Zero Marginal Product of Labor?

Debate has returned to Tyler Cowen's odd notion that current high levels of unemployment are at least partly explained by some workers having a zero marginal product.  

I find the idea perplexing.   It goes like this.    Both real GDP and employment fell during the recession.   Real GDP was $13,363 billion in the fourth quarter of 2007 and fell 553 billion to $12,810 billion by the second quarter of 2009.    Employment, on the other hand, was 146.3 million  in December of 2007 and fell 8.3 million to 138 million by December 2009.  The decrease in real GDP was 4.1 percent and the decrease in employment was 5.7 percent.

While those figures suggest that some of the lost workers weren't exactly pulling their weight, the real puzzle is with the recovery.    Real GDP has risen by $469 billion to $13,279 billion.   It is only $85 billion less than it was in late 2007.   The 4.8 percent increase by the third quarter of 2010 has brought real GDP up to only .64 percent short of where it was before the recession began.

But what of employment?   Employment has increased by 1.25 million, up  to 139.6 million.   But that is only a slightly less than 1 percent increase, and employment remains 4.8 percent below where it was before the beginning of the recession.     So, .6 percent less output is produced with 4.8 percent fewer workers.

Cowen is claiming that those 7 million workers have zero marginal products.    They must not have been adding anything to production, because what they were producing can apparently be produced without their help.  Or, more exactly, he is claiming that this should be the baseline assumption.

In my view, the proper baseline assumption is that if 7 million of the 14.5 million unemployed were put to work, so that employment returned to the level of late 2007, the 5 percent expansion of employment would result in output increasing by 5 percent too, roughly to $13,942 billion.   This, of course, would assume that the marginal product of the unemployed workers is equal to the average product of the currently employed workers.

It is conceivable that none of these people could contribute anything to the production of any good or service.   They would simply get in the way of existing workers.     Or, it could be that they could generate some physical product, but that it would be of no use to anyone.   The marginal physical product of some or all of those workers might be positive, but the marginal value product would be zero.

I think the most likely result is that the marginal value product of those workers is somewhat less than the average product of all current workers, so that real GDP would rise by less than 5 percent.    (While I favor money expenditure targeting, and so, favor allowing market forces to determine employment and production, I certainly think that what should happen is that employment rise closer to 10 percent and return to its trend growth path of the Great Moderation.)

Now, some might argue that sure, those unemployed workers could produce something, but that their employers won't be able to sell it.     So, the workers have a zero marginal product.

Actually, production adds to real GDP whether it is sold or not.   But the proper determination of whether or not the workers have zero or positive marginal value products depends on what people would be willing to pay for the additional output.     Zero marginal value product workers can either add nothing to physical output, or else, no one is willing to pay anything for what they can add to output.

Of course, outside of perfect competition, firms hire where wages are equal to marginal revenue product, not marginal value product.   For example, if firms hired more workers and produced more output, and could continue to sell current levels of output at existing prices, and charge lower prices for additional output, they might be willing to try.   If unemployed workers were willing to accept lower wages, then perhaps employment would be increased.   Heck, maybe the lower paid employees could buy the lower priced products.

But, of course, this is unrealistic.   Few firms could price discriminate in that fashion.   And so, to lower prices to sell more output requires lower prices for their current volume of output.   And operating on that basis doesn't simply require paying less to newly employed workers, but also lower pay for current workers.   In other words, by far the best way to increase the amount firms can sell is for money expenditures to grow so that more can be sold at current prices and wages.  

Cowen's zero marginal product of labor argument is that even if firms could sell 5 percent more, and tried to hire 5 percent more workers, their actual level of production would be unchanged because those workers have zero marginal products.   They would produce nothing.   The firms, then, would respond by raising their prices by approximately 5 percent.    And competition for those workers that are currently employed (because their marginal products are well above zero,) would result in their receiving pay increases of approximately 5 percent.

I don't believe it.