Misconceptions about bank liquidity: a 5-point clarification

First, let’s agree on a few definitions. There are different concepts of liquidity, and it’s important not to mix them up:

– market liquidity » refers to the ease with which assets can be sold rapidly on a market, without this having a significant impact on their price.

– macroeconomic liquidity » refers, if we take the definition given by Patrick Artus (2012), to the sum of 2 components: the liquidity created by the central bank and the liquid assets – in the sense of market liquidity – that non-financial agents choose to keep in their portfolios

– the « liquidity of a bank (or company) » reflects its ability to meet its obligations/disbursements as they fall due

– the concept of « bank liquidity » refers to all « liquid » assets in the sense of market liquidity that a bank possesses, i.e. all assets that can be rapidly mobilized without loss of value (note the subjectivity of such an estimate).

Here, we focus on the latter concept. In the following, we will take a strict definition of a bank’s liquidity (which we may also refer to in this sense as « bank cash »): we will refer to it as the sum of credit institutions’ holdings with the central bank. This definition therefore encompasses both the amounts in a bank’s current account with the central bank (commonly referred to as « reserves », and therefore including both required reserves and excess reserves) and the amounts placed by the bank with the central bank’s deposit facility. This definition is very strict in the sense that it includes only highly liquid assets, i.e. the cashless component of the monetary base. It could be broadened to include money-market assets held by banks, but this would run counter to the terms usually used.

 » Bank liquidity » and « bank liquidity »: individual bank behavior has no significant influence on the total level of bank liquidity in the system.

When we talk about liquidity in banking economics, the concept can be integrated at the level of a single bank or at the level of banks considered as a whole (i.e. the banking system). The essential idea here is that a bank can modify the amount of liquidity it possesses, whereas liquidity in the banking system will not be influenced by individual bank behavior (with 2 marginal exceptions, cited in footnote 1). In other words, the banking liquidity circuit is a closed one, in which leaks are exogenous to the banking system, as are replenishments (dependent on the will of the central bank).

Let’s take a concrete example to understand this idea. Suppose bank A buys an asset worth 100, say a government bond. You and I have an account with a bank, but the banks have their account with the bank of banks, i.e. the central bank. So, if A buys this asset from another bank B, bank A will see its account with the central bank decrease by 100 and bank B will see its current account with the central bank increase by 100. The result is a fall in bank liquidity for A, offset by an increase in bank liquidity for B: total liquidity in the system has not moved. Now let’s suppose that A buys this asset from a household or company. The operation is carried out by crediting the account of the bank where the household or company has its account, which is the same as if the bank had bought the asset directly from that bank. So, if the buyer has an account with bank B, bank A will be debited 100 from its account in exchange for the asset, and bank B will increase its account by 100 with its new deposit of 100. It’s important to remember that scriptural central bank money is a currency that only makes sense within the banking system, so it’s as if things happened in this order: first, bank A buys the security in exchange for scriptural money, then the household or company hands over this scriptural money to bank B in exchange for a « credit note », or more simply, a « deposit ». Depending on its own needs, Bank B will either hold these reserves or purchase various securities. When a bank buys an asset from a household or company, the bank’s reserves are simply transferred to the household or company’s bank account in return for the deposit. The same applies when a bank « sells » a loan to another bank. There’s no need to elaborate on this as it concerns credit transactions, since there is no risk of confusion in this case. The idea is that reserves do not leave the circuit through bank behavior[1].

The only agents that can influence the level of bank reserves are non-banks and the central bank[2].

Let’s use concrete examples to explain how bank reserves vary. If non-banking economic agents increase their demand for banknotes (during the festive season, for example), then reserves will fall. This is because the bank will have to « buy » bills from the central bank, and the bank’s current account will be debited in exchange for these bills. Similarly, when an agent has to pay the Treasury (which generally has an account at the central bank), the operation will result in a transfer from the agent’s bank account to the Treasury’s account, i.e. a debit from the bank’s current account to credit the Treasury’s current account at the central bank. These factors are often referred to as « autonomous factors », since they do not result from the central bank’s behavior. Bank reserves may also decrease or increase as a result of « institutional factors » , which in turn result from the central bank’s behavior. The central bank has a monopoly on issuing the monetary base: it can withdraw or add liquidity at any time (all other things being equal in normal circumstances). If it wants to withdraw liquidity, it can sell the securities it holds[3] to banks; if it wants to add liquidity, it buys them or carries out reverse repo operations (for temporary liquidity). In all cases, the decision to grant or withdraw liquidity rests with the central bank. This is currently the case in the eurozone, where major movements in the ECB’s « reserves » account are the result of actions taken by the central bank, mainly the long-term refinancing operation (LTRO).

We sometimes see articles directly or indirectly stating that the large amount on the central bank’s « reserves » account reflects a preference for liquidity on the part of banks (implying that a different behavior on the part of banks would change this amount), or that banks would choose between storing reserves and lending, with an increase in reserves meaning that they are not lending[4]. Or, in another form, that excessive excess reserves mean that banks are not lending. The above explanations make it clear that these assertions are erroneous, since they stem from confusion about the concept of bank liquidity.

There is no immediate correlation between the level of bank reserves and the level of lending in the economy: banks do not « lend out » reserves.

As we said earlier, bank reserves are a closed circuit. When a bank makes a loan, it creates money in the sense that, schematically, it creates a deposit that did not previously exist. Suppose bank A grants a loan of 100 to Mr. Durant. It will credit Mr. Durant’s account with it with 100 and create a « receivable from Mr. Durant » on its assets side. It has therefore not lent reserves to Mr. Durant, but simply created a receivable on its assets side and money on its liabilities side. The level of reserves has not changed, only the distribution, assuming the existence of compulsory reserves (which we’ll neglect later[5]). This example clearly shows that the same level of reserves can correspond to several levels of lending in the economy: banks do not « lend » reserves.

Of course, it would be incorrect to say that the quantity of credit is totally independent of the level of reserves. Bank A could repeat the same operation several times, with the only constraint being that it must be prepared for « leaks » outside its internal circuit: it will potentially have to keep a quantity of reserves in case Mr Durant carries out transactions with customers of bank B and in case he asks for banknotes. In the first case, it will have to pay bank B with a commonly accepted currency, i.e. central bank money (i.e. reserves, as explained above). To do this, it either decides to maintain a minimum level of reserves to satisfy this requirement, or it borrows from other banks or from the central bank once the requirement has materialized[6]. Clearly, if the central bank anticipates a 10% leakage of credit and does not want to borrow the quantity of money in question when the requirement materializes, it will grant credit for a maximum amount of 10 times its reserves. So, according to this approach, the maximum amount of credit is correlated with the level of reserves at individual level[7]. But if banks do not follow this approach, the limit no longer depends on the level of reserves, since each bank can always obtain reserves either from its peers or from the central bank at the marginal lending facility or discount window (or in the form of a current account overdraft) when the banking system’s reserves are insufficient (extreme case). In the latter case, it is clear that the level of reserves does not restrict the bank’s ability to lend, but may alter the profitability of its loans if the bank has to go to the central bank. The main point remains: there is no immediate correlation between the level of reserves and the level of credit. When credit picks up again, reserves do not « leave » the central bank, but remain at the same level.

When banks « use up » excess reserves, this does not mean that reserves are decreasing, it means that they are leading to an increase in the money supply.

Current quantitative easing policies have greatly increased the level of reserves in the banking system. Many economists advocate that banks should « use » these reserves, and the meaning of this phrase is sometimes misunderstood. This principle is easy to understand after what we have explained above. Banks’ reserves cannot be reduced by their own behavior, but what is expected of them is that they « use » these reserves by increasing credit or their purchases of securities from companies or households. In the first case, as we said earlier, the maximum potential quantity of credit has increased as a result of the rise in reserves: banks can no longer be reluctant to increase credit because of fears of « leakage » and therefore a potential lack of liquidity. In the case of asset purchases, this will translate into an increase in the money supply if banks buy assets from households or businesses (who will therefore increase their deposits). In the latter case, therefore, it is an increase in the demand for money that will lead to inflationary pressures, primarily on assets. Indeed, it is only when the monetary base « passes » into the money supply that inflationary pressures appear. The aim of this article is not to comment on the economic impact, but simply to point out that when banks « use up » their reserves, this does not mean that they are diminishing. The important thing to watch is therefore the change in the money supply following the change in the monetary base: if the latter has « passed through » into the former, then we can deduce something about bank behaviour[8].

Commercial banks do not finance central bank asset purchases via their reserves: don’t confuse the chicken with the egg!

This confusion stems from a disturbing article I read on the Le Monde.fr blog: « The ECB finances itself through banks, whether they like it or not ». The author of this post, who is nonetheless CEO of a New York investment bank, indicated in his post that the ECB finances « the bulk of its loans to financial institutions through bank deposits ». To do so, he relies on a tantalizing paradigm, based on the fact that when bank reserves increase, bank lending increases at the same time. The author even goes so far as to argue that banks in rich countries, which have plenty of liquidity, finance banks in poor countries indirectly via the ECB’s « reserves » account and its bank lending activities. This line of reasoning can lead to total confusion when we read « an increase in the ECB’s needs in the form of purchases of Italian or Spanish bonds raises the question of its financing: does it intend to do this partly in the form of minimum reserves? ». The author of this article, like the readers who failed to notice this error, have probably not fully grasped who is the chicken and who is the egg in the monetary system. The central bank is not financed by the banks, since the money on its assets side in the « reserves » line already has a corresponding asset. In other words, when the banks’ reserve account increases, it’s because the central bank has issued this central bank money: either by buying a security from the banks, or by entering into a reverse repurchase agreement, via a repo or refinancing operation. It has created central bank money simply by crediting the account of the bank in question (« electronic » money creation). The causality necessarily runs in this direction, since, as we explained earlier, the central bank has a monopoly on issuing the monetary base. In a way, the central bank is the chicken, and the banks’ « reserves » account is the egg. It makes no sense to say that reserves finance the central bank’s asset purchases[9].

Notes: