In early 1991 the Bank of Canada and the Government of Canada jointly announced targets for reducing the rate of inflation on the way to price stability. In late 1993, following the election of a new government and upon the appointment of a new governor, price stability was confirmed as the longer term goal of monetary policy and the targets for controlling inflation were extended.

In the first part of the paper I set out the background situation leading up to the introduction of the targets in February 1991 and the objectives that the targets were intended to achieve. The following section focuses on the initial experience with the targets¹, the announcement of the extension of the targets in December 1993, and the subsequent experience. Section 3 lays out how the Bank of Canada dealt with some of the technical issues that had to be resolved in designing a target path for reducing and controlling inflation. The final section draws some tentative conclusions from the Canadian experience.

I.  The Background Leading up to the Announcement of the Targets and the Objectives for the Targets

Monetary policy in Canada, as in many other countries, has as its goal the achievement and maintenance of price stability. By fostering confidence in the value of money, monetary policy makes its contribution to the ultimate objective of public policy--a well-functioning economy.

Even though the achievement and maintenance of price stability is the longer term goal of monetary policy, central banks have found it useful to have an intermediate nominal target to help them guide monetary policy over the short and medium term. Such a target acts as an anchor to help the monetary authorities avoid cumulative errors. For countries on a fixed exchange rate regime, the exchange rate can serve as a nominal anchor to the system, provided that the partner country has a domestic nominal anchor. For countries on a flexible exchange rate regime, monetary aggregates, credit aggregates and nominal spending are all potentially useful intermediate targets for monetary policy. The ability of such variables to function effectively as intermediate targets is based both on the stability of their empirical relationship to the goal variable and on their relationship to the instrument of monetary policy.

In Canada, the narrow monetary aggregate M1 was used as the intermediate target of monetary policy between 1975 and 1982. It was dropped as a target in 1982 primarily because innovations by Canadian financial institutions resulted in the introduction of new financial instruments which significantly weakened the linkage of M1 to nominal spending. With the demise of M1 as a target, the Bank embarked on a protracted empirical search for an alternative monetary aggregate target, but no aggregate was able to bear the weight of being a formal target. Instead, the Bank used both monetary and credit aggregates in the lesser role of information variables or guides to policy. Although nominal spending was considered as a possible intermediate target, there were a number of practical reasons for not choosing to make nominal spending the formal target of policy. Thus, from 1982 to 1991 monetary policy in Canada was carried out with price stability as the longer term goal but without intermediate targets or any specified time path to the longer term goal.

The inflation-reduction targets were jointly announced by the Bank and the government on February 26, 1991. The government's announcement came as part of its annual budget, while the Bank issued a press release and a background note setting out practical details regarding the operation of the targets.

The targets set out an explicit path towards price stability. The first guidepost was set for the end of 1992 (22 months after the announcement) and provided for a 12-month rate of increase in the CPI of 3 percent, to be followed by 2 1/2 percent for mid-1994 and 2 percent by the end of 1995, each with a band of plus and minus 1 percent. It was specified that after 1995 there would be further reductions of inflation until price stability was achieved.

The targets were intended to achieve both near-term and longer term objectives. In the near term they were intended to prevent the price shocks that were then buffeting the economy from leading to a further wage-price spiral and to help reduce the prevailing inflation expectations. In particular, the sharp rise in oil prices following the invasion of Kuwait and the anticipated rise in the price level resulting from the replacement of the existing federal sales tax at the manufacturers' level by a broader VAT-type goods and services tax (GST) at the beginning of 1991 raised concerns about a deterioration of inflation expectations and the possibility of additional upward pressure on inflation. By providing a clear indication of the downward path for inflation over the medium term, the key near-term aim of the targets was to help firms and individuals see beyond these shocks to the level of prices to the underlying downward trend of inflation at which monetary policy was aiming and to take this into account in their economic decision making.

In the longer term, the inflation-reduction targets were designed to make more concrete the commitment of the authorities to the goal of achieving and maintaining price stability and to add to the credibility of that goal. By making explicit the path to price stability, the announcement of the targets was aimed at facilitating the achievement of the longer term goal of price stability. To quote from the background note to the targets, "if Canadians begin to base their economic decisions on this declining path for inflation, the objectives can be readily achieved and will contribute to lower interest rates."

Although the Bank of Canada had long been on record regarding the ultimate goal of price stability, it was becoming apparent that this general commitment to price stability was not having the desired effect on inflation expectations and that it would be useful to buttress it by either an intermediate target or an explicit downward path for inflation. In the absence of a viable intermediate target, this argued for inflation-reduction targets as a way of making more concrete and more credible the nature of the commitment.

Moreover, since the targets were presented as a joint commitment of the Bank and the government to inflation reduction and price stability, it was made clear that the government was supportive of the price stability goal. It had earlier been argued that the credibility of this goal and the willingness of individuals and companies to act on it was lessened by the concern that the government was not fully committed to it and that such a goal would therefore not be achievable. This concern was heightened by the fact that the longer term inflation projections in the budgets of preceding years did not seem to be consistent with the Bank's announced goal of price stability. The projections in the 1991 budget, in contrast, were entirely compatible with the inflation-reduction targets.

To the extent that the credibility of the inflation-reduction objective was enhanced by the targets, the adjustment of inflation expectations could be accelerated. The costs of adjustment to lower inflation could thus be reduced and monetary conditions could be easier than otherwise would have been the case. Although it was not clear that the announcement of the targets would have any such short-run effect on expectations, it was anticipated that, at a minimum, the achievement of the first milestone at the end of 1992 would enhance the credibility of the subsequent targets and hence facilitate the achievement of these targets as well as the movement to price stability thereafter.

Finally, by providing information on the specific objectives to which the monetary policy actions of the Bank would be directed, the targets were intended to make the Bank's actions more readily understandable to financial market participants and to the general public. They would thus provide a better basis than before for judging the performance of monetary policy and would further improve the accountability of the Bank.

II.  The Initial Experience with the Targets and Their Extension

The announcement of the targets in early 1991 was initially met with considerable skepticism. Business economists and forecasters expressed doubt that the target of 3 percent at the end of 1992 (2 to 4 percent band) would be met. However, the rate of inflation came down rapidly in 1991 as the cyclical weakness of the economy and the major restructuring of Canadian industry in response to increased globalization put strong downward pressure on price increases and then on wage increases. As a result, changes in expectations regarding future rates of inflation appeared to become widespread in the latter part of 1991. In the event, the 12-month rate of increase in the CPI at the end of 1992 was 2.1 percent (2.0 percent for the CPI excluding food and energy), just above the bottom end of the band (Figure 1).

As noted earlier, there were both near-term and longer term objectives that the targets were expected to achieve. The near-term objective, to avoid a resurgence of inflation pressures following the once-off price increases related to the oil price increase and the introduction of the GST, was clearly achieved. Of course, the oil price increase was fortuitously reversed in early 1991 but, on the other hand, unexpected increases in excise taxes, especially on tobacco, put added upward pressure on price levels. Moreover, the substantial depreciation of the Canadian dollar beginning in late 1991 and continuing through most of 1992 provided a further price level shock to the system. The absorption of the large shocks to prices from indirect taxes and the currency depreciation without a wage-price spiral can be considered a major accomplishment of policy over the period. The clear statement by the Bank in the context of the targets that the first-round effects of such shocks would be accommodated but not the second-round effects probably contributed, along with the sluggishness of the economy, to the avoidance of a wage-price spiral in response to the shocks.

As for the longer run objective of the targets--to facilitate the achievement and maintenance of price stability--considerable progress was clearly made. Nonetheless, it is not possible to be definitive about the relative importance in the achievement of the first target at the end of 1992 of the sluggishness of the economy (i.e., gaps in the output and labor markets) and changes in expectations resulting from announcement of the targets. On the one hand, it appears that an augmented Phillips curve equation for price inflation which uses survey estimates for expected inflation is able to track the broad downward movement of inflation quite well. This might suggest that there is no need to resort to explanations involving increases in the credibility of policy to explain the pace of disinflation in Canada. On the other hand, to some extent the decline in the survey estimates of inflation might be associated with the targets. Also, one can be somewhat skeptical about such projections in the uncharted territory of low inflation on the basis of an equation estimated over a period in which inflation was rarely below 4 percent. One should therefore be cautious about using such equations for price inflation to infer that the targets have played little or no role in the disinflation.

Of course, it was not anticipated that the initial announcement of the targets would, in and of itself, have a major effect on expectations. It is evident that words without actions are not particularly helpful. However, there was a distinct possibility that actions with words can have a stronger effect than actions without words. Moreover, even if the announcement of the targets did not have much direct effect on the achievement of the first milestone at the end of 1992, it was expected that with the achievement of the first target the credibility of the policy would be enhanced and expectations of inflation weakened such that the subsequent targets could be achieved more easily. And this may well have been the case.

In December 1993, on the occasion of the announcement of the appointment of the senior deputy governor of the Bank as the new governor, the Bank and the newly-elected government issued a joint statement on the objectives of monetary policy. In this statement, the government and the Bank recommitted themselves to price stability as the goal of monetary policy. They also agreed to extend the 1 to 3 percent band (which was the target for end-1995) for three more years through the end of 1998, with the decision on the definition of price stability put off until 1998. There were two reasons for the extension: (i) given that it has been a long time since Canada has had such low rates of inflation, it would be helpful to have more experience in operating under such conditions before an appropriate longer term objective is determined; and (ii) some time is needed to enable Canadians to adjust to the improved inflation outlook.

In the revised targets more emphasis is placed on the bands than on the mid-points. At the time of the initial announcement in 1991, the focus was on providing a downward path for inflation that was as simple and as credible as possible, in order to influence expectations. As a result, the midpoints of the bands were emphasized and the inflation-reduction path did not stretch too far into the future. With inflation having fallen to the bottom of the bands during the past couple of years and seeming likely to remain there in the near future, the emphasis in the revised targets was put on the inflation bands, thereby indicating that the Bank cannot control inflation all that closely.

The Bank has made it very clear that in achieving the inflation-control targets it is focussing on the underlying or trend rate of inflation, or what might be called the momentum of inflation. If it appears that the momentum of inflation is developing in such a way as to result in the trend rate moving outside the bands, the Bank will take action before inflation reaches the limit of the band. On the other hand, if there is a movement in the measured rate of inflation that is likely to be transitory and not affect the underlying rate, there is no need to respond to it.

In particular, although the 12-month rate of increase in the total CPI through much of 1994 was virtually zero, the Bank focused on the fact that the reduction in excise taxes on cigarettes in early 1994 accounted for a decline of about 1.3 percent in the total CPI. Operationally, therefore, the emphasis has been placed on the CPI excluding food, energy and the effect of indirect taxes, which has been posting a rate of increase between 1½ percent and 1¾ percent. At mid-1994, the date of the second milestone, the rate of increase of total CPI was at 0.0 percent while that of the CPI excluding food, energy and the effect of indirect taxes was at 1.8 percent, near the bottom of the band.

III.  Issues that Required Resolution in Developing a Targeting Framework

In establishing targets for inflation reduction, there were a large number of decisions that had to be made, some technical and some of a more substantive nature. In this section I examine a number of the issues that arose in the course of planning the framework for the targets.

(a) How fast should the decline of the target path for inflation be?

This issue is, of course, simply a variant of the long-standing debate in the economics literature on gradualism versus "cold shower" policies. In that literature, the existence of long-term contracts, lags in adjustment of behavior, and lags in the adjustment of inflation expectations all argue for more gradual disinflation. On the other hand, credibility notions and, in particular, the possibility that a more rapid disinflation could break entrenched expectations could argue for a more rapid disinflation.

In the consideration of various target paths for inflation reduction, similar factors came into play. The lags in the response of the Canadian rate of inflation to changes in monetary policy have traditionally been long, both as a result of institutional characteristics (such as the widespread use of two- and three-year contracts in labor markets) and expectational sluggishness. At the same time, the path had to depict a clear downward trend in inflation if there was to be any gain in credibility. In the event, it was decided to opt for a rather gradual path, in which the rate of inflation fell by 1 percentage point in total over 3 years after the initial 3 percent target at the end of 1992. The gradualism of the targeted pace of disinflation was related to the fact that we were moving down into ranges of price inflation not seen for almost three decades. There was therefore considerable uncertainty about how the economy would react and how feasible a more rapid pace of disinflation would be.

(b) What measure of inflation should be used?

The two principal contenders for this role were the CPI, or one of its variants, and the GDP deflator. The CPI was chosen, in large part because it is the most commonly used measure of inflation and it is the most relevant measure for most Canadians. The widespread acceptance and understanding of the CPI by Canadians was particularly important given the role of the targets as a means of communicating the aim of monetary policy. There were also more technical reasons for preferring the CPI: it is available on a monthly basis and hence can be tracked regularly; it is published in timely fashion without long delays; and it is almost never revised.

In order to emphasize trend movements in inflation and not short-term fluctuations, the targets focussed on the 12-month rate of increase of the CPI. Although the targets are specified in terms of the total CPI, the Bank in practice used the CPI measure which excludes food and energy as the basis for its policy actions since it is a better measure of underlying inflation. Since, over time, the CPI and the CPI excluding food and energy have tended to move together, achieving the target path for the latter would also achieve it for the former. Nonetheless, if there were a divergence in the trend movements of the two series, it is the overall CPI that would provide the basis for targeting over the longer term.

(c) What is the response to changes in indirect taxes?

In designing the targets, an approach to large once-off price changes arising from increases in indirect taxes had to be developed. It was decided that the first-round or direct effects on the CPI resulting from large and unexpected increases in indirect taxes would be accommodated, but not their second-round effects. Thus, "tax-induced drift" from unexpected changes in indirect taxes would be permitted in the price level but not in the ongoing rate of inflation. Increased emphasis has been put on the CPI excluding the volatile food and energy components and adjusting for sharp changes in indirect taxes as a way of making operational this approach. However, since the approach was adopted to deal with unusually large or unpredictable tax changes, it was decided (consistent with the treatment of trends in food and energy prices) that if indirect tax increases came to occur with such frequency and predictability that they constituted a trend, that trend would not be permitted to impinge on the targets.

(d) What should be done in response to very large price shocks?

Certain major price shocks would justify a change to the entire target path but this should be done only in very unusual circumstances. For example, a very large increase in oil prices that could not be prevented from spilling over into a broad range of other prices or a widespread natural disaster would be situations that might call for a reconsideration of the entire target path to see whether it remained feasible, and perhaps for the establishment of a revised target path to price stability. This provision is not intended to deal with the more typical demand and supply shocks that buffet the economy.

(e) The relationship between the targets and monetary policy

The way in which the target path was framed and explained is importantly related to the nature of monetary policy in a number of ways.

(i) Choice of the first target date

Monetary policy, as is well known, operates with long and variable lags. Hence, an attempt to achieve targets for the rate of inflation (or, equivalently, the price level) in the very near term would be either impossible or totally inappropriate. Indeed, earlier work at the Bank of Canada had ruled out the near-term price level or rate of inflation as a target for monetary policy because it would result in large swings in the economy as well as potential instrument instability. It seemed appropriate to set the first target almost two years away, at the end of 1992, and to focus on rates on inflation in the future, not the present, in taking policy actions.

(ii) The use of bands

The difficulty in forecasting inflation and the lack of precision in predicting the effects of monetary policy actions on inflation also suggested that bands be set around the targets. Thus, bands of plus and minus 1 percent were established around the target path for inflation. These ranges were in fact smaller than called for by empirical work done at the Bank.

There is a trade-off in such circumstances between the probability of successful achievement of the targets and the usefulness of the targets as a communications device designed to influence expectations. The wider are the bands, the higher is the probability of successful achievement of the targets but the less useful are the targets in changing behavior. In the event, it was decided to use somewhat narrower bands to avoid the problem that overly wide bands might leave the impression that the authorities were not serious about bringing inflation down.

(iii) Deviations from the target

Unexpected price shocks, for example a substantial increase in the prices of raw materials, can result in temporary deviations from the target band. Although monetary policy actions cannot be expected to reverse such effects in the very short run, actions can be taken to put in train adjustments needed to return the trend rate of inflation to the target path. If such shocks were large enough or occurred close to the target date, they might push the rate of inflation above the band at the announced target date. In such circumstances, policy actions would have to be directed to ensuring that the target rate of inflation announced for the following period was met. In other words, along the path to price stability base drift would be accepted for the price level but not for the rate of inflation.

The target path for inflation reduction does not in itself give direct guidance on how the instrument of monetary policy should be set. In making judgments on the necessary actions to achieve the target path, it is helpful to monitor a variety of indicators that stand between the policy instrument and the rate of inflation. The Bank of Canada has focussed closely on estimates of excess demand or supply (or "gaps") in goods and labor markets as key inputs into the inflationary process. It also follows closely such variables as the rate of expansion of money (especially the broader aggregate M2+, which is the best financial indicator of future rates of inflation), the growth of credit, the rate of increase of total spending and wage settlements as guides to policy action.²

IV.  Some Tentative Conclusions

Although it is too early to draw definitive conclusions from the Canadian experience, there are a number of points worth noting on the basis of the Canadian experience and that of other countries.

(i) It would be best if the targets follow from a legislative mandate to achieve price stability. Failing that, it is useful to have a formal commitment by the government to the targets.

(ii) The targets were expected to aid the credibility of the price stability policy, especially after the first target was achieved. The importance of achieving the first target suggests that such targets should not be introduced at a time of upward pressure on underlying inflation, but rather at a time when there is a realistic chance of getting inflation down. Furthermore, one should fix a downward path for the targets that is not so steep that it appears impossible to achieve, but sufficiently steep that it shows a clear downward trend for inflation.

(iii) The usefulness of the bands was apparent in the outcome of the first and second milestones. With the unexpected sluggishness of the economy, the rate of inflation fell faster and further than initially anticipated, and this in spite of the fact that monetary conditions were easing for most of the period. The focus on the CPI excluding food, energy and the effect of indirect taxes as an operational target has been especially helpful in dealing with the indirect tax shocks to the system.

(iv) Inflation-reduction targets aimed at bringing about disinflation can be thought of as allowing for a form of automatic stabilization in response to demand shocks and hence for partial stabilization of output fluctuations. Thus, when inflation is above the targets monetary conditions will tighten, and the demand pressures will thereby be mitigated. Conversely, in a weak economic situation with inflation declining in line with the targets there is scope for easing in monetary conditions and hence providing support for the economy. Thus, as inflation has declined in Canada, monetary conditions have eased, with both interest rates and the external value of the Canadian dollar down significantly from their peaks. And with wage and price inflation under control, the nominal depreciation has been translated into a real depreciation. Similarly, the inflation-control targets, which aim at maintaining a very low rate of inflation by establishing a 1 to 3 percent band from 1995 to 1998, have the same scope for automatic stabilization as do the inflation-reduction targets. When demand pressures push the rate of inflation toward the top of the band, monetary conditions will be tightened. Conversely, with a weak economic situation inflation will decline below the bottom of the band, leading to an easing of monetary conditions. The latter will provide support to aggregate demand.

(v) By allowing major price movements resulting from supply shocks to affect the price level but not the rate of inflation, the operation of the targets permits an appropriate monetary policy response to supply shocks.

(vi) The targets have played a useful role in communicating and making more concrete the Bank's policy of moving to price stability. The publicity attached to the announcements of both the original and revised targets was very helpful in this regard.

(vii) Communication of the targets and of inflationary developments in relation to them is an essential element in achieving credibility and in being held accountable. A number of countries, such as New Zealand, the United Kingdom and Sweden, issue periodic reports on inflationary developments with respect to the targets. And the Bank of Canada will be introducing a semi-annual monetary policy report beginning in the spring of 1995. These reports, which have both backward-looking and forward-looking perspectives, have received considerable attention and careful scrutiny by the press, the financial markets, and parliamentary committees. It may well be that their most important contribution will be to signal prospective inflationary pressures and the need for timely policy action, at a time when actual rates of inflation (which are of course a lagging indicator) are still relatively subdued.

(viii) Although expectations of inflation remain very subdued in Canada there appears to remain some lingering uncertainty about the longer term. This shows up in long-term bond yields which, although down appreciably from their peaks, remain high relative both to recent rates of Canadian inflation and to their U.S. counterparts. This has been a disappointment for the Canadian authorities. These relatively high rates are partly attributable to fiscal problems in Canada as well as to political uncertainties.

In assessing 3½ years of inflation reduction and control targets, it is clear that they have not painlessly shifted inflation expectations. But they have given a concreteness to the Bank's objectives and they have given economic agents a clear basis for making their plans. And with monetary policy having to operate in a forward-looking fashion, the targets provide a justification for taking timely action. The targets also keep the monetary policy focus firmly on the nominal side of the economy and ensure that the Bank is held accountable for what it can deliver rather than being held responsible for all problems on the real side of the economy.

I would also note that although inflation-reduction targets are neither necessary nor sufficient to achieve price stability, it appears on the basis of the Canadian experience thus far that they can be very helpful in the attainment of this goal. As a form of pre-commitment by the authorities to the reduction of inflation, they can perhaps be considered as a way of dealing with the time-inconsistency problem. A comparison of inflationary developments in countries which use inflation-control targets and those in countries without such targets will provide interesting material for future researchers about the role of formally announced targets in affecting expectations and helping to achieve price stability. However, it will be necessary to await the inflation results over a whole business cycle before drawing firmer conclusions about the role of the targets on the basis of the experience of these countries. In the short run, it will be interesting to see the extent to which targets provide a useful internal discipline for central banks, facilitating timely action in response to the buildup of inflationary pressures.