Remarks at the Laboratory for Aggregate Economics and Finance, University of California, Santa Barbara

Transcription

1 The Zero Lower Bound: Avoidance and Escape Remarks at the Laboratory for Aggregate Economics and Finance, University of California, Santa Barbara William T. Gavin Vice President and Economist Federal Reserve Bank of St. Louis Santa Barbara, California October 27,

2 The Zero Lower Bound: Avoidance and Escape By William T. Gavin 1 (Statistical results presented here come from The Zero Lower Bound and the Dual Mandate, an ongoing research project with Francesco Carli and Ben Keen) Today I am going to talk about the monetary policy process. In particular I want to talk about the policy process that took us to a zero interest rate and what, if anything, the Fed can do to get interest rates and the economy back to normal. Before I begin, I want to thank Finn Kydland and the Laboratory for Aggregate Economics and Finance for inviting me to talk to you today. Also, I want to thank him for all he has taught me about economics and helping me to think more clearly about the role of money and taxes in aggregate fluctuations. There are three parts to my talk today. First, I describe the current economic situation and explain why I think that we have fallen into a zero interest rate trap. Second, I explain what economic models say about why we have fallen into this trap and how we might have avoided it. Understanding how to avoid the zero lower bound suggests policies to escape from it. Escaping involves the issue of inflation targeting. In the third and final part of this talk, I discuss some popular myths about inflation targeting. The Zero Lower Bound The term zero lower bound refers to the idea that the market interest rate cannot be negative. A negative interest rate would mean that a lender would pay someone to borrow their money. For example, if the interest rate were a negative 2 percent, one could borrow $100 today and pay back $98 after one year. A better alternative for the lender is to hold the $100 in cash. Figure 1 shows a history of two interest rates: the yield on 10-year U.S. Treasury bonds in red and the overnight interest rate on bank reserves the federal funds rate in blue. The 10-year yield was between 2 and 2½ percent in It rose in fits and starts to over 14 percent in It rose so much in the 1960s and 1970s because the Federal Reserve allowed the inflation rate to rise. Neither the Fed nor other policymakers in the executive and legislative branches understood the full costs of inflation. The blue line depicts the overnight rate on bank reserves, which are deposits of commercial banks and thrifts at the Federal Reserve. These reserves are used to meet reserve requirements and to act as a buffer stock for funds flowing through the payment system. The federal funds rate is the rate that banks pay to each other to borrow funds on deposit at the Fed. (Note that this is not the discount rate, which is the rate that banks pay when they borrow directly from the Fed.) 1 William T. Gavin is a Vice President and Economist in the Research Department at the Federal Reserve Bank of St. Louis. The ideas and opinions in this lecture are his and do not necessarily reflect official positions of the Federal Reserve Bank of St. Louis, the Federal Reserve System, or the Board of Governors. 2

3 The Federal Reserve can control the federal funds rate because it is the monopoly supplier of bank reserves. As you can see, before 1980 the federal funds rate was often above the 10-year rate. The reason was that the Fed was in a regime in which inflation was not controlled. The Fed would lower interest rates to stimulate the economy and inflation would rise. As inflation became a problem the Fed would have to raise rates very high to discourage borrowing and spending, but inflation kept going up because the Fed was not clear about its responsibility to control inflation. Notice that there were two periods before 1980 in which the Fed kept the federal funds rate relatively low as the expansion proceeded, in 1971 to 1973 and again in 1976 to In both of these episodes the Fed kept the rate low in an attempt to speed up the recovery. In both instances, the economy did not grow faster, but inflation did. The lesson learned then was that the Fed could not control inflation if it relied too much on low interest rates to stimulate the economy. As we will see, that lesson was not quite correct. It all depends on what people think the Fed is going to do in the future. In the 1970s, the public did not trust the Fed to control inflation. In October 1979, a crisis occurred in the international currency markets: The dollar was falling relative to other major currencies, and the international community insisted that the United States control inflation and restore confidence in the dollar. Fed Chairman Paul Volcker responded with a dramatic change in the policy procedures that shifted control from interest rates to bank reserves. This change allowed the federal funds rate to trade above 18 percent during a period of double digit unemployment rates. This drastic policy worked. It stopped inflation from rising and began a 30-year trend of declining inflation and bond yields. A key point to remember is that inflation expectations are the main determinant of the trend in bond yields. Paul Volcker retired in 1987 and Alan Greenspan became the Fed Chairman. During his reign federal funds rarely traded at a rate higher than the 10-year bond yield. In the period from 1992 to 1994, the Fed held the federal funds rate well below the 10-year bond yield during the recovery from the 1991 recession. Again, following the 2001 recession, the Fed held the overnight rate well below the 10-year yield. Both episodes were characterized as jobless recoveries and the federal funds rate was held down in an attempt to stimulate faster real growth. During this period, speeches by Federal Reserve officials indicated that, although the Federal Reserve did not have an explicit inflation target, they wanted inflation to be low and would do whatever was necessary to prevent a 1970s rerun of high and rising inflation. It worked: Interest rates remained relatively low, and inflation appeared to remain under control. Inflation in consumer prices has been modest about 1 percent on average over the past year. The outlook for inflation is more of the same, inflation at or below this rate. The Fed s inflation objective (1.7 to 2 percent) is revealed in the range of Federal Open Market Committee (FOMC) forecasts for longer-run inflation. Countries that have been explicit about their inflation 3

4 objective Germany 25 years ago, New Zealand and Canada 20 years ago, and many others since have chosen targets at or near 2 percent. Although the Fed does not have an explicit inflation objective, it appears from recent policy statements that the FOMC begins to worry about deflation when reported inflation falls as low as 1 percent. The bottom line is that the Fed feels that in this situation, with inflation on the low side of 2 percent and the unemployment rate at 9.6 percent, it would like to lower the interest rate. The problem is that the federal funds rate is already at zero. However, there are good reasons to worry about more monetary stimulus. Some people worry about runaway inflation, pointing to rapid money growth excess reserves have grown by $1 trillion and the government is expected to run large budget deficits far into the future. But other people worry about deflation. 2 The Fisher equation gives some insight into this concern. The Fisher equation states that the nominal interest rate is equal to the real interest rate plus the expected inflation rate. Real interest rates tend to go up and down with economic growth. Currently, the economy is growing, recovering from a deep recession, and real interest rates are rising. If the Fed holds the federal funds rate at zero as real interest rates rise, the inflation rate will eventually fall. If inflation is falling and the economy remains short of full employment, the political reality is that the Fed cannot raise interest rates. This happened in Japan in the mid 1990s. The longer the Bank of Japan held the overnight rate at zero, the lower inflation fell and the lower bond yields fell. Eventually, people stopped forecasting that inflation and the overnight rate would rise and the Japanese economy settled into an equilibrium with short-term rates at zero, government bond yields at 2 percent, and a slight deflation less than 1percent per year. The Japanese people and their government adapted to that outcome by accepting slight deflation as part of the policy objective. The Monetary Policy Process and the Zero Lower Bound To analyze policy options facing the Fed today, we use a model that has it origins in Kydland and Prescott (1982) one of the two papers cited in the 2004 Nobel prize in economics. 3 The modified version I am going to use is one that is widely used by central bank policy advisors all around the world. This model assumes that prices do not adjust immediately to clear markets contrary to the original version by Kydland and Prescott but it doesn t make any difference for the results here. I report only effects of monetary policy on inflation and market interest rates. These results hold in virtually all forward-looking macroeconomic models, which may have different implications for the effect of policy on real variables such as GDP and job growth, but 2 For an essay on this topic see Bullard (2010). 3 Our computational model is a slightly modified version of the one described in detail in Gavin, Keen, and Pakko (2005). See Reifschneider and Williams (2000) and Coibion, Gorodnichenko, and Weiland (2010) for examples of different models that produces results consistent with those reported here. 4

5 which have nearly identical implications for the effect of monetary policy on interest rates and inflation. To analyze monetary policy in a model, it must be represented as an equation. The Taylor rule is widely used both inside and outside the Fed to represent policy. It is given as: R R y y * * * t 1 t 1 ( t t ) y( t t) (1) where R is the federal funds rate that is, the interest rate that banks with excess reserves charge to lend to other banks with reserve shortfalls. It is this interbank lending rate that the Fed can control because it is the monopoly supplier of bank reserves. Here we use π to represent the inflation rate and y to represent the logarithmic level of GDP. The starred terms represent the equilibrium levels of the federal funds rate, the inflation target, and the level of GDP that is expected when the economy is operating at full employment. This concept is called potential GDP and is constructed and published by the Congressional Budget Office. The equation represents a policy rule in which the central bank reacts to both deviations of inflation from target (inflation gap) and output from potential (output gap). John Taylor (1993) showed that this rule did a fairly good job of replicating the history of the federal funds rate during the early Greenspan years (1987 to 1992), if he set the inflation target at 2 percent and equal weight on both deviations (θ π = θ y = 0.5). The Taylor rule has become a popular way to characterize central banks use of short-term money market interest rates to implement monetary policy when the central bank has a dual mandate in the United States the legislation instructs the Fed to achieve price stability and to promote full employment. The problem for the Fed with this part of the mandate is that there are no levers to pull that will directly promote job growth. The best the Fed can do is to create a monetary and financial environment that promotes the greatest economic efficiency. There is also a problem with using this rule because the Fed does not have an explicit inflation target. If the target is vague or subject to shifts, then the response to the policy will be different than if the Fed is committed to an explicit inflation objective. We represent this vagueness by allowing the inflation target to fluctuate around 2 percent in response to economic shocks and the Fed s reaction to them. The upshot is the inflation rate can wander away from 2 percent for long periods at a time. The Computational Experiments. Our models include utility-maximizing households, profitmaximizing firms, and a government that uses some version of the Taylor rule to operate monetary policy. The model includes economic shocks to technology, to aggregate demand, and to monetary policy. The size and pattern of shocks are calibrated to match the volatility in U.S. output, inflation, and interest rates. In each of these experiments, we assume that the government has an inflation target of 2 percent, but some policies do better than others in achieving the 5

6 target. Some of the policies have wider swings in interest rates than others, so are more likely to hit the zero lower bound. In each case we run the experiments for 2500 years or 10,000 quarters. We then record the number of times that the model predicts zero interest rates. We also record the number and length of episodes in which the interest rate stays at zero for more than one quarter. Results are displayed in histograms, which report the number of episodes on the vertical axis and the length of the episodes on the horizontal axis. Note that the current episode, which began in December 2008, is already 2 years old. Japan has been at the zero lower bound for about 15 years. In our first set of experiments, we look at how putting more or less weight on the output gap or in other words, trying to fulfill the full employment part of the mandate more or less agressively affects the likelihood of hitting the zero lower bound. The best we can do to avoid the zero lower bound is to put no weight on output or employment. Using the Taylor rule with no weight on the output gap results in hitting the zero lower bound about 4 percent of the time. Using John Taylor s suggestion, which puts equal weight (0.5) on both output and inflation, does almost as well. It results in hitting the zero lower bound 6 percent of the time, with a few more long episodes lasting as long as five years. Business economists usually estimate that the Fed s weight on the output gap is larger than recommended by Taylor, perhaps as large as 1 or 2: With θ y = 2 the model predicts that the interest rate is at the zero lower bound 16 percent of the time. As Figure 2 shows, putting more weight on the output gap causes wider swings in the federal funds rate and greatly increases the chances of hitting the zero lower bound. If the weight is as large as 2 with an inflation target of 2 percent, the interest rate is quite likely to hit the zero lower bound in every recession. The problem with the Taylor rule is that it targets the short-run inflation rate. When the Fed misses the target for any reason, the target miss is forgiven and the future target is kept at the target rate. The reason the dual mandate causes such wide swings in the interest rate and the inflation rate is because output and employment are much more volatile than inflation. By putting more weight on output, the policy transmits the fluctuations in output and employment into wide swings in the inflation rate and interest rates. The Taylor rule offers no remedy for these fluctuations. The solution is for the central bank to find a way to commit to offsetting the swings in inflation that are induced by its mandate to promote full employment. Kydland and Prescott (1977) showed that it can be difficult, if not impossible, for governments to commit to following good long-run policies if the optimal short-run policy runs counter to it. A clear example is the government s policy to insure homes destroyed by natural disasters. Take flood insurance: When a flood destroys some homes, we all want to help the victims of the flood. The problem is that when people see the government rebuilding flooded communities, they are more likely to build homes in the flood plains. People like to live near water. The government can say that they will not help next time, but it is not credible. The next time a flood comes, we help the victims. Since it cannot commit to bailing out flood victims, the government now has regulations that prohibit building in flood plains. 6

7 As long as people believe that there is a tradeoff between full employment and inflation, they will always want more inflation when there is some unemployment. Without an explicit target, it is easy to let the inflation objective move up and down with shocks to the economy. If these shifts are not offset, then they induce uncertainty about inflation and interest rates at all horizons. Recent research has shown that a price path can be used as a commitment mechanism for an inflation target. 4 Suppose the Fed announced a 2 percent inflation target and also announced that it would adjust the short-run target to achieve the target inflation rate on average over longer periods. This is equivalent to announcing a path for the price level growing at 2 percent each year. Surveys of business economists show that the median respondent predicts that the inflation rate will be perhaps as low as 1 percent over the coming year. If the Fed had a price path target growing 2 percent per year and inflation came in at 1 percent in 2010, then the Fed would be expected to aim at a rate somewhat above 2 percent until the average rate went back up to 2 percent. If this policy is credible, interest rates will rise immediately to reflect the higher inflation expectations. There is also a problem with current policy that is revealed in surveys of individual economists inflation forecasts. Not only has the average forecast come in at the low end of the FOMC s desired range, but also the individual forecasts are becoming more disperse some economists are forecasting higher inflation because of faster growth in bank reserves and other economists are forecasting lower inflation, perhaps because of the Fisher equation. Ideally, everyone should be on the same page, expecting inflation to be equal to the Fed s inflation objective. To modify the Taylor rule to take account of a commitment to the inflation target, we add a term for the price level path target: R R y y p p * * * * t 1 t 1 ( t ) y( t t) p( t t), (2) * * * where p is the logarithm of the price level. Note also that p p and p p. 7 t t t 1 t t 1 The third term is the gap between the current price level and the price level path target. In equation (1) there is a time subscript on the inflation target because the Federal Reserve does not have an explicit inflation target. Consequently, the inflation target has evolved over time and with shifts in the economy. In equation (2) the inflation target is a constant. The price path is a reference line that loses much of its operational content as an indicator if the inflation target is allowed to shift from period to period. In the experiments reported in Figure 3, the policymaker is aggressively pursuing the full employment mandate the weight on the output gap is 2. Results from the Taylor rule from Figure 2 are repeated here as a benchmark to show how much policy improves when we add a price level path target with a weight of 0.1 on a price gap. Doing so almost eliminates the 4 Svensson (1999) shows that the discretion solution to a price path target is isomorphic to the commitment solution to an inflation target.

8 likelihood of hitting the zero lower bound. Choosing a weight of 0.2 resulted in no occurrences in 10,000 quarters, there was not a single instance of hitting the zero lower bound. To summarize the key results from our computational experiments, we find that one can avoid the zero lower bound by giving up the dual mandate, something that is not considered politically feasible, or one can commit to an inflation target. In fact, committing to an inflation target is the only way that we can avoid the zero lower bound and achieve the dual mandate in forwardlooking macroeconomic models. So why hasn t the Fed adopted an inflation target or better yet, a long-run average inflation target? The answer is simply that there are serious misunderstandings about inflation targeting. Four Myths about Inflation Targeting Myth 1: The Fed does not have an inflation target. Current and former members of the FOMC have argued that the Fed does not need a numerical objective for inflation. They argue that the numbers are uncertain and subject to revision. Different inflation indexes give different signals, even over extended periods, and there is no clear consensus among the members about what the objective should be. But the U.S. dollar is a paper currency. The inflation rate is determined by Fed policy. The policy it chooses leads to an expected inflation objective that is state contingent. By state contingent I mean that expectations about the Fed s inflation objective go up and down with news about output, inflation, interest rates and fiscal policy. These expectations also can change with changes in the membership of the FOMC and other positions of political leadership. A just do it target works fine when there are only small shocks to the economy. As recent experience shows, it doesn t work so well when shocks are big. Myth 2: A price path target would increase the chances of having episodes of deflation. This is just not true. In all our experiments, we also saved the results for inflation. As we did with the zero lower bound on interest rates, we recorded episodes of negative inflation deflation. Figure 4 shows two cases of the Taylor rule with the aggressive output gap policy and alternative weights on a price path. The Taylor rule in its original form (but with a higher weight on output) has more incidences of deflation, especially long periods of deflation for up to a decade. Putting weight on the price gap reduces the incidence of deflation. Myth 3: A price path target is not an inflation target. It is arithmetic. The inflation rate is the percentage change in the price level. A price path is a mechanism for making the short-run inflation target contingent on past errors in the targeting process. It works to reduce volatility in interest rates and inflation by conditioning expectations to be consistent with long-run policy objectives. 8

9 Myth 4: The Fed cannot choose a long-run average inflation target because it hasn t yet tried a short-run target. There is one counterexample in history. The Riksbank in Sweden abandoned the gold standard in 1931, and with advice from Knut Wicksell created a consumer price index and targeted it from September 1931 until World War II broke out. 5 As a matter of record, Sweden had the mildest downturn of any developed economy during the 1930s. The other interesting evidence can be found in recent examples of inflation targeting. Bean et al. (2010) argues that people have confused inflation targets with price path targets; that is, they believe that the annual inflation targets are actually long-run average inflation targets. This alone suggests that people will have no trouble accepting that the central bank has a multiyear inflation objective. Another way to interpret recent experience is that people have reacted to inflation targeting as if central banks were committed to the policy. Conclusions We argue that the current economic and policy situation is the direct result of policy aimed at dual objectives for price stability and full employment in an environment with an uncertain inflation objective. Once the interest rate hits zero, a policymaker using the Taylor rule has a problem managing expectations about future inflation. The problem is that there is a stable outcome with zero interest rates and a mild deflation trend. An example of this outcome can be seen in Japan (1995 to 2010). This is not necessarily a bad outcome as long as people expect the inflation rate to be zero or less. There is also a concern that attempting to expand the supply of bank reserves with very large purchases of Treasury securities may lead to a replay of the 1970s, during which the Federal Reserve and other government officials appeared to ignore the inflationary consequences of excessive money growth. Both sets of beliefs appear to exist simultaneously in today s market as surveys of inflation expectations show more dispersion in beliefs about future inflation. The only way to achieve the dual mandate in the forward-looking models used at central banks is to commit to a clear inflation objective. Managing expectations is the key to successful monetary policy. In this talk I have argued that the key is to manage expectations about the long-run average inflation rate. We show that trying to pursue a dual mandate for price stability and full employment will likely lead an economy to the zero lower bound if the central bank is not committed to an inflation objective. 5 See Black and Gavin (1990) for a description of this price level targeting policy. 9

10 References Bean, Charles, Matthias Paustian, Adrian Penalver and Tim Taylor. Monetary Policy after the Fall, paper presented at the Federal Reserve Bank of Kansas City Annual Conference, Jackson Hole, Wyoming, 16 September Black, Susan, and William T. Gavin. "Price Stability and the Swedish Monetary Experiment," Federal Reserve Bank of Cleveland Economic Commentary, December 15, Bullard, James. Seven Faces of The Peril, Federal Reserve Bank of St. Louis Review, September/October 2010 pp Coibion, Olivier, Yuriy Gorodnichenko, and Johannes F. Wieland, The Optimal Inflation Rate in New Keynesian Models, NBER Working Paper 16093, June Gavin, William T., Benjamin D. Keen and Michael R. Pakko The Monetary Instrument Matters, Federal Reserve Bank of St. Louis Review, September/October 2005, 87(5), pp Kydland, Finn E., and Edward C. Prescott Rules rather than discretion: The inconsistency of optimal plans, Journal of Political Economy 85 (1977), Kydland, Finn E., and Edward C. Prescott. Time to Build and Aggregate Fluctuations, Econometrica 50(6) (November 1982), Reifschneider, David, and John C. Williams, Three Lessons for Monetary Policy in a Low-Inflation Era, Journal of Money, Credit, and Banking 32 (November 2000, pt. 2), Svensson, Lars E.O. "Price Level Targeting vs. Inflation Targeting: A Free Lunch?" Journal of Money, Credit, and Banking (August 1999, pt. 1), Taylor, John B. "Discretion versus Policy Rules in Practice," Carnegie-Rochester Conference Series on Public Policy 39 (1993),

11 Figure 1. US Interest Rates: 1955 to Percent Annual Rate Overnight Federal Funds Rate 10 Year Treasury Bond Yield 11

12 Figure 2. Histogram of Zero Lower Bound Episodes and the Weight on the Output Gap Frequency of Episode of length X θ y = 2 θ y = 1 θ y = 0.5 θ y = More Length of Episode X (quarters) 12

13 Figure 3. Histogram of Zero Lower Bound Episodes with Some Weight on the Price Gap Frequency of Episode of length X Taylor rule with weight θ y = 2 and θ p = 0 Taylor rule with weight θ y = 2 and θ p = More Length of Episode X (quarters) 13

14 Figure 4. Histogram of Deflation Episodes and Price Level Path Targeting 70 Frequency of Episode of length X Taylor rule with weight θ y = 2 and θ p = 0 Taylor rule with weight θ y = 2 and θ p = Length of Episode X (quarters) 14