I am trying to sort out my thinking on unemployment in the Recalculation Story. I think that a basic question is this: when workers lose jobs because a sector needs to shrink, this creates a pool of unemployed workers. Why don’t firms fish in that pool? More thoughts below.Engaging in market work means that you exploit comparative advantage. You do something you are good at. This gives you income, which allows you to buy goods and services that other people can produce more cheaply for you than you can produce yourself.

Being unemployed means that you do not engage in market work. Why not? Because you cannot find a firm that will pay you enough to make you believe that you are exploiting your comparative advantage.

In the spirit of the latest Nobel Prize that went to the economists known by the initials DMP, we can think of this as a matching problem. Presumably, there is some firm that will offer you enough pay to get you to leave the couch and go to work. But you do not take just any job, because you are waiting/hoping for a better match.

This DMP model is given a workout by Marc Bils, Yongsung Chang, and Sun-Bin Kim (older, ungated version here).

In the DMP model, there is something called “match rent,” which is the profit a worker gets from being in a good match. What the Bils paper says is that it is hard to characterize match rents in a way that fits observed behavior.

If workers are being laid off in large numbers rather taking pay cuts, then the DMP model presumes that those workers had low match rents to begin with. Otherwise, they would have taken a pay cut to stay on the job. The fact that we observe layoffs rather than wage fluctuations suggests that match rents are low.

However, if match rents are low, then we should expect a relatively large amount of voluntary separations. Workers should often be on the brink of quitting to take a different job. We do not seem to observe this. (Or do we? The JOLTS data shows a lot of turnover. Also, there is a lot of churn that takes place within a particular organization–workers respond to internal postings, accept promotions, get reorganized, and so on.)

So Bils says that the DMP model has a problem, because it wants match rents to be low to explain fluctuations in employment, but it wants match rents to be high to explain a relatively low rate of quits relative to layoffs.

I suggest a different solution, which is that match rents are high but they can change drastically. My thinking is based on the Recalculation Story with Garett Jones workers. If workers are producing organizational capital rather than output, then there can be large changes in the value that firms place on workers, so that formerly high match rents can suddenly plummet. I ran that idea by Bils in an email, and he responded that it fails to generate a Beveridge Curve. If I am correct that unemployment is caused by job separations due to plunging match rents in some sectors, that should not cause the rest of the economy to stop posting vacancies. If anything, more vacancies should appear as firms try to fish in this large pool of unemployed workers. But empirically, vacancies decline as unemployment goes up.

So we are back to the question with which I began: in a recession, why don’t firms fish in the pool of unemployed workers? Some possibilities:

1. Unemployed workers are “lemons.” They do not generate high match rents, and they are not heavily committed to working, so they are liable to quit soon after you hire them. Bils in his email suggested that he and his colleagues have worked on incorporating something like this into a model.

2. The “wedge” between worker compensation and take-home pay is high. The wedge consists of taxes and also health care benefits that cost more to the firm than their value to healthy young people.

3. Hiring workers is an expensive capital investment. It takes a long time for a worker to learn enough about the organization to be able to contribute in a positive way.

I think my view of the world is as follows:

There are some workers for which match rents are low, and as Bils pointed out in his email we do observe that layoffs are more prevalent among workers who in general tend to have volatile work experience. In other words, layoffs occur most where match rents are lowest. Call these the weakly-attached workers.

However, in our complex economy, there are many workers for whom match rents are very high. For small shocks, these workers tolerate pay cuts (which can come in subtle forms, such as longer waits for promotion, more onerous work schedules, less autonomy, etc.) in order to keep their jobs. Call these the strongly-attached workers.

The Recalculation Story is that sometimes shocks can be large enough to cause match rents to drop so far that even strongly-attached workers experience layoffs. The layoffs mean that the drop in match rents is permanent, and this represents a significant drop in wealth. This is the sense in which Tyler Cowen is correct that we are now poorer than we thought we were.

Finally, although match rents are high, the cost of investing in workers is also high. In the DMP model, this might be expressed as saying that the matching technology is expensive. The expensive matching technology explains why firms don’t fish more in the pool of unemployed workers.

Because the matching technology is expensive to operate, it takes a long time for some strongly-attached workers to get into good matches. Meanwhile, weakly-attached workers are affected by the wedge between compensation and take-home pay. This wedge has been rising over time. Moreover, there may be a tendency for the wedge to be higher in an economy where wealth has declined due to a fall in match rents.