Here I will explain why this idea simply does not make sense, but no more casinos and gaming chips. No more senior croupiers and casino cashiers. I will dispense with the metaphor and instead stick to a more prosaic explanation, looking at interactions between banks and central banks.
All banks have their own accounts with the central bank. Often these are called “reserve accounts”, although in Australia they are called “exchange settlement accounts” (ESAs). As the Australian terminology suggests, the primary function of these accounts is to facilitate settlement of transactions that take place between banks. To keep it simple here, I will stick to the terminology of “reserve accounts”.
To see how this works, imagine I make a $100 purchase from a shop on my credit card. If the shop banks with the same bank as I do, all that happens is that our bank increases the balance of my credit card by $100 and also increases the balance in the shop’s bank account by $100. With a couple of simple accounting entries and no movement of any physical currency, the transaction is complete. In fact, as was discussed in the casino money post, this simultaneous $100 loan advance to me and $100 deposit raising for the shop has effectively “created” an additional $100 of money in the economy that was not there before.
Things are slightly more complicated if the shop banks with a different bank. In that case, my bank (let’s call it “Bank C”) will still debit $100 from by credit card, but it now has to make a $100 payment to the shop’s bank (“Bank W”). The way this is done is through the two banks’ reserve accounts: my bank will have its balance with the central bank reduced by $100 and the shop’s bank will have its balance increased by $100. In practice, this does not happen for every individual transaction. On any given day, there will be a large number of payments to and from every bank. So, banks will net all of these transaction each day and then transfer money between their reserve accounts to cover the net amount owed one way or another.
One problem that can then arise is that a bank may not have a sufficient balance in its reserve account to make the settlement payments. The typical solution is for the bank to borrow what it needs from another bank. Sticking with the simple example of the $100 purchase and ignoring the fact that the figures involved here are unrealistically small, my bank may only have $50 in its reserve account and may therefore choose to borrow $50 from another bank. This loan results in $50 being transferred into its reserve account, bringing the balance up to $100 which it can then transfer to Bank W’s reserve account. It may even be that Bank C could borrow from Bank W, in which case it is effectively making good on the payment of the $100 by a combination of a $50 reserve account transfer and a $50 loan. As a last resort, if Bank C is unable to borrow from another bank, it will have to borrow from the central bank. Acting as a lender of last resort in this way is part of the function of a central bank.
Looking at all of these scenarios, purely internal account entries at Bank C, reserve balance transfers between Bank C and Bank W and combinations of reserve transfers and inter-bank loans, there is one very important point to note. At no point did the aggregate balance across the reserve accounts of all banks change. All that happened was that balances were shuffled from one bank to another. In fact, contrary to the view of certain financial commentators, it is impossible for a bank to lend its reserves to the private sector. Banks can certainly make loans, but these loans never take money from a bank’s reserve account and give it to an individual or a business (who, after all, do not have accounts with the central bank). If anything, these loans may move some of their reserve balances to the reserve account of another bank and so the actions of one bank can affect that bank’s reserve balance, but not the total amount of bank reserves.
So, how can aggregate reserve balances change? The only way this can happen is through the transactions of entities, other than banks, which deal directly with the central bank. This means the central bank itself or the government treasury, which also banks with the central bank. In its management of monetary policy, the central bank will routinely conduct “market operations” which involve lending to banks (which increases aggregate reserve balances) or borrowing from banks (which decreases aggregate balances). Similarly, the execution of fiscal policy by the treasury will also involve government spending (which increases aggregate balances), debt issuance or taxation (both of which reduce aggregate balances). To give an example of the mechanics behind one of these processes, imagine you are due a tax refund and you receive a cheque from the tax office. Once you deposit this cheque at your bank, the bank will increase your account balance by the amount on the cheque and then put a claim in to the tax office for that amount. This claim will be met by means of a transfer into the bank’s reserve account. For accounting clarity, there will also be a debit of the account with the central bank associated with the tax office, but in a modern economy with fiat money the government is effectively the monopoly issuer of currency (just like the casino is the monopoly issuer of chips in the last post) and so there is no sense in which the amount credited to the bank’s reserve account actually has to come from anywhere. It is simply a matter of the central bank saying that the bank’s balance has increased.
So, treating the central bank and the treasury as part of a broad “government sector”, the only way aggregate reserve balances can change is through the actions of the government sector, not through the dealings of banks with the private sector. Reflecting on what governments have done through the financial crisis can then solve the supposed riddle of the large reserves, particularly in the US and the UK. Governments around the world have made extensive use of fiscal stimulus packages, putting their budgets into deficit. Where has all of this stimulus spending ended up? In the reserve accounts of banks. Furthermore, some central banks have conducted extensive “quantitative easing” programs, which is just a fancy name for buying large volumes of bonds of various types (including corporate bonds and mortgage-backed securities). Without going into the reasons or the efficacy of quantitative easing, the relevant point here is that the process of central banks buying bonds directly results in increases in bank reserve balances. Are banks then hoarding these balances? Of course not. There is nothing they can do to reduce reserve balances by lending. The balances will not go down again until governments run surpluses or central banks start selling down their bond holdings again (or both). Does it matter that these reserve balances are so high? Not in the least.
There is one last point to clarify in dispelling the myth of reserve hoarding and that is the notion of “excess reserves”. In some countries, including the US, banks are required to have a minimum balance in their reserve accounts as a percentage of their total deposits. In the US, this minimum percentage is 10%. So, a bank with $1 billion in deposits must maintain at least $100 million in reserves with the US Federal Reserve (the US central bank). Any balances in excess of this amount constitute “excess reserves”. So, a bank with $1 billion in deposits and $150 million in reserves has $50 million in excess reserves. Unlike aggregate reserve balances, the actions of individual banks can affect aggregate excess reserves. Recall that bank lending effectively creates new money in the form of deposits. Now imagine that this bank with $1 billion in deposits writes a whole lot of new loans totalling $200 million. For simplicity imagine that all of the money is spent at businesses which bank with the same bank (much like the first $100 shopping scenario above), so this bank now has $1.2 billion in deposits and a minimum reserve requirement of $120 million. It still has $150 million in reserves, but its excess reserves have now dropped from $50 million to $30 million. Of course, it is not the case that the bank has spent the $20 million difference, it’s just that this $20 million is no longer considered “excess”.
Before we start getting too generous and conclude that commentators bemoaning reserve hoarding were actually talking about excess reserves, there are a couple of points to note. First, many countries, including the Australia and the UK, do not have minimum reserve requirements (and the article I linked to at the top refers to banks in the UK). Second, the relationship between excess reserves and bank lending is not straightforward. Many banks have reacted to the crisis by making greater use of the bond market. The reason for this is that, deposits can be withdrawn at call and so poses risks for the bank (so-called “liquidity risk”), just like the risk the casino runs that too many people will cash in their chips at the same time. Bonds on the other hand do not have to be repaid for a specified period of time, typically three to five years and so help to reduce the bank’s liquidity risk. For the same reason, banks have been tempting depositors with attractive interest rates to switch to term deposits. These term deposits are not subject to reserve requirements in the US and other countries that have minimum reserve requirements. So this shift in liquidity can increase the banks excess reserves whether or not they continue to lend.
So while finance commentators can reasonably ask whether bank lending has been swung too far from the easy money days of lending too freely to excessively tight lending practices, worrying about banks “hoarding reserves” is to completely misunderstand the workings of the banking system.
* I originally linked to this article but, as has been pointed out to me, not only does that article misunderstand the mechanics of central bank reserve accounts, but the writing was rather impenetrable. I have therefore replaced the link with another that is equally misguided, but easier to read.