By James Tobin
Paul Volcker, while chairman of the Board of Governors of the federal reserve system (1979–1987), was often called the second most powerful person in the United States. Volcker and company triggered the “double-dip” recessions of 1980 and 1981–1982, vanquishing the double-digit inflation of 1979–1980 and bringing the unemployment rate into double digits for the first time since 1940. Volcker then declared victory over inflation and piloted the economy through its long 1980s recovery, bringing unemployment below 5.5 percent, half a point lower than in the 1978–1979 boom.
Volcker was powerful because he was making monetary policy. His predecessors were powerful too. At least five of the previous eight postwar recessions can be attributed to their anti-inflationary policies. Likewise, Alan Greenspan’s Federal Reserve bears the main responsibility for the 1990–1991 and 2001 recessions.
Central banks are powerful everywhere, although few are as independent of their governments as the Fed is of Congress and the White House. Central bank actions are the most important government policies affecting economic activity from quarter to quarter or year to year.
Monetary policy is the subject of a lively controversy between two schools of economics: monetarist and keynesian. Although they agree on goals, they disagree sharply on priorities, strategies, targets, and tactics. As I explain how monetary policy works, I shall discuss these disagreements. At the outset I disclose that I am a Keynesian.
Few monetarists or Keynesians would disagree with this dream scenario:
•First, no business cycles. Instead, production—as measured by real (inflation-corrected) gross national product—would grow steadily, in step with the capacity of the economy and its labor force.
•Second, a stable and low rate of price inflation, preferably zero.
•Third, the highest rates of capacity utilization and employment that are consistent with a stable trend of prices.
Monetary policies are demand-side macroeconomic policies. They work by stimulating or discouraging spending on goods and services. Economy-wide recessions and booms reflect fluctuations in aggregate demand rather than in the economy’s productive capacity. Monetary policy tries to damp, perhaps even eliminate, those fluctuations. It is not a supply-side instrument (see supply-side economics). Central banks have no handle on productivity and real economic growth.
The second and third goals frequently conflict. Should policymakers give priority to price stability or to full employment? American and European monetary policies differed dramatically after the deep 1981–1982 recession. The Fed “fine-tuned” a six-year recovery and recouped the employment and production lost in the 1980 and 1981–1982 downturns. Keeping a watchful eye on employment and output, and on wages and prices, the Fed stepped on the gas when the economic engine faltered and on the brakes when it threatened to overheat. During this catch-up recovery the economy grew at a faster rate than it could sustain thereafter. The Fed sought to slow its growth to a sustainable pace as full employment was restored.
European central banks, led by the German Bundesbank, were more conservative. They did little to help their economies catch up. They regarded active monetary stimulus as dangerously inflationary, even when their economies were barely emerging from recession. They were determined never to finance more than sustainable noninflationary growth, even temporarily. Europe recovered much more slowly than America, and its unemployment rates have ratcheted up from the 1970s.
Priorities reflect national dreams and nightmares. German horror of inflation, for example, dates from the 1923 hyperinflation and from a second bout of inflation after World War II. Priorities also reflect divergent views of how economies work. European monetary authorities were acting like monetarists, Americans like Keynesians, although both would disavow the labels.
Here is the crucial issue: Expansionary monetary policy, all agree, increases aggregate spending on goods and services—by consumers, businesses, governments, and foreigners. Will these new demands raise output and employment? Or will they just raise prices and speed up inflation?
Keynesians say the answers depend on circumstances. Full employment means that everyone (allowing for persons between jobs) who is productive enough to be worth the prevailing real wage and wants a job at that wage is employed. In these circumstances more spending just brings inflation. Frequently, however, qualified willing workers are involuntarily unemployed; there is no demand for the products they would produce. More spending will put them to work. Competition from firms with excess capacity and from idle workers will keep extra spending from igniting inflation.
Monetarists answer that nature’s remedy for excess supply in any market is price reduction. If wages do not adjust to unemployment, either government and union regulations are keeping them artificially high or the jobless prefer leisure and/or unemployment compensation to work at prevailing wages. Either way, the problem is not remediable by monetary policy. Injections of new spending would be futile and inflationary.
Experience, certainly in the Great Depression and also in subsequent recessions, indicates that downward adjustments of wages and prices cannot avoid damage to output and employment. [Editor’s note: for another view, see great depression.] Moreover, wage and price cuts may actually reduce demand by generating expectations of further disinflation or deflation.
A. W. Phillips’s famous curve (see phillips curve) showed wage inflation varying inversely with unemployment. Keynesians were tempted to interpret it as a policy trade-off: less unemployment at the cost of a finite boost in inflation. Milton Friedman convinced the economics profession in 1968 that if monetary policy persistently attempts to bring unemployment below “the natural rate of unemployment” (the rate corresponding to Keynes’s “full employment”), it will only boost the inflation rate explosively. Friedman’s further conclusion that monetary policy should never concern itself with unemployment, production, or other real variables has been very influential. But in situations of Keynesian slack, as recent American experience again confirms, demand expansion can improve real macroeconomic performance without accelerating prices.
Here too the monetarist-Keynesian controversy is exemplified by Federal Reserve and Bundesbank policies in the 1980s. The issue is this: how actively and frequently should policymakers respond to observed and expected departures from their targets? Friedman wants them to follow the same routine regardless of the economic weather, increasing the money supply at a constant rate. In his view, trying to outguess the economy usually exacerbates fluctuations.
While not all monetarists endorse Friedman’s rule, they do stress the importance of announced rules enabling the public to predict the central bank’s behavior. In principle, announced rules need not blind policymakers to changing circumstances; they could specify in advance their responses to feedback information. But it is impossible to anticipate all contingencies. No central bank could have foreseen the opec shocks of the 1970s and decided its responses in advance. Any practicable rule is bound to be simple. Any reactive policy, like the Fed’s fine-tuning after 1982, is bound to allow discretion.
Relation to Fiscal Policy
In monetarists’ view, government budgets have important supply-side effects for good or ill but have no demand-side role unless they trigger changes in monetary policy. In Keynesian theory, fiscal policy is a distinct demand-side instrument. The government affects aggregate demand directly by its own expenditures and indirectly by its taxes.
Prior to 1981, presidents and Congresses in making annual budgets considered their macroeconomic effects. In the 1980s budget making became slow and cumbersome, and the explosion of deficits and debt made countercyclical fiscal policy very difficult. Since then, the burden of stabilization policy has fallen almost entirely on monetary policy. The one main exception, not necessarily intentional, is the timing of President George W. Bush’s tax cuts, which were, in essence, activist fiscal policy after 2001.
Monetary and fiscal policies are distinct only in financially developed countries, where the government does not have to cover budget deficits by printing money but can sell obligations to pay money in the future, like U.S. Treasury bills, notes, and bonds. In the United States, Congress and the president decide on expenditure programs and tax codes and thus—subject to the vagaries of the economy—on the budget deficit (or surplus). This deficit (or surplus) adds to (or subtracts from) the federal debt accumulated from past budgets. The Federal Reserve decides how much, if any, of the debt is “monetized”—that is, takes the form of currency or its equivalent. The rest consists of interest-bearing treasury securities. Those central bank decisions are the essence of monetary policy.
Mechanics of Monetary Policy
A central bank is a “bankers’ bank.” The customers of the twelve Federal Reserve banks are not ordinary citizens but “banks” in the inclusive sense of all depository institutions—commercial banks, savings banks, savings and loan associations, and credit unions. They are eligible to hold deposits in and borrow from Federal Reserve banks and are subject to the Fed’s reserve requirements and other regulations.
At year-end 2003, federal debt outstanding was $7,001 billion, of which only 11 percent, or $753 billion, was monetized. That is, the Federal Reserve banks owned $753 billion of claims on the U.S. Treasury, against which they had incurred liabilities in currency (Federal Reserve notes) or in deposits convertible into currency on demand. Total currency in public circulation outside banks was $664 billion at year-end 2003. Banks’ reserves—the currency in their vaults plus their deposits in the Fed—were $89 billion. The two together constitute the monetary base (M0 or MB), $753 billion at year-end 2003.
Banks are required to hold reserves at least equal to prescribed percentages of their checkable deposits. Compliance with the requirements is regularly tested—every two weeks for banks accounting for the bulk of deposits. Reserve tests are the fulcrum of monetary policy. Banks need “federal funds” (currency or deposits at Federal Reserve banks) to pass the reserve tests, and the Fed controls the supply. When the Fed buys securities from banks or their depositors with base money, banks acquire reserve balances. Likewise the Fed extinguishes reserve balances by selling treasury securities. These are open-market operations, the primary modus operandi of monetary policy. These transactions are supervised by the Federal Open Market Committee (FOMC), the Fed’s principal policymaking organ.
A bank in need of reserves can borrow reserve balances on deposit in the Fed from other banks. Loans are made for one day at a time in the “federal funds” market. Interest rates on these loans are quoted continuously. Central bank open-market operations are interventions in this market. When the Federal Reserve (or other central bank) conducts an open-market operation, it typically buys treasury bills, paying for them with reserves, or sells them, taking reserves in payment. Open-market operations thus amount to interventions in the federal funds market. Banks can also borrow from the Federal Reserve banks themselves, at their announced discount rates, which are in practice the same at all twelve banks. The setting of the discount rate is another instrument of central bank policy. Nowadays it is secondary to open-market operations, and the Fed generally keeps the discount rate close to the federal funds market rate. However, announcing a new discount rate is often a convenient way to send a message to the money markets. In addition to its responsibilities for macroeconomic stabilization, the central bank has a traditional safety-net role in temporarily assisting individual banks and in preventing or stemming systemic panics as “lender of last resort.”
Tactics: Operating Procedures
Through open-market operations, the FOMC can set a target federal funds rate and instruct its trading desk at the Federal Reserve Bank of New York to enter the market as necessary to keep the funds rate on target. The target itself is temporary; the FOMC reconsiders it every six weeks or so at its regular meetings, or sooner if financial and economic surprises occur.
An alternative operating procedure is to target a funds quantity, letting the market move the funds interest rate to whatever level equates banks’ demands to that quantity. This was the Fed’s practice in 1979–1982, adopted in response to monetarist complaints that the Fed had been too slow to raise interest rates in booms to check money growth and inflation. The volatility of interest rates was much greater in this regime than in the interest-rate-target regime.
How is the Fed’s control of money markets transmitted to other financial markets and to the economy? How does it influence spending on goods and services? To banks, money market rates are costs of funds they could lend to their customers or invest in securities. When these costs are raised, banks raise their lending rates and become more selective in advancing credit. Their customers borrow and spend less. The effects are widespread, affecting businesses dependent on commercial loans to finance inventories; developers seeking credit for shopping centers, office buildings, and housing complexes; home buyers needing mortgages; consumers purchasing automobiles and appliances; credit card holders; and municipalities constructing schools and sewers.
Banks compete with each other for both loans and deposits. Before 1980, legal ceilings on deposit interest restricted competition for deposits, but now interest rates on certificates of deposits, savings accounts, and even checkable deposits are unregulated. Because banks’ profit margins depend on the difference between the interest they earn on their loans and other assets and what they pay for deposits, the two move together.
Banks compete with other financial institutions and with open financial markets. Corporations borrow not only from banks but also from other financial intermediaries: insurance companies, pension funds, and investment companies. They sell bonds, stocks, and commercial paper in open markets, where the buyers include individuals, nonprofit institutions, and mutual funds, as well as banks. Households and businesses compare the returns and advantages of bank deposits with those of money market funds, other mutual funds, open-market securities, and other assets.
Thanks to its control of money markets and banks, the Fed influences interest rates, asset prices, and credit flows throughout the financial system. Arbitrage and competition spread increases or decreases in interest rates under the Fed’s direct control to other markets. Even stock prices are sensitive, falling when yields on bonds go up, and rising when they fall.
The Fed has less control over bond yields and other long-term rates than over money market and short-term rates. Long rates depend heavily on expectations of future short rates, and thus on expectations of future Fed policies. For example, heightened expectations of future inflation or of higher federal budget deficits will raise long rates relative to short rates because the Fed has created expectations that it will tighten monetary policy in those circumstances.
Another mechanism for transmitting monetary policy to the demand for goods and services became increasingly important after 1973. Since 1973 foreign exchange rates have been allowed to float, and obstacles to international movements of funds have steadily disappeared. An increase in U.S. interest rates relative to those in Tokyo, London, and Frankfurt draws funds into dollar assets and raises the value of the dollar in terms of yen, pounds sterling, and deutsche marks. American goods become more expensive relative to foreign goods, for buyers both at home and abroad. Declines in exports and increases in imports reduce aggregate demand for domestic production. High interest rates and exchange appreciation created a large and stubborn U.S. trade deficit in 1981–1985. Since 1985, as the interest advantage of dollar assets was reduced or reversed, the dollar depreciated and the U.S. trade deficit slowly fell. A similar pattern recurred between 1995 and 2002 as the exchange rate appreciated and the trade deficit widened as a proportion of GDP by more than 50 percent compared with its 1980s nadir. As of the end of 2004, two years of depreciation of the dollar have yet to reverse the trade deficit.
Targets: Monetary Aggregates or Macroeconomic Performance?
People hold dollar currency because it is the means of payment in many transactions. But checkable deposits are usually more convenient. They are not confined to particular denominations, cannot be lost or stolen, pay interest, and generate records most of us find useful.
The use of deposits in place of currency greatly economizes on base money. The $89 billion of bank reserves at year-end 2003 supported about $629 billion in checkable deposits. (The $572 billion of other assets behind those deposits were banks’ loans and investments. In this sense banks “monetize” debts of all kinds.) These deposits plus the $664 billion in circulating currency provided a stock of transactions money (M1) of $1,293 billion. But time deposits and deposit certificates, though not checkable, are close substitutes for transactions deposits in many respects. So are money market funds and other assets outside banks altogether. Consequently the Fed keeps track of a spectrum of monetary aggregates, M1, M2, M3, each more inclusive than the preceding one.
The same open-market operations that move the monetary base up and down and interest rates down and up change the quantities of M1 and other monetary aggregates. Operations that reduce federal funds rates and related short-term interest rates add to bank reserves, thus also to bank loans and deposits. In 2003 reserve requirements averaged about 10 percent of checkable deposits. Thus a one-dollar increase in the bank reserves component of the monetary base meant roughly a ten-dollar increase in the deposit component of M1. In contrast, a one-dollar increase in the currency component of the monetary base is always just a one-dollar increase in M1.
If there were no change in the ratio of deposits to currency that the public preferred in 2003, a $1.00 increase in the monetary base would mean a $1.70 increase in M1. This is the “money multiplier.” It does not stay constant, for several reasons. The Fed occasionally changes the required reserve ratio. Banks sometimes hold excess reserves, and sometimes borrow reserves from the Fed. The public’s demand for currency relative to deposits varies seasonally, cyclically, and randomly. In fact, the public’s demand for currency relative to demand deposits has increased significantly over the past fifteen years. In 1990, the money multiplier was $2.71 per new dollar of monetary base. In 2003, it was about 40 percent smaller. This reflects developments in transactions technology and financial institutions that allow people and firms to keep more of their liquid funds in forms other than checkable deposits and still pay their bills easily. Thus, the Fed’s control of M1 is imprecise, and its control of broader aggregates is still looser.
Monetarists urge the Fed to gear its operations to steady growth of a monetary aggregate, M1 or M2. Under congressional mandate the Fed twice a year announces target ranges for growth of monetary aggregates several quarters ahead. In the 1970s the FOMC tried to stay within these ranges but often missed. Monetarist criticism became especially insistent when money growth exceeded Fed targets during the oil shocks. In October 1979 Chairman Volcker warned the public that the Fed would stick to its restrictive targets for monetary aggregates until inflation was conquered. Three years later, however, the Fed stopped taking the monetary aggregates seriously.
Monetary aggregates are not important in themselves. What matters is macroeconomic performance as indicated by GDP, employment, and prices. Monetarist policies are premised on a tight linkage between the stock of money in dollars—say, MI—and the flow of spending, GDP in dollars per year. The connection between them is the velocity of money, the number of times per year an average dollar travels around the circuit and is spent on GDP. By definition of velocity, GDP equals the stock of money times its velocity. The velocity of MI was 8.5 in 2003. If it were predictable, control of MI would control dollar GDP too. But MI velocity is quite volatile. For the years from 1961 to 1990, average annual velocity growth was 2.2 percent, with a standard deviation of 3.6 percent. That is, the chance was about one in three in any year that velocity would either rise by more than 5.8 percent or decline by more than 1.4 percent. For the 1991–2003 period, the average annual growth of velocity was 1.6 percent, with a standard deviation of 5.9 percent. (M2 velocity is less volatile, but M2 itself is less controllable.)
Velocity depends on the money management practices of households and businesses throughout the economy. As transactions technologies and financial institutions have evolved and an increasing array of money substitutes has arisen, velocity has become less stable and monetary aggregates have become less reliable proxies for aggregate spending and economic activity. The 1981–1982 recession was deeper than the Fed intended because the FOMC stuck stubbornly to its monetary aggregates targets while velocity was precipitously falling.
Accounting for aggregate demand as the product of a money stock and its velocity is inadequate shorthand for the complex processes by which monetary policies are transmitted—via interest rates, banks, and asset markets—to spending on GDP by households, businesses, and foreigners. The Fed does better by aiming directly at desired macroeconomic performance than by binding itself to intermediate targets.
Scott Sumner on Monetary Policy, EconTalk podcast, November 9, 2009
Scott Sumner on Money and the Fed, EconTalk podcast, January 2, 2012
Jeffrey Rogers Hummel, “The Myth of Federal Reserve Control Over Interest Rates” at Econlib, October 7, 2013
Jeffrey Rogers Hummel, “What’s Wrong With the Taylor Rule?” at Econlib, November 2, 2014