What is a bank? (highly, highly speculative)
When we are young, we are told a bank is a place you put your money to keep it safe. If you don’t find that hilarious, then this post is probably not for you. In fact, this post is probably not for you simpliciter. Oh well.
Something I’ve been thinking about recently is how we should think about banking. In philosopher-speak, I’ve been wondering about the ‘essence’ of banking. What makes a bank, well, a bank?
First stab: maturity transformation. Banks borrow short and lend long. At the extreme end, they borrow deposits (which have a maturity of zero and must be constantly rolled over) and lend for 25+ years. Banks accommodate the preference of consumers to hold short-term, low-risk and liquid financial assets, and borrow for big long-term projects (buying a house, expanding a business etc.).
But it’s not logically necessary for banks to exist in order to do some of this. Imagine Bank A offers as 10 year loan to Company B. Bank A funds itself in the wholesale market with 1-year bonds. Alternatively, Company B could just offer 1-year bonds. Cut out the middleman!
Now there may be very good reasons why we want someone in the middle. And, indeed, there are. When you have banks transforming the short-term debt into long-term, the ability to roll over the debt no longer has the same direct relationship to the perceived future prospects of individual companies at whatever time their debts happen to mature. This is unquestionably a good thing. But it makes the point that maturity transformation is not really about borrowing short and lending long – the lending long only occurs because we put the intermediary in there. We need to look at what the point of the whole system is, which is investing in long-term illiquid assets.
If we had a world where institutions that are both debtors and creditors were banned, it’s not like maturity transformation would stop. You’d probably just end up with with everyone issuing short-term debt and still going on to buy houses and invest in new factories and so on (but with much higher interest rates). It wouldn’t be pretty, but there’s no logical contradiction in it.
The key service that a bank provides is it pools the illiquidity risk of of all the companies it lends to, who would otherwise have much shorter term debt maturity profiles. A bank effectively provides a kind of insurance against short-term illiquidity by assuming the relevant risk of the entire portfolio; meaning people can buy more illiquid assets than they otherwise would be able to safe in the knowledge they don’t have to roll over their liabilities so often. Since increasing our living standards often requires investing in long-term projects, the provision of this insurance product is hugely important for economic growth. It also is what allows the short-term assets to be low-risk and liquid.
Of course, like any insurance company, a bank is highly exposed to any event that is correlated across all the debtors in its portfolio… like, I don’t know, an increase in demand for money that is not offset by a corresponding increase in supply. You can insure against the fact that there might otherwise be runs on the liabilities of a few companies in the portfolio at any existing time, but you can’t insure against the conditions that will lead to a run on all of them any more than you can insure against a fire that wipes out an entire city. You can’t insure against falling aggregate demand.
This is why banks are so vulnerable when aggregate demand falls. It ruins the insurance model. The tricky thing about insurance is that the reason it is so good is that it changes the decisions we make for the better, by getting others more suited than us to bear the risks that are (for want of a better term) incidental to the projects we want to undertake. But as soon as that insurance vanishes, that project is no longer a good idea. The rational organisation of economic activity in a world where persons and companies have to individually bear the risk of rolling over short-term liabilities is completely different to one in which that risk is pooled together, which is why everything goes haywire when the risk-sharing breaks down.
Banks affect the rational structure of economic activity through pooling together the risk that individuals couldn’t roll over short-term liabilities, and they do this to the enormous greater good of mankind. But you can’t insure against outcomes correlated through the entire risk pool, so you can’t insure against illiquidity risk caused by falling aggregate demand.
Maybe thinking about banks in this way makes it that bit clearer just how vital it is to have stable aggregate demand.