Bank capital allocation is normally a pretty dry subject – and that is to people who work in the area. To the rest of you it must be sleep inducing. That is, until you consider the real effects that it can have.

This is a follow on piece to yesterday’s article.

On a macro level, of course, the differences in the capital charges (between Economic, Regulatory and Total) drive some really odd behaviour. The effects on an unregulated institution (or one without effective capital regulations) are that, in the absence of regulatory restrictions, they can hold assets with less of a capital charge than a bank. This means that, if the bank lends the money out in the first place, the loan is worth more to an unregulated institution than to a bank. Selling the loan then means that both parties make money on the deal. It also means that it is worth trying to get the home loan into a different regulatory capital category – say as a traded instrument rather than a loan.

Banks therefore bundle(d) these up into corporate vehicles (securitisations) and tradeable instruments (CDOs etc.). If the bank can get the loans off their books then they can make the money from the initial fees from lending, selling the loans and also from managing the loans once they are off the bank’s balance sheet.

If the loan is sold and the bank retains little or no credit risk (i.e. if the loan goes bad then the bank does not have to pay) then the incentives are simple – write as many loans as possible regardless of the credit risk provided you can then sell them. If housing prices go up then there is no problem at all – everyone (including the borrower) stands to make money. The bank gets a fee for originating the loan, the purchaser of the loan makes money off the interest and, if the loan goes bad gets the remaining funds from the sale of the house and the borrower gets to live in the house and may well be given some money at the end when it is sold.

If prices go down, though, it is another story. The originating bank still makes money, the borrower (in the US at least due to the without recourse lending) gets to live in the house until foreclosure but the loan purchaser gets stuck with any losses, as has happened in the US over the last year.

The point here is that the excess and disproportionate requirements of regulatory capital is one of the main things driving this. People will always try to make money and should be (justly) condemned for doing so when it involves fraud or negligence, but the system itself should not create the incentives in the first place.

Creating a CDO or securitisation costs money. There are legitimate reasons for them to exist (some banks are better at lending than borrowing and some investors want to be able to lend specifically for housing for example), but having them drive adverse lending behaviour can cost vast amounts more money. Reducing the incentives for them to exist means fixing the regulatory capital weights.

Until that is done merely requiring more capital in banks will increase, not reduce

, the problem.