so what's the argument against stimulus?
What makes you think they should all go work in health care? That is what needs to be calculated.
Posted by: Sonic Charmer at Oct 4, 2009 10:13:03 AM
I would agree with Boonton that the recalculation process for real estate is being allowed to happen. Some of the recalculation process for the auto industry was also allowed to occur, just at a slower pace. I would also agree with Tyler Cowen that a recalculation process for universities also needs to occur. And to that I would add banking and our prison system. But if the recalculation process in the banking and education sectors happens later than recalculating real estate, why isn't that better than all three happening at once? And the bigger question is where are these jobs supposed to go?
Posted by: David C at Oct 4, 2009 10:51:28 AM
To me 'recalculation' seems to imply that we will have a depression because its going to take a while to get that 10% of workers who were in the home business to get trained to work in healthcare. OK fine, so what's the argument against stimulus?
The argument is against specifically spending money to prop up home prices if they need to fall, or against spending money on zombie companies like GM and Chrysler that have been dying for years (and would have gone bankrupt in 2010 if the crisis, which only sped up their demise, hadn't hit.)
The other argument is that putting off the day of reckoning sometimes ensures that the day of reckoning never comes, and that the government just spends more and more to prop up an industry that will never come back.
Posted by: John Thacker at Oct 4, 2009 11:26:48 AM
Hmmm... I guess the Austrian point is that no one (with the emphasis on "one") knows the solution of the recalculation problem, @Boonton, David, John. Who says that construction *needs* to shrink? I read somewhere that the U.S. population is expected to reach 500 at the end of the century. Or the automobile industry: GM and Ford are doing pretty good in Asia.
Posted by: JSK at Oct 4, 2009 12:19:50 PM
I don't know why you even quote Ryan Avent. He is quite possibly the most willfully ignorant blogger in the world.
Posted by: Danny at Oct 4, 2009 12:23:26 PM
"Recalculation" = micro applied to the whole real world economy since the marginalist revolution.
Macro without "recalculation" = economics with a spike through the head.
Posted by: Greg Ransom at Oct 4, 2009 12:33:09 PM
"Recalculation" = micro applied to the whole real world economy since the marginalist revolution.
Macro without "recalculation" = economics with a spike through the head.
Posted by: Greg Ransom at Oct 4, 2009 12:36:08 PM
Boonton -- read some Roger Garrison and Friedrich Hayek.
You don't have any idea how much coordination across time and between adjusting relative prices and between savings, investment and consumption you are leaving out.
Your failure of imagination here is immense -- like failing to imagine the whole of the cosmos beyond the earth, moon and sun.
Posted by: Greg Ransom at Oct 4, 2009 12:49:54 PM
I read somewhere that the U.S. population is expected to reach 500 at the end of the century.
Someone predicts massive depopulation?
Posted by: Constant at Oct 4, 2009 2:20:11 PM
Ah, 500 million.
Posted by: JSK at Oct 4, 2009 3:01:59 PM
What makes you think they should all go work in health care? That is what needs to be calculated.
Because that's how I set up my hypothetical. Of course, that sounds a bit like today. After all we see healthcare is a great field to enter, pay is still high and forecasts for future demand are strong. The housing market appears, at best, to be maybe starting to recover very slowly.
But if the economy can suddenly determine that its calculations were all wrong that makes any action more risky than we originally thought.
The argument is against specifically spending money to prop up home prices if they need to fall, or against spending money on zombie companies like GM and Chrysler that have been dying for years (and would have gone bankrupt in 2010 if the crisis, which only sped up their demise, hadn't hit.)
OK but only a small portionof the stimulus was spent on 'proping up home prices' (whatever that means) and the bailouts of GM and Chrysler were more like controlled bankrupticies. Most of the stimulus was dispersed boosts in demand cross a broad array of the economy (tax cuts, unemployment/food stamps, state aid etc.) The other side of the stimulus, the monetary stimulus via low interest rates and easy money from the Federal Reserve has no bias towards 'zombie companies'. If, say, healthcare is the next big thing and housing was the old thing low interest rates benefit health investment.
Greg
You don't have any idea how much coordination across time and between adjusting relative prices and between savings, investment and consumption you are leaving out.
I think you are misunderstanding me here. I presented a simple hypothetical of an economy shifting quickly from housing to health only for illustrative purposes. I'm not saying this is what has been happening for the last year or two, only presenting a hypothetical to flesh out what you guys are trying to articulate.
Here's the problem I have. I think most people accept that a crash can happen in a sector should something unexpected happen. For instance, a mad cow scare could harm the beef industry just when everything seemed to be going perfect.
It's quite another thing to assert that we must endure a depression because one day everyone woke up and said houses are boring and something else is the next big thing. We are seeing a general glut, not a crash in a particular sector. It seems strange to assert that this is all just a rational 'recalculation' that we need to sit and wait out. More of a problem, the people who do talk like this like to blame Greenspan for the bubble and crash even though inflation was not a major issue in the 2000's. They seem to enjoy prescribing pain as punishment for our economic sins. I think Keynes and the rest of mainstream economics (incl. Friedman) were right to reject this economic version of S&M as a viable theory.
Posted by: Boonton at Oct 4, 2009 4:18:08 PM
Boonton
[What makes you think they should all go work in health care?] Because that's how I set up my hypothetical.
No it isn't. In your hypothetical, 20% of workers are involved in health care (and 20% in housing) at time 1. Then something happens and it becomes time 2, at which point all the housing workers are laid off. You claim that those housing workers should/will all be redirected to health care (and I take the conclusion to be: if/since we know this, may as well stimulate it). I'm pointing out that we don't know it. But now you're saying it follows from your hypothetical setup. But it doesn't. Even in your hypothetical, it would need to be 'calculated' where those housing workers should go. You can't know this from the state of affairs at time 1. You can't use the state of affairs at time 1 to predict where capital should be directed at a different time and state of affairs 2. After all, if you could, the housing bubble would never have burst. It was doing fine at time 1, so why would it not be doing fine at time 2?
After all we see healthcare is a great field to enter, pay is still high and forecasts for future demand are strong.
Wait, so now you're not talking about the hypothetical anymore, but what we 'see' now? Well okay then. So you're saying that based on current prices, demand, reasonable forecasts etc., it appears that health care is a good field to be in? Well hey, maybe you're right. Maybe those price signals are flowing to you well, and maybe you've interpreted them correctly. But I thought that's precisely what Kling means by 'calculation'. If the price signals are working and accurate, there's no argument in there for the government to do anything. (And if they're not, then government action can distort things and overshoot.)
Posted by: Sonic Charmer at Oct 4, 2009 5:35:34 PM
how is the economy supposed to find the next near equilibrium state when people are hoarding and when viable businesses are being killed for lack of financing?
Posted by: babar at Oct 4, 2009 8:33:22 PM
Sonic
1. Then something happens and it becomes time 2, at which point all the housing workers are laid off. You claim that those housing workers should/will all be redirected to health care (and I take the conclusion to be: if/since we know this, may as well stimulate it).
The something I said was:
Now suppose, all in the sudden, the economy wakes up and discovers homes are not so cool anymore.
I should have been more specific. Let's say the 'answer' to the recalculation is healthcare. So right now the home workers are useless but the healthcare workers are. As a result home based income falls but healthcare rises. But, and I suppose the recalculation idea would be that it takes time for the economy to discover that. What does that mean?
Well in an economy with instant recalculation the demand for health would go up as soon as the demand for homes goes down. Health workers would see wage inflation while home workers would see unemployment. Even assuming it takes a few years to train for healthcare jobs, in theory at least, the training could be financed by loans and such from an economy seeking more healthcare.
Kling's recalculation, though, allows the economy to instantly 'discover' it put too much into homes but takes time to figure out where else it goes. The 'instant' variation is just the old Say's Law 'recessionas/depressions/general gluts' are impossible school. Demand destroyed in one area must be instantly created somewhere else.
Kling, though, would allow demand to disappear from the economy for a while until everything is figured out. During that time the home workers are in a horrible position. Not only are they unemployed but they cannot even consider training for a new career because the economy doesn't know what their new place will be yet.
To me this all sounds like a restatement of Keynesism. Instead of 'animal spirits' and fear driving the economy into recession its a momentary mass 'brain fart'. I think the idea of fear is more accurate here. What happened in 2008 wasn't a mass scratch of the head. In a state of simple uncertainity people may go any which way with some buying while others sell. In a state of panic people tend to move together and in one direction, to the exit.
But here's the problem. Whether you want your theory with Keynesian or Klingian flavor, they both seem to lead to the same place. If demand has disappeared there's no reason why stimulus couldn't and shouldn't be used to replace it. It's perfectly fair to say the gov't shouldn't be trying to save particular companies or industries like AIG or the US auto industry...but what you take with one hand you have to give with the other. The policy place this leads to is fewer AIG's and GM bailouts and more stimulus package whose benefits are diffuse and diverse rather than concentrated in one particular place which may be on the bad end of the 'recalculation'.
Posted by: Boonton at Oct 4, 2009 8:37:52 PM
I should have been more specific. Let's say the 'answer' to the recalculation is healthcare.
That's fine but we can't know this until the 'recalculation' actually occurs. If we prescribe it, through government, we might be wrong, and that will be bad. If you include it in your hypothetical, you're begging the question.
If demand has disappeared there's no reason why stimulus couldn't and shouldn't be used to replace it.
'Stimulus' may answer the recalculation wrongly, and redirect resources to wasteful/bad/unwanted places.
But, Kling says he's in favor of stimulus, so I'm speaking for myself here I guess.
The policy place this leads to is fewer AIG's and GM bailouts and more stimulus package whose benefits are diffuse and diverse rather than concentrated in one particular place which may be on the bad end of the 'recalculation'.
I agree this would be the least-bad sort of 'stimulus', I suppose.
Posted by: Sonic Charmer at Oct 4, 2009 9:25:14 PM
Well, it seems no one else has yet offered the obligatory Austrian complaint: How come Tyler comes out against Austrian business cycle theory as being relevant to the current crisis, and yet now defends Kling who took ABCT and relabeled it "recalculation"? (I'm not knocking Kling; he acknowledged early on the affinity between his views and ABCT.)
Posted by: Bob Murphy at Oct 4, 2009 9:27:51 PM
That's fine but we can't know this until the 'recalculation' actually occurs. If we prescribe it, through government, we might be wrong, and that will be bad. If you include it in your hypothetical, you're begging the question.
I assumed that was the answer for the hypothetical but didn't assume the gov't or anyone else knows that. In fact, I pointed this out when I sketched out the differences between an economy that 'knows' the answer instantly versus Kling's recalculation economy that requires time to learn the answer.
'Stimulus' may answer the recalculation wrongly, and redirect resources to wasteful/bad/unwanted places.
Here's the problem with stimulus bashing, until the economy figures out what it wants to do you are going to have unutilized resources, including millions of people out of work. Stimulus closes the output gap that is itself inherently destructive to both labor and capital (see, for example, studies on what long term unemployment does to people).
This anti-stimulus argument would make sense in an economy where time was not required for 'recalculation'. There the new answer is known instantly and any effort to stimulate anything besides the next NEW THING(tm) is automatically wasteful.
But Kling's model requires time which means wasted resources in the meantime. In fact, by spurring different types of economic activity, stimulus might even allow the economy to stumble upon the 'answer' faster.
I'm seeing a bigger problem with recalculation. Why does the economy suddenly decide a major piece of itself was deeply misguided but it is uncertain about a new direction? In other words, demand disappears from housing but doesn't appear anywhere else for quite a while. Why is the economy so unsure of everything except the fact that one factor, say housing, was wrong? In rational conditions of uncertainity speculators would be placing their bets among many different industries seeking out the next big thing. Why wouldn't these losing bets serve to stimuluate the economy naturally to full employment until the new thing is found and all the money flows to it?
Why, again, should demand simply disappear?
Posted by: Boonton at Oct 5, 2009 9:01:53 AM
You think 'stimulus' is good to make use of unutilized resources while the recalculation takes place. I think it's bad because it can hinder, delay, or even prevent the recalculation from successfully going through.
You find it unnatural that 'recalculation' requires more time than a bubble-burst. Actually, seems to me this is the most natural thing in the world and it's odd to think otherwise. What takes longer, building a building or demolishing a building?
You ask why 'demand' should disappear. Actually I don't think it did. It just became redirected.
Posted by: Sonic Charmer at Oct 5, 2009 9:07:22 PM
If demand was simply redirected then losses in one area would become gains in another. That would follow the 'instant' model I presented, one without a time lag. So home based jobs and income would disappear but healthcare jobs would boom. If unemployed home workers weren't able to take healthcare jobs, then the income of existing healthcare workers would go up. It would become economical for home workers to be retrained as healthcare workers...even if it means that companies pay home workers to go to medical and nursing school or the financial industry loans money to unemployed workers to be retrained.
The recalculation argument is based on demand in one area disappearing. OK no problem. But then that demand stays gone for a while until the economy can figure out where it should go. Hence income of the home workers drops but the healthcare workers see no increase. We all sit in misery until the economy 'decides'.
What I think this model doesn't answer is why? If the economy is uncertain where the boom will come next wouldn't investors try numerous possible industries? Instead of the instant model we'd have a probing model. A few jobs in healthcare, a few in cell phones, a few in video games, social networking sites, electric cars and so on. Then when the economy realizes the real money is in healthcare those probes fail and the money moves to healthcare driving change.
This model, IMO, might have been more applicable after the dot com bust. The economy knew the new tech was great but also knew it was overdone so it looked for other things to take its place. Housing, finance, healthcare and other things were it. A recalculation based recession isn't such a bad thing.
But this recession didn't feel like that. There wasn't 'probes' to see where the next big thing would be but nothing. This was panic and that is the model Keynes developed and worked with. 'Recalculation' to me sounds like doublespeak some Soviet spinagent would come up with to avoid the true nature of event (sorta like "No it wasn't a nuclear meltdown, recalculations demonstrated that the reactor's core had dispersed over a larger geographic footprint than original design intentions...")
Posted by: Boonton at Oct 6, 2009 10:24:02 AM
The recalculation argument implies a rejection of the New Classical Economics, in both their monetary and Real Business Cycle form. According to the New Classical Economics, the economy works as modeled by the theory of perfect competition, so that macroeconomic models must use the theory of perfect competion as its microfoundations and be consistent with it. But according to the theory of perfect competition, if there are excess workers in a market, the real wage in that market will fall and some of the exess workers will voluntarily move to other markets in which real wages are higher, all of this without any of the structural unemployment that recalculation posits developing.
Posted by: Roland Buck at Oct 6, 2009 3:12:01 PM
If demand was simply redirected then losses in one area would become gains in another.
Quite right. And this was seen. Gains in government debt, for example.
What I think this model doesn't answer is why? If the economy is uncertain where the boom will come next wouldn't investors try numerous possible industries?
It's not my model so I can't speak for it. But if investors had shifted their demand to safe money they might be (and of course were, in the event) less willing to 'try' a bunch of stuff just for the heck of it. The recalculation model seems to posit investors shifting demand away from 'returns' and toward 'safety'. This is consistent with both a 'sitting on the sidelines' story and the fact that there wasn't necessarily a lot of 'probing', at least not until price signals give some indication of good returns to be had somewhere or other.
Posted by: Sonic Charmer at Oct 6, 2009 8:30:38 PM
Basically this is just Keynesian economics. The 'rush to safety' is the same thing as a lack of 'animal spirits' or simply panic. The attempt by everyone at the same time to increase their liquidity (ultra safe investments or cash) puts the economy in a liquidity trap.
Posted by: Boonton at Oct 7, 2009 8:58:02 AM
"More fundamentally, there is strong reason to believe that a recovery in the real economy is salutary to the financial sector. When people are employed and buying things, loan defaults fall and asset prices are likely to rise. Both of these developments would surely be helpful to stressed financial institutions. This is, I believe, a key lesson of the Great Depression."
Posted by: buy r4 ds at Jan 29, 2010 12:25:59 AM
Ok this is a bit over my head. What is it all about? I was browsing for water coolers in Google when I landed here. Weird...
Posted by: counter bar stools at Mar 16, 2010 10:25:26 AM
