By Dan Kervick
Provoked partly by some recent posts by Paul Krugman, which seem to imply that understanding the institutional structure of the banking system is irrelevant to gauging the effectiveness of the monetary policies implemented by the central bank, questions have arisen again about the relationship between bank lending and bank reserves. One of the issues raised can be framed by asking, “Do banks lend their reserves?” And as with so many questions in economics, the answer to this question depends on disambiguating the question, clearly distinguishing parts from wholes, and avoiding fallacies of composition and division.
An individual bank certainly can lend out its reserves. The total reserves of a Fed member bank consist of the balance in the bank’s reserve account at the Fed plus the sum of its vault cash. When a bank customer borrows money from a bank, that borrower may ask for all or part of the loaned amount in cash. If that happens, the bank will remove some cash from its vault and give it to the borrower. Suppose the amount borrowed is $10,000 and it is all taken in cash. Then the bank books a new loan asset worth $10,000, and concurrently books a $10,000 reduction in its cash assets. The net balance sheet effect is zero. (There are different ways in which the interest on the loan can be handled in the accounting, and so I’m leaving that out.)
But in most cases the borrower will take the borrowed funds on account. The bank will credit the borrower’s account with $10,000, which in turn represents a liability of the bank. Again, the bank will book a $10,000 loan asset, and so the total balance sheet effect will be zero. In the first case the balance sheet effect was zero because a reduction in cash assets was offset by an increase in loan assets. In this second case the balance sheet effect is zero because the increase in the bank’s loan assets is offset by a corresponding increase in the bank’s liabilities.
But even if the borrower takes the loan in the form of the deposit balance, as soon as the borrower begins spending the money in the account, some of the bank reserves will likely leave the bank. It is possible that they would not leave the bank, if the borrower’s spending all goes to depositors at the same bank. In that case, the result of the settlement and clearing of the payments will just be a reduction in the bank’s liability to the borrower and an increase in its liabilities to the payees. But in most cases, many of the people the borrower pays will be depositors at other banks. The settlement of these payments will thus require a settlement between the two banks, resulting in a Fedwire transfer of reserve funds from the borrower’s bank to the payee’s bank.
So yes, individual banks can lend their reserves. But the more important question is what happens to the reserves of the banking system as a whole in response to expanded bank lending. And in this case the answer to the question is that those reserves will not appreciably change.
Consider the first case, where the borrower has taken the loaned amount in cash. As the borrower proceeds to spend the $10,000 in cash that was removed from the first bank’s vault at the time of the loan, the businesses who receive these cash payments will make routine deposits of their cash receipts at their own banks, at which point it goes back into the cash vaults in the banking system. Similarly, if the borrower’s funds are in the form of a deposit account balance, then as the borrower makes payments by check, debit card or other instrument against that account, the clearing of the payments will result in reserve transfers from the borrower’s bank to the reserve accounts of other banks.
The end result is that an aggregate increase in commercial bank lending is unlikely to diminish in any appreciable way the total quantity of bank reserves; it just changes the pace at which those reserves circulate from bank to bank. People sometimes look at the large current excess reserve holdings in the banking system, and say, “Why aren’t those banks lending their reserves out?” But that question bespeaks a misunderstanding of the way the banking system functions.
If banks in the aggregate were not already carrying excess reserves prior to an expansion in bank lending, then we would see reserve balances going up, not down. An increase in deposit account balances is going to mean an increased volume of interbank payments, and increased bank demand for reserve account liquidity to make those payments. The added liquidity demand would prompt bank liquidity managers either to sell more Treasury securities to the Fed, or “loan” more securities via repurchase agreements, or (less profitably) borrow dollars from discount window. Reserve balances would then go up.
If, however, banks are carrying abundant excess reserves, it is possible for the banks in the aggregate to expand their lending and create more deposit liabilities without acquiring any new reserves at all. In that case, we would not see reserves decreasing; we would just see a decrease in the ratio of reserves to deposits and reserves to loans.
As it happens, though, we currently have a situation in which there are abundant excess reserves, and in which loans are increasing while reserves are also increasing. This is due to QE, which consists of Fed purchases of treasury securities, agency debt and other mortgage backed securities. We are therefore also seeing large injections of additional reserves along with reductions in bank holdings of treasury and agency securities.
Cross-posted from Rugged Egalitarianism