Stephen Williamson suggests here, here, and here that blog readers browse his recent paper, Liquidity, Monetary Policy, and the Financial Crisis: A New Monetarist Approach.
Here are some thoughts I had upon reading it.
Williamson sets up an imaginary universe in which there are two types of people who live and produce in differing market environments. The various frictions that these two types must bridge in order to purchase each other's production leads to a role for fiat money, and later banks.
The first environment is a Walrasian centralized market in which trade is easily executed. The second is a decentralized market. Trade is difficult in the decentralized market because buyers and sellers meet each other in a haphazard and anonymous manner. Those who inhabit the centralized market must trek over to the decentralized market so as to trade for consumption goods. Unfortunately for them, they can't bring the goods they have produced in their centralized market habitat since these are perishable. Adding to this already inconvenient proposition, they cannot use credit to make purchases in the decentralized market. This is because trade in the decentralized market is anonymous and cannot be tracked, and therefore the inhabitants of the decentralized market won’t take the credit of the inhabitants of the centralized market. This leaves only one way to transact – fiat money. Or, as Williamson points out, they can use bank deposits.
I find Williamson’s banks a bit odd. The denizens of the centralized market “make deposits in the form of goods” at a bank. The bank then purchases currency and interest-bearing assets with these goods. This is odd for a few reasons. In the real world, individuals don’t deposit goods at banks – they deposit currency. Why abstract from the real nature of banks in such an odd manner? Secondly, Williamson has already defined all consumer goods as perishable, presumably to prevent inhabitants of the centralized market from taking their goods over to the decentralized market and bartering them. But if goods are perishable, how can banks accept a deposit of goods only to on-sell them? By definition, haven’t they already perished?
Before they must make the trek to the decentralized market to satisfy their desire for consumption, the inhabitants of the centralized market are informed of what sort of payment medium their anonymous trading partner will require: cash or deposits. If the latter, they withdraw the necessary cash from the bank and bring it over to the decentralized market. If the former, they bring their cheque book. Using this model, Williamson goes on to demonstrate all sorts of implications, including using it to understand the credit crisis.
Back to Williamson’s banks. These aren't actually banks, at least not modern ones. Rather, I’d call them “banks of deposit.” Whatever is deposited remains there, although it might be swapped out for some other asset. Alternatively, think of Williamson’s banks as money-market mutual funds. People invest their savings in money market fund, and the fund in turn invests these funds in some pre-existing asset. The point being is that both banks of deposit and money market funds are purely passive – they only create deposits when they are approached by an investor who already has liquid funds.
But modern banks – the ones who were in the credit crisis – actively create credit. Sure, you can bring your funds to the bank and hand them over to the teller in return for a deposit. But most of the deposits that banks put into circulation are not provided to those who have liquid funds or already-existing consumption. Rather, those who have neither of these obtain deposits in the form of a loan. Real banks don't need to wait for a money to arrive in their till before creating a deposit, they create deposits out of nothing. Because banking is much more complex than Williamson’s model seems to propose, it's hard to know how relevant his model is to a world which spawned the 2008 credit crisis.