"How Darwinian should an economy be?" asks Gilles Saint-Paul.

Econ 101 says: maximally so. If competition is fierce only efficient firms will survive and so we'll achieve an optimum allocation of resources. That's the first theorem of welfare economics.

However, there's an assumption here - that the environment is stable, so that firms that are maximally adapted to their environment yesterday will be maximally adapted today. But this might not be the case, says Professor Saint-Paul. If the enviroment keeps changing, firms that were optimal yesterday won't be optimal today. If this is the case, then fierce competition and strong Darwinian selection will give us instability but not optimality. He says:

If selection is too extreme in the current environment, the firms that are best adapted to a given future environmental change, yet performing poorly in present circumstances, will be very scarce, and it will take longer for the economy to produce a large number of such firms in the new environment.

He gives the example of a cobweb cycle. Low wages will select in favour of labour-intensive firms, but this will lead to higher labour demand and rising wages which select against labour intensive firms. Thus we get a boom-bust cycle of inflation and deflation.

But we don't need theory to see his point. We only need to remember the financial crisis. In the good times, Darwinian selection (in terms of pressures upon CEOs rather than selecting firms) favoured banks who borrowed in wholesale markets and invested in mortgage derivatives. As Citi's Charles Prince said: "As long as the music is playing, you've got to get up and dance." But when the environment changed, such strategies were disastrous. Banks that were well adapted to an environment of high liquidity were wiped out when the environment changed.

In this sense, says Saint-Paul, there can be a trade-off between optimality and stability; strong competition and selection gives us optimality as long as the environment stays the same, but also volatility when it changes. How severe this is will, he says, depend among other things upon the size and persistence of shocks.

I'd add that it also depends upon the nature of networks between firms. The collapse of a few banks had catastrophic effects because these were hubs with tight interconnections with both other banks and non-banks dependent upon credit. By contrast, the failure of (say) Phones4U earlier this year had no macro effects because it was part of a looser network. As Haldane and May said (pdf), "disassociative structures are likely to support a larger number of coexisting banks, and can make the network more robust to random losses."

You might think this trade-off can be mitigated by good management: if managers can see shocks coming they can adapt in advance. In the real world though, this might not be possible; as Ormerod and Rosewell pointed out in 2006, "firms have very limited capacities to acquire knowledge about the true impact of their strategies" - a claim corroborated by the banking crisis.

The point of all this is not necessarily to argue for state direction of the economy. If bosses can't know much, nor perhaps can governments: Orgel's second rule says that "Evolution is cleverer than you are."

Instead, it is perhaps to defend the real world in which market forces don't select very strongly. As Nick Bloom and colleagues have shown, there is a large variation in corporate efficiency around the world. From a conventional point of view, this is sub-optimal. But from a Darwinian point of view, it might not be. As Andrew Lo says, what look like suboptimal strategies might in fact be second-best adaptations which permit survival in a changing environment. And the same lack of selection that allows inefficient firms to survive also supports a diversity of firms which stabilizes the economy in the face of shocks.

We know that cognitive and cultural diversity are good things. So too is ecological diversity.