流動性の罠とデフレーションからの脱出：The Foolproof Way 他
by Lars E. O. Svensson
Existing proposals to escape from a liquidity trap and deflation, including my “Foolproof Way,” are discussed in the light of the optimal way to escape. The optimal way involves three elements: (1) an explicit central-bank commitment to a higher future price level; (2) a concrete action that demonstrates the central bank’s commitment, induces expectations of a higher future price level and jump-starts the economy; and (3) an exit strategy that specifies when and how to get back to normal. A currency depreciation is a direct consequence of expectations of a higher future price level and hence an excellent indicator of those expectations. Furthermore, an intentional currency depreciation and a crawling peg, as in the Foolproof Way, can implement the optimal way and, in particular, induce the desired expectations of a higher future price level. I conclude that the Foolproof Way is likely to work well for Japan, which is in a liquidity trap now, as well as for the euro area and the United States, in case either would fall into a liquidity trap in the future.
In the last decade or two, central banks all over the world have been quite successful in achieving low and stable inflation. Average annual inflation in the industrialized countries has fallen to a level below 2 percent. Average inflation has not been so low since the 1950s, as shown in Table 1. In emerging countries, inflation is now the lowest since the 1960s (International Monetary Fund (2003b)).
Table 1. Average annual inflation in industrial countries
(consumer price index, percent)
(Source: International Monetary Fund (2003b).)
These gains against inflation are good news. They have brought substantial benefits in terms of reduced distortions, less uncertainty, and improved resource allocation. But they also raise new risks. Unanticipated negative shocks to demand or supply can always cause recessions and lower inflation − and, starting from a low inflation level, even deflation. In such a situation, the appropriate response by central banks is to lower interest rates and this way stimulate the economy out of recession and too low inflation. But with low inflation or even deflation, a negative interest rate may be required to provide sufficient stimulus to the economy, whereas nominal interest rates cannot fall below zero. The economy might then become caught in a liquidity trap and a prolonged recession and deflation.
This paper begins with a discussion of the causes and consequences of a liquidity trap and deflation, with some emphasis to Japan’s experience since the 1990s. It then discusses policy options for preventing a liquidity trap and deflation from occurring and for escaping from a liquidity trap and deflation if they have already occurred. Whereas policy for avoiding a liquidity trap and deflation is less controversial, there is a fair amount of controversy about the range of policies to escape from a liquidity trap and deflation, including my own proposal, the “Foolproof Way” (Svensson (2001, 2002)).
Causes and Consequences of a Liquidity Trap and Deflation
How can a liquidity trap and deflation arise, and why are they a problem? An increasing number of central banks aim both to stabilize inflation around a low level and to keep output close to its potential level. But monetary policy operates under considerable and unavoidable uncertainty about the state of the economy and the size and lag of the economy’s response to monetary-policy actions. Unanticipated shocks to demand and supply are unavoidable. Because of the lags in the effect of monetary-policy actions, good central banks are forward-looking, use available information about the economy and anticipated shocks to construct forecasts of inflation and output, and respond to these forecasts as best as they can. A substantial realized or anticipated negative shock to aggregate demand, because of, for instance, the bursting of an asset-price bubble, a correction of overoptimistic growth and productivity expectations, increased doubts about future pensions and benefits due to demographic developments and/or reckless fiscal policy, increased uncertainty for geopolitical or other reasons, will lower both actual inflation and output as well as forecasts of future inflation and output. If initial inflation is low, this may be all that is needed for not only a temporary recession but a temporary deflation.
When central-bank forecasts indicate recession and too low inflation or even deflation, the appropriate response is to lower the central bank’s “instrument rate,” the short(-maturity) nominal interest rate it uses to implement monetary policy – the federal funds rate in the United States. A lower short nominal interest rate, combined with sluggish private-sector inflation expectations, will lowers the short real interest rate – the nominal rate less expected inflation. Expectations of lower future short real rates then lower longer(-maturity) real rates, the rates that matter for consumption and investment decisions and thereby aggregate demand. The lower real rates, with some lag, stimulate aggregate demand and output and bring the economy out of recession. Increased aggregate demand and increased inflation expectations then increase actual inflation, also with some lag. On occasion, a competent or lucky central bank may even be able to preempt the recession and too low inflation more or less completely. Such successful preemption is a central banker’s dream.
If the nominal interest rate is initially low, which it is when inflation and expected future inflation are low, the central bank does not have much room to lower the interest rate further. But with deflation and expectations of deflation, even a nominal interest rate of zero percent can result in a substantially positive real interest rate that is higher than the level required to stimulate the economy out of recession and deflation. Nominal interest rates cannot fall below zero, since potential lenders would then hold cash rather than lend at negative interest rates. This is the so- called “zero lower bound for interest rates.”
In particular, conventional monetary policy seems unable to provide sufficient stimulus to the economy and address recession and deflation once the zero lower bound for interest rates has been reached. The problem is that the economy is then satiated with liquidity and the private sector is effectively indifferent between holding zero-interest-rate Treasury bills and money. In this situation, standard open-market operations by the central bank to expand the monetary base by buying Treasury bills lead the private sector to hold fewer Treasury bills and more money – but this has no effect on prices and quantities in the economy. When this “liquidity trap” occurs, expanding liquidity (the monetary base) beyond the satiation point has no effect. If a combination of a liquidity trap and deflation causes the real interest rate to remain too high, the economy may sink further into a prolonged recession and deflation.
Prolonged deflation can have severe negative consequences. The real value of nominal debt rises, which may cause bankruptcies for indebted firms and households and a fall in asset prices. Commercial banks’ balance sheets deteriorate when collateral loses value and loans turn bad, and financial instability may threaten. Unemployment may rise, and if nominal wages are rigid downwards, deflation means that real wages do not fall but increase, further increasing unemployment. All this may contribute to a further fall in aggregate demand, a further increase in deflation, a further increase in the real interest rate, and bring prices and the economy down in a deflationary spiral. Therefore, a liquidity trap with the associated risk of a prolonged recession or even a deflationary spiral is a central banker’s nightmare.
Japan’s recent experience provides a stark warning of the dangers of a liquidity trap and deflation. Japan has already lost a decade to economic stagnation and deflation. Without effective policy measures, it may very well lose another decade. Whatever the reasons for Japan’s initial recession and stagnation, most observers of Japan’s experience have concluded that the reason for the prolonged stagnation and deflation is due to policy mistakes and an inability to take decisive and coordinated action to resolve Japan’s problems. The policy measures that have been tried have not succeeded in ending stagnation and deflation. Expansive fiscal policy, with a big fiscal deficit, has not ended stagnation but has lead to huge national debt, close to 150 percent of GDP at the end of 2001 and still increasing (International Monetary Fund (2002)). With regard to monetary policy, Bank of Japan lowered the interest rate to zero and kept it there from February 1999 to August 2000, and again from March 2001 until now. From March 2001, after long indecisiveness, it also attempted a so-called “quantitative easing,” a substantial expansion of the monetary base. During two years up to the spring of 2003, the monetary base was increased by about 50 percent (Bank of Japan (2003)). But these steps were not sufficient to induce a recovery. As the Japanese economy faces expectations of further deflation, the real interest rate remains positive and too high. But, as will be discussed and as many frustrated observers have repeated, there are more effective policy measures for ending recession and deflation that the Japanese authorities have declined to apply.
Deflation in Japan measured with the GDP deflator began in 1995. Since 1999 the GDP deflator has been falling at a rate of between 1 and 2 percent per year. Deflation in the consumer price index began in 1999, and since then the consumer price index has been falling at a rate slightly less than 1 percent per year (International Monetary Fund (2003a)). Thus, deflation in Japan is still relatively modest and has not started to increase dramatically. This indicates that the problem is not a dramatic deflation in itself but the recession, the zero lower bound and the liquidity trap preventing monetary policy to provide sufficient stimulus to the economy.
Japan does not only have a macroeconomic problem of recession and deflation; it also has many structural and microeconomic problems, especially in the financial sector (for instance, Kashyup (2002)). Ending recession and deflation is not a substitute for solving those structural problems and undertaking structural reforms. But it can be easier to solve those problems and undertake the necessary reforms in a growing economy with positive inflation. So far, the Japanese authorities have demonstrated the same inability to handle the structural problems as the macroeconomic ones.
During the Great Depression, deflation in the United states during the three years from 1930–1932 was more dramatic, about 10 percent per year. During the same time, industrial production fell by 50 percent and GDP by almost 30 percent. There is broad agreement that monetary factors and mistakes by the Federal Reserve played a crucial role both in the onset and prolongation of the Great Depression (Meltzer (2003) and International Monetary Fund (2003a)).
In the United States today, low inflation and a sluggish recovery from recession in 2002 and 2003 has also led commentators and policymakers to worry about the risk that new unfavorable shocks could topple the United States into a liquidity trap and even deflationary spiral (for example, Ahearne et al. (2002); Bernanke (2002)). In the euro area, low inflation and recession in Germany has led some commentators and policymakers to be concerned about the risk of Germany experiencing deflation (for example, The Economist (2002b); Issing (2002)). In May 2003, the International Monetary Fund issued a report from a special task force on deflation in the world (International Monetary Fund (2003a)) and a certain media frenzy was notable.
Escaping from a Liquidity Trap and Deflation
Many researchers and policymakers have recently discussed the consequences of the zero bound, a liquidity trap and deflation, how to avoid becoming trapped, and how to escape if trapped, often with specific references to Japan.2 There seems to be considerable agreement on how to avoid the zero bound and a liquidity trap and minimize the risk that it happens. Many papers recommend an explicit positive symmetric inflation target (say 2 percent per year), to give a sufficient margin to deflation. Many central banks already conduct forward-looking inflation targeting, trying to take preemptive actions if inflation forecasts are too low or too high relative to the inflation target. Another possibility is to set a target path for the price level in the future, perhaps rising at 2 percent per year, although no central bank currently implements explicit price-level targeting as distinct from price-level targeting (more on this below). Svensson (1999a) has suggested that prudent central banks should prepare in advance a set of emergency measures, to be used at preannounced indications of an imminent liquidity trap. Some of these emergency measures will be further discussed below.
Less agreement exists on how to escape from a liquidity trap and deflation, if the economy has already fallen into a liquidity trap and the real interest rate is too high for appropriate stimulus of the economy. This section will discuss a variety of practical proposals for such escape. These proposals include: announcing a positive inflation target; announcing a price-level target path; expanding the monetary base via open-market operations in Treasury bills and more unorthodox assets; reducing long interest rates via a ceiling on long interest rates or via a commitment to keep the instrument rate equal to zero for a substantial time in the future; depreciating the currency by foreign-exchange interventions; introducing a time-varying exchange-rate target; introducing a tax on money; introducing more expansionary fiscal policy; affecting intertemporal substitution of consumption and investment by time-variable tax rates; and, finally, a policy of combining a price- level target path, a currency depreciation and a crawling peg, and an exit strategy that makes up my Foolproof Way to escape from a liquidity trap.
The Optimal Way to Escape from a Liquidity Trap
Given that the central bank cannot reduce the nominal interest rate below zero, what is the best way to escape from the recession and deflation? The real interest rate is the difference between the nominal interest rate and expected inflation. Thus, even if the nominal interest rate is constant at zero, the central bank can affect the real interest rate, if it can affect private-sector inflation expectations. If the central bank could manipulate private-sector beliefs, it would make the private sector believe in future inflation, the real interest would fall, and the economy would soon emerge from recession and deflation.
The problem is that private-sector beliefs are not easy to affect. A few decades back, when inflation was high, central banks would often promise low future inflation, but the private sector often paid little attention. Often, high inflation continued to rule. Similarly, if a central bank in a liquidity trap promises high inflation in the future, the private sector may doubt either the ability or the will of the central bank to achieve that future inflation. The central bank may be tempted to cheat, that is, to promise high future inflation to get out of the liquidity trap, but once out renege on the promise and keep inflation low. Indeed, the situation can be described as one of multiple equilibria. If the private sector is pessimistic and expects deflation, the real interest rate will remain high and the recession and deflation will be longer. If the private sector is optimistic and expects deflation to be replaced by inflation, the real interest rate will be lower and the recession and deflation will be shorter.
Let us consider the best possible rational-expectations equilibrium in this situation, that is, when the private sector believes in the central bank’s promise and the central bank lives up to its promise. Suppose that the central bank prefers to keep inflation close to a given small but positive explicit or implicit inflation target and output close to potential output. In the recession and deflation, output is below potential and inflation is below target. Sometime in the future, the liquidity trap will end, inflation will return close to target, and output will return close to potential. For the bank, it would be better to overshoot the inflation target intentionally in the future, since this would correspond to higher inflation expectations and a lower real interest rate and help the economy out of the current liquidity trap. The loss of higher-than-target future inflation would be compensated by higher output and less deflation in the current liquidity trap.
Thus, the best possible rational-expectations equilibrium is one where the central bank intentionally conducts more expansionary policy and causes a higher inflation in the future so as to shorten the current recession and deflation. This policy also implies keeping the nominal interest rate at zero for some period even after the recession and deflation is over. Rational private-sector expectations of this policy will then lower the real interest rate in the liquidity trap. The basic insight into the nature of this optimal policy is due to Krugman (1998). The precise derivation of the optimal policy in some specific circumstances is presented in Jung, Teranishi and Watanabe (2001) and Eggertsson and Woodford (2003).
As Krugman has emphasized, the problem is that this optimal policy may not be credible. Once the recession and deflation is over, the central bank may renege on its promise of a future expansion and instead keep inflation low and close to its target rate. Indeed, if the private sector’s preferences agree with the bank’s, the private sector would also prefer that, once the recession and deflation is over, inflation is held close to its low target rate. But if this outcome is anticipated, private-sector inflation expectations will remain low and the recession and deflation will be longer. The central bank would need to commit itself to the future monetary expansion, and also communicate this commitment to the private sector. But with the normal instrument, the instrument rate, already constant at zero, it is difficult to demonstrate any commitment. Thus, it is natural to discuss
proposals for ending deflation according to the nature of the commitment to future monetary expansion they involve, and how such commitment can be demonstrated in the current period and be effective in inducing private-sector expectations of a higher future inflation.
Announcing a Positive Inflation Target or a Price-Level Target Path
Several authors have proposed the announcement of a sufficiently positive inflation target as a commitment to a higher future inflation rate (for instance, Bernanke (2000); Krugman (1998); Posen (1998)). In line with the optimal policy of a future overshooting of the normal inflation target, this target should be higher than normal for a few years. Krugman (1998) has stated that the central bank should “credibly promise to be irresponsible,” by which he means setting an inflation target higher than might otherwise be desirable. For Japan, Krugman (1998) has suggested a relatively high 4 percent inflation target for 15 years. Posen (1998) has suggested a more modest initial inflation target of 3 percent, to be reduced to 2 percent after a few years.
However, the mere announcement of an inflation target and a future monetary expansion need not be credible with the private sector and therefore need not affect inflation expectations, in the absence of any commitment mechanism or any action supporting the announcement. Thus, this method is therefore more likely to work if it includes published inflation forecasts, transparent inflation reports, public hearings, and other elements increasing the commitment to the inflation target. Even so, the private sector may expect a higher than normal inflation target to be adjusted downward once the liquidity trap is over. In particular, for a central bank like the Bank of Japan or the Federal Reserve, that have for many years publicly resisted announcing an inflation target, the announcement of any inflation target may be interpreted as an unconvincing “gallows speech,” to be disregarded when the liquidity trap is over.
Another possibility is to announce an upward-sloping target path for the price level, perhaps rising at 1–2 percent per year, as suggested for Japan in Svensson (2001) and more recently by Bernanke (2003). The practical difference between these two approaches is that if inflation falls short of the inflation target in one year, the inflation target for the next year does not change. However, with a price level target, lower inflation in one year must be counterbalanced by a higher rate of inflation in future years to return to the desired price level path. In the context of escaping from a liquidity trap, a price-level target offers an advantage above an inflation target, since long-term inflation expectations matter more than short term. Long real interest rates are long nominal rates less long- term inflation expectations. If a central bank with an inflation target is expected to undershoot its inflation target for a couple of years and then return to it (which a central bank in a liquidity trap might be expected to do), then long-term average inflation is lower, since the bank does not compensate in the future for past misses. However, if a central bank with a price-level target is expected to undershoot its target for a couple of years and later return to it, long-term inflation expectations are unaffected by the initial misses. Furthermore, if deflation occurs and the price level falls further below the target, inflation and inflation expectations will rise to get back to target. Thus, further deflation automatically lowers the real interest rate even if the nominal rate is constant (at zero, for instance).
A price-level target path could even start above the current price level with a “price gap” to undo. As emphasized by Bernanke (2000, 2003), several years of zero or negative deflation may have resulted in a price level below previous expectations that has increased the real value of debt and deteriorated balance sheets for banks and firms. For Japan, this price gap may be 10–15 percent or more.
Thus, a price-level target, if credible, has an advantage in corresponding to more desirable long- term inflation expectations as well as the undoing of a price gap. Indeed, Eggertsson and Woodford (2003) show that the optimal rational-expectations equilibrium for escaping from the liquidity trap is best expressed as a price-level target path rather than an inflation target that disregards past misses. Indeed, they argue that, from a credibility point of view, it is better to follow a price-level target before a liquidity trap occurs than to announce it once the liquidity trap occurs. This is a general argument in favor of price-level targeting rather than inflation targeting. No central bank currently implements explicit price-level targeting, although Sweden did so during part of the 1930s (Berg and Jonung (1999)).3
Announcing an inflation target or a price-level target will lower the real interest rate and be expansionary only to the extent that the targets are credible with the private sector. Since the standard policy tool of a lower short interest rate is neutralized by the zero bound, it is natural to look for other instruments of monetary policy that can potentially demonstrate the central bank’s commitment.
Expanding the Monetary Base
Although the zero lower bound prevents lowering the nominal interest rate below zero, the central bank can still expand the monetary base (Benhabib, Schmitt-Grohé and Uribe (2002); Bernanke (2000); Clouse et al. (2003); Meltzer (2001); Orphanides and Wieland (2000)). However, the precise mechanism through which an expanded monetary base will alter expectations is not altogether clear.
For example, Meltzer (2001) suggests that an expanded monetary base will affect a number of other asset prices and interest rates in an expansionary direction, even if short nominal interest rates are zero, especially depreciating the domestic currency. But, in a liquidity trap, Treasury bills and money are approximately perfect substitutes, and open-market operations increasing private holdings of money and reducing private holdings of Treasury bills would have little or no effects on other asset prices and interest rates.
Therefore, an expansion of the monetary base would increase inflation expectations and reduce the real interest rate only if it is seen as a permanent expansion. Indeed, Krugman (1998) expressed the desirable future monetary expansion in terms of an increased future money supply. In principle, the central bank could expand the monetary base without limit, by continually buying domestic and foreign government debt, and if these are exhausted, other domestic and foreign assets. Such a dramatic policy would eventually affect private-sector expectations and have a dramatic effect on the domestic price level and the exchange rate and certainly put an end to deflation.
The problem is, again, why an expansion of the monetary base today should be viewed as a commitment to increased money supply in the future. While the liquidity trap lasts and the interest rate is zero, the demand for monetary base is perfectly elastic and excess liquidity is easily absorbed by the private sector. However, once the liquidity trap is over and the nominal interest rate is positive, demand for money will shrink drastically, in most cases requiring a drastic reduction of the monetary base. It is difficult to assess how much the monetary base would have to be expanded before inflation expectations and inflation take off. Beyond some unknown threshold, deflation may be quickly replaced by hyperinflation. As noted above, the Bank of Japan has expanded the monetary base by about 50 percent in the two years prior to the summer of 2003; given this step, it will definitely have to contract the monetary base once the liquidity trap is over. Thus, a commitment not to reduce the monetary base at all in the future is not credible, but a commitment to reduce it by less than otherwise is a more complex matter.4 The private sector may anticipate that the central bank will immediately back off any expansion of the monetary base if it fears igniting inflation, which in turn could make the initial commitment to monetary-base expansion not credible, implying that initial monetary base expansion has little or no effect; as has indeed been the case for the substantial increase of the monetary base in Japan.
Reducing Long Interest Rates
Even if short nominal interest rates are zero in a liquidity trap, long nominal interest rates need not be. As already noted, it is longer real interest rates, rather than short real rates, that affect consumption and investment decisions. Thus, a reduction of long nominal interest rates could, everything else equal, reduce long real rates and hence be expansionary and contribute to an escape from the liquidity trap. Several researchers and policymakers have therefore suggested open- market operations in long bonds as a way of reducing long interest rates (for instance, Clouse et al. (2003); Lebow (1993); Meltzer (2001)).
It is difficult to determine how large an open-market operation would be needed to reduce the long interest rate, because of difficulties in estimating the determinants of the term premium of interest rates (that is, the difference between long and short interest rates and its dependence on the degree of substitutability between short and long bonds). However, Bernanke (2002) has proposed an elegant operational solution to this problem. The central bank simply announces a low (possibly zero) interest-rate ceiling for government bonds up to a particular maturity, and makes a commitment to buy an unlimited volume of those bonds (that is, potentially the whole outstanding volume) at that interest rate. This commitment by the central bank is readily verifiable – since everyone can verify that the central bank actually buys at the announced interest rates – and achieves the desired impact on the long interest rate, without a need to specify the precise magnitude of the open-market operation required. The central bank may have to buy the whole outstanding issue of the long bond, though.
Another way to reduce long bond rates, proposed by Orphanides and Wieland (2000), relies on the expectations hypothesis that long bond rates are related to expectations of future short nominal rates. They suggest a commitment by the central bank to maintain the short nominal interest rate at zero for a substantial time in the future, even if the economy recovers. This proposal is in line with the optimal way to escape from a liquidity trap that was described above, which involves a zero interest rate also after the economy has recovered. But as discussed earlier, it is not clear that this commitment can be made credible.
Even if the central bank may be able to reduce long bond rates, this may not provide sufficient stimulus to the economy. That is, without the creation of long-term inflation expectations, the resulting long real interest rate may still be too high.
A Tax on Money
Goodfriend (2002) and Buiter and Panigirtzoglou (1999) have proposed an unorthodox way of eliminating the zero bound on nominal interest rates by introducing a tax on money. Such a tax would allow negative nominal interest rates in equilibrium, and allow the central bank to achieve the desired stimulating negative interest rate.
It is technically feasible to introduce a tax on commercial-bank reserves in the central bank and on electronic money, such as consumer cash cards. However, introducing a tax on currency requires technological innovations like electronic chips in the notes or a lottery that determines what numbered notes in a series become worthless in each period. It could also imply the inconvenience of notes circulating with the same denomination but trading at different discounts. One might also anticipate some public resentment against a system that would makes some of the money in people’s pockets conspicuously worthless.
Fiscal policy is an obvious policy alternative in a liquidity trap, when traditional monetary policy is ineffectual. However, the effectiveness of this policy depends to a considerable extent on the reactions of the private sector.
For example, if the initial level of government debt is high and a higher debt is deemed unsustainable, a policy of higher government debt may cause the private sector to anticipate tax increases or government benefit reductions in the near future. A resulting increase in private-sector saving will then reduce any impact of the expansionary fiscal policy.
Depending on the degree of independence of the central bank, the private sector might also anticipate that increased budget deficits will be financed by the central bank, which would presumably lead to inflation expectations. However, in Japan, expansionary fiscal policy over a number of years has led to a dramatic increase in the government debt, without stimulating the economy out if its recession and liquidity trap. For Japan, a further bond-financed fiscal expansion may be neither effective nor accepted by lenders without substantial interest-rate increases, which would defeat the stimulus. A money-financed fiscal expansion – that is, a budget deficit financed by the printing press, or more precisely, by the central bank buying the government bonds issued to finance the deficit – may still be expansionary, since a money-financed fiscal expansion need not necessarily be followed by eventual tax increases or expenditure cuts. Bernanke (2003) proposes that a price-level target for Japan is combined with a money-financed fiscal expansion. But, again, the expansion of the money supply need even in this case not be permanent and credible, since in the future concern about too high inflation may induce the central bank to reduce the money supply and increase the outstanding government debt.
Fiscal policy can also be used in another way in a liquidity trap, namely to lower the real interest rate net of taxes and subsidies (Saxonhouse (1999); Feldstein (2002)). A temporary reduction in the value-added tax combined with a temporary investment tax credit will reduce the after-tax real interest rate. By combining these policies with a temporary surcharge on the income and corporate tax, these tax changes can be fully financed and need not affect the budget deficit. One potential problem with such temporary tax changes is that they need not be credible. That is, the private sector may believe that the government will not reverse the tax cut as soon in the future as promised, taking into the account that the government may be tempted to prolong any stimulating effect by postponing the reversal. But an anticipated more permanent tax reduction will have less effect on the after-tax real interest rate. From this point of view, a temporary tax reduction that is less than fully financed may be more credible.