May 09, 2003

Transaction cost economics

(Government health warning: involves lengthy technical arguments about economics: read at own risk)

I said a while back that I wanted to respond to Stephen Karlson's arguments about transaction cost economics (TCE). Professor Karlson provides a nice summation of the basic argument of TCE when he says

"I suggest, to the contrary, that a relatively simple notion, institutions evolve to reduce transaction costs, will provide much more understanding of the evolution of cooperation in general, including culture, than continued interest in the transgressive and the bizarre."

Almost certainly, the vast majority of people studying the New Institutional Economics would agree with the core statement of this sentence, the notion that institutions come into being and change over time in order to reduce transaction costs. I don't, but I'm in a small minority. Others in this minority include Jack Knight (whose ideas I steal wholesale in this post), Jean Ensminger, Jim Johnson, Margaret Levi, and sometimes Doug North, depending on which side of the bed he's gotten up on that morning.

The minority argument goes something like this. To say that institutions evolve in order to reduce transaction costs, is to make a functionalist argument. It's to say that institutions come into being in order to meet general social needs, and more specifically to facilitate exchange. But economists have a notoriously cynical take on human nature; they believe that humans are motivated not by broad social imperatives, but rather by their individual self interest. This isn't to say that the two can't coincide; as economists since Smith have argued, individual self-interest should lead to socially optimal outcomes in free markets. But where's the "free market" in institutions?

The alternative perspective suggests that in the absence of perfect competition, we're likely to get bargaining, of the kind that can be modeled in simple mixed-motive coordination games, wherever there are potential gains to cooperation. In other words, everyone may recognize that there are benefits to cooperation - we're all better off if we work together. But even so, we are likely to have different preferences over which kind of cooperation we want, as long as different forms of cooperation have different distributional benefits. This will apply to choosing institutions too - as institutions affect both whether cooperation is possible in the first place, and the conditions under which the proceeds of cooperation are shared. In this perspective, actors are driven less by the desire to create institutions to reduce transaction costs, than by the desire to create institutions that provide them with the greatest distributional benefits. They will prefer institution A, which gives them a bigger slice of a smaller pie, to institution B, which provides them with a smaller slice of a bigger pie, as long as their payoffs from institution A are bigger in absolute terms than their payoff from institution B. In other words, when individual self interest clashes with overall social benefits (such as those from institutions that reduce transaction costs), individual benefit wins.

Under this set of assumptions, what institutions are likely to be chosen? Most probably, those institutions that give the lion's share of distributional benefits to more powerful actors. Simple game theory suggests that a good index of power in bargaining games is the "breakdown values;" i.e the payoffs that players get if they fail to reach agreement. Those actors who have many other options in the case that agreement isn't reached will have greater bargaining power than actors who have few other options, because the former can credibly threaten to break off negotiations if they don't get what they want.

There are circumstances under which individuals may end up coordinating on the superior institutional equilibrium anyway. First, it may be that the "best" equilibrium (say, in terms of reducing transaction costs) is also in the interests of the most powerful actors. Alternatively, actors may not have clashing interests over which equilibria are chosen (it's a pure coordination game rather than a mixed motivation game). In either case, actors have no reason not to coordinate on the best equilibrium. Second, it may be that there aren't real power disparities between different actors (they have similar breakdown values), so that no actor can credibly threaten to break off negotiations if she doesn't get her way. Third, it may be that weaker actors can offer side-payments to stronger actors in order to get them to coordinate on superior institutions.

The point isn't that these are impossible; but that each possibility involves quite restrictive conditions. I strongly suspect that these conditions aren't very often met in the real world. This may be my bias as a political scientist (if the professional deformation of economists is to look for efficiency, the professional deformation of political scientists is to look for power relations). But there are an awful lot of institutional arrangements out there that demonstrably have very little to do with efficiency or transaction cost reduction, and an awful lot to do with furthering the interests of social elites. Just look at the pervasiveness of corruption in various societies in the developing and developed world, at the practices of authoritarian regimes, and so on. Or, if you want to be tendentious (and I do), you could point to the Bush administration's eagerness to push tax cuts that disproportionately favour the rich, and don't do very much to improve economic growth in general, if they in fact do anything at all.

Posted by Henry at May 9, 2003 08:46 PM | TrackBack

Actually, you are writing far more clearly than Jack often does (actually did—he has improved immensely).

My only complaint about your example is not going to the next step. It appears that many in the Bush administration would like to institutionalize their preferences so that even if they lose direct control over the institutions for budgeting, structural deficits will institutionalize lower spending for many years except on debt payments.

Posted by: ArchPundit at May 11, 2003 07:13 PM

Good point on the institutionalization of preferences - something I should have spelled out and didn’t. Thanks. On Jack and clarity; I’ve cheated a little by leaving out the complex part of Jack’s theory - how informal institutions come to instantiate these power relations, even in the absence of central mechanisms. This is where he gets hard to understand; in part because the processes involved are so tricky to capture on paper. We’re revisiting this set of issues at the moment in a paper we’re writing on the broad theoretical relationship between institution-induced expectations and trust.

Posted by: Henry at May 11, 2003 09:02 PM

On Bush and the deficit, it’s not clear who’s preferences are being institutionalised - see my stuff on the’responsibility of being responsible’ over at Bonobo Land.

On more intersting questions about path and non-path dependant processes, I’m only a layman here, but hasn’t this got something to do with deterministic and stochastic processes. And isn’t it the case that not all path-dependant processes as a family are so closed and predictable as the Polya Urn stuff. Some of the processes need to be run to actually find out what’s going to happen (as Woolfram has found out with his CA toys). Stochastic processes assume you can predict outcomes with degrees of probability using systems of differential equations. Have I got it about right?

For the participant there may well be dificulties in principle in deciding in which type of system they are participating. If it’s an alogrithmic type structure their may also be a computational problem in running through an infinite number of moves in finite time.

What I’m really getting at, is that the unintended consequences bit is really important in looking at social strucures. In your Polya Urn paper you say:

“Jack Knight (1992)defines power as involving the ability to constrain another’s choice set. Under
this definition, we may see how the “distance” of particular urns need not be a happenstance feature of social structure; instead, it may reflect the power of external actors to make it more costly for A to adopt certain solutions.”

Of course many players will try to do this but I had understood strategic uncertainty (following Arthur) as implying the impossibility in principle for those ‘in power’of knowing whether constraining A from adopting certain solutions would in fact turn out to benefit them. Only after the programme has been run will they know.

From the sublime to the ridiculous, case in point, the euro. It is a constraining institution, which in principle imposes lock-in. But how many surprise ‘white balls’ need to come out of the urn, before players start to draw conclusions and attempt to jump ship.

ie the really interesting thing about paths is not about identifying how we get locked into them, this is the easy bit (although as you note many are still in denial on this), the interesting thing is about how, and under what conditions, we change them.

I tend to follow Per Bak, and think in terms of self-organising criticality and cascade processes.

Finally, I once read a piece by the late Rene Thom where he suggested that one whole bag of problems with the model of science we use comes from our inheritance - via Bacon and Descartes - of an idea of nature as a passive agent. But what if nature were an active participant - the relation bacteria-antibiotic comes immediately to mind - and the rules of the game were constantly changing? Thom’s model seems to satisfy a lot better the intuition that science is an on-going, but in fact ‘objectless’, process.

See ya.

Posted by: Edward Hugh at May 12, 2003 06:49 AM

Oh, yea, I forgot to say: I think you’re absolutely right about the importance of redundancy.

Posted by: Edward Hugh at May 12, 2003 06:57 AM
Post a comment

Remember personal info?