This post will seem simple to some, but I want to correct a slight confusion I’ve seen over the last several years in the economics blogosphere. (I was motivated to write because of an exchange with Nick Freiling, who loves the Austrian School but thought I had made a basic mistake in a recent piece I wrote about the Federal Reserve’s policies.) Specifically, Freiling and many others have challenged the standard claim that commercial banks lend out reserves when they make loans to customers. The critics argue that since the public will generally end up depositing their checks with their own respective banks, the granting of loans will merely rearrange which banks hold certain levels of reserves, but the banking system as a whole can’t “lend them out” because there would be nowhere for them to go. Hence, the critics allege, the talk of the Fed (say) raising the interest rate that it pays on reserves in order to discourage lending is nonsense; in Freiling’s words, there is (allegedly) no tradeoff between loans and reserves.
This argument from the critics is wrong. It rests on a confusion between micro-incentives and system-wide outcomes. In particular, the interest rate that the Fed pays on reserves can most definitely affect the willingness of commercial banks to make loans on the margin.
Before jumping directly into the issue, let me start with an analogy with actual currency held in people’s wallets or purses. (I see Nick Rowe thought of the same analogy last summer.) Forget about banks. Suppose there are $100 billion in actual currency in the economy, held by a population of 100 million people, and that this is the only money that these people use. That means on average each person holds $1,000 in currency.
Now notice that in this world, it is impossible for the community as a whole to get rid of cash and into other assets or goods. For example, if Bill wants to spend $100 on a TV, then Bill’s cash balance goes down by $100 but the merchant who sold him the TV sees his cash balance go up by exactly $100. Assuming that nobody loses (or deliberately destroys) currency, and that the authorities don’t print more, the community as a whole will always hold $1,000 per person on average, regardless of the willingness to hold cash, prices, expectations of price inflation, etc. This is a simple division problem.
So did I just “prove” that things like interest rates, prices, expectations of future prices, etc. have no effect on Bill’s willingness to spend money? For example, suppose the government’s printing office announces that in exactly one year, it will create an additional $100 billion in currency, which it will drop by helicopter onto the population. Will that announcement have any impact on the economy right now? People might think it would cause a rush away from currency and into assets like real estate or stocks, but isn’t that impossible for the community as a whole to achieve?
The answer is that any individual–faced with the new information–will be less willing to hold currency, and will try to substitute into assets that retain their value amidst inflation. In the act of trying to (say) buy gold with paper dollars, the price of gold gets bid way up. (The gold sellers are aware of the new information as well, and adjust their asking price accordingly.) In the new equilibrium, well before the new currency is printed and dropped on the economy, prices of assets, goods, and services have all risen to varying degrees. People still hold (on average) $1,000 in currency, though its distribution might be different depending on the differing preferences of people in the community.
However, the crucial point is that each person is now holding a “smaller” cash balance when measured by its purchasing power in “real” terms. In other words, even though the average person is still holding $1,000 as before, now he or she is holding fewer “apples’ worth” of cash, or fewer “ounces of gold worth” of cash. Since people ultimately value cash for its ability to fetch goods and services in the marketplace, the purchasing power of a cash balance is really the economically relevant factor.
Now step back and realize what we’ve just shown: Even though it is a trivial tautology to say that the community in general can’t “get rid of” currency balances, nonetheless the micro-level analysis of the demand to hold currency is still quite relevant. When an important element of the demand to hold currency changes, the rest of the economy must adapt in order to restore equilibrium and make the community collectively willing to hold all of the currency in existence.
By analogy, this is how to think about the commercial banking system and bank reserves. If we assume for the sake of argument that the public leaves the same amount of money in the banks (in the form of checking accounts), and if we further assume that the central bank doesn’t add new reserves to the system, then it is a tautology to say that the total amount of bank reserves is constant–the banking system in general can’t “lend out” bank reserves. Nonetheless, the micro-level analysis of an individual commercial bank’s demand to hold reserves versus making a loan is still quite relevant.
In particular, if the Federal Reserve (say) started paying a 2.5% interest rate on reserves rather than the current 25 basis points, this would have a huge impact on the economy. An individual bank would be much more reluctant to advance a particular loan at a given interest rate, if it knew it could earn a guaranteed 2.5% by keeping those funds parked at the Fed. This is relevant even though the reserves in question would most likely end up in the hands of another bank soon after the loan is made. The individual bank doesn’t care about the reserves of the system as a whole; it cares about its own earnings.
In conclusion, although there are certain assumptions we can make under which the total amount of bank reserves in the system are constant, this constant level of reserves can support a greater or smaller amount of total bank loans. The policies of the central bank–including the interest rate it pays on reserves–definitely affect the size of the total loan balances.