One initial tangential point is that what I am referring to are no loan loss reserves taken by banks. In a country like Canada where reserve requirements were abolished in the 1990s, loan loss reserves might be the only “reserves” that comes up in conversation outside of macroeconomic circles. A loan loss reserve is an estimate of future credit losses on a bank’s loan book — they are holding a “reserve” against the value of assets that are held at book value.
For simplicity, I am not going to attempt to get into the details of payments systems. For out purposes, we can simplify the situation to be that private banks hold deposits at the central bank, and when they need to transfer money to each other, this is done via a transfer of between deposit accounts. If the private bank needs to wire money to/from the central government, the deposit balance is reduced/increased. Note that although the banks will sometimes transact on their own account, they are mainly acting as agents on behalf of their customers. That is, if a customer needs to pay a $100 tax bill to the central government, the bank will reduce the customer’s deposit balance by $100 and then “transfer” the money to the government (which just reduces the private bank’s balance at the central bank).
(If we do want to worry about payment systems, there is a new entity introduced into transactions. Banks — both private and the central bank — send payments to each other via the payments system. However, net transactions are supposed to net out to zero, and so if we look at end-of-day balance sheets, the payments system is not supposed to be visible. If it is, something has probably gone horribly wrong.)
Dealing With Misunderstandings
Once we realise that “reserves” are just a type of asset held by private banks, a lot of points of confusion should be dissipated. The most typical problem is the idea that “banks lend reserves” to non-banks. Since non-banks are normally precluded from holding deposits at the central bank, there is no way for a bank to transfer a “reserve” to that customer. (Banks can “lend reserves” to each other in inter-bank markets.)
Another issue I see is the belief that banks need reserves in order to deal with depositors’ transfers elsewhere. In particular, this comes up in the context of extending loans — since loan amounts typically are used to buy things, banks allegedly need to hold bank reserves ahead of loan extensions.
I would argue that this is the result of a hidden assumption: money is a something resembling a commodity, and needs to be held before transfer. This misses credit relations. Payments systems typically allow the equivalent of overdrafts: banks can run deficits during the day. They only need to settle up to desired balance levels by the day’s close. If it looks like the bank is going to have a balance that is either “too high” or “too low,” it undertakes offsetting transactions in money markets (including inter-bank markets) to steer the balance back to the desired level.
If we look at the Canadian pre-2020 system, it is entirely possible that the “desired balance” at the end of the day is $0. This means that if look at end-of-day balance sheets (the only ones that are published), settlement balances would always be close to $0 (with misses that were relatively insignificant compared to the size of bank balance sheets).
Banks do not need to hold settlement balances (reserves) to meet cash inflows/outflows, they just need the ability to generate cash in money markets (typically by having liquid assets that can be sold or borrowed against).
Required Reserves Do Not Provide Liquidity
Let us imagine we are in a country that has reserve requirements. (With the abolition of reserve requirements in the United States, this might have to be a developing country.) Regulatory requirements vary from country to country. But my guess is that most follow a system similar to the old system in the United States: banks are required to hold a certain amount of deposits at the central bank, with the amount a fraction of the bank’s deposit liabilities. For example, if the reserve ratio is 10%, for every $100 in (certain classes of) customer deposits, the bank needs to hold $10 in deposits at the central bank.
However, getting a handle on a bank balance is sheet is not easy, give the rapidity of transactions. In the United States, this meant that the reserve requirements were based on deposits as of certain date in the recent past. Until the next reserve calculation date, the amount of required reserves are fixed.
Does this necessity to hold required reserves help their ability to meet customer outflows? Of course not! Imagine that you held $2000 in deposit at your bank, and then the government decided it was concerned about your liquidity, and decreed that your bank balance is not allowed to drop below $1000. Yay — you now only have $1000 available to spend.
To what extent there are reasons for a government to impose reserve requirements, they certainly do not help banks’ liquidity positions. The lack of good justifications for reserve requirements explains why the developed countries have quietly abolished them without anything interesting happening. (The reason why people expect things to happen is because of laughable Economics 101 money multiplier models.)
With the United States finally abolishing reserve requirements, one can only hope that Economics 101 textbooks will catch up to this. Within a few decades, people might finally stop nattering on about them.