My recent post on Endogenous Money attempts to reconcile the Post-Keynesian view that banks control the money supply with the classical/loanable-funds view that central banks do. I find that banks create money only when they originate zero-duration loans (loans that pay a variable rate of interest). When it comes to loans with duration (loans that pay a fixed rate of interest), the classical view offers a better description of how banks operate. That’s because banks hedge the loan duration in order to keep interest rate risk within pre-defined limits. In the process of hedging, banks effectively match willing borrowers with willing savers.
Unsurprisingly, I am getting pushback from both sides. Post-Keynesians object to any restrictions on bank money creation, even though such restrictions do not, in any way, invalidate the most important macro implication of endogenous money, namely that the economy does not self-correct to full employment.
Neither do loanable-funds proponents accept the more limited endogenous money view. An admission that banks create money endogenously effectively means that central banks do not control the money supply. This undermines the notion that monetary policy transmits to the economy through short-term interest rates given by the intersection of money supply and demand. The irony here is that that the restrictions I place on bank money creation mean that central banks do control the economy, but not through short-term interest rates, but through the size of the central bank balance sheet.
In this Q&A, I will respond to two challenges by loanable-funds proponents.
- Banks maintain a liquidity buffer of highly liquid assets, including cash and reserves. Does this mean that banks acquire funds first and lend later?
- Probably, the strongest evidence in support of endogenous money is the unstable money multiplier. For example, as the Fed embarked on QE in the aftermath of the Great Financial Crisis (GFC), the money multiplier plummeted, meaning that banks did not lend-out the new reserves. However, could there be an alternative explanation? The Fed started paying interest on reserves in 2008. Could that be the reason for the decline in the money multiplier?
First, what is bank liquidity? Here is a helpful definition from the OCC:
Liquidity is the risk to a bank’s earnings and capital arising from its inability to timely meet obligations when they come due without incurring unacceptable losses. Bank management must ensure that sufficient funds are available at a reasonable cost to meet potential demands from both funds providers and borrowers.
The most fundamental bank obligation is the promise to convert, on demand, bank deposits into cash or central bank reserves. We use bank deposits as money precisely because we perceive this promise to be risk-free. Bank liquidity management is the first line of defense behind this promise. Next in line is bank capital, which enables banks to absorb losses. Once capital is exhausted, there comes government deposit insurance, which is the ultimate backstop.
This illustrates that bank liquidity management is part of a complex hierarchy of mechanisms for mitigating counterparty risk in the banking system. Bank liquidity management generally involves the following:
- Sources of liquidity on the liability side of the balance sheet. In the US, banks have access to instant liquidity from Fed Funds markets as well as intraday credit and overdraft privileges at the Fed. Instant liquidity is used over very short periods, primary for settling payments and meeting reserve requirements. As to funding assets on balance sheets, banks use wholesale funds, retail deposits, other borrowed funds and capital. Wholesale funds, such as FHLB advances, repos and brokered deposits, are available at a very short notice, but such funding sources are more volatile in times of market destress. Retail deposits take time to raise and are more costly due to the operational and marketing footprint but offer a more stable source of funds.
- Sources of liquidity on the asset side of the balance sheet. In addition to cash and reserves, banks also hold other liquid assets such as government securities. Such assets are not subject to large price swings and can be sold at a short notice with little or no loss. Alternatively, banks can raise funds by pledging government securities as collateral in repo transactions. Asset liquidity is further enhanced by a range of securitization opportunities, whereby banks can pool loans into securities to be sold in capital markets.
- Policy limits. Banks must have a liquidity policy that sets limits with respect to various sources of liquidity. For example, a bank may decide that wholesale funds may not exceed 40% of total liabilities or that 10% of the balance sheet must be held in liquid assets. In certain instances, regulators may actually dictate those limits as is the case with liquidity coverage ratios for big banks.
- Cashflow forecast. Banks project near-term inflows and outflows of funds under various scenarios and raise funds as needed and/or add liquid assets to the balance sheet to meet the liquidity policy limits.
- Bank Capital. It is important to note that a bank can raise funds on an as-needed basis only if it is well capitalized. Bank capital is ultimately what gives rise to liquidity because it signals creditworthiness to other market participants, other banks and last but not least, the central bank. For this reason, the liquidity management policy should always be considered in conjunction with the capital management policy.
So does the fact that banks manage liquidity and hold liquid assets disprove endogenous money? In other words, does liquidity management mean that banks must obtain funds before extending loans? First, nowhere does the OCC say that. Instead, funds must be available at a reasonable cost to meet potential demands. This is entirely consistent with the notion that banks obtain funds as needed, not prior to lending.
There are several reasons for that. Raising funds always comes at a cost. If the bank is unable to find borrowers, such funding costs eat into profits. Instead, liquidity management allows banks to do the exact opposite — it enables banks to lend first and raise funds later, which is a way to maximizes profits.
Furthermore, cash and reserves are only a small portion of liquid assets. Banks attempt to minimize cash and reserves because of the low rate of return (vault cash earns zero and reserves did not pay interest prior to 2008). Instead, banks meet their liquidity ratios primarily by holding government securities (or government-guaranteed securities such as GNMA MBS) that earn a higher rate of return. If a bank finds itself short of funds, such securities can be pledged, at a very short notice, in repo transactions. On the other hand, if a bank finds itself holding excess funds, the immediate course of action is to pay down wholesale liabilities (or lend to other banks that need funds on as-needed basis) as opposed to lend more to private sector. No bank keeps excess idle cash around — that’s banking 101.
The main take-away from bank liquidity management is that banks manage to ratios. For example, rapid loan growth may depress the liquid asset ratio and increase reliance on wholesale funds. This prompts the bank to buy government securities and raise rates on retail deposits in an attempt to bring the liquidity ratios within policy limits. The important thing is that all of this happens ex-post. In other words, loan growth comes first. Loan growth is driven by the capital budget, not by the availability of funds. In conclusion, bank money creation (aka, bank lending) is always capital-constrained, but never reserve-constrained.
Now, let’s switch gears and talk about the money multiplier. Chart 1 shows the money multiplier in the US (blue line equal to M2/MBase).
If the loanable-funds view of money creation is true, one would expect to see two things: i) a flat money multiplier and ii) a multiplier that does not correlate in any way with the supply of reserves. As the story goes, the central bank injects reserves into the banking system. Banks lend-out the reserves, expanding the money supply in a consistent fashion based on minimum reserve requirements and payment settlement needs. To be more precise, one would expect to see a stair-step profile, where each step corresponds to a change in the regulatory regime (e.g., new reserve requirements / different rules guiding daily intrabank settlements, intraday credit and overdrafts at the Fed).
Instead, the data show anything but a stable money multiplier. Unstable money multiplier means that banks are in control of the money supply, not the central bank. Effectively, that’s another way of saying that the endogenous money view holds. The evidence post-GFC, when the Fed massively expanded its balance sheet as a result of QE, is even more dramatic. Changes in the money multiplier post-GFC show strong negative correlation with the Fed balance sheet. Basically, QE did not prompt banks to lend-out the new reserves because banks faced other challenges such as capital constraints and lack of creditworthy borrowers. As a result, the money multiplier fell.
Coincidentally, the Fed began paying interest on reserves in 2008. Could that have something to do with the fall of the multiplier? In the words of loanable-funds proponents, the Fed was rewarding banks not to lend. Well, not exactly. Banks incur significant operating and funding costs. The 25 bps the Fed paid on reserves were woefully insufficient to cover these costs. Furthermore, from the standpoint of an individual bank, getting interest on reserves is no different than selling funds in the Fed Funds market, the return is the same. Pre-GFC, the money multiplier was high despite a robust Fed Funds Market that paid in excess of 5%. The fact of the matter is that banks make money by lending to the private sector at much wider spreads that are sufficient to pay for funding and operating costs. I don’t believe any creditworthy borrower was turned away by a well capitalized bank simply because of the 25 bps at the Fed.
The reason the Fed pays interest on reserves is to maintain control over the Fed Funds rate, its preferred monetary policy tool, even when there are excess reserves in the banking system. In 2015, the Fed began raising interest rates in pursuit of its monetary policy objectives. If interest on reserves had anything to do with the money multiplier, one would think that the multiplier would fall further. Instead, the opposite happened. The money multiplier rose despite a very respectable 2.5% interest on reserves. That’s because the Fed had also started winding-down its balance sheet (mechanically, this causes the multiplier to rise, which is consistent with the endogenous money view). Furthermore, banks were better capitalized, and demand for loans increased as the economic recovered from the crisis. In other words, favorable economic conditions enabled banks to increase the money supply despite decline in reserves.
For further evidence let’s look at the UK (Chart 2) and the Eurozone (Chart 3). UK banks are not subject to minimum reserve requirement, and the BoE has always paid interest on its liabilities. Similarly, the ECB has always paid interest on reserves (Chart 3).
In either case, the negative correlation between the central bank balance sheet and the money multiplier is quite clear. When the central bank expands the balance sheet through QE, the multiplier falls. When the balance sheet shrinks, the multiplier rises. On the other hand, there is no relationship between the multiplier and the interest the central bank pays on reserves. In the Eurozone, interest on reserves became negative starting in 2014. You would think banks would be running head over heels to lend those funds to avoid the negative interest, but that’s not what happened. The multiplier fell continuously as the ECB continued to expand its balance sheet through QE with little to no impact on bank lending.
In conclusion, even if you believe that the money-multiplier process better describes how banks operate, it is quite evident that central banks have no control over the magnitude of the multiplier, but commercial banks do. This is just another way of saying that commercial banks control the money supply, which is ultimately what endogenous money means. We get to the same place, meaning that the implications to the macro economy are the same, regardless of which process of money creation you chose.
