Statistical Implications of Inflation Targeting
Chapter

13 Regime-Oriented Data Coverage Standards: Case Study of Inflation Targeting in Israel

Author(s):
Carol Carson, Claudia Dziobek, and Charles Enoch
Published Date:
September 2002
Share
  • ShareShare
Show Summary Details

The International Monetary Fund’s Special Data Dissemination Standard (SDDS) was initiated in 1996, in the aftermath of financial crisis in Mexico, to guide the production and dissemination of better macroeconomic data. In that brief time, the SDDS has gained increasing acceptance by subscribing countries and users alike. International lenders and credit rating agencies regard subscription to and compliance with the SDDS as important features for countries wishing to gain access to international capital markets. Subsequent international financial crises, of yet greater frequency and severity, have augmented the perceived need for macroeconomic data that meet such a standard.

At its inception, IMF Directors viewed the SDDS as a flexible concept that ought to be adapted to changing circumstances. They called for reviews of the standard in order to provide opportunities to make needed adjustments. To be sure, if the SDDS is to be of any value it must provide at any point in time a rigid standard for data dissemination that participants must observe. But over time the detailed provisions of the standard must be reviewed and adjusted periodically to maintain relevance and, if possible, to introduce improvements.

This chapter suggests a conceptually simple framework for improving the data coverage requirements of the SDDS, namely a regime-oriented data coverage standard (called RODS here for syntactical simplicity). The objective of RODS is to further improve the transparency of macroeconomic policy by requiring the dissemination of data that are of greatest relevance for evaluating given policy measures implemented at a given time, in a given place.

This chapter illustrates the RODS concept for the case of inflation targeting. The authors’ decision to concentrate on a particular regime is based on their comparative advantage and on their assessment that the details are very important in considering a modification of the data standard. Of course, the RODS concept augments, but does not invalidate, existing data coverage standards, which could become, in large part, a common core for all regimes. Details of regime-specific coverage for other regimes must be addressed by experts in those other regimes.

This chapter now turns to an example of data coverage in an inflation-targeting regime, using the authors’ experience in Israel as a case study. The next section presents a general overview of our ideas for data dissemination, including attributes of the regime itself. The chapter then presents the data. It concludes with some thoughts on implementing the RODS concept.

Regime-Oriented Data Standard for Inflation Targeting

There are two key elements of primary interest in an inflation-targeting regime: parameters of the inflation target itself, and inflation forecasts. Information about the inflation target itself should be disseminated not only to enable the public to assess performance in attaining the target, but also because the nature of the target itself is an important indicator of the degree of commitment to the inflation-targeting framework, especially in the initial period after adoption. Information about the inflation-targeting regime would include items such as (1) the value of the target; (2) the horizon over which the target must be attained; (3) the specific variable that is being targeted; (4) whether there are “escape” clauses; (5) the nature of provisions to ensure accountability, such as the publication of inflation reports; and (6) possible sanctions if the target is missed. The format for the dissemination of data on the details of the inflation-targeting regime might be a framework similar to the IMF’s Exchange Arrangements and Exchange Restrictions publication, rather than time-series data. This framework might include a website that is continuously updated as details of inflation-targeting regimes are changed by policymakers, and a hard copy to be prepared, say, annually.

This section turns now to the inflation forecast and related forward-looking data, the key feature of RODS for inflation targeters. Forecasts are important in inflation targeting because of the relatively weak control that the monetary authorities have over inflation itself as compared with other possible monetary policy targets such as the exchange rate or even the money supply. The lags in the effect of monetary policy on inflation imply that policy settings—for example, central bank interest rate decisions—must be based on forward-looking considerations to a greater extent in inflation-targeting regimes than in monetary or exchange-rate regimes. Indeed, in an inflation-targeting regime it is both natural and practical to view the inflation forecast as the intermediate target of monetary policy. Furthermore, in regimes where some flexibility of inflation targeting is allowed (Svensson, 2000, provides an early discussion of flexible inflation targeting), forecasts of other variables, especially the output gap and unemployment, are needed.

The forecasts that are used, at least implicitly, in policy decision making are generally judgmental assessments based on a variety of underlying forecasts that in turn are derived from alternative methodologies. For example, a given inflation forecast at a central bank may be some judgmental consensus based on forecasts from an econometric model or several such models, judgmental forecasts of central bank staff experts or policy board members, market-based forecasts, and forecasts provided by outside forecasters from the academic or business community. Finally, a variety of indicators may provide some information about the likely development of the inflation path, even though the indicators are not aggregated into a quantifiable inflation forecast. For example, a growing share of unindexed domestic currency assets in the public’s portfolio is likely to indicate greater credibility of an inflation-targeting policy, especially in countries where a high degree of indexation or exchange rate linkage prevailed during earlier high-inflation periods. Decision theory suggests that combining a number of forecasts, especially from different methodologies, leads to improved policy decisions (see, for example, Armstrong, 1989). But how to combine a possibly large number of forecasts from a smaller, yet non-negligible, number of methods is also an important issue.

With this in mind, we propose that the regime-oriented component for inflation targeters be composed of three subcomponents: (1) details of the regime itself; (2) data on inflation forecasts; and (3) data on key indicators of inflation that are not expressed as inflation forecasts. Not much more can be said at this stage about the details of how the standard for each of these components might be established. Such a standard must be based on a broad survey of the types of information used in policymaking in the various inflation-targeting countries. While some such survey information is available, the establishment of a standard would require a more substantial effort, and it is beyond the scope of the present work. Instead, this chapter provides as a case study an overview of the information that is provided in the monthly monetary decision meetings at the Bank of Israel.

Israel as a Case Study for Regime-Oriented Data Dissemination

The main topic of this chapter is a systematic review of the data on indicators that are used at the Bank of Israel in formulating its inflation-targeting monetary policy. This set of indicators serves as one example of the kinds of indicators that might be included in a data coverage standard for an inflation-targeting regime. Indicators can be divided into four broad classes: market-derived information, forecasts of professional forecasters, model-based forecasts, and other indicators. For each type of data series, four key issues are considered: (1) a basic exposition of the particular data series, including some conceptual justification for looking at it; (2) how the data are constructed at the Bank of Israel, with some discussion of the available series; (3) how the series is used in policymaking; and (4) practical problems that arise in constructing and using the data.

Information Derived from Financial Market Prices

Information derived from financial market prices is an area where Israel has an advantage relative to many other countries. The use of this type of information has been developed in Israel because the very noisy nature of the economy made econometric modeling efforts difficult for many years (this is changing now), while financial markets were reasonably well developed compared with other emerging market economies.

Market-based inflation expectations

Market-based inflation expectations (MBIEs) can be constructed from the yields on nominal bonds and bonds indexed to the consumer price index (CPI) that are traded in reasonably well-functioning markets. MBIEs over a given time horizon are calculated as the “difference” between the yield on a plain vanilla nominal bond for that horizon and the yield on a CPI-indexed bond for that horizon, where the two bonds are (nearly) identical in all other respects. The basis for this type of calculation is the well-known Fisher equation relating real and nominal interest rates and inflation expectations.

Figures 13.1 through 13.3 include data on MBIEs for Israel. Figure 13.1 shows a time series plot of 12-month-ahead MBIEs for the period during which inflation in Israel reached low single-digit levels. Actual inflation for the previous 12 months is also shown, along with the corresponding inflation target and inflation forecasts of professional forecasters. There is a close relationship between the MBIEs and past inflation. The degree of dependence of inflation expectations on the recent past is not yet well understood, especially in light of the regime change that has taken place. The MBIEs are not as well correlated with actual inflation over the forecast horizon (not shown here), but that is to be expected during periods of general macroeconomic turbulence (financial crises abroad, geopolitical problems in the region, and so forth) or, more specifically, of a monetary regime change whose specific features, such as timing or degree of seriousness, are not always well known in advance. Figure 13.2 contains data on inflation expectations “yield curves”—that is, each line contains the spectrum of MBIEs for a given date (actually average of business days over a given month) spanning the available time horizon of MBIEs at the time the curve was constructed. In Israel the periods spanned by such yield curves have lengthened over time (not shown in this figure) because the maximum maturity of nominal bonds was increased in various steps during the period, whereas indexed bonds have always been available for longer terms to maturity than nominal bonds. Figure 13.3 is a comparison of inflation forecasts from the market (MBIEs), professional forecasters, and a macroeconometric model maintained by the Monetary Department. One interesting feature of this figure is the fairly long period during which the MBIEs were the lowest of the various forecasts, suggesting the absence of a positive inflation risk premium in the MBIEs (see below).

Figure 13.1.Israel: Market-Based Inflation Expectations (12 Months Ahead), Forecasts of Professionals for the Next 12 Months, Actual Inflation, and the Inflation Target1

(Percent)

Source: Bank of Israel.

1 Monthly averages; actual inflation is current consumer price index vis-à-vis 12 months ago.

Figure 13.2.Israel: Market-Based Inflation Expectations—“Yield” Curves

Source: Bank of Israel.

Figure 13.3.Israel: Inflation Expectations for 12 Months—Model, Market, and Forecasters

Source: Bank of Israel.

The MBIEs are among the most important indicators for monetary policy decision making in Israel. The presence of reasonably well functioning markets in indexed bonds with a minimal indexation lag is a key factor in the prominence of these indicators. A succession of large and varied macroeconomic shocks has hit the Israeli economy in the past 15 years, including disinflation, real and financial reform, massive immigration, and political fluctuation; these have hampered conventional modeling efforts for a long time. Recently, as the economy has stabilized, modeling efforts have been more successful, and modelers are taking advantage of the availability of MBIEs by incorporating them in model specification.

MBIEs are affected by a number of important deficiencies. Because of the prominent place of MBIEs in monetary policy decision making, the Bank of Israel has spent considerable effort on minimizing the possible distortions caused by these problems. The key problems and some remedies that have been applied or considered are as follows:

  • Tax treatments of the nominal and real bonds are not identical. Yields on Treasury bills (nominal, pure discount, bonds of up to one-year maturity) are not taxed at all. The inflation-adjusted coupon yields of longer-maturity nominal bonds and the real coupon on indexed bonds are taxed, but at differing rates for different investors. It is not clear what the tax rate is for the “marginal” investor in the coupon bonds. The Bank of Israel has conducted a number of studies to try to ascertain, as nearly as possible, the relevant tax rates, though the results are far from clear-cut.

  • The interest rate differential is likely to include an inflation risk premium. The risk premium almost surely raises the nominal returns above the indexed returns by more than inflation expectations, although the opposite possibility cannot be entirely ruled out. A recent study by Stein (2001), using the capital asset pricing model (CAPM), finds that the average inflation risk premium from 1995 to 1999 was approximately 40 basis points. Nevertheless, as mentioned, Figure 13.3 provides some indication that this risk premium is not always positive.

  • Indexation on the indexed bonds is not perfect, so there is some nominal component in them. The terms of Israel’s indexed bonds provide for indexation from the last known CPI value before the bonds were issued up to the last known CPI value before they were redeemed, so the indexation is much better than that on comparable bonds in some other countries such as the United Kingdom and the United States, where there is an indexation lag of up to eight months. Furthermore, under conditions of near price stability, the lack of perfect indexation is not likely to have a significant effect on relative returns, unlike periods of high inflation, when this factor can be very important.

  • Poor liquidity, especially in the indexed bond market, is likely to generate an illiquidity premium on the indexed bonds. While this factor may be presumed to offset in part any inflation risk premium on nominal bonds, there is no presumption about the relative magnitudes of the two premiums.

  • The precise maturities of nominal and real bonds for a given time horizon may not be perfectly matched. For example, inflation expectations for a one-year-ahead horizon may be based on a nominal bond with maturity between 11 and 12 months, because series are issued monthly, while series of real bonds may be issued at lower frequency than a month. The one-year real yield is generally calculated as the average of yields on all of the series between 10 and 14 months.

Each of these problems has some non-negligible degree of importance, so the Bank continues to make an effort to limit their severity where possible and to keep them in mind when using MBIEs in decision making. Nevertheless, the Bank feels that the problems are not so severe as to prevent the MBIEs from serving as a key indicator for monetary policy. It is worth recalling that there have even been proposals to target MBIEs, including an analytical paper by Boschen (1988) laying out the advantages of such a policy framework, and strong support for such an approach from Milton Friedman (1992), who went so far as to say that targeting MBIEs “may be more politically feasible than my own earlier proposals for structural change” (page 227). In view of the practical problems outlined above, and also because the conceptual “monkey in the mirror” problem raised in Bernanke and Woodford (1997) may be important for forecast targeting, the Bank feels that targeting MBIEs would be carrying things too far.

Market-based interest rate expectations

Market-based interest rate expectations (MBREs) are implied forward spot rates calculated from yield curves based on the expectations theory of the term structure of interest rates, augmented where possible by estimates of risk premiums. Such MBREs may be obtained from the nominal yield curve (that is, from rates on unindexed bonds), and from the real yield curve (from rates on indexed bonds). For brevity, this chapter considers only the nominal yield curve, which is more familiar to readers in most countries. Because the Bank of Israel controls the level of the very short nominal spot rates at each moment, and the public understands this well, the implied forward spot rates represent the public’s expectations of future interest rate decisions by the Bank of Israel. Calculation of these implied forward interest rates thus provides policymakers with a good indication of the public’s view of coming policy measures. The profile of expected future short rates can then be compared with model simulations of the time path of the bank’s interest rate that is required to hit the inflation target.

Figure 13.4 presents the “yield curve” of nominal MBREs for the average number of business days from February 1 to February 20, 2002.

Figure 13.4.Israel: Implied Forward Nominal Interest Rates

(February 1–20, 2002; average percent)

Source: Bank of Israel.

The figure indicates expectations of monetary tightening during most of 2002.

The main problems with MBREs are a subset of problems with MBIEs. The most important issue is that the term structure of interest rates is likely to include risk premiums, over and above expectations of future spot rates, and while there are no major differences in tax treatment of different nominal bonds at present, such differences may arise in the future.

Implied volatilities from options markets

A variety of options are traded in Israeli markets. Because the focus of this presentation is monetary policy, it does not consider options on individual stock prices or stock price indexes; the discussion is confined to foreign currency options (that is, options on the exchange rate). Three main categories of forex options may be distinguished in Israel: (1) forex options traded on the Tel Aviv Stock Exchange (TASE); (2) over-the counter forex options, where one of the parties is usually a bank and the options are often part of some tailored risk management package for a customer; and (3) forex options written by the Bank of Israel.

The forex options that have been of greatest interest for monetary policy are the ones written by the Bank of Israel. These are options on the new sheqel/dollar exchange rate; they include three- and six-month call options, issued weekly in volumes of $10 million and $5 million, respectively, and three-month put options, issued weekly with a volume of $10 million. An important feature of all of these options is that their strike price is forward-at-the-money—that is, the strike price is an approximation to the forward new sheqel/dollar exchange rate. Because the market for forward forex is not sufficiently well organized, the forward rate is approximated by the difference between a nominal new sheqel interest rate for the relevant horizon and a comparable U.S. dollar-denominated interest rate.

Because the options written by the Bank of Israel are forward-at-the-money, the premium reflects pricing of pure risk, and it is possible to extract relatively “clean” implied volatilities for the market’s expectation of the probability distribution of the forward new sheqel/dollar exchange rate. These implied volatilities are a measure of the uncertainty about the path of this exchange rate.

The main use of information derived from these options is to assess the market’s degree of uncertainty about exchange rate developments. This can be expressed in a variety of ways, including time series plots of the implied volatilities, probabilities of depreciation of given amounts, or even exchange rate “fan charts.” Note that so far, all of this information is based on the Black-Scholes options pricing formula (see discussion below).

Figure 13.5 presents a time series plot of implied volatilities from Bank of Israel three-month call options and the ratio of implied volatilities from three- and six-month call options (adjusted for the expected increase in volatility over the longer horizon under the [questionable] assumption of lognormality of the exchange rate). An interesting feature of this body of data is the stability of the series in spite of significant economic and political shocks that have hit the Israeli economy over the past year.

Figure 13.5.Israel: Implied Volatility Derived from Three-Month Forex Options, and Ratio Between Six-Month and Three-Month implied Volatility

Source: Bank of Israel.

1Raw ratio divided by √2, to reflect the expected effect on uncertainty due to passage of time.

There are at least two problems with these options. First, the Black-Scholes formula is used even though some of its key assumptions are not fulfilled. For example, Israel maintains a crawling band exchange regime, so when the exchange rate nears one of the bounds of the band, the stochastic process of the underlying asset is not likely to be lognormal. And even when the exchange rate is well within the band, its distribution appears to be fat-tailed relative to the log normal benchmark. In addition, although the underlying asset, forex, is traded continuously during business hours and then some, the options themselves are not tradable.

Forex options traded on the TASE are also of interest for monetary policy because generally there are a number of options that are nearly identical except for the strike price, so it is possible to extract the so-called “smile” from these options and interpret it. The smile indicates where the market believes greater risk to be concentrated. Figure 13.6 presents “smile” plots for a number of dates that have been carefully selected to tell interesting stories. Each curve is identified by the date for which it was constructed and the representative spot exchange rate for that date. The issue of interest is as follows: On December 23, 2001, the Israeli government and the Bank of Israel announced a package of policy measures that included a radical, one-time interest rate reduction of 2 percentage points, from 5.8 to 3.8 percent; fiscal retrenchment in the form of a budget cut equal to about 1.5 percent of GDP; and some reform measures in financial markets. A typical smile curve before this announcement (December 10) is fairly symmetric but indicates a fat-tailed distribution of the underlying exchange rate. The first two smile curves shown for dates after the announcement (December 30 and January 6) are quite asymmetric, indicating that the public was willing to pay higher premiums for options with exercise prices in the direction of appreciation. But following substantial depreciation of the Israeli sheqel, during January and the beginning of February, the smile curves became flat.

Figure 13.6.Israel: Implied Volatility from Call Options According to Strike Price (Smile)

(Percent)

Source: Bank of Israel.

The smile curves are based on traded options, so they have a distinct advantage over the Bank of Israel options. Nevertheless, they also suffer from a number of problems. The various options used to construct a given smile curve are not completely identical; for example, there is some variation of the expiration dates, although all are between one and two months, where there is generally a concentration of options trading. Longer options are not well traded, and there is a paucity of observations on premiums. Moreover, not all of the options from a given curve are traded at or near close of business, so some prices may be “outdated” (the Bank is currently considering the construction of a database of offers to buy or sell at a given time).

In spite of the various problems, the Bank of Israel does feel that the implied volatilities and smile curves give some reasonable indication of the degree of exchange rate uncertainty, and hence are an indicator of the credibility of monetary policy in promoting price and financial stability.

Information from Professional Forecasters

The professional macroeconomic forecasting industry is a relatively new phenomenon in Israel. Until the early 1990s, various large firms employed macroeconomists to provide analysis and forecasts for internal uses, but they generally did not publish forecasts, certainly not on a systematic basis. Since the mid-1990s there has been a growing trend toward the preparation and publication of macroeconomic forecasts by independent forecasters and by forecasters employed by commercial banks and other large firms to publish forecasts of key macroeconomic variables or make forecasts available to the Bank of Israel for internal policy analysis.

The main variables that are of interest from professional forecasters are the inflation rate, the Bank of Israel’s interest rate, and the exchange rate. Forecast horizons generally include the coming few months, the end of the current calendar year, and 12 months ahead. The Monetary Department prepares a weekly table summarizing these forecasts over various horizons.

Forecasts of professional forecasters serve as key indicators of how monetary policy is perceived by the financial industry in particular, and by the informed business community at large. The main problems with professional forecasts are that—even though identified individuals prepare them—there is no information on how they are prepared or what assumptions are being made about important exogenous variables. In addition, some of the forecasts are not systematic, because forecasters move from one employer to another, the forecasts are not always produced regularly, and sometimes certain forecasters do not bother to update their forecasts, even when significant events occur.

As noted, Figure 13.1 compares the inflation forecasts of professional forecasters with those of the market and with recent actual inflation. As with MBIEs, there is a close correspondence between 12-month-ahead forecasts and actual inflation over the previous 12 months. Figure 13.7 provides the consensus (that is, average) professional forecasters’ forecasts for three variables over a 12-month horizon. The variables are the inflation rate, the Bank of Israel interest rate, and the exchange rate.

Figure 13.7.Israel: 12-Month Forecasts of Professional Forecasters

Source: Bank of Israel.

Information from Macroeconometric Models

This topic is covered briefly, because it is likely to be quite familiar to most readers. Israel has experienced a variety of major economic shocks since the mid-1980s, including a major economic stabilization program that reduced inflation from high triple-digit figures to low double-digit figures, major structural reforms in the business sector following that program, gradual but highly significant reforms in financial markets, a massive wave of immigration, frequent alteration of the monetary and exchange regime, and developments of all sorts on the political front. For this reason, the estimation and use of macroeconometric models has not been an easy task, and the Bank of Israel has placed a high weight on market-based indicators for monetary policy relative to model-based indicators, by international standards. This situation has been changing rapidly in the past three or four years as the economy has become more stable. A few “reasonable” and functional models have recently been estimated at the Bank, but there is still no consensus to regard one of them as the Bank of Israel model, with status similar to, say, the models used in Canada or the United States.

Key advantages of models are their transparency—that is, a clear and replicable specification of the transmission mechanism of monetary policy and the ability to conduct controlled experiments via alternative model simulations. Their use in policy formulation is obvious and familiar. Key problems with models are their lack of flexibility and their derivation from very specific assumptions about the transmission mechanism, reflecting views that may not always be in consensus.

Inflation forecasts from a baseline projection of the model currently maintained by the Monetary Department of the Bank of Israel are shown in Figure 13.3. Following the substantial easing of monetary policy on December 23, 2001, and subsequent depreciation of the sheqel, model-based inflation forecasts have increased well beyond the inflation target range and beyond the forecasts of professional forecasters. While differences of opinion make a horse race, different forecasting methodologies provide substantial food for policymakers’ thoughts.

Information from Portfolio Composition

The composition of liquid asset portfolios provides further information for assessing the credibility of an inflation-targeting regime, primarily during periods of disinflation.

In Israel, the overall asset and liability portfolio of the nonfinancial private sector can be divided into three broad categories by type of indexation: (1) unlinked sheqel (local currency) instruments; (2) instruments indexed to the price level, usually the CPI; and (3) instruments denominated in foreign currency (in actual forex and in sheqels indexed to the exchange rate). Growth of the unindexed sheqel sector is viewed as an indication that the public views disinflation efforts as being credible in view of announced target paths. Furthermore, greater willingness by the public to take a position in longer-term-to-maturity unindexed assets, as opposed to shorter-term ones, and a greater share of fixed interest rate instruments as opposed to variable rate ones, are also interpreted as greater credibility of the disinflation efforts.

Interpretation of these developments in portfolio composition must be done with special care because there are often other factors affecting such compositions, from long-term influences such as financial reforms to short-term shocks. In general, therefore, inferences about the credibility of the regime is based only on steady, trend-like movements in portfolio composition and not on sudden shifts.

Figure 13.8 presents the composition of the nonfinancial public’s portfolio of financial assets for two widely disparate dates. The rise in the share of unindexed assets is the result of a long-term trend that has been taking place since serious disinflation efforts began in Israel in the mid-1990s.

Figure 13.8.Israel: Composition of Public’s Financial Assets Portfolio

(Percent)

Source: Bank of Israel.

Moving Toward a Regime-Oriented Data Standard

The experience of one inflation-targeting country is not a sufficient basis for even beginning to formulate specific proposals for a data standard that would be relevant for all inflation targeters. The authors’ proposal, therefore, is to begin a process of broad-based exploratory work whose goal would be the formulation of such a standard. The first stage of the work would be to survey all of the current inflation targeters (so far there are no “ex-targeters”) in order to determine the range of indicators they use in the formulation of monetary policy. The various indicators could then be grouped into categories by alternative classification modalities. Examples of these modalities include

  • economic categories such as inflation, interest rates, and real activity;

  • sources of data (market-based, model-based, outside forecasters, and so on); and

  • length of horizon for forecasts—for example, up-to-target horizon or beyond-target horizon.

It would also be useful to inquire as to the relative importance of principal groups of these indicators. The results of this survey could then be used to propose a standard for all inflation targeters. The nature of the standard presumably would depend on the nature of the variety of the indicators that are used. For example, if nearly all countries look at forward nominal interest rates, a fairly specific dissemination standard, such as a 12-month-ahead T-bill forward, could be prescribed. At the other extreme, if there is a very large variety of practice in inflation expectations, the standard could prescribe that each country disseminate the two or three (for example) most important forecasts of different types (model-based, market-based, professional forecasters, and so forth, but not a large number of forecasts of the same type).

The ultimate objective of data dissemination standards is to improve the transparency of policy. Therefore, whatever the details of the construction of the RODS for inflation targeting, the objective of the standard should be to provide a net (as in fishing net) that catches the most important indicators used regularly in each inflation-targeting country. Because a substantial amount of work on inflation targeting has already been done in the IMF and elsewhere, it might not be unreasonable to expect that an initial proposal for a standard could be developed by the end of 2002.

The authors are grateful to Ofer Klein for substantial contributions to this chapter.

    Other Resources Citing This Publication