Monetary policy implementation and signalling June 1997

Review of submissions and decision

Discussion document

Financial Markets Department
Reserve Bank of New Zealand

June 1997


In March the Bank issued a Discussion Paper that reviewed technical aspects of the implementation of monetary policy in New Zealand. A change from targeting settlement cash balances to targeting the overnight interbank interest rate (the cash rate) was proposed, and submissions on the proposal were invited.

It was noted in the Discussion Paper that changing the operating system would only make sense if the apparent advantages outweighed any risks. Following consideration of submissions and discussions with a number of interested parties, we have decided that there are risks that might not be outweighed by benefits. Accordingly, at least for the foreseeable future, we will be staying with the current operating system. Some relatively minor modifications to the details of that system might be considered in due course, but the essential character of the current system will not be affected.

Under headings that reflect where most of the discussion was focused, this paper isolates the key issues that were addressed in the review and explains the reasons for the resulting decision.

Is there a signalling problem?

Signalling, in the New Zealand monetary policy context, refers to the Reserve Bank providing information on the intended state of monetary conditions, both now and potentially also into the future. All central banks use signalling. In New Zealand, signalling has a special character because:

  • We provide a considerable amount of information about our view of future inflation pressures, sufficient to allow financial markets to understand and anticipate our policy concerns.
  • Unlike most central banks we don't set, or target directly, an interest rate or the exchange rate (the main components of monetary conditions) in our operations. Financial markets cannot therefore read directly from our operations the state of monetary conditions that the Bank desires.
  • We believe that both the exchange rate and interest rates matter for transmitting monetary policy actions to the economy. So the "monetary conditions" that matter for us cannot be approximated by interest rates or the exchange rate alone.

From time to time we have experienced problems with signalling, with monetary conditions moving in a way that was not intended when a Bank statement was made, and claims of lack of clarity. Part of any lack of clarity may have been associated with confusion over the respective roles of interest rates and the exchange rate; part may have been due to the use of qualitative rather than quantitative terminology by the Bank ("tighter" monetary conditions, for example, rather than "xyz basis points tighter").

The proposed cash rate target operating system offered two modifications to the signalling structure (although the proposal was not seen primarily as a solution to signalling problems). First, it would have involved targeting directly a component of monetary conditions in our operations. Second, in doing so it would have provided quantitative signals.

Several of the submissions argued that these modifications were unnecessary for effective signalling and, more importantly, could even be counterproductive.

  • Unnecessary because from December 1996 we have started to use quantitative terminology when referring to monetary conditions (both actual and desired). Thus, for example, we indicated in December that monetary conditions could be easier by 50 basis points.
  • Counterproductive because the cash rate target operating system could focus public and market attention on the interest rate component of monetary conditions when, as noted, we believe that more than interest rates matter. Although the Discussion Paper indicated that we would continue to use an index of monetary conditions (also introduced in December) as the focal point for discussions of actual and desired conditions, with actions on the cash rate simply being an implementation matter, it was felt that the distinction might not be clear. Thus, for example, a change in the cash rate target to offset a shift in the exchange rate could have been misinterpreted as a policy signal even though our policy stance might not have changed.

Is there a leverage problem?

Leverage refers to the ability of the Reserve Bank to achieve an intended change in monetary conditions through a shift in policy instruments or "levers", in the event that signalling has not by itself achieved the desired result. Leverage is required when, for example, financial markets form a different view as to where the stance of monetary policy is heading than the view held by the Bank.

Two elements of leverage can be distinguished. One is the effectiveness, or power, of policy instruments to affect monetary conditions. The other is the predictability of the effect on monetary conditions of a given shift in the instruments.

It is in the latter regard that cash rate target system was thought to have potential benefits; there is no doubt that existing instruments have ample power. The Discussion Paper noted that there is a degree of uncertainty in the linkage between the settlement cash target - the main instrument or lever in the current system - and resulting monetary conditions. A $5 million adjustment in the settlement cash target at one point in time might have a trivial impact, whereas the same adjustment in different circumstances might have a substantial impact.

On balance, submissions accepted that adjustments to a cash rate target would have a more predictable impact on monetary conditions than adjustments to the settlement cash target. However, the gain was not rated as significant for two reasons:

  • There would still be considerable uncertainty as to the exchange rate reaction to any given shift in interest rates, with the result any improvement in the predictability of leverage over monetary conditions as a whole could well be modest.
  • Improvements to the Bank's signalling made since December (see previous section) were likely to have increased the predictability of the current levers, and reduced the likely need to use them.

Are there operational efficiency gains that are sufficient to motivate change?

The Discussion Paper noted that there are several features of the current system that involve costs for the financial system. Any policy implementation system is likely to involve costs for the financial system. The Discussion Paper estimated that the difference between the costs associated with the current system and those likely under the cash rate system proposal would be in the order of $2 million per annum, with some additional savings on resource costs also possible.

Although these cost savings look substantial against everyday yardsticks, they are not sufficiently large to warrant pursuing in their own right if that involves running risks on the efficient achievement of our monetary policy objective. As already noted, several submissions identified areas of risk, and a number argued that the cost savings, while welcomed, were less important than achieving optimum effectiveness in signalling.

We have reconsidered whether some of the savings in operational efficiency might be able to be achieved without moving away from the settlement cash target system. We are not confident that partial implementation of the proposals can extract the benefits of reduced costs without other problems arising. However, we will continue to explore this issue.

Other issues

Of the range of other issues covered in the Discussion Paper and addressed in submissions, the one most relevant to our final decision - in addition to those already covered above - relates to the "prominence" that the Reserve Bank would have under the proposed arrangements.

The proposed system would have had the Bank formally considering the setting of the target cash rate on a frequent (say weekly) basis. Even if the cash rate target were not actually adjusted very frequently, each formal review would involve a decision which would be of interest to financial markets and would accordingly attract public attention in a way that does not attach to our current weekly informal reviews.

Several submissions argued that this would risk giving the Bank and changes in the operating levers more prominence than is consistent with the proper medium-term focus of monetary policy. It would certainly be of concern if the overall stance of monetary policy appeared to be up for review and debate at very short intervals. In fact it was not intended that the overall stance of policy would be reviewed any more frequently than is currently the case. However, it might not be as easy as had been assumed to make the distinction in the public mind between actions on the cash rate target that are simply operational - ie intended only to offset exchange rate changes so leaving overall monetary conditions approximately unchanged - and those that are more substantive in character.

Some also argued that the increased formal association of the Bank with interest rate adjustments would risk "politicising" interest rate adjustments unnecessarily.

Overall assessment

The Discussion Paper indicated that there was no one particular benefit that would alone warrant a change to a cash rate targeting system. Our current techniques for implementing monetary policy have generally been effective; our experience in achieving and maintaining a low inflation rate demonstrates that. Instead, the expected gains that were identified, while individually not substantial, were collectively thought worth pursuing.

Doubts have been cast on that conclusion by a number of the submissions, mostly in relation to areas where perceptions are important. In particular, perceptions of the effectiveness of the Bank's signalling after recent modifications (and the likely impact on that of the cash rate targeting scheme), and perceptions of the consequences of more frequent formal reviews of settings, worked against the net gains of the proposed system.

Perceptions are important to the success of monetary policy implementation and are also hard to judge objectively. As much as anything else, it is for this reason that we were keen to seek advice when we issued the Discussion Paper. Against this background, it was decided that there was sufficient uncertainty about the net gains of the proposal to warrant not proceeding to implement the proposed system.