It’s curious how little guidance our standard models give us
about what should happen in labour markets as we come out of a recession. Employment will increase, of course, but what
should happen to the nominal wage? It’s
a bit unfair to be criticizing newspapers for relying on gut feelings, if
economic models aren’t giving clear testable hypotheses, at the very least.
Consider the sticky-nominal-wage models. In those models, nominal wages are sticky
downwards, not upwards. It’s unlikely
that nominal wages will fall in a recovery, and there’s a reasonable chance
they’ll remain unchanged, if only because there’s still an excess supply of labour
so firms don’t have to raise wages to attract labour yet, but as to whether
they’ll rise, all we can really say is that if they do, they can’t rise any
faster than the price level does, since in the s-n-w models the real wage
should be falling , at least through the early stage of the recovery.
One framework for thinking about the nominal wage has gained
some traction recently: that’s the monopsony labour market model . The Obama White House Council of
Economic Advisors brought out a briefing paper that suggested that monopsony was
responsible for a lot of undesirable outcomes, including a decline in the
labour share of national income. The
arguments in it are pretty sketchy (and to be fair it’s not intended to be a
piece of technical analysis – it doesn’t even set out a formal monopsony model)
but a certain number of policy analysts think it fits what they see as going on
in the aggregate labour market.
Formally, Monopsony is the buyer-side counterpart of
Monopoly. Joan Robinson came up with the idea back in the 1930s.
A monopoly is a single seller; a monopsony is a single buyer.
When we set up the diagram for a monopoly, we draw a
downward sloping demand curve and along with it a downward sloping marginal
revenue curve. In the case of a
monopoly, the market demand curve is also the firm’s demand curve since as the
only seller, it doesn’t have to share the market with anyone. If the market were competitive, the market
price would be at the intersection of the demand curve and the supply curve,
which we usually assume is the same as the monopolist’s marginal cost
curve. Every individual firm selling
into that market would have to charge the same price since, if it tried to
raise its price, its customers would simply turn to one of the other suppliers. If the firm is a monopoly, though, it can
pick a price/quantity point anywhere on the market demand curve and set the
price at that level. The price will be
above the competitive price, but there’s nothing which will force the
monopolist to cut its price since it doesn’t have to share the market with any
competing suppliers.
To find the profit maximizing price, it picks the quantity
of output at which the marginal revenue curve cuts the marginal cost curve, and
reads the price off the demand curve, directly above that quantity point.
The marginal revenue curve is obviously key to the monopoly
analysis. Typically when we talk about
monopoly we’re talking about a single price monopolist, meaning that it has to
charge the same price for each unity of output that it sells. The demand curve tells us that, if it wants
to increase the quantity it sells, it will have to cut the price it charges not
just for the additional units it’s planning to sell but for every unit,
including those it could have sold at a higher price. That means that its revenue will increase by
the additional amount the new sales bring in, but fall to the extent that it’s
had to cut the price on units it could have sold at the higher price. The Marginal Revenue curve nets out those two
effects and tells us whether cutting the price will increase or decrease total
revenue, and by how much. So the
important thing in the basic monopoly model is the fact that because every unit
it sells is identical to every other unit, the single-price monopolist has to
set the same price on each unit.
In the case of monopsony, there’s a single buyer. Just as the market demand curve constrains
the price/quantity choice of the monopolist (i.e. the (P,Q) pair it picks has
to be on the demand curve), there is a supply curve which constrains the
price/quantity choice of the monopsonist (we’re assuming that the suppliers of
the thing the monopsonist wants to buy are small, perfectly competitive firms). The monopolist decides how much it wants to
sell, the monopsonist decides how much it wants to buy. Just as the monopolist has to charge the same
price for each unit it sells, the fact that the units it is buying are
identical means that the monopsonist will pay the same price for each unit it
buys. If the monopolist wants to sell
more it has to cut the price it charges for each and every unit; if the
monopsonist wants to buy more it has to raise the price it pays for each and
every unit.
That last bit gives us the Marginal Expenditure curve, the
monopsony counterpart to the Marginal Revenue curve. It starts from the same vertical intercept as
the supply curve and, if the supply curve is positively sloped so too is the
marginal expenditure curve, lying above the supply curve in the same way as the
marginal revenue curve lies below the monopolist’s demand curve. The monopsonist’s marginal expenditure curve reflects
the fact that in order to buy more the monopsonist must not only pay a higher
price for each new unit it buys, it must pay a higher price for each unit it
had previously bought at a lower price.
So the marginal expenditure curve shows how its expenditure rises as it
buys more, and lies above the supply curve for every unit of output.
The monopsonist is assumed to have a marginal benefit curve
for the thing its buying, which has the usual downward slope. The optimal quantity is found at the point
where the marginal expenditure curve cuts the marginal benefit curve, and the
price the monopsonist will be willing to pay per unit is read off the supply
curve at that quantity. Note that
because the marginal expenditure curve lies above the supply curve, the price
will be below the marginal benefit the monopsonist derives from buying that
quantity of output. Just as the gap
between price charged and marginal cost is a measure of the degree of monopoly
power, the gap between price paid and marginal benefit is a measure of the
degree of monopsony power.
The monopsony model has been mentioned in labour economics
pretty commonly, although generally not taken as too important. In that setting you’re assuming a single
buyer of a particular type of labour – a company town, for example – and the
marginal benefit curve for the buyer is the Value of the Marginal Product of
Labour curve. The key assumption is that
the labour supply curve being faced by the employer is positively sloped,
rather than horizontal, which means that the firm can cut the wage it offers
without losing all of its workers (which is what would happen if the labour
supply curve were horizontal). The
assumption that the employer pays the same price for each unit of labour it
hires means that in order to hire more labour it has to raise the price it pays
every unit, including the units it was previously able to hire at lower
wages. That gives us a positively sloped
Marginal Expenditure curve, in this context called a Marginal Factor Cost
curve, lying above the Labour Supply curve and steeper than the supply curve so
that the gap between the two widens as we move to higher and higher levels of
labour, and an intersection between VMPL and MFC at a point where the VMPL is
above the wage (the wage being read off the labour supply curve). So if you’ve got a monopsonistic labour
market, labour isn’t paid the full value of its marginal product and the wage
will always be on the labour supply curve.
The reason the monopsony model was always regarded as more
of a curiosum in labour economics was that it’s hard to come up with real world
examples where it might hold. That’s
probably why most studies which used it dealt with only a few types of labour –
hospital nursing markets in the US probably being the most common.
It’s come back into fashion among a few analysts who have
argued that, even if they’re not sole buyers of labour, a large number of firms
face upward-sloping labour supply curves and hence MFC curves lying above their
particular labour supply curves. So
whereas standard theory assumes that the wage/labour combination will be on the
VMPL curve, monopsony theorists argue that it will be on the Labour Supply
curve. Then in an economic recovery, as
the VMPL curve moves to the right, the wage should also rise.
There’s a growing literature claiming to find evidence for
monopsony in the sense of finding upward-sloping firm-specific labour supply
curves, some at the micro level and a few at the macro. In general these studies can’t actually test
for the existence of a gap between the VMPL and the wage, since the VMPL is
unobservable in the structures they’re using, so most of them conclude that
firms could exercise monopsony power.
At the aggregate level, there’s a paper by Glover and Short
which concludes that firms have monopsony power over older workers. The idea here is that the costs of changing
jobs, especially if it involves moving, are higher for older than for younger
workers, so it would take a larger wage cut to make it worth an older worker’s
while to up and leave, which makes the supply curve of older workers more
steeply positively sloped than that for younger workers and hence gives their
employers greater monopsony power over them.
The fact that older workers are more experienced and so generally have a
higher VMPL increases the gap between their particular VMPL and their
particular wage.
At the level of a particular sector, there’s a recent paperby Goolsbee and Syverson which drew a certain amount of
academic interest. The title is “Monopsony
Power in Higher Education: A Tale of Two Tracks”, and according to the
abstract:
“ This paper tests for and measures monopsony power in the U.S. higher education labor market. It does so by directly estimating the residual labor supply curves facing individual four-year colleges and universities using school-specific labor demand instruments. The results indicate that schools have significant monopsony power over their tenure track faculty. Its magnitude is monotonic in rank, being greatest over full professors and smaller for associate and assistant professors. For non-tenure track faculty, however, universities do not seem to have any monopsony power and instead face perfectly elastic residual labor supply curves. Universities’ market power over tenure track faculty does not differ between public and private schools nor between female and male faculty.”
So while one might well be skeptical of the overall
monopsony story, when you’re assessing the evidence, don’t forget those poor,
downtrodden exploited full professors.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.