Monetary policy accountability and monitoring

Michael Reddell, Economics Department

This article outlines the framework used for monitoring the conduct of monetary policy in New Zealand. It explains both the legal framework contained in the Reserve Bank of New Zealand Act and the way the monitoring system works in practice. The formal New Zealand model places considerable explicit focus on mechanisms to hold the monetary policy decision maker (the Governor) to account. The formal structures, and the conventions that have developed in the years since the Act was passed, mean that the Bank is subject to extensive ongoing monitoring of its monetary policy processes and decisions, especially by the Bank's Board.

Monetary policy, directed to "achieving and maintaining stability in the general level of prices, is the primary function of the Reserve Bank.1 This article explains and assesses the ways in which the Reserve Bank is, and can be, held to account for its conduct of monetary policy2. Section 1 outlines the formal framework and the thinking behind it. Section 2 looks at the various ways in which the Board and other entities actually undertake monitoring and help, in practice, to hold the Bank to account for the conduct of monetary policy. Section 3 offers some concluding perspectives.


The central banking reforms of the late 1980s, contained in the Reserve Bank of New Zealand Act 1989 (hereafter "the Act"), were an ambitious attempt to balance the insights of two important strands in contemporary thinking about the best way to manage monetary policy.

On the one hand, there was a growing acceptance that central banks with operational autonomy over monetary policy were likely, on average, to produce better monetary policy decisions. But, on the other hand, the extensive public sector reforms in New Zealand in the late 1980s were shaped by a desire to establish better governance arrangements for public agencies and departments. In that model, Ministers set objectives and hold departmental chief executives to account for delivering the objectives3. The vision was one of clear and specific goals, operational autonomy, and focused accountability, all supported by a high degree of transparency.

The Statutory Provisions

In establishing a regime for monitoring the performance of the Bank, and for holding it to account, in its conduct of monetary policy, the Act outlines important, and distinct, roles for the Minister of Finance, the Governor, and Board of Directors. It also establishes key roles for two documents: the Policy Targets Agreement and the Monetary Policy Statements.

Under earlier Reserve Bank legislation, the powers of the Bank had rested ultimately with the Board (essentially the corporate model, with executive responsibilities delegated to the Governor). Under that legislation, the Bank was required to implement the monetary policy of the government of the day.

In the 1989 Act, the Bank was given the ability to adjust the instruments of monetary policy without any routine political involvement. As part of this shift, responsibility for the exercise of the Bank's powers and the conduct of its functions was vested explicitly in the Governor4. The explicit intention, in making this change, was to provide a clear focus for accountability5.

The Board's new role focused on two dimensions: advising the Minister on the appointment (and reappointment) of a Governor, and monitoring and providing advice on the Governor's performance.

The Governor and Board members are both appointed by the Minister of Finance. However, whereas the parliamentary term is three years, Board members are appointed for staggered five year terms. The Minister cannot appoint as Governor someone whom the Board has not recommended6.

Before appointing a person as Governor, the Minister is required (section 9) to "fix, in agreement with that person, policy targets for the carrying out by the Bank of its primary function during that person's term of office. The Board has no formal role in that process.

Once the Minister has agreed policy targets with the Governor, a set of procedures set out in the Act is designed to help monitor performance and enable the Governor to be held to account formally for the Bank's conduct of monetary policy.

Section 15 of the Act requires that the Bank deliver to the Minister and publish at least every six months a monetary policy statement. In this document the Bank is required, inter alia, to review and assess recent monetary policy and to articulate "the policies and means by which the Bank intends to achieve the policy targets".

The Board is required (section 53) to "determine whether policy statements made pursuant to section 15...are consistent with the Bank's primary function and the policy targets agreed to with the Minister". That is, they must make some judgement on each and every Monetary Policy Statement contemporaneously.

Sections 49 and 53 provide that the Minister may seek the removal of the Governor7 (or the Board may recommend that the Minister do so) if he is satisfied on any of several counts. These include, inter alia, the following which bear directly on monetary policy:

  • That the Bank is not adequately carrying out its functions (the primary function being monetary policy); or
  • That the performance of the Governor in ensuring the Bank achieves the policy targets has been inadequate; or
  • That a Monetary Policy Statement is inconsistent in a material respect with the Bank's primary function, or with any policy target fixed in the Policy Targets Agreement.

Consistent with its primary ongoing role as a monitoring agent on behalf of the Minister, the Board is charged with keeping under "constant review" matters relating to the first two points above. Moreover, the Board is required to advise the Minister whenever it is satisfied that any of these three conditions has been met. Whether or not dismissal is sought is entirely at the discretion of the Minister.

Note that the Act does not allow the Governor to be dismissed simply for failing to meet the policy targets. The criteria in the Act refer explicitly to the performance of the Bank and the Governor in pursuit of those targets.

In 2000, as part of the broader process of accountability, the government commissioned an independent review of monetary policy, undertaken by the eminent Swedish academic economist Lars Svensson. Following that review, various amendments were made to better align the legislative provisions with the monitoring role of the Board. In particular, a non-executive director is now required to chair the Board (the Governor had previously chaired regular meetings of the Board) and the Board is now required to prepare a (public) annual report, outlining the Board's assessment of aspect of the performance of the Bank and the Governor which it is required to monitor.

The Policy Targets Agreement

How the monitoring framework is applied depends to some extent on the form the targets take. The Act requires that targets be agreed "for the carrying out by the Bank of its primary function" - monetary policy, directed to maintaining a stable general level of prices - during the Governor's term of office. That provides considerable latitude to the negotiating parties. In particular, the Act does not specify that the targets must be for inflation itself.

Monitoring the conduct of monetary policy, and holding the Governor to account for his performance, would be relatively straightforward if the policy targets could be specified in terms of something over which the Reserve Bank had direct, or relatively direct, control (in the jargon, an "output" of Reserve Bank decisions and processes).

If the policy targets were to be expressed in terms of, say, maintaining a fixed exchange rate (or a band), or a target for some measure of the monetary base (say, primary liquidity or settlement cash), the Board's monitoring of performance against the Policy Targets Agreement could be done largely by reference to actual outcomes. In the early discussions about granting the Reserve Bank operational autonomy, The Treasury was keen an operational regime for monetary policy that allowed exactly that sort of tight ex post monitoring8.

In fact, all parties eventually concluded that purely output-based performance measures were not feasible. Medium-term targets that allowed very tight and clear-cut formal monitoring and accountability would not produce economically sensible outcomes (for the things we all really care about - such as inflation, nominal income, or the variability of GDP). Nor would they provide useful information to markets, and firms and households, on which to base their own planning.

And so the choice settled on inflation targets. Specifying the targets in terms of something closer to the outcomes everyone cares about has some obvious attractions. Communicating with the public about what monetary policy is doing is rather easier - indeed, this may be the single biggest gain from the international shift towards inflation targeting. But because monetary policy affects inflation only slowly, and indirectly, and because it is not the only influence on inflation, inflation targets have never been seen as an ideal basis for holding monetary policy decision makers formally to account. What has evolved is a regime that attempts to provide as much formal accountability as is consistent with encouraging good policy and maximising the chance of getting good economic outcomes. Good outcomes are what matter most - accountability models matter in so far as they help deliver good outcomes, and avoid very bad ones.

Since the Act came into effect there have been seven Policy Targets Agreements.
The Policy Targets Agreements have had several common features:

  • They have each specified a range of inflation rates which monetary policy should aim to achieve.
  • Each quantitative target has been specified largely in terms of the CPI, (although some early ones made explicit allowance for problems in the measurement of inflation; in particular, the fact that until the mid 1990s interest rates were included directly in the CPI)
  • Each has recognised that at times inflation will be, and should be, outside the normal range and each has outlined procedures, more or less general, for handling, and accounting for, those episodes.

The Policy Targets Agreements have consistently been forward-looking in their orientation: they set out to specify an agreed framework which the Governor should use in making monetary policy decisions, bearing in mind that today's decisions affect inflation up to two years from now. The PTAs have not sought to deal comprehensively with all eventualities, and have been documents of no more than one or two pages.

At the same time, the approach adopted in the Policy Targets Agreement has evolved. The provisions of the first PTA were consistent with the heavy focus on formal accountability and reflected a vision that sought to minimise the Bank's use of discretion. Under those provisions, the Bank had the right to seek to renegotiate the inflation target when, say, an oil shock hit which would take inflation outside the target range for a time. That model was quite quickly replaced. Since the end of 1990 the PTAs have not required renegotiation in the event of shocks. Instead, the Bank has been required to state publicly how it proposes to handle such shocks, and how it will ensure that core inflation pressures remain consistent with the target once the direct effects of the price shock has passed.

Since 1999, the Policy Targets Agreements have stated explicitly, that in implementing monetary policy the Bank "shall seek to avoid unnecessary instability in output, interest rates and the exchange rate". Elements of this approach had long been implicit in the targeting framework - for example, avoiding undue short-run economic disruption had consistently been the basis for "looking through" the direct impact of one-off price shocks.

The main innovation in the current PTA, signed in 2002, was to narrow the inflation target range (from 0 to 3, to 1 to 3 per cent) and to specify it explicitly as a target for "future CPI outcomes.....on average over the medium-term. These changes were seen as giving monetary policy a little more scope for flexibility.

In each case, when the Policy Targets Agreement has been changed, the Bank, the Board and (by implication at least) the Minister have had to reflect on just what the changes mean for the inflation targeting regime as a whole, and specifically, what the changes mean for reaching judgements on the Governor's performance in the conduct of monetary policy9.

What makes monitoring monetary policy hard?

Like any activity, whether in the public or private sectors, where decisions today affect outcomes well into the future, and then only indirectly and subject to other influences, monetary policy is complex. That makes the decision maker's job hard, and probably makes the job of those charged with monitoring the decision maker's performance even harder.

Some of the specific features of monetary policy that are relevant to the devising a workable monitoring and accountability model include:

  • No single CPI inflation rate, in isolation, sums up in all circumstances what one would want monetary policy to achieve (eg, oil shocks mean that in some circumstances inflation should be above, or below, the target range)
  • The Reserve Bank has no ability to control inflation directly.
  • Even the instrument the Bank sets directly (the OCR) does not directly affect inflation; rather it operates through various and often complex "transmissions channels", affecting spending behaviour, investment activity and so on, before actually affecting inflation.
  • Evaluating the pressure monetary policy is imposing on the economy, and hence the outlook for inflation, relies on estimating a whole series unobserved (and often unobservable) variables; (eg gaps between output, unemployment, interest rates and exchange rate and the sustainable medium-terms levels for those variables).
  • In making policy today, we do not have firm data about where the economy stands today - official data are available only with a lag, and in many cases are subject to non-trivial revisions.
  • Policy actions today take up to around two years to have their full effect on inflation. This means, for example, that policy decisions today are made with only partial knowledge of the impact of decisions taken even twelve months previously.
  • All relationships between economic variables are estimated only with (often considerable) margins of error. The broad structure of the how the economy works in known only imprecisely, let alone the details. And these relationships change through time.
  • Should inflation move outside a target range there are meaningful choices to make about how quickly to seek to return it to the target range,
  • We are attempting to influence human behaviour indirectly; not manage an engineering relationship,
  • The world does not stand still while monetary policy gradually modifies inflation outcomes over a two year horizon. Events that were not forecast happen in the interim, both in New Zealand and abroad, and those events affect actual inflation outcomes. Those events will at times also affect the policy stance one would wish, with hindsight, to have adopted.

While there are specific issues arising out of each particular Policy Targets Agreement, this list of considerations and challenges is not specific to the current Policy Targets Agreement. Indeed, many of the same sorts of issues, differing only slightly in detail, would arise if, say, the PTA specified targets for something other than inflation10.

These considerations all help explain why the Act focuses on monitoring and evaluating the Bank's conduct of monetary policy, not on (inflation) outcomes in isolation. The lags between OCR choices and inflation outcomes (two years or so hence) mean that in the course of any five year term there are only a couple of genuinely independent inflation outcomes. There simply are not enough independent observations for the role luck and exogenous shocks to average out, leaving the role of the Governor's skills and judgements clearly apparent in the data.

Monitoring by the Board

Since the Act came into effect in 1990, the Board has the primary body with formal responsibility for monitoring the conduct of monetary policy.

The Bank's Board faces some unique challenges, and some generic ones. Few other central banks have a monitoring body quite like our Board, so there are few models to follow in holding the Bank formally to account. On the other hand, many of the Bank's Board members have had considerable experience as directors in a commercial environment. One of the main roles of any Board is to evaluate the performance of the chief executive, and (conceptually at least) to isolate the relative roles of skill and luck in contributing to the outcomes stakeholders observe.

Some of the items the Reserve Bank's Board might be expected to concern themselves with in fulfilling the monetary policy monitoring role include:

  • The processes the Governor uses to gather and interpret economic information.
  • The choices the Governor makes in allocating resources areas of the organisation relevant to monetary policy (including judgements he makes on whether to seek more, or fewer, resources, when the five-yearly funding agreement is negotiated).
  • The means the Governor uses to ensure that he is exposed to alternative perspectives.
  • The quality of the people the Governor appoints to advise him on policy choices.
  • The way in which the Governor applies section 3 and 4 of the PTA (dealing with deviations from the target range, and the avoidance of unnecessary instability).
  • The way in which the Governor thinks about and responds to the uncertainties around monetary policy.
  • The ability of the Governor to articulate the reasons for his policy choices, and his ability to convince others of his case.
  • The processes the Governor uses to assess past policy and learn from experience.
  • The stability through time in the Governor's policy choices.

In seeking to perform this role, the Board would be expected to use a variety of sources of information in addition to the Bank's own analysis and interpretative perspectives. These would typically include:

  • Information on how other central banks handle similar issues, and about how they react to similar shocks.
  • Forecasts made by other agencies.
  • Commentaries by other economists and commentators.
  • Developments in financial market prices.
  • Indications about the confidence that markets and key stakeholders have in the Governor's performance in the role.

Inflation outcomes matter, but not in isolation. A Governor could be running a prudent and appropriate monetary policy, and yet end up with inflation consistently outside a CPI target range - if he was unlucky in the series of shocks he faced during his term. On the other hand, monetary policy could be run rather imprudently, on a consistent basis, and yet over a five year term, a Governor could find that inflation was typically inside a CPI target range.

So when the Board considers actual inflation outcomes, they are attempting to sift through the murk. Inflation outcomes outside the target range should raise questions and prompt analysis, even if it at the time the relevant OCR decisions were made the Board was content that the contemporary Monetary Policy Statements were consistent with achieving the inflation targets. What was it, for example, that the Governor missed, and how reasonable was it that the Bank, with the resources at its disposal, should have missed it? Given the asymmetry of resources and economic expertise available to the Governor, the fact that the Board accepted at the time that any particular Monetary Policy Statement was consistent with the PTA, should not stop them asking questions about the conduct of policy later, with the benefit of the passage of time, and actual inflation outcomes.

Inflation outcomes inside the target range should also prompt questions from the Board. In evaluating the Governor's conduct of monetary policy, the Board should be seeking to use as much data as possible, and assess (formally or otherwise) the role that good or bad policy choices, as opposed to sheer good or bad fortune, played in all inflation outcomes.

Of course, few of the indicators or issues the Board is likely to look at are unambiguous. That is so almost in the nature of assessing anything as complex as monetary policy. Absent an egregious error, which any reasonable Governor should have avoided, the judgements the Board is charged with making are more likely to involve an accumulation of evidence about the way in which the Governor and Bank have handled monetary policy over time, rather than the evaluation of a single decisive choice.

The framework for monitoring the conduct of policy, and thinking about how to evaluate performance, has developed over the years since the Act came into force. In addition, the Bank's internal monetary policy decision making processes have become considerably more structured, and more amenable to formal monitoring than they were in the early inflation-targeting years.

The issues first came into sharp focus in the mid 1990s when inflation first went above the top of the (then) target range, and was forecast to remain around the top of the range for some time. In considering how to respond to that situation, the Bank's Board formally recognised the importance of the lags, the uncertainty, and the role of exogenous shocks, and identified that its role was primarily to evaluate what the Governor had done rather than the inflation outcome itself.

The Board noted that the inflation target was something which the Bank was to be "constantly aiming" for, and that deviations from the target range neither could nor should be instantly corrected. This perspective has recently been reaffirmed by the Board.

In conducting monetary policy, then, the Governor is expected to be constantly forward looking, focusing on the horizons where monetary policy can work most efficiently and effectively11. Bygones (recent inflation outcomes) matter to current policy only to the extent that they affect likely future inflation outcomes.

The Board is required (and the Minister may) assess his performance in (a) making/implementing the decisions, and (b) communicating and accounting for policy in the Monetary Policy Statements. As the framework has evolved, the focus of Board monitoring has involved a mix of testing and scrutinising the processes Bank is using, as well as seeking to understand, and being able to question, the stance of policy adopted by the Bank and the role that stance has played in actual inflation outcomes12.

In practice, the Board uses a wide variety of information and skills to undertake its evaluation and monitoring role. Much of this involves conventions that have developed over the years that the Act has been in place, and procedures adopted by the Governor, to help the Board have confidence in the processes and judgements that are part of monetary policy.

Key dimensions include:

  • The Board meets regularly (typically monthly). The Governor himself is a member of the Board, and as a matter of practice has his Deputy and Assistants present for some or all of the Board meetings and discussion13. All this exposes the Board regularly to key senior managers involved in advising the Governor.
  • Publications. The Act requires that Monetary Policy Statements be issued at least every six months, but the Bank currently issues them quarterly (and also provide brief policy comments halfway between each one). Moreover, we have chosen to publish forecasts, including those for inflation itself, and for GDP, interest, and exchange rates. The Act does not require the publication of forecasts (some inflation targeting countries, notably Australia, do not publish them). These practices give the Board a large amount of public material to assist in their evaluation and monitoring.
  • The Board is provided with all the background documents that the Bank's Monetary Policy Committee and the OCR Advisory Group (OCRAG) used in the formal process leading up to OCR decisions (including material on the views of markets and other commentators).
  • The Board is provided with a presentation after each Monetary Policy Statement explaining the background to the relevant OCR decision
  • The Governor seeks written advice from each member of OCRAG. These pieces of advice are prepared independently of each other, and copies (names removed) of the individual pieces of advice are provided to the Board.
  • Board members themselves bring a wide range of mutually complementary experiences and professional backgrounds to the task. Since the Act was passed there has always been a professional macroeconomist serving as a member of the Board (who has particularly assisted in evaluating each Monetary Policy Statement), many directors have commercial backgrounds, and others have had backgrounds in various dimensions of public policy.
  • The Board is provided with material on the Bank's budget etc, providing it with information on (and opportunity to question) resources devoted to the monetary policy functions.
  • For the last five years or so, the Governor has engaged external advisers as members of OCRAG, for limited terms, to provide independent non-staff input directly to the policy advice process.
  • The Governor has engaged outside "peer reviewers" to sit in on the forecast week process, many of whom have been senior overseas central bankers. The comments of these experts have been provided to the Board, and on occasion the Board has met separately with those experts.
  • The Governor provides to the Board copies of many of the supporting or interpretative pieces of analysis prepared for the Monetary Policy Committee between forecast rounds (including those, for example, on the comparative accuracy of the Bank's forecasts).

This list is dominated by measures which help provide assurance about the Bank's processes. This is, perhaps, both inevitable and appropriate given, for example, the pervasive uncertainties affecting an assessment of any one, or any series of, OCR decisions. The Board has also focused on inflation outcomes, especially on occasions when inflation has been, or threatened to, move outside the target range.

The Board has a privileged position which makes it better placed to hold the Governor to account for the conduct of monetary policy than any other domestic entity, and probably at least as well placed as any official monitoring entity evaluating other countries' monetary policies.

One of the Board's most important roles is that of recommending a suitable candidate for Governor. In any complex field, choosing the right person - someone with the temperament, character, and skills for the job - is almost always the best way of improving the chances of good outcomes. In that respect, monetary policy is no different. However, while a Governor is in office, with the extensive powers of a single decision maker, the monitoring and accountability procedures that have been developed over the years have a useful role in reducing the risk of sustained monetary policy errors. They provide a basis for ongoing conversations with the Governor, and also help the Board accumulate more systematically the perspectives that enable them to seriously review and evaluate performance at the time any decision on a possible reappointment has to be made.

Other Monitoring

Perhaps more than any other area of public policy, monetary policy is conducted under intense and continuous public scrutiny. There are formal dimensions to this (the Board has the key formal role), but extensive and continuous informal monitoring also occurs. In normal circumstances, the informal monitoring may be at least as important, and in all circumstances provides additional data for the formal monitoring bodies.

The Treasury is the principal agency charged with providing advice to the Minister of Finance on macroeconomic policy. They provide regular briefings to the Minister on issues relating to monetary policy. If there were concerns that monetary policy was running seriously astray, The Treasury and the Minister could be expected to be active, doing their own analysis and asking questions of the Board and the Governor. The power to commission a performance audit (section 167) provides considerable leverage. The Bank is now required to publish a Statement of Intent each year, and is required to consult with, and have regard to, the views of the Minister of Finance in bringing together this document - issues that could be directly relevant to monetary policy might include strategic research priorities and resource allocation choices

Parliament also has a role in monitoring the Bank's conduct of policy. By statute, each Monetary Policy Statement stands referred to Parliament's Finance and Expenditure Committee. That Committee, with the assistance of its own professional adviser, can choose to extensively scrutinise the extent to which policy was, and has been, conducted consistent with the Policy Targets Agreement. Statements of Intent also stand referred to the select committee. The Office of the Auditor General - an agent of Parliament, rather than of the executive - is also able to commission quite extensive investigations into the activities of the Bank14. At least every five years, the Bank's Funding Agreement needs parliamentary approval, providing an opportunity to scrutinise resource allocation choices.

Financial markets, the business media, and other economic commentators all play a part in scrutinising and making sense of the Reserve Bank's monetary policy choices. It is not difficult to make monetary policy choices that turn out to be wrong - indeed, in the nature of things, many will turn out to have been less than ideal. But the presence of the extensive market commentary, on every major piece of data and on OCR decisions themselves, means that if the Bank takes a position that even a significant minority of outsiders disagree with, the difference is likely to be highlighted. This not only allows for public debate and scrutiny, but also provides information that the Board themselves can (and does) use in questioning and evaluating the Governor.

In assessing Monetary Policy Statements the Board should expect to draw some comfort if all independent forecasters and commentators are relatively supportive of particular choices the Governor has made. And looking back, after possible adverse inflation outcomes have emerged, it is harder for the Board to be seriously critical of the Governor if, at the time he made his choices, most other commentators would have taken the same stance.

Of course, market commentators neither have the resources the Governor has, nor do they have the incentives to devote much resource to evaluating, after the event, whether the Governor was, in fact, correct in his judgements. And, in any case, a diversity of view among market commentators, investors, and forecasters is a fairly normal feature - highlighting again the way in which reasonable people can, at times, reach quite different conclusions about the appropriate stance of monetary policy.

Some concluding thoughts

The Reserve Bank of New Zealand Act sets out an ambitious framework for setting monetary policy goals and holding the Governor to account for the conduct of monetary policy. Its vision - the mix of operational autonomy and clear accountability (formal and informal) - is clearly articulated in legislation. The framework is built around the provisions that make the Governor the single decision maker, assign to the Board a primary role of monitoring and reporting on the Governor's performance, and allow the Governor to be dismissed for inadequate performance.

To support that formal framework, extensive procedures for disclosing information to the Board, and enabling them to challenge both processes and policy judgements, have been developed. The combination of a single decision maker, and the privileged access the Board has to enable it to understand and evaluate the choices and judgements the Governor is making, are central to the way the system works in practice.

Actual accountability is, of course, multi-dimensional. The formal provisions of the Reserve Bank Act help to strengthen the challenge and review process, especially in the context of a single decision maker. Long lags, and the pervasive uncertainty, all make it hard to reach quick or definitive judgements on the conduct of policy. However, the Board's privileged access to and insights on the policy process, combined with the monitoring undertaken by markets, the media, and Parliament all mean that the Reserve Bank of New Zealand is subject to extensive ongoing scrutiny, formal and informal. New statutory obligations on the Board, including the need to provide a public Annual Report, have only served to strengthen the monitoring regime in recent years.

Appendix: International approaches to monitoring monetary policymakers

The New Zealand legislation devotes considerable attention to balancing operational autonomy and a formal system for monitoring and accountability. The approach remains relatively unusual internationally. This box briefly outlines the approaches adopted in the legislation governing three other inflation-targeting central banks; those of Australia, Canada, and the United Kingdom.

In New Zealand, the Governor is personally responsible for all the monetary policy decisions taken by the Bank. The position in Canada is rather similar: although the Board is responsible for the management of the Bank, the legislation specifically empowers the Governor with full authority to act, although making him accountable to the Board15. By contrast, in Australia the Bank's Board itself is responsible for making interest rate decisions (in practice on the advice of the Governor), and in the United Kingdom, interest rate decisions are made by a Monetary Policy Committee, compromising both senior officials of the Bank of England and external appointees named by the Chancellor of the Exchequer.

The Governor of the Reserve Bank of New Zealand can be removed from office during his term on grounds of poor performance, including in the conduct of monetary policy. This is highly unusual internationally. In the other three countries summarised here, for example, more traditional provisions, which emphasise the independence of the central bank, mean that a Governor (and in the UK case, other members of the MPC) can be removed only on grounds such as serious misconduct, bankruptcy, or incapacity.

In New Zealand and the United Kingdom, there is a place in the legislation for the establishment of a formal monetary policy target. In the United Kingdom, the Treasury is required to advise the Bank of England at least once a year of the target it is required to pursue, and neither the Governor nor the Monetary Policy Committee has any formal role in devising that target. The targets to date have explicitly required the Monetary Policy Committee to account publicly for deviations from the target. In New Zealand, a Policy Targets Agreement must be reached as part of (re)appointing a Governor, and the Governor is a full party to any negotiations leading up to the new Agreement. The formal accountability framework is built around the PTA.

By contrast, in neither Canada nor Australia is a formal target required under any legislation. In both countries, the inflation targets are now jointly-determined with the Minister of Finance. In Australia this process is now linked to the timing of the appointment of the Governor. The targets provide a benchmark for market monitoring, and a clear signal to firms and households, but are not part of a formal accountability structure for the central bank.

A further area of difference relates to publication requirements. Of the four countries, only in New Zealand (following a US model) is the publication of a Monetary Policy Statement required. Moreover, the Board must satisfy itself that each Statement is not inconsistent with the Policy Targets Agreement16. In each of the United Kingdom, Australia, and Canada regular documents analysing inflation developments and the challenges facing monetary policy are published by the central bank, and attract considerable attention and scrutiny from markets and commentators.

The different focuses of the legislation are perhaps clearest in matter of formal monitoring entities. In New Zealand, the Bank's Board exists largely to monitor the performance of the Governor, and has a formal responsibility to report on that performance. There is nothing similar in the other English-speaking inflation targeting economies. In the Bank of England's case, the Court of Directors is explicitly not responsible for monitoring anything to do with monetary policy. In Australia's case, the Reserve Bank of Australia's Board is the decision making body, (and the Secretary of the Treasury is a voting member of that Board).

This brief summary highlights the difference of emphasis that shaped thinking the character of the New Zealand legislation. The New Zealand reforms took place at a time when the governance and management of the entire central government sector was being reformed, to place considerable emphasis on formal accountability for achieving mandated objectives.

By contrast, in each of the other three countries considered here the legislation is focused more traditionally on safeguarding the independence of the central bank. The price of that greater independence is less-stringent formal accountability for the Governor and/or the monetary policy decision making body. In those countries, inflation targets play a substantial role in the informal monitoring of the central bank, undertaken by markets, media, and politicians more generally. However, it is probably fair to see those targets as having a role more heavily weighted towards communicating clearly the goal that the central bank will be aiming at, rather than to providing leverage to hold decision makers formally to account for their conduct of policy. In practice, perhaps, and in normal times, the extensive public scrutiny all developed country central banks now face, means that the implications of the different models may be small. But if legislation matters mainly in periods of tension, then the New Zealand system is designed to offer better protection against an errant central banker, while the other regimes tend towards protecting a little better against an errant Minister.

1 Section 8 of the Reserve Bank of New Zealand Act

2 The article does not discuss at all the monitoring and accountability for the Reserve Bank's other functions. The Board has a somewhat similar role to play in respect of those functions, but undertakes its monitoring against less-specific performance goals and benchmarks.

3 These issues, and the debates that led to the model finally adopted, are covered in depth "Origins and early development of the inflation target." Reserve Bank of New Zealand Bulletin, Vol 62 No 3 (September 1999) and in chapter 5 of the recently-published history of the Reserve Bank; J Singleton et al, Innovation and Independence: The Reserve Bank of New Zealand 1973-2002, AUP, Auckland, 2006

4 The Minister of Finance also has certain reserve powers relevant to monetary policy (in particular sections 12, 16 and 17 of the Act) although these are not considered further in this article.

5 Note that neither the Act, the PTA, nor the Governor's employment contract has ever tied the Governor's remuneration to inflation outcomes.

6 The Minister can reject a nomination made by the Board, but in this case the Board would be required to provide another nomination.

7 Removal from office is by the Governor-General, by Order in Council ,on the advice of the Minister

8 See references in footnote 3.

9 For example, one of the issues raised in the current PTA was the meaning of "future inflation outcomes...on average over the medium term". In November 2002 the Governor gave a speech ("The Evolution of Monetary Policy in New Zealand") which noted that the Bank would interpret this to mean that its setting of the OCR should be such that projected inflation should be "comfortably inside the target range", in the latter half of a three year forward horizon.

10 Note that a number of these considerations are relevant to most corporate chief executives, and to corporate boards responsible for employing and monitoring those chief executive. Even with a clear long-term goal (maximise shareholder value), actions and decisions taken today affect outcomes over several years, or even decades, and those outcomes are themselves affected by other events beyond the control (or forecasting) of the chief executive.

11 That is, inter alia, taking into account the constraints, such as they are, of section 4(b) of the PTA.

12 The Board has the right (section 53(2) to provide advice to the Governor on any areas of the Bank's operations. However, it has carefully avoided offering advice in advance on the specific monetary policy decisions, partly to avoid compromising its ability to credibly act as an ex post monitor on the Governor's conduct of policy.

13 Although, of course, no staff are present when the Board formally evaluates the Governor's performance.

14 The State Services Commission also has some investigatory powers.

15 Recent Governors have chosen to make interest rate decisions collectively, and by consensus, in the (executive) Governing Council.

16 In the United Kingdom, the Monetary Policy Committee is required by statute to announce its interest rate decision (a level of transparency not imposed elsewhere).