Inflation has been around since antiquity. The Roman Empire was particularly known for debasing its currency (by minting their coins with increasingly impure precious metals) to cover state expenditures, especially in times of war. The increasing nominal money supply and pressure from imports paid for with silver resulted in steady inflation of prices and wages.
The invention of paper money made it easy for governments, particularly those facing a balance of payments crisis, to monetize their debt and get caught in a vicious cycle of inflation. This occurred most famously in the Weimar Republic (prices doubling about every two days) and in Zimbabwe (prices doubling about every day). The most extreme example in recorded history was the Hungarian pengő (prices doubling about every 15 hours): “…by the time the pengo was replaced by the forint in August 1946 the dollar value of all Hungarian banknotes in circulation was just one-thousandth of one [US] cent.”
Some small level of inflation is seen by present-day central banks as desirable for various reasons. Most obviously, it discourages hoarding of capital and therefore fuels economic growth. The ability to freely print money also gives central banks a greater level of control over the economy through setting monetary policy. The Federal Open Market Committee (FOMC), which sets monetary policy in the United States, has a varying inflation target but it’s generally around 2% per annum.
The equation of exchange states:
where M represents the supply of money, V represents the velocity of money, P represents prices, and Q represents the real value of transactions. This equation relates the “real” economic factors V and Q with the “nominal” factors M and P particular to the monetary system. The monetarist school of thought holds that because V and Q are determined by the “real economy” they are out of our direct control, so inflation in prices P should be managed by controlling the size of the money supply M. This is by no means an uncontentious assertion, but the preference for monetary rather than fiscal intervention is deeply ingrained into today’s economic orthodoxy so it will serve as a useful model for us. So then how does the money supply expand under our current system, and what controls are in place to control this expansion?
The conventional view is that banks take money from depositors (in exchange for safety and account services) and lend some of that money out to borrowers (in exchange for interest). The original depositors still have their money on paper (as long as not too many people try to withdraw their money at once), and the borrowers now also have money to spend. The M1 money supply has therefore increased. Some of this borrowed money is spent and deposited back into the banking system, where it can be loaned out again. The main obstacle said to prevent this lending cycle from going on forever and producing infinite money is the reserve requirement imposed on banks by regulators to ensure that sufficient reserves (say, 10% of deposits) are on hand at all times to satisfy withdrawals of demand deposits. The inverse of the reserve requirement is called the money multiplier; a 10x money multiplier means that a deposit of ten dollars results in nine of them being lent out to borrowers, with one dollar held in reserve. Supposedly, by adjusting reserve requirements a central bank can control how quickly the money supply grows.
Numerous economists and financial commentators have pointed out that this conventional textbook view bears little resemblance to how banks lend money in the real world. Banks do not and can not lend out their reserves to anyone except for other banks and a few other types of entities involved in short-term money markets. Instead, when a commercial loan is disbursed, a new asset (the value of the loan to the bank) and a new deposit liability (the money placed in the borrower’s account) are added ex nihilo to the bank’s balance sheet. This newly-created “commercial bank money” does require corresponding “reserve money” (physical bank notes in the vault or reserves on deposit with the central bank) in case the borrower wants to withdraw cash, but any solvent bank has no difficulty in acquiring sufficient reserves by borrowing them from other banks or, in an emergency, the lender of last resort. Reserve requirements mostly serve to ensure that sufficient liquidity is present to settle transactions between banks at the end of each business day. Reserve requirements play little role in limiting commercial lending and the Federal Reserve seldom changes reserve requirements to effect changes in monetary policy. Some countries, including Canada, the United Kingdom, New Zealand, Australia, Sweden, and Hong Kong have abolished traditional reserve requirements altogether. Post-Basel III, large banks are required to hold sufficient so-called high-quality liquid assets to cover 30 days of outflows (the liquidity coverage ratio) but, again, this is meant to ensure sufficient liquidity to settle flows between banks, not to limit commercial lending. The factor primarily responsible for limiting excessive lending is not a bank’s reserve requirements, but a bank’s capital requirements: a minimum ratio of equity to debt meant to ensure that the bank can remain solvent if some of its loans fail. The equity of a commercial bank is at best only distantly connected to the strength of its reserves.
Given adequate capitalization, a commercial bank can create new money as it pleases, and the money market will ultimately provide it with whatever reserves it needs to meet its reserve requirements. So then how does the central bank control the money supply in order to hit its inflation target? It has to incentivize or disincentivize lending. It turns out to be very difficult to anticipate and control the supply and demand of money, so in modern times the Fed instead tries to control the interest rates of loans throughout the banking system. In a low interest rate environment, money for investment is easily obtained and the economy “heats up,” increasing prices. By raising interest rates, the central bank makes loans less attractive for borrowers and the economy “cools down,” decreasing prices (or increasing them more slowly than before). This is called the monetary transmission mechanism.
In order to influence commercial lending rates, the Federal Reserve manipulates the federal funds rate, the average interest rate that banks charge each other for overnight loans to cover their reserve requirements.
This rate perhaps doesn’t seem prima facie particularly significant, but the short maturity of the loan (overnight) and extreme creditworthiness of the borrowers (banks) makes the federal funds rate an essential indicator for the theoretical interest rate of an uncollateralized loan with close to zero financial risk. This risk free interest rate is used throughout the financial system as a base for determining rates on all kinds of debt instruments. For example, the prime rate used for home loans, student loans, credit cards, and other lines of credit runs about 3% above the federal funds rate. The Fed has the most direct impact on short-term, low-risk interest rates because reserve lending competes most directly with short-term, low-risk debt. As you go farther out on the yield curve (or look at higher-yield debt), the Fed’s impact is more muted. Would-be arbitrageurs will tend to narrow the spread between the fed funds rate and other forms of debt, but they have to assume some risk to do so, so there will be a premium involved which depends on the macroeconomic outlook of the market. Thus, in practice the Fed has a difficult time controlling long-term rates given the tools they have.
So how does the Fed manipulate the fed funds rate? Individual banks determine amongst themselves the rates that they charge each other for overnight loans, but the Fed can constrain and influence these rates. It sets a ceiling on the federal funds rate by offering at all times to lend to member banks at the so-called discount rate, which is set higher than the target federal funds rate, in exchange for suitable collateral. Banks with excess reserves can’t charge other banks much more than the discount rate because they’re competing with the discount window at the Fed. Similarly, the Fed sets a floor on the federal funds rate by paying interest on excess reserves. Banks typically won’t loan reserves to each other at less than the IOER rate because they would make more money by simply depositing the reserves with the Fed. The Fed also operates a repo facility where the Fed can take out or provide what are essentially collateralized loans, typically just for the night. The rates on these loans serve to constrain the fed funds rate just like the discount rate and IOER rate, albeit only up to a limited volume determined daily by the Fed.
Within these bounds, the Fed has various tools for pressuring the federal funds rate in one direction or the other, though it mainly engages in open market operations, wherein the Fed trades securities (especially short-term Treasury debt) with its member banks in exchange for reserves, at market prices. Banks with excess or insufficient reserves are often happy to take such a deal, particularly with Treasury bills because they compete directly with reserve lending in the short-term risk-free money market. When the Fed buys Treasurys from the banks in exchange for reserves, the amount of excess reserves in the system goes up, driving down the cost of borrowing reserves. Conveniently, buying large quantities of Treasurys also creates a scarcity in the market for government debt, allowing the Department of the Treasury to get away with offering lower rates on new issues of debt. Conversely, when the Fed sells Treasurys to banks in exchange for reserves, the amount of excess reserves in the system decreases and the cost of borrowing reserves goes up. The extra Treasurys on the market make it harder for the government to sell new debt, and they have to raise rates on new debt to attract buyers.
So, to summarize, who creates money? The central bank creates some money to fund government operations, manage financial crises, and ensure sufficient reserves. But mostly commercial banks do, creating it out of thin air on their balance sheets. Curbing this ability was the focus of the failed sovereign money movement in Switzerland earlier this year. The ability of commercial banks to lend in the United States is primarily limited by capital requirements to ensure that lenders don’t go bust when borrowers default. But there’s a lot of wiggle room within that basic safety limit. If the Fed wants to speed up or slow down commercial money creation to manage inflation, they have to incentivize lenders by manipulating interest rates.
If you’re interested in learning more, this Creative Commons textbook has more details. There are also numerous reports on monetary policy at federalreserve.gov. If you’re interested in the mechanics of the hyperinflation mentioned at the beginning of this post, I highly recommend picking up the fantastic and highly accessible Big Debt Crises by Ray Dalio of Bridgewater Associates fame. (Also available for free as a PDF.)