[Note for bibliographic reference: Melberg, Hans O. (1997), Five short ideas, http://www.oocities.org/hmelberg/papers/970609.htm]
Five short ideas
Nurses, Keynes, Hirschman, Popper and Akerlof
by Hans O. Melberg
Introduction
Time is scarce and my memory is imperfect. Hence, to save time, and to avoid forgetting
some of these thoughts, I have decided to write down five ideas which I originally
intended to write as five separate observations. I have to warn potential readers that
these ideas are not scrutinised too closely. There are probably lots of background
assumptions and variables that I have ignored. As before, criticism and comments are
welcome.
1. Nurses
In the newspaper Vårt Land there was recently an article about the shortage of
qualified nurses and pre-school teachers in Norway. The paradox, however, was that a
survey revealed that several thousand people who were educated as nurses and pre-school
teachers had decided not to work within their profession. Thus, the problem - it seems -
was not the number of people trained, but the number of people who chose to work within
their profession.
My first reflection was that this is what happens when the government sets the wage. If
the laws of the free market were allowed to work, the wages of nurses would rise until the
shortage was eliminated (assuming that the supply is not backward bending - in which case
the system is unstable). However, a second reflection also hit my mind: The problem was
not only the wage. In fact, there is an interesting vicious circle here: When there is a
shortage of nurses, those who work as nurses are under constant pressure to work fast.
This, in turn, makes it less attractive to be a nurse (at a given wage). In short, few
nurses makes it even less attractive to work as a nurse. Some nurses may even decide to
quit because of the stress (and some qualified nurses who would otherwise accept the given
wage will not work since it is too stressful). But, when some nurses quit it becomes even
more stressful for those who are left - leading to even more nurses to quit their work. In
short, we have an bad circle.
I do not claim that this circle is the main factor behind the lack of qualified nurses
in Norway. I do, however, believe it is one possible mechanism. The interesting point
about this mechanism, is its potential for creating instability. We like to think about
the free market as self-correcting and stable. This example shows a mechanism which leads
to the opposite conclusion - the market mechanism left to itself would be
"explosive" and unstable (For another example of the same, consider a Cob-Wed
structure in a market in which the suppliers increase the supply when the price is
low e.g. a family which makes more cogs in order to achieve their needed income when the
price is lower). This may justify government intervention and some form of central
planning. As an extreme example, imagine the government sent a letter to 3000 qualified
nurses telling them to where to work. This could solve the co-ordination problem between
the nurses which prevents them from working within their profession. But, of course, in a
liberal society this kind of ordering mechanism is unacceptable). However, I do not want
to stretch the conclusion too far. There are many other factors at work, and a single
mechanism leading to instability does not make the whole aggregate market unstable.
Nevertheless, knowledge of these mechanisms may be of use when we are analyzing a problem
(such as the lack of nurses) and prescribing cures - e.g. try to reduce stress as opposed
to increase the wage in order to attrach more nurses.
2. Was Keynes wrong?
In the book The Unemployment Crisis Layard, Nickell and Jackmann discuss the
problem of why there is persistent rationing of jobs (unemployment). One answer to this is
the so called dis-equilibrium models based on rationing in the goods market (the firms
cannot sell as much as they want, thus they do not employ "enough" labour to
clear the market). For example, Malinvaud, and Barro and Grossman have built models based
on this. However, LNJ reject this approach because the models imply "an arbitrary
rigid price which prevents perfect competitors from selling all they want to: we know of
no mechanism that could sustain such a price" (p. 33). The argument seems valid.
There is little reason to believe that the producers are rationed on the goods market
since they are free to lower their price in order to sell more (true, there are some
mechanisms, such as menu-costs, which create some degree of rigidity. However this is not
enough to sustain a dis-equilibrium price over time).
Does this mean that Keynes was wrong - that there is no such think as Keynesian
unemployment (unemployment caused by low aggregate demand)? I need to think more about
this (especially about Clower's dual decision hypothesis, and the co-ordination problems
which occur when all firms try to reduce costs (wages) in order to reduce price and sell
more). However, there is an alternative interpretation of Keynes. To examine this we have
to assume that we live in a world of uncertainty. Imagine a firm which decreases the price
of its product in order to sell more. What if customers take this as a signal that the
price is going to decrease even more in the future? If expectations are formed in this
way, a price reduction may actually lead to a decrease in sales. Once again we have a
mechanism which leads to instability - an adjustment away from equilibrium after a shock.
However, once again I think there are many other factors at play, and the mechanism
described above need not dominate.
3. Was Hirschman wrong?
I recently wrote an article in the student journal Observator in which I argued
that Hirschman's analysis of exist and voice may serve as a justification for a monopoly.
The argument was that in a monopoly the customer was forced to complain (use voice) when
something was wrong, while in a market with many firms the customer could simply stop
buying the product (exit). The reason this may justify a monopoly, is that voice may be a
better communicator of information than a decrease in profit, since the firm will know
exactly what the customers are dissatisfied about. Hence, a monopoly may be best because
it forces people to use voice, not exit. However, Ole J. Røgeberg pointed out to me that
this argument is flawed.
The argument is flawed because profit maximizing firms would seek the information
communicated by voice as long as it was profitable. Film-producers test their movies on
small audiences before they release the film; Ice-cream producers test their products
before they are released; Most companies try to encourage feedback from customers by
rewarding those who respond - you get chocolate if you complain to the chocolate producer!
Hence, the voice I thought could be gained by having a monopoly, is already exploited in
the free-market. There might, of course, be imperfections in this mechanism (if people
have high discount rates they will not be motivated by the reward, or of the ability to
use voice effectively is unevenly distributed such as among parents of children in a
school), but the pro-monopoly argument also has many flaws. For example, why would a
monopoly care about voice (the complaints)? Hence, I think Hirschman's analysis represents
a weak case for monopolies.
4. Unfalsifiable theories, Popper
In a previous observation I argued that even unfalsifiable theories might be assigned a
probability of being true and hence should not be discarded (see Why
unfalsifiable theories are also worth considering). I still think there is some value
to this argument, but my interpretation of Popper may be incorrect - or at least
incomplete. To understand why, consider the following arguments (from The Economist,
31. May, 1997, p. 31):
The Monopolies and Merger Commission in Britain has recommended to ban recommended
retail prices across a wide range of electrical goods. The reason was that the firms used
these recommended prices to discourage price competition between retailers. One piece of
evidence used by the commission, was that the price of many products did not vary much
between retailers. Typically the prices of 90%, 80% and 75% of (respectively) the washing
machine Hotpoint WM22, the tumble-dryer Zanussi TD520, and Bosh's dishwasher SMS 4452 were
within a narrow 5% band. This indicates a small degree of price competition. However, the
producers complained and argued that "the similarity in prices is evidence not of
collusion but of fierce competition" (p. 31). This is interesting: We have one piece
of evidence (the similarity of prices) and two radically different interpretations.
I guess it is this kind of problem that Popper (and Kolstø - see the previous
observation) was worried about. I also think they were right to worry about this, but I
still believe one might find evidence to discriminate between the plausibility of the two
theories (although one might not be able to falsify one completely). For example, if the
firms are correct that it is the fierce competition which is producing the uniform prices,
they should not be afraid of the ban on recommended retail prices. Thus, their resistance
against this move makes me believe that RRPs actually do hinder competition - and the
interpretation of the uniform prices as evidence of collusive behaviour is more plausible
than the "fierce-competition" argument.
5. Prices as signals
In economics one often speaks about prices as signals. The standard theory is that prices
are good conveyers of information i.e. good signals. A higher price signals an increased
demand - and investors will expand the capacity in that sector. In this way the price
signals make sure that the structure of the economy is adjusted to the preferences of the
population. However, the theory isn't as easy as it may appear.
One exception is presented by Akerlof in his article on adverse selection and lemonds.
In this article he shows how the price signal does not fulfil its function. The problem
occurs when there is asymmetric information and the price is taken as a sign of quality.
For example, in the market for second-hand cars the seller knows the quality of the car,
while the buyer has less information. The price demanded for a second-hand car may then
become a signal of the quality of the car (the seller will demand more for good quality
cars). However, when the price is a signal of quality the seller of a bad car may try to
bluff: To charge a high price in order to fool the buyer into believing that this is a
good second hand car. Now, to make a long story short, the buyer knows he might be bluffed
so he will never pay as much as the seller of good quality cars demand for their cars. The
end result is a market failure (caused by asymmetric information): There is no market for
good quality second-hand cars.
There are many more examples of signals in economics. Spencer's model of education as a
signal; signalling in game-theory (to signal your type e.g. by drinking beer to signal
that you are tough even if you really want Fanta!); Veblen argues that we buy expensive
products to signal that we are rich. All of these exemplify the complexities of signalling
in economics. Prices are not always perfect signals leading to market equilibrium. In
fact, I mentioned another possible exception to the question about Keynesian unemployment.
A decrease in price may be taken as a signal not to buy more, but to wait in order for the
price to fall even more!
I now want to discuss a more general frame for signals in economics, and I want to give
an example of market failure which may be original (!). The first question concerns this:
What do we want our signals to achieve? The price mechanism serves at least two functions.
First, it allocates goods at any point in time (those who pay most get the goods). Second,
it directs investments since goods in high demand will fetch a high price which, in turn,
will lead to more investment in that sector. One might call this the static and the
dynamic function of prices. The interesting questions are now: First, do the static and
the dynamic function of prices conflict. Second, are price signals a good way to allocate
resources? Third, is the price mechanism a good way to co-ordinate the adjustment of size
of the various sectors in the economy?
The first question is quite well explored. From Tinnbergen we know that the same
mechanism cannot be expected to maximize two objective functions at the same time i.e.
there is probably a conflict between static and dynamic efficiency. For example,
Schumpeter has argued that monopolies are dynamically efficient, but statically
inefficient. The precise mechanism is that monopolies earn enough profit to invest in
research and development, while firms in a perfectly competitive market do not have the
same resources to invest and improve efficiency (This argument has many weaknesses, but I
will not discuss these here).
As for whether the price mechanism is a good way to allocate resources, this is a
normative question. Consider, for example, an economy with individuals who have different
abilities. Assume person A earns $5000 a month, while person B gets $2500. Assume also
that a certain good - say a ticket to a football game - would increase A's utility by 10
units, while it would increase B's utility by 15 units. It is fully possible in this
economy that A gets the ticket, even though B would benefit more in terms of utility from
the ticket. The ticked does not go to the one with the highest utility valuation because
income is unevenly distributed. There are, of course, many hidden assumptions in this
example. For example, there are problems of interpersonal comparison (and there is an
incentive problem if we start compensating lack of ability since this would make it less
profitable to work in order to acquire skills). Moreover, the example is not a general
proof that a market economy in general does not tend to give goods to the people who will
gain the most utility from the good. In short, there are many more factors at work than
can be discussed in a single paragraph!
Third, dynamic efficiency. In the example of nurses (question 1) I argued that the free
market would solve the problem of a shortage of nurses by increasing the salary of nurses.
I now want to examine this mechanism a bit closer in order to see if this mechanism is
efficient. The key is this: While it seems necessary to increase the salary to attract new
nurses, it is not necessary to increase the salary of the nurses who are already working
on order to make the "nurse" sector bigger. However, the free market does both:
A lack of nurses drives the wages of all nurses up. I wonder whether this represents an
inefficiency which could be avoided. Then again, it does not really represent a dynamic
inefficiency (a rise in wages will lead to more nurses - unless the supply curve is
backward bending). The problem is more that the cost is too high compared to what it could
have been in an imaginary ideal (?) world in which we could target the signals more
precisely.
[Note for bibliographic reference: Melberg, Hans O. (1997), Five short ideas, http://www.oocities.org/hmelberg/papers/970609.htm]