Jose Mourinho yesterday issued a challenge to methodological individualism. When asked about Diego Costa, he replied (15'30" in): "I like my team more than my players."

This is a very reasonable attitude to have to a team that includes John Terry and Diego Costa. But it expresses an important truth - that collective entities can have a property (likeability) that its component individuals do not.

This corroborates something I've said before. The most successful football managers have a collectivist mentality: the decline of British football managers and rise of individualism might be related.

But there's another point here. Mourinho's paradox applies widely.

In football, for exmaple, good individual players do not add up to a good team; think of England in the 00s. Conversely, mediocre individuals can make a decent team; Stoke have achieved mid-table respectability despite have several players who refute the hypothesis that Neanderthal man has gone extinct.

The same is true of financial markets. Sometimes, rational individuals can produce a market than looks irrational*.This happens if rational but ill-informed traders chase noise, or in beauty contest-type behaviour, where individuals worry that others will worry. And conversely, experiments in agent-based economics show us that stupid individuals can produce rational markets. "Mr Market" is a misleading metaphor.

The same can be true in other contexts.Granovetter's riot model (pdf) shows us that violent disorder can emerge not because all individuals are violent but simply if large numbers imitate a few others. Adam Smith's invisible hand theory says that a benign social order can emerge from individual self-interest. Marx's theory of exploitation shows how Christian Victorian gentleman can produce a viciously inegalitarian economy. Schelling's spatial segregation model shows how ethnic segregation can happen even if people aren't especially racist. And I suspect that non-sexist individuals can also generate a patriarchal society.

All these are examples of what Mourinho said - that aggregates can have properties which individuals do not. This can happen - as in my examples - because of emergence. Or - as in Mourinho's case - it can happen because a coach consciously shapes a team. And in other cases, it might be a mix of the two: Irving Janis's examples of groupthink, I suspect, arise partly from bad luck and partly bad leadership.

All this is a challenge to methodological individualism; what's the point of studying individuals' motives and rationality if these are not necessarily related to social phenomena?

I suspect the question can be answered. Careful studies of actual individual behaviour, for example, can tell us how prone they are to the sort of peer effects that can cause aggregate fluctuations or mispricing. And they can also help in the design of better incentive mechanisms: they show us, for example, that big financial incentives can backfire.

I don't have conclusive answers here. My point is rather that, not for the first time, football managers have done a better job of drawing our attention to big issues of social theory than so-called serious thinkers.

* I'm glossing over the question here of whether we can apply descriptions of individuals (such as rational or irrational) to impersonal things.