Endogenous money is the notion that banks create money out of thin air. To be more precise, banks create money in the act of lending. Banks buy promissory notes from borrowers in exchange for demand deposits. A demand deposit is a promise, issued by a bank, that is redeemable upon demand into paper currency or central bank reserves. In effect, demand deposits are a type of IOU that we, in turn, accept as money to trade goods and services. To illustrate, when you use a credit card to pay at a restaurant, the bank debits your credit card account and credits the deposit account of the restaurant — voila, your credit-card loan created the restaurant’s deposit. Under the endogenous money view, banks do not intermediate between borrowers and savers but stand ready to facilitate borrowers with new money creation.
Endogenous money runs counter to the classical (aka, loanable funds) view according to which banks lend-out savings. In other words, banks must acquire loanable funds from willing savers before lending them out to willing borrowers. Because banks are not subject to 100% reserve requirement, they are able to lend the same funds to multiple borrowers, thereby multiplying the money supply. Under this view, banks truly intermediate between willing borrowers and savers with loanable funds (aka, reserves) originating from some exogenous source (eg, gold discoveries under the Gold Standard or the central bank today).
In this post, I will attempt to do two things:
- First, I will argue that the true nature of money creation lies between the two views. The endogenous money view applies to zero-duration loans such as credit cards, but when it comes to loans with duration (eg, 30-year fixed-rate mortgage), the classical view offers a better description of how banks operate.
- Next, I will discuss the implications for capital markets and the macro economy.
Let’s start with duration. Duration is a financial term that measures the price sensitivity of an asset or liability (A/L) to changes in interest rates. If the A/L price never changes, regardless of interest rates, such A/L has zero duration. If the price does change with changes in interest rates, such A/L has duration. A simple rule of thumb: A/Ls with a fixed rate of interest have duration (e.g., Term CDs, 30-year fixed-rate mortgages); A/Ls with a variable rate of interest have zero duration (eg, demand deposits, credit card loans). In practice, A/Ls with duration enable borrowers to lock the interest on their loans and savers to lock the expected return on their savings.
- Zero-duration Assets and Liabilities
Bank deposit liabilities have zero duration because deposits generally pay a variable rate and are redeemable upon demand. When a bank extends a variable interest loan, it adds a zero-duration asset to its balance sheet (credit-card loan) and a zero-duration liability (demand deposit). There is no duration mismatch, hence no interest rate risk. The bank pays variable interest on the demand deposits and receives variable interest plus spread on the loan, which is how banks make money. This illustrates that banks can facilitate zero-duration borrowers simply by issuing demand deposit, effectively creating new money eh nihilo. In other words, the endogenous money view holds.
Proponents of the classical view will counter that borrowers will likely spend the funds and the newly created deposits will leave the bank as a result, forcing the bank to look for outside funding. Going back to the credit card example, if the restaurant owner banks with a different institution, the credit-card issuing bank (Bank A) must convert the new deposit into central bank reserves and send those reserves to the bank of the restaurant owner (Bank B).
Reserves are liabilities issued by the central bank that commercial banks use to meet reserve requirements and settle interbank transactions. The chart below shows how reserves fit into the hierarchy of money.
At the bottom of the pyramid, commercial banks supply the non-bank private sector with deposits by extending credit to private borrowers. The central bank supplies the banking sector with reserves either by extending credit to banks or by buying government bonds. Think of bank deposits and central bank reserves as two different currencies that are fully convertible into each other at a fixed exchange rate of 1 to 1.
It is important to note that the central bank has two goals with respect to the supply of reserves. First, it seeks to facilitate settlement of interbank payments. For that purpose, the central bank is willing to massively expand its balance sheet intra-day by creating reserves to fund bank overdrafts due to interbank payments. Each commercial bank has an account at the central bank with intra-day overdraft privileges. To facilitate an interbank payment, the central bank simply debits the account of the payer-bank and credits the account of the payee-bank. Even if a bank overdraws its account at the central bank during the day, that’s okay as long the account is replenished with reserves by the end of the day.
Going back to our example, if Bank A overdraws its account at the central bank as a result of the restaurant payment to Bank B, the central bank will simply add an overdraft asset to its balance sheet that is due from Bank A and create an offsetting reserve liability that is credited to Bank B. Before the end of the day, Bank A could borrow the newly created reserves from Bank B and pay back the overdraft to the cental bank. Upon such payment, the reserves, that were created earlier in the day, are destroyed; the central bank balance sheet shrinks back to its original size, and last but not least, the restaurant payment from Bank A to Bank B has been successfully settled.
One could argue that a bank does not need retail deposits at all but could fully fund its balance sheet by borrowing from other banks (wholesale borrowings). The very reason the wholesale interbank market exists is because loans and deposits are not equally distributed across the banking industry. In practice, retail deposits have a number of advantages as a funding source, plus regulators require banks to fund at least a portion of their balance sheet with retail deposits, hence banks compete for deposits. However, this competition does not in any way invalidate the fact that loans create deposits and the banking system as a whole, as long as it is creditworthy, is self-funding.
One way or another, a creditworthy bank can always obtain reserves as needed to meet settlements and required reserves and pass the related costs onto its borrowers. In other words, banks’ capacity to originate loans by issuing new demand deposits is unconstrained by reserves [1].
An interesting question to consider is whether creditworthy banks are constrained in any way by the demand for deposits. What if people don’t want to hold deposits? The problems is that people cannot destroy deposits unless a borrower pays off a loan (or a deposit is converted into cash, but that constraint affects only non-creditworthy banks). When it comes to deposits, the non-bank private sector has the hot potato problem. In the aggregate, depositors cannot rid themselves of deposits, they can only pass them around, which lends further support to the endogenous money view.
Another argument for the classical view looks at interest rates, which relates to the central bank’s second objective with respect to reserves. The central bank, as the monopoly supplier of reserves, can operate under one of two regimes. The central bank can either fix the supply of reserves, letting interest rates in the interbank market equilibrate demand, or it can fix the price of reserves via its interest rate target, letting the supply of reserve fluctuate with demand.
The first approach was abandoned in the 1980s because it was found to be impractical as interest rates and the money supply fluctuated unpredictably, making it impossible for the central bank to achieve its price stability and full employment mandates. Since then, central banks have adopted the second approach, whereby they supply reserves as needed to achieve the interest rate target.
According to the classical view, given that central banks target interest rates and that banks need new loans in order to issue new deposits, central banks affect the demand for borrowings and, by extension, the money supply vis-a-vie interest rates. According to this argument, the central bank’s rate target ultimately determines money creation by banks, with the money supply being no different had banks operated under the classical view.
The problem is that interest rates are only one of many factors that determine the demand for borrowings. Income-growth and needs expectations are much more important drivers. This means that banks create money pursuant to borrowers’ income-growth and needs expectations, and central banks simply play catch-up to supply reserves as needed.
This became obvious in the aftermath of the Great Financial Crisis. Central banks massively increased the supply of reserves, significantly above reserve requirements, but that did not prompt banks to lend more. The money multiplier stopped working, and the money supply stagnated. The reason: people feared declining incomes and even 0% interest rates could not compel them to borrow more.
The take-away here is that the endogenous money view holds with respect to zero-duration loans. Banks are not constrained by reserves nor do central banks control the money supply. Creditworthy banks can obtain reserves on demand and the banking sector, as a whole, determines the money supply through lending activities because it is zero-duration loans that create deposits, not the other way around.
2. Bank Loans with Duration
When it comes to loans with duration, banks operate very differently. When a bank originates a 30-year fixed-rate mortgage, it creates an instant duration mismatch between its assets and liabilities. This exposes the bank to interest rate risk. If interest rates were to rise, the variable interest it pays on deposits goes up while the fixed interest it receives on mortgages stays the same. This causes the bank’s net interest margin to decline and even turn negative, resulting in losses to the bank.
Unlike credit risk, which is knowable and can be priced in the spread banks charge on loans, interest rate risk is unknowable because future interest rates are unpredictable. For a bank, going long on duration is no different than placing a casino bet. Banks are not in the business of betting on rates but in the business of locking spreads. For this reason, as soon as a bank commits to originating a loan with duration, it hedges the related interest rate risk by selling duration in capital markets. Furthermore, bank regulators require banks to manage interest rate risk within pre-defined risk tolerances. From the standpoint of both regulators and shareholders, it’s simply a matter of prudent banking.
Generally, there are two ways that banks sell duration, both of which cancel the money created upon the origination of the loan. The simplest duration sale involves the securitization of loans with duration into bonds and the sale of the bonds directly to savers in capital markets. As soon as the saver buys the bond, the money used for the purchase is destroyed. Basically, the saver returns the bank’s IOUs in exchange for the bond, at which point the IOUs are canceled.
The second type of duration sale involves an interest rate swap. Under an interest rate swap, two parties agree to exchange interest cashflows on some notional amount. In our case, the bank agrees to pay fixed interest to the other party while the other party, the duration buyer, agrees to pay variable interest to the bank.
To see why the swap cancels the money created upon the origination of the loan, think in terms of going long or short money. When I receive variable interest, I am long money (as if I held a bank deposit that pays variable interest). On the other hand, when I agree to pay variable interest, I am short money (as if I sold a bank deposit to someone else and the variable interest stream that comes with it). As soon as the bank hedges with an interest rate swap, the variable swap payer goes short money, canceling the long-money position created upon the origination of the loan. In other words, the net money position held by the non-bank private sector is unchanged as a result of the loan.
Someone could argue that private actors place huge bets on interest rates all the time. If the non-bank private sector wants to go long duration and short money, doesn’t this risk appetite facilitate bank money creation? Yes, but the question is which comes first. Does the demand for duration enable the bank to make the loan or does the new money issued upon the origination of the loan create its own demand for duration?
The answer is the former for two reasons. In terms of timing, banks hedge duration as soon as they commit to making a fixed-rate loan. In other words, banks hedge rate locks before the loans are actually originated. This means that the non-bank private sector must go short money first in order for the bank to be in position to create money later.
Furthermore, interest rates in capital markets always equilibrate the supply and demand for duration. Even if banks did not originate fixed-rate loans, capital markets will continue to trade duration unabated. It is only a matter of price or the rate of interest at which one party is willing to borrow and another party is willing to save. However, the reverse is not true. Banks will not be in position to originate fixed-rate loans unless they can transfer the duration to willing savers in capital markets. When it comes to loans with duration, banks truly intermediate between willing borrowers (suppliers of duration) and willing savers (buyers of duration). In other words, the endogenous money view does not hold, meaning that banks are not in position to create money ex-nihilo.
This framework is a hybrid between the classical view and the endogenous money view. Capital markets are the markets for duration (as opposed to loanable funds under the classical view) with banks truly intermediating between willing borrowers and savers. Money markets (or the markets for zero-duration A/Ls) are distinctly different. Banks act as market-makers, standing ready to facilitate zero-duration borrowers with newly created money even in the absence of willing savers.
3. Macro-economic implications
Why does this matter? Two reasons. First, this framework reveals the true nature of interest rates. Interest rates are the price of duration in capital markets. Duration here can be viewed as desire to borrow or save at a fixed rate of interest. Changes in interest rates ensure that such desired borrowings and savings are always in equilibrium. In other words, the classical view that interest rates equilibrate savings and investment does apply but only to savings and investments with duration. Disequilibrium between desired borrowings and savings can occur only in money markets because banks can facilitate zero-duration borrowers by creating money ex-nihilo. Such disequilibrium causes NGDP fluctuations that, in turn, enforce the ex-post identity between total borrowings and savings as well as the identity between money supply and demand. In other words, money supply creates its own money demand through nominal incomes.
The second implication redefines equilibrium. Endogenous money means that money does not exist ex-ante but must be created first before any trade of goods and services can occur. In advanced economies, there are only two sources of money: banks extending zero-duration loans or central banks buying assets with duration [2].
This has profound implications to output and employment. Setting the central bank aside for a moment, you need zero-duration borrowers for money to flow into the economy, generating spending and nominal incomes in the process. Who are these borrowers? These are consumers who use credit cards to facilitate purchase prior to, but in anticipation of, future incomes. These are firms that draw on bank credit lines to produce an inventory of goods and services in anticipation of future revenue. The newly created money flows to workers in the form of wages, who now have the money to buy the very same goods and services they themselves produced. Importantly, in both instances, consumer and firms borrow and spend prior to, but in anticipation of, future incomes and revenue. This reveals the fundamental feature of a monetary economy. Incomes must be anticipated ex-ante to be generated ex-post. In other words, a monetary economy must bootstrap in order to achieve any level of output and employment.
With respect to equilibrium, bootstrapping means that demand comes first. Agents form spending plans pursuant to their income expectations for the current period plus desired borrowings less desired savings. Banks finance this aggregate demand by creating money. Agents spend the money generating nominal incomes. Agents use such nominal incomes to redeem the zero-duration borrowings incurred at the beginning of the period, destroying the respective money balances in the process.
This process of money creation and destruction is what enables a monetary economy to operate. In a monetary economy, bank-financed demand creates its own nominal supply with prices, output, employment, as well as the money balances persisting at the end of the period, being the unknown residuals. This means that any observed state is an equilibrium state (we live in an infinite-equilibria economy), but such state may not be consistent with full employment. To achieve a state of full employment, aggregate demand must be bootstrapped for the potential incomes of the unemployed.
This is where the central bank and fiscal deficits come into play. The second source of money creation is the central bank buying assets with duration. Using its balance sheet, the central bank can prevent disequilibrium in money markets, thereby eliminating NGDP fluctuations. Furthermore, the central bank can bootstrap the economy toward full employment by monetizing fiscal deficits. In the absence of such monetization, bond-financed fiscal deficits have no impact on NGDP. As I explain above, capital markets perfectly match borrowers and savers, which means that bond-financed fiscal deficits crowd-out private borrowers. However, when the central bank buys government bonds, the respective deficits are, in effect, financed with newly created money. As a result, deficit spending becomes immediately additive to NGDP, bootstrapping the economy toward full employment.
It should be self-evident that understanding money is critical to understanding how a monetary economy operates. In my view, such understanding will chart a path toward sustainable prosperity without unemployment, inflation as well as asset booms and busts. I have to caveat that my views are fairly unorthodox and do not fit in any of the existing econ boxes. I put forth this new framework in several SSRN papers, summarized in this post: Alternative Understanding of the Business Cycle. As to policy prescriptions, I summarize my views here: Thoughts on Fed Targets and Tools. Finally, I find strong evidence of crowding-out and the bootstrapping effect of deficit monetization: The Empirical Case for Crowding-out and Japan — more dramatic evidence of crowding-out and the importance of central bank balance sheets.
PS: Please see Endogenous Money Q&A for these follow-up questions:
- What term point separates money markets from capital markets?
- What is the macro significance of repo markets?
- Does the duration of central bank balance sheets matter?
- Why does the interest rate swap cancel the money of the variable-swap payer? What does going short and long-money mean?
Endogenous Money — Q&A Part II discusses:
- Bank liquidity management
- The unstable money multiplier, which offers the strongest evidence in support of endogenous money, and interest on reserves.
[1] This assumes a normally functioning reserve market. However, if the interbank reserve market grinds to a halt due to build-up of counterparty risk and if the central bank refuses to supply additional reserve by expanding its balance sheet, you end up with bank failures and reduction of the money supply. However, this constraint is more relevant to the destruction of money, not to its creation.
[2] Proponents of MMT would argue that fiscal deficits are a third source of money creation. That would be the case if fiscal deficits were financed with zero-duration debt or if the central bank was required to finance the public debt; however, neither of those conditions applies to advanced economies due to central bank independence and self-imposed restrictions on fiscal authorities. For more, please refer to this post: Does MMT describe the operational realities of our monetary system?