Banking and the law of large numbers
14/11/2011 § 3 Comments
From my point of view, liquidity creation, duration transformation, and simple diversification are all attempts to make the law of large numbers work for us. They are largely successful attempts not to buy liquidity, immediacy, and insurance from those who want to sell them, but rather to create them out of whole cloth–via clever applications of the principles of probability.
I probably liked it because probabilities and blackjack intrigue me. Relying on having enough transactions to keep actual claims on reserves fairly close to predicted, human society has figured out how to grease the wheels of commerce.
I wanted to add two things. What I remember of Doug Henwood’s account of Marx’s critique is that the problem is with proliferation of financial instruments. It isn’t fractional reserve per se, it’s the continual development of ever-more complex and opaque fractions of fractions. The recent experience with mortgage-backed securities bears out the critique. First, the investors in CDOs didn’t know what was in the bundles they were buying. Secondly, what people claimed was in the bundles of mortgages actually wasn’t, because the necessary legalities hadn’t been observed.
That gives rise to the second point. Fractional reserves is another case in which the concept is great, but one can take it too far. I don’t hand out pieces of paper saying ‘backed by Bill’s house’, but the bank effectively is. Somewhere between the two is a social optimum. Searching for that sweet spot of maximum liquidity, immediacy, and insurance (with appropriate preference weightings) has proven to be a messy affair.