ZMP v. Sticky Wages

I find myself in the unusual position of being closer to Paul Krugman (and Scott Sumner, less surprising) than Tyler on the question of Zero Marginal Product workers.

The ZMP hypothesis is too close to a rejection of comparative advantage for my tastes. The term ZMP also suggests that the problem is the productivity of the unemployed when the actual problem is with the economy more generally (a version of the fundamental attribution error).

To see the latter point note that even within the categories of workers with the highest unemployment rates (say males without a high school degree) usually a large majority of these workers are employed. Within the same category are the unemployed so different from the employed? I don't think so. One reason employed workers are still fearful is that they see the unemployed and think, "there but for the grace of God, go I." The employed are right to be fearful, being unemployed today has less to do with personal characteristics than a bad economy and bad luck (including the luck of being in a declining sector, I do not reject structural unemployment).

To see the importance of sticky wages consider the following thought experiment: Imagine randomly switching an unemployed worker for a measurably similar employed worker but at say a 15% lower wage. Holding morale and other such factors constant, do you think that employers would refuse such a switch? Tyler says yes. I say no. If wages were less sticky the unemployed would be find employment

By the way, the problem of sticky wages is often misunderstood. The big problem is not that the wages of unemployed workers are sticky, the big problem is that the wages of employed workers are sticky. This is why stories of the unemployed being reemployed at far lower wages are entirely compatible with the macroeconomics of sticky wages.

Although I don't like the term ZMP workers, I do think Tyler is pointing to a very important issue: firms used to engage in labor hoarding during a recession and now firms are labor disgorging. As a result, labor productivity has changed from being mildly pro-cyclical to counter-cyclical. Why? I can think of four reasons. 1) The recession is structural, as Tyler has argued. If firms don't expect to ever hire workers back then they will fire them now. 2) Firms expect the recession to be long – this is consistent with a Scott Sumner AD view among others. 3) In a balance-sheet recession firms are desperate to reduce debt and they can't borrow to labor hoard. 4) Labor markets have become more competitive. Firms used to be monopsonists and so they would hold on to workers longer since W<MRP. Now that cushion is gone and firms fire more readily. What other predictions would this model make?

It would be interesting to know why Paul Krugman thinks productivity has become counter-cyclical but I believe he has yet to address this important topic.

Addendum: Paul Krugman gives his answer and The Economist offers a review with many links.


On your 1-4, #3 (which have you mislabeled as another #4) doesn't seem to fit the data, since firms have hoards of cash. It's also a way of saying "there is ZMP using the firm's now-higher discount rate." #1 and #2 may be true, but they don't explain why productivity is shifting in a new way over the course of the cycle. #4 explains employment behavior, not productivity data. Of the four, I don't see them as alternative contenders to explain the productivity fact. They are geared toward explaining the employment facts only.

Keep in mind it's not ZMP *vs.* sticky wages but rather which new assumption should be added to the AD/sticky wages framework. I still don't see which new assumption -- which would explain the *productivity* data, namely the increasing countercyclical nature of productivity for a given level of unemployment -- you are proposing in lieu of ZMP.

Tyler, why do you refer to your hypothesis as zero marginal product? The basic competitive econ says wages approximate (are less than) marginal revenue product = marginal physical product * revenue the employer receives from the sale of the additional physical product. The failure to hire today says nothing about MP alone. I understand your back of the envelope calculations that the MP of the laid-off must have been about zero, since production has rebounded but employment has not (and I understand the objections to that calculation that Sumner raises). But it seems that you abandon your calculation (that the ZMP workers could literally produce zero today) with your comments that their MP could be positive if the macro economy improved. That makes sense if you are talking about MRP, in which case these workers have positive MP today, but the employers cannot sell that incremental output at prices high enough to justify additional employment. And if the economy improves, the hiring calculation changes. But if the employee literally has a zero marginal product, why would anyone hire him in the better economy of the future? And if the marginal product is not an attribute of each potential employee, but rather is a conditional attribute based on the state of the macro economy, then what does ZMP really add to our set of useful concepts? We already have the MRP concept that covers that ground.

I agree with Alex. I think that it is evidence that companies that use open books management have laid off fewer employees.

I think that another issue is that society changes slowly and so absorbs excess labor slower that it could. An example is that many people mow their own lawns for cultural reasons even though the price to having some else do has dropped to were it would be utility gain by paying someone to do it.

Here's why I don't buy the "business used to hoard ZMP workers but now have stopped" argument. From the unemployment rates for various professions posted from the WSJ yesterday:

% unemployment rates in 2007, 2010

waiters/waitresses 7.2 11.7
cooks 7.8 13.5
maids, household workers 6.7 12.0
food prep workers 8.6 13.6
bartenders 7.0 11.8
taxi drivers,chauffeurs 3.5 10.7
chefs/head cooks 4.4 7.7
dental assistants 2.0 5.7

Anyone who has worked in a cubicle farm knows that there are some people just hiding out, not doing much work. But anyone who has worked in a restaurant knows that it is almost impossible to get away with not constantly doing something -- and in a restaurant, unlike a cubicle farm -- "constantly doing something" is going to be productive above your rate of pay -- unless there are no customers in which case the problem is the business. Restaurants don't hoard workers. They watch their margins constantly and tell workers, who are usually hourly, to go home when business is slow. If they didn't, they would go out of business. So the only way there are much fewer cooks, waiters, chefs, etc. is if there are fewer restaurants.

Same with taxi-drivers, chauffeurs. Who was hoarding ZMP taxi drivers? Chauffeurs are a luxury item. I can imagine people are spending less on luxury because the economy still sucks, but I don't think the Chauffeur firms were hoarding drivers or suddenly got rid of the ones who didn't know their way around town.

Dental assistants. What is a ZMP dental assistant doing with their day? Getting it on with the dentist? OK, maybe.

"So the only way there are much fewer cooks, waiters, chefs, etc. is if there are fewer restaurants." - or if restaurants have less business.

The change in firms' behavior to a "just-in-time inventory approach to human resources" probably does explain "the increasing countercyclical nature of productivity for a given level of unemployment", but it doesn't show that the result of that approach is shedding ZMP workers, only that firms are more cautious about capital and more dubious about the value of having long-term employees, justifiably or not.

The AMP hypithesis is amenable to empirical testing: Just reemploy a statistically valid sample of the unemployed and measure their output.

in the real world....
many firms are small, the owner knows and hired people; many owners don't like to fire people - it is very unpleasant (many readers of this blog must have had to fire a grad student at one point, which is much less painfull....)
you can do more with less; if you know about rapid protoyping in the industrial design process, CAD, low cost goods from china, you really can do more with less
lower quality is acceptable; look at the quality of websites, in terms of grammar and spelling, vs 'traditional" print literature; even in printed materials, spell and grammar checks are lacking in material from reasoanbly sized firms, much less startups like the one I work at
true story: "guage blocks" are precision metal blocks used by machinists when you need to know that 1" is really 1 inch, exactly.
Guy says to his boss, need a set of guage blocks.
Ok, get some quotes
Guy comes back: well known US firm Starrett, 2 grand; stuff on the web from China, 200 bucks.
As economists, I think perhaps you can realize that it is difficult to justify the extra 1800 dollars, which is good high paying jobs in MA
I will never take seriously the economist view that free trade is good until free trade is not something that helps economists (as relatively well paid people, they get cheap goods at wal mart and cheap labor for nannies, gardens, resturants, etc) but hurts economists (dragon mom reared chinese economists with phd's working for 25 K a year; that starts to happen, you will be surprised at how many ivy league and UC school economists find deep theoretical reasons why free trade is bad....

Sorry Alex, Barack Obama has come out in favor of ZMP.

And I spoke too soon about "counter-cyclical regulation" and crickets.

"As noted, at $1 per hour there would be no unemployed. Why isn't this obvious policy?"


Because the cost of health care in the USA is $5 per person per hour, 3 times as much as Europe. This makes America an uneconomic place to employ people for most manufacturing jobs.

And why is this:

1} Insurance company overheads are 25% by their own admissiom.

2) Doctors' offices and hospitals add another 25% overhead of their own to deal with the insurance companies

3) Malpractice insurance and litigation adds another 10%

4) The cost of 1-4 triples the cost of health care for health care workers too, adding another 10% to the bill.

It does not matter who pays for it all; the basic problem is there are several million too many people employed in health care who don't treat patients.

A National Health System would simply put all doctors and nurses on the payroll. There would be no insirance companies, no billing, nd with no billing thee can be no fraud. The cost reduces to 100-25-25-10-10 = 30% of what it is now, for the same care.

Until Republicans embrace this idea, the American economy is doomed.

Comments for this post are closed