Coasian Hell: Back Into Hell

The Coasian Hell series were some of the most popular posts ever on this blog, and I’m quite proud of them, so how about a follow-up?

One Coasian objection to the concept might be that the actors in the system can trade their way around to an efficient outcome through side-bargaining between them. This is the argument, often referred to as the Coase theorem, from Ronald Coase’s The Problem of Social Cost, while most of the Coasian hell concept is based on his The Nature of the Firm.

Steven Medina’s paper The Coase Theorem at Sixty goes deeply into the intellectual history here. As it turns out, Coase didn’t formulate it as a theorem, didn’t call it that, and as so often, included it in the original paper to make a very different point than the one it’s usually put to work for. Hey, I called that blog post In The Eternal Inferno, Fiends Torment Ronald Coase With The Fate Of His Ideas for a reason.

Rather than arguing unconditionally that private bargaining would arrive at an efficient solution, if it wasn’t for those pesky transactions costs, Coase’s point was that this was unrealistic and in practice everything would come down to the appropriate design of institutions. Even if the answer was a market, this begged the question. What kind of market? How would it work? What consequences would its workings have? Well, that’s what I’m talking about!

Something interesting I took away from Medina’s paper is that the idea of transaction costs is under-specified and poorly theorised. Economists following Coase tended to model them as if they were a simple ad-valorem tax, like a fixed percentage commission. If this helped with the maths, and made it possible to use the substantial theoretical framework they already had for analysing the incidence of taxation, it helped mightily with the institutional economics, in that it pretty much avoided thinking about it.

As such, it emptied the original ideas of much of their substance. If we stick with a relatively limited definition to begin with, imagine that the transaction cost is something like a bid-offer spread rather than a flat-rate commission. Being a forced buyer in an illiquid market where spreads are wide and prices often post far from whatever mid-point might be defined is one way of saying that you’re being exploited, and that the result of such a trade is unlikely to represent any kind of efficient optimum.

Opening up the definition further, though, gives us more. The costs of transaction can be thought of as including the information problems I was talking about in the original series. Also, one player’s costs are another player’s revenue. One way of looking at the Coasian hells series is that these are cases where the transactions costs are more interesting than the market itself, and rather than understanding them as a tax on market activity, we should think whether the market exists to generate transactions revenue. This question needs to define transactions cost/revenue broadly – if the game is find-the-lady, the problem is not the house edge.

A really interesting point in Medina’s review is that some experimental econ researchers in the Coasian tradition found that lowering the costs of transaction, in the narrow definition, could have perverse effects, in that the actors in the market began to cycle from one solution to another without finding a stable equilibrium. This reminds me of the distinction between transactional, contractual, and institutional timescales in this post. It’s not obvious, for example, that the enormous reduction in transactions costs in finance has improved anything, although it has enormously increased the volume of transactions and the size of the sector.

It also reminds me of the criticism, which Medina quotes, that the zero transaction costs world would be extremely strange. There would, for example, be little point using Coasian insights to inform your search model of the labour market, as zero transactions costs is equivalent to assuming equilibrium, so all markets would instantly clear and therefore nobody would be searching for a job. As with most economic arguments that assume all markets are in equilibrium, it essentially assumes away everything interesting.

What I’m driving at here is that as the insights in The Nature of the Firm were transformed from being a new means of inquiry to a positive recommendation that things ought to be more like that, and all steps towards that goal were themselves improvements, the insights in The Problem of Social Cost were also so transformed, into a campaign to push towards a more transactional world. Rail privatisation both tried to transform the railways into a nexus of contracts, and hoped that side-bargaining would sort out the problems that resulted. Fascinatingly, though, this effort to actualise the zero transactions cost environment generated a ton of transactions costs, and somebody got paid with that money. Making the Coasian hell heat up also required someone to stoke the furnaces with subsidies from the taxpayer.

Perhaps the effort to move towards the ZTC environment foundered on the problems of the firm as bundle of contracts, or vice versa, but I think it would be more accurate to say that the whole project foundered on the appropriately submerged problems both papers identified but we forgot to remember. The theorem, as it emerged from the intellectual machinery of the economics profession, argued that given minimal property rights and zero transactions costs, actors could bargain their way to an optimal solution. Sixty years on, perhaps the question is how the transaction costs environment determines what your effective rights actually are, and whose fault that is.

Coase, of course, originally pointed out that the question was to choose the best institutional design, which wouldn’t necessarily be a market.