William R White: Comments on Marvin Goodfriend's paper: Financial stability, deflation and monetary policy

Speech by William R White for the Ninth International Conference at the Institute of Monetary and Economic Studies, Bank of Japan, on 'The Role of Monetary Policy under Low Inflation: Deflationary Shocks and their Policy Responses', 3-4 July 2000

BIS speech  | 
03 July 2000

This was a very enjoyable paper to read, but one which covered a very wide range of economic territory. Since it would be impossible to evaluate its content systematically, let me reorder Goodfriend's contents with a view to addressing only two questions. First, how does a country get itself into a deflationary situation? There will be no surprise if I suggest that the answer can be succinctly expressed in two words: "boom" and "bust". Second, how does a country get itself out of a deflationary situation? With respect to each question, I will focus (albeit not exclusively) on Goodfriend's recommendations regarding monetary policy.

Getting into deflation?

Goodfriend makes a simple but fundamental point concerning the problems which can arise in the "boom" phase of an economic cycle. Justified optimism can turn into excessive optimism; rational enthusiasm can turn into irrational exuberance. The "good news" of low inflation can blind both policymakers and market participants to emerging problems. In Mexico in 1994, economic crisis followed a period of high consumption and rapid credit creation associated with welcome policies of fiscal restraint, structural reform and the privatisation of banks. In Asia in 1997, the promise of benefits arising from deregulation and globalisation led to rapid credit creation and unsustainable levels of investment. The experiences of the United States in the 1920s and Japan in the 1980s are no less relevant. In all of these cases, low inflation was not sufficient to prevent dangerous financial excesses from building up.

While I certainly agreed with Goodfriend's conclusion that low inflation is no panacea, he perhaps places too little emphasis on the deflationary dangers of rapid asset price increases and the associated vulnerability of the financial system; indeed, property prices are not mentioned at all in his paper. Goodfriend's basic story, valid as far as it goes, is rather one of "sticky" inflation (due to credibility effects) which finally accelerates, leading to the need for a sharp monetary policy response. Recession follows and, with inflation initially low, deflation is the next step.

While Goodfriend's narrative could be richer on the financial asset side, he does in the early part of his paper put welcome emphasis on associated problems of liquidity in financial markets. He notes, for example, that when "narrow liquidity" rises, "broad liquidity" rises too. I would interpret this statement as being consistent with what BIS publications have been saying for some years.1 That is, when monetary policy eases, credit and other risk spreads tend to narrow as a spirit of optimism begins to permeate the markets. In recent years, moreover, this tendency for risk to be underpriced has been exacerbated by the growing degree of competition in financial markets encouraged by the search for shareholder value, on the one hand, and the provision of public safety nets, on the other.

Confronted with such tendencies in the "boom" period, how should monetary policy react? Here I take issue with Goodfriend. He suggests (referring in a footnote to the Taylor Rule) that policymakers:

"should make every effort not to be fooled in the first place into being insufficiently preemptive during an economic expansion. To guard against this possibility a central bank should benchmark its policy against a rule that has performed reasonably well in the past".

This seems to me to miss the basic point. We begin with the assumption that CPI inflation is low, and normal rules would say there would be no need for a policy response. Yet, underlying this benign set of circumstances, there would nevertheless be a developing problem in the financial system which could threaten deflation if it were to materialise.

That being said, Goodfriend is more than aware that a central bank can become the victim of its own rhetoric. Having itself suggested that a low inflation rate is not just necessary, but also sufficient to ensure sustainable growth, how can a central bank raise interest rates in the face of low CPI inflation? Indeed, the dilemma is made even worse if one considers that moderate interest rate increases may not be sufficient to deal with these financial excesses. Sharply higher rates to deal with a bubble, which must be burst because it cannot be deflated, will have very little political support in a world conditioned to believe that "all is well".

What is also missing from Goodfriend's paper is the question of whether a central bank ought to burst an asset price bubble. This question has a long history and there is clearly no right answer. On the one hand, it is tempting to say that a bubble should be burst as soon as one knows it is a bubble. On the other hand, since such certainty will come only very late in the day, a cynic might feel a better course would be for policymakers to stay out of the way and simply let nature take its course. Indeed, this response might also be recommended if the policymaker felt there would always be some grounds for believing that observed prices could be justified in terms of underlying fundamentals.

The "bust" phase of the cycle can also lead to policy conundrums. Goodfriend rightly notes that the market dynamics evident in the upturn are likely to go sharply into reverse. As the economy weakens, whether due to higher interest rates or the simple weight of past excesses, credit and liquidity spreads will widen. Goodfriend emphasises in the early part of his paper the impact of lower asset prices on net worth and the attempt by affected market-makers to pass on their increased vulnerability in the form of higher transaction costs. This might well happen. However, the effects of margin calls in response to leverage, portfolio insurance, and a common response to shared risk management techniques (recommended to all by the supervisory authorities) might play an even bigger role in contributing to large and potentially disruptive price movements in financial markets. Markets do seem in recent years to have become less atomistic and potentially more vulnerable to overshooting as macroeconomic excesses resolve themselves.

Goodfriend's assessment of the macroeconomic processes likely to be in play during the downturn also seem sound. They have a rather Austrian flavour in that he emphasises over-investment during the boom phase, the necessity for retrenchment later, and the need to face up to the problems associated with declining asset values but unchanged nominal liabilities. Indeed, using recent experience in the United States as an example, one could easily identify even more causes for concern. The heavy buyback of shares by established "old economy" firms has been accompanied by a significant expansion of IPOs in "new economy" industries. The capacity of actual rates of return to meet investors' expectations remains to be tested. Consumers have also invested heavily in such durable expenditures as cars and houses, which makes further expenditures eminently postponable. And finally, the widening current account deficit in the United States implies that other aspects of the hangover from a boom will be a lower dollar, a negative terms-of-trade shock, and a further need to reduce domestic absorption. Together with all the financial strains likely to emerge in the "bust" phase of such a cycle, these developments could pose a significant challenge to policymakers.

In such circumstances, I would contend that some policy reactions might clearly be recommended but others are less clear-cut. On the one hand, lender of last resort facilities and emergency liquidity assistance should be available to institutions that need them provided they meet appropriate criteria. A more difficult question is what to do about asset prices in a crisis. In general terms, it is of course clear that no attempt should be made by policymakers to maintain asset prices at some predetermined level. I think most would agree that those in a capitalist system who have made errors of judgment should be made to pay the price of those misjudgments. But what is more complex is the issue of what central banks should do in the specific circumstances of a collapse of prices due to a drying up of liquidity in a particular market, perhaps even one of global importance? Should they act as a "market maker of last resort" or not? An even more important question is whether there should be a general reduction in interest rates in such circumstances. Goodfriend seems to me to take an even-handed view. He recognises that "hidden" inflation can emerge as the economy turns down, presenting an unpleasant picture of stagflation. Indeed, if the external value of the currency were to decline sharply in the case of an economic slowdown, this stagflation problem might be sharply accentuated. However, what receives less attention than it should in Goodfriend's analysis is the potentially different effects of monetary easing in different markets. For example, when the Federal Reserve rightly eased in the face of the liquidity problems faced by many financial markets in autumn 1998, equity prices rose sharply in many countries to even more unsustainable levels. Clearly, the actual monetary policy response had some desirable features but also some less desirable features. In this crisis stage of the cycle, there are again no easy answers.

Getting out of deflation?

The conventional wisdom would be that there should be more reliance on fiscal policy, since it works, and less reliance on monetary policy, since it does not work given the assumed existence of a liquidity trap. Goodfriend takes a defiantly contrary view. While I agree with a great deal of what he says, on balance I would come down more on the side of conventional wisdom.

The issue of monetary policy and the zero lower bound (ZLB) problem is of particular concern in Japan today. Goodfriend feels that it could be resolved by imposing a negative rate of return on banks' excess reserves and also on currency. Practically, I doubt that this issue can be easily dealt with. Most currency is returned to banks and then central banks through large retailers (especially those selling food) who might feel inappropriately singled out for such attention by the tax authorities.

Nevertheless, Goodfriend is right to note that the monetary authorities, faced with deflation, could buy massive amounts of assets both domestic and foreign to break the downward asset price spiral and to stimulate price increases more generally. As Milton Friedman famously noted: "Inflation is always and everywhere a monetary phenomenon". Yet there would be complications given such a policy. The assets purchased might fall in value rather than rise, leaving the central bank open to losses and needful of injections of capital from the Treasury. This would be a clear threat to central bank "independence", though how much actual harm might be done is also open to question.

A central bank having a clear mandate to pursue price stability, avoiding both inflation and deflation, should not be much concerned about accepting balance sheet losses in the pursuit of its objective. That is what public policy is all about; positive externalities require the public sector to pay certain costs. However, were a central bank to embark upon such a course in the interest of "jump-starting" the economy, it might at the same time wish to subject its actions to the discipline of a regime of either inflation or price level targeting. To avoid the excesses that developed in the late 1930s in Japan, it would be important to establish that "emergency" measures were precisely that. Once economic growth had been re-established, there would have to be a politically agreed framework to ensure that inflation was not accepted as a normal means to oil the wheels of the economy. In my view, price level targeting has special attributes since it allows a more expansionary policy if prices have been declining for some time. Moreover, since the target is expressed in level terms there should be no danger that an initial large increase in the rate of change of prices would be thought of as a permanent phenomenon.

If Goodfriend perhaps overestimates the role of monetary policy in deflationary conditions, he perhaps underestimates the role of fiscal policy. I know of no convincing empirical evidence that fiscal multipliers are less than one, much less zero. Nevertheless, as Goodfriend implies, the way in which fiscal stimulus is conducted surely matters. Many public sector expenditures in Japan, particularly those designed to support the construction and property sectors, are not expected to generate a positive social rate of return. This implies that the increased liabilities of taxpayers are not offset by an increase in valuable assets, which is clearly an invitation to the Ricardian Equivalence problem that Goodfriend rightly worries about.

If monetary and fiscal policies have a role to play in a deflationary period, so too does bank restructuring. Arguably, deflation having negative implications for growth2 only occurs when financial instability is a major complicating factor. In the first part of his paper, Goodfriend correctly notes that forbearance can lead to heavy costs: $120 billion rather than $20 billion had the savings and loan crisis in the United States been promptly addressed. I would only add, based on the successful experience of the Nordic countries in the early 1990s, that a prompt official response should also be definitive, free from political interference and wholly transparent with respect to the burden of costs above all. In many countries, bank restructuring has fallen well short of these requirements. When, as Goodfriend states, it is the general public that refuses to face up to reality, then the public perhaps gets both the government and the form of bank restructuring that it deserves.

A last set of policies required to escape from a deflationary situation is not in fact mentioned in Goodfriend's paper. For the sake of completeness, these policies might be noted here. If one believes, as the Austrian school of thought did, that profits are the key to the capitalist system, then steps must be taken to keep profits up even as prices fall. Excess capacity must be written off and bankruptcies must be faced squarely. Deregulation might well allow profitable diversification and investment in sectors that had previously been unexploited. And, finally, real wages must adjust downwards in response to altered circumstances. Were nominal wages to remain static as prices fell, the implications for rising unemployment would become as clear as they became in the United States in the early 1930s.

Yet it must also be admitted that wages are income and that income drives consumption; falling real wages thus also have a downside. President Hoover was fully aware of this latter possibility as he campaigned against wage cuts at the beginning of the 1930s in the United States. This is a further reason for suggesting that, faced with serious deflationary tendencies, all of the weapons in the macroeconomic arsenal should be used to their full effect to ensure that aggregate demand is maintained. The concept of "creative destruction" has a certain intuitive appeal, but it should be remembered that the phrase was coined well before the onset of the Great Depression.


Bank for International Settlements (1999): 69th Annual Report, June.

Borio, C, S Kennedy and S D Prowse (1994): "Exploring aggregate asset price fluctuations across countries: measurement, determinants and monetary policy implications". BIS Economic Papers No. 40, April.

1 For an early reference see Borio et al (1994) and successive BIS Annual Reports. See especially BIS (1999), Graph V.8.

2 See BIS (1999), Chapter IV. Since the late 1800s periods of deflation have also generally been periods of positive (if moderated) real growth. The principal exception to this was the first few years of the Great Depression of the 1930s.

Related information