"Central banks' initial impact on interest rates is self-reversing. But how long does the process take?"
t is almost an article of faith among economists, politicians, and the public alike that central banks, such as the Federal Reserve (Fed), have total control over some critical interest rates. This faith applies across the political spectrum and, with only a few notable exceptions, across the policy positions that divide academic economists. Yet this near universal belief is inconsistent not only with standard economic theory and received economic history, but also with what is still taught in many mainstream economic texts. As Milton Friedman
put it as late as 1998, "After the U.S. experience during the Great Depression, and after inflation and rising interest rates in the 1970s and disinflation and falling interest rates in the 1980s, I thought the fallacy of identifying tight money with high interest rates and easy money with low interest rates was dead. Apparently, old fallacies never die."
The fixation on interest rates as the central bank's daily operating target emerged—or, more accurately, re-emerged after having been temporarily discredited by monetarism—during the 1980s. Financial deregulation and assorted other factors caused the predictable relationship between the money stock and such variables as the price level to break down in the short run, as money demand (and its reciprocal, velocity) behaved more erratically than in the past. So central bankers threw up their hands, concluded that monetary targeting was unworkable and returned to focusing on interest rates.
There is no denying that central banks have some impact on interest rates, in both the short and the long run. But the complexity of and qualifications to that impact have been swept into a memory hole, submerging a host of additional questions. How strong or weak is the initial effect on interest rates, and can its strength vary over time or with institutional arrangements? How long does the effect operate after a change in the rate of monetary growth? Does any impact on real rates eventually disappear completely, or can a constant rate of monetary growth make it permanent? Exactly what interest rates are initially affected by monetary policy: only short-term rates or rates across a broad spectrum of maturities extending even into real assets? Addressing these questions exposes serious constraints on the magnitude, duration, and predictability of the Fed's actual control over interest rates. The conventional impression of precise central bank management of even short-term rates turns out to be, at best, highly exaggerated.
The Received Theory: Short Versus Long Run
Friedman helped banish at least one popular fallacy about central bank control over interest rates. Economists now recognize the crucial distinction between nominal and real interest rates, real rates being adjusted either ex-post for actual inflation or ex-ante for anticipated inflation. Central banks can affect nominal interest rates indirectly through their impact on inflation. If the Fed pursues an expansionary monetary policy, then, once the higher inflation is fully anticipated, nominal interest rates will rise to offset the negative impact on real rates. And the Fed, through a tight monetary policy, can eventually lower nominal interest rates. In other words, high nominal interest rates can be a sign of either tight or easy money, depending on inflationary expectations.
For a basic example of how expectations of inflation affect interest rates, see Interest Rates, by Burton G. Malkiel, in the Concise Encyclopedia of Economics.
This long-term relationship between inflation and nominal interest rates is generally known as the Fisher effect, after the American economist Irving Fisher. Yet the effect received widespread acknowledgment only with the work of Friedman and the Great Inflation of the 1970s. The Fisher effect, of course, is not the current basis for the belief that central banks control interest rates. Rather, the dominant consensus involves a short-run, immediate impact that works in the opposite direction on both nominal and real rates, with an expansionary monetary policy lowering interest rates and vice versa. In 1969, Friedman introduced the term liquidity effect for this initial impact. But Friedman was neither the first nor the only economist to point out that the liquidity effect should be temporary. Lower interest rates encourage increased real investment and income, and, as the increased money stock simultaneously drives up the price level, real interest rates should return approximately to their previous level in what Friedman called the "income-and-price level effect"—later shortened to intermediate income effect. This analysis became conventional fare in money and banking textbooks.
Thus, an increase in the growth rate of the money stock will initially lower both real and nominal interest rates. But real rates eventually rise toward their previous level because nominal rates go higher due to anticipated higher inflation. This self-reversing sequence is just a manifestation of the approximate long-run neutrality of money, a proposition accepted even by most New Keynesian economists today. In long-run equilibrium, money affects only nominal magnitudes and not real factors. The long-run neutrality of money opens the first chink in the central bank's alleged control over interest rates. If the liquidity effect is a self-reversing phenomenon, then how can the Fed keep real rates below (or above) their equilibrium level for prolonged periods without simultaneously generating accelerating inflation (or accelerating deflation)?
Unfortunately, there is no agreement on why the liquidity effect operates at all. Friedman's final discussion of this question appeared in 1982, in a book he wrote with Anna J. Schwartz. Friedman and Schwartz offered two channels for what they relabeled the "impact effect" of monetary policy. They still called the first channel, attributed to John Maynard Keynes, the liquidity effect; but for clarity, I will adopt Phillip Cagan's term and call it the portfolio effect. It arises from how changes in the quantity of money alter people's total portfolio of assets. Before an unanticipated increase in the money stock drives up the average price level, it raises people's relative holdings of money above their desired level. Therefore, they will shed the excess by buying other assets, financial and real, to rebalance their portfolios. As the prices of those other assets consequently rise, the yields or interest rates on them temporarily fall. An unanticipated decrease in the money stock will have the reverse impact.
Friedman and Schwartz distinguished this portfolio effect from what they called the "first-round loanable funds effect." Before Keynes, the loanable funds effect was the traditional explanation for the initial influence of money on interest rates. In a more generalized form, this effect dates back to the eighteenth-century Irish-French economist Richard Cantillon and is generically known as a Cantillon effect. This effect hinges on where newly created money is injected into the economy. Since the conventional way is for central banks or private banks to increase the supply of loanable funds, interest rates fall, ceteris paribus. The reverse similarly occurs when banks decrease the supply of loanable funds, although, for simplicity, I will hereafter confine the discussion to monetary increases (unless otherwise stated).
Empirically distinguishing between these two channels is almost impossible. Most mainstream discussions seem to equivocate between the two ways that money can have a first-round impact on real and nominal interest rates: They base their theoretical justifications on a portfolio effect but their policy analysis on a Cantillon effect within financial markets. But whichever effect dominates, they are both self-reversing. Sooner or later prices will rise, increasing the nominal demand for money and restoring interest rates to their long-run equilibrium level.
Whether the new equilibrium level of real interest rates is exactly the same as the old depends partly on whether the economy is responding to a one-shot increase in the money stock or to an increase in money's rate of growth. Don Patinkin's Money, Interest, and Prices remains a definitive demonstration of money's long-run approximate neutrality. Patinkin persuasively concludes that after a one-shot increase in the money stock, real (and nominal) interest rates will remain unchanged in the long-run, so long as there is no exogenous change in any other variable, particularly the velocity of money.
What about an increase in money's growth rate? Here, because the effect will be an increase in the inflation rate, the analysis is less straightforward. We know that higher inflation means that non-interest-bearing money depreciates faster. Therefore, people will wish to hold less money, creating a one-time increase in money's velocity—that is, a decrease in money demand. Patinkin refers to such changes in velocity as "shifts in liquidity preference" and concludes, from a purely theoretical perspective, that whether they increase the real interest rate, decrease it, or leave it unchanged depends on how people reallocate their spending. Keynesian economists Robert Mundell and James Tobin have argued that an increase in velocity will decrease consumption and increase saving, causing real interest rates to fall. As a result, if the inflation rate rises, say, from three to eight percent, they predict that the Fisher effect will push nominal interest rates up by less than five percentage points, with real rates slightly lower than before the change in inflation. Most economists regard this Mundell-Tobin effect on real rates, if it exists at all, as very slight because non-interest-bearing money is a small fraction of people's total wealth.
Duration and Magnitude of Interest Rate Impacts
In short, central banks' initial impact on interest rates is self-reversing. But how long does the process take? A great deal of research has been done on the timing of the Fisher effect, and the conclusion is that institutions matter. But for the post-World War II period, the Fisher effect operates quickly enough to produce an obvious correlation between inflation and nominal rates in countries worldwide; and, sometimes, inflationary expectations can take hold quickly enough to swamp the liquidity effect, causing nominal rates to rise rather than fall in anticipation of a dramatic acceleration in monetary growth.
In contrast, very little research has been done on how long it takes the intermediate income effect to return real interest rates to their previous level. The most extensive discussion I have found is in the 1982 Friedman and Schwartz volume, which even considers the possibility of a cyclical adjustment that temporarily causes nominal and real rates to overshoot the long-run equilibrium. Yet given the current rapidity of the Fisher effect, it makes little sense to insist that real rates remain very far from long-run equilibrium for prolonged periods. After all, the Fisher effect is a partial return of real rates to equilibrium. The only way for the portfolio effect to keep real rates down is for the monetary expansion to remain partly unanticipated so that the price level has not yet fully adjusted. A Cantillon effect operating directly through the loan market not only is more abrupt, but also could be more prolonged because of where it pinpoints the continuous monetary injection. But notice that this prolonged lowering of real rates is accompanied by a rise in nominal rates, which still runs counter to simplistic, popular notions about the Fed's control over interest rates.
Whatever the duration of the initial liquidity effect on lowering interest rates, even greater problems confront the conventional view of the magnitude of the effect, however long it lasts. Deirdre McCloskey argues that the effect is tiny. Pointing out that "a little person in a large market cannot move the price very much," she builds her argument on the microeconomic distinction between a price taker—a firm so small relative to the size of the market that it has almost no influence on market price—and a price searcher—a relatively large firm that can have a significant impact on the price of its product. The Federal Reserve directly controls only the monetary base—that is, the total quantity of dollar reserves in the banks and dollar currency in the hands of the general public. While the base is critical for affecting the price level, she observes that in 1998, it was about $500 billion and rose by about $40 billion over the subsequent year. Estimating the size of global assets, real and financial, at $280 trillion, McCloskey calculates the elasticity of the demand curve that the Fed was facing in its effort to move those rates as 14,000. She concludes: "That's no influence."
Eugene Fama gives a more conservative statement of the same argument: "In recent years total U.S. credit market debt, as reported in Federal Reserve Flow of Funds tables, is in excess of $50 trillion." He thereby omits all real assets and all equity finance, ignores the rest of the world, and looks at the Fed relative to the size of only the U.S. credit market. "Prior to the financial crisis of 2008, total financial assets held by the Fed are less than $1 trillion, or less than two percent of the U.S. market. In response to the financial crisis of 2008, total financial assets held by the Fed jump to over $2 trillion and are almost $2.5 trillion at the end of 2010. This is huge by historical standards, but still less than five percent of the U.S. market. Several large banks (J.P. Morgan Chase, Bank of America, Citibank, and Wells Fargo) have balance sheets comparable in size to the Fed's." In other words, the belief that central banks have control over real interest rates violates the most fundamental and widely accepted principles of price theory.
The strength of this criticism, admittedly, depends on which is more powerful for the Fed's first-round impact on interest rates: the portfolio channel or the loanable-funds channel. The criticism appears devastating for the portfolio effect. As Friedman and Schwartz point out, "reported interest rates are only a few of a large set of rates of interest, many implicit and unobservable." Among myriad ubiquitous connections, "the interest rate connects stocks with flows, the rental value of land with the price of land." Interest rates also "connect the holders of financial assets to holders of real assets." This point by Friedman and Schwartz coincides with the long-standing insistence of Austrian capital theorists that the loanable funds market is just a small segment of the total time market; while the interest rate in the loan market provides an observable benchmark, it is dominated by internal rates of return within the capital structure. In defense of the central bank's influence, Friedman and Schwartz suggest that "because interest rates connect large stocks to relatively small flows, they can display wide variations as a result of apparently trivial changes." Yet that still leaves unclear why the small flows resulting from changes in base money should be singled out as uniquely powerful over other, often larger, flows. One such larger flow is the inflow of foreign savings into the United States. In 1998, this exceeded $200 billion, five times the base increase noted by McCloskey in the same year.
If we turn to the loanable-funds channel, then a stronger case exists for the magnitude of the central bank's affect on interest rates—but only for securities that the central bank actually purchases. For the Fed, that means that it has the most impact on the Federal funds rate, the rate at which banks loan each other reserves and the next-most on Treasury securities. But consider the theoretical quandary this creates. To the extent that the various parts of the loan market are segmented, and the longer it takes for interest-rate changes to be transmitted across maturities and financial sectors, the more powerful the central bank's impact on particular interest rates, but the weaker its effect on the economy overall, will be. The less markets are segmented and the more rapid the transmission of changes, the weaker will be the impact of central banks on interest rates overall, bringing us back to the McCloskey-Fama challenge.
The combination of the loanable-funds channel and some market segmentation along with inevitable lags is why the Fed, after its creation in 1914, was able to smooth out seasonal fluctuations of short-term interest rates. It was why, during World War II, the Fed was able to peg the rate on Treasury bills at 0.375 percent in a full-fledged price support and to keep Treasury bonds at 2.5 percent. It did so by relying on a massive expansion in the monetary base at a time when the Depression decade had already left market interest rates at historical lows and the demand for money exceptionally high, and when wartime controls over prices and resource allocation took the United States in long strides toward a command economy. And it is why central banks in relatively closed economies with exchange controls and high barriers to the international exchange of goods and savings have more influence over interest rates than those operating in periods of globalization. But it in no way undermines the overwhelming dominance of real factors in ultimately determining real interest rates.
Empirical and Historical Evidence
A vast empirical literature, accumulated over the decades since World War II, has attempted to determine the strength of the liquidity effect. But none of it is truly decisive. Why? Any correlation between central bank targets and market interest rates can result from either the market following the central bank or the central bank following the market. Even during periods when the Fed seems to have the most control over the Fed funds rate, the Fed does not consistently hit its target. "[T]he evidence says that Fed actions with respect to its target rate," concludes Fama, in one of the most recent empirical studies, "have little effect on long-term interest rates, and there is substantial uncertainty about the extent of Fed control of short-term rates. I think this conclusion is also implied by earlier work, but the problem typically goes unstated in the relevant studies, which generally interpret the evidence with a strong bias toward a powerful Fed."
For more discussion of the Federal Reserve's interest rate policies, see the EconTalk podcast David Laidler on Money. See also the EconTalk podcasts with John Taylor.
Of course, accurate expectations about future Fed policy could, in theory, cause short-term market rates to anticipate changes in the target rate. Indeed, some economists believe that expectations can magnify the Fed's influence on interest rates, in what is sometimes referred to as "open-mouth" operations. Ironically, the ascendancy of the rational expectations hypothesis in the 1970s led to the opposite conclusion. Any anticipated change in monetary policy was viewed as having absolutely no effect on real rates, and some went so far as to imply that the Fed was utterly impotent at affecting any real variable under any circumstances. Today, the economics profession has pretty much rejected this extreme view of short-run monetary neutrality. Indeed, as a result of the financial crisis, a few have careened to the opposite extreme, believing that un-anchored expectations through irrational waves of positive feedback can drive market prices willy-nilly. Between these two extremes, McCloskey got it exactly right: "Market demands and supplies, after all, depend on expectations, and maybe some denial of price theory can for a while hold back its force. But Greenspan, like King Canute, cannot hold back the tide of global supply and demand for funds, not for long."
The Taylor Rule, with its unrealistic assumption that, without Fed intervention, the underlying real interest rate is constant, has reinforced the widespread belief that Alan Greenspan was responsible for the low interest rates that preceded the financial crisis. Indeed, sometimes the charge degrades into meaningless circularity: "Why were interest rates so low? Because of Greenspan's expansionary monetary policy. How do you know Greenspan's policy was expansionary? Because interest rates were so low." This, despite a steady fall in the growth rate of the monetary base, from ten percent in 2001 to below five percent in 2006. Moreover, the increase in the base was overwhelmingly dwarfed in size by the net inflow of savings from abroad. By 2006 alone, that annual inflow was in the neighborhood of $800 billion, far exceeding the mere $200 billion increase in the base for the entire half decade.
Even the dramatic explosion of the monetary base initiated by Ben Bernanke in September 2008 in response to the financial crisis is not the exception that it might appear to be. As the Fed increased bank reserves and currency in circulation by $2.5 trillion over the five years since, it also was, for the first time, paying banks interest on their reserves deposited at the Fed. Although the 0.25 percent rate it pays is quite low, it has consistently exceeded the yield on Treasury bills, one of the primary securities in the Fed's balance sheet. Thus, at least $2 trillion of the base explosion represents interest-bearing money that, in substance, is government or private debt merely intermediated by the Fed. This may have noticeable impacts on some implicit bid-ask spreads within financial markets and on the allocation of credit. But because it involves no net increase in loanable funds or any major, temporary increase in the net portfolio of people's real wealth, it should have no liquidity effect on interest rates through either channel. The Fed can have no more impact on market rates through pure intermediation—borrowing with interest-earning deposits in order to purchase other financial assets—than can Fannie Mae or Freddie Mac. The remaining $500 billion increase in non-interest-bearing money (what economists call "outside money") represents only a slightly more rapid increase than in the decade before the crisis, and nearly all of that has been in the form of currency in the hands of the public.
None of this is to deny the existence of the Fed's initial, short-run liquidity effect. As I conceded at the outset, central banks can affect interest rates somewhat. What is incorrect is the now-common but simplistic belief that the liquidity effect is so powerful that it allows the Fed to put interest rates wherever it wants, irrespective of underlying real demands and supplies in the economy. Nor do I deny that central banks have other far-reaching economic repercussions, often detrimental. But in a globalized world of open economies, the tight control of central banks over interest rates is a mirage. Central banks remain important enough players in the loan market that they can push short-term rates up or down a little. But in the final analysis, the market, not central banks, determines real interest rates.
Among the exceptions I would single out: Daniel L. Thornton, whose several articles on the subject include "Monetary Policy: Why Money Matters and Interest Rates Don't," Federal Reserve Bank of St. Louis, Working Paper Series, no. 2012-020A (July 2012): PDF file; Eugene F. Fama, "Does the Fed Control Interest Rates?" Chicago Booth Research Paper No. 12-23; Fama-Miller Working Paper (June 29, 2013); John H. Cochrane, "Inside the Black Box: Hamilton, Wu, and QE2," Comments at the NBER Monetary Economics Meeting (March 3, 20111): PDF file; and Deirdre N. McCloskey, "Other Things Equal, Alan Greenspan Doesn't Influence Interest Rates,"Eastern Economic Journal, 26 (Winter 2000): 99-101; Nick Rowe, "Interest Rate Targeting as a Social Construction," Worthwhile Canadian Initiative blog (November 9, 2009).
Milton Friedman, "Reviving Japan,"Hoover Digest (April 30, 1998). Also see "Friedman on Money,"EconTalk (August 28, 2006). This is fully consistent with Friedman's earlier writings on the subject, cited below.
Milton Friedman, "Factors Affecting the Level of Interest Rates," in Proceedings of the 1968 Conference on Saving and Residential Financing (Chicago: United States Saving and Loan League, 1969), pp. 11-27. Reprinted in Thomas M. Havrilesky and John T. Boorman, eds., Current Issues in Monetary Theory and Policy (Arlington Heights, IL: AHM Publishing, 1976), pp. 362-78, and available online as PDF file. Friedman had already sketched out these views in his 1967 presidential address to the American Economic Association, "The Role of Monetary Policy," American Economic Review, 58 (March 1968): 1-17.
Milton Friedman and Anna J. Schwartz, Monetary Trends in the United States and United Kingdom: Their Relation to Income, Prices, and Interest Rates, 1867-1975 (Chicago: University of Chicago Press, 1982), pp. 477-500. [Those pages are available online as a PDF file through the NBER].
Phillip Cagan, The Channels of Monetary Effects on Interest Rates (New York: National Bureau of Economic Research, 1972).
Don Patinkin, Money, Interest, and Prices: An Integration of Monetary and Value Theory, 2nd ed. (New York: Harper and Row, 1965). Although the first edition was published in 1956, it was the second edition that received wide attention.
Robert Mundell, "Inflation and Real Interest," Journal of Political Economy, 71 (June 1963): 280-83; James Tobin, "Money and Economic Growth," Econometrica, 33 (October 1965): 671-84. Harry G. Johnson arrived at the same conclusion employing what is now widely known as the Solow growth model (see Robert Solow) in "The Neo-Classical One-Sector Growth Model: A Geometrical Exposition and Extension to a Monetary Economy," in his Essays in Monetary Economics (London: Allen and Unwin, 1967).
McCloskey, "Other Things Equal, Alan Greenspan Doesn't Influence Interest Rates": 99, 101.
This quotation appears in the earlier, August 18, 2012 draft of Fama's working paper, "Does the Fed Control Interest Rates?" cited in n. 1 above. Fama dropped the statement from the latest, June 29, 2013, version of the working paper, and the earlier draft is no longer available online.
Friedman and Schwartz, Monetary Trends in the United States and United Kingdom, pp. 486, 26-27, 500.
Several economists have made this telling point: Thornton, "Monetary Policy: Why Money Matters and Interest Rates Don't": 27; Cochrane, "Inside the Black Box: Hamilton, Wu, and QE2": 7.
Jeffrey A. Miron, "Financial Panics, the Seasonality of the Nominal Interest Rate, and the Founding of the Fed," American Economic Review, 76 (March 1986): 125-40.
Fama, "Does the Fed Control Interest Rates?" most recent version (June 29, 2013): 19-20.
Daniel L. Thornton, "The Fed and Short-Term Rates: Is It Open Market Operations, Open Mouth Operations or Interest Rate Smoothing?" Journal of Banking and Finance, 28 (March 2004): 475-98; and Thornton, "Can the FOMC Increase the Funds Rate Without Reducing Reserves?" Federal Reserve Bank of St. Louis, Economic Synopsis, no. 28 (October 6, 2010).An influential survey and critique of the rational expectations literature is Robert J. Shiller, "Can the Fed Control Real Interest Rates?" in Rational Expectations and Economic Policy, ed. by Stanley Fischer (Chicago: University of Chicago Press, 1980), pp. 116-56.
In 2006 the U.S. current account deficit was $798 billion, whereas the financial account surplus was only $780 billion, reflecting the fact that the international balance of payments is the most imprecise of all macroeconomic aggregates; Bureau of Economic Analysis: Table. The total net inflows for 2001 through 2006 came to $3.5 trillion. The common term "global savings glut" may not be the most precise description of this inflow that peaked at 6 percent of GDP, but the sheer numbers suggest it represents a far better explanation for the period's low interest rates. Martin Wolf, Fixing Global Finance, updated edn. (Baltimore: John Hopkins University Press, 2010) offers one of the most convincing cases for the global-savings glut thesis of Alan Greenspan and Ben Bernanke.
Most recently, the financial crisis has inspired the Fed to go beyond mere control over the money stock into determining where credit flows, in what is essentially the new central planning. See Jeffrey Rogers Hummel, "Ben Bernanke versus Milton Friedman: The Federal Reserve's Emergence as the U.S. Economy's Central Planner,"Independent Review, 15 (Spring 2011): 485-518; reprinted in David Beckworth, ed., Boom and Bust Banking: The Causes and Cures of the Great Recession (Oakland, CA: Independent Institute, 2012); Hummel, "The New Central Planning," review of The Federal Reserve and the Financial Crisis by Ben S. Bernanke, Wall Street Journal (March 29, 2013): A-13.