Problems with neoclassicism

Gabriel isn't impressed with my claims regarding neoclassical competition theory:

Oh, yeah, and AJE is wrong. There’s nothing wrong with mainstream theory and its concept(s) of competition.

Part of this is a slight miscommunication on my part. When I originally said that

Regulators use the term “competition” as an adjective that labels the conditions of various market states. This is consistent with textbook economics, but is a faulty (and retreating) type of economic reasoning.

I interpret Gabriel's response as maintaining the internal consistency of textbook competition theory, whilst accepting that it might not be the most sensible way to study real markets for regulatory purposes, and thus a mistake to utilise. But he thinks i've overstated by claiming that the theory itself is faulty. As promised, a clarification. I meant "faulty" to mean "not up to the task", and thus my critique is aimed at the utilisation of the theory, not the theory itself. Mea culpa.

However, I don't think that neoclassical competition theory is internally consistent. In Boettke, Coyne and Leeson's "Saving government failure theory from itself: recasting political economy from an Austrian perspective", they present Demsetz's critique of the standard monopoly diagram. The argument is pretty straightforward - the inefficiency of monopoly stems from the existence of dead weight loss. But why would a monopolist leave potential gains from trade on the table? All they need to do is perfectly price discriminate. The response is that this is very costly, but these are transaction costs, and should be part of the marginal cost curve. If the MC curve reflects the costs of price discrimination, P=MC and the inefficiency criteria disappears. Over to Gabriel.

Comments

[Gabriel]

You are right, in a sense. If there's perfect information (including information about everyone's characteristics), then yes.

I see this as a particular case of what Usher says in his Coase theorem paper: with perfect information and costless bargaining (and I add a commitment device, e.g. property rights) we're always going to get efficiency, although the split of the surplus is not necessarily determined. (In other words, you can always get to the land of Oz, eh, I mean the core.)

On the other hand, here's some token defense of the textbook monopoly model:

>> "The response is that this is very costly, but these are transaction costs, and should be part of the marginal cost curve."

Not so fast.

Let's say that the monopolist knows for certain the true market demand curve, in which case the standard analysis (restricted to one price for everyone, if you will) goes through.

Now, let's say that this market demand comes from a large number (continuum of measure 1?) of buyers with various willingness-es to pay.

What we want is a mechanism that will satisfy:

a) participation constraints (buyers can abstain from buying, monopolist can stop production);

b) revelation principle (everyone tells voluntarily their true willingness to pay => perfect discrimination);

c) The monopolist makes at least as much profit as by charging a single price to everyone.

It's not obvious to me that such a mechanism exists, even if we were to relax (c)!

Assuming such a mechanism exists (with or without (c)), I'm not sure how to translate this into marginal cost considerations.

[Michael Greinecker]

If the monopolist could perfectly price discriminate, you would be right. We agree that that is usually not the case, so monopolies and taxes are usually not first best efficient.

"But why would a monopolist leave potential gains from trade on the table? All they need to do is perfectly price discriminate. The response is that this is very costly, but these are transaction costs, and should be part of the marginal cost curve. If the MC curve reflects the costs of price discrimination, P=MC and the inefficiency criteria disappears."

You don't get rid of problems by reinterpreting curves. The problem is that "transaction costs" may be part of the monopolists perceived cost curve, but not part of the technology socially available.

We know there exists an income distribution so that when the monopolist does marginal cost pricing while ignoring "transaction costs", the resulting allocation Pareto dominates the one in which the monopolist maximizes profits. A planner may not have the information to implement the Pareto gain, but if we are satisfied with maximizing consumer and producer surplus in this partial equilibrium framework, we will be fine. Your "transaction cost" part of the cost function

[Michael Greinecker]

Ooops, got cut off.

...Your "transaction cost" part of the cost function is a cost of socially unproductive price discriminating. Price discrimination is not a product that gives utility directly. So for welfare considerations, only the cost of producing the actual good matters.

[Gabriel]

Michael,

I agree but maybe, for the sake of discussion, we could consider constrained optima, constrained on the presence of monopoly?

AJE,

What I was trying to say is that I'm not sure I know which transactions would bring about perfect discrimination (their cost aside for the moment).

[Gabriel]

Is this thread dead? I was still looking out for replies...

[aje]

I'm not sure I know which transactions would bring about perfect discrimination (their cost aside for the moment)

Individual negotiation - e.g. second hand cars, antiques, houses

[Cheap Jordan 1]

Thanks for sharing. This website is to I too have to help. Very good.