Bryan Caplan  

Keynesianism as a ZMP Story

Compared to What?... The "Amazon" Tax...
I'm not a fan of Tyler Cowen's view that many unemployed workers have Zero Marginal Product (ZMP).  Cowen and Lemke:
In essence, we have seen the rise of a large class of "zero marginal product workers," to coin a term. Their productivity may not be literally zero, but it is lower than the cost of training, employing, and insuring them. That is why labor is hurting but capital is doing fine; dumping these employees is tough for the workers themselves -- and arguably bad for society at large -- but it simply doesn't damage profits much. It's a cold, hard reality, and one that we will have to deal with, one way or another.
Funny thing: Keynesians have been quick to dismiss the ZMP story.  But old-school Keynesianism is a ZMP theory of a particular sort.  Consider this remark by Paul Krugman
Caplan frames the argument in terms of the nasty effects of raising labor costs. Um, we have a problem with demand, not supply; time to reread Keynes on wages.*
As far as I can tell, Krugman is saying that when there's insufficient Aggregate Demand, cutting wages won't increase employment.  And how is that possible?  If workers' marginal product is zero (or less).  After all, as long as a worker has a positive marginal product, there will be some wage low enough to make it profitable to employ him.

Keynesians could certainly respond that, unlike Tyler, they think that workers' ZMP is an artifact of low demand, not a feature of a modern skill-based labor market.  But before they do, I'd like them to admit that the Cowen-Keynesian debate isn't about the reality of ZMP workers.  It's about whether the ZMP status of the unemployed is structural or cyclical.

Either way, I'll stick to my view that a 5-10% fall in real labor costs would rapidly return us to 4% unemployment - and remain baffled by economists who hastily second-guess the textbook solution to a modest labor surplus.

* Here's my original reply to Krugman.

Comments and Sharing

COMMENTS (13 to date)
Tyler Cowen writes:

You wish to say you are not a fan of the ZMP view, but both the original and subsequent expositions explicitly make your point of structural vs. cyclical and suggest it is a bit of both with the division uncertain.

James Oswald writes:

MP does not have to be 0 for the ZMP story to hold, it just has to be lower than the wage workers will accept. Raising the minimum wage can trigger ZMP-like behavior, as can wage stickiness.

Michael Fisk writes:

I'm thinking that effective ZMP is more an artifact of binding wage floors than a lack of practical skills - if it were the latter, a firm planning for the long run would consider it worth their while to take some of these workers, pay to train them, and reap the benefits in future months/years. If skills were all these ZMP workers were lacking, then we'd see noticeable returns from government job training programs - gains that have yet to materialize, despite decades of attempts.

The question then would be whether a reduction in wages alone would be enough to provide incentive to an employer to bear the responsibility for training a near-ZMP worker, or whether the typical cost burden of worker-expected, employer-provided benefits alone exceed the expectations for average marginal product within a certain time horizon.

Nathan Smith writes:

"Either way, I'll stick to my view that a 5-10% fall in real labor costs would rapidly return us to 4% unemployment"

I doubt it. I think the basic problem is that, to be productive, workers need to be paired with capital, and capital formation is being prevented by uncertainty about the future. The debt deal might have helped there, but it will take time for business investment to regain lost ground sufficiently to keep all American workers busy, even if the wage falls.

There might be ways to use workers with a lot less capital, but it would take a long time for the market to figure them out, and it won't be worth doing that unless they expect it to be worthwhile to use low-cost workers for a little longer. So, ZMP is an exaggeration, but the marginal market wage is probably a lot more than 5-10% above the typical unemployed worker's marginal product, for the time being.

Colin k writes:

The only companies I know of whose profits are labor-constrained right now are those in exceptionally hot sectors, like iPhone app development. If labor rates for auto manufacturers or residential homebuilders dropped by 10%, logic suggests this would go towards increased margin and not employment. The only time you need labor is when you can't service demand. If you could figure out a way to teach mediocre programmers to be good ones, let alone teaching UAW members to become mediocre programmers, you Gould make a large fortune.

[duplicated "http: in url fixed--Econlib Ed.]

Chris Koresko writes:

Cowen and Lemke: Their productivity may not be literally zero, but it is lower than the cost of training, employing, and insuring them.

Bryan Caplan: After all, as long as a worker has a positive marginal product, there will be some wage low enough to make it profitable to employ him.

It looks like there's a confusion of terminology here: C&L include wages in their definition of marginal product, while Caplan apparently doesn't. No wonder they don't reach the same conclusions about it.

Tom West writes:

I think that the definition of a ZMP worker is fairly broad. You can slash your R&D or Quality Assurance department by a lot and the revenue won't go down for a year or two.

Were the workers there ZMP? Pretty clearly yes if you look at the next year's balance sheet, no matter *what* they're paid (which is why I think Bryan is wrong). I suspect that there are a *lot* of companies that are basically coasting on previous institutional capital that ZMP workers provided in order to protect themselves from being destroyed in a very uncertain market.

These companies may hit very hard in another two years or so when they've "used up" their store but for now they've increased security by throwing their ZMP workers overboard.

Let's hope for their sake that their competitors did the same.

Dan Lemnaru writes:

Who's to say that rising wages are not exactly what the economy is missing? It's a known fact that in the US real wages have been more or less stagnant, while productivity has grown for decades. That in itself may be a sign of what's wrong with the system, the debt etc.

Anyway, imagine a world where a king has say 10 subjects. Initially all that the 10 subjects can do through their labor is to barely feed themselves and their master. After a while, they are able to increase productivity by 100%, so that they could all be fed properly and even have a surplus. The king however keeps their salaries (rations) unchanged. The king himself eats and wastes a lot, but even so there are surpluses which he piles in a big heap.

The king decides that 2 of his subjects should stop working in the field because there's just too much food, and there's no need for it (from his POV). One he keeps employed for entertainment and other purposes. He stops paying the other one - "You're no longer needed. I own everything. You may not use anything that you don't own. Good bye!"

He's the "zero marginal product" worker, his existence is no longer required for his society to continue to function.

Cutting wages to everyone else will not make him any more likely to be employed again.

One solution for him to survive would be for the king to decide to use some of the extra production and raise wages. The workers would now have some leftovers, enabling them to hire him -- in return getting a taste of king-like entertainment, housekeeping etc. every once in a while.

Another solution would be simply to have the king feed him anyway, but this is unlikely to be seen as fair by the others, risking an uprising in the long run.

Yet another solution would be to reduce the number of working hours by decree, by about 1/8th, while keeping the daily wages constant. ZMP guy's services would now be needed again to maintain production.

Lord writes:

There is a sense in which they are ZMP. Since the wages of the employed can't be cut, the wages of the unemployed must be eliminated to make up for them, so instead of a 10% across the board cut, we end up with 90% with no cut and 10% with a 100% cut. Naturally, no one is going to work for nothing.

How does cutting wages not work? When the cut is nominal rather than real for prices will also fall. Krugman has also pointed out that when you no longer have control over money, grinding deflation is the only possibility, but that can hardly be called a cure; it isn't fast or painless.

Lord writes:

There is a cure though, cut real wages through inflation. Inflation would not have to reach 10%, but would rapidly lower unemployment.

Mr. Econotarian writes:

While the economy is clearly poor, I think unemployment is marginally worse due to the extension of unemployment insurance length slowing personal career recalculations, plus the minimum wage (+payroll taxes) putting a trap door below which low marginal product (LMP) employees fall through.

I'd like to see a comparison of unemployment of those with a high school degree compared with state & local minimum wage levels. Although who knows if "illegal" aliens are counted properly in either data point.

Robbl writes:

I finally realized where the right wing is coming from in this whole discussion.

You say " I'll stick to my view that a 5-10% fall in real labor costs would rapidly return us to 4% unemployment"

The left says it an AD problem.

I think real labor costs have fallen by 10%. If employers aren't able to find workers today at -10% real from what they would have had to pay in 2008, they aren't trying very hard.

You clearly think something different.

It is similar to discussing the laffer curve. Everyone agrees that there is a curve, the argument is where you are on the curve.

Jeff writes:

Keynes knew Marshallian economics very well. The classical cure for a recession was falling real wages. But Keynes argued that the same demand shortfall that drove nominal wages down would drive prices down as well, so that real wages would not change.

You can argue that Keynes is wrong about this. But what Krugman is saying is a fair interpretation of what Keynes wrote. Few economists today really know what Keynes was all about, because the experience of the 1970s so discredited his theory that it isn't really taught in grad schools any more.

Oh, I just now noticed that Lord's comment above makes much the same argument. His second comment is also correct.

Comments for this entry have been closed
Return to top