So why exactly does almost ALL tech live in Silicon Valley?

Ownership and ideas

Critical mass there is obviously akin to clustering. But they are taking it a stage further, in that they're not arguing that it's just more convenient or efficient, as in clustering, but that there's got to be some minimum amount for it all to work at all.

Obviously, if this valuable knowledge is the preserve of some priestly caste then that caste is going to be pretty well rewarded. Yet as this paper goes on to discuss there's more to it than that. We should, with contribution goods, start to see that priestly caste not just being well rewarded workers but actual owners of the firm itself (Terence Kealey):

The conversion of scientific knowledge into new tradable goods and services confronts obvious transactional difficulties between scientists and technologists, technologists and entrepreneurs, and entrepreneurs and financiers.

Cooperation between these elements entails high costs of transacting and is likely to involve the formation of firms with internal labour markets and specially designed incentive arrangements to mitigate them.

Hansmann’s (1996) proposition that ownership rights tend to be assigned to the group that faces the highest transactions costs might suggest, for example, the development of scientist-owned firms or firms with significant control rights in the hands of the knowledge creators and users.

It's a standard of this part of the economic literature that if someone has some asset (and we assume that certain forms of knowledge are such assets) that is essential to the operation of the organisation or firm then they should be the or an owner of the organisation or firm.

Please note that this isn't about desserts or it being just. This is purely about organisational function.

Imagine some startlingly original idea: you can't tell people what it is in great detail because they could just walk off with it. But no one can invest in the idea without knowing what it is and thus judging its likely success. We thus have one of Sir Terry's imp arses to deal with. The solution to which is to bring the idea generator into the firm as a principal. Which rather neatly explains why those nerds are being granted vast chunks of the firms they run even as they gobble up the venture capital.

Or, as Paul Walker, one of the economists working on this sort of thorny problem, puts it:

Brynjolfsson [a US academic, and Director of the MIT Center for Digital Business] considers a situation where an scientist-entrepreneur has some expertise needed to run a firm but no value can be created without both the knowledge asset of the scientist-entrepreneur and the physical assets of a firm.

He assumes that no comprehensive contract can be written between the entrepreneur and the firm. If the scientist-entrepreneur does not own the firm and he makes an investment in effort and creates value, he can be subject to hold-up by the other party since he needs the firm's physical assets.

If the scientist-entrepreneur owns the firm then clearly the hold-up problem ceases to exist.

The most obvious interpretation of Brynjolfsson model is as a model of a labour-owned firm (scientist-owned firm in this case). Brynjolfsson argues that it is optimal to give the entrepreneur ownership of the physical assets of the firm since he has information that is essential to its productivity.

This result is obviously just an application of Hart and Moore’s proposition that an agent who is ‘indispensable’ to an asset should own it (Hart and Moore 1990).

Here, firms are owned by the indispensable human capital (a scientist), or, as is more usual, by a small section of the human capital, e.g. a partnership between a number of scientists.

It's fair enough to consider at least the R&D part of the Valley as being scientists rather than simply engineers, so the model does seem to have some value to me.

However, please note something important here. No economist is going to go around shouting that his model solves everything: economists have a name for people that do that and it's “non-economists”.

A model explains only those aspects of reality that the model was set up to explore. So the mapping of any one model over reality is never going to be one to one.

We do think that this model, where knowledge is a combination of good and also where the asset is really the knowledge holders themselves, explains some aspects of why the top end of the tech business is so concentrated in Silicon Valley and also why people get paid so blindingly much.

But nobody is arrogant enough to suggest that it explains all behaviour in the industry, nor that it even explains all about the clustering and the incomes. Only that it shines a light on some of it. ®