Arnold Kling  

A Theory of Economic Recovery

Charlton Controversy, Continue... The Lower Right Quadrant...

Reading Paul Seabright has made me want to once again dismiss the theory of aggregate demand and instead push Recalculation.

There are two patterns that must emerge in a complex modern economy. One is a pattern of labor specialization. People need to develop general human capital (broadly-applicable skills) and specific human capital (on-the-job know-how), and this human capital has to be useful, as determined by the market.

The other pattern is financial trust. We want to have safe, liquid savings in a world where entrepreneurs undertake risky, illiquid investments. Financial intermediaries can issue safe liabilities while holding risky assets, provided (a) that they are operating wisely and (b) that we trust them.

In normal times, these patterns change gradually. People shift their assets, but slowly. The mix of labor demand shifts, but not too quickly for people to adjust.

Both the Great Depression and our current Great Recession have involved disruptive shifts in these patterns. Currently, the pattern of financial trust that built up in the asset-backed securities market has been shattered. The pattern of labor specialization has been sharply disrupted also, in ways that are both obvious and subtle. The obvious disruption is the shrinkage in industries linked to housing construction, sales, and finance. But the subtle disruption is that organizations are pressed by the recession into accelerating some of the pruning that they would have undertaken anyway.

I am slowly reading Paul Kennedy's Freedom from Fear, about the 1930's. One point he makes is that the agriculture sector was probably the weakest. My view is that the U.S. had not yet adapted the structure of farming to the internal combustion engine. Once we did so, we needed many fewer farmers, because farming could be concentrated in land that had the most comparative advantage for growing crops. Farms no longer needed to be adjacent to cities.

Similarly, I think that we have not yet adapted to a structure of production that is efficient given the Internet. Many patterns of specialization that existed as of 1995 are obsolete today, and the Great Recession represents a sudden collapse of many of those patterns.

When the economy recovers, new patterns of financial trust will have emerged. Also, new patterns of labor specialization will be evident.

The traditional macroeconomic villains are weak aggregate demand and sticky wages. They may be part of the problem today. However, I think they are at most just a small part.

[UPDATE: I think that it is very hard to separate empirically a Recalculation story from a sticky-wage and weak nominal demand story, but macroblog has a post by Menbere Shiferaw and John Robertson that I view as providing some support for the view that changes in the pattern of specialization (or restructuring) are playing a role in the current recession.

My view of the 2001 recession is that the recalculation effects were relatively large in the labor market. In my macro lectures (and in this essay written in 2003), I pointed out that using hours worked as the cyclical indicator, that recession was long and deep (what others called a "jobless recovery" I said we should call a productivity-cushioned recession).

On the other hand, although the stock market took a large hit in 2000-2001, there was much less institutional disruption in the financial sector. That certainly accounts for the less-panicky government response, and either the lack of financial disruption or the lack of panic helps to explain the fact that employment did not drop as badly back then.]

Comments and Sharing

CATEGORIES: Macroeconomics

COMMENTS (19 to date)
John Alcorn writes:

Arnold, You have a knack for identifying core mechanisms and placing them in broad historical perspective. A question, re: recalculation in the construction sector - What do you make of Casey Mulligan's findings that housing is still driven by a long-term increase in demand (smaller recalculation that people think)? (See his blogposts at Supply and Demand.)

jsalvati writes:

Consider that the burst of the tech bubble also involved a big change in the kinds of activities that were economic but did not precipitate a bad recession. That fact is difficult to fit into the Recalculation theory. I find Sumner's view much more persuasive.

AJ writes:

Wonderfully put.

Now combine this with the insight that highly leveraged derivatives and financial instruments were almost all priced heavily on the smaller disruptions occurring continuosly over time and not the "black swan" of the larger disruption. Viola, a financial collapse and crisis which gets far more attention than the real disruption and significant adjustment taking place.


AJ writes:

Research which supports the notion that there's much more to a recession than aggregate demand:

In the late 1980's, I did some never published research just before severing most ties to academia which showed that market share changes within markets/sectors tend overwhelmingly to occur in the recovery phase of recessions. Long term market winners recover better, while losers don't recover or do so much less. I interpret this to mean that recessions coalesce/accelerate structural changes in the economy where groundwork is already laid.


Ted writes:

This is an unfalsifiable theory so I'm not sure why we should care. But a few questions first:

(i) What about all the bubbles that didn't result in any massive recalculation? Were these just "different"?

(ii) Really, the Great Depression? You are telling me you believe that with more aggressive monetary policy and without NIRA we wouldn't have recovered due to "recalculation"? You are telling me if the Fed pushed up expectations to increase demand even more we would still be in the GD due to recalculation? That seems absurd.

(iii) Your theory isn't any different that a super-charged labor market rigidities thesis. To generate search and matching frictions of that severity you would require implausibly high rigidities. How do you square that with estimates from the data? I suppose it seems plausible that rigidities can be asymmetric across the business cycle - but they shouldn't be THAT asymmetrical.

(iv) What do you mean that you've seen a shrinkage in industries like "sales"? Sales is so generic I have to take that to mean anything related to aggregate demand. And yes, aggregate demand is still weak. You are telling me if the Federal Reserve began to push up AD that sales wouldn't rebound - that we would be caught in a recalculation trap? I don't buy that.

(v) To some extent human capital limits inter-industry labor market mobility. But, let's be frank, a finance background can easily be transmitted to other areas of business. Construction workers can relatively easily move into other fields, especially if "sales" are pushed up and employers need workers. Furthermore, at a minimum, the Federal Reserve by increasing growth could pump up the commercial real estate market.

You're theory seems to require implausibly high labor market rigidities when a straight-forward demand driven recession does much better to explain this. Why should I buy a theory that requires implausibly high estimates of rigidities, one that doesn't seem particularly consistent with history, and one that could be told with a much simpler and reasonable theory?

AJ writes:

jsalvati writes:
Consider that the burst of the tech bubble also involved a big change in the kinds of activities that were economic but did not precipitate a bad recession. That fact is difficult to fit into the Recalculation theory...

Good observation, jsalvati. The burst of the tech bubble was a huge disruption and huge recession to telecom, internet related businesses, and all suppliers of labor, services, and products to them. There was no overall or "average" recession, because the Fed dramatically pumped up money and thereby pumped up the housing and housing/related and interest sensitive sectors even more. Very few recessions are recessions for every sector/industry. In fact, I know of only two in last 100 years that hit every sector (Great Depression and our own Great Recession).


jb writes:

The Internet killed off a fair amount of demand for commercial construction (retail stores, etc), but not home construction. But the financial sector has a much lower risk tolerance, which makes acquiring the finance to build housing tracts difficult.

A lot of people are out of jobs, and moving in with friends and relatives because they can't afford houses, so that helped create the glut (as well as unrealistic expectations on the % of people who could find homes)

In addition, some people with good jobs and decent credit still can't buy homes, because of the risk-averseness in mortgage underwriters.

So the glut in homes will continue for a while. Some people need to save up more and build up bigger downpayments, some need to retrain/find work in new sectors (possibly lower paying ones), and then save up lots of money for downpayments.

And until all of that unwinds, I doubt we'll see a lot of reduction in the housing glut, or the commercial building glut. Those two sectors are going to be hurting for a while.

Chris Koresko writes:

This discussion looks fascinating and I'm trying to follow the arguments, but am having a hard time of it. If you don't mind clarifying a few things for the non-specialist I'd be grateful.

To the extent that I understand the Recalculation story it seems to make some sense, at least for normal recessions. I don't see the basis of Ted's statement that Arnold's picture is unfalsifiable, especially in light of AJ's claim to have research results which seem to support it. Ted, would you mind elaborating?

Is the duration of a Recalculation closely related to the time it takes for a typical worker to relocate and/or re-train for a new job? For new companies to be started and grow to significant size? Or some other, measurable phenomenon?

My picture of the Great Depression is strongly influenced by the Higgs 1997 paper on regime uncertainty. That paper argues that the Depression probably ended some time during WWII. But it seems to me that the war must have involved at least two Recalculations, one at the beginning when the US was arming itself, and one at the end when the process was reversed. So if Recalculation were a big factor, wouldn't the recovery have been delayed until after the war ended?

I have heard that one of the most fundamental facts that any macroeconomic theory must account for is the steadiness of long-term economic growth: brief disruptions like recessions don't seem to have any impact over the long haul. What do you think of a picture in which long-term productivity is determined almost entirely by technology, which progresses steadily, but short-term output is punctuated by periodic Recalculations which are needed to take full advantage of new technology?

Thanks for any clarification and/or insight.

Justin Dailey writes:


I think what Arnold is saying that the severe downturn is the trigger which accelerates the pace of change. In normal circumstances, things proceed at a more gradual and adaptable pace because the pressure isn't as high.

As a hypothetical example, consider a firm in which many departments could restructure, becoming more efficient but only slightly so. You are running one of those departments, and you have 48 workers under you. You could reorganize and eliminate Bob and Shelia - who are hard workers and well liked, but just aren't all that necessary to the success of the firm. The savings aren't large, and the economy is strong. You might be tempted to leave them be, if everything is fine, and you care about their wellbeing and the other 46 employees might be very upset about the reorg. Let's face it - who wants to shrink their empire AND look like a bad guy? The small cost savings probably won't get you promoted. Perhaps you are content as is, and since people change jobs a lot you prefer to not replace them as they move on, and otherwise reorganize in small steps rather than rapidly and disruptively.

But now if sales are suddenly down 10%, and your boss is screaming for you to cut costs or else the company folds, the reorganization takes place immediately - in fact, business is down enough that you're down to 40 employees. This is the severe recession that Sumner would have mitigated through stabilizing NGDP expectations. Over time, even as business recovers, necessity has forced you to figure out how to operate with a smaller crew, and things are working pretty well. Your organization can even handle some recovery in workload, either via better processes or somewhat longer work hours. Now, you don't need to rehire. Your process is streamlined, and your investment in organizational capital has paid off. Bob, Shelia and the rest have been unemployed for two years - their skills don't match the market's needs. You'll never need to rehire them. It is going to take some time to figure out what their next job will be.

Had the Sumner Fed jumped and stabilized demand in 2008, you would have seen Bob leave in 2011 and Shelia leave in 2013 of their own accords - to new jobs, retirement, whatever - and perhaps by 2015 you would have had the restructuring that we see now in 2010. But it would have been managable, not disruptive change. Unfortunately, since Bush didn't appoint Scott Sumner as Fed Chair, NGDP crashed, the efficiency changes have been made, and just boosting base money isn't going to give Bob and Shelia their old jobs back. You would get some inflation which would help reduce labor market rigidities and increase employment in that sense, but my view is that the proportion of unemployment related to recalculation has been growing steadily since early 2009, and a large proportion is now of that type.

Obviously lots of things are going on in the economy, but I think recalc is a dominant theme for 2010.

By the way, the last bubble seemed to have a recalculation element. Unemployment steadily grew from 3.8%-6.3% over 2000-2003. Part of that recalculation is that demand growth for manufactured goods fell, even as manufacturing productivity remained very high. Although I can't prove this with data, it seems to me that cheap credit eventually directed many of the unemployed manufacturing workers into construction and directed many others into finance/real estate jobs. As these sectors grew, other areas which had to get lean in 2001-2003 could begin to add employees, which drove unemployment down through 2007.

As Scott Sumner notes, there was already recalculation occurring from 2007-2008, as unemployment rose from 4.6% to 5.5% or so by mid year. Lots of people in finance/real estate/construction needed to find a new line of work. But the recalcuation was confined to the hurting industries. Without the demand shock, unemployment would have probably peaked between 6%-7%, stayed stuck for awhile until the bankers and construction workers found something new to do (or for home construction to gradually recover from an overhang and bank balance sheets to be repaired). Once there was a demand shock, the recalculation spread economy-wide, and now it is largely calcified. A sudden jump in NGDP will be more inflationary over the coming year than it would have been over 2009, and with less real effect.

So in sum, I think you're correct that demand explains the initial severity of the downturn, but the severe downturn in turn sharply accelerated existing organizational changes, allowing firms to maintain output with less labor input (evident in the productivity statistics), and so simple demand growth isn't going to bring those workers back to their old jobs. Many sectors need fewer workers and perhaps healthcare needs more - but you can't just go from, say, accounting to healthcare. You need lots of human capital development to make the switch and that takes time.

Foobarista writes:

jsalvati, my guess is we're having the recession now that we should have had in 2001-2002 (sure, we had a bit of one, but not a big one), but didn't because the Fed massively goosed the economy. The recalculation started then and is continuing, but this time, the "throw cash at the problem" solution has played itself out and is off the table.

jimdew writes:

I don't understand why the "aggregate demand" theory and the Recalculation theory are mutually exclusive. Can't both be true? Of course, if both are true, this leaves policy makers with a problem since we know little about the proportional contributions of these and other factors. We also know little about their relationship and interaction with each other. Perhaps an economy is too intricate to be neatly described by some tidy theory.

Adam writes:

Yes, decentralized adjustment and recalculation are big parts of the problem. In the 30s, electricity and gasoline engines were replacing physical labor, on-farm and off. But the new economy had no place to put the unskilled and independent farm and city labor. Result: huge unemployment and the CCC. The CCC did seem successful in keeping the unemployed on rural, isolated projects at less than minimum wage.

I've often thought that WWII was critical to the recalculation after the 30s--not for a bogus Keynesian stimulus but for the practical and technical education it provided to the 16 million who served. Armed forces training took independent, generalist farm boys and drilled them to understand both teamwork and specialization. At a minimum, the team approach suited veterans for assembly lines that sprouted up throughout the country after WWII. Others who knew only buggy whips and shovels before the war, got trained in the skills of new trades such as mechanics, heavy machinery operators, electricians, and radio specialists. These new skills were ideally suited to the blossoming of the new industrial-electronic-gasoline economy after WWII.

The challenge now is getting retrained to maximum value in our new internet world.

Chris T writes:

This is close to my own view that the 'Great Recession' is largely explainable by changes in technology (particularly computers and information technology).

I do not understand why some people insist that there is only ever one cause to recessions. Why can't individual recessions have different causes?

Chris Koresko writes:

Just a follow-on to my own ignorant musings:

There seems to be some support here for Recalculation as a chief cause of the current recession, and for technological change as the source of the need for the Recalculation.

Question: What do you guys think of the idea that the Recalculation is needed in response to the changing regulatory environment? After all, we have had a hike in the minimum wage, impending financial reform, cap-and-trade, a stimulus package, TARP, bailouts of home mortgage borrowers, the government purchase of GM, and strong-arm tactics being used liberally against private actors. Surely all of that must cause businesses to have to rethink a lot of the assumptions they had held before.

Essentially, I'm suggesting here that our current Recalculation is driven by Higgsian regime uncertainty. If so, it might be possible to distinguish it from a "normal" tech-driven Recalculation because it'll keep going as long as the regime uncertainty does, rather than ending in 12-18 months or whatever.

Does this make any sense?

MernaMoose writes:

Chris K.,

I was going to throw that in but you beat me to it. I'm quite certain it's a major contributor to the overall recalculation that's going on. We're dealing with all kinds of new regulations in the private sector, all the time.

Technology drives changes in the labor market but so do government regulations. It's not always easy to tell which is dominant in a particular company, let alone the overall economy.

But personally I think the government regulation question is going to drive this down turn to last much longer than it has to.

Nobody knows what's coming out of Washington next. We don't for example know how ObamaCare is really going to work or how much it's going to cost. Corporations therefore don't know what their margins have to be in order to survive, let alone thrive. Hence they don't know how to identify the core product lines they should be investing in.

Therefore investment is slow to stagnant, because clear "this is the future" stories are hard to come by.

Meanwhile, the "off-shoring of US jobs" problem" is tied directly into this whole story for obvious reasons.

I was working in Telecom in the late '90's (and got out one step ahead of the big crash). Telecommunications in the US didn't go bust, we just off-shored the work, largely to China and Japan. In some part it was because the technology had matured and was ready to go the commodity route. But it also happened sooner than it would have because, the cost of doing business in the US grows higher each year. Because government agencies impose new regulations each year and seem to believe the sky's the limit.

I wonder, if we are in fact reaching the limit that the growing load of regulations must inevitably lead to. I take western Europe as a pretty good indicator of where we're probably headed and how bad it's probably going to get.

Foobarista writes:

Chris K, there's definitely some - or maybe a whole lot - of that going on. My wife sells small businesses, and this year has been downright dreadful. She has lots of sellers, and lots of "looking" buyers, but precious few who will actually pony up and buy businesses. Regulatory uncertainty of various sorts are topics of frequent discussion with her buyers. The lack of credit is a problem, but not as much as a problem as people simply don't know how much it'll cost to run their businesses or how much taxes they'll pay, so they don't know how to value the business revenue.

The only businesses that are selling reliably are cash-focused businesses with lots of under-the-table labor.

Paul Seabright writes:

What I like about Recalculation is that it tells a story about behavior in abnormal times that is plausibly continuous with behavior in normal times, while being compatible with mechanisms that explain why such behavior produces qualitatively different outcomes in abnormal times. The story is that everyone has to specialize in some way, and we are never sure whether we have got it right: the signals the markets sends us are inherently ambiguous. A demand shock could be a signal about people's preferences for my particular skills or a signal about a temporary shift in demand for a broad class of services, or a combination of both. Imagine I set up a construction business; it's not a commodity service because in lots of ways - where I work, the type of work I do, my quirks and charms as an artisan, how reliable I am for my customers - I am subtly differentiated from my competitors. I get a rent for that, and that rent, even if it's not large, makes me rationally reluctant to abandon my work just when demand is weak; I prefer to hang in there in the (usually justified) expectation that conditions will improve. But in normal times, if demand stays weak I will soon conclude that this isn't the line of work for me. The process is painful, maybe even tragic for me and my family, but it leaves few ripples on the national pond. In abnormal times I can't conclude any such thing: demand may be temporarily down and I might be foolish to abandon this line of work. Indeed, my response is non-monotonic in the demand shock: up to a certain point the larger the fall in demand the more likely I am to quit, but after that point a really large shock convinces me that the market is not, after all, telling me personally that I made the wrong specialization choice.

Now the difficult question: why don't I nevertheless accept in the meantime the current price that clears the market for my services? Two parts to the answer: first, my opportunity cost of time has gone up because by accepting to work for the lower price I reduce my ability to search, and the marginal value of search has increased because of the greater difficulty in inferring the long-run value of my specialization decision from current conditions. Secondly, there is a signaling disincentive: by accepting the current market-clearing price for my services I thereby signal something very unfavorable about the quality of those services: that I am close to bankruptcy, perhaps, that I am desperate and maybe likely to cut corners, that no-one else values my services; none of which would make me attractive to one of the rare buyers of my services who is still out there. So instead I hold on, cutting back in the meantime on the services of my baby-sitter, gardener and pizza deliverer (who now have decisions of their own to make just like the ones I have recently been through).

Key to this is that there is no such thing as a market clearing wage in general (my services are subtly different from other peoples). It might be rational for me to accept the true market-clearing wage for my own services if I or anyone else could figure out what it is, but that's not what the market will offer me. I'll get instead an offer of remuneration that is based on a set of confused signals about the match between my services and the long-run preferences of the buyers of my services. And I will have the incentive to engage in lots of individually rational but socially inefficient signaling behavior that involves staying in my chosen line of work for what could be much longer than I would do in normal times.

topcat writes:

The author of "Freedom from Fear" is David Kennedy, a professor at Stanford, not Paul Kennedy, a professor at Yale.

John Schroy writes:

" ... When the economy recovers, new patterns of financial trust will have emerged. ..."

I think that this is the key point to be considered when discussing the timing of economic recovery from the current crisis.

However, new patterns of 'financial trust' must be based on actual institutional reforms as well as sociological changes in the business community, which take time.

I believe that the current crisis is a 'game changer', like the 1930s or the Johnson-Nixon-Ford-Carter years. It will take time to push the 'reset button'.

For a discussion of crises that are 'game-changers', see: Economic recovery may wait until 2016.

Comments for this entry have been closed
Return to top