I got a response to a posting on efficiency wages about a month ago
and kept meaning to reply with a long post to make clear exactly what
I meant by "efficiency wages don't undermine neoclassical economics."
The thought kept making me delay. :) But here it is, and it IS long...
Some time in late December, John Buttrick wrote:
> For fixed capital and assuming some unemployment at the start, profit
> maximizing and an ordinary prod fnc: X =f(N), with N being workers "off
> the street". They can be made into "standardized workers": N=Lg(w) where
> w is real wages. So profit=F(L,w) and the most profitable pair of values
> for L,w imply unemployment. It will vanish as capital increases but not
> rapidly because labour is also increasing. If employed workers share,
> unemployment will be somewhat reduced. If employed workers and employers
> are asked to vote on a political full-employment package, they will oppose
> on grounds of self interest since such a package would reduce their wages.
> Sort of like discrimination where those who do it, gain. It seems to me
> that an efficiency wage model is anything but conservative in implications.
> It leads "naturally" to a class-type society, for example.
I think this use of the words "class society" underpins exactly what I meant
about E-W theory not being radical. What you are suggesting here is that
employed or "insider" workers will want to guard their wages against the
"oppressed" class of unemployed, secondary-labor-market, etc. workers.
This IS after all a conclusion that follows from the wage being determined
after labor demand: higher wages means less employment and lower output.
So what would be the optimal policy solution? Under a shirking model, the
government should develop better monitoring technology. Under a specific
capital model, it should work on standardizing machinery and reducing
workplace idiosyncracies; a rent sharing model would create a better
outcome for workers as a whole, but would also suggest that certain workers
are destined to remain privileged. One could make the case that an
"effort wage" simply reflected a hedonic locus of workplaces where some
jobs required greater incentive than others, but ditto about the privilege.
The defect in all of this is that it assumes that workers would be best off
at an even-wage equilibrium, where employment and output are maximized.
Although wage-earners would gain if labor demand were inelastic (it is),
efficiency wages according to the given theories create stratifications
among wage-earners; therefore it would be most equitable to get rid of them
and move to the lower-wage full-employment equilibrium.
So these theories basically imply that individually rational behavior
creates socially inefficient behavior by deviating from the Marshallian
outcome. In this vein I can't see them as anymore an undermining of the
neoclassical theory than the existence of monopolies: the proscription
is that they exist and should be gotten rid of.
A more radical theory in the traditional (Marxist) sense is that labor
demand is not exogenous to the wage equation and that the aggregate
wage level significantly effects the demand for an individual firm's
product; therefore in the aggregate sense, W<VMP even though as the
individual firm sees it, it is paying the workers at or above their
marginal products taking the wages of other firms as given. The
existence of W<VMP is necessary and sufficient for there to be a
class conflict between workers and capitalists, since it is equivalent
to saying that workers are being exploited.
[A side note on terminology: I had said that efficiency wages imply
W>VMP, and was corrected that employers are paying workers a higher-
than-market VMP. Let me re-state what I meant in a more precise way:
if there were a way to control for shirking, quitting, or strikes,
then the wage would be lowered. The extra pay is not for higher
output but to buy off the workers against the threat of quits, etc.
Therefore, from the point of view of physical production, W>VMP
even though under these production-related threats employers still
pay equal to the marginal product of the workers that they hire; it
is higher, though, than what they would be able to give if they could
pay as many workers as they could the wage that those workers would
be willing to work for, so in that sence, these PLM workers are
supposedly not being exploited.]
I do not deny that efficiency wages exist. Sometimes they are helpful,
for example when workers in primary markets (eg auto in the WWII period)
are able to get primary wages and then fight for them to be expanded to
the rest of the economy. Sometimes they are harmful, since they can
allow wages to be determined by non-productivity-related characteristics
(eg, race, gender, etc.). One thing they are certainly not is a damning
of capitalism. I am happy to agree that sections of the working class
maintain privilege because of wage structures, but focussing on this
as somehow undermining neoclassical economics (unless by "neoclassical"
we only mean a handful of Chicago straw-men) can actually lead us to
ignore the more significant problem of exploitation by capitalists.
Smash the pretty blonde stewards,
Tavis
P.S. Ken, thanks for the sig slogan. I'm soon going to develop a "MIM"
program in Lisp along the veins of Elisa except that it responds to
entire messages and not just lines -- and is based on keywords like
"armed peasants" and "working-class thugs" instead of "mother," "father,"
etc. (I'll base it on the Elisa code, which I have).
