RBA Annual Conference – 2004 Discussion

1. Robert N McCauley[1]

The author has written a broad and balanced account. This paper makes the reader think about this critical period of monetary policy-making differently. My remarks concern the main thesis, monetary and fiscal policy at a zero overnight interest rate, and abstaining from inflation targeting.

The main thesis

This paper argues that the Bank of Japan shrank from doing the right thing because it required the adoption of unconventional policies that might lose the Bank its newly acquired independence. Taka Ito frankly recognises that this argument has a problem: the Bank of Japan adopted two and a half of the recommended unconventional policies: increasing excess reserves, buying government bonds and buying shares – with the last counted as one-half because the justification was couched in terms of financial stability rather than monetary policy. Moreover, the author seems to argue that the other unconventional policy of buying foreign bonds, or in Svensson's formulation, depreciating the exchange rate, was in effect adopted in the form of heavy intervention by the Ministry of Finance, matched by a broadly parallel rise in excess reserves at the central bank. This leaves only inflation targeting as the unadopted unconventional policy.

In passing it should be noted that inflation targeting does not belong on the list of unconventional measures, either logically or historically. Taka argues that an inflation target would have been achieved through the above-named unconventional measures, implying that inflation targeting is not logically just another such instrument. Historically, it is hard to call inflation targeting an unconventional policy in a world in which it is recommended so widely.

In the end, it appears that the Bank of Japan was unwilling to adopt inflation targeting, but not because it was unwilling to adopt the measures that the author feels would have served to hit the target. Rather, it was unwilling to commit to pushing these unconventional measures to hit pre-announced targets within a certain time. In other words, the Bank of Japan was willing to adopt the measures (means) urged by its critics, but not in the service of an announced, symmetric and time-bound inflation commitment (end) that they would consider inflation targeting. The Bank of Japan feared that it could not deliver on such a commitment.

Monetary and fiscal policy at a zero short-term interest rate

This paper devotes considerable space to arguments that are not central while leaving most of the central economic arguments off-stage. On the first count, much of the discussion of the costs of deflation, particularly on the interest burden on corporations and the government, really concerns the costs of unanticipated disinflation. The transition to deflation differs from disinflation because of the zero lower bound in nominal wage growth, if any, and nominal interest rates. The discussion of the first suggests that the bonus system allowed nominal wages to fall. A graph of real wages over the period of deflation would be useful. Similarly, some evidence on real interest rates, at the short and long term, would enrich the discussion of deflation and interest burdens. Whatever the deficiency, from Taka's point of view, of the Bank of Japan's commitment to keep nominal interest rates at zero (no ceiling for inflation, no time frame for achievement), what is remarkable, and what would surprise Keynes most in the story, is how nominal 10-year yields got down to a half per cent. What is called the ‘policy duration effect’ seems to have been powerful indeed. The same low Japanese government bond yields imply, however, that market participants believed until quite recently that the unconventional measures (rising excess reserves, bond buying and foreign exchange intervention) would not reliably whip deflation in five years.

Regarding the off-stage argumentation, reference is made to what ‘mainstream’ economists urged, appealing to a broad agreement on the underlying economics. On the basis of this paper alone, the presumption of shared analysis seems to encompass an embrace of both base monetarism and the related usefulness of the distinction between sterilised and unsterilised intervention at zero interest rates. At a deeper level, there is a presumption that anything that can affect the price level is potentially and properly a policy to be pursued by the central bank. Let us consider these in turn.

My starting point is the near-perfect substitutability of excess reserves at the Bank of Japan, and the treasury or financing bills of the Ministry of Finance at zero interest rates. Both carry essentially zero interest rates, offer great liquidity and pose much the same credit risk. In effect, when the Bank of Japan set the overnight rate at zero, it gave the Ministry a printing press. One argument for a difference is that the excess reserves are permanent additions to the monetary base and so will support a higher nominal GDP. But the commitment of the Bank of Japan to maintain effectively zero interest rates until inflation returns implies that the oversized current balances at the Bank of Japan are temporary.

If one compares the situation of 6 trillion yen in excess reserves with the one of 31 trillion in excess reserves, the difference is 25 trillion yen of short-term government debt in private hands in the first case, and 25 trillion yen in overnight deposits at the Bank of Japan in the second case. It is very hard to see why the behaviour of private parties would differ between these two cases. Viewed as a commitment device, excess reserves per se seem weak since they can be rapidly withdrawn at low cost.

If one accepts this near-perfect substitutability, then unsterilised versus sterilised intervention becomes a distinction without a difference. In one case, the private sector holds the short-term, essentially non-interest-bearing paper of the Ministry of Finance, and in the second case, the Bank of Japan holds the Ministry paper and the private sector holds the short-term, zero-interest claim on the Bank of Japan.

Taka and I agree that intervention can affect the yen. I would interpret the effect as running through the portfolio balance effect: unlike Ministry paper and Bank of Japan balances, Ministry paper and US Treasury notes are not perfect substitutes, so that their relative supplies affect their relative price, the exchange rate. We may differ if Taka attaches importance to the rough parallel in the increase in Ministry's holdings of dollars and the Bank of Japan current balances. I would attach importance only to the dollar holdings at zero interest rates.

At another level, there is an unstated presumption that anything that can affect the price level is properly in the field of action for the central bank. This presumption covers the purchase of bonds, risky domestic assets and foreign exchange, all of which can affect inflation. However, such unconventional ‘monetary measures’ are really the province of the government, both conceptually and practically.

  • Purchase of bonds: The supply of duration to any government bond market is fundamentally a treasury responsibility. In the late 1990s, there was a debate within the Ministry of Finance over whether to issue a smaller share of debt in the form of long-term bonds, but those responsible for debt management resisted. Conceptually, having the central bank issue overnight liabilities to buy 10-year bonds must be recognised as poor man's debt management, ‘poor’ because the balance sheet of the central bank is smaller than that of the government. A bold debt management policy would have not only ended the issue of new long-term fixed-rate debt (in favour of floating-rate or indexed-linked debt), but also bought outstanding bonds back. Banks would have been denied the easy option of earning profits by taking duration but not credit risk. Practically, the Ministry of Finance stands to gain in tax revenues from a revival of nominal growth, so it makes sense for it to bear the (opportunity) cost of a rise in interest rates.
  • Purchase of equities, etc: Buying risky assets puts the government in the position of favouring certain issuers over others. Like tax and expenditure policy, this is highly charged politically. The first best solution is to follow the model of Herbert Hoover's Reconstruction Finance Corporation, with explicit legislative authority for an entity to purchase risky assets – just as legislation underpinned the Japanese government's recapitalisation of the banks.
  • Buying foreign bonds: Looking across countries, the assignment of foreign exchange risk-bearing to the government seems to be the exception rather than the rule. But in Japan, foreign exchange policy has been the responsibility of the Ministry of Finance. Its purchases of foreign exchange require financing that is subject to legislative approval. Having the Bank of Japan buy foreign bonds would not have been any more effective in economic terms (by the argument above) but would have evaded this legislative control.

Expediency could be seen to override these arguments. The debt managers insist on pumping out long-term bonds, so the central bank should buy some of them and risk the need for recapitalisation; the legislature cannot agree on the government buying equities or real estate, so the central bank should do so; the legislature cannot agree to increase the debt limit, so the central bank should buy foreign exchange on its own account. Is it obvious, however, that the legislature would be happy to recapitalise the central bank ex post for losses from risk-taking that the legislature would not appropriate funds for ex ante? More fundamentally, we need to think twice about recommending putting tax-payers money at risk without the consent of their representatives. It is a fine judgement whether events constitute such an emergency as to require an override of parliamentary prerogatives.

The most powerful lever on the price level of these unconventional measures is intervention in the exchange rate. Foreign exchange intervention allowed global equity investors to return to their neutral weightings on Japan without pushing up the exchange rate. If Japan's recovery continues and the economy exits from deflation, this policy deserves credit. Would the Bank of Japan have been well-advised to count on this outcome within a given period of time? There was the usual uncertainty about transmission (in this case the pass-through from foreign prices to the domestic price level). Beyond this, however, was the question of whether trading partners would acquiesce in the exchange rate management. Would it have been safe to presume an acceptance in Washington of the proposition that recovery in Japan is so much in the US interest that it should be allowed even at the expense of US net exports? Was the ultimate acceptance a function of the larger trans-Pacific relationship involving non-economic elements? It seems fair to ask: had the central bank been willing to manage the exchange rate on its own account to break the deflation, was it clear ex ante that it would be able to do so?

Abstaining from inflation targeting?

Is there an argument for a central bank that shares the low inflation goal of formal inflation targeters to abstain from inflation targeting? Baltensperger, Fischer and Jordan (2002) have argued that central banks that enjoy a measure of goal independence may refrain from binding themselves with an inflation target. This argument can go beyond Ken Kuttner's weighting on output stability to the possibly competing goal of financial stability. While it is possible to translate financial instability into output instability, as have some RBA papers, it seems to me that concern for financial stability can in practice compete with inflation stability.

It is not clear how these considerations might apply to the Bank of Japan on the far side of its current policy. One can imagine that the Bank chooses inflation targeting and thereby underscores its measure of goal independence.


The views expressed in this note are those of the author and do not necessarily reflect those of the BIS. [1]


Baltensperger E, AM Fischer and TJ Jordan (2002), ‘Abstaining from inflation targets: understanding SNB rhetoric in the inflation targeting debate’, Swiss National Bank, mimeo.

2. General Discussion

An important theme raised in the discussion was the wider implications that could be drawn for monetary policy from the Japanese experience. One participant highlighted the need for complementarity between a central bank's actions and communication. This follows from the importance of managing the public's and financial markets' expectations to best achieve successful outcomes, and also highlights a key reason why the Bank of Japan's (BoJ) use of unconventional policies proved unsuccessful. The problem was not the policies themselves, but that the BoJ's communication led the public to believe that these policies would be prematurely reversed.

A number of participants suggested that the BoJ had difficulty dealing with its independence. Borrowing from the title of Ken Kuttner's paper, one participant suggested that, like adolescents, the BoJ wanted independence, but once they had received it, they were not quite sure how to proceed, nor were they prepared to take responsibility for their actions. Another facet of the ‘independence trap’, referred to in Ito's paper, was the feeling of insecurity, which had led the BoJ to reject courses of action suggested by outsiders. A prominent example of this was the reaction to Paul Krugman's suggestion that the BoJ should announce a 4 per cent inflation target for 15 years. While the merits of the proposal were open to debate, one effect of Krugman's (undoubtedly bold) suggestion was to discourage the BoJ from taking bold action of any sort.

There was some discussion of the roles of different institutional factors in the debate over introducing inflation targeting. One participant suggested that the presumption of the debate seemed to be that the introduction of inflation targeting was entirely the decision of the BoJ. This contrasted with the international experience, where the introduction of inflation targeting was commanded by the government and not the central bank, though Australia was an obvious exception. The author responded that there was some support within the Ministry of Finance and among politicians to introduce inflation targeting, but this did not make it into the mainstream because the Finance Minister at the time, Kiichi Miyazawa, opposed the idea.

The role of academic economists in the debate over inflation targeting was also raised. One participant suggested that Ito's paper gave the impression that within Japan academics were on one side of the debate (supporting inflation targeting) and central bankers were on the other side (opposing inflation targeting). Yet the arguments presented by some academics, that Japan's difficulties reflected a supply-side shock and the impact of sizeable gains in China's competitiveness, indicate that there was some support from the academic community for the BoJ. That being said, such supply-side arguments were criticised by a number of participants.

In terms of the recent economic upswing in Japan, one participant cautioned against the risk of attributing the improvement in performance solely, or even largely, to monetary policy. The participant contended that performance had improved for other reasons, such as the flow-on effects from the continued strong growth in China and also balance sheet recovery in the corporate sector. Ito's paper was seen as particularly important in this regard, as it laid out the policy mistakes that had occurred, at least prior to the cementing of the most recent ZIRP.

One participant suggested that it was now opportune for the BoJ to introduce inflation targeting: the cyclical recovery had created some credibility for the BoJ in financial markets; the banks' bad debts were declining; and, there was greater confidence in equity markets. Another participant agreed, suggesting that the real challenge for the BoJ will come when it is ready to cease quantitative easing and move away from the ZIRP, at which time it may need to develop a new target. A number of participants thought that the introduction of inflation targeting could be of significant assistance in this regard.