Arnold Kling  

An Empirical Disagreement

PRINT
A Different Mr. Negative... Why Should We Restrict Immigra...

Bryan focuses our disagreement:


when 10% of the workers in an occupation lose their jobs, or 5% of firms in an industry go out of business, continuity isn't merely a convenient assumption. It's a hard fact.

So now it boils down to an empirical issue about how prevalent are discontinuities. That is a difficult issue to settle in a blog, and I certainly must allow for the possibility that I am wrong. But some examples of discontinuities that come to mind:

1. It seems that troubled firms, rather than laying off one worker at a time to lower costs at the margin, "clump" their layoffs into large batches.

2. Similarly, a company like Borders, rather than closing a few unprofitable stores, goes out of business completely.

3. When a company outsources customer support, it does not do so incrementally, by shifting cutting back its domestic customer call center by one or two workers and replacing them with a few people in India. Instead, the tendency is to move the entire call center.

I think that the better argument for continuity as a "hard empirical fact" would likely be that in spite of the fact that individual firms behave in discontinuous fashion, the aggregate market for labor acts as if it were continuous. I think that is likely to prove a better defense of continuity. However, it would not rescue Bryan's "wing-walking" hypothesis in which unemployment could be eliminated by having every worker remain in his or her existing job at a moderately lower wage.


Comments and Sharing





COMMENTS (13 to date)
Mark writes:

Not to discredit your argument, but I do believe that Borders closed a few unprofitable stores first, and then after months of failing to find a buyer, it shut down completely.

Daniel Milstein writes:

[Comment removed pending confirmation of email address and for policy violations. Email the webmaster@econlib.org to request restoring your comment privileges. A valid email address is required to post comments on EconLog and EconTalk.--Econlib Ed.]

Glen S. McGhee writes:

Continuity and discontinuity are constructs, not hard empirical facts (see especially Stephan Fuchs, 2001).

From one point of view (say, the population ecology perspective of Freeman and Hannan), the emphasis is on business' discontinuity, which masks underlying continuities.

Aggregated statistics on labor force participation rates run in the other direction. The bias is on taken-for-granted continuities, without much concern for micro-level discontinuities.

Ludwig Fleck in the 1930s showed how this worked for scientific knowledge, leading to the birth of the sociology of knowledge.

david writes:

There is "clumpy" churn all the time, though. It just averages out during moderate periods. The churn being net negative for a sustained period of time seems to be, like any aggregate, approximately continuous.

Thomas DeMeo writes:

What Bryan is suggesting would transfer a massive amount of business risk from organizations to their employees. The resulting changes in behavior would not be pretty.

Vinnie writes:

I think Bryan's version would be more realistic in fields where organizational capital (as I believe you call it) is produced. In my own field, for example, civil engineering, it seems more likely and would be more preferable for an organization to stay in tact during a lull, rather than laying people off. The sum of the parts is often more valuable than the individual contribution of any one member, and it would make sense to keep the team in tact, even if that means spreading the workload thinner and cutting pay. In practice, I don't think this happens very often, whether I think it should or not. If bosses are reluctant to cut pay even in this situation--which, in my experience, they are--I would guess that Arnold's version is more accurate. But this is all very anecdotal and may not be useful to the discussion.

Troy Camplin writes:

A series of very small dscontinuities may also look like continuity from a distance. However, catastrophe theory shows that there are many different paths from one state to another, in a continuum from continuous to larger and larger discontinuities. From the point of view of the economy as a whole, a few discontinuities here and there won't cause an economy-wide discontinuity -- but large clump of them will, with the effect of a disruptive, reorganizing recession. All of this is consistent with the view that the economy is a natural, self-organizing process, or spontaneous order.

GIVCO writes:

Re: #1 & #3. A company also aggregates absolute and offshore-driven layoffs because their employment attorneys tell them to. If you fire anyone but a straight white man under 40, you must be prepared to have a non-discriminatory reason that you can evidence, and broad economic-driven layoffs are the best evidence.

Glen S. McGhee writes:

Bravo, Troy!

"A series of very small dscontinuities may also look like continuity from a distance. ... catastrophe theory shows that there are many different paths from one state to another, in a continuum from continuous to larger and larger discontinuities. From the point of view of the economy as a whole, a few discontinuities here and there won't cause an economy-wide discontinuity -- but large clump of them will, with the effect of a disruptive, reorganizing recession. All of this is consistent with the view that the economy is a natural, self-organizing process, or spontaneous order."

The Luhmannian view in this last sentence (autopoietic economies) runs through Fuchs, Against Essentialism, and is dedicated to Luhmann.

All of this, of course, starts from the idea that these are social systems.

The Sheep Nazi writes:

I had thought the big discontinuity in a recession was that the rate of new hires collapses, not that the rate of layoffs increases (though it does.) In fact, I am surprised that Dr. Kling did not mention that specifically -- it seems like an obvious response to Caplan.

Chris Koresko writes:

AK paraphrasing BC: ...unemployment could be eliminated by having every worker remain in his or her existing job at a moderately lower wage.

Another issue with this approach is the assumption that the wage would have to be only moderately lower. In some processes wages are not the dominant cost of production, so when those become wealth-destroying, a very large wage reduction might be needed to restore profitability. In fact, the wage reduction could have to be larger than the wage itself.

One possibly silly example: In the past it made economic sense for men to herd cattle on horseback to distant markets. That's rarely done now, and when it is, the men are usually paying for the privilege, i.e., they are doing it as part of a vacation.

I'm not enough of a historian to know how abruptly the cowboys disappeared, but in principle it could have happened pretty fast, e.g., due to new rail lines getting built.

Dan Carroll writes:

It seems to me that #1 and #3 are examples in favor of sticky wages.

1. Firm's clump layoffs because firing one person doesn't correct an out-of-alignment cost structure, but firing many does. Further, cost structures don't appear out of alignment until the tide goes out in a demand-driven recession. Of course, it is always difficult to fire one person without cause, for legal and organizational reasons, which is probably the dominant reason for the clumping of layoffs.

2. Borders went out of business because of a high fixed cost structure, where a 10% reduction in revenues puts the enterprise at risk. Since the fixed cost structure is at the store level, closing a few stores doesn't solve a problem of a company-wide demand reduction. Wages weren't really an issue, though fixed costs are another way to define sticky prices. However, I think we can concede Borders as an example of PSST at work, as it would have gone out of business even without a recession (though it may have taken longer).

3. Outsourcing call centers to India are lumpy for similar reasons layoffs are lumpy. As are re-insourcing those call centers to the US after it becomes clear that the outsourcing didn't work as well as expected. This is a continual process that occurs in recessions as well as booms, and has not been particularly prevelent in the recent recession.

Bankruptcy is an inherently discontinuous process brought about by fixed principle debt contracts. Layoffs are discontinuous due to the nature of labor laws and cultural propensities. All of these add up to sticky prices and sticky wages, which reduce economic flexibility in the face of both nominal shocks and real (PSST) shocks. Therefore, I'm having trouble understanding what you are arguing about.

A=A writes:
1. It seems that troubled firms, rather than laying off one worker at a time to lower costs at the margin, "clump" their layoffs into large batches.

Anecdote/single data-point in support of this:

In the financial panic of Sept-Dec. 2008, I worked for a professional services firm. Low utilization led to the announcement that 1 week hence, 10% of client-facing positions would be cut.

Why an arbitrary number like 10%? Why not 6.35%? Perhaps because 10% is the standard corporate headcount reduction percentage when times are tough, and it was in our interest to signal to clients that we shared their pain in similar ways. The reason was psychological, not mathematical.

Within that week, I anonymously suggested to our C-level management that rather than reduce headcount, we instead temporarily reduce the salaries of employees, making the strategy's advantages and consequences very clear: greater morale via demonstrated commitment to our people (our #1 asset, we frequently were told!), layoff-by-attrition, shared-pain, retained capacity for the eventual bounce-back, etc.. That the idea was supported by a wide range of economists was unknown to me at the time.

The executives never announced the existence of my suggestion. I have no idea whether they ever considered it. I survived the cut, but some coworkers I liked were not so lucky.

Glenn's comment "Continuity and discontinuity are constructs, not hard empirical facts" is dead-on. Review the philosophy of Nassim Taleb and repeat after me: a model is not reality, but rather an incomplete, imperfect representation of reality - and this is why we have error rates in statistics (and as Taleb points-out, error rates themselves have error rates, recursively).

I agree, Arnold, that your best argument for a macroeconomic continuous function in spite of the discontinuous microeconomic data (like mine, above), is the same as that of quantum physics: discontinuities at the quantum level manifest themselves to such a fine degree at the classical level as to reasonably-assume the truth is continuous -- even though the whole-truth is in-fact (strangely) composed of discontinuous events.

Comments for this entry have been closed
Return to top