This section summarizes the operational practices of inflation targeting central banks. In most respects, emerging market countries and industrial countries operate monetary policy similarly under inflation targeting. In all cases, countries need to use a good dose of judgment. However, countries with more developed statistical models can rely relatively more on those models than can countries with less developed statistical models, owing to data shortfalls and ongoing structural changes. In addition, emerging market countries must deal with more pronounced economic and financial shocks and less developed financial markets.

In an inflation targeting framework, the modalities of the operation of monetary policy follow from the lag between a policy change and its impact on inflation. These lags have led to the description of monetary policy under inflation targeting as “constrained discretion,” which stands in contrast to the “automatic pilot” approach seen to hold under strict exchange rate or monetary aggregate targeting (Bernanke and others, 1999). This extra degree of discretion lends added importance to the operation of monetary policy.

Monetary policy under inflation targeting works as follows:

• The inflation forecast is usually updated on a regular schedule and based on the latest economic data, indicators of market sentiments, model results, and judgment.

• The desired policy stance—in terms of the level of the operating target required to narrow any gap between the inflation forecast and the inflation target, and taking into account the output gap and other factors—is determined using models and judgment.

• Other factors—such as international developments and political events—that could bear on the timing and magnitude of changes in the policy stance are also considered.

• The change in policy stance is announced and the operating target is adjusted to its desired level.

Inflation Forecasting

Inflation forecasting is especially challenging for emerging market countries. Inflation forecasts play a key role in the conduct of policy under inflation targeting because of the lags between monetary actions and their impact on inflation. Indeed, the inflation forecast can be viewed as the intermediate target of policy (Svensson, 1997; Bogdanski and others, 2000). Emerging market countries rely less on statistical markets to forecast inflation owing to data shortfalls, ongoing structural changes, and their vulnerability to shocks.

Many inflation indicators are common to all inflation targeting countries. An inflation indicator refers to observable data that have been previously shown to provide useful signals regarding future changes in inflation. Indicators include aggregate demand and supply variables, monetary aggregates, interest rate and exchange rate measures, inflation, price measures, and expectations.

Surveys of inflation expectations conducted by the private sector or even by the central bank itself are useful in the inflation targeting regime. Surveys can directly sample market and consumer inflation expectations (New Zealand), or the central bank can summarize surveys of market inflation expectations (United Kingdom), Qualitative surveys of private sector conditions are utilized in Canada and New Zealand. In addition, informal canvassing of market participants by central bank staff can provide useful information.

Yields on government bonds indexed to inflation can be a useful indicator (Price, 1997). Movements in the spread between the yields on indexed and non-indexed bonds have been used by the central banks of Australia, Canada, Israel, New Zealand, and the United Kingdom to gauge the impact of policy changes and other developments on inflation expectations and/or the inflation risk premium.¹⁵ These data should be used with care, however, because the market for indexed debt is thin and there may be large and variable risk premiums. Moreover, in some countries technical issues, such as tax treatment of returns on indexed debt compared with nominal debt, can complicate the extraction of information on inflation expectations. Expected future inflation rates can also be derived from nominal and real forward interest rates (Söderlind and Svensson, 1997).

For emerging market countries, aggregate demand and supply indicators have greater weight. The greater weight of these indicators reflects the importance of real shocks and the lack of alternative market-based indicators. Supply shocks seem to be especially important for middle-income emerging market countries (Agénor and others, 1999).

Emerging market countries can benefit from the introduction or refinement of inflation indicators. Several central banks have worked with statistical agencies to define core measures of inflation that provide information on trend movements of inflation. Central banks can also identify or develop new inflation indicators based on market conditions.

The structural models used by central banks to forecast inflation have important common elements. These include an open economy demand curve, a Phillips curve, an international asset market equilibrium condition, and a monetary policy reaction function (Bogdanski and others, 2000; Clinton, 2000). These elements have been used extensively in empirical work on central bank behavior, monetary policy rules, and sacrifice ratios (Taylor, 1999). Also, they lie at the heart of some of the new large macro-econometric models developed in central banks, including the Bank of Canada’s Quarterly Projections Model (QPM) and the U.S. Federal Reserve Board’s FRB/U.S. model.

Large-scale macroeconomic models are used by several industrial countries. These models can serve as useful guides to policy for countries that have had inflation targeting in place for some time and have stable economic relationships. Canada and New Zealand use large dynamic models to produce forecasts (Drew and Hunt, 1998). These models are calibrated to provide theoretically plausible simulation results because the parameters generated by estimation on the basis of a good fit to historical data proved to be less useful for conducting policy simulations. Development of large-scale models has taken many years, notwithstanding the relatively long period of time that inflation targeting has been in place. Most industrial countries also utilize small structural models that capture a particular sector or key variable to assess developments in key sectors (Drew and Hunt, 1998).

Emerging market countries, in contrast, rely less on quantitative models. This is because the years of experience with inflation targeting are too few to generate reliable estimates of such a model, and because of ongoing changes in structural relationships (Leone, 1999; Leiderman and Bar-Or, 2000). However, most emerging market inflation targeting central banks have found it worthwhile to devote extra resources to development of such models for greater future use. Brazil, the Czech Republic, and Israel all work with three or four equation models along the lines of those commonly used by industrial country central banks (Bogdanski and others, 2000; Clinton, 2000).

Models can facilitate monetary policy in important ways other than by producing inflation forecasts (Isard and Laxton, 2000a). Small-scale models can help central banks think through policy transmission channels. More important, models can serve as a framework for policy discussion and can even assist in the presentation of the inflation forecasts that are so essential to transparency. Fan charts, showing the probability distribution of inflation forecasts over a range of outcomes, are a recent improvement in presentation that is facilitated by quantitative models (Allen, 1999).

Time series models impose less economic structure but provide better short-term forecasts, as well as a consistency check for the larger structural models. Univariate models of the aggregate inflation index, components of the index, or other key inflation indicators are relatively easy to produce and allow for frequent assessment of economic conditions. In New Zealand, output from these models is used to provide forecasts for the first two quarters beyond the most recent historical data. Multivariate time series models (e.g., vector autoregression) allow for the simultaneous interaction between time series with few, if any, structural assumptions. Such models are also used by the central banks of Brazil, Chile, and Israel.

In practice, inflation forecasts are based on a combination of indicator variables, quantitative economic models, and qualitative judgment. Inflation targeting central banks typically first analyze the output from their economic models and other assorted inflation indicators, and then apply judgment based on qualitative information gathered from their contacts in all sectors of the economy to arrive at a view regarding the appropriate stance for monetary policy.

Policy Transmission Channels

The policy channels between changes in the monetary stance and inflation for open economies generally operate through direct and indirect channels (Svensson, 1998; Ball, 1999). In the short run, an increase in interest rates directly reduces inflation by appreciating the exchange rate, which passes through into lower imported goods prices. Over the longer term, an increase in interest rates indirectly reduces inflation first by dampening expectations and thereby business investment, and second, by reducing consumption and thereby causing an exchange rate appreciation that switches expenditures from traded to non-traded goods. Of course, the central bank cannot count on the exchange rate to transmit a policy change to a prespecified change in inflation, even if over time the exchange rate has been the main short-run policy channel. The markets will determine whether a given change of policy will influence inflation primarily via the exchange rate or through market interest rates. In addition, monetary policy may influence inflation indirectly by altering wealth and thereby aggregate demand.

Emerging market countries with higher inflation rates seem to have channels characterized by downward price stickiness and rapid pass-through from the exchange rate to inflation (Table 5.1). In Brazil, Chile, and Israel, long experiences with high rates of inflation and indexation often led to downward price inertia and almost contemporaneous pass-through from exchange rate changes to inflation until the inflation targeting framework became credible (Bogdanski and others, 2000; Landerretche and others, 1999; Leiderman and Bar-Or, 2000; Leone, 1999). Exchange rate pass-through seems to be slower for industrial countries—that is, around six to eight quarters (Menon, 1995). For New Zealand, the direct exchange rate channel, which operates in the short term, is seen as having been muted by exchange rate hedging and by a lengthening of the lags between policy and inflation. For Brazil, the indirect channel of monetary policy through demand is seen to take two or three quarters to affect inflation.

Table 5.1.

Important Features of Monetary Policy Transmission Channels

Sources: Andrés and others (1998), Borio (1997), Corbo (2000), de Fiore (1998), Haldane (2000), Leone (1999), Landerretche and others (1999), Leiderman and Bar-Or (2000), Jonsson (1999), Van’t dack (1999).

Table 5.1.

Important Features of Monetary Policy Transmission Channels

		Important Features
Emerging Market Countries
	Brazil	Interest rate affects inflation with minimum lag of six months with rapid exchange rate pass-through.
	Chile	Indexation has led to downward price inertia and quick transmission (approximately three quarters) from exchange rate and wage shocks to inflation.
	Czech Republic	The transmission mechanism has been weakened by the fragility of the financial sector.
	Israel	Widespread indexation has led to relatively quick pass-through from exchange rate to prices. More recently, pass-through has lengthened.
	Poland	Credit channel is weak due to the structure of the banking sector and the relatively undeveloped financial markets.
	South Africa	Interest rate changes affect inflation with a lag and linkages between broad money and inflation are weak.
Industrial Countries
	Canada	The transmission mechanism is well developed; lags typically range from six to eight quarters, but they have varied somewhat over time.
	Finland	The move to flexible exchange rates has weakened the transmission mechanism and has increased exchange rate variability.
	New Zealand	The transmission mechanism is well developed: lags typically range from six to eight quarters, but they have varied somewhat over time.
	Spain	Variables in the real exchange rate appear to be the most important of the monetary transmission channels.
	Sweden	The transmission mechanism is well developed; lags typically range from five to eight quarters.
	United Kingdom	Monetary policy has its maximum effect on output after around a year, and on inflation after around two years.

Sources: Andrés and others (1998), Borio (1997), Corbo (2000), de Fiore (1998), Haldane (2000), Leone (1999), Landerretche and others (1999), Leiderman and Bar-Or (2000), Jonsson (1999), Van’t dack (1999).

View Table

Table 5.1.

Important Features of Monetary Policy Transmission Channels

		Important Features
Emerging Market Countries
	Brazil	Interest rate affects inflation with minimum lag of six months with rapid exchange rate pass-through.
	Chile	Indexation has led to downward price inertia and quick transmission (approximately three quarters) from exchange rate and wage shocks to inflation.
	Czech Republic	The transmission mechanism has been weakened by the fragility of the financial sector.
	Israel	Widespread indexation has led to relatively quick pass-through from exchange rate to prices. More recently, pass-through has lengthened.
	Poland	Credit channel is weak due to the structure of the banking sector and the relatively undeveloped financial markets.
	South Africa	Interest rate changes affect inflation with a lag and linkages between broad money and inflation are weak.
Industrial Countries
	Canada	The transmission mechanism is well developed; lags typically range from six to eight quarters, but they have varied somewhat over time.
	Finland	The move to flexible exchange rates has weakened the transmission mechanism and has increased exchange rate variability.
	New Zealand	The transmission mechanism is well developed: lags typically range from six to eight quarters, but they have varied somewhat over time.
	Spain	Variables in the real exchange rate appear to be the most important of the monetary transmission channels.
	Sweden	The transmission mechanism is well developed; lags typically range from five to eight quarters.
	United Kingdom	Monetary policy has its maximum effect on output after around a year, and on inflation after around two years.

Sources: Andrés and others (1998), Borio (1997), Corbo (2000), de Fiore (1998), Haldane (2000), Leone (1999), Landerretche and others (1999), Leiderman and Bar-Or (2000), Jonsson (1999), Van’t dack (1999).

Policy channels can be rendered less effective by corporate and bank balance sheet problems and underdeveloped financial markets. Policy channels in the Czech Republic have been stifled by the highly interest inelastic lending policies of banks burdened by nonperforming loans and borrowing practices of unprofitable enterprises. In Poland, the credit channels of policy have been weakened by undeveloped financial markets and strong capital inflows.

Policy Implementation

All inflation targeting countries use a very short-term interest rate as the operating target. The operating target can be defined as the money market indicator that best captures the authorities’ intentions (Table 5.2). The popularity of an overnight interest rate operating target implies that central banks have chosen to broadly accommodate fluctuations in the demand for bank reserves or settlement balances. The growing popularity of short-term interest rates as an operating target is part of the larger trend away from money stock targets (Borio, 1997; Van’t dack, 1999) and need not be inherent to the inflation targeting framework.

Table 5.2.

Operating Targets and Main Instruments of Monetary Policy

Sources: Central bank websites Borio (1997)Van’t dack (1999), and Landerretche and others (1999).

Table 5.2.

Operating Targets and Main Instruments of Monetary Policy

		Operating Target	Main Instruments
Emerging Market Countries
	Brazil	Overnight interbank interest rate (SELIC)	Open market operations (OMOs) using Treasury bonds or bonds issued by the central bank
	Chile	Real overnight interbank rate tied to CPI with 20-day lag	OMOs through issuance of central bank paper and repurchase agreements (repos)
	Czech Republic	Two-week repo rate	Daily two-week repo auctions
	Israel	Interest rate on short-term loans to and deposits with banks	Fixed-term and daily discount-window loans to banks, and OMOs with Treasury bills
	Poland	28–day National Bank of Poland bill rate	OMOs with central bank bills
	South Africa	Overnight repo rate	OMOs with government securities
Industrial Countries
	Australia	Overnight interest rate	OMOs with repos and out left transactions in government securities
	Canada	Overnight interest rate	Operating band enforced through standing facilities
	Finland	Monthly repos and outleft OMOs	OMOs with government securities
	New Zealand	Overnight cash rate	Lending or borrowing of overnight money to commercial banks
	Spain	Money market overnight interest rate	OMOs with government securities and central bank paper
	Sweden	Weekly repo rate	OMOs with central bank paper
	United Kingdom	Short-term repo interest rate	OMOs with repos, government securities, and other eligible local authority and bank bills

Sources: Central bank websites Borio (1997)Van’t dack (1999), and Landerretche and others (1999).

View Table

Table 5.2.

Operating Targets and Main Instruments of Monetary Policy

		Operating Target	Main Instruments
Emerging Market Countries
	Brazil	Overnight interbank interest rate (SELIC)	Open market operations (OMOs) using Treasury bonds or bonds issued by the central bank
	Chile	Real overnight interbank rate tied to CPI with 20-day lag	OMOs through issuance of central bank paper and repurchase agreements (repos)
	Czech Republic	Two-week repo rate	Daily two-week repo auctions
	Israel	Interest rate on short-term loans to and deposits with banks	Fixed-term and daily discount-window loans to banks, and OMOs with Treasury bills
	Poland	28–day National Bank of Poland bill rate	OMOs with central bank bills
	South Africa	Overnight repo rate	OMOs with government securities
Industrial Countries
	Australia	Overnight interest rate	OMOs with repos and out left transactions in government securities
	Canada	Overnight interest rate	Operating band enforced through standing facilities
	Finland	Monthly repos and outleft OMOs	OMOs with government securities
	New Zealand	Overnight cash rate	Lending or borrowing of overnight money to commercial banks
	Spain	Money market overnight interest rate	OMOs with government securities and central bank paper
	Sweden	Weekly repo rate	OMOs with central bank paper
	United Kingdom	Short-term repo interest rate	OMOs with repos, government securities, and other eligible local authority and bank bills

Sources: Central bank websites Borio (1997)Van’t dack (1999), and Landerretche and others (1999).

Some open economy central banks have used an average of an interest rate and the exchange rate (a monetary conditions index) as an operating target or as a way to facilitate the signaling of policy intentions. Monetary conditions indices (MCI) have been used because in a small open economy it can be difficult to predict whether a policy adjustment will influence aggregate demand and inflation through the exchange rate or interest rates. The monetary conditions index is a weighted combination of the exchange rate and short-term interest rate less their values in a base period and can be calculated in nominal or real terms. The weights are determined by the relative influence of the exchange rate and benchmark interest rate for real aggregate demand. In practice, the usefulness of the monetary conditions index as a means of publicly communicating the central bank’s monetary policy stance has been complicated by the expectation by some market players that all exchange rate shocks (even on a daily frequency) might be offset by interest rate adjustments to keep the index fixed, and by the fact that different types of exchange rate shocks can have different implications for the desired path of the index (Freedman, 1994; Hunt, 1999). Canada was the first to use such an index in the late 1980s, but reduced its role as a signaling device in the spring of 1998. New Zealand reported a projected path and band width of a monetary conditions index, but shifted to an interest rate operating target in mid-1998 to smooth interest rate volatility, which was complicating communication of the desired policy stance.

All of the inflation targeting central banks employ market-based instruments of monetary policy to maintain the operating target interest rate at the desired level. Open market operations (OMOs), conducted on an outright or repo basis, are the prevalent monetary instrument for central banks that operate under the inflation targeting framework (Table 5.2). Most central banks use open market operations to maintain domestic liquidity conditions in line with the announced operating interest rate. These open market operations take the form of central bank purchases and sales of Treasury or central bank securities, usually with a commercial bank counterparty.

Inflation targeting central banks appear to rarely use direct instruments for active liquidity management, if at all. Direct instruments of monetary policy, such as reserve requirements, loan limits, and liquidity asset ratio requirements, can only be adjusted infrequently. This aspect of direct instruments implies that they may not be effective in offsetting inflationary shocks and in signaling the policy intentions of the central bank. In contrast, market-based instruments can be adjusted quickly in response to inflationary shocks, and are more easily fine-tuned to achieve the desired policy stance and signal changes in the policy outlook (Alexander and others, 1995). These considerations suggest that market-based instruments are extremely useful for implementing monetary policy under inflation targeting.

Most inflation targeting countries adjust short-term interest rates in response to deviations of inflation or expected inflation from the target and the output gap. In general, central bank interest rate setting behavior has been characterized in many studies as a Taylor rule under which interest rates are adjusted in response to deviations of current and lagged variables such as inflation and the output gap (Borio, 1997; Taylor, 1999). Under inflation targeting, interest rate response functions are more forward-looking than under other policy frameworks, owing to the lags between policy changes and the inflation outcome and the desire to avoid excessive output volatility (Isard and others, 2000; Armour and Côté, 2000). In Brazil, the central bank employs three interest rate response functions in its model simulations; an exogenous interest rate path, a Taylor rule with linear combination of system variables, and interest rate paths generated by optimal response functions.

In practice, central banks complement such an approach with judgment that uses all information pertaining to the inflation forecast. Emerging market central banks often must use more judgment because their models are subject to uncertainties. Information on many variables is available only with a lag and could be subject to revision (Orphanides, 1998). Shocks may be larger and more persistent than for industrial countries (Agenor and others, 1999), and their models are subject to parameter instability (Brainard, 1967).

Inflation Targeting and IMF Conditionality

The prospect that countries with IMF-supported programs may adopt inflation targeting has prompted a reconsideration of monetary policy conditionality in those countries. The traditional monetary policy conditions for IMF programs are a floor on the central bank’s net international reserves (NIR) and a ceiling on the central bank’s net domestic assets (NDA), or in some cases on base money (Schadler and others, 1995). Under flexible exchange rates, a floor on net international reserves serves to limit foreign exchange market intervention, while a ceiling on net domestic assets or base money is intended both as a check against either the sterilization of foreign exchange market intervention or excessive monetary expansion that would jeopardize price stability. Given the uncertainty that characterizes all econometric relationships, a breach of the net international reserves floor or the net domestic assets ceiling does not necessarily signal an overexpansionary monetary policy stance: instead, it is a signal that the monetary policy stance may need to be reassessed. Such a reassessment—in the context of considering a waiver for a breach of a performance criterion—must take account of the full range of factors that are relevant to achieving the program’s objective (Mussa and Savastano, 1999).

If the traditional configuration of performance criteria is used in a country following inflation targeting, it is important to distinguish between the conduct of monetary policy and program monitoring. Central banks typically steer a short-term interest rate as the operating target rather than net domestic assets and pay attention to a range of inflation indicators for their monetary policy decisions. The floor on net international reserves and the ceiling on net domestic assets provide a check that these do not conflict with program objectives. But given the greater transparency associated with inflation targeting, together with the IMF’s increasing transparency, there is a risk of confusion between the inflation objective that guides monetary policy and the traditional performance criteria used to monitor the IMF-supported program.

The IMF has, on an experimental basis, adopted a reviews-based approach to monetary conditionally under inflation targeting as an alternative to the traditional configuration of performance criteria. Under this approach, monetary policy would be subject to quarterly reviews focusing on indicators of inflation prospects, including recent inflation outturns and various leading indicators of inflation. A floor for net international reserves would also be retained under this approach to limit the use of the IMF’s resources for foreign exchange market intervention.

A variant of the reviews-based approach is currently in use in Brazil, the only inflation targeting country supported by an IMF arrangement. A “consultation mechanism for inflation” is included as a condition for monetary policy, as well as a floor on net international reserves and an indicative ceiling on net domestic assets (Blejer and others, 2000). Based on the government’s announced annual inflation targets, quarterly bands of ±1 percentage point for the inner band and ±2 percentage points for the outer band have been established. Should the inflation rate exceed the upper limit of the inner band, the central bank will discuss the appropriate policy response with the IMF. In case of a breach of the upper limit of the outer band, the authorities will complete consultation with the IMF’s Executive Board on the proposed policy response before requesting further financing under the program.

Changing Economic Relationships

Structural changes seem minor in the early stages of inflation targeting, but as the framework gains credibility, empirical evidence suggests that the linkages between inflation and the level of economic activity weaken. The experiences of both emerging market and industrial inflation targeting countries suggest that a new inflation targeting framework gains credibility gradually as the central bank demonstrates its ability to achieve the targets (Johnson, 1998). Moreover, the results of (Bernanke and others, 1999), (Fillion and Leéonard, 1997), (Landerretche and others, 1999), and Leiderman and Bar-Or (2000) indicate that inflation becomes less responsive to the level of economic output and the exchange rate following the introduction of the targets.¹⁶ This also suggests that while expectations may respond quickly in a high-inflation environment to monetary policy changes, the transmission channels appear to become more complex and indirect when inflation is low. Similar conclusions can also be drawn for the relationship between monetary aggregates and inflation.

As the links between inflation and output weaken, the central bank may want to take an “eclectic” approach to formulating its outlook for inflation, and in deciding on appropriate monetary actions. Such an approach may help uncover, through experience, the rules of thumb appropriate for the inflation targeting framework (Bertocchi and Spagat, 1993). This suggests that following the establishment of credibility it might be prudent to monitor a broad range of indicators of inflation with the aim of looking for consistent signals across indicators. This would be particularly relevant in periods when the economy is operating near potential and authorities cannot place much confidence in traditional measures of the output gap to guide monetary policy.

The growing credibility of the inflation targeting framework also appears to be associated with a weaker relationship between the exchange rate and inflation. Domestic prices and inflation expectations are now less sensitive to exchange rate movements in Chile, for example, which has contributed to longer lags in its monetary policy transmission mechanism.

Breaches of the Inflation Target

Communicating ahead of time on how the central bank will respond to breaches of the target may reduce uncertainty. Clifton (1999) argues that inflation targeting central banks must communicate in advance how they will treat the boundary points in order to minimize uncertainty with respect to the conduct of monetary policy and contribute to the credibility and transparency of the monetary policy framework.

Target Floor

Central banks have eased monetary policy in response to bottom breaches only after it has become clear that the banks’ outlook for future inflation was signaling that inflation would stay below target. In Sweden, central bank officials were surprised in the latter half of 1996 and early 1997 by the extent of disinflationary pressures prevailing in the economy. These were mainly attributed to the one-time effect of a reduction in value-added taxes, and to declining mortgage rates that were judged as not lowering underlying inflation expectations. It became evident that future inflation would move below the target range, however (Bernanke and others, 1999). The authorities did not wish to risk engaging in instrument instability for the sake of formally hitting a target. Instead, they placed more emphasis in public communications on trends in “core” inflation rates and their own internal forecasts of inflation, even though the targets were defined in terms of the headline consumer price inflation rate.

A consensus seems to exist in favor of symmetric responses to breaches of the floor and ceiling of the inflation target range that arise from demand shocks in order to limit variability in employment and output. Given that deflation can persist, breaches of the floor warrant a decisive response to prevent costly employment and output losses. Equally, breaches of the ceiling require a strong response to maintain credibility, especially for emerging market countries. Moreover, given the lags between monetary policy actions and their effects on inflation, central banks should not—and typically do not—wait for breaches to occur before taking action. Rather, acting once it becomes clear that a breach is likely may help keep the number of actual breaches to a minimum, with consequent payoffs to the credibility of monetary policy.

Countries that are disinflating have been prepared to accept an inflation outcome below the tolerance zone, however. Table 5.3 and Figures 5.1 and 5.2 summarize the number of occasions inflation targeting central banks have seen inflation move away from the target in the period where the inflation targets were interim targets on a path to a long-run inflation objective. Canada, the Czech Republic, Israel, and New Zealand did not hesitate to take advantage of unexpected opportunities to lock-in disinflationary outcomes—either by allowing inflation to undershoot until the target range caught up to the decline in the inflation rate (Canada and New Zealand), or by ratifying the decline in the inflation rate with lower interim targets (Israel). The four countries seized these opportunities to achieve the long-run inflation objective more quickly than originally anticipated—that is, they were “opportunistic disinflators.” In a disinflationary phase, such an approach can be a useful means of building credibility for the new inflation targeting framework because “over-achievement” of the interim targets can signal the central bank’s determination to bring inflation under control.

Table 5.3.

Quarterly Breaches of Inflation Targets

Sources: Central bank inflation reports; Bernanke and others (1999)

¹Excludes the United Kingdom, which operates with a point target.

Table 5.3.

Quarterly Breaches of Inflation Targets

		Disinflation Period		Long-Run Target Period
		Above Target	Below Target	Above Target	Below Target
		(Number of Quarters)
Industrial Countries1
	Australia	…	…	0	0
	Canada	0	8	0	0
	Finland	…	…	0	0
	New Zealand	0	4	5	0
	Spain	0	0	0	0
	Sweden	…	…	0	12
Emerging Markets
	Brazil	0	0	…	…
	Chile	0	0	…	…
	Czech Republic	2	4	…	…
	Israel	4	2	…	…
	Poland	2	0	…	…

Sources: Central bank inflation reports; Bernanke and others (1999)

¹Excludes the United Kingdom, which operates with a point target.

View Table

Table 5.3.

Quarterly Breaches of Inflation Targets

		Disinflation Period		Long-Run Target Period
		Above Target	Below Target	Above Target	Below Target
		(Number of Quarters)
Industrial Countries1
	Australia	…	…	0	0
	Canada	0	8	0	0
	Finland	…	…	0	0
	New Zealand	0	4	5	0
	Spain	0	0	0	0
	Sweden	…	…	0	12
Emerging Markets
	Brazil	0	0	…	…
	Chile	0	0	…	…
	Czech Republic	2	4	…	…
	Israel	4	2	…	…
	Poland	2	0	…	…

Sources: Central bank inflation reports; Bernanke and others (1999)

¹Excludes the United Kingdom, which operates with a point target.

Industrial Countries: Annual Inflation Rates and Inflation Targets

(In percent)

Sources: IMF, International Financial Statistics; national authorities.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Industrial Countries: Annual Inflation Rates and Inflation Targets

(In percent)

Sources: IMF, International Financial Statistics; national authorities.

• Download Figure
• Download figure as PowerPoint slide

Industrial Countries: Annual Inflation Rates and Inflation Targets

(In percent)

Sources: IMF, International Financial Statistics; national authorities.

• Download Figure
• Download figure as PowerPoint slide

Emerging Market Countries: Annual Inflation Rates and Inflation Targets¹

(In percent)

Sources: IMF, International Financial Statistics; national authorities.¹ Vertical lines indicate start date and end date of transition periods toward adoption of full-fledged inflation targeting framework (see Table 4.1 of text).² During the first half of the 1990s, inflation peaked at 6,038 percent in the second quarter of 1990.³ During the early 1990s, inflation peaked at 1,135 percent in the first quarter of 1990.⁴ During the first half of the 1990s, inflation peaked at 425 percent in 1991.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Emerging Market Countries: Annual Inflation Rates and Inflation Targets¹

(In percent)

Sources: IMF, International Financial Statistics; national authorities.¹ Vertical lines indicate start date and end date of transition periods toward adoption of full-fledged inflation targeting framework (see Table 4.1 of text).² During the first half of the 1990s, inflation peaked at 6,038 percent in the second quarter of 1990.³ During the early 1990s, inflation peaked at 1,135 percent in the first quarter of 1990.⁴ During the first half of the 1990s, inflation peaked at 425 percent in 1991.

• Download Figure
• Download figure as PowerPoint slide

Emerging Market Countries: Annual Inflation Rates and Inflation Targets¹

(In percent)

Sources: IMF, International Financial Statistics; national authorities.¹ Vertical lines indicate start date and end date of transition periods toward adoption of full-fledged inflation targeting framework (see Table 4.1 of text).² During the first half of the 1990s, inflation peaked at 6,038 percent in the second quarter of 1990.³ During the early 1990s, inflation peaked at 1,135 percent in the first quarter of 1990.⁴ During the first half of the 1990s, inflation peaked at 425 percent in 1991.

• Download Figure
• Download figure as PowerPoint slide

Other disinflating countries have treated breaches symmetrically. The Bank of Israel opted in early 1999 to partially reverse the post-Russia crisis lightening of its policy stance to limit uncertainties regarding the objectives of policy, and offset short-run losses in employment and real output. Following the cuts, however, the discount rate was still high by both Israeli historical and international standards—at 12 percent in April 1999—and the inflation target for 1999 was significantly undershot.

Breaches of the Target Ceiling

Breaches of the target ceiling may not seriously damage the inflation targeting framework provided transparency and a firm policy response maintain the public’s confidence. Inflation moved above the target range in New Zealand in the mid-1990s when the Reserve Bank of New Zealand underestimated the strength of the economy and the associated inflation pressures (Bernanke and others, 1999). The Reserve Bank responded with a significant tightening of its policy stance, resulting in sharply higher interest rates and a major firming of the exchange rate. Several years of higher interest rates followed, as the authorities fought to keep the inflation rate within the target range.

In Israel, inflation overshot the 8 percent target in 1994 when the economy performed more strongly than had originally been expected. Inflation expectations did not rise because authorities had already begun to tighten the stance of monetary policy, however—although the exchange rate was in line with the central tendency of the crawling peg when it became clear that achievement of the target was at risk. Indeed, this episode was useful because it helped cement the inflation target as the primary objective for monetary policy.

In Poland, inflation moved above the target range in the second half of 1999. Here, too, the central bank misjudged the extent of inflationary pressures present in the economy and conceded in its December 1999 inflation report that policy had been too loose. The bank raised its official interest rate by 100 basis points in September 1999 and again by 250 basis points in November 1999, after the bank’s inflation outlook signaled that a target breach was imminent. The yield curve shifted somewhat higher but remained inverted. The target breach did not appear to raise inflation expectations significantly, possibly because the central bank announced ahead of time that a breach was imminent and it clearly demonstrated its resolve to take the actions needed to address the situation.

Conducting monetary policy transparently and acting promptly in a preemptive manner may prevent potential breaches from ratcheting up inflation expectations. None of the observed breaches of the inflation targets significantly undermined the credibility of the inflation targeting frameworks. In part, this may reflect the transparent nature of the frameworks, which provide the public with a firm understanding of monetary policy objectives and the motivations behind the stances adopted by central banks. Moreover, in some cases (Canada and Poland), the central banks went so far as to publicly signal potential breaches of the targets in advance and acted promptly, when required, to forestall the breach. For example, in May 1995 the Bank of Canada publicly signaled in its Monetary Policy Report that an uptick in the inflation rate brought about by one-off exchange rate pass-through effects on the price level would soon be reversed. This helped to ensure that expectations did not become unhinged as the inflation rate rose above the midpoint of the target range. The fact that the bank’s projection proved correct may have contributed to the market’s subsequent willingness to accept easier monetary conditions (Clinton and Zelmer, 1997). The ability of an inflation targeting framework to absorb unexpectedly high inflation outcomes in a buoyant economic environment has not been fully tested, however.

Responding to Significant Economic and Financial Shocks

Emerging market countries are inherently more vulnerable to shocks owing to their more concentrated production base, greater openness (in most cases), less developed financial systems, and often less-entrenched policy credibility. Perhaps most important are large shifts in international capital flows that trigger rapid movements in the exchange rate, especially for countries still building credibility of the inflation targeting framework, and pronounced swings in the spreads between domestic and international interest rates. As a result, central banks in emerging market countries tend to rely on qualitative information obtained from their contacts with market participants and the private sector generally and on information from market indicators—for instance, spreads between nominal and real return bonds—when available.

Clear explanations by the central bank of the rationale underlying its policy stance may help nudge expectations in its favor. In situations where the central bank and the market disagree over the appropriate course of monetary policy, the central bank might find it advantageous to communicate the economic rationale underlying its desired policy stance and be clear about the way in which policy is implemented (Zelmer, 1996). Over time, even though a successful inflation targeting framework is unlikely to reduce the possibility of shocks, a proven track record of attaining the inflation targets should help market participants develop greater confidence in the motivations underpinning monetary actions.

Exchange Rate Movements

Central banks aim to ensure that exchange rate movements do not destabilize inflation expectations or domestic financial markets. The practical difficulty is identifying the cause of the exchange rate movement, which determines whether a policy response is warranted (e.g., in the case of a temporary external shock) or not (e.g., a permanent improvement in domestic productivity).

Many inflation targeting countries responded to exchange rate fluctuations triggered by the 1997–98 Asian/Russian financial crises. The central banks of Canada, Chile, Israel, and New Zealand raised official interest rates, in part because of a concern that the currency depreciation might feed through into domestic inflation expectations (Table 5.4). Chile and Israel, moreover, raised interest rates to accommodate higher risk premiums in the context of their exchange rate bands. This concern is especially important for emerging market countries with large and rapid pass-through effects from exchange rate movements to prices and inflation expectations. In New Zealand’s case, the central bank’s willingness to tolerate higher interest rates may have also been motivated by its focus on its monetary conditions index and the time taken to fully understand the consequences of the Asian shock for the New Zealand economy.

Table 5.4.

The Policy Response of Inflation Targeters to the Asian/Russian Crises

Sources: Central bank inflation reports and Bank for International Settlements, Annual Report 1999.

Table 5.4.

The Policy Response of Inflation Targeters to the Asian/Russian Crises

		Impact of Crises		Response of Monetary Authorities
Inflation Targeting Countries		Rapid Exchange Rate Depreciation	Higher Medium- and Long-Term Interest Rate Spreads	None	Foreign Exchange Intervention	Increase in Official Interest Rates
Emerging Markets
	Chile	Yes	Yes		X	X
	Czech Republic	No	No			X
	Israel	Yes	Yes			X
Industrial Countries
	Australia	Yes	Yes		X
	Canada	Yes	Yes		X	X
	Finland	No	No	X
	New Zealand	Yes	Yes			X
	Spain	No	No	X
	Sweden	No	No	x
	United Kingdom	No	No	X

Sources: Central bank inflation reports and Bank for International Settlements, Annual Report 1999.

View Table

Table 5.4.

The Policy Response of Inflation Targeters to the Asian/Russian Crises

		Impact of Crises		Response of Monetary Authorities
Inflation Targeting Countries		Rapid Exchange Rate Depreciation	Higher Medium- and Long-Term Interest Rate Spreads	None	Foreign Exchange Intervention	Increase in Official Interest Rates
Emerging Markets
	Chile	Yes	Yes		X	X
	Czech Republic	No	No			X
	Israel	Yes	Yes			X
Industrial Countries
	Australia	Yes	Yes		X
	Canada	Yes	Yes		X	X
	Finland	No	No	X
	New Zealand	Yes	Yes			X
	Spain	No	No	X
	Sweden	No	No	x
	United Kingdom	No	No	X

Sources: Central bank inflation reports and Bank for International Settlements, Annual Report 1999.

Domestic financial market stability is another rationale for limiting exchange rate movements prompted by volatile capital flows. The Czech National Bank noted in its inflation report that it adopted a cautious approach toward easing its monetary policy stance in 1998, partly because of concerns that events in Russia might otherwise spill over into its domestic markets. Similarly, the central banks of Canada, Chile, and Israel report that they were forced to raise interest rates aggressively in the latter half of 1998 when turbulence in the foreign exchange market unsettled local fixed-income markets. These central bank actions were meant to contain risk premiums in domestic medium and longer-term interest rates that were rising as investors (domestic and foreign) sought protection against possible losses from any further unexpected exchange rate depreciation. Tactical maneuvers like these are undertaken not necessarily because a tighter monetary stance is desired, but to avoid the more dramatic tightening that might otherwise occur if market expectations were to become destabilizing (Murray and others, 2000).

Financial fragility might limit the scope for allowing exchange rate movements. Central banks may want to moderate exchange rate movements if there is a risk that they could undermine weak balance sheets in the private sector—by triggering losses at banks and nonfinancial corporations with large open foreign exchange exposures (Lane and others, 1999; Stone, 2000). For example, in the case of highly leveraged balance sheets, monetary policy aimed at exchange rate stabilization must account for the negative impact on growth of bankruptcies. Trade-offs between currency stability and growth will depend on the extent of leverage and the split between domestic and foreign debt. Exchange stabilization is especially difficult if foreign currency or dollarized claims make up a large share of balance sheets (Mishkin, 2000).

Tactics for Coping with Shocks

Options in response to a shock have ranged from doing nothing to a mixture of foreign exchange intervention and tighter monetary policy. Finland, Spain, Sweden, and the United Kingdom were able to come through the shock period associated with the Asian and Russian financial crises without having to respond, since they were buffered to some extent by the capital flows into Europe in the period leading up to the introduction of Economic and Monetary Union. Australia, Canada, and Chile initially engaged in sterilized foreign exchange intervention to moderate the pace of exchange rate depreciation on the grounds that an increase in official interest rates was not warranted given the outlook for inflation domestically. Canada and Chile resorted to increases in official interest rates, however, when it became clear that foreign exchange intervention alone would not be sufficient to restore a sense of two-way risk in the foreign exchange market. Chile also relaxed restrictions on foreign capital inflows (BIS, 1999). Israel and New Zealand accepted higher interest rates without engaging in foreign exchange intervention. The Bank of Israel noted its reluctance to engage in such intervention because of concerns that it might encourage market participants to speculate against the central bank. Moreover, by not intervening, the bank hoped to build public awareness of exchange rate risk and encourage the private sector to properly hedge this risk so that the latter would be better positioned to cope with shocks in the future. The Czech Republic adopted a more gradual reduction in official interest rates than it would have chosen based exclusively on the inflation outlook.

Tactical responses to financial market shocks may cause inflation to move below the target. In the cases of the Czech Republic and Israel, a cautious approach to monetary easing (Czech Republic) during the Asian and Russian financial crises and large increases in interest rates that were only gradually unwound (Israel) were followed by a sharp fall in the inflation rate to levels below the target in the subsequent six-month period (see the mid-1999 inflation reports for those countries).

Exchange and capital controls were not used to cope with these shocks. Although such controls can help buy a limited amount of time in the short run to introduce appropriate policies, over the longer run they are less effective than sound macroeconomic policies and a strong well-supervised financial sector (IMF, 1999b).

Coping with Asset Price Bubbles

Inflation targeting central banks generally only look at asset prices and associated capital flows to the extent that they might affect inflation. The dominant view is that monetary policy should focus on inflation in the markets for goods and services; asset prices are taken into account only insofar as they might affect inflation—for example, through wealth effects on demand (Kent and Lowe, 1998). Nonetheless, some central banks have expressed concerns about “irrational” asset price bubbles, since a sharp reversal of these prices might, in some cases, undermine the stability of the financial system if there is a built-in reliance on the continuation of favorable price trends in asset markets. The prevailing view, however, is not to focus on asset prices directly in the conduct of monetary policy, but rather to take steps to insulate the financial system from the deleterious effects of asset price volatility. Nonetheless, given the shallowness of financial markets in emerging market countries, central banks might be more inclined to take a more protective stance toward sudden, rapid declines in asset prices even if it means that inflation moves outside the target range.¹⁷’¹⁸