Why banks are special — a reconsideration.
David Andolfatto has a very interesting paper that attempts to reconcile the mainstream and heterodox views on banking. The main thrust of his paper is to push back against the heterodox argument that banks are special because they can extend credit ex-nihilo (i.e., without needing to secure funds first). He constructs a model to illustrate that this distinction is trivial because capital markets effectively accomplish the same thing (borrowers issue promises ex-nihilo to be sold in exchange for savers’ labor or endowment of goods). Accordingly, banks do not differ from capital markets in a material way and are special only to the extent that they are better at funding investment. Therefore, money-creating banks can be safely ignored for the purpose of studying the business cycle.
I would argue that banks are special for an entirely different reason. I recently published a working paper titled Gain-Loss Utility, Equilibrium and Interest Rates. The paper reveals the fundamental role played by banks and the importance of money for the business cycle. My argument boils down to two words: bootstrapping and duration.
Practically all mainstream models assume that money exists ex-ante. Such cash-in-advance approach is justified if money is some exogenous object such as gold. In David’s case, the central bank supplies the old generation with money in the form of social security payments. However, in a fiat-money economy, money does not exist ex-ante. Instead, it must be created endogenously before any trading can occur, which is where banks come in.
To illustrate endogenous money, imagine an economy with agents and a single bank. Every agent uses a credit card, issued by the bank, to pay for purchases. Every time you swipe the credit card, you are effectively borrowing from the bank and the newly-created money is being deposited into the seller’s bank account. Upon receipt of incomes, agents pay-off their credit card balances. The redemption of credit-card debt results in the destruction of the corresponding money balances.
The important take-away from this simple illustration is that we borrow and spend prior to receipt of incomes but in anticipation of such incomes. In the aggregate, it is this bank-financed demand that generates nominal incomes. This means that a monetary economy bootstraps, whereby incomes must be anticipated ex-ante in order to be generated ex-post. This is also precisely the reason why banks are special. Only banks can act as the flux-capacitors that enable the economy to bootstrap.
To see why capital markets cannot help with bootstrapping, let’s first define duration. Duration measures the change in the value of an asset or liability with a change in interest rates. Assets and liabilities that do not have a term (meaning that they can be redeemed upon demand) or that carry a variable rate of interest have zero duration. Assets and liabilities with a term and a fixed rate of interest have duration. In the illustration above, the credit-card debt that bootstraps the economy has zero duration because agents intend to redeem such borrowings upon receipt of incomes, usually within the same period. For capital markets to be able to facilitate zero-duration borrowings, there must be zero-duration savers on the other side. But what if every agent in the economy bootstraps? In other words, every agent is a zero-duration borrower, and there are no zero-duration savers. In that circumstance, only banks can facilitate zero-duration borrowings. Banks are the market-makers for zero-duration borrowings, effectively stepping into the shows of the zero-duration saver.
In reality, not all agents bootstrap. Entrepreneurs do, but workers often do not. Because of job uncertainty, workers prefer to spend out of prior-period incomes. In other words, workers are the zero-duration savers. However, capital markets are not of much help here, either. Simply, workers do not want to invest their zero-duration savings in their employers’ IOUs. Zero-duration savings fund expenditures on a daily basis, hence agents must have the certainty that the funds are safe and available. This means that workers want to hold their zero-duration savings in a zero-duration asset that is also risk free. Only money, including bank-issued demand deposits, can foot that bill.
With respect to borrowings and savings with duration, capital markets rule. Duration allows long-term borrowers and savers to lock-in interest rates beyond the current period. To illustrate, if you expect an investment in a new factory to return 6% over 10 years, you want to fund that investment at a cost below 6%. Obviously, you wouldn’t want to borrow from a bank at a variable rate of interest that could exceed 6% in the future. By the same token, if you are a long-term saver and your minimum required rate of return over some time horizon is 4%, you want an asset that is expected to return at least 4%. Accordingly, you wouldn’t want an interest-bearing demand deposit that pays a variable rate of interest. The supply of duration by borrowers (or their willingness to borrow at a fixed rate of interest) and the demand for duration by savers (or their willingness to save at a fixed rate of interest) give rise to capital markets. In other words, I think of capital markets as the markets for duration as opposed to the markets for loanable funds.
How does this relate to unemployment and the business cycle? As the market-makers for zero-duration borrowings, banks enable the economy to bootstrap, and bootstrapping determines the level of employment. Furthermore, bootstrapping is precisely the reason why expectations are self-fulfilling. As a result, any disruption to banks’ capacity to lend has severe consequences for the economy, which explains why banking crises lead to severe downturns.
Banks also play a critical role with respect to the business cycle. Long-term savers with negative income-growth expectations (i.e., savers with negative minimum required returns) will not go to the capital markets to purchase assets with duration but will hold bank-issued money instead, giving rise to what I refer to as asset money demand. To see why, consider the fact that paper cash prevents interest rates on demand deposits from falling below zero. Zero is greater than the negative minimum returns required by such savers. By holding bank deposits, they effectively engage in a risk-free arbitrage. They keep all the upside, should interest rates rise in the future, but none of the downside, should interest rates fall.
Due to this duration mismatch, interest rates in capital markets will rise above the level that equilibrates desired borrowings and savings. This results in a shortage of desired borrowings that, in turn, translates into insufficient aggregate demand. Insufficient aggregate demand depresses nominal incomes below expectations, resulting in a fall in prices and employment. Only if banks go long on duration (by buying bonds or extending loans at a fixed rate of interest) could they offset the duration mismatch due to asset money demand. If banks go long on duration above and beyond what’s required to meet asset money demand, interest rates in capital markets will fall below the equilibrium level. This results in excess desired borrowings and, by extension, excess aggregate demand. Excess aggregate demand boosts nominal incomes above expectations, resulting in a rise in prices and employment.
In conclusion, banks are the beating heart of a monetary economy. They are critical institutions that cannot be ignored for the purpose of studying the macro economy. Their specialness does not come solely from their ability to create money ex-nihilo. After all, anyone can write IOUs ex-nihilo, which is exactly what borrowers in capital markets do. Banks are special because they are the market-makers in the market for zero-duration borrowings and savings. In other words, they have infinite capacity to finance aggregate demand by creating money, thereby enabling the economy to bootstrap.
 There is widely-held belief that returns on assets traded in capital markets dominate the returns on money. Not true. For simplicity, let’s set risky assets aside (for a detailed discussion of risky assets, please refer to the paper I referenced above). The yield on the 10-year Treasury is identical to the expected return on money over a 10-year time horizon, as projected by the 1-month forward rate curve. However, there is no guarantee that the 1-month forward rate curve will correctly predict the future path of interest rates on money. Only by buying the bond can a saver lock that expected return.
 Bootstrapping also means that an economy without an employer of last resort does not self-correct to full employment. Instead, any level of unemployment could be consistent with a steady state. To illustrate, unemployed agents will not bootstrap their demand (nor are banks going to extend them credit) because such agents do not anticipate incomes in the current period. As a result, aggregate demand in the current period will be insufficient to provide them with jobs.
 In practice, bank regulators require commercial banks to manage interest rate risk by duration-matching their assets and liabilities. In other words, commercial banks cannot go long on duration (or at least, they are not supposed to). Only the central bank has the balance sheet that can withstand a duration mismatch. Accordingly, only the supply of base money (or gold, under a gold standard) can offset asset money demand.