A bit more on Coasian hells

As a follow-up to this 8,000 pageview hellbeast, here’s an effort to dissect its central argument a bit more. I see maybe three points within it.

First, there is an argument from transaction cost. This is simply that all those contracts need some lawyering. In itself this is probably the weakest of the three, and the one most likely to be ignored because “the Internet!” or “Blockchain!” or whatever. A stronger version is that the number of pairwise contracts that need looking after scales with the square of the size of the organisation, and therefore even low transactions costs will become a problem for fairly small organisations. (Thanks, reader, for correcting my maths!) Even so, if it’s just the cost of drawing up documents, it probably shouldn’t matter other than as a gross-error check.

Second, there is an argument from time. Lawyers will tell you that a transaction implies a contract, but I would draw a distinction between a transactional relationship and a contractual one. In important ways, the implicit contract involved in a transaction is over as soon as the transaction is. A genuinely contractual relationship binds the parties for part of the unknowable future. Both parties therefore take on risks, commit to actions in advance, and have to consider how they would alter the contract if that became necessary. This means either that someone can make decisions on a discretionary basis, or that the contract will eventually become too restrictive. To put it another way, if it lasts, it will evolve towards organisation, or else it won’t last. This is actually much more important.

It’s possible that there are three time-frames here: transactional, contractual, and organisational or institutional time. The first is meant to be instant, or as short as possible. The second is a longer but fixed period. The third is…well, it’s meant to stick around. A lot of trouble arises from not respecting these distinctions. We’ve been talking about trying to substitute contracts for organisation, but selling contracts as if they were transactions, or relying on a transactional relationship for a contractual or institutional one, is also pretty terrible.

Third, there’s an argument from uncertainty/information. At the time of signing, you don’t know what will happen in the future. You don’t even know what you will do in the future as well as you think you do. And – this is crucial – it may well be impossible to state literally everything in the contract that the parties would want to. The proof is simply that the parties to the contract don’t have perfect knowledge, and therefore can’t know all the things they would want to put in it. Incompleteness is inevitable, and it increases with the length of the proposed contract.

Points two and three converge. We can’t specify the future. We also cannot specify terms we don’t know. This suggest either that we should ignore them – reduce the problem to a succession of transactions – or else internalise them into an organisation. The second means putting up with bureaucracy while the first means putting up with instability and also deskilling, as information that can’t be contractualised vanishes into air.

1 Comment on "A bit more on Coasian hells"

  1. I think there’s a related, but extra element, that comes up towards the end of this piece:


    Speaking of complex things, you can’t control – and further, can’t write contracts about, something you don’t have the institutional capacity to understand. And this is where I come down on the side of (at least some level of) internalisation. The only way to construct/maintain an institutional capacity to understand is to construct/maintain at least some level of institutional capacity to implement it yourself.


Leave a Reply

Your email address will not be published. Required fields are marked *