One of the big questions in macroeconomic forecasting now is: what's going to happen to business investment? For me, this highlights the importance of thinking about heterogeneity rather than representative agents.

On the one hand are those who believe increasing confidence and high replacement demand will cause a big rise in capital spending: the Bank of England seems to side with these. On the other are those, like the guys at Fathom Consulting, who fear that credit constraints will hold it back; this weeks figures showed that bank lending to non-financial firms fell again in January.

Here is where heterogeneity comes in. Profitable investment opportunities aren't distributed evenly across firms; some have them, some don't - which is why some thrive and others fail. If these opportunities are concentrated among credit-constrained firms then investment will be weak. But if they are concentrated among cash-rich firms - and non-financial firms are sitting on £284.8bn of UK sterling bank deposits - then investment could rise a lot.

You might object that only a few firms are credit-constrained; the last CBI survey found only 4% of manufacturers saying that an inability to raise external finance limits their investment. But this is a survey of existing firms; it's quite plausible that potential new firms are more constrained.

You might also object that capital markets should solve this problem, because credit-constrained firms (or potential start-ups) can sell to cash-rich firms. But two things make this only an imperfect solution: asymmetric information and commitment problems. There's a big difference between having a good idea and persuading conservative men in suits to buy it. And working for others, especially in a bureaucratic culture, might drain one of the energy required to make such a project work.

Heterogeneity, then, matters. Here are three other ways it does:

1. If an adverse demand shock hits cash-rich firms and households, the effects might be more modest than if it hits highly-indebted ones, where it might trigger a cascade of bankruptcies.

2. Paul Geroski and Paul Gregg found that in the 1989-91 recession, the worst-hit 10% of firms accounted for 80% of the drop in sales and 85% of the cut in employment. This tells us that recessions are about what happens in a minority of firms.

3. Right now, household debt, at £1.4 trillion, is equivalent to 1.3x disposable income. But the distribution of this matters. An economy in which everyone had debt of 1.3x income - equivalent to a mortgage of around £30,000 for a household on median income - would be a different thing from one in which two-thirds had no debt and one-third debt of almost 4x income.

And here's the thing. We know much less about the heterogeneity of firms and households than we do about aggregates and averages. It could be that the latter helps to give us a pretense of knowledge, rather than real knowledge.

Another thing: Yes, I know all this should be familiar to Austrians who have always been sceptical about macro aggregates. But it bears repeating.