1. Introduction

The past two decades have seen the increasingly widespread adoption of inflation targeting as the framework for monetary policy. While inflation-targeting regimes share a number of similarities, most importantly the focus on an inflation rate as the objective of monetary policy, there are a number of differences in terms of their practical implementation.

One approach which has been the boilerplate for a number of emerging market countries, including that of Brazil, is that practised by the Bank of England. This approach places a strong reliance on model-based forecasts (while still providing some scope for judgmental adjustments). On paper, the Australian approach lies towards the other end of the spectrum. It takes more explicit account of the dual mandate of price stability and full employment. The inflation target is to maintain ‘consumer price inflation between 2 and 3 per cent, on average, over the cycle.’ This definition of the inflation target is perhaps the most flexible amongst the inflation-targeting countries. Despite these differences, the outcomes in Australia in terms of both inflation and economic growth have been at least as good as those in other countries, and the degree of predictability of the monetary policy decision is on a par with other countries.

This paper describes the Australian approach to inflation targeting by drawing together various papers and speeches published by the Reserve Bank of Australia over the past decade and a half. The next section provides details of the inflation-targeting framework in Australia, focusing on the aspects of the framework in Australia that provide scope for greater flexibility. In doing so, it compares the Australian framework to that of the UK. Section 3 describes three episodes over the past decade that demonstrate the flexibility of the Australian framework.

The global financial crisis is providing a significant stress test of the inflation-targeting framework, including in Australia. Section 4 discusses the challenges to the framework and provides a preliminary assessment of how the framework has withstood the challenge. In doing so, it discusses the vexed issue of the appropriate response of monetary policy to asset prices. It also briefly covers two other open issues: the possibility of a price-level target, and the usefulness of the current ‘state of the art’ technology, warning about the dangers of an over-reliance on the technology.

2. Institutional Details of the Inflation-targeting Framework in Australia

The legislation establishing the Reserve Bank of Australia in 1959 specifies that the goal of the Bank is to set monetary policy to best contribute to:

1. the stability of the currency of Australia;^[2]
2. the maintenance of full employment in Australia; and
3. the economic prosperity and welfare of the people of Australia.

The legislation clearly states that monetary policy has both nominal and real objectives, similar to the goals specified for the Federal Reserve Board.^[3]

Unlike a number of other inflation-targeting countries, Australia's adoption of an inflation-targeting framework was evolutionary rather than revolutionary. To a large extent, the evolutionary shift reflected the absence of a catalytic crisis such as the radical program of economic reform that occurred in New Zealand following a prolonged period of poor economic outcomes, or in the case of the UK and Sweden, the sudden departure from the ERM.  It also reflected a recognition that previous monetary frameworks in Australia (which covered the gamut of possibilities) had not been successful in maintaining low inflation (Graph 1).

In part for this reason, at the inception of the inflation target there was no change to the legislated framework, which has not materially altered since its inception in 1959 (Table 1).^[4] Moreover, the existing legislative framework was regarded as satisfactory: the Reserve Bank of Australia (RBA) has always had a high degree of legislated independence (Macfarlane 1996). However, the operational independence was not always as high. The Australian experience reinforces the notion that legislation does not guarantee independence (although it can certainly hinder it); it is the practice which matters most.

In Australia (as in Sweden and Finland), the inflation target was adopted first by the Reserve Bank in 1993, as an operational interpretation of the price stability goal of its legislated mandate.^[5] The inflation-targeting framework in Australia was subsequently verbally endorsed by the government of the day, but was not formally endorsed until 1996, when a new government signed a letter of agreement with a new Governor, upon his appointment. This Statement on the Conduct of Monetary Policy reiterated the Reserve Bank's broad goals stipulated in the Reserve Bank Act, and endorsed the inflation target as the practical interpretation of the medium-term goal of price stability. The status of the Statement has been reinforced by the issuance of a second Statement (almost identical to the first) upon the re-appointment of Governor Macfarlane in 2003 and the appointment of current Governor, Glenn Stevens, in 2006, as well as upon the only change of government that has occurred since the formalisation of the inflation target. The joint announcement of the inflation target between the Government and the central bank helps demonstrate that there is unlikely to be any inconsistency between the setting of monetary and fiscal policy.

The Statement in Australia is very similar to the Letter of Remit from the Chancellor to the Bank of England in defining the practical interpretation of the goal of price stability. The latter is required by legislation, whereas the Statement is not.

The Statement clarifies the possible tension between the real and nominal goals that are embodied in the Reserve Bank Act. It recognises that price stability is the main contribution that monetary policy can make to sustained growth in output and employment, but also recognises that monetary policy-makers need to be cognisant of the effects of monetary policy on the real economy.

In the UK, New Zealand and Europe, there is a clear lexicographic ranking of the real and nominal objectives, with the primary goal of the RBNZ, the Bank of England and the ECB being price stability, while real economy goals are given a subordinate weighting. For example, the Bank of England Act 1998 states that the objectives of the Bank of England shall be (a) to maintain price stability, and (b) subject to that, to support the economic policy of Her Majesty's Government, including its objectives for growth and employment.

The recognition that growth/employment outcomes were an important consideration for the central bank was initially seen as setting the RBA apart from other inflation-targeting central banks where the rhetoric (at least) reinforced the primacy of price stability. But despite these apparent differences, in practice the experience suggests that there has been little substantive difference in this aspect of the framework.

In the event of demand shocks, there is not a large conflict between the real and nominal objectives; the monetary response is the same to meet both objectives, and the actions of all the inflation-targeting central banks would not be significantly different. There may be some difference at the margin in terms of the speed with which inflation is returned to the target range. The Australian system potentially puts more emphasis on minimising short-term output variability and tolerating greater inflation variability. In doing so, it needs to rely less on transmission channels such as the exchange rate which can have a more immediate effect on inflation, than working through the output gap.

In the case of most supply shocks, where the required response to meet the two objectives may differ in the short term, each of the frameworks is sufficiently flexible to cause no material difference in the central bank's response. Only in the event of a large supply shock (such as the oil price shocks of the 1970s) might any difference in response across the different frameworks become apparent.

All of the central banks have both inflation and output (employment) in their objective functions and practice some form of flexible inflation targeting (Svensson 1997). That is, none of the central banks are an inflation ‘nutter’ where inflation only is in the objective function.^[6] The dual mandate of the Reserve Bank of Australia would again tend to suggest a higher weight on output stabilisation than the other central banks, in that if a shock were to drive inflation and output away from target in divergent directions, it would tolerate a slower return of inflation to target (i.e. inflation variability would be higher) to reduce output variability.

To some extent, the framework in Australia is similar to an approach of targeting nominal income growth, without the attendant problems that may beset the latter. These include the problem of public communication, where the public is likely to be more understanding of inflation than the more nebulous concept of nominal growth, as well as the problem that nominal income is often subject to sizeable revision by the statistical agency.

Accountability

In exchange for the independence and greater discretion that an inflation-targeting framework permits a central bank, it is important that the central bank be held accountable on a regular basis to the public. To some extent, financial markets are continuously holding a central bank accountable for its actions. Any deviation from a ‘sensible’ course for monetary policy would likely be penalised by movements in the longer end of the yield curve or by movements in the exchange rate.

However, more formally (and democratically), the Statement on the Conduct of Monetary Policy requires that the Governor and other senior central bank officers appear before a parliamentary committee twice a year to explain their thinking and actions on monetary policy. These appearances are in public, and allow the public, via the parliament, to question whether the central bank's actions remain consistent with the inflation target.

The RBA's accountability is collective rather than individual as is the case with the Bank of England. The Board as a whole is accountable, in large part reflecting its composition where the majority of members are business people rather than practising economists. Individual accountability in the RBA's case may compromise the ability of the business members of the Board to take decisions in the national interest rather than their sectoral interest (Stevens 2007).

There are no explicit penalties imposed on the RBA should inflation move outside the 2 to 3 per cent range. In part, this reflected a concern that a hard-edged band with explicit penalties would introduce a discontinuity in the payoff function which might induce unnecessary instability in monetary policy settings. The penalties in the Australian framework are less explicit. The RBA is increasingly uncomfortable at an increasing rate as inflation moves away from the desired level.

In the UK, should inflation fall below 1 per cent or rise above 3 per cent the Monetary Policy Committee is required to write a letter of explanation to the Chancellor. The accountability in Australia is, in this regard, less formalised, with the penalties being imposed via financial markets, the parliamentary testimonies and through loss of reputation.

Communication

More generally, the adoption of inflation targeting has coincided with a large increase in the written and spoken output of central banks. Increased communication and transparency is beneficial for any monetary policy framework but it has played a particularly prominent role in inflation-targeting regimes. In part this reflects the starting point of many central banks that adopted inflation targeting: they generally had a poor inflation history and low credibility with the public and financial markets. Thus a high degree of communication and transparency was necessary to build credibility as quickly as possible, to enhance the transmission of monetary policy and to provide an anchor for the public's expectations of future inflation. In this respect, Mervyn King (1997) has characterised inflation targeting as ‘trust building by talk’ (see also Kuttner and Posen 1999). Indeed, the inflation target's role as a coherent framework for the central bank to communicate with the public may be one of its greatest benefits.

There are two important issues for the central bank to communicate to the public. Firstly, the central bank should use every opportunity to reiterate the target itself and emphasise their commitment to it, to increase the likelihood that it acts as an anchor for inflation expectations. Secondly, it needs to communicate a general understanding of the central bank's reaction function.

Speeches by the Governor and senior officials of the RBA have been a primary vehicle to enhance the status and understanding of the inflation-targeting framework in Australia. Aspects of the framework may also be conveyed in the press release that the RBA issues announcing changes in monetary policy, which includes detailed reasoning for the policy change. From the inception of the inflation-targeting regime, the press releases have made explicit reference to the inflation target as the primary justification for the monetary policy decision.

The other major source of information on the central bank's reaction function has been the quarterly inflation reports or their equivalent. The RBA had traditionally published a detailed assessment of its monetary policy actions only in its annual report, with very brief accounts of current economic conditions published quarterly in the RBA Bulletin. However, with the inception of the inflation target, the quarterly assessments on its view of the current and future state of the Australian economy gradually increased in length and depth of analysis, to reach their present form where they are published as a stand-alone document – the quarterly Statement on Monetary Policy.^[7] Moreover, over time the RBA has provided more detail about its forecasts such that currently numerical forecasts for output and inflation are presented each quarter. In this sense too, the RBA's communications strategy has evolved more gradually than other central banks, but practice has converged over time.

The inflation target has served as a useful organising framework for this document. Analysis is provided on all of the variables that feed into the RBA's overall outlook for inflation and output, with the concluding section focusing on the outlook for inflation and the risks to that outlook.

Ex post it is important to explain why observed CPI outcomes are, or are not, consistent with the central bank's view of inflation, and some time is spent in the RBA's quarterly Statement on Monetary Policy explaining the details and implications of the quarterly CPI outcomes. Whether or not the central bank accurately forecasts the next quarter's outcomes to the nearest decimal point, however, implies a degree of knowledge and precision that a central bank does not possess. Thus the focus of the public and financial markets should be on the accuracy of the central bank's inflation forecast at the policy horizon.

In recent years, the RBA has enhanced its communication in a number of ways.^[8] While the details of the RBA Board's decision had been made public in a press release at the time of changes in monetary policy, they had not been made public when the stance of monetary policy was left unchanged. Since 2007, a detailed press release follows every decision.

Moreover, since the middle of this decade, the introduction to the Statement on Monetary Policy had summarised the deliberations taken by the RBA Board at their monthly meetings on monetary policy but the detailed minutes of the meeting were not published. Again, since the end of 2007, the minutes of the Board meeting are published with a two-week lag. The individual views of the Board members are not attributed, which is appropriate given the composition of the RBA Board, where only two of the members are full-time central bankers, with the majority of the others being business people.

It is however possible that there is a limit to transparency. If the central bank's communications suggest that it has greater knowledge or greater precision in its inflation control than it does in reality, then when this becomes apparent and the public's expectations are disappointed, the central bank's credibility may be damaged. There is likely to be some advantage from being modest in communicating what the central bank knows and can do. This does not mean that the central bank can hide behind a veil of uncertainty but rather that it should acknowledge that uncertainty and convey to the public the implications of the uncertainty for its inflation forecasts and policy actions.

Flexibility

There are a number of aspects of the Australian framework which have the potential to deliver greater flexibility in the practical implementation of the inflation target. These include the dual mandate specified in the goals, the emphasis on a (thick) point target rather than a band, and the specification of a horizon for the target.

The ‘2 to 3 per cent’ specification may appear to suggest that the inflation target in Australia is a narrow band. However, the intent of this language is that the target is a ‘thick point’ rather than a range, or in other words, the desirable rate of inflation for the Australian economy is ‘2 point something’ (Stevens and Debelle 1995).

Relatedly, and at the heart of the output/inflation variability trade-off, is the horizon over which inflation is returned to target. The RBA's inflation target is defined ‘over the course of the cycle’, ‘allowing for the natural short-run variation in inflation’. While the allowable short-run variation is not defined, the aim is that in most circumstances, the Bank's inflation forecast should lie between 2 and 3 per cent at the monetary policy horizon, that is, the maximum impact of changes in monetary policy on inflation.

In circumstances where the forecast lies outside the range over the policy horizon, the forecast path for inflation should be such that inflation would be expected to return to between 2 and 3 per cent within a reasonable period, that is, the trend in inflation should be clearly back toward the target range. In Australia, there is thus no specific horizon for the inflation target, allowing greater flexibility in policy-setting, and less variability in output. The Bank of England has been more explicit in focusing on an optimal horizon of eighteen months to two years (Batini and Haldane 1999; Bean 2003).

3. Flexibility in Practice

To illustrate the practical application of the inflation-targeting framework in Australia and its flexibility, it is useful to focus on the operation of monetary policy in three particular episodes (Graph 2). The first in 1994 demonstrates the pre-emptive monetary policy response to demand shocks. The second is the Asian crisis where there was a major negative shock to external demand and a large depreciation of the exchange rate. The third is the introduction of the goods and services tax in 2000 which boosted the price level by 3 per cent in one quarter.

3.1 Responding to demand shocks

By the middle of 1994, signs of inflationary pressure were evident as economic growth was accelerating toward 6 per cent. The labour market had tightened as the unemployment rate had declined by around 3 percentage points in two years, and wage pressures were building. Inflation expectations in financial markets had increased by around 2 percentage points. However, underlying inflation was still at 2 per cent.

In anticipation of a rise in inflation, and reflecting its inflation forecasts, the Bank raised the cash rate by 275 basis points in three moves over the second half of 1994. The press releases that accompanied these tightenings emphasised the need to act pre-emptively to curtail the upward movement in inflation. Subsequently, in 1995, inflation did indeed rise to slightly above 3 per cent in underlying terms, but by less than would have been the case had the tightening in monetary policy not taken place. Indeed, the financial markets anticipated significantly more tightening than actually occurred, reflecting their lack of faith in the credibility of the relatively new inflation-targeting framework.

In 1996, as demand pressures were easing, inflation was forecast to move lower, even though inflation at the time was still 3.1 per cent. The stance of monetary policy was moved back toward a more neutral setting, reflecting the inflation forecasts.

In this situation, the overall monetary policy decision was relatively straightforward as the required movement was the same to meet both inflation and output goals, as is the case in the event of demand shocks.

This episode also demonstrates the flexibility offered by the medium-term nature of Australia's inflation target. In 1994, monetary policy could have been tightened sufficiently to ensure that underlying inflation did not rise above 3 per cent. But the rise was forecast only to be temporary after which inflation was expected to fall back below 3 per cent, hence the additional tightening was not delivered. As a result, there was less volatility in output. The critical issue in determining the extent of the tightening was whether inflation was forecast to return to the 2 to 3 per cent range within the policy horizon (around 18 months).  In both instances, the flexibility of the framework also allowed the decision of whether the target was in jeopardy to be reassessed as the situation evolved.

3.2 The Asian crisis and exchange rate shocks

At the onset of the Asian crisis, the Australian economy was growing at around trend rates, with domestic demand beginning to accelerate, and underlying inflation at 1.6 per cent. Monetary policy had been eased over the prior year or so in anticipation of the decline in inflation that subsequently occurred. Thus the shock hit the Australian economy at a time when it was in reasonable shape with the stance of monetary policy already relatively expansionary.

Exports to east Asia accounted for around one-third of Australia's exports. In the year following the onset of the crisis, Australia's exports to the region declined by nearly 20 per cent, directly subtracting around 1 percentage point from aggregate growth. Thus the decline in output in the east Asian region represented a significant negative demand shock to the Australian economy. Australia's terms of trade also fell sharply as commodity prices declined, further exacerbating the decline in export demand.

In the event, inflation rose by less than was forecast, in part because of the decline in the pass-through of the exchange rate depreciation, as well as a greater-than-expected disinflationary impulse from the Asian region which put downward pressure on import prices.

If policy had been set to ensure that inflation did not rise above 3 per cent, the rise in interest rates would have exacerbated the contractionary shock to foreign demand. With the benefit of hindsight, given the lower-than-expected inflation outcomes, this would have resulted in a significant undershooting of the inflation target.

The flexible inflation target served as a useful framework to think about the Asian crisis. Strong consideration was given to the goal of output stabilisation because the inflation target in the medium term was not felt to be in jeopardy. In addition, the policy credibility that had built up since the adoption of the inflation-targeting regime also allowed the Reserve Bank greater flexibility in its policy response.

3.3 Price-level shocks

On 1 July 2000, a 10 per cent broad-based goods and services tax was introduced, replacing a wide range of indirect taxes that had been levied primarily on goods. As a result, the price level measured by the CPI rose by 3 per cent. Year-ended CPI inflation was thus boosted by 3 percentage points until this rise in the price level dropped out of the calculation one year later. Underlying measures of inflation were similarly boosted because of the broad-based nature of the tax. In the September quarter 2000, the CPI inflation rate was 6.1 per cent, while the weighted median inflation rate (before accounting for the effect of the tax) was 5.4 per cent.

The increase in the price level was fully anticipated by households and financial markets. However, the increase in the CPI did not feed through to medium-term inflation expectations or into wage outcomes.^[9] This was helped by the provision of income tax cuts at the same time (and adjustment of benefits) which offset the effects of the rise in the price level so that real incomes were not harmed.

The Bank did not seek to offset the effect of the GST on the price level. It assumed that the boost to the price level would be once-off, and that the credibility of the inflation-targeting framework would ensure that inflation expectations remained anchored at the target rate. It announced this strategy well in advance of the event, to help condition expectations that the rise in inflation would only be temporary. Again, the assessment was made that the inflation target was not in jeopardy in the medium term with year-ended inflation forecast to be within the targeted range once the effect of the GST had passed.

In the event, the Bank's strategy was successful and the inflation target proved to be credible. Almost all of the public discussion at the time on the appropriate setting for monetary policy focused on the inflation outcomes excluding the influence of the changes in the tax rate (Graph 4). The price level shock was mostly regarded as a non-issue. Medium-term inflation expectations remained stable.

Monetary policy was tightened reflecting the combination of other sources of price pressure, including strong economic growth, a sharp rise in oil prices and the depreciation of the exchange rate.

4. Current Challenges

The Global Financial Crisis

A long-standing criticism of the inflation-targeting framework was that it had not been properly tested. I find this criticism somewhat disingenuous. Firstly, the framework itself has contributed to the stability. Secondly, the inflation-targeting period had not been completely benign. It included the Asian crisis, the Russian debt default and the collapse of LTCM, large falls and large rises in oil prices. In Australia's case, it also included a large depreciation and large appreciation of the exchange rate. As shown above, the targeting framework handled all these events well.

Moreover, for Australia, the period also included one of the largest rises in the terms of trade in the country's history. In the past, large terms of trade increases in Australia had led to serious macroeconomic problems with a rapid inflationary boom, followed by bust. This time around, the inflation rate rose to 5 per cent for a year or so, but has since returned to the target.

Notwithstanding this, the inflation-targeting framework is clearly being severely tested in the current circumstance. While the inflation rate in most countries is currently consistent with the target, many forecasts assume a rapid fall and even deflation, with, for example, the Bank of England currently forecasting a prolonged undershooting of the target. At the same time, in many countries, conventional tools of monetary policy have been exhausted with policy interest rates at zero, resulting in the widespread application of unconventional policy responses.

One obvious question to ask is whether other monetary frameworks would have fared any better. In that regard, it is worth noting that the three major economic areas, none of which have an explicit inflation-targeting framework have suffered at least as large an economic dislocation as the inflation-targeting countries, and in the case of Japan, considerably larger. I do not see a strong argument that previous frameworks would have delivered any markedly different outcomes. The policy response by inflation-targeting central banks has been very rapid as the crisis has unfolded, notwithstanding the fact that in some cases, the current level of inflation was above the target range.

However, a possibly more pertinent question to ask is whether the inflation-targeting framework itself led to any mistakes that contributed to the crisis and that would have been avoided under other frameworks. One criticism that has been advanced is that the supposedly narrow focus on CPI inflation has led central banks to neglect developments in other prices, particularly asset prices. Thus, the long-running debate on the role of asset prices has come to the fore again.

Before discussing the asset price issue, again it is worth repeating that the issue is whether inflation targeting itself led to monetary policy settings being easier than would have been the case in other frameworks. I do not see a strong argument in favour of this proposition. Asset price booms and busts and credit-related booms have occurred under many different monetary regimes, including in highly regulated financial systems.

Asset prices

The debate prior to this crisis can be (perhaps simplistically) characterised as between those who argued that an inflation-targeting central bank should care about asset prices to the extent that they affected the forecasts of output and inflation over the policy horizon, and those who argued that additional attention needed to be paid to asset prices and the possibility of credit imbalances. The prominent variant of the latter was the Borio-White criticism that central banks had an asymmetric response to asset prices where asset prices booms were not resisted but asset price busts were responded to vigorously by monetary policy, thereby sowing the seeds of the next asset price bubble and bust. Note that this argument was applied to many central banks, most obviously the US, not just those who targeted inflation.

In my view, the current episode vindicates the position that monetary policy, narrowly defined as the setting of the policy interest rate, should be confined to targeting inflation. Set interest rates primarily to achieve the inflation goal as that, in itself, contributes to sizeable social gains. A departure from that runs the risk of losing the nominal anchor that the inflation target provides.

But other tools, most notably the much-touted (although not clearly defined) macro-prudential instruments, should be used to address asset price and credit imbalances. I do not think that a slightly tighter setting of interest rates would have prevented the development of the imbalances that have led to the current financial crisis. When human psychology is such that optimism about asset price rises is at the fore, then an excessively stringent setting of interest rates would be required to suppress the optimism. The Australian and Scandinavian experience in the late 1980s shows the sort of interest rate settings required to achieve such an outcome. In that example, a credit boom and bubble-like asset price dynamics took hold and only a very high setting of real interest rates ultimately curtailed that, but at the cost of a historically high level of unemployment.

I do not think it would be socially acceptable or desirable to endure the level of unemployment that would come with the high interest rates necessary to pop the bubble. It is asking too much of the single monetary policy instrument, namely, the targeted short-term interest rate to target both financial excesses and inflation.

Nor do I believe there is much to be achieved by ‘leaning against the wind’. The wind that is blowing in most episodes of credit booms is generally at least gale force. Setting interest rates a bit higher in such circumstances is likely to be close to futile when such credit dynamics take hold. Again, what would be the point of undershooting the CPI inflation target and enduring a higher than desirable level of unemployment with little to be gained. How would such actions be explained to the public?

Instead I would argue that other instruments should be deployed that more directly address, at the source, the provision of the excess credit that leads to these imbalances. That is the discussion which needs to be (and is) taking place. These instruments would be a supplement to the setting of the monetary policy instrument to achieve the inflation target. I am not arguing that these alternative instruments will be successful in countering asset price bubbles and credit imbalances, because I think bubbles are a permanent feature of the landscape resulting from entrenched human behaviour. Rather, instruments other than the interest rate tool are likely to be more effective without some of the attendant social costs.

This assumes that it is possible to identify that excesses are indeed occurring. I do not see this as being as large an obstacle as others, most notably Alan Greenspan (at least his earlier pre-crisis views). Identification of over-inflated asset prices and excess credit expansion is no more or less easy than the identification of consumer prices pressures. Moreover, a number of macro-prudential instruments under consideration would have some measure of automatic stabilisation built in, reducing the need for discretionary decision-making.

Price-level or inflation target

The inflation target in Australia is defined on average over the [business] cycle, which, if taken literally, suggests that it may be interpreted as a price-level, rather than an inflation-rate, target. However, the averaging refers more to the distribution of inflation outcomes than to a strict average of CPI outcomes. That is, the intent is that over the course of the business cycle, the bulk of the distribution of year-ended inflation outcomes should lie between 2 and 3 per cent, not that the annualised average inflation rate from the start of the business cycle to the end should necessarily lie between 2 and 3.

Practically, this translates into a policy of letting bygones be bygones and allowing ‘base’ drift in the price level. A shock to the price level which temporarily lowers the inflation rate below 2 per cent does not imply that monetary policy will be set to ensure an offsetting period of high inflation.

While no country has adopted a price-level target, Canada is actively considering it, and I am sure my co-panellist, Dave Longworth, will discuss this issue. Hence I will only touch on this briefly.

Ball, Mankiw and Reis (2003) argue that a price-level target rather than an inflation-rate target should be the optimal goal for a central bank. The primary justification for their proposal is that an inflation-rate target is costly because it does not permit long-run predictability of the price level, which has first-order welfare effects in their models. Ball et al's argument depends importantly on their assumption that agents have completely forward-looking expectations. After a positive shock to the price level, knowing that the central bank has a price-level target, agents fully anticipate the disinflation required to return the price level to target. As the disinflation is fully anticipated, there are no expectational errors and hence there is no cost in terms of lost output from the disinflation.

In a model where there are costs to disinflation (which would include a model with backward-looking expectations or less than full credibility of the central bank) the gains from full predictability of the price level may not offset the costs of occasional disinflations following positive price level shocks.^[10]

More generally, I think the strongest argument against a price-level target is that the absence of long-term indexed contracts suggests that the benefits of long-run price predictability are not that large.

The technology

Inflation targeting is often characterised as inflation-forecast targeting, most notably by Lars Svensson. The approaches taken in forecasting inflation by the various central banks have some similarities and some differences. The Bank of England, RBNZ, Riksbank and Norges Bank, inter alia, make significant use of large scale macro-models, whereas the RBA puts greater reliance on single-equation models. A DSGE model exists but is only a small input to the policy process.

The RBA's inflation forecast is based on a suite of single-equation models of inflation supplemented with judgmental adjustments. In this respect, the RBA's approach to inflation forecasting corresponds to Blinder's recommendation: ‘Use a wide variety of models and don't ever trust any of them too much’ (Blinder 1999, p 12).

The RBA has used larger macroeconomic models in the past, but generally found their performance to be unsatisfactory from a monetary policy perspective. In part, this is because the large scale of the models diminishes the ability to observe the key macroeconomic relationships central to the policy decision.

In the UK, the Bank of England also makes extensive use of a large-scale macro-model, and a significant portion of the policy formulation process is taken up with the discussion of the necessary judgmental adjustments to the equations of the model to reflect current developments. In addition, various smaller alternative models are used to provide information on relationships not well handled by the central model.^[11]

As always, these models are based on the average outcomes over the estimation period. At any point in time, the central bank may have some knowledge about the current and future residuals in the equation which are not easily incorporated into the modelling framework. Hence, all of the central banks make judgmental adjustments to the forecasts generated by the models.

To some extent, the models tell you the easier part, which is a summary of the main inter-linkages in the economy. The harder part is often knowing where exactly (or even approximately) the macroeconomic conjuncture is, and the extent to which behavioural changes are moving the economy away from the model's framework. This is where a central bank always has to rely on the detailed knowledge of its staff, often supplemented with an extensive liaison program with the business community (which the RBA has found to be invaluable). The human element, I believe, is a critical part of the policy process.

More generally, I am uncomfortable with the reliance placed on New Keynesian models in the policy process.^[12] As a heuristic device to ensure consistency of thinking, then I think such models can be beneficial. But I do not see them as having contributed much to the real-world policy debate. I have not seen them as providing insights that were not already self-evident.

Moreover, I am sceptical of many of the underlying assumptions and calibrations in these models. The models appear to have worked well in the period of relative stability that has characterised the Great Moderation. But then, random-walk models of inflation worked pretty well during this period too. I fear that parts of the central bank community have become blinded by the science and the technique, and that there has been excess allocation of intellectual resources to this pursuit (which perhaps could have better been focused on financial stability issues).

In the current circumstance, these models have not been useful because they generally do not handle credit well, if at all. Obviously work is being done to address this and Mike Woodford's paper yesterday indicated substantial progress. But it seems to me that these models need to add a healthy dose of Post Keynesianism to their New Keynesian foundations.

When the input of models would perhaps have been most beneficial, we have been trying to apply linear frameworks to a world which in the current circumstances is at its most non-linear.

5. Conclusion

Inflation targeting in Australia has coincided with a period of low and stable inflation, and a prolonged economic cycle with a high average rate of growth, which has only recently come to an end. Monetary policy alone clearly cannot take the credit for these outcomes, but one can argue it has been supportive of them. Monetary policy appears to have been broadly successful in its aim to ‘let the economy grow as fast as possible, consistent with the inflation target’.

Australia's inflation-targeting framework has remained basically unchanged since its adoption in 1993. In 1993, the Australian framework was at the flexible end of the spectrum of inflation-targeting practice, and was criticised for being too lax. The economic outcomes suggest that the greater flexibility of the Australian approach has not been at the expense of economic performance. Equivalently, the experience also suggests that a rigid application of an inflation-targeting framework may not be necessary, and that there may be elements of the Australian approach which may be applicable to emerging market economies considering adopting an inflation target.

Over the years, there has been a convergence in practice, arguably toward the more flexible end of the spectrum, with an increased focus on medium-term outcomes. To some extent, this shift has been possible because of the increased credibility of the inflation-targeting central banks. While some differences may still remain across central banks in terms of the structure of the frameworks and in communication practices, in practice, these differences are generally superficial and have not appeared to have had a discernible impact on economic outcomes.

The inflation-targeting framework is being severely stress-tested at the moment. I would argue that it is withstanding the test reasonably well, and at least as well as the alternatives. I do not accept the argument that the supposedly narrow focus of an inflation target has contributed to the current conjuncture. Rather I would argue that the current situation reinforces the idea that the standard operation of monetary policy should continue to be directed at the inflation target, but that it needs to be supplemented by other tools to address the ever-present threat of financial imbalances.

References

Ball L, NG Mankiw and R Reis (2003) ‘Monetary Policy for Inattentive Economies’, NBER Working Paper No 9491.

Batini N and AG Haldane (1999), ‘Forward-Looking Rules for Monetary Policy’, Bank of England Working Paper No 91.

Batini N and A Yates (2001), ‘Hybrid Inflation and Price Level Targeting’, Bank of England Working Paper No 135.

Bean C (2003), ‘Asset Prices, Financial Imbalances and Monetary Policy: Are Inflation Targets Enough?’, paper presented at the BIS Conference on Monetary Stability, Financial Stability and the Business Cycle, 28–29 March, Basel.

Blanchflower D (2009), ‘The Future of Monetary Policy’, Open Lecture at Cardiff University, 24 March.

Blinder AS (1999), Central Banking in Theory and Practice, MIT Press, Cambridge MA.

Debelle G (1998), ‘Inflation Targeting and Output Stabilisation’, in M Blejer, A Ize, A Leone and S Werlang (eds), Inflation Targeting in Practice: Strategic and Operational Issues and Application to Emerging Market Economies, IMF, Washington DC, Chapter 6.

Debelle G (1999), ‘Inflation Targeting and Output Stabilisation’, RBA Research Discussion Paper No 1999-08.

Fraser B (1993), ‘Some Aspects of Monetary Policy’, Reserve Bank of Australia Bulletin, April, pp 1–7. Available at <https://www.rba.gov.au/publications/bulletin/1993/apr/pdf/bu-0493-1.pdf>

King M (1997), ‘Changes in UK Monetary Policy: Rules and Discretion in Practice’, Journal of Monetary Economics, 39(1), pp 81–97.

Kuttner KN and AS Posen (1999), ‘Does Talk Matter After All? Inflation Targeting and Central Bank Behavior’, Federal Reserve Bank of New York Staff Report No 88.

Macfarlane IJ (1996), ‘Making Monetary Policy – Perceptions and Reality’, Reserve Bank of Australia Bulletin, October, pp 32–37. Available at <https://www.rba.gov.au/publications/bulletin/1996/oct/pdf/bu-1096-3.pdf>

Pagan A (2003), ‘Report on Modelling and Forecasting at the Bank of England’, Report to the Court of Directors of the Bank of England. Available at <http://www.bankofengland.co.uk/archive/Documents/historicpubs/qb/2003/qb030106.pdf>

Stevens, GR (1999), ‘Six Years of Inflation Targeting’, Reserve Bank of Australia Bulletin, May, pp 46–61.
Available at <https://www.rba.gov.au/publications/bulletin/1999/may/pdf/bu-0599-2.pdf>

Stevens GR (2007), ‘Central Bank Communication’, Address to the Sydney Institute, 11 December.
Available at <https://www.rba.gov.au/speeches/2007/sp-gov-111207.html>

Stevens GR and G Debelle (1995), ‘Monetary Policy Goals for Inflation in Australia’, in AG Haldane (ed), Inflation Targeting, Bank of England, London, pp 81–100.

Svensson LEO (1997), ‘Optimal Inflation Targets, “Conservative” Central Banks, and Linear Inflation Contracts’, American Economic Review, 87(1), pp 98–114.

Williams JC (1999), ‘Simple Rules for Monetary Policy’, Board of Governors of the Federal Reserve System, Finance and Economics Discussion Series No 1999-12.

Endnotes

This paper draws on a number of articles I have written on this topic over the years and has benefited from the comments of many colleagues. The opinions expressed are those of the author, and not necessarily those of the Reserve Bank of Australia. [1]

These goals were determined for a fixed exchange rate regime. With the floating of the currency in 1983, the first of these goals has been interpreted to mean price stability, rather than literally the stability of the exchange rate. [2]

The similarity reflects the fact that the goals of both central banks were determined immediately after the Second World War (in Australia's case, these goals had been given to the Commonwealth Bank in 1945), with the Depression still relatively fresh in the minds of the legislators. [3]

At that time, the central banking and trading functions of the (then) government-owned Commonwealth Bank were separated, with the Reserve Bank of Australia created to handle the former and the Commonwealth Bank continuing as a commercial bank. The Commonwealth Bank's role as Australia's central bank dates back to the early part of the 20^th century. [4]

A case can be made that the first public exposition of the inflation target came in 1993 in a speech by then Governor Fraser (1993): ‘My own view is that if inflation could be held to an average of 2–3 per cent over a period of years, that would be a good outcome’. Such a formulation was repeated and refined in subsequent speeches by Fraser and the inflation target was gradually elevated in prominence in the eyes of the public and financial markets. In particular, the inflation target was prominent in the justification for the interest rate increases in the second half of 1994. For more details, see Stevens (1999). [5]

Note that even an ‘inflation nutter’ central bank will have a positive weight on output in its reaction function, even with a zero weight on output in its objective function (see Debelle 1998). [6]

The 1996 Statement on the Conduct of Monetary Policy reinforced the status of these documents as a primary means of accountability and communication for the RBA. [7]

See ‘Central Bank Communication’, Address by Governor Glenn Stevens to the Sydney Institute, December 2007. [8]

Some wage agreements did have a clause which allowed for larger increases should the boost to the CPI be greater than expected. [9]

Batini and Yates (2001) and Williams (1999) examine this issue more formally. [10]

For more detail on forecasting at the Bank of England, see Pagan (2003). [11]

Blanchflower (2009) addresses this concern in more eloquent detail. [12]