-A___\WPS2 5( POLICY RESEARCH WORKING PAPER 2511 M onetary Policy under Inflation targeting is a flexible policy framework that allows Flexible Exchange Rates a country's central bank to exercise some degree of An Introduction to Inflation Targeting discretion without putting in jeopardy its main objective of maintaining stable prices. Pierre-Richard Agenor The World Bank World Bank Institute Economic Policy and Poverty Reduction Division U December 2000 iC; P oLHcY RESEARCH WORKING PAPER 2511 Summary findings In the past few years a number of central banks have exercise some degree of discretion without putting in adopted inflation targeting for monetary policy. Agenor jeopardy its main objective of maintaining stable prices. provides an introduction to inflation targeting, with an In middle- and high-income developing economies that emphasis on analytical issues and the recent experience can refrain from implicit exchange rate targeting, it can of middle- and high-income developing countries (which improve the design and performance of monetary policy have relatively low inflation to begin with and reasonably compared with other policy approaches that central well-functioning financial markets). banks may follow. After presenting a formal analytical framework, Not all countries may be able to satisfy the technical Agenor discusses the basic requirements for inflation requirements (such as adequate price data, adequate targeting and how such a regime differs from money and understanding of the links between instruments and exchange rate targeting regimes. targets of monetary policy, and adequate forecasting After discussing the operational framework for capabilities), but such requirements should not be inflation targeting (including the price index to monitor, overstated. Forecasting capability can never be perfect, the time horizon, the forecasting procedures, and the and sensible projections always involve qualitative role of asset prices), he examines recent experiences with judgment. inflation targets, providing new evidence on the More important, and often more difficult, is the task convexity of the Phillips curve for six developing of designing or improving an institutional framework countries. that would allow the central bank to pursue the goal of His conclusions: Inflation targeting is a flexible policy low, stable inflation while maintaining the ability to framework that allows a country's central bank to stabilize fluctuations in output. This paper-a product of the Economic Policy and Poverty Reduction Division, World Bank Institute-is part of a larger effort in the institute to understand the dynamics of monetary policy in developing countries. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Maria Gosiengfiao, room J4-282, telephone 202-473-3363, fax 202-676-9810, email address mgosiengfiao@worldbank.org. Policy Research Working Papers are also posted on the Web at www.worldbank.org/research/workingpapers. The author may be contacted at pagenor@worldbank.org. December 2000. (94 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Policy Research Dissemination Center Monetary Policy under Flexible Exchange Rates: An Introduction to Inflation Targeting Pierre-Richard Agenor* The World Bank Washington DC 20433 JEL Classification Numbers: E44, F32, F34. *I would like to thank, without implication, Esteban Jadresic, Brian Kahn, and Murat Ucer for useful comments on an earlier draft, and Nihal Bayraktar for excellent research assistance. Contents 1 Introduction 3 2 Inflation Targeting: A Conceptual Framework 5 2.1 Strict Inflation Targeting ..................... 5 2.2 Policy Trade-offs and Flexible Targeting ..... . . . . . . . 11 2.3 Inflation Targeting in an Open Economy ..... . . . . . . . 15 3 Comparison with Intermediate Target Strategies 19 3.1 Monetary vs. Inflation Targeting ...... . . . . . . . . . . 19 3.2 Exchange Rate vs. Inflation Targeting ...... . . . . . . . 20 4 Basic Requirements for Inflation Targeting 22 4.1 Central Bank Independence and Credibility ..... . . . . . 22 4.2 Absence of de facto Exchange Rate Targeting ..... . . . . 24 4.3 Transparency and Accountability ...... . . . . . . . . . . 25 5 The Operational Framework of Inflation Targeting 27 5.1 Establishing Inflation Targets ....... . . . . . . . . . . . 28 5.1.1 The choice of a price index ...... . . . . . . . . . . 28 5.1.2 Width of the target band ....... . . . . . . . . . . 33 5.1.3 Horizon of the inflation target ...... . . . . . . . . 34 5.1.4 Forecasting procedure ....... . . . . . . . . . . . 35 5.2 Interest Rates Rules in Practice ....... . . . . . . . . . . 37 5.3 Asset Prices and Inflation Targeting ...... . . . . . . . . . 39 6 Recent Experiences 40 -6.1 Industrial Countries .......... .. .. . .. .. .. .. . 41 6.2 Developing Countries ......... .. .. . .. . .. .. . . 44 7 Some Unresolved Analytical Issues 48 7.1 Asymmetric effects .......... .. .. .. . .. .. .. . 48 7.1.1 Non-quadratic policy preferences ..... . . . . . . . 48 7.1.2 The convex Phillips curve ...... . . . . . . . . . . 50 7.2 Uncertainty and optimal policy rules ...... . . . . . . . . 55 7.3 Endogenizing Reputation and Credibility ..... . . . . . . . 59 8 Summary and Conclusions 62 2 1 Introduction There is growing acceptance among both policymakers and economists that the pursuit of price stability (defined as maintaining a low and stable rate of inflation) is the main medium- to long-run goal of monetary policy. The first reason is the recognition that a high and variable inflation rate is socially and economically costly. These costs include price distortions, lower savings and investment (which inhibits growth), hedging (into precious metals or land) and capital flight (into foreign assets). The second is that experience has shown that short-term manipulation of monetary policy instruments to achieve other goals-such as higher output and lower unemployment-may conflict with price stability. The attempt to achieve these conflicting goals tends to generate an inflationary bias in the conduct of monetary policy without, in the end, achieving systematically higher output and employment. To achieve the goal of price stability, monetary policy in many countries was for a long time conducted by relying on intermediate targets such as monetary aggregates or exchange rates. During the 1990s, however, several industrial and developing countries have begun to focus directly on inflation itself. This new approach to the problem of controlling inflation through mon- etary policy is known as inflation targeting.' It essentially makes inflation- rather than output or unemployment-the primary goal of monetary policy. It also forces the central bank to predict the future behavior of prices, giving it the opportunity to tighten policies before sustained inflationary pressures develop. A large literature has examined the practical experience of industrial countries with inflation targeting (see, most recently, Bernanke, Laubach, Mishkin, and Posen (1999), and Schaechter, Stone, and Zelmer (2000)). The purpose of this paper is to provide an overview of analytical issues associated with inflation targeting, with a particular focus on the policy and structural context of developing countries and their recent experience. Whether infla- tion targeting has a wider applicability to developing economies has indeed been a matter of debate in recent years, with authors like Masson, Savastano and Sharma (1997) taking a rather cautious view. It has been argued, for instance, that poor data on prices and real sector developments, the absence of reliable procedures for forecasting inflation, the difficulty of maintaining de 1As discussed below, two major reasons why countries chose to implement inflation tar- geting over alternative monetary policy frameworks were exchange rate crises and money demand instability. 3 facto independence for the central bank, and the lack of an anti-inflationary history may preclude the establishment of a transparent framework for con- ducting monetary policy and therefore any attempt at inflation targeting. However, others (including Mishkin (2000) and Morande and Schmidt-Hebbel (1999)) have adopted a more favorable position at least for the case of high- and middle-income developing countries, where the financial system is suf- ficiently developed to permit the use of indirect instruments of monetary policy. Understanding the terms of this debate is essential because several developing countries have in recent years adopted floating exchange rates (of- ten as a result of unsustainable exchange rate pressures on their adjustable peg regimes) and must therefore find another nominal anchor to guide do- mestic monetary policy over the medium and long term. The remainder of the paper is structured as follows. Section II presents an analytical framework for inflation targeting in both closed and open economies, based on the important work of Svensson (1997b, 1999b). The closed-economy model provides the starting point for understanding the na- ture of an inflation targeting regime; it is then extended to an open-economy setting to highlight the role of the exchange rate in the transmission process of monetary policy. Section III compares inflation targeting regimes with money supply and exchange rate targeting regimes and highlights the risks associated with pursuing implicit exchange rate targets. Section IV identifies three basic requirements for implementing an inflation targeting framework, namely, central bank independence, the absence of implicit targeting of the exchange rate, and transparency in the conduct of monetary policy. The operational framework of inflation targeting is the focus of section V. It discusses, in particular, issues associated with the measurement of inflation (including sources of imperfection in traditional measures), whether a target band for inflation is more appropriate than a point target, the time hori- zon of monetary policy, the inherent difficulties associated with forecasting inflation, and whether asset prices should be taken into account in assess- ing inflationary pressures. Section VI reviews the recent experience of both industrial and developing countries with inflation targets, with a particular emphasis on the latter group. The last section focuses on some unresolved analytical issues in the design of inflation targeting regimes, namely the role of nonlinearities and asymmetric effects (related to both the form of policy preferences and structural relationships, most notably the Phillips curve), uncertainty (about behavioral parameters and the transmission process of monetary policy), and the treatment of credibility and reputation in empiri- 4 cal macroeconomic models of inflation. New results regarding the convexity (or lack thereof) of the Phillips curve are also presented for six developing countries. The conclusion summarizes the main results of the analysis and offers some final remarks. 2 Inflation Targeting: A Conceptual Frame- work The first step in understanding the nature of an inflation targeting framework is to analyze the relation between explicit policy goals, policy instruments, and preferences of the central bank (which affect the form of its reaction func- tion).2 This section begins by examining the link between inflation targets and the nominal interest rate (viewed as the main instrument of monetary policy) when the central bank is concerned only about deviations of actual inflation from its target value. The analysis is then extended to consider the case in which both output and inflation enter the central bank's loss func- tion. In both cases the analysis focuses on a closed economy; open-economy considerations will be discussed later on. 2.1 Strict Inflation Targeting Following Svensson (1997b), consider a closed economy producing one (com- posite) good. The economy's structure is characterized by the following two equations, where all parameters are defined as positive: 7rt- 7rt-1 = alYt-1 + Et, (1) Yt = OlYt-i -2(it-1 -7t-1) + yt, 31 < 1, (2) where 7rt -Pt - Pt-i is the inflation rate at t (with Pt denoting the logarithm of the price level), Yt the output gap (defined as the logarithm of actual to potential output), and it the nominal interest rate (taken to be under the direct control of the central bank). Et and qt are independently, identically distributed (i.i.d.) random shocks. 2In what follows the term "instrument" is used in a broad sense to refer both to the operational target of monetary policy and to the actual instrument(s) available to achieve this target. 5 Equation (1) indicates that changes in inflation are positively related to the cyclical component of output, with a lag of one period. Equation (2) relates the output gap positively to its value in the previous period and negatively with the ex post real interest rate, again with a one-period lag. In this model, policy actions (changes in the nominal interest rate) affect output with a one-period lag and, as implied by (1), inflation with a two- period lag.3 The lag between a change in the policy instrument and inflation will be referred to in what follows as the control lag or control horizon. The assumption that the central bank controls directly the interest rate that affects aggregate demand warrants some discussion. In principle, what affects private consumption and investment decisions is the cost of borrowing, that is, given the characteristics of the financial structure that prevails in many developing countries, the bank lending rate. In general, bank lending rates depend on banks' funding costs, a key component of which is either the money market rate or (ultimately) the cost of short-term financing from the central bank.4 Thus, to the extent that bank lending rates (and money market rates) respond quickly and in a stable manner to changes in policy rates, the assumption that the central bank controls directly the cost of borrowing faced by private agents can be viewed simply as a convenient shortcut.5 What, then, is the evidence? Figure 1 reports impulse response functions of a one-standard deviation increase in the central bank's discount rate (taken to be the policy rate) in a group of six developing countries for which data were readily available. These responses are obtained from a bivariate vector autoregression (VAR) model that includes the policy rate and the money market rate (see Appendix B for details). The figure shows 3Note that introducing a forward-looking element in equation (1) would imply that monetary policy has some effect on contemporaneous inflation; this would make the solu- tion of the model more complicated but would not affect some of the key results discussed below. See Appendix A, which dwells on Clarida, Galf and Gertler (1999), for a discussion. Nevertheless, it should be kept in mind that the assumption of model-consistent expecta- tions also has drawbacks; in particular, it downplays the role of model uncertainty-which, as discussed later, may be very important in practice. 4Bank lending rates also depend on the perceived probability of default of potential borrowers and, in an open economy, the cost of funding on world capital markets. See Agenor and Aizenman (1998) for a model that captures these features of bank behavior. 5Note that if aggregate demand depends on longer-term interest rates, a similar effect would arise. This is because longer-term rates are driven in part by expected future movements in short-term interest rates, which are, in turn, influenced by current and expected future policy decisions of the central bank. 6 that (except for Uruguay) market interest rates respond relatively quickly and significantly to changes in official interest rates. These results suggest therefore that it is a reasonable analytical approximation to assume, as is done here, that the central bank controls directly the interest rate that affects aggregate demand. The central bank's period-by-period policy loss function, Lt, is taken for the moment to be a function only of inflation and is given by Lt = (7rt _ ~r) 2 (3) 2 where 7r is the inflation target. An alternative assumption would be to assume that the the price target is specified in terms of the price level, as opposed to the inflation rate. The conventional view is that a price level target entails, on the one hand, a major benefit in that it reduces uncertainty about the future level of prices. On the other, if the economy is subject to (supply) shocks that alter the equilibrium price level, attempts to disinflate and lower the price level back to its pre-shock value may generate significant real costs and increased volatility in inflation and output.6 In practice, as discussed later, all inflation-targeting central banks have opted to define their price objective in terms of the inflation rate; accordingly, it will be assumed in the present that the price target is indeed specified in terms of the inflation rate. The central bank's policy objective in period t is to choose a sequence of current and future interest rates {ih}h=t so as to minimize, subject to (1) and (2), the expected sum of discounted squared deviations of actual inflation from its target value, Ut: minUt = E 6 Lh < 6 < 1, (4) h=t h=t where 6 denotes a discount factor and Et the expectations operator condi- tional upon the central bank's information set at period t. The most direct way to solve this optimization problem is to use dynamic programming techniques. As shown by Svensson (1997b), however, problem 6This argument, however, has been challenged in some recent papers, including Dittmar, Gavin, and Kydland (1999), Svensson (1999a), and Vestin (2000). The lat- ter two studies, in particular, show that under certain conditions price-level targeting may deliver a more favorable trade-off between inflation and output variability than does inflation targeting. 7 (3) can be recast in a simpler form, which allows a more intuitive derivation of the optimal path of the policy instrument. To begin with, note first that, because the nominal interest rate affects inflation with a two-period lag, 7rt+2 can be expressed in terms of period t variables and shocks occurring at periods t + 1 and t + 2. Equation (1) can thus be written as 7rt+2 = 7rt+l + alyt+i + Et+2. Updating (2) in a similar manner and substituting the result in the above expression for Yt+i yields 7rt+2 = (7rt + alYt + E+,) + a,il[3Yt - /2(it - lrt) + '1t+ll + Ct+2, that is 7rt+2 = al7rt + a2Yt - a3it + Zt+2, (5) where Zt+2 = Et+2 + ±t+l + a1?7t+i, a,1 =+ 091±32, a2 = al(l + 01), a3 = a132. FRom (5), it is clear that the interest rate set at period t by the central bank will affect inflation in year t + 2 and beyond, but not in years t and t + 1; similarly, the interest rate set in period t + 1 will affect inflation in periods t + 3 and beyond, but not in periods t + 1 and t + 2; and so on. The solution to the optimization problem described earlier can therefore be viewed as consisting of setting the nominal interest rate in period t (and then t + 1, t + 2, ...) so that the expected inflation in period t + 2 (and then t + 3, t + 4, ...) is equal to the target rate. Put differently, because from (5) ?rt+2 is affected only by it and not by it+,, it+2,**., the problem of minimizing the objective function Ut in (4) boils down to a sequence of one-period problems, 82 min -Et(7rt+2 - T)2 ± Xtr (6) it 2 subject to (5), with xtE { E minhEt [(7rh+2 ] } 8 Because xt in (6) does not depend on it, the central bank's optimization problem at period t consists simply of minimizing the expected, discounted squared value of (7rt+2 - *) with respect to it: 32 min 2 Et(7rt+2 - ft)2. (7) Note that, from standard statistical results,7 Et(7rt+2 _ )2 = (7rt+21t - r)2 + Vt(7rt+2), (8) where 7rt+21t =Et7rt+2. This expression indicates that the central bank's op- timization problem can be equivalently viewed as minimizing the sum of expected future squared deviations of inflation from target (the squared bias in future inflation, (7't+2[t -,Tr)2) and the variability of future inflation condi- tional on information available at t, Vt(7rt+2). Because Vt(7rt+2) is independent of the policy choice, the problem consists in minimizing the squared bias in future inflation. Using (5), the first-order condition of problem (7) is given by 32Et { (7rt+2 - } = -62a3(7rt+21t - 7r) = 0, implying that 7rt+21t =XT (9) Equation (9) shows that, given the two-period control lag, the optimal policy for the central bank is to set the nominal interest rate such that the expected rate of inflation for period t + 2 (relative to period t + 1) based on information available at period t be equal to the inflation target. To derive explicitly the interest rate rule, note that from (5), because Etzt+2 0= , 7rt+21t is given by lrt+21t = alwt + a2yt - a3it, (10) which implies that, given the definition of a1, - -(7rt+21t - 7rt) + ai/327rt + a2yt a3 7This standard result is E(x - X*)2 = (Ex - X*)2 + V(X), that is, the expected squared value of a random variable equals the square of the bias plus the conditional variance. Decomposition (8) will prove useful for the discussion later on of the role of uncertainty. 9 This result shows that, in particular, because interest rate changes affect inflation with a lag, monetary policy must be conducted in part on the ba- sis of forecasts; the larger the amount by which the current inflation rate (which is predetermined up to a random shock, as implied by (1)) exceeds the forecast, the higher the interest rate. The fact that the inflation forecast can be considered an intermediate policy target is the reason why Svensson (1999b) refers to inflation targeting as inflation forecast targeting. The use of conditional inflation forecasts as intermediate targets in the policy rule is optimal, given the quadratic structure of policy preferences.8 The inflation forecast can readily be related to the current, observable variables of the model. To do so requires setting expression (10) equal to ir and solving for it: -? al7rt+ a2yt tt= a3 Given the definitions of the ah coefficients given above, this expression can be rewritten to give the following explicit form of the central bank's reaction function: it = rt + b, (-Ft - ) + b2Yt, (l where 1 b2 1+ Q132 / 2 Equation (11) indicates that it is optimal for the central bank to adjust the nominal interest rate upward to reflect current inflation (to a full extent), the difference between current and desired inflation rates, as well as increases in the output gap. As emphasized by Svensson (1997b, p. 1119), the rea- son why current inflation appears in the optimal policy rule is not because current inflation is a policy target but because it helps (together with the contemporaneous output gap) predict future inflation, as implied by (10). It is also important to note that rule (11) is certainty-equivalent: the same interest rate rule would be optimal in the absence of shocks. Although the central bank cannot prevent temporary deviations of actual inflation from its 'As noted by Bernanke and Woodford (1997), this result does not imply that the central bank should react mechanically to private-sector forecasts. The reason is that there is a risk of "perverse circularity", which stems from the fact that private agents may find it optimal to forecast inflation equal to the announced policy target, depriving thereby their forecasts of any informational value for the central bank. 10 target value, it can ensure that the effects of such shocks do not persist over time.9 In equilibrium, actual inflation in year t + 2 will deviate from the inflation forecast 7rt+2It and the inflation target, 7r, only by the forecast error Zt+2, due to shocks occurring within the control lag, after the central bank has set the interest rate to its optimal value: Tt+2 - 7rt+21t + Zt+2, or 7rt+2-7r - Zt+2- (12) The fact that even by following an optimal instrument-setting rule the central bank cannot prevent deviations from the inflation target due to shocks occurring within the control lag is important in assessing the performance of inflation targeting regimes in practice. 2.2 Policy Trade-offs and Flexible Targeting Consider now the case in which the central bank is concerned not only about inflation but also about the size of the output gap. Specifically, suppose that the instantaneous policy loss function (3) is now given by (7'rt - -R)2 Ay2 Lt = + A8, ) > 0, (13) where A measures the relative weight attached to cyclical movements in out- put.'" The expected sum of discounted policy losses is now given by Ut = Et {Z6h-t [(rh 2) + AYh]} (14) Deriving the optimal interest rate rule when both inflation and output enter the objective function is more involved than was previously the case. Essentially, the problem of minimizing (14) cannot be "broken down" into a 9This, of course, results from the fact that shocks have been assumed to be i.i.d. In practice, however, shocks are often persistent; as discussed later, this may have important implications under parameter uncertainty. 10Note that, because the "bliss level" of the output gap is zero, there is no built-in inflationary bias in this specification; see Cukierman (1992) and the discussion below. 11 series of one-period problems because of the dependence of current inflation on lagged output and the dependence of current output on lagged infla- tion. Using standard dynamic programming techniques, Svensson (1997b, pp. 1140-43) showed that the first-order condition for minimizing (14) with respect to the nominal interest rate can be written as 7rt+2t= - Yt+t, (15) where /c > 0 is given by /c 2 1 - p + F/(1 + /j)2 + 4A/al2} and = A(1 - 6) Condition (15) implies that the inflation forecast 7rt+21t will be equal to the inflation target * only if the one-period ahead expected output gap is zero (yt+llt = 0). In general, as long as A > 0, 7rt+21t will exceed (fall short of) * if the output gap is negative (positive). The reason is that if the output gap is expected to be negative for instance at t + 1, the central bank will attempt to mitigate the fall in activity by lowering interest rates at t (given the one-period lag); this policy will therefore lead to higher inflation than otherwise at t + 2, thereby raising the inflation forecast made at t for t + 2. The higher A (the relative weight on output fluctuations in the policy loss function) is, the larger the impact of the expected output gap on the inflation forecast will be." An alternative formulation of the optimality condition (15) can be ob- tained by noting that, from (1), with EtEt+1 = 0, Yt+ijt = Irt+21t - 7rt+llt Substituting this result in (15) and rearranging terms yields 7rt+21t - *r = C(7rt+llt -*), 0 < C 2 < 1- (16) "The policy loss function (13) can be further extended to account for interest rate smoothing by adding the squared value of changes in it. As shown by Svensson (1997b), an instrument-smoothing objective would make the inflation forecast deviate further from the inflation target-this time to reduce costly fluctuations in interest rates. 12 This expression indicates that the deviation of the two-year inflation fore- cast from the inflation target is proportional to the deviation of the one-year forecast from the target; when A = 0, c = 0 and the previous result (equation (9)) holds. The implication of this analysis is that, when cyclical movements in output matter for the central bank, it is optimal to adjust gradually the in- flation forecast to the inflation target. By doing so, the central bank reduces fluctuations in output. As shown again by Svensson (1997b, pp. 1143-44), the higher the weight on output in the policy loss function is (the higher A is), the more gradual the adjustment process will be (the larger c will be). The interest rate rule can be derived explicitly by noting that, from (1) and (2), t+llt = 2rt + ClYt, 7rt+21t = 7Tt+llt + ±lYt+11t, Yt+1It = lyt- 2(t - 7t)- Substituting the first and third expressions in the second yields 7rt+21t = 7rt + Ce(1 + ±1)yt - a,2(it - 7t). (17) Equating (16) and (17) and rearranging terms implies that it= 7rt + b1(7rt -,r) + b2yt, (18) where bl, c b'2 - ; 0 b' 1 2' 02__ _ from which it can be verified that b' = bi and b' = b2 when A = 0 (and thus c = 0). Equation (18) indicates that the optimal instrument rule requires, as before, the nominal interest rate to respond positively to current inflation and the output gap, as well as the excess of current inflation over the target. However, an important difference between reaction functions (11) and (18) is that the coefficients of (18) are smaller, due to the positive weight attached to cyclical movements in output in the policy loss function.12 This more gradual response implies that the (expected) length of adjustment of current inflation to its target value, following a disturbance, will take longer than the minimum two periods given by the control horizon. The time it takes for expected inflation to return to target following a (permanent) unexpected '2Note also that in both cases the parameters characterizing the optimal policy rule continue to be independent of variances of the shocks affecting inflation and output. This is because certainty equivalence holds in both cases (see the discussion below). 13 shock is known as the implicit targeting horizon or simply as the target horH- zon. Naturally, the length of the implicit target horizon is positively related not only to the magnitude of the shock and its degree of persistence but also to the relative importance of output fluctuations in the central bank's objec- tive function. As can be inferred from the numerical simulations of Batini and Nelson (2000), it also depends on the origin of the shock-whether it is, for instance, an aggregate demand shock or a supply-side shock. This is because the transmission lag of policy adjustments depends in general on the type of shocks that the economy is subject to, and the channels through which these shocks influence the behavior of private agents. A simple illustration of the concepts of control lag and target horizon is provided in Figure 2. Suppose that initially the rate of inflation is on target at * and the output gap is zero. Rkom (11) and (18), under either form of inflation targeting, the initial nominal interest rate is thus equal to -k. Suppose that the economy is subject to an unexpected random shock at t = 0 (an increase in, say, government spending) that leads to an increase in the inflation rate to 7ro > ir. As implied by the reaction function under both strict and flexible inflation targeting, the central bank will raise immediately the nominal interest rate; but, because inflation is predetermined (monetary policy affects inflation with a two-period lag), actual inflation remains at 7ro in period t = 1. The behavior of inflation for t > 1 depends on the value of A. If A = 0 (the central bank attaches no weight to movements in the output gap) inflation will return to its target value at exactly the control horizon, that is, in period t = 2. The nominal interest rate increases initially to io = 7ro + bi(nro - ) and returns to ir at period t = 1 and beyond; the output gap does not change at t = 0 but falls to y, < 0 in period t = 1, before returning to its initial value of 0 at period t = 2 and beyond. By contrast, with A > 0, convergence of inflation to its target value may take considerably longer; the figure assumes, to fix ideas, that convergence occurs at t = 8.13 The interest rate increases initially to io = iro + bl(iro - r) < io, which limits the faUl in the output gap to y < yi. Although falling over time, the interest rate remains above its equilibrium value *r until period t = 6 (given the two-period control lag) whereas the output gap remains negative until period t = 7. In general, the higher A is, the flatter will be the 13With an instrument-smoothing objective in the policy loss function, returning inflation to its target value could take even longer because the central bank is also concerned about large movements in interest rates. Note that, strictly speaking, convergence of actual inflation to target when A > O occurs only asymptotically, for t -* oc. 14 path of inflation, interest rates and the output gap for t > 1. Thus, the central bank's output stabilization goal has a crucial effect not only on the determination of short-term interest rates but also on the speed at which the inflation rate adjusts toward its target after a shock. It can also be shown that policy preferences affect the variability of output and inflation; and in the presence of supply shocks, flexible inflation targeting entails a trade-off between inflation variability and output-gap variability. By varying the relative weight attached by the central bank to the two policy goals in its loss function, it is possible to derive an "optimal policy frontier" (or optimal trade-off curve), which can be defined (following Fuhrer (1997a, p. 226)) as the set of efficient combinations of inflation variability and output variability attainable by policymakers."4 The slope of the output-inflation variability frontier is also related to the slope of the aggregate supply curve (Cechetti and Ehrmann (1999)): the flatter the aggregate supply curve, the larger the increase in output variability that accompanies a reduction in inflation variability. In addition, the higher the relative weight attached to output fluctuations in the policy loss function, the longer it will take for inflation to converge to its target value following a shock. 2.3 Inflation Targeting in an Open Economy In an open economy, the exchange rate is an essential component of the transmission mechanism of monetary policy; it affects the target variables of monetary policy (inflation and the output gap) through a variety of channels. There is a direct exchange rate channel via the impact of prices of imported final goods on domestic consumer prices with, generally, a relatively short lag. There are also two indirect channels, operating through both aggregate demand and aggregate supply. By altering the real exchange rate, the nomi- nal exchange rate affects aggregate demand, typically with a lag (due to the time it takes for consumers to respond to relative price changes); this affects the output gap and, with another lag, inflation. The exchange rate may also affect aggregate supply (with or without a lag), because costs of production may depend on the cost of imported intermediate inputs, whereas nominal wages may depend on (actual or expected) changes in consumer prices caused by exchange rate changes (see Agenor and Montiel (1999, Chapter 8)). In 140f course, the existence of a long-run tradeoff between the variances of output and inflation does not imply a long-run tradeoff between the levels of these variables. In the present setting, such a tradeoff only exists in the short run. 15 turn, the exchange rate is affected by interest-rate differentials, foreign dis- turbances, and expectations of future exchange rates and risk premia that depend on domestic factors, such as the size of the domestic public debt or the degree of credibility of the inflation target. The exchange rate is thus impor- tant under inflation targeting in an open economy, both in transmitting the effects of changes in policy interest rates and in transmitting various distur- bances."5 Because foreign shocks are transmitted through the exchange rate, and the exchange rate affects consumer price inflation, stabilizing exchange rates has remained an important consideration under inflation targeting. These various channels can be captured in a relatively simple generaliza- tion of the closed-economy model presented earlier. Suppose now that the economy produces two goods, tradables and nontradables, with the foreign- currency price of tradables set on world markets. The economy's structure is characterized by the following set of equations: 7tN = Let + ajYt-1 + Et, (19) Yt = -i2(it-1 - 7rt-1) + 03(Aet,-_ 7r1) + n7t /3 > 0, (20) ?rt = 6z7N+ (1-6)Aet, O < 6 < 1, (21) it it + Eet+1 - et + (t, (22) Etet+l = et - O(Aet - 7N), 0 > 0, (23) where et denotes (the logarithm of) the nominal exchange rate, 7T the infla- tion rate in nontradables, i* the world interest rate, and (t an i.i.d. random disturbance. Equation (19) is a Phillips-curve relationship, which is now assumed to hold only for the nontraded good sector. It differs from (1) in two respects: there is no lagged effect of nontradable inflation, and the rate of depreciation of the nominal exchange rate is taken to have a direct and immediate impact on the rate of increase in prices of nontraded goods. As noted earlier, this effect may reflect the supply-side impact of changes in the price of imported "5The effects of interest rates and exchange rates on aggregate demand may also depend on the structure of indebtedness of the economy. For instance, in a country with a large foreign debt, exchange-rate changes may have important wealth and balance sheet effects, possibly offsetting their direct effects on aggregate demand. 16 intermediate goods. Equation (20) is the aggregate demand for nontraded goods; it has a form similar to (2) with two modifications: there is no own lagged effect (01 = 0) and changes in the real exchange rate (as given by the difference between the rate of nominal depreciation and the rate of nontrad- able inflation) are assumed to affect positively the demand for home goods with a lag. Equation (21) defines aggregate inflation as a weighted average of inflation in nontradables and tradables; for simplicity, the world price of tradables is assumed constant so that its rate of change is zero. Equation (22) is the uncovered interest parity condition, which relates domestic inter- est rates to the world interest rate (assumed constant), the expected rate of depreciation of the nominal exchange rate, and a serially uncorrelated ran- dom term. Finally, (23) relates expectations of future nominal depreciation to contemporaneous movements in the real exchange rate: if nontradable inflation is rising faster than the rate at which the nominal exchange rate is depreciating, the current real exchange rate is appreciating; this, in turn, creates expectations of a future nominal depreciation. There are two types of issues that can be explored by studying inflation targeting rules in an open-economy setting. The first is whether the exchange rate channel matters for output stability. To address this issue, suppose that the policy objective is given by (14), which assumes that the central bank targets aggregate inflation, 7rt. Solving the model given by (19)-(23) and (14) using the same dynamic programming approach proposed by Svensson (1997b), it can be shown that inflation targeting can destabilize output in an open economy. The reason is the effect of changes in the nominal exchange rate on inflation through tradable prices. Because it is the fastest channel from monetary policy to inflation in this model, large movements in the exchange rate can produce excessive fluctuations in output by inducing large changes in interest rates."6 Clearly, because the traded and nontraded sectors may react differently in the short run to movements in the (real) exchange rate, the destabilizing effect on aggregate output can be mitigated if the central bank attaches different weights to fluctuations in sectoral output in its objective function (see Leitemo (1999)). In general, however, simulation studies have tended to corroborate this prediction. The second issue that can be addressed with an open-economy model is 16Ball (1999) was one of the first to establish this result. Jadresic (1999) also showed that targeting the overall price level may destabilize output in a model with staggered price setting if policymakers cannot observe current realizations of aggregate output and inflation. The generality of this result, however, is unclear at this stage. 17 whether targeting inflation in nontradable prices only is more appropriate than targeting aggregate inflation. The instantaneous policy loss function given by (13) assumes that the central bank targets aggregate inflation, 7rt. If instead the central bank chooses to target nontradable inflation, its instan- taneous loss function would take the form'7 (L N - N)2 A2 (24) Lt = t2 _.+-2. 24 To analyze this issue, consider for instance a shock unrelated to fundamen- tals that causes a persistent depreciation of the nominal exchange rate-say, a large and sustained outflow of short-term capital due to an adverse shift in confidence (Bharucha and Kent (1998)). The immediate effect is an increase in inflation in the traded goods sector. If, for instance, firms producing home goods use imported intermediated inputs (or if nominal wages are indexed to the overall price level) inflationary pressures will also develop in the non- tradable goods sector and prices there may also rise, compounding the initial increase in tradable prices. Targeting aggregate inflation may involve sub- stantial adjustment in the interest rate and increased volatility in output. By contrast, if the central bank is targeting only nontradable inflation, the adjustment of the interest rate would be less of a lower magnitude, and out- put and nontradable inflation would be less variable-albeit at the cost of greater variability in the nominal exchange rate and aggregate inflation. However, as can be shown by solving the model described earlier using either (13) or (24) as the policy loss function, whether nontradable inflation targeting is strictly preferable to aggregate inflation targeting depends in general on the nature of the shocks hitting the economy, in addition to their relative size. In fact, targeting nontradable inflation may produce undesirable outcomes when the economy is subject to shocks other than to the exchange rate. For instance, in response to demand or supply shocks, a central bank 17A more general specification than (24) would account for the possibility that the central bank is also concerned about large shifts in competitiveness. Its period-by-period policy loss function would therefore look like this, in case of aggregate inflation targeting: (Ir _ -F)2 AyN2 2(et W >O. It is intuitively clear that concerns about real exchange rate fluctuations would also affect the optimal instrument rule-in the sense of making policy changes more gradual than they would otherwise be-as shown earlier when minimizing output fluctuations was introduced as an additional policy objective. See Svensson (1999b) for a discussion. 18 with a nontradable inflation target is likely to attempt to restore inflation on its targeted path rapidly. This would occur through large adjustments in the interest rate-which would entail greater volatility in the exchange rate and aggregate inflation. In sum, whereas an aggregate inflation target may induce excessive volatil- ity in the interest rate (and thus output) to offset exchange rate shocks, a nontradable inflation target may induce excessive volatility in the exchange rate as the policy instrument is adjusted to offset supply or demand shocks. Indeed, in the simulation results presented by Bharucha and Kent (1998), neither aggregate inflation targeting nor nontradable inflation targeting pro- duced consistently lower volatility in both product and financial markets across all types of shocks. 3 Comparison with Intermediate Target Strate- gies Price stability as a medium- to long-term goal can be achieved, in principle, not only by focusing directly on the final objective itself, the inflation rate or the price level, but also by adopting either a pegged nominal exchange rate or a monetary target as an intermediate goal. This section reviews these two alternative frameworks for monetary policy and compares them with inflation targeting. 3.1 Monetary vs. Inflation Targeting Monetary targeting presumes the existence of a stable relationship between one or more monetary aggregates and the general level of prices. When this is the case, monetary policy can be directed at a particular rate of growth in the monetary aggregate (the intermediate objective) compatible with low inflation. Specifically, monetary targeting requires adequate knowledge of the parameters characterizing the demand for money. In an economy undergoing rapid financial liberalization, however, these parameters (notably the interest elasticity of money demand) may be highly unstable. In such conditions money ceases to be a good predictor of future inflation; that is, the relation between the intermediate target and the final objective becomes unstable. Similarly, in a context of disinflation, the demand for money may be subject to large and unpredictable shifts; as a consequence, the information content 19 of money for future inflation will be very low. Both arguments suggest that relying on monetary aggregates can be potentially risky. In addition, suppose that monetary targeting is viewed as minimizing money growth variability around the money-growth target-a characterization that is fairly adequate if the policy loss is quadratic. As shown by Svensson (1997b), this policy goal may be in conflict with the objective of minimizing inflation variability; that is, there often is a conflict between stabilizing inflation around the inflation target and stabilizing money growth around the monetary target. In fact, monetary targeting will in general imply greater inflation variability than inflation targeting. By inducing higher volatility in interest rates, it also leads to increased variability in output (Clarida, Gaif, and Gertler (1999))." Several industrial countries have indeed adopted inflation targeting after abandoning (or being abandoned by) their monetary targets due to increased distortions in the link between the money supply and overall prices, as doc- umented for instance by Estrella and Mishkin (1997). 9 It is worth noting, however, that although some researchers have argued that the relationship between monetary aggregates and prices has also weakened in developing countries (see for instance Mishkin and Savastano (2000, p. 22) for Latin America) systematic formal evidence on this issue remains limited (particu- larly for the late 1990s) and subject to different interpretations. The study by Arrau, De Gregorio, Reinhart and Wickham (1995), for instance, showed that the alleged instability in money demand documented in several studies focusing on developing countries during the 1980s may well have been the result of an omitted variable, namely financial innovation. 3.2 Exchange Rate vs. Inflation Tairgeting Many countries (particularly in the developing world) have viewed pegging their nominal exchange rate to a stable low-inflation foreign currency as a means to achieve domestic price stability, through a "disciplining mechanism" '8See McCallum (1999) for a further discussion of the lack of efficiency of monetary targeting. "9It has also been argued that, in practice, the lack of stability and predictability in the assumed relationships between interest rates and the target monetary aggregate, and between the target aggregate and inflation, have been well recognized in those countries that have pursued monetary targeting. See, for instance, the discussion of German mone- tary policy in Bernanke et al. (1999). Studies of the reaction function of the Bunbesbank suggest also that real variables have had a significant influence on policy decisions, in addition to monetary variables. See Clarida, Galf and Gertler (1998a). 20 with two dimensions. First, to the extent that higher domestic relative to foreign inflation results in a real exchange rate appreciation, the demand for domestic goods would fall and induce a cyclical downswing that would put downward pressure on domestic prices. Second, to the extent that wage- and price-setting decisions anticipate these consequences of wage and price increases being too high, they would make higher domestic inflation less likely to occur in the first place. In a sense, countries that target their exchange rates (against an anchor currency) attempt to "borrow" the foreign country's monetary policy credibility. However, the experience of recent years has shown that in a world of high capital mobility and unstable capital movements, conventional pegged exchange rates have proved fragile (see Agenor and Montiel (1999)). Most im- portantly, simply pegging the exchange rate did not prove to be a substitute for maintaining monetary stability and credibility at home. In fact, recent experiences suggest that exchange rate pegs can be sustainable only when they are credible, and credibility is to a large extent determined by domestic macroeconomic policies. FRom that perspective, an inflation targeting regime may operate better than an exchange rate targeting framework. It may even be argued that, to the extent that the domestic currency in many develop- ing countries has been attacked because the central bank had an implicit or explicit exchange rate objective that was not perceived to be credible, the adoption of inflation targeting may lead to a more stable currency if it signals a clear commitment to macroeconomic stability and a freely-floating exchange rate. It is worth emphasizing that a key characteristic of inflation targeting regimes compared to other approaches to controlling inflation is that the ad- justment of policy instruments relies on a systematic assessment of future (rather than past or current) inflation, as opposed to an arbitrary forecast. Under this regime, the central bank must explicitly quantify an inflation tar- get and establish precise mechanisms to achieve this target. This implies that there is an important operational difference between an inflation targeting regime, on the one hand, and monetary and exchange rate targeting, on the other.20 Changes in monetary policy instruments usually affect the money 20Note also that there is an important difference between exchange rate targeting and monetary targeting, in the sense that while it is possible to deviate temporarily from monetary targets if the underlying relationships appear to have changed, it is generally not possible to depart temporarily from an exchange rate peg (or a target band, for that matter) without there being a loss of credibility and possibly a currency crisis. 21 supply and the exchange rate faster than inflation itself; as discussed earlier, this leads to the existence of a control lag and a reaction function that relates the policy instrument to an inflation forecast. The implication, as pointed out by Haldane (1998), is that the credibility of an inflation targeting regime depends not on achieving a publicly-observable, intermediate target that is viewed as a leading indicator of future inflation (as is the case under mone- tary or exchange rate targeting), but rather on the credibility of a promise to reach the inflation target in the future. This in turn depends on whether the public believes that the central bank will stick resolutely to the objective of price stability. Credibility and reputation of the monetary authorities may play therefore an even more crucial role in dampening inflation expectations under inflation targeting. At the same time, because performance can only be observed ex post, the need for transparency and accountability becomes more acute under inflation targeting, in order to help the public assess the stance of monetary policy and determine whether deviations from target are due to unpredictable shocks rather than policy mistakes. 4 Basic Requirements for Inflation Tirgeting There are three basic requirements for implementing an inflation targeting regime. The first is a high degree of central bank independence (not so much in choosing the inflation target itself but rather in the choice and manipula- tion of policy instruments), the second is the absence of a de facto targeting of the nominal exchange rate (or, equivalently, the predominance of the in- flation target), and the third is increased transparency and accountability. 4.1 Central Bank Independence and Credibility Inflation targeting requires that the central bank be endowed by the political authorities with a clear mandate to pursue the objective of price stability and most importantly a large degree of independence in the conduct of monetary policy-namely, in choosing the instruments necessary to achieve the target rate of inflation.21 This implies, in particular, the ability to resist political 21Several countries (such as Israel and the United Kingdom) have followed a contractual approach to inflation targeting; the government sets an inflation target in a contract with the central bank, and gives the central bank operational independence so that it can manipulate its policy instruments to achieve the agreed target. 22 pressures to stimulate the economy in the short term and the absence of "fiscal dominance," that is, a situation in which fiscal policy considerations play an overwhelming role in monetary policy decisions. In countries where systematic reliance on seigniorage as a source of revenue is high (a situation quite common in some developing countries where government borrowing from the central bank is large) such requirements are difficult to satisfy.22 In such conditions, fiscally-induced inflationary pressures will undermine the effectiveness of monetary policy by forcing, for instance, the central bank to maintain low interest rates in an attempt to prevent unsustainable public debt dynamics. Alternatively, as discussed by Beddies (1999), the central bank can be forced by the government to adopt an inflation target that is dictated by seigniorage requirements as opposed to price stability.23 Inflation targeting calls not only for a high degree of central bank inde- pendence, but also for a sufficient degree of credibility-or more properly (following Drazen and Masson (1994) and Agenor and Masson (1999)) an adequate anti-inflation reputation. Independence, credibility and reputation are of course related but may evolve differently over a given period of time. In countries where the financial system is perceived to be highly vulnerable (to, say, exchange rate shocks, as discussed below) and the central bank is perceived to be likely to inject liquidity to prevent a full-blown crisis, the credibility of an announced inflation may be seriously undermined-even if the central bank is deemed independent. Lack of confidence in the policy- makers' commitment to (or ability to maintain) low inflation may be one of the reasons why inflation tends often to display a strong degree of persis- tence in developing countries, as illustrated by the auto-correlation functions for twelve developing countries (with the exception of India, Korea and the Philippines) displayed in Figure 3.24 But establishing credibility or improv- ing reputation, particularly in countries with a history of high inflation and macroeconomic instability, is a difficult process. Analytically, it has been 22The ability of the central bank to conduct an independent monetary policy is also hampered in some countries by severe weaknesses in the financial system, which may force the central bank to inject repeatedly large amounts of liquidity to support ailing banks. 23Note that, in practice, it has proved difficult to establish a close and stable relationship between fiscal deficits and inflation in developing countries. As discussed by Agenor and Montiel (1999), this may have been the result of changes in deficit financing rules or shifts in expectations over time. 24f course, inflation persistence may also be the result of backward-looking inflation expectations or overlapping and asynchonized wage and price contracts. 23 shown most notably by Walsh (1995) and Svensson (1997a)-that inflation targets can be used as a way of overcoming credibility problems because they can mimic optimal performance incentive contracts; and by increasing the accountability of monetary policy, inflation targeting may reduce the in- flation bias inherent in discretionary policy regimes. Moreover, as argued by Walsh (1999), the public announcement of inflation targets may itself help to improve the credibility of the central bank when its policy preferences are uncertain. It is important to keep in mind, however, that the link between inflation performance and the degree of de jure central bank independence does not appear to be particularly strong in developing countries-at least for the 1980s-as illustrated in Figure 4 for different measures of independence and as shown more formally in some recent research (see for instance Sikken and De Haan (1998), and De Haan and Kooi (2000)). What matters, in any case, is de facto independence. To the extent that, for instance, a lack of actual autonomy translates into uncertainty about the central bank's preferences over output and inflation, delegating monetary policy involves a trade-off between credibility and stabilization and an optimal contract can perform better than an inflation target (Beetsma and Jensen (1998)).25 4.2 Absence of de facto Exchange Rate Targeting Adopting a low and stable inflation rate as the main objective of monetary policy requires in principle the absence of any commitment to a particular value of the exchange rate, as is the case under a floating exchange rate regime. In practice, however, in many of the developing countries that have opted for a de jure flexible exchange rate, monetary authorities have contin- ued to pay considerable attention to the value of the domestic currency- often adopting a de facto target path or band. There are various reasons for the central bank to be concerned with nominal exchange rate movements, even when its degree of independence (and thus its ability to commit itself only to the pursuit of price stability) is high. As noted earlier, the exchange rate has a direct impact on inflation and plays a key role in transmitting monetary policy shocks to prices. If the pass-through effect is indeed high, the central bank may be tempted to intervene on the foreign exchange market 25However, Muscatelli (1998, 1999) has argued that neither inflation targeting nor an optimal contract is likely to be superior to Rogoff-type conservatism if central bank pref- erences are uncertain. 24 to limit currency fluctuations. A high degree of nominal exchange rate insta- bility may also be of concern to policymakers to the extent that it translates into a high degree of variability in the real exchange rate and distorted rel- ative price signals to domestic producers. Another important reason is that in (partially) dollarized economies (like Peru or Turkey for instance) large fluctuations in exchange rates can lead to banking and financial instabil- ity by inducing large portfolio shifts between domestic- and foreign-currency denominated assets. Finally, in countries where the corporate and banking sectors hold large foreign-currency liabilities, exchange rate depreciations can have significant adverse effects on their balance sheets. This was, indeed, one of the important features of the Asia crisis.26 When limiting (or preventing) exchange rate fluctuations is a stated or an implicit policy target, it will be usually very difficult for the central bank to convey to the public its intention to give priority to price stability over other objectives of monetary policy in a credible and transparent manner. Private agents are likely to discount heavily public pronouncements; and the lack of credibility will translate into higher inflation expectations. Thus, the absence of (implicit or explicit) commitment to a particular level for the exchange rate-or, equivalently, giving the inflation target unambiguous pri- ority over other policy objectives-is an important prerequisite for adopting inflation targeting. In fact, a credible commitment to an inflation targeting regime in developing economies, by enhancing macroeconomic and financial stability, may well provide a greater degree of stability to a flexible nominal exchange rate than a pegged arrangement that is subject to recurrent spec- ulative pressures (and possibly frequent crises and forced devaluations) due to perceived inconsistencies in macroeconomic policy. 4.3 lransparency and Accountability Openness and transparency in the conduct of monetary policy are important ways to improve credibility in an inflation targeting framework. By mak- ing the central bank publicly accountable for its decisions, they raise the incentive to achieve the inflation target and enhance therefore the public's confidence in the ability of the monetary authorities to do so. And by expos- 26See, for instance, Alba et al. (1999). It should be noted, however, that these last two points can also be viewed as calling for adequate regulation and supervision of the domestic banking system, not necessarily as arguments in favor of a rejection of inflation targeting as a policy regime. 25 ing to public scrutiny the process through which monetary policy decisions are taken, they may lead to improved decision-making by the central bank and enhanced credibility (see Briault, Haldane and King (1997)). The fact, for instance, that monetary authorities must announce policy changes and explain the reason for these changes to the public may increase the effective- ness of monetary policy under inflation targeting. By reducing uncertainty about the central bank's preferences, transparency may lead to a lower ex- pected rate of inflation and a lower propensity to respond to supply shocks (see Eijffinger, Hoeberichts, and Schaling (2000)).27 For instance, Faust and Svensson (1998) examined the role of trans- parency in a model similar to the one developed by Cukierman and Meltzer (1986). The central bank in the model is tempted to deviate from an an- nounced inflation target due to fluctuations in an idiosyncratic component of its employment target. The employment target is private information to the central bank and unobservable to the public. Private agents, nev- ertheless, observe macroeconomic outcomes (contemporaneous inflation and employment) and imperfectly infer the central bank's employment target. This inference process affects the central bank's perceived reputation by the public, which in turn affects private inflation expectations. In this setting, increased transparency allows the public to infer the bank's employment tar- get with greater precision, thereby rendering the central bank's reputation and the private-sector inflation expectations more sensitive to its actions. This, in turn, increases the cost for the central bank of deviating from the announced inflation target and pursuing its idiosyncratic employment tar- get. Consequently, increased transparency induces the central bank to stick more closely to the announced policy. It provides an implicit commitment mechanism that reduces the temptation by the monetary authorities to act in a discretionary fashion and deviate from the announced policy. A potential problem with accountability in an inflation targeting frame- work is related to the difficulty of assessing performance only on the basis of inflation outcomes. The reason is that (as indicated earlier) there is a lag between policy actions and their impact on the economy; it is thus possible (or tempting) for the central bank to blame "unforeseen" or totally unpre- dictable events for inadequate performance, instead of taking responsibility for policy mistakes. To mitigate this risk, in inflation-targeting countries 27This, of course, assumes that the public is familiar with the inflation target and the specific price index upon which the target is based; see the discussion below. 26 the central bank is required usually to justify its policy decisions and explain publicly differences between actual outcomes and inflation targets.28 In what is arguably the strongest case of accountability of a central bank in inflation- targeting countries, the government in New Zealand can dismiss the central bank governor if the inflation target is breached, even temporarily (see Archer (2000)). In practice, openness and transparency have been promoted by the regular publication of an Inflation Report, which sets out the central bank's analysis of recent economic developments, and a forecast of inflation (as well as other variables, including output) over the coming year or years. In the words of Archer (2000, p. 16): It is harder for the central bank to "cheat" on its mandate when it is forced to lay out an internally consistent basis for the decisions to be made. To be sure, a good publicist can make almost any position sound reasonable, but when it matters financial markets seem to have good noses for spin-doctoring. Accountability has been promoted by providing public explanations (in the form of a public letter from the governor of the central bank to the gov- ernment) of the reason(s) why the rate of inflation deviated from the target by more than a given percentage on either side, how long these deviations are expected to persist, and what policies the central bank intends to implement to bring inflation back to target. 5 The Operational Framework of Inflation Tar- geting Establishing an operational framework for implementing an inflation target- ing regime requires monetary authorities to take various steps. They must, in particular, specify a price index to target and quantify the target. They must also determine under what circumstances escape clauses or exemp- tions to the inflation target are warranted (taking into account the potential credibility loss that their discretionary nature may involve), decide ways through which they can convince the public that achieving the inflation tar- get takes precedence over all other objectives of monetary policy and devising a forward-looking operating procedure in which monetary policy instruments 28The distinction between goal independence and instrument independence, alluded to earlier, implies that the latter is essential for accountability of the monetary authorities. 27 are adjusted (in line with the assessment of future inflation) to achieve the target. This section reviews these various requirements, with a particular emphasis on their implications for developing countries. 5.1 Establishing Inflation Targets Establishing inflation targets requires the central bank, in particular, to spec- ify an index to target, set a level for the target, and establish an explicit quantitative target for inflation for some periods ahead-a process that re- quires deciding whether to set the target as a point or as a band (for instance, between 2 and 4 percent) or as a point, and choosing the number of quarters or years over which the target will be averaged. They must also set up a model or methodology for inflation forecasting that uses indicators contain- ing information on future inflation. 5.1.1 The choice of a price index How to measure inflation is naturally a key issue in the context of infla- tion targeting. A first question to ponder in this context is whether the price index chosen should reflect the prices of goods and services for current consumption only, or prices relevant for both current and future consump- tion. Economic theory suggests that a correct measure of inflation is an index that captures prices of both present and future consumption, because only such an index can recognize the scope for intertemporal substitution in consumption (Alchian and Klein (1973)). In practice, however, price in- dices do not include prices of present claims over goods and services for future consumption-except indirectly when they incorporate prices of ser- vices provided by consumer durables that can be acquired today. In most cases, markets for future goods and services do not exist, and producing a comprehensive price index as theory would suggest is simply not feasible.29 A second question that must be considered is related to the fact that con- sumer price indices typically found in developing countries are not adequate measures of the cost of living, despite several advantages (general acceptance 291n addition, using asset prices (which can be viewed as reflecting the current money prices of claims on future consumption) is fraught with practical difficulties. Note, however, that excluding asset prices from the price index used for inflation targeting does not necessarily preclude the possibility of taking these prices into account in the formulation of monetary policy; see the discussion below. 28 among the public, availability on a monthly basis, publication without long delays, and infrequent data revisions, unlike, say, GDP deflators). In almost all of these countries, consumer price indices are typically fixed-weight aver- ages of prices of a "basket" of goods and services. Price data are gathered by statistical agencies through regular visits to specified sales outlets. The weights are based on the distribution by product of consumer expenditure and are revised at regular intervals as the composition of household expendi- ture changes. These indices are not designed to measure the effect of changes in the types, amounts and quality of the goods and services households pur- chase. Specifically, in measuring inflation, existing price indices suffer from three main sources of bias, whose relative importance is of course likely to vary across countries: * Substitution bias, which arises because consumers can be expected to switch expenditure away from more expensive products toward cheaper products, and away from more expensive stores toward cheaper ones. Regular reviews of the weights in the index and of the selection of outlets chosen for sampling purposes helps to reduce substitution bias, but the bias is likely to persist in the periods between these reviews. * Quality bias, which arises because the quality of goods and services within a particular category may change, but the change may not be fully reflected in computing the price index. If quality increases over time (as tends to be the case for many products), failure to account for it fully will lead to overstatement of inflation. * New products bias, which stems from the fact that new goods are con- tinually introduced whereas existing ones are continuously witdrawn. To the extent that new products widen choice opportunities for house- holds and allow them to achieve a given level of utility at a cost that may be lower than previously, failure to account for them will lead to an overestimate of inflation. These sources of bias can be quite significant in practice. For instance, the Boskin Report (see Boskin et al. (1996)) estimated that, in the United States, substitution bias (including outlet substitution bias) was about 0.5 percent a year in terms of CPI inflation whereas quality and new products bias represented about 0.6 percent per annum in 1995-96. In general, how- ever, accounting for bias can be very difficult. One reason is that estimates 29 of bias tend to be very unreliable; another is that the intensity of bias may vary significantly over time.30 Inflation targets are thus generally set with some degree of arbitrariness on the basis of price indexes that suffer from potentially significant measurement problems. In addition, particularly in developing countries, there is a considerable amount of transitory noise that makes the measurement of inflation difficult. In many of these countries (as in most industrial countries), the consumer price index is a Laspeyres index, that is, an arithmetic mean of the relative prices of the goods and services contained in a representative basket of goods, this mean being weighted ac- cording to expenditure shares in a base period.3" The annual inflation rate is thus measured as the weighted average of the rates of change for individ- ual prices. Thus, if the price of a particular component is subject to large short-term movements, then the measured rate of overall inflation will also reflect the corresponding movements-to an extent that reflects the relative weight of that component in the consumption basket. In countries where domestic food prices (which can be highly volatile) are a large component of the consumption basket, such problems can be significant. Figure 5 illus- trates the high correlation between variability in food prices and variability in consumer prices in developing countries. Shocks to the exchange rate, wages, controlled prices and external inflation that affect price developments can also be beyond the control of monetary authorities. To alleviate the problem of noisy or erratic short-run movements in prices, inflation-targeting central banks have often used a measure of core (or long- terrn) inflation as their policy target.32 Such measures are generally calcu- 30See Gordon (1999, p. 17), who suggested that estimates of inflation bias published in the Boskin Report (of which he was a co-author) were probably understated, in part because the sharp declines in the prices of new products were not fully accounted for. Cellular phones, for instance, appeared in 1985 in the United States as a consumer product but was introduced in the CPI only in 1998-by which time there were already 55 million subscribers to cellular phone services in the country. Between 1985 and 1997, the price of these services dropped dramatically. 31Formally, if p' denotes the price of good iat period t (i = 1,...n), qt the quan- tity of good i purchased at period t, the Laspeyres price index between periods tand t±+ 1 isP~t/t+1 = 1pt+1qt /i=lptqt. This index tends to overstate inflation, be- cause it assumes no substitution in consumption expenditures in response to changes in relative prices. By contrast, a current-weighted Paasche index, defined as P = t/t+1 Ei=1 Pt+lq+1 / Eni=lptqt+1, uses current quantities as weights and thus tends to over- state substitution and to understate inflation. 32See Roger (1998) and Wynne (1999) for recent reviews of the literature on core infla- 30 lated by removing some components of the price indexes, particularly those that appear to be substantially more volatile than prices of other goods and services. The argument is that short-term movements in these prices result from rapid adjustment to frequent real shocks that are often reversed, and so they contain substantially less information about the long-term trend in inflation. By choosing an adjusted price index, instead of the general index of consumer prices, central banks may thus avoid unnecessary fluctuations in interest rates, employment and output. In general, there are three types of price disturbances that inflation- targeting central banks have attempted to remove, the first two being par- ticularly relevant for developing countries: * Disturbances associated with temporary changes in the rate of infla- tion. Examples are movements in highly volatile, but over time offset- ting, CPI components (such as food and energy prices), shocks to the exchange rate, and changes in indirect taxes or regulated prices that lead to jumps in the price level, without sustained, longer-term effects on the inflation rate. o Shocks to inflation with possibly long-lasting effects, but where a strong monetary response would produce very large fluctuations in the real economy. Examples include large changes in prices of imported inputs or natural disasters. . Disturbances associated with the direct effects of policy changes on inflation, in cases where the targeted price index contains components (such as mortgage payments) that vary directly with the level of short- term interest rates. The reason is that a tightening in monetary policy may translate into a spurious rise in inflation. The identification of these various types of disturbances makes it possible to distinguish, as proposed by some economists, between core inflation (as defined above), cyclical inflation (due for instance to food prices), and tran- sitory inflation (price increases generated by specific shocks, such as changes in indirect taxes and imported energy prices, or shocks to the exchange rate and controlled prices). In practice, there are several techniques for calculating core inflation (see Wynne (1999)). One method, used by the Bank of Canada, consists of using tion. 31 weights that are linked to the relative price variability of individual goods and services in the consumption basket; each expenditure weight is divided by the standard deviation of its relative price change (calculated as the difference between the inflation rate of the sub-component and the overall index) and they are all renormalized (using a single- or double-weighted method) to ensure that they add up to unity. The higher the standard deviation (that is, the higher the volatility) of any given component, the lower its adjusted weight. In this way, all components are kept in the basket but the effect of the more volatile ones on the overall index is reduced. Nevertheless, the practical difficulties involved in the calculation of ad- justed measures of underlying inflation should not be under-estimated.33 In particular, truly temporary shocks are rare in practice and characterizing them as such can be the result of an incorrect assessment of general equilib- rium effects. Consider, for instance, a rise in indirect taxes that lowers the rate of inflation during a limited period of time, after which inflation returns to its previous level. Such changes in taxes will generally also affect the rest of the economy; they may reduce aggregate demand and production, which could have an additional downward effect on inflation in the longer term. They may also alter inflation expectations or wage settlements and thereby price-setting decisions. These "second-round" effects may be protracted and must be taken into consideration when establishing an inflation forecast. The result may be that the proper response to an increase in indirect taxes is not only to refrain from increasing interest rates, but indeed to lower them, if the inflation forecast has fallen below target. A similar argument can be made in relation to temporary energy price shocks. Mechanically excluding certain components of the consumer price index from the policy reaction function can translate into a loss of potentially important information regarding future inflation developments. It is also worth noting that there is a potential for conflict or uncertainty resulting from differences in price movements between inflation in the to- tal CPI and adjusted inflation. If the target is not easy for the public to understand and follow, these differences can create credibility problems, par- ticularly in countries with a history of high inflation and lack of reputation of the central bank (see the discussion below of the Brazilian experience). 33In particular, it should be noted that there remains some disagreement in the literature regarding the very definition of core inflation. Quah and Vahey (1995), for instance, define core inflation as the component of inflation that does not affect real output in the long run. They propose a structural VAR approach to calculate it. 32 In such conditions, a headline inflation target, being more familiar to the public, would make monetary policy more transparent-even if it is subject to some defects in measuring underlying price pressures. 5.1.2 Width of the target band Many inflation-targeting countries have chosen to specify their inflation tar- get in the form of a band (sometimes around a central value) rather than a single point target. A key reason for doing so is the recognition that, be- cause economic behavior and outcomes are not completely predictable, it is difficult to hit continuously a single point target for inflation. Trying to do so could cause interest rates to be excessively volatile and could destabilize financial markets. By recognizing uncertainty, a target band for inflation can stabilize price expectations. At the same time, a band introduces some de- gree of flexibility in the conduct of monetary policy. It gives the central bank some discretion over which point in the band it aims for in taking its policy decisions and allows it to accommodate transitory shocks to inflation-and this ability can provide a partial substitute for targeting a core inflation rate (Orphanides and Wieland (2000, p. 1383)). This flexibility is important not only to stabilize output fluctuations but also to accommodate movements in the nominal exchange rate. A large depreciation, for instance, because it puts upward pressure on domestic prices (as discussed earlier) would require a rise in policy interest rates that may be sizable with a point target. But the depreciation may weaken corporate balance sheets if firms are highly in- debted in foreign currency and push them into bankruptcy; the rise in interest rates would only exacerbate the problem. On the contrary, with a sufficiently wide band, the inflationary impact of exchange rate market pressures could be accommodated (at least to some extent) without large and potentially destabilizing policy changes. The foregoing discussion suggests that the width of the band should de- pend on the variability of the rate of inflation (which itself depends on the nature, size and variability of underlying shocks to prices and changes in behavioral patterns), the policy horizon (the further ahead the horizon, the more difficult it is to forecast inflation and predict the effects of monetary policy actions), as well as the degree of discretion that the central bank main- tains in the conduct of monetary policy. FRom the latter perspective, there is obviously a trade-off between credibility and flexibility in the choice of the optimal target band: the wider the bands are, the more likely it is that the 33 target will be achieved but the less credible the target will be.34 Another issue is whether the band should be asymmetric or not. In practice, inflation-targeting countries have tended to adopt symmetric bands upon announcement of a central target. The "signal" that central banks have attempted to convey in doing so is that they view deflation as being as costly as inflation. There are cases, however, where asymmetry may be desirable. If, for instance, the central bank believes that the credibility loss associated with overshooting the target is higher than the loss incurred by undershooting it, an asymmetric band may provide a somewhat higher degree of flexibility.35 5.1.3 Horizon of the inflation target As discussed earlier, because inflation is subject to unpredictable shocks and monetary policy operates with a lag, inflation-targeting central banks can- not achieve continuously their target. At what time horizon, then, should monetary policy aim to achieve the inflation target? Suppose, for instance, that the inflation forecast indicates that with un- changed interest rates inflation will rise above target in the first year of the forecast period and fall below target in later years. If interest rate changes implemented today affect inflation only after a lag of (say) two years, then it is too late to prevent a rise in inflation above target in the first year and interest rates should be lowered so as to bring inflation in the second year and beyond up to target. But suppose now that the control lag is only six months. Increasing interest rates today may therefore allow the central bank to bring inflation down toward target in the second half of the first year of the forecast. At the same time, however, the increase in interest rates today is likely to reduce inflation even further below the forecast at the beginning of the second year; and if the lag between policy actions and their effect on inflation are distributed over a period of time, rather than being concentrated at a particular lag length, inflation will remain below forecast for at least part of the second year. The central bank has to use interest rates to manage the 34In some sense, the choice of an optimal inflation target band is reminiscent of the problem of choosing the optimal width of an exchange rate target zone, as discussed for instance by Cukierman, Kiguel, and Leiderman (1994). See, however, Orphanides and Wieland (2000) for a different perspective. 35As discussed later, asymmetric policy preferences may also lead to an asymmetric optimal instrument rule, regardless of whether a formal band is in place or not. 34 path of inflation relative to the inflation target, starting from the point in the future at which monetary policy decisions taken today start affecting inflation. Thus, as noted earlier, the target horizon cannot be shorter than the control horizon. At the same time, the central bank needs to take into account the trade- off that may arise between reducing deviations of inflation from target and preventing a high degree of variability of output. Suppose that there has been an adverse supply shock which causes inflation to increase above tar- get. How quickly should the central bank aim to bring inflation back to target? As discussed earlier, there are two conflicting considerations (see Figure 2). On the one hand, the quicker the disinflation, the shorter the period during which inflation is above target; on the other, the quicker the disinflation, the greater the potential fluctuations in output. The optimal length of the target horizon depends therefore also on policy preferences, in addition to the magnitude of the shocks affecting inflation. The greater the relative importance attached to output stabilization relative to inflation sta- bilization, the longer the optimal length.36 In addition, however, if there is an inflationary bias, a longer targeting horizon will hamper the credibility of monetary policy, because agents will understand the central bank's greater temptation to inflate and create surprise inflation. FRom that perspective, there is also a trade-off between credibility and flexibility (Garfinkel and Oh (1993)). Whether or not a finite-period targeting procedure is desirable de- pends then on the cost of limiting flexibility so as to reduce the severity of the credibility problem. 5.1.4 Forecasting procedure Inflation targeting requires the monetary authorities to have the technical and institutional capacity to model and forecast inflation, understand the monetary transmission mechanism, and estimate the time lag between the adjustment of monetary instruments and their effect on output and prices. This, in turn, implies greater reliance on forward indicators of inflation than in other monetary policy regimes and a continuous assessment of the rela- tionship between the instruments of monetary policy and the inflation target. 36fSee Batini and Haldane (1999) and Isard and Laxton (2000) for recent quantitative research on policy tradeoffs of this type. Note also that, in practice, stabilizing movements in output can also be achieved by a temporary widening of the target band, if there is one in place. 35 Establishing an accurate forecast of inflation and a relatively precise es- timate of how inflation is likely to be affected by changes in monetary policy instruments is not an easy task when the structure of the economy is unstable and behavioral relationships unreliable.37 Moreover, the higher the level of inflation, the more volatile the inflation rate tends to be (as shown in Figure 6) and thus the more difficult it is to predict accurately future price devel- opments. In such conditions, setting point targets may actually damage the central bank's credibility. In practice, inflation-targeting central banks have used both economic models (generally small in size) and qualitative judge- ment regarding how economic behavior and relationships are likely to differ in the future from what has been observed in the past. Moreover, they have increasingly recognized in their inflation reports the uncertainty surrounding economic forecasts, that is, the fact that forecasts embody probability dis- tributions of future values of economic variables. The Bank of England for instance, and many other central banks in its wake-including those of Brazil, Chile, and Sweden-have used a "fan chart" to present their forecasts (see, for instance, Blix and Sellin (1998)). By de-emphasizing, or even not pro- viding, point forecasts, the inherent uncertainties associated with forecasting can be communicated to the public. It should be noted, however, that releasing an inflation forecast does not necessarily enhance the transparency of the decision process of monetary pol- icy; repeated forecasting errors can eventually reduce the credibility of the central bank by raising doubts about its ability to assess the future behav- ior of prices.38 Another issue relates to how best to establish and release forecasts to the public. Many inflation-targeting central banks-with the exception of the Bank of England, which recently began releasing inflation projections based on market expectations of future interest rates-publish forecasts based on the assumption that policy interest rates will remain con- stant over the projection horizon. To the extent that the predicted path for inflation is too high relative to target, this gives a signal that interest rates will be raised in the near future-which may influence economic conditions, 371n many developing countries, administered or controlled prices are an important component of aggregate price indices and thus important factors in the short-run behavior of inflation. A proper inflation forecasting procedure needs also to incorporate explicit assumptions about the timing and magnitude of changes in these prices. 38Releasing inflation forecasts can also be problematic if they have an undesirable effect on mominal wages, for instance by becoming a floor for centralized wage negotiations. 36 potentially making the forecasts internally inconsistent.39 Moreover, basing the inflation forecast on endogenous monetary policy responses would make the forecast itself irrelevant, because it would always be on target at the horizon over which the central bank acts to achieve its goals. Public interest would therefore shift to the implied sequence of policy rates embedded in the forecasts. However, central banks may not be willing to release projections of future short-term interest rates, because this may run the risk that any deviation of policy rates from the forecast level (even if caused by new in- formation) could be perceived as inconsistent behavior by the central bank and harm its credibility. Finally, information regarding the expected path of interest rates is not necessarily useful to the public unless it is specified how short-term interest rates are determined in the first place. Knowledge of the central bank's policy reaction function is essential to determnine the implications of new information for the setting of policy instruments. 5.2 Interest Rates Rules in Practice As shown earlier, under "strict" inflation forecast targeting the optimal rule is to set the interest rate so as to bring expected inflation in line with the inflation target at the control horizon. By contrast, under "flexible" inflation forecast targeting, the optimal rule is to close less than fully any gap between expected inflation and the inflation target. In practice, however, inflation- targeting central banks have tended to use simple feedback interest rate rules rather than complicated, optimal targeting rules. Simple feedback rules have some clear advantages. First, they tend to be more robust when there is uncertainty about the true structure of the economy; second, the may be advantageous on monitorability and credibility grounds. This is important in the context of inflation targeting. However, simple rules are also subject to one major weakness: they tend to use information (notably about the economy's structure) less efficiently than an optimal rule. Two additional issues to consider in this context are, first, whether simple rules display evidence of "flexible" inflation targeting; and, second, whether they embed forward-looking components. Regarding the first issue, some recent research has indeed attempted to determine whether the behavior of inflation-targeting central banks can be characterized as "strict" or 'flexible" 39Market observers may also wonder why the "future" tightening implied by publishing a forecast that shows inflation moving above target has not been implemented yet-and this may hurt the central bank's credibility. 37 inflation targeting. This has been done in some cases by testing whether the output gap enters significantly in an empirically-estimated interest rate rule. Leiderman and Bar-Or (2000), for instance, estimated the policy reac- tion function of the Central Bank of Israel using monthly data for the period 1994-99 and found that interest rates are positively related to current infla- tion, lagged changes in output, and the previous period's interest rate. It is important to note, however, that the fact that there is a significant coefficient attached to output in an estimated interest rate reaction function does not necessarily imply that in practice monetary policy is guided in part by an output stabilization objective. To the extent that current inflation is only an imperfect proxy for future inflation, and to the extent that the latter does depend on the level of economic activity, it is possible for output to appear in the reaction function with a significant coefficient-without, by itself be- ing indicative of a departure from "strict" inflation targeting. Indeed, in the derivation of the optimal interest rate rule in Svensson's model given above, it was shown that even when the output gap does not enter the policy loss function (A = 0), it will (together with current inflation) affect the value of the instrument; the reason is that both variables help predict future inflation. The second issue relates to the role of forward-looking components in actual policy reaction functions. This is important because simple forecast- based rules have some clear similarities with the type of optimal flexible inflation-forecast targeting rules derived earlier; monetary policy under both types of rules seeks to offset deviations between expected inflation and the inflation target at some horizon.40 The Reserve Bank of New Zealand's pol- icy projections, for instance, are based on an explicit, forward-looking pol- icy reaction function that takes a relatively simple form (Archer (2000)). Some econometric studies (see Clarida, Gall and Gertler (1998a, 1998b)) have shown that a reaction function that includes a forward-looking measure of inflation, the contemporaneous output gap, and lagged interest rates (to account for smoothing considerations) characterizes monetary policy fairly well for the United States, Japan and Germany. The kind of policy rule that underlies this specification is what Clarida, Galf and Gertler call "soft- hearted" inflation targeting: in response to a rise in expected inflation rel- ative to target, the central bank raises nominal interest rates sufficiently to 40Batini and Haldane (1999), Isard and Laxton (2000), Parrado (1999), and Rudebush and Svensson (1999), have analyzed the dynamic properties of both simple and optimal forecast-based rules using simulation techniques. 38 push up real rates, but there is also a pure stabilization component to each rule. Mehra (1999) also estimated a policy reaction function for the Federal Reserve Board that contains both backward- and forward-looking elements; the funds rate is shown to depend on actual inflation, increases in expected future inflation, the expected output gap, and the long-term bond rate. The latter variable, which captures the effect of long-term inflation expectations on policy decisions, became particularly important after 1979, the beginning of the Volcker-Greenspan era. 5.3 Asset Prices and Inflation Targeting The role of asset prices-stock prices, house prices, and bond yields-in the conduct of monetary policy has been the subject of renewed debate in indus- trial countries (see Cecchetti, Genberg, Lipsky, and Wadhwani (2000)). Part of the reason is the growing importance of equities in households' financial wealth, as documented for instance by Boon, Giorno, and Richardson (1998), and the high degree of volatility exhibited by stock prices in recent years. In the specific context of inflation targeting, (unexpected) movements in asset prices may affect the central bank's inflation forecast and there are at least two reasons why that should be so (see Bernanke and Gertler (1999), and Smets (1997)). First, changes in asset prices may have a direct impact on aggregate demand. For instance, changes in stock prices may affect private consumption expenditure (through their impact on wealth) and investment spending-by affecting the ability of firms to pledge collateral and raise funds. Second, higher house prices may increase the ability of households to borrow and spend, whereas sharp changes in exchange rates may affect the demand for foreign goods and the ability of firms to sell domestic goods abroad. To the extent that movements in asset prices are not the result of changes in underlying fundamentals, offsetting these changes through adjustments in monetary policy instruments (essentially a "lean against the wind" interest rate policy) may avoid costly fluctuations in output and prices. Second, as- set prices are strongly influenced by expectations of future returns, which in turn are related to expectations of future economic activity, inflation and monetary policy. Thus, even if their impact on aggregate demand is limited, they may contain useful information about current and future economic con- ditions. This information may be used to improve the inflation forecast on which the direction of monetary policy is based. However, the information contained in asset prices movements may be 39 limited because they may reflect erratic changes in expectations. To what extent this is actually the case may be difficult to gauge, because existing asset price models are based on unobserved variables; their empirical predic- tions are subject to wide margins of error. This makes it difficult to identify the "right" price (reflecting, say, future profit growth rates or productivity shocks) and therefore what is an erratic movement or speculative bubble. In such conditions, incorporating asset prices systematically in monetary policy feedback rules may be unwarranted (see Fuhrer and Moore (1992)). More- over, the risk premium that is typically embedded in asset prices tends to vary over time. Basing monetary policy on a broader, asset-based measure of prices or monetary conditions may actually lead to greater variability in current and future output and inflation. 6 Recent Experiences As noted in the introduction, in recent years a number of industrial and developing countries have adopted an inflation targeting regime. To put matters in perspective, consider the results of the survey of monetary policy frameworks conducted in late 1998 by the Bank of England and summarized by Sterne (1999). 91 countries, of which 43 were considered developing coun- tries, were covered. Almost all the central banks surveyed targeted either the money supply, the exchange rate, or inflation directly. Several of them (in both industrial and developing countries) stated that instability in the veloc- ity of money (induced in many cases by financial liberalization) led them to abandon money targeting. In 1998 there were 54 countries with an explicit inflation target (compared to 8 in 1990), 37 with an explicit monetary aggre- gate target, and 47 with an explicit exchange rate target. Between 1990 and 1998, explicit targets (whether for money, inflation, or the exchange rate) became more widespread. Of the 54 countries that had inflation targets, 13 (or 14 percent of the sample) had inflation targets only; thus, many countries had several explicit targets.41 The majority of countries that abandoned their exchange rate target (including Mexico in 1995 and Russia in 1998) did so in response to a currency crisis. The fastest growing regime was the combined use of explicit money and inflation targets (24 percent of the countries in the 4"The use of multiple explicit targets may reflect the fact that, in practice, targets are viewed more as benchmarks (against which outcomes are measured) by monetary authorities as opposed to objectives that must be achieved. 40 sample in 1998)-more than the combined total of inflation targets only (14 percent) and money targets only (5 percent). In a related study, Morande and Schmidt-Hebbel (1999) listed a total of 45 countries that, over the 1990s adopted some form of inflation target: 12 industrial countries, 12 transition economies and 21 developing countries. They also identified 34 countries that targeted primarily money and 36 that targeted primarily the exchange rate. Although there is some scope for disagreement with the classification schemes used in these two studies, the broad message is quite clear: a growing number of countries have turned to inflation targeting as a monetary policy framework. There is now a large literature assessing country experiences with inflation targeting, particularly in the industrial world; among the most recent publi- cations are those of Bernanke, Laubach, Mishkin, and Posen (1999), Blejer, Ize, Leone, and Werlang (2000), Landerretche, Morande and Schmidt-Hebbel (1999), and Schaechter, Stone, and Zelmer (2000). Accordingly, the focus of this section will be rather brief and will focus on developing countries-most notably Brazil, Thailand, and South Africa. 6.1 Industrial Countries The first country to introduce inflation targeting was New Zealand in late 1989. Other countries to follow suit include Canada (February 1991), Is- rael (January 1991), the United Kingdom (late 1992), Australia, Finland, and Sweden (all in 1993), and Spain (January 1995).42 Adoption of infla- tion targeting followed either from increased instability in the relationship between monetary aggregates and prices, or exchange rate crises. In the United Kingdom, for instance, inflation targeting was introduced after the country left the European Exchange Rate Mechanism in September 1992. With the exception of Israel and Spain, all countries implementing inflation targeting did so in the context of a highly flexible exchange rate regime. Is- rael for instance has maintained since 1991 a diagonal exchange rate band; the presence of this band has acted as a constraint on the inflation objective, 42See Allen (1999) for a review of the United Kingdom's experience with inflation target- ing, Freedman (2000) for Canada, Archer (2000) for New Zealand, Leiderman and Bar-Or (2000) for Israel, and the various chapters in Blejer, Ize, Leone, and Werlang (2000). Schaechter, Stone, and Zelmer (2000) provide a general overview. Note that the Euro- pean Central Bank's monetary policy framework also incorporates elements of inflation targeting. 41 and there have been episodes during the past few years when the level of interest rates necessary to achieve the inflation target was inconsistent with the level required to maintain the nominal exchange rate within the band (Leiderman and Bufman (2000)). Spain also maintained an exchange rate commitment during the 1990s and therefore cannot be viewed as a "pure" inflation targeter. In any case, both Spain and Finland dropped formal tar- gets for domestic inflation when they joined the European single currency in 1999. Table 1 provides a concise summary of the basic elements of the inflation targeting regimes in eight industrial countries with explicit targets for infla- tion, including the date their inflation target was first introduced. The table reveals the following facts: 3 Most industrial countries specify their inflation target as a range, and the desired target is usually the middle of that range-even if this is not always made explicit. Intermediate target ranges were established in several countries (such as New Zealand) to reflect an initially high level of inflation and the desire for gradual disinflation. * Australia, the United Kingdom and Spain explicitly exclude mortgage interest payments from their headline measures; Australia, in addition, also excludes government-controlled prices and energy prices. Canada excludes indirect taxes as well as food and energy prices. Finland excluded housing capital costs, indirect taxes and government subsidies; Sweden has no formal exemptions. * Formal escape clauses (that is, conditions under which the inflation target may be missed) are rarely specified and used. Shocks to inflation have been typically accommodated by allowing fluctuations within the target band and/or by relying on a measure of core inflation. * The inflation target typically represents an understanding between the government and the governor (or president) of the central bank and is not a legislated objective, except in New Zealand. It can be changed either by an amendment to central bank legislation or by a decision of the government. * Only three of the eight central banks listed in the table now define a term over which the inflation target is to be met, following a transitional 42 period during which an explicit horizon was specified. In most cases the target horizon is indefinite. * Responsibility for setting the inflation target varies across countries. The target is determined by the central bank in four of eight countries, whereas the government sets the target in two countries (Australia and the United Kingdom). The decision is joint in the two remaining cases (Canada and New Zealand). * In six of the eight countries, the central bank provides (either volun- tarily or by mandate) a formal report on inflation prospects. Four out of eight central banks make available to the public explicit inflation forecasts. The available empirical evidence also reveals that inflation was on a down- ward trend in many of these countries prior to the introduction of inflation targets and that inflation persistence has continued to drop significantly after the adoption of inflation targets (Siklos (1999)). There is also some evidence suggesting that the degree of risk aversion to inflation volatility in inflation- targeting countries increased significantly during the 1990s (Cechetti and Ehrmann (1999)). However, the evidence does not conclusively indicate whether inflation targeting has led to a lowering of inflationary expecta- tions and enhanced credibility, thereby mitigating the real output costs that disinflation typically entails. In fact, sacrifice ratios (which measure the cu- mulative loss in actual output relative to trend output, per percentage point fall in inflation) do not appear to have been much affected by inflation tar- geting.43 In addition, as noted by Bernanke, Laubach, Mishkin, and Posen (1999), there appears to be no strong evidence that the introduction of in- flation targets dramatically affected inflation expectations, as revealed either by surveys or by the level of long-term nominal interest rates. Inflation ex- pectations have come down, in most cases, mainly because inflation-targeting central banks were able to demonstrate that they were capable of achieving 43It should ne noted, however, that the computation of the sacrifice ratio is fraught with difficulty-as can be inferred from the previous discussion on the measurement of potential output. Empirical estimates often fail to disentangle the impact of disinflation policies from other shocks. Conceptually, it also suffers from serious shortcomings; in particular, it focuses only on short-run costs and therefore does not capture the longer- run benefits of disinflation. 43 and maintaining low inflation. A recent quantitative analysis of the per- formance of inflation targeting in New Zealand, Canada, and the United Kingdom conducted by Honda (2000) provides some support for this view. He estimated an unrestricted VAR model with four variables (inflation, de- viations of output from trend, the short-term nominal interest rate, and the rate of change of the nominal exchange rate) over a sample period prior to the introduction of inflation targeting and simulated dynamically the model over the period following the adoption of inflation targeting. The test is that if predicted values from the model tend to overestimate actual values of inflation and the other variables in the system, then inflation targeting can be deemed as having had some favorable impact on inflation and other macroeconomic variables. However, Honda found no evidence that inflation targeting had an effect on either inflation or any other variable in the system. Of course, the power of alternative stability tests in a VAR context is an open issue, but Honda's analysis appears to be a promising way of analyzing the impact of inflation targeting on actual inflation performance. 6.2 Developing Countries Few developing countries have so far adopted inflation targeting, but the number of potential candidates appears to have risen significantly in the past two or three years. Not all developing countries that have floated their currency in recent years have adopted an (explicit or implicit) inflation tar- geting framework. Mexico is a case in point. Carstens and Werner (2000) have argued, for instance, that the high frequency and incidence of exogenous shocks (both domestic and external, such as movements in wages and pub- lic sector prices, and nominal exchange rate fluctuations) to domestic prices makes the adoption of an explicit inflation targeting framework unfeasible in the case of Mexico. In Peru, the central bank announces an annual inflation target since 1994 but does not have a formal inflation targeting regime. Nev- ertheless, both countries have been viewed by some observers (see Mishkin and Savastano (2000)) as moving toward the adoption of such a framework. By contrast, Chile has a long history of announcing inflation targets (one year ahead) that goes back to 1990, although it started to make its infla- tion report publicly available only recently (May 2000).44 More recently, "Because Chile maintained a crawling peg exchange rate regime until September 1999, some observers have argued that the adoption of a full-fledged inflation targeting frame- 44 Brazil, Thailand, and South Africa adopted such a regime, and several other countries (including those in East Asia that adopted a floating exchange rate following the 1997-98 financial crisis) are currently pondering the feasibility of such an approach.45 Chile's experience is well documented by Landerretche, Morande and Schmidt-Hebbel (1999); accordingly, this section provides a brief discussion of the cases of Brazil, Thailand, and South Africa. Table 2 summarizes the main characteristics of inflation targeting regimes in Chile, Brazil, South Africa, and Thailand. After the real crisis of January 1999 and the adoption of a floating exchange rate, Brazil adopted on June 21, 1999 an inflation targeting framework for monetary policy (see Bogdanski, Tombini, and Werlang (2000)). The inflation targets as well as the tolerance bands are set by the National Monetary Council, on the basis of a proposal by the Finance Minister. These targets must be set in principle two years in advance. The price index chosen for setting the inflation targets is the Broad Consumer Price Index (IPCA), reported by the National Bureau of Geography and Statistics (IBGE), although since March 2000 monthly mea- sures of core inflation have also been published (see Figueiredo (2000)). The choice of a full index was dictated by the need to mitigate potential credibil- ity problems associated with perceptions of deliberate manipulation of the index. The operational framework of inflation targeting consists of a set of mod- els to support the monetary policy decision process (including a small-scale structural econometric model, and short-term univariate forecasting models), supplemented with surveys of market expectations of major macroeconomic variables (including inflation and growth). The econometric model focuses on the aggregate demand channel of monetary policy, with a direct effect of changes in nominal interest rates, and an indirect effect through relative prices induced by the effect of changes in the nominal interest rate on the exchange rate. The targets are considered to be met whenever the observed accumulated inflation during the period January-December of each year (measured on the basis of the variations in the IPCA index) falls within the band. In case the target is missed, the central Bank Governor must issue an open letter to the Finance Minister explaining the causes of the breach, the measures work is only a recent phenomenon; see Schaechter, Stone and Zelmer (1999, p. 12). 45In countries like Korea, for instance, the central bank is now required by law to set a price stability target every year in consultation with the government. 45 to be adopted to ensure that inflation returns to the target band, and the period of time that will be needed for these measures to have the intended effect. Since September 1999, the central bank also issues a quarterly inflation report (simultaneously in Portuguese and English) that provides information on current and future inflation performance, and the effect of monetary policy on prices. Minutes of the monthly Monetary Policy Committee meetings are released with a two-week lag, which is soon to be reduced to one week. In June 1999, official inflation goals were announced for the coming three years: 8 percent in 1999, 6 percent in 2000, and 4 percent in 2001 (accumulated annual variations by year end). Tolerance intervals were set at ± 2% for each year. As shown in Figure 7, inflation has been on a downward path since the introduction of the inflation targeting regime and has remained well within the target band. South Africa adopted a full-fledged inflation targeting framework in Feb- ruary 2000, at a time when the country's inflation performance was continu- ing to improve-falling to 5.2 percent in 1999 from 8.6 percent in 1997 and 6.9 percent in 1998.46 A key reason to do so was the perceived growing in- stability between monetary aggregates and inflation in a context of financial liberalization (Schaechter, Stone and Zelmer (2000)). The inflation objective is currently expressed in terms of an adjusted inflation rate, the headline con- sumer price index excluding mortgage interest costs. Escape clauses include "major unforeseen events" deemed to be outside the control of the central bank. The target range for the year 2002 is a band of 3 to 6 percent average inflation. This medium-term target was set in view of estimated lags between monetary policy decisions and their impact on inflation: changes in interest rates in South Africa have been estimated to take from 18 to 24 months to affect inflation. The relatively large width of the band was chosen to account for the fact that the targeted price index includes a large number of volatile items, whose movements are beyond the control of the central bank. At- tempts to improve transparency have taken the form of public explanations of the monetary policy stance following meetings of the Monetary Policy Committee of the Reserve Bank (although full minutes of the deliberations are not yet released), the creation of a Monetary Policy Forum that meets twice a year (in May and November), and the publication of a semi-annual 46Before that, the Reserve Bank had an informal goal of maintaining inflation at a level that would be more or less in line with the average inflation rate in South Africa's major trading partners and international competitors. 46 Monetary Policy Review (twice a year beginning in 2001), which will rep- resent the Reserve Bank's inflation report. At this stage, inflation forecasts and the underlying forecasting models are not released to the public. Following the collapse of the Thai baht in July 1997 (after almost five decades of pegged exchange rate regimes), Thailand adopted a floating rage regime under which monetary targeting was followed at first. Quarterly mon- etary base targets (agreed with the IMF) were implemented and achieved through daily liquidity management. However, growing evidence of instabil- ity in the relationship between money, output and inflation led Thailand to announce the adoption of inflation targeting as a monetary policy framework in May 2000. Under this regime, monetary policy is decided by the Monetary Policy Board (MPB), which was appointed in April 2000 and comprises both high-level central bank officials and external experts. An amendment to the Bank of Thailand has recently been passed to give the MPB an official status and operational independence in conducting monetary policy. At the same time, price stability was reaffirmed as the overriding objective of monetary policy. The price index targeted by the central bank is an index of core infla- tion, which excludes raw food and energy prices (both components deemed to be highly volatile) from headline inflation. The target core inflation rate is set in the form of a band of between 0 and 3.5 percent for 2000-01 and is explicitly aimed at allowing the Bank of Thailand to cushion temporary shocks and reduce interest rate volatility. At the time of the adoption of inflation targeting, headline and core inflation were running at, respectively, 2 percent and 1.2 percent on a yearly basis. Monitoring of deviations from the target band is done on the basis of quarterly averages of core inflation, to avoid the high degree of volatility that may characterize monthly obser- vations. If core inflation breaches the target band, the MPB must explain why the breach occurred, identify the measures that it intends to take in response, and the amount of time required to bring inflation back within the desired range. Finally, the first inflation report was issued in July 2000 and is to be issued on a quarterly basis; risks associated with inflation forecasts are shown in fan charts, as is now done by several other inflation-targeting central banks. 47 7 Some Unresolved Analytical Issues Analytical and operational aspects of inflation targeting regimes continue to generate a large amount of research. How best to measure core inflation, for instance, continues to be actively discussed. This section focuses on three important issues of debate in the analytical literature. The implications of asymmetries (related to both policy preferences and the slope of the trade-off between inflation and the output gap) for the design and operation of an in- flation targeting regime are considered first. The implications of uncertainty (regarding the variables to be measured, behavioral parameters, and policy lags) are explored next. Finally, the treatment of credibility and reputational factors in empirical models of inflation is examined. 7.1 Asymmetric effects 7.1.1 Non-quadratic policy preferences The central bank's instantaneous policy loss functions (3) and (13) were taken to be symmetric; positive output gaps, for instance, were considered to be just as costly as negative output gaps. Adopting this approach can be justified not only because of its tractability but also because it can be viewed as a reasonable approximation of an underlying utility-based welfare function (Rotemberg and Woodford (1999)). In general, however, the short-run cost of disinflation may matter a great deal to the central bank and may lead to a situation in which a higher weight is attached to negative output gaps-for instance by adding a term that is linear in -y in (13). As a result, even if there is no long-run trade-off between output and inflation, the optimal inflation rate may be greater than 7r. More fundamentally, Orphanides and Wieland (2000) recently questioned the use of a quadratic objective function for policymakers in the analysis of inflation targeting regimes. They begin by noting that (as discussed earlier), in practice most inflation-targeting central banks specify a target band as opposed to point targets; they then argue that the existence of a target range implies a nonlinear optimal policy rule. A simple way to specify policy pref- erences that are consistent with a target band is to write the instantaneous loss function as, instead of (13): Lt (itr2 )+ Ay2 + I yti (25) 48 where 0 > 0. As shown by Orphanides and Wilcox (1996), this specifica- tion implies an asymmetric policy response: as long as inflation is relatively close to target, the optimal interest rate policy is simply to stabilize output. Otherwise, policy should keep inflation within a range (that varies positively with A), and then wait for favorable supply shocks (positive shocks to 6t) to move it closer to the desired value ir. Orphanides and Wieland (2000) considered in fact a more general spec- ification than (25), namely, a zone-quadratic policy loss function, that is, a loss function that assigns a quadratic loss to inflation deviations outside an explicit target band and a (near) zero loss as long as inflation fluctu- ates within the band. As a consequence, if the central bank assigns some weight to fluctuations in output (as is generally the case in practice), the output objective will dominate during periods when inflation is within the band and will lose (some) importance when inflation is outside the band. They argued that this specification of policy preferences is consistent with the often-observed tendency of central banks to show overwhelming concern with inflation only when it is beyond some range. Using numerical analy- sis Orphanides and Wieland also found that, with nonquadratic preferences (and a nonlinear inflation-output trade-off, as discussed below) uncertainty due to unexpected shocks has important effects on the width of the target band and on the relative size of the policy response inside and outside the band. In particular, the optimal policy rule under uncertainty does not call for a mechanical response only when inflation falls outside the band. Instead, it is optimal to respond to inflation deviations already within the band and continue to do so more aggressively if inflation continues to evolve outside the desired range.47 Although intuitively appealing, the results of Orphanides and Wieland deserve further scrutiny. There appears to be growing evidence of nonlinear- ities regarding aggregate behavioral functions (such as the Phillips curve, as discussed below), but the empirical research on the structure of policy pref- erences is less conclusive. Some other recent studies have, in fact, supported the standard approach. Chadha and Schellekens (1999), for instance, sug- gested by comparing several alternative specifications of the central bank's loss function that the assumption of quadratic losses may not be too far off the mark. Of course, such results may also be model specific. 47In addition, they showed that in the case where the Phillips curve is linear, the width of the band increases with the variance of shocks to inflation. 49 7.1.2 The convex Phillips curve As discussed earlier, establishing forecasts in the context of inflation target- ing requires a reasonable understanding of the structure of the economy and adequate parameterization of some key aggregate relationships, most notably between interest rates and output, and between output and inflation. There is a growing literature on the Phillips curve suggesting that the relation- ship between changes in the output gap and inflation may be nonlinear in industrial countries. Specifically, it suggests that the Phillips curve has a convex shape, that is, positive deviations of aggregate output from potential (or booms) tend to be more inflationary than negative deviations (reces- sions) are disinflationary. As discussed in detail by Dupasquier and Ricketts (1998), there are several analytical models that may justify the existence of an asymmetric relationship between inflation and the output gap. One of these models emphasizes the role of capacity constraints, that is, the fact that (some) firms may find it difficult to increase their capacity to produce beyond a certain range in the very short run. As a result, when aggregate demand is increasing, the impact on inflation will tend to be greater than during periods of low demand, and the Phillips curve will have a convex shape. Taking into account these nonlinearities in inflation targeting is impor- tant because they may imply the need for possible asymmetries in monetary policy decisions-for instance, the policy response to deviations of inflation above target may need to be stronger than those that occur when inflation is lower than target. Recent analytical research on this issue includes the study by Schaling (1999), who extended Svensson's basic closed-economy model presented earlier to consider the case of a convex Phillips curve of the type proposed by Chadha, Masson, and Meredith (1992), and Laxton, Mered- ith, and Rose (1995). He shows that the optimal instrument-setting rule is asymmetric and implies a higher level of nominal interest rates than the rule derived by Svensson. Specifically, suppose that the Phillips curve takes the convex form 7rt - 7rlt-1 = 1(Yt-1; ) = lYt-l (26) 1- alCyt-l where, as before, a, > 0 and 0 < w < 1 is a parameter that measures the degree of convexity of the curve. This specification is consistent with the capacity constraint argument stated earlier. The 0(.) function is such that 50 0(0; w) = 0; and the slope of the output-inflation trade-off is given by ot °ej-1 [1-alwYt-l]2 ' with 0'(0; w) = a1 and limit values of limO' = a,, lim 0' = oo, lim 0' = 0. w-S0 yt-I1/a la yt-iQ-00 The quantity l/law represents, thus, an upper bound that the output gap cannot exceed in the short run. As w -4 0, (26) tends to a linear rela- tionship between the (lagged) output gap and (changes in) inflation. Figure 8 summarizes the properties of the curve (26). In addition to (26), suppose that the output gap evolves in a way similar to (2) with, for simplicity, f2 = 1: yt = lyt-1 - (it-I - 1rt-1) + qt (27) where the aggregate demand disturbance, nt, is again assumed to be an i.i.d. shock. Assuming that the monetary authorities pursue "pure" inflation target- ing, and that the loss function takes the quadratic form (3) given earlier, it can be shown that the optimal interest rate rule can again be found by solv- ing a period-by-period problem (as shown in (7)) and that the first-order condition is again 7rt+21t = * (see Schaling (1999)). The instrument rule, however, is now a nonlinear relationship given by r 1 _ alw(,3yt - rt) (7t _ ir) + 1 + 1 - alw(2,31yt - rt)t 0= where rt = - 7rt denotes the real interest rate. FRom this equation it can be verified that, for w -O 0, the interest rate rule boils down to 1 rt =-(7rt - ) + (1 + O)yt, a, which is identical to (11). In general, however, the optimal short-term interest rate is a nonlinear function of both deviations of current inflation from target and the output gap.48 As shown by Schaling (1999), positive deviations from 48It should be noted that Bean (2000) proposed a specification of a convex Philips curve that results in a linear policy rule. In general, however, the optimal rule will also be nonlinear if the Phillips curve is nonlinear. 51 the inflation target imply larger movements in interest rates (in absolute terms) than negative deviations. In the same vein, positive output gaps imply larger increases in interest rates (in absolute terms) than negative output gaps. In addition, the nonlinear rule implies that increases in interest rates (in response, for instance, to positive output shocks) are larger than those obtained with the linear rule-implying that the latter underestimates the appropriate level of the policy instrument. The notion that the output-inflation trade-off depends on the initial state of the economy is crucial in an inflation-targeting framework. However, all the available evidence pertains to industrial countries (see most notably Lax- ton, Meredith, and Rose (1995), and Pyyhtia (1999)). Tables 3 and 4 present estimates for six developing countries (Colombia, Korea, Mexico, Nigeria, the Philippines, and Thrkey) of linear and nonlinear Phillips curves (using ordi- nary least squares in the first case and nonlinear least squares in the second), along the lines of Laxton, Meredith, and Rose (1995), using quarterly data. The explained variable is the rate of inflation INFL = A4 ln P, where P is the consumer price index, and the explanatory variables are (see Appendix B for details): * Lagged inflation (one and two periods), to capture persistence and expectations. * the output gap, GAP, given by (the log of) the ratio of actual output, Y, to potential output, YT. The latter is measured using two different filters: the standard Hodrick-Prescott filter (with a smoothing parame- ter of 1600) and a quadratic trend. For the nonlinear Phillips curve, the following specification is adopted: f (GAP) = GAP- where w is a parameter to be estimated.49 49As noted by Laxton, Meredith, and Rose (1995, p. 353), because positive output gaps have larger effects on inflation than negative output gaps, using detrended actual output may result in downward-biased estimates of potential output. To correct for this bias, they measured the output gap in their empirical results as ln[Y/(1 + a)YT] where a is a parameter to be estimated. Here, however, preliminary regressions yielded an estimated value of a that was very small and never significantly different from zero. Accordingly, the output gap was specified simply as ln(Y/YT). 52 * Current and lagged values of the rate of change of the nominal effec- tive exchange rate, to assess the direct pass-through effect of currency depreciations. * Foreign inflation, measured by the price of imported goods. In addition, following Leiderman and Bar-Or (2000), the possibility of a nonlinear effect of exchange rate movements on inflation is tested. The assumption is that the strength of the pass-through effect depends on the state of the economy; the extent of pass-through may be larger in booms than in recessions, when falls in aggregate demand tend to put downward pressure on prices.50 To capture this effect, an interactive term given by the product of the nominal effective exchange rate and the output gap is added in the regression. Consider first the results for the linear Phillips curve given in Table 3. Overall, the regressions fit the data fairly well. Lagged values of inflation are significant in all cases, whereas the value of the interactive term is significant only for the Philippines and Mexico (using the Hodrick-Prescott estimate of trend output). Foreign inflation is significant only for Korea; the rate of depreciation of the nominal effective exchange rate, by contrast, is highly significant (both current and one-period lagged values) except for Colombia and Nigeria. However, the output gap is not significant in any equation, regardless of the detrending method used or the lag considered.51 Consider now the results for the nonlinear Phillips curve shown in Table 4. The explanatory power of the regressions (as measured by the adjusted R-squared and the standard error of the regression) is about the same as before, and the degree of significance and size of the coefficients estimated for all variables, except the output gap, are very similar. Regarding the output 50An alternative assumption worth exploring in future research is, as suggested by Taylor (2000), the possibility that low inflation itself may result in a lower pass-through, due to reduced perceived persistence of exchange rate movements. Another possibility to consider is that the strength of the pass-through effect may decline over time as the inflation targeting regime gains credibility. This may explain why in Chile, for instance, the large exchange rate depreciation that followed the Asian and Russian crises in mid-1997 and mid-1998 did not translate into major upward pressure on domestic prices. 51Higher-order lags than those shown in the table were also tried, with no success. In addition, following Pyyhtia (1999), the model was estimated by entering separately positive and negative output gaps; the results remained qualitatively similar to those reported in the table. 53 gap, the results now suggest that the coefficient w is significantly different from zero in virtually every case and regardless of the detrending technique used. The coefficient of the function f(GAP), however, is significant only at the 10 percent level, and only for the Philippines (where the current value of GAP is used, with a quadratic trend used to measure potential output) and Nigeria (where a two-period lag is used, with the Hodrick-Prescott filter used). The linear and nonlinear specifications are compared using a non-nested test, the J test developed by MacKinnon, White, and Davidson (1983), which is briefly discussed in Appendix B. The results indicate that for Colombia, the linear model is rejected against the nonlinear alternative, regardless of whether the Hodrick-Prescott filter or quadratic trend filter is used. For Korea, both model hypotheses fail to be rejected, suggesting that the data do not provide enough information to discriminate between them. For Mexico, even though the linear model is rejected against the nonlinear one when the Hodrick-Prescott filter is used, both models are rejected when the quadratic trend method used. For Nigeria, the nonlinear model is accepted against the linear alternative when the Hodrick-Prescott filter is used; but failure to reject both models occurs when the quadratic trend method is used. In the case of the Philippines, the nonlinear model is accepted against the linear specification. By contrast, in the case of Turkey, the nonlinear model is rejected against the linear alternative using the Hodrick-Prescott filter and at a 10 percent significance level. Both models cannot be rejected when the quadratic trend is used in the regressions. Although suggestive, the above results should be viewed as preliminary. There is clearly a need for further testing to assess other possible sources of nonlinearities in the Phillips curve, as discussed by Dupasquier and Ricketts (1998). Although this is a difficult task, the source of the nonlinearity in the Phillips curve is important from a policy viewpoint, because they may have very different policy implications. With nonlinearities, the policy rule is state contingent; parameters (elasticities) that capture the response of interest rates to movements in the output gap and deviations of current inflation from target depend on the current levels of inflation and the output gap. This has considerable practical implications in the present context, given the relative lack of experience that inflation-targeting central banks (particularly those in developing countries) have had with the conduct of monetary policy during a cyclical downturn. 54 7.2 Uncertainty and optimal policy rules The implications of uncertainty for the design of policy rules under inflation targeting has received considerable attention in the most recent literature. Analytically, it is convenient to distinguish between four sources of uncer- tainty: * Uncertainty about some of the deterninants of inflation, which relates to the fact that some economic series are unobservable and must be es- timated. The most obvious example in the present case is the measure- ment of potential output in the calculation of the output gap. Potential output is very often approximated by trend output (as done earlier), but alternative detrending techniques may sometimes give large dis- crepancies.52...3 o Uncertainty about the parameters of the model, which can be inter- preted in two ways: either the underlying model itself is uncertain, or the "true" model is deterministic but policymakers do not know it for sure and must estimate it. Econometric techniques normally provide a sense of the degree of uncertainty that accompanies empirical esti- mates, because they yield not only point estimates of parameters but also their variances and covariances. * Uncertainty about the transmission lag and the timing of policy ac- tions. The transmission lag depends on a variety of economic and insti- tutional factors, such as the degree of development of financial markets and the intensity of competition on both the supply and demand sides, the degree of trade openness, and the composition of private agents' financial wealth.54 52Note that the error in measuring potential output is not necessarily problematic if it takes an additive form and is uncorrelated over time. 53A related problem is the significant revisions in economic time series (due to changes in seasonal adjustment factors, re-definitions, and so on) that often occur after a preliminary release of data-particularly those dealing with the real sector. Large revisions in the variables entering the instrument rule, in particular, may complicate the use of preliminary data as a basis for policy decisions. 5'As discussed earlier, there are lags in the response of aggregate demand to changes in interest rates, and lags in the response of inflation to the output gap. There are also lags in the response of inflation expectations to policy changes, as well as the response of inflation to changes in inflation expectations. In addition, as noted ealier, there are lags in 55 * Uncertainty about the nature and degree of persistence of shocks that the economy is subject to, that is, whether the shocks are on the supply or demand side of the economy, and whether they are temporary or permanent. It has been shown that, whereas some of these sources of uncertainty cause optimal policy to become more cautious, others have the opposite effect. To illustrate these results and their implications for inflation targeting, it is sufficient to consider two analytical examples: uncertainty about parameters and uncertainty about the degree of persistence of macroeconomic variables, including inflation. In a seminal paper, Brainard (1967) argued that when the policymaker is uncertain about the effect of its actions, it may be optimal to adopt a more gradual policy stance than under certainty (or, more generally, under certainty equivalence, which holds in a linear model with a quadratic loss function and additive shocks). To illustrate the implications of Brainard- type uncertainty, consider again Svensson's model of strict inflation targeting described earlier (equations (1)-(4)). To simplify the analysis, suppose that output affects immediately inflation, a, = 1, 31 = 0, and that there are no supply shocks (Et = 0 for all t). The behavioral equations of the model therefore become 7rt- rt-1 CelYt, (28) Yt =-02(4t-1-' t-1) + 77t, (29) where the demand shock qt is once again an additive, serially uncorrelated shock with zero mean. Substituting (29) in (28) yields 7rt+l = 1t - 72t + nt+l, where -y = (1 +±a112) and 1Y2 a 102. Assuming the same intertemporal loss function as before (equation (4)) and setting for simplicity the target ir = 0, the optimal interest rate rule can be shown to be it=-7t (30) 72 the response of aggregate demand to changes in relative prices induced by exchange rate changes, and lags in supply response to exchange rate-induced movements in the domestic price of imported inputs. 56 As noted earlier, this rule is certainty equivalent: the same interest rate rule would be optimal in a world with no uncertainty about aggregate demand shocks. But suppose now that the central bank does not know for sure the values of 'Yl and 72; all that is known is that these parameters are drawn from independent, normal distributions with means 71 and Y2 and variances oa and a 2, respectively. In this case, it can be shown (see Martin (1999)) that the optimal instrument rule becomes it= Y t2 +2rt. (31) Y2 2 This equation shows that as uncertainty about the parameters in the transmission process of policy shocks increases (that is, as a 2 rises), the opti- mal instrument response to movements in current inflation becomes smaller.55 The fundamental reason for this result is the following. As shown earlier (equation (8)) the per-period loss function can be decomposed into the sum of the squared expected deviation of each variable from its target (or the squared bias), and the conditional variance of that variable. With additive uncertainty, the variance is independent of the policy rule, and so policy de- cisions aim only at minimizing expected deviations in inflation. By contrast, under uncertainty about the parameters of the model, the variance of (fu- ture) inflation depends on the level of nominal interest rates; in particular, large movements in the policy instrument in response to deviations between actual and targeted inflation tend to reduce bias, as implied by the first term in (8)-at the cost, however, of increasing the variance of inflation, the sec- ond term in (8). The central bank will therefore internalize this effect by choosing a lower optimal level of interest rates.56 A similar result would hold in a more general, linear-quadratic setting in which the central bank pursues several policy objectives simultaneously. By and large, model-based simulation studies, for instance by Ha (2000), Martin and Salmon (1999), and Sack (2000), have confirmed the practical importance of the Brainard effect. The optimal interest rate rule calls for more gradual adjustment in the presence of parameter uncertainty. In par- ticular, Ha (2000) examined, using a numerical model for New Zealand, the impact of uncertainty about the transmission lag of monetary policy (that 55Note that as U2 tends to zero, (31) becomes identical to (30). 56It is important to note, however, that Brainard himself qualified this result by showing that it does not necessarily hold when the covariance between parameters (the policy multiplier, in his example) and the additive disturbance is sufficiently negative. 57 is, lag uncertainty regarding the timing of policy actions, as opposed to the effect of these actions, as in Brainard's analysis) on the setting of monetary policy instruments in the context of inflation-forecast based rules. His analy- sis showed that indeed less aggressive policy rules are more robust, in the sense that they are less affected by uncertainty about the monetary policy transmission lag. However, at the same time, more aggressive rules tend to produce lower inflation variability. Sack (2000) showed that accounting for parameter uncertainty in deriving an optimal interest rule from an estimated VAR model and a quadratic loss function in inflation and unemployment helps to explain the observed tendency for the Federal Reserve to change its policy rate (the federal funds rate) only gradually, despite the absence of interest rate smoothing considerations by the central bank. By contrast, several authors have shown that uncertainty about the de- gree of persistence in the economy can lead to a more aggressive policy than under certainty. Notable studies are those of Soderstrom (1999) and Shuetrim and Thompson (1999). In Soderstrom's model, for instance, it is uncertainty about the degree of persistence of inflation itself that may lead to this result. Without full information regarding inflation persistence, a cautious mone- tary policy may result in inflation not approaching the target at the desired rate, or even diverging from the target. The central bank can lower this risk by implementing large adjustments in interest rates, thereby reducing uncer- tainty regarding the path of inflation. This more aggressive policy leads to the expectation that inflation will return to target more quickly, so that the implicit targeting horizon is shortened.57 The sharp differences in these results suggest that more quantitative re- search is needed to fully understand the impact of uncertainty on policy rules under inflation targeting. In particular, it may well be that the impact of uncertainty on the optimal policy rule may depend on which parameter, or which behavioral relationship, is being considered in a structural model. Put differently, uncertainty about particular parameters may be of relatively lim- ited importance for the conduct of monetary policy, whereas others may have an unduly large effect on the setting of policy instruments. Identifying which parameter(s) matter(s) may well be model-specific. 5"Note that if the central bank cares only about stabilizing inflation, the implicit tar- geting horizon is alread as short as possible (that is, equal to the control lag). In this case, it is not affected by uncertainty about the persistence of inflation. 58 7.3 Endogenizing Reputation and Credibility The foregoing discussion highlighted the need, in setting policy instruments in an inflation targeting framework, to account for the trade-off between the cost of having inflation above target (which is reduced if disinflation is faster) and the cost of output fluctuations (which is increased if disinflation is faster). The trade-off arises because of the assumption that to reduce inflation the central bank must necessarily induce a (temporary) reduction in output. In that regard, it is important to note that the extent and duration of the re- duction in output depend crucially on the credibility of the commitment to the inflation target and its evolution. The more credible the commitment becomes over time, the faster inflation expectations will fall, and the lower will be the output cost of reducing inflation. More generally, accounting for changes in credibility in forecasting inflation and simulating policy shocks may be important in the first stages of implementation of an inflation tar- geting framework, particularly in countries where the degree of confidence in the central bank's commitment to low inflation is not well established. Nevertheless, there have been relatively few attempts to endogenize cred- ibility in empirical macroeconomic models used for forecasting and policy analysis under inflation targeting. One reason may be the difficulty of opera- tionalizing theoretical concepts of credibility and its determinants. A simple approach is to assume that the expected inflation rate, lt/t-1, is a weighted sum of lagged inflation, 7rt-1, and the inflation target, 7r, with relative weights of 1 - at and at, respectively: 7rt/t-1 -(1-Ut)7rt-1 + ±Ut. (32) The weight on the announced inflation target ot can be viewed as a mea- sure of policy credibility, which can be modeled as { At-1 when 'P 0 at = Aatl+T(7rtj--, 7rt-2-l, *--) -* , (33) Aa t- + 1 when - 1 where 0 < A < 1 and IF(.) is a function of past inflation forecast errors, with the properties that 'I -* 1 when these errors become very small and I- 0 when these errors become very large. Expectations are thus com- pletely backward-looking when T -* 0 (because at -O 0 as well), whereas when 'I -* 1 full credibility is achieved (at -* 1), with expectations depend- ing only on the announced target. Isard, Laxton and Eliasson (2001) provide 59 a more elaborate treatment of this approach to endogenizing credibility. A key conclusion that emerges from simulation studies based on expectations formation schemes similar to (32) and (33) is that endogenous policy credi- bility strengthens the case for the type of forward-looking inflation forecast- based rules discusser earlier. An alternative and conceptually appealing approach to endogenizing cred- ibility and learning in a tractable manner is the one pursued by Drazen and Masson (1994) and Agenor and Masson (1999), who view credibility as con- sisting of two elements: an assessment of the central bank's "type" (which could be termed reputation), and (for a given type) an assessment of the prob- ability that the central bank will actually decide to stick to the announced policy (maintain inflation close to target) in the presence of adverse shocks to prices.58 In this setting, inflation expectations will reflect assessments about the central bank's "type", as captured by the relative weights that the au- thorities attach to each of their policy objectives, which are not known by the public. In addition, private agents-knowing that random shocks will alter the balance of costs and benefits associated with maintaining the inflation rate close to target-will reevaluate on the basis of observed variables the probabilities that a particular type of policymaker will decide to stick to the inflation target in the future. Put differently, if there is significant persistence in the effects of policies, then a restrictive policy carried out today (which lowers inflation but also reduces output) may make it less likely that such a policy will be continued in the future. To illustrate this approach, consider again the per-period policy loss func- tion under flexible inflation targeting described earlier (equation (13)). The first component of (lack of) credibility, the probability that the central bank put a high weight on limiting output fluctuations, can be modeled using Bayesian updating, under the assumption that there are two possible types of policymaker, each with a known set of weights on its objectives: a "weak" central bank, who sets A = Aw, and a "tough" central bank, whose value of A is AT < Aw. Thus, the policy loss function can be written as: Lt 2 Ay2 ' A Aw (34) 2 2'AT < Aw 5"An alternative approach consists in defining credibility as the ability of the central bank to precommit its actions, that is, its capacity to convince private agents that it will carry out policies that may be time inconsistent. See, for instance, Cukierman (1992). 60 In each period private agents calculate ex ante the likelihood of each type deciding to deviate from the inflation target. If inflation remains close to target ex post, this would give information about whether the central bank is weak (even if the shocks cannot be observed), so that initial priors about that probability are updated on the basis of the relative likelihood that each policymaker would have deviated from target, given the distribution of the unobserved shocks. In this framework inflation expectations depend on the probability that the central bank is weak or tough as well as the ex ante probability that a given type will decide to deviate from the inflation target as a result of random supply shocks, as captured by Et. Private agents do not observe the supply shock; they form their expectations using an information set that includes variables known as of the end of t - 1, that is, the lagged values of output, interest rates, and inflation, and whether the policymaker has al- lowed inflation to deviate from target or not. The central bank observes the supply shock, and chooses whether or not to adjust interest rates and keep inflation close to target. In general, the central bank will allow inflation to deviate significantly from target when a negative shock to output is so large that the cost of maintaining inflation close to its target value will exceed those associated with flexibility (that is, higher inflation). Formally, let LT be the value of the loss function if inflation is kept close to target through changes in interest rates, and L D the value when it is allowed to deviate significantly from target with no change in policy instruments. The central bank will maintain interest rates constant when LD - LTF < 0. Using, as be- fore, equations (1) and (2) to characterize the economy, it can be established on the basis of the results derived by Agenor and Masson (1999) that for -D - L F to be negative requires the supply shock Et to be relatively large compared to a threshold value, 5h, which depends on the type of central bank and other variables in the model. The threshold level is lower for a weak central bank than for a tough central bank (EW < E). The expected inflation rate in this setting is the product of the probabil- ity of deviating from target, Pt, and the size of the deviation from target, xt, which may or may not be constant over time. In turn, the private sector's as- sessment of Pt is equal to the probability of a weak central bank, Ot, times the probability that a weak central bank will devalue, pw, plus a corresponding term for a tough central bank: Pt = StP+ (1- t)p[ (35) 61 Expected inflation is thus ptxt [= [tPt + (1 - Ot)pT]xt. (36) Given knowledge of the authorities' objective function and of the distri- bution of the supply shock, the private sector can calculate the probabilities pw and pT, which can be defined as Pt = Pr(Et > Eh)- Finally, the probability of a weak policymaker, Ot, can be assumed to be updated using a Bayesian rule, starting from a prior estimate At-1. Specif- ically, because private agents observe the absence of deviations of inflation from target at time t - 1, they will revise 7rt-I on the basis of the relative likelihoods that the two types would have chosen not to deviate from the inflation target: ( 1 Pt-)Ot o- + ( 1 11 ) Ot-1) The above approach may prove fruitful as a way to endogenize credibility and reputation in macroeconomic models designed to predict inflation and analyze the performance of alternative policy rules, particularly in countries where, to begin with, the degree of confidence in the central bank's ability to maintain its comrnmitment to price stability is relatively low. However, other approaches are also possible, and sorting out the advantages and limitations of each of them remains a matter for further research. 8 Summary and Conclusions In the past few years a number'of central banks (in industrial and developing countries alike) have adopted an explicit inflation targeting framework for the conduct of monetary policy. This paper attempted to provide a (mostly ana- lytical) discussion of the issues involved in designing such a framework, with some emphasis on the particular features and recent experience of developing countries. The first part of Section II described an analytical framework for a closed economy, due essentially to Svensson (1997b, 1999b). The analysis showed two main results. First, because of lags in the transmission process of short-term interest rates to prices, inflation targeting implies inflation 62 forecast targeting. The central bank's forecast becomes an explicit interme- diate target. FRom that perspective, then, inflation targeting can be viewed as a monetary policy framework under which policy decisions are guided by expected future inflation relative to an announced inflation target; the forward-looking instrument rule takes into account lags in the transmission process. Second, if the central bank, in addition to seeking to achieve its inflation target, aims at stabilizing output, it should allow for a slower ad- justment of the inflation forecast to the target value compared to a situation in which the inflation target is the only goal. Extension of the analysis to an open-economy setting showed that, given the critical role that the exchange rate plays in the transmission process of monetary policy, inflation targeting may lead to a relatively high degree of output volatility by inducing excessive fluctuations in interest rates. Section III compared inflation targeting regimes with money supply and exchange rate targeting regimes. It was noted that monetary targeting re- quires a stable relationship between monetary aggregates and the price level, but that such stability had become elusive as a result of financial liberal- ization and abrupt changes in inflation expectations. Regarding exchange rate targeting, it was noted that a number of developing countries have been forced to abandon their exchange rate pegs in recent years as a result of un- sustainable speculative pressures. These developments led in many cases to the adoption of inflation targeting as an operational framework for conduct- ing monetary policy. Section IV discussed three basic requirements for implementing an infla- tion targeting framework, namely, central bank independence, the absence of implicit exchange rate targeting, and transparency in the conduct of mone- tary policy. It is now well recognized that openness and transparency play an important role in achieving credibility in monetary policy. The announce- ment of inflation targets communicates the central bank's intentions to the financial markets and to the public, and in so doing helps to reduce uncer- tainty about the future course of inflation. In that sense, transparency and accountability act as constraints on the temptation to adopt discretionary policies. The risks associated with the pursuit of an implicit exchange rate target were also highlighted. It is crucial for the central bank to be able to convince the public that the inflation target will take precedence over other policy goals in case of conflict. In particular, if the central bank is also con- cerned about fluctuations in the nominal exchange rate (as may be the case in countries where the pass-through rate to domestic prices is high or when the 63 short-term foreign-currency liabilities of the private sector are large) there is a risk that inflation targets may lack credibility-undermining the operation of the inflation targeting framework. The operational framework of inflation targeting was discussed in Section V. Among the issues reviewed were the measurement of inflation (includ- ing sources of imperfection in traditional measures), whether a target band for inflation should be chosen, the time horizon of monetary policy, difficul- ties associated with forecasting inflation, and whether asset prices should be taken into account in targeting inflation. It was noted, in particular, that the width of the inflation target range depends on the variability of shocks to inflation, the policy horizon, and the desired speed of adjustment to eco- nomic disturbances-which itself depends on the relative weight attached by the central bank to output fluctuations and the desired degree of interest rate smoothing. In addition, there is also a trade-off between credibility and flexibility: if the band is made too wide to provide more flexibility (that is, more scope to accommodate transitory shocks to inflation), the inflation target may lose credibility and inflation expectations may remain high. It was also noted that there may be several reasons why central banks may want to respond to financial prices in their pursuit of price stability. In par- ticular, shocks to financial prices that are not driven by fundamentals may destabilize the economy through their effects on aggregate demand, in which case the central bank may want to offset them. however, it was also noted that asset prices are determined by arbitrage equations in which expecta- tions of future returns play an important role. If expectations are subject to large and unpredictable shifts, these prices may contain limited additional information about current and future economic conditions. Finally, it was pointed out that inflation forecasts are in practice based on a combination of quantitative models (generally small in size), indicator variables (such as survey expectations), and qualitative judgement. Section VI reviewed the recent experience of industrial and developing countries with inflation targets, with a particular emphasis, in the latter group, on Brazil, Thailand, and South Africa. The evidence clearly suggests that the adoption of an inflation targeting regime was not by itself sufficient to dampen inflation expectations and help countries to deliver consistently better inflation performance. Nevertheless, for a subset of countries (most notably Canada, New Zealand, Sweden, and the United Kingdom), infla- tion persistence dropped significantly after the adoption of inflation targets, perhaps as a result of a significant and credible shift in policy preferences. 64 The last section focused on some ongoing issues of debate in inflation targeting, namely the role of nonlinearities and asymmetric effects (related to both the structure of policy preferences and structural relationships, most notably the Phillips curve), uncertainty (about behavioral parameters and the transmission process of monetary policy), and how to account for cred- ibility and reputational factors in empirical macroeconomic models. It was argued that the first two issues have important implications for the design of optimal instrument rules and the operation of a state-contingent targeting regime. Specifically, it was shown that whereas uncertainty about parame- ters may lead to more caution in the manipulation of policy instruments, uncertainty about the degree of persistence of inflation may lead to a more aggressive, as opposed to a more gradual, interest rate policy. New results regarding the convexity of the Phillips curve were also presented for six de- veloping countries (Colombia, Korea, Mexico, Nigeria, the Philippines, and Turkey). Comparison of the linear and nonlinear specifications using the MacKinnon-White-Davidson non-nested test provided mixed support for the latter. Finally, it was argued that accounting for changes in credibility in forecasting inflation and simulating policy shocks may be important, partic- ularly in countries implementing an inflation targeting framework from an initial position of low confidence in the central bank's commitment to price stability. The main conclusion of this paper is broadly in line with several existing studies: inflation targeting is a flexible policy framework that allows the cen- tral bank to exercise "constrained discretion", as emphasized by Bernanke, Laubach, Mishkin, and Posen (1999, p. 293). In middle- and high-income developing economies that have relatively low inflation to begin with, rea- sonably well-functioning financial markets, and can refrain from implicit ex- change rate targeting, it has the potential to improve the design and perfor- mance of monetary policy compared with alternative operational procedures available to central banks. There are, of course, technical requirements (such as adequate data on prices, sufficient understanding of the links between instruments and targets of monetary policy, and the ability to forecast rela- tively well price developments) that may not be satisfied in all countries and to the same extent. But such requirements should not be overstated; fore- casting capability, for instance, can never be perfect and sensible projections always involve qualitative judgement. A more important and difficult task, in many cases, may be to design or improve the institutional framework in order to allow the central bank an effective degree of independence in pursu- 65 ing the goal of low and stable inflation without undue pressure to stabilize output fluctuations or alleviate the public debt burden through low interest rates. 66 Appendix A The Optimal Instrument Rule under Forward-Looking Expectations This Appendix discusses the derivation of the optimal interest rate rule implied by inflation targeting under rational, forward-looking expectations. Following Clarida, Galf, and Gertler (1999), consider a closed economy in which the Phillips curve and aggregate demand are given by the following equations: 7rt =- a yt + a2Etirt+i + -t, (Al) yt = Etyt+l - 3(it - Et7rt+l) + 77t, (A2) where 0 < a2 < 1 and, as before, 7rt is the inflation rate, yt the output gap, it the nominal interest rate, and Et and t are disturbances that obey Et Pest-i + Vt, (A3) 71t =P7t7-i + (t, (A4) where 0 < pe, p, < 1 and vt and (t are random shocks with zero mean and constant variances, given respectively by a2 and a . Equation (Al) relates inflation to the current value of the output gap (as opposed to the lagged value) and the one-period ahead expected future infla- tion rate. It can be derived from a Calvo-Taylor model with staggered nomi- nal wage and price setting; see Roberts (1995) and Fuhrer (1997a, 1997b)).59 Equation (A2) relates output to its future value and the ex ante real interest rate. It can be derived by log-linearizing the first-order condition for con- sumption maximization obtained in a representative agent model and impos- ing equilibrium of the goods market (see Clarida, Galf, and Gertler (1999)). For simplicity, changes in interest rates are assumed to affect immediately current output. The key difference-which is readily apparent by iterating (Al) and (A2) forward-in this specification with forward-looking expectations is that out- put depends on current and future values of the real interest rate (and thus 59It can also be shown that a, is inversely related to the degree of price rigidity; the longer prices are held fixed (on average), the less responsive is inflation to cyclical fluctu- ations in output. 67 on future policy decisions), whereas inflation depends on current and future values of the output gap. This dependence makes calculation of the optimal instrument rule more involved. The central bank's policy objective is, as in (14): min Ut = Et { ah-t [(rh _*)+AY2 } (A5) where A > 0. Under discretion, the central bank tales expectations as given in solving its optimization problem. To derive the optimal policy rule proceeds in two stages: * The objective function (A5) is minimized by choosing Yt and 7rt, given the Phillips curve (Al). This is possible because no endogenous state variable appears in the objective function; thus future inflation and output are not affected by today's policy decisions and the central bank cannot directly affect private expectations. * Conditional on the optimal values of Yt and 7rt, the value of it implied by the aggregate demand equation (A2) is determined. The first stage consists thus in choosing Yt and 7rt to minimize (7rt-*)2 + +Xt 2 where xt = Et +E [(2rh -7r) + Ayh] } subject to (using (Al)): 7t = alYt + Zt, where zt = a2Et7rt+l + Et. The first-order conditions for the first-stage prob- lem are -7rt - f + At = °, \Yt - Atal = O, where pt is a Lagrange multiplier. Combining these conditions gives 7rt = * - ( A )Yt- (A6) 68 Substituting this expression in (Al) for yt yields 2 Irt =-yC7rt - T) + a2Et7rt+l + Et, that is alir Ce2AEt7rt+l AEt __t __ + + . + + (A7) Art + 2 +a2 I I A + alQ This equation can be solved by using the method of undetermined coef- ficients. Conjecturing a solution of the form 7rt = Kl1 r + /c2Et, (A8) implies that 7rt+l = ±lk + 1c2Et+l, so that, using (A3): Et7rt+1 = 'l'r + ± 2EtEt+l = Kllr + K2P6t. Substituting this expression in (A7) and rearranging terms yields 7rt = (R + . +(1 ± 021t2Pe) (A9) Equating coefficients in (A8) and (A9) yields a ±2+ COK A(I+ a2i,2P,) 1=Al + a:˘/z 2 c A ( + a 2Etp) which can be rearranged to give a 2 A A(l -a2) + a2<1, K2 = (1- a2P,) + a2 ' Thus, the solution for 7rt is 7r- = ±1,R + ASEt, (AlO) where A(l - a2P,) ± al Substituting (AIO) in (A6) yields Yt = (-( )(7rt-r) =-(cl) {(cl - 1)Yr + AOEt}, 69 that is Yt = A1r - ajOet, A = a, (1-c) (All) The second stage of the solution procedure consists in rewriting (A2) as it= {(Etyt+1 - Yt) + 7t} + Et7rt+1, (A12) and substituting for Etyt+l - Yt. From (A3) and (AIO), Et7rt+l is given by Et7rt+j = Kir + AOEtEt+j = K17r + AOpeEt, (A13) whereas from (A3) and (All): ,tyt+l = Ai - aQOpe&t Using this expression together with (All) yields Etyt+l- Yt = l(l - pe)OEt. (A14) Equation (A13) can be rewritten as OcEt =-l1r + Et7rt+j Ap, Ap, Substituting this result in (A14) yields Etyt+l- yt = aiKi(l - P)T + a< (l - pj Et7rt+l. (A15) Ap, Ap, Substituting (A15) in (A12) and rearranging terms yields the optimal interest rate rule: it=P) + la'( -P Et7rt+l + 77t} + Et7rt+l, or equivalently it = Q* + 5Et7rt+l + (A16) where air,,(l- Pe) + = a1ct(l Pe&) > f3Ap ' Ao/3p, 70 Thus, the optimal policy rule (A16) also calls for inflation forecast tar- geting. Because E > 1, an expected increase in future inflation calls for a more than proportional increase in the current nominal interest rate, in order to lower the real interest rate today and reduce aggregate demand now. If A -+ 0 (strict inflation targeting) then 1/E -+ 0, AOp, 0 and tc -* 1, so that (A13) yields Et7rt+l = which is analogous to what Svensson's model would predict in the absence of a lag between changes in the output gap and inflation (see (9) in the text). Put differently, strict inflation targeting is optimal-in the sense of equating the inflation target and the one-period ahead expected value of inflation-only if the central bank has no concern for output fluctuations. Otherwise, convergence to inflation back to target following a shock will be gradual, as implied by the discussion of flexible inflation targeting in the text. The optimal rule also calls for completely offsetting aggregate demand shocks because they do not imply a short-run trade-off between output and inflation.60 A more general specification of the Phillips curve (Al) and the aggregate demand equation (A2) in which both backward- and forward-looking com- ponents are present can be considered, as for instance in the New Keynesian models discussed by Fulurer (1997a, 1997b): lrt = alYt + ,o7rt_- + (1 - p)a2EtIrt+1 + et, Yt = -8lYt-1 + (1 - O)EtYt+l - 12(it - Et7rt+l) + qt, where 0 < p, 0 < 1. Equations (1) and (2) can be obtained from the above specifications by setting y = 0 = 1. Following Clarida, Gali, and Gertler (1999, pp. 1691-95) it can be shown that the qualitative properties of the optimal instrument rule derived earlier remain essentially the same. 60Equations (A10) and (All) can be used to illustrate the output-inflation variability tradeoff that arises in the presence of supply shocks. In general, a} = a2022, and a2 = A292a2. When A -+ 0 (the case of strict inflation targeting), 0 -l/c, and a2 - 0, y- ca/al. And when A -+ oo (the case when stabilizing output is the only policy goal), 9 -+ 0 and (by applying L'Hospital's rule) AO -S 1/(1 - C2pe). As a result, aV - 0 and a,r - ore/(l - a2P,)- 71 Appendix B Data Sources, Causality Tests, and Nonnested tests for the Phillips Curve The data used to produce Figure 1 are obtained from the quarterly data- base compiled by Agenor, McDermott and Prasad (2000). The bivariate VAR includes the money market rate and the discount rate (in that order). Estimation periods are 1995:1-1999:4 for Colombia, 1978:1-1999:3 for Ko- rea, 1978:1-1996:4 for Malaysia, 1983:2-1996:4 for Tunisia, 1986:2-1996:2 for Turkey, and 1994:1-1999:4 for Uruguay. The optimal lag length, chosen on the basis of the Akaike Criterion, was 2 for Colombia, 2 for Korea, 2 for Malaysia, 2 for Tunisia, 1 for Turkey, and 5 for Uruguay. Standard, bivariate Granger causality tests showed also that the discount rate causes the money market rate only in Korea (F-statistic = 3.122, P-value =0.049). There is two-way causality in Malaysia and Turkey, and no statistically significant evidence of causality in Colombia, Tunisia and Uruguay.61 The variables used in the regressions reported in Tables 3 and 4 are defined as follows. Inflation (INF) is the log difference of consumer price index (IFS code 64) between periods t and t-4. GAP is the log difference between the seasonally-adjusted industrial production index (IFS code 66c or 66ey) and its trend value, calculated by using either the Hodrick-Prescott method or a quadratic trend (obtained by regressing output on a constant term, time and time squared). The growth rate of the nominal effective exchange rate (NEERGR) is the log difference of the nominal effective exchange rate (IFS code eneer) between periods t and t-4. An increase in the nominal effective exchange rate is a depreciation. The growth rate of import prices (PMGR) is the log difference of the import price index (IFS code 75 or 75d) between periods t and t-4. For Nigeria, imported inflation is measured by the growth rate of export prices in developed countries (PXGR DEVED), which is the log difference of the export price index (IFS series code 74d and country group code 110) of these countries between periods t and t-4. The estimates of the linear and nonlinear Phillips curve models reported in Tables 3 and 4 are compared using the J-test proposed by MacKinnon, 61These results should be treated with some caution because during the estimation pe- riod monetary policy procedures changed significantly in some of the countries considered. In Turkey, for instance, the repurchase rate has become the main interest rate instrument in recent years. 72 White, and Davidson (1983). Specifically, suppose that the hypotheses to be tested are: Ho: y = ao +,30f (x, y) + eo, H1 : y = a,i + /lx + e1, where y is the endogenous variable, x a vector of exogenous variables, ai, Bi and -y the parameters to be estimated, and ei error terms with classical properties. The basic idea of the J -test is that if one of the models (say, the model specified under Ho) is the correct one, then the fitted values from the other model (corresponding to H1) should not have any explanatory power when estimating the Ho model. In practical terms, testing the Ho model against the H1 model proceeds as follows. First, estimate the H1 model and retrieve the fitted values. Then estimate the Ho model including the fitted values from the H1 model. If the fitted values from the H1 model enter significantly in the Ho model, reject the Ho model. Similarly, the H1 model can be tested against the Ho model by estimating first the Ho model and calculating the fitted values, then estimating the H1 model including the fitted values from Ho. If the fitted values from the Ho model enter significantly in the H1 model, reject the H1 model. If the test fails to reject only one model, this model is said to be superior to the other one. If both models are rejected against the alternative, this suggests that another model is needed. If the tests fall to reject both models, the data do not provide enough information to discriminate between the two alternatives. The results are reported in the text. 73 References Agenor, Pierre-Richard, and Joshua Aizenman, "Contagion and Volatility with Imperfect Credit Markets," ," IMF Staff Papers, 45 (June 1998), 207-35. Agenor, Pierre-Richard, and Paul R. Masson, "Credibility, Reputation, and the Mexican Peso Crisis," Journal of Money, Credit, and Banking, 31 (February 1999), 70-84. Agenor, Pierre-Richard, C. John McDermott, and Eswar Prasad, "Macroeco- nomic Fluctuations in Developing Countries: Some Stylized Facts," World Bank Economic Review, 14 (May 2000), 251-86. Agenor, Pierre-Richard, and Peter J. Montiel, Development Macroeconomics, Princeton University Press (Princeton, New Jersey: 1999). Alba, Pedro, Amar Bhattacharya, Stijn Claessens, Swati Ghosh, and Leonardo Hernandez, "The Role of Macroeconomic and Financial Sector Linkages in East Asia's Financial Crisis," in Financial Crises: Contagion and Market Volatility, ed. by Pierre-Richard Ag6nor, Marcus Miller, David Vines, and Axel Weber (Cambridge University Press: 1999). Alchian, Armen A., and Benjamin Klein, "On a Correct Measure of Inflation," Journal of Money, Credit, and Banking, 5 (February 1973), 173-91. Allen, William A., Inflation Targeting: The British Experience, Handbook in Central Banking Lecture Series No. 1, Bank of England (January 1999). Archer, David D., "Inflation Targeting in New Zealand," unpublished, Central Bank of New Zealand (March 2000). Arrau, Patricio, Jose De Gregorio, Carmen Reinhart, and Peter Wickham, "The Demand for Money in Developing Countries: Assessing the Role of Financial Innovations," Journal of Development Economics, 46 (April 1995), 317-40. Ball, Laurence, "Policy Rules for Open Economies," in Monetary Policy Rules, ed. by John B. Taylor, Chicago University Press (Chicago, Il.: 1999). Batini, Nicoletta, and Andrew G. Haldane, "Forward-Looking Rules for Mon- etary Policy," in Monetary Policy Rules, ed. by John B. Taylor, Chicago University Press (Chicago, IE.: 1999). Batini, Nicoletta, and Edward Nelson, "Optimal Horizons for Inflation Target- ing," Working Paper No. 119, Bank of England (July 2000). Batini, Nicoletta, and Edward C. Nelson, "Optimal Horizons for Inflation Tar- geting," unpublished, Bank of England (July 1999). Bean, Charles, "The Convex Phillips Curve and Macroeconomic Policymaking under Uncertainty," unpublished, London School of Economics (February 2000). 74 Beddies, Christian H., "Monetary Policy and Public Finances: Inflation Targets in a New Perspective," IMF Staff Papers, 46 (September 1999), 293-314. Beetsma, Roel M., and Henrik Jensen, "Inflation Targets and Contracts with Uncertain Central Bank Preferences," Journal of Money, Credit, and Bank- ing, 30 (August 1998), 384-403. Bernanke, Ben S., and Mark Gertler, "Monetary Policy and Asset Price Volatil- ity," Economic Review (Federal Reserve Bank of Kansas City), 84, No. 4 (December 1999), 17-51. Bernanke, Ben S., Thomas Laubach, Frederic S. Mishkin, and Adam S. Posen, Inflation Targeting: Lessons from the International Experience, Princeton University Press (Princeton, New Jersey: 1999). Bernanke, Ben S., and Frederic S. Mishkin, 'Inflation Targeting: A New Frame- work for Monetary Policy?," Journal of Economic Perspectives, 11 (Spring 1995), 97-116. Bernanke, Ben S., and Michael Woodford, "Inflation Forecasts and Monetary Policy," Journal of Money, Credit, and Banking, 29 (November 1997), 653- 86. Bharucha, Nargis, and Christopher Kent, "Inflation Targeting in a Small Open Economy," Discussion Paper No. 9807, Reserve Bank of Australia (July 1998). Blejer, Mario I., Alain Ize, Alfredo M. Leone, and Sergio Werlang, eds., Inflation Targeting in Practice, International Monetary Fund (Washington DC: 2000). Blix, Marten, and Peter Sellin, "Uncertainty Bands for Inflation Forecasts," Working Paper No. 65, Sveriges Riksbank (- 1998). Bogdanski, Joel, Alexandre A. Tombini, S6rgio R. Werlang, "Implementing In- flation Targeting in Brazil," unpublished, Banco Central do Brasil (January 2000). Boon, Laurence, Claude Giorno, and Pete Richardson, "Stock Market Fluctua- tions and Consumption Behavior: Some Recent Evidence," Working Paper No. 98-21, OECD (November 1998). Boskin, Michael J., E. Dulberger, Robert J. Gordon, Zvi Griliches, and Dale Jorgenson, Toward a More Accurate Measure of the Cost of Living, Final Report to the Senate Finance Committee from the Advisory Commission to Study the Consumer Price Index (Washington DC: 1996). Brainard, William, "Uncertainty and the Effectiveness of Policy," American Eco- nomic Review, 57 (May 1967), 411-25. Briault, Clive, Andrew Haldane, and Mervyn King, "Independence and Account- ability," in Towards more Effective Monetary Policy, ed. by Iwao Kuroda, 75 St. Martin's Press (New York: 1997). Carstens, Agustin, and Alejandro M. Werner, "Mexico's Monetary Policy Frame- work under a Floating Exchange Rate Regime," in Inflation Targeting in Practice, ed. by Mario I. Blejer, Alain Ize, Alfredo M. Leone, and Sergio Werlang, International Monetary Fund (Washington DC: 2000). Ceccheti, Stephen G., "Inflation Indicators and Inflation Policy," in NBER Macroeconomics Annual 1995, ed. by Ben S. Bernanke and Julio J. Rotem- berg, MIT Press (Cambridge, Mass.: 1995). Ceccheti, Stephen G., and Michael Ehrmann, "Does Inflation Targeting Increase Output Volatility? An International Comparison of Policymakers' Prefer- ences and Outcomes," Working Paper No. 7426, National Bureau of Eco- nomic Research (December 1999). Cecchetti, Stephen G., Hans Genberg, John Lipsky, and Sushil Washwani, As- set Prices and Central Bank Policy, Centre for Economic Policy Research (London: 2000). Ceccheti, Stephen G., and Erica L. Groshen, "Understanding Inflation: Impli- cations for Monetary Policy," Working Paper No. 7482, National Bureau of Economic Research (January 2000). Chadha, Bankim, Paul R. Masson, and Guy Meredith, "Models of Inflation and the Costs of Disinflation," IMF Staff Papers, 39 (June 1992), 395-431. Chadha, Jagjit S., and Philip Schellekens, "Monetary Policy Loss Functions: Two Cheers for the Quadratic," Working Paper No. 101, Bank of England (September 1999). Clarida, Richard, Jordi Gali, and Mark Gertler, "Monetary Policy Rules in Prac- tice: Some International Evidence," European Economic Review, 42 (June 1998a), 1033-68. , "Monetary Policy Rules and Macroeconomic Stability: Evidence and some Theory," Working Paper No. 6442, National Bureau of Economic Research (March 1998b). Forthcoming, Quarterly Journal of Economics. , "The Science of Monetary Policy: A New Keynesian Perspective," Journal of Economic Literature, 37 (December 1999), 1661-707. Cukierman, Alex, Central Bank Strategy, Credibility, and Independence, MIT Press (Cambridge, Mass.: 1992). , "Towards a Systematic Comparison between Inflation Targets and Mon- etary Targets," in Targeting Inflation, ed. by Andrew G. Haldane, Bank of England (London: 1995). Cukierman, Alex, Miguel A. Kiguel, and Leonardo Leiderman, "Choosing the Width of Exchange Rate Bands: Credibility vs. Flexibility," Discussion Pa- 76 per No. 907, Centre for Economic Policy Research (January 1994). Cukierman, Alex, and Allan H. Meltzer, "Theory of Ambiguity, Credibility, and Inflation under Discretion and Asymmetric Information," Econometrica, 54 (September 1986), 1099-128. Debelle, Guy, "Discussion," in Monetary Policy and Inflation Targeting, ed. by Philip Lowe, Reserve Bank of Australia (Sydney: 1997). De Haan, Jakob, and Willem J. Kooi, "Does Central Bank Independence Really Matter? New Evidence for Developing Countries using a New Indicator," Journal of Banking and Finance, 24 (April 2000), 643-64. Dittmar, Robert, William T. Gavin, and Finn E. Kydland, "The Inflation- Output Variability Tradeoff and Price-Level Targets," Quarterly Review (Fed- eral Reserve Bank of St. Louis), 81 (January 1999), 23-31. Drazen, Allan, and Paul R. Masson, "Credibility of Policies versus Credibility of Policymakers," Quarterly Journal of Economics 104 (August 1994), 735-54. Dupasquier, Chantal, and Nicholas Ricketts, "Non-Linearities in the Output- Inflation Relationship: Some Empirical Results for Canada," Working Paper No. 98-14, Bank of Canada (August 1998). Eijffinger, Sylvester C., Marco Hoeberichts, and Eric Schaling, "A Theory of Central Bank Accountability," Discussion Paper No. 2354, Centre for Eco- nomic Policy Research (January 2000). Estrella, Arturo, and Frederic S. Mishkin, "Is there a Role for Monetary Aggre- gates in the Conduct of Monetary Policy," Journal of Monetary Economics, 40 (October 1997), 279-304. Faust, Jon W., and Lars E. 0. Svensson, "Transparency and Credibility: Mon- etary Policy with Unobservable Goals," Working Paper No. 6452, National Bureau of Economic Research (March 1998). Figueiredo, FRancisco, "Evaluation of Measures of Core Inflation," unpublished, Banco Central do Brasil (May 2000). FReedman, Charles, "The Canadian Experience with Targets for Reducing and Controlling Inflation," unpublished, Central Bank of Canada (March 2000). Fuhrer, Jeffrey C., "Inflation-Output Variance Trade-Offs and Optimal Mone- tary Policy," Journal of Money, Credit, and Banking, 29 (May 1997a), 214- 34. , "The (Un)Importance of Forward-Looking Bebavior in Price Specifica- tions," Journal of Money, Credit, and Banking, 29 (August 1997b), 338-50. Fuhrer, Jeffrey C., an George Moore, "Monetary Policy Rules and the Indicator Properties of Asset Prices," Journal of Monetary Economics, 29 (April 1992), 303-36. 77 Garfinkel, Michelle R., and Seonghwan Oh, "Strategic Discipline in Monetary Policy with Private Information: Optimal Targeting Horizons," American Economic Review, 83 (March 1993), 99-117. Gordon, Robert J., "The Boskin Commission Report and its Aftermath," un- published, Northwestern University (September 1999). Ha, Yuong, "Uncertainty about the Length of the Monetary Policy Transmission Lag," Working Paper No. 2000/01, Reserve Bank of New Zealand (February 2000). Haldane, Andrew G., "Some Issues in Inflation Targeting," Working Paper No. 74, Bank of England (January 1998). Hansen, Eric, "Price Level versus Inflation Rate Targets in an Open Economy with Overlapping Wage Contracts," Working Paper No. PB96-01, Federal Reserve Bank of San Francisco (September 1996). Herrendorf, Berthold, "Inflation Targeting as a Way of Precommitment," Oxford Economic Papers, 50 (July 1998), 431-48. Honda, Yuzo, "Some Tests on the Effects of Inflation Targeting in New Zealand, Canada, and the UK," Economics Letters, 66 (January 2000), 1-6. Isard, Peter, and Douglas Laxton, "Inflation Forecast Targeting and the Role of Macroeconomic Models," in Inflation Targeting in Transition Economies: The Case of the Czech Republic, ed. by Warren Coats, International Mone- tary Fund (Washington DC: 2000). Isard, Peter, Douglas Laxton, and Ann-Charlotte Eliasson, "Inflation Targeting with NAIRU Uncertainty and Endogenous Policy Credibility," Journal of Economic Dynamics and Control, 25 (January 2001), 115-48. Jadresic, Esteban, "Inflation Targeting and Output Stability," Working Paper No. 99/61, International Monetary Fund (April 1999). Jensen, Henrik, "Targeting Nominal Income Growth or Inflation," Discussion Paper No. 2341, Centre for Economic Policy Research (December 1999). Kiley, Michael T., "Monetary Policy under Neoclassical and New Keynesian Phillips Curves, with an Application to Price Level and Inflation Targeting," Discussion Paper No. 27, Board of Governors of the Federal Reserve System (May 1998). Landerretche, Oscar, Felipe Morande, and Klaus Schmidt-Hebbel, "Inflation Targets and Indexation in Chile," Working Paper No. 55, Central Bank of Chile (December 1999). Laxton, Douglas, Guy Meredith, and D. Rose, "Asymmetric Effects of Economic Activity on Inflation," IMF Staff Papers, 2 (June 1995), 344-74. 78 Leiderman, Leonardo, and Hadas Bar-Or, "Monetary Policy Rules and Trans- mission Mechanisms under Inflation Targeting in Israel," Working Paper No. 71, Central Bank of Chile (May 2000). Leiderman, Leonardo, and Gil Bufman, "Inflation Targeting under a Crawling Band Exchange Rate Regime: Lessons from Israel," in Inflation Targeting in Practice, ed. by Mario I. Blejer, Alain Ize, Alfredo M. Leone, and Sergio Werlang, International Monetary Fund (Washington DC: 2000). Leitemo, Kai, "Inflation Targeting Strategies in Small Open Economies," un- published, University of Oslo (June 1999). MacKinnon, J., H. White, and R. Davidson, "Tests for Model Specification in the Presence of Alternative Hypotheses: Some Further Results," Journal of Econometrics, 21 (January 1983), 53-70. Martijn, Jan Kees, and Hossein Samiei, "Central Bank Independence and the Conduct of Monetary Policy in the United Kingdom," Working Paper No. 99/170, International Monetary Fund (December 1999). Martin, Ben, "Caution and Gradualism in Monetary Policy under Uncertainty," Working Paper No. 105, Bank of England (December 1999). Martin, Ben, and Salmon, "Should Uncertain Monetary Policymakers Do Less?," Working Paper No. 99, Bank of England (August 1999). Masson, Paul R., Miguel A. Savastano, and Sunil Sharma, "The Scope for Infla- tion Targeting in Developing Countries," Working Paper No. 97/130, Inter- national Monetary Fund (September 1997). Mayes, D., and B. Chapple, "Defining an Inflation Target," in Targeting Infla- tion, ed. by Andrew G. Haldane, the Bank of England (London: 1995). McCallum, Bennett T., "Issues in the Design of Monetary Policy Rules," in Handbook of Macroeconomics, ed. by John B. Taylor and Michael Woodford, Elsevier (North Holland: 1999). Mehra, Yash P, "A Forward-Looking Monetary Policy Reaction Function," Eco- nomic Quarterly (Federal Reserve Bank of Richmond), 85 (Spring 1999), 33-53. Mishkin, Frederic S., "Inflation Targeting in Emerging Market Countries," Work- ing Paper No. 7618, National Bureau of Economic Research (March 2000). Mishkin, Frederic S., and Miguel A. Savastano, "Monetary Policy Strategies for Latin America," Working Paper No. 7617, National Bureau of Economic Research (March 2000). Morande, Felipe, and Klaus Schmidt-Hebbel, "The Scope for Inflation Targeting in Emerging Market Economies," unpublished, Central Bank of Chile (June 1999). 79 Muscatelli, V. Anton, "Optimal Inflation Contracts and Inflation Targets with Uncertain Central Bank Preferences: Accountability through Independence," Economic Journal, 108 (March 1998), 529-42. , "Inflation Contracts and Inflation Targets under Uncertainty: Why we Might Need Conservative Central Bankers," Economica, 66 (May 1999), 241- 54. Orphanides, Athanasios, and David W. Wilcox, "The Opportunistic Approach to Disinflation," unpublished, Board of Governors of the Federal Reserve System (May 1996). Orphanides, Athanasios, and Volker Wieland, "Inflation Zone Targeting," Eu- ropean Economic Review, 44 (June 2000), 1351-87. Parrado, Eric, "Inflation Targeting and Exchange Rate Rules in an Open Econ- omy," unpublished, New York University (June 1999). Pyyhtia, Ilmo, "The Nonlinearity of the Phillips Curve and European Monetary Policy," Working Paper No. 17/99, Bank of Finland (November 1999). Quah, Danny, and Shaun P. Vahey, "Measuring Core Inflation," Economic Jour- nal, 105 (September 1995), 1130-44. Roberts, John M., "New Keynesian Economics and the Phillips Curve," Jounal of Money, Credit, and Banking, 27 (November 1995), 975-84. Roger, Scott, "Core Inflation: Concepts, Uses and Measurement," Working Pa- per No. 98/9, Reserve Bank of New Zealand (July 1998). Rotemberg, Julio, and Michael Woodford, "Interest Rate Rules in an Estimated Sticky Price Model," in Monetary Policy Rules, ed. by John B. Taylor, Chicago University Press (Chicago, Il.: 1999). Rudebusch, Glenn, and Lars E. 0. Svensson, "Policy Rules for Inflation Target- ing," in Monetary Policy Rules, ed. by John B. Taylor, Chicago University Press (Chicago, Il.: 1999). Sack, Brian, "Does the Fed Act Gradually? A VAR Analysis," Journal of Mon- etary Economics, 46 (August 2000), 229-56. Schaechter, Andrea, Mark R. Stone, and Mark Zelmer, "Practical Issues in the Adoption of Inflation Targeting by Emerging Market Countries," unpub- lished, International Monetary Fund (August 2000). Schaling, Eric, "The Non-Linear Phillips Curve and Inflation Forecast Target- ing," Working Paper No. 98, Bank of England (July 1999). Shuetrim, G., and C. Thompson, "The Implications of Uncertainty for Monetary Policy," in Monetary Policy under Uncertainty, ed. by Benjamin Hunt and Adrian Orr, Central Bank of New Zealand (Wellington: 1999). Sikken, Bernd Jan, and Jakob De Haan, "Budget Deficits, Monetization, and 80 Central Bank Independence in Developing Countries," Oxford Economic Pa- pers, 50 (July 1998), 493-511. Siklos, Pierre L., "Inflation-Target Design: Changing Inflation Performance and Persistence in Industrial Countries," Quarterly Review (Federal Reserve Bank of St. Louis), 81 (March 1999), 47-58. Smets, Frank, "Financial Asset Prices and Monetary Policy: Theory and Ev- idence," in Monetary Policy and Inflation Targeting, ed. by Philip Lowe, Reserve Bank of Australia (Melbourne: 1997). Soderstrom, Ulf, "Shoud Central Banks be more Agressive?," Working Paper No. 84, Sveriges Riksbank (May 1999). Sterne, Gabril, "The Use of Explicit Targets for Monetary Policy: Practical Experiences of 91 Economies in the 1990s," Quarterly Bulletin (Bank of England), 39 (August 1999), 272-81. Svensson, Lars E. O., "Optimal Inflation Targets, 'Conservative' Central Banks, and Linear Inflation Contracts," American Economic Review, 87 (March 1997a), 98-114. , "Inflation Forecast Targeting: Implementing and Monitoring Inflation Tar- gets," European Economic Review, 41 (June 1997b), 1111-46. , "Open-Economy Inflation Targeting," Working Paper No. 1989, National Bureau of Economic Research (October 1998). "Price Level Targeting vs. Inflation Targeting: A Free Lunch?," Journal of Money, Credit, and Banking, 31 (August 1999a), 277-95. , "Inflation Targeting: Some Extensions," Scandinavian Journal of Eco- nomics, 101 (September 1999b), 337-61. Svensson, Lars E. O., and Michael Woodford, "Implementing Optimal Pol- icy through Inflation-Forecast Targeting," unpublished, Princeton University (November 1999). Taylor, John B., "Low Inflation, Pass-Through, and the Pricing Power of Firms," European Economic Review, 44 (June 2000), 1389-408. Vestin, David, "Price-Level Targeting versus Inflation Targeting in a Forward- Looking Model," unpublished, Stockholm University (May 2000). Walsh, Carl E., "Optimal Contracts for Central Bankers," American Economic Review, 76 (March 1995), 150-67. , "Announcements, Inflation Targeting and Central Bank Incentives," Eco- nomica, 66 (May 1999), 255-69. Wynne, Mark A., "Core Inflation: A Review of Some Conceptual Issues," in Measures of Underlying Inflation and their Role in the Conduct of Monetary Policy, Bank for International Settlements (Basle: 1999). 81 Figure 1 Response of Money Market Rate to One Standard Deviation innovations in Discount Rate (+/- 2 standard error) 8 Colombia 2 Korea 6 , - 1.5 -. 4 2~~~~~~ U - -'' ° -, - - --- - s*,~~~~~~~~~~~~~. -2 - 0- -4 -6 | ,|F:,tI ,I , ,tt-0.5 tI,I,^||*,, ,,I 1 5 9 13 17 21 1 5 9 13 17 21 0.8- 1 Malaysia 0.7 Tunisia 0.8 - - -. 0.6 - - 0.5 0.6 0.4 0.4 0.3 0.2~~~~~~~~~~~~~~~~~. - S ~~~~~~~~~~~~0 0~ -0.2 I , I I -0.2 I I I I I I I I I-0.2 1 5 9 13 17 21 1 5 9 13 17 21 25 Turkey 2 Uruguay 20 1 / . , - -- 15 -3 1 5 9 13 17 21 1 5 9 13 17 21 -15 -3~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~* S I5 ~ .S - 1 5 9 13 17 21 1~ ~ ~ ~~~~~~~ 5 9 3 17 2 Figure 2 Behavior of Inflation Following a Shock (in percent) (flexible inflation targeting) (strict inflation targeting) 0 1 i, 2 3 4 5 6 7 8 9 10 2Peods io 7X T-- --------X---°-- L____ _____ I I 0 1 2 3 4 5 6 7 8 9 10 Pefiods 0~~~~~~~ y1' o 1 2 3 4 5 6 7 8 9 10 Periods Figure 3 Autocorrelation Function of Inflation Rate (quarterly annualized growth rates) 1.2 Chile Colombia 1 (1978:1-1999:4) 1 (1978:1-1999:4) 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 India 1 Korea 0.8 (1978:1-1999:4) (1981:1-1999:4) 0.8 0.6 0.6 0.4 0.20. ~0 4 -0.2 -0.20 -0.4 ~~~~~~~~~~~~~~~~-0.2 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 alaysia 1.2 Mexic (178:1-1999:4) 1978:1-1999:4) 0.81 0.6 0.8 0.6 0.4 0.4 0.2 0.2 0 1 2 3 4 5 6 7 8 9 10 11 121 2 3 4 5 6 7 8 9 10 11 12 Figure 3 (concluded) Autocorrelation Function of Inflation Rate (quarterly annualized growth rates) Morocco Nigeria 1 978:1-1999:4) (1978:1-1999:4) LL ~~~~~~~~~~~~~~~~~~~1 I' 0.8 0.6 0.5 0.4 0 -0.5 1 2 3 4 S 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 I Philippines l Tunisia 1 [ (1978:1-1999:4) 008 0.5 06 Ii 0 -05 0 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 key 1.2 rruguay 78:1-1999:4) 1 78:1-1999:4) 0.8 0.8 0.6 0.6 0.4~~~~~~~~~~~~~~~. 0.2~~~~~~~~~~~~~~~~~~~0 10 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 526 7 8 9 10 11 0. 0~~~~~~~~~~~~~~~~~. 3 4 6 7 8 910 1 121 2 3 4 5 6 7 8 9 10 11 12 Figure 4 Average Inflation Rate and Central Bank Independence Index 1/ (Averages over 1980-89) 150 - 120 - * 100 0 100 - 80 - Cu - c 60 - o *.c:420 - 0 0 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.2 0.4 0.6 0.8 1 1.2 Index based on turnover rate Index based on political vulnerability 150 - 150 100 100 50 500 0 0~~~~~~~~~0 S 50 * 50 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Index based on legal dependence Index based on lending limit Source: Sikken and Haan (1998). Note: The countries are Barbados, Chile, Colombia, Costa Rica, Ethiopia, Ghana, Greece, Honduras, India, Indonesia, Israel, Kenya, Korea, Malaysia, Malta, Mexico, Morocco, Nepal, Pakistan, Philippines, Portugal, South Africa, Tanzania, Thailand, Turkey,Uganda, Uruguay, Venezuela. Figure 5 Developing Countries: Volatility of Consumer Price Inflation and Food Price Inflation (Averages over 1981-98) 40 Correlation coefficient: 0.971 c 30 0 CL 0 20~~~~~~~ 0. c~~~~~~~~ . 20 _ m~~~~~~ 10~~~~~ 0 0 .o 0@-0 n2 10 _*h ; 0o I I I 0 10 20 30 40 Standard deviation of consumer price inflation Note: The countries are Bangladesh, Benin, Botswana, Burkina Faso, Burundi, Central African Republic, Chad, Chile, Colombia, Congo Republic, Costa Rica, Dominican Republic, Ecuador, Egypt, Arab Rep., El Salvador, Ethiopia, Fiji, Gabon, Gambia, Ghana, Guatemala, Haiti, Honduras, India, Indonesia, Jamaica, Jordan, Kenya, Korea, Lesotho, Madagascar, Malawi, Malaysia, Mauritius, Mexico, Morocco, Namibia, Nepal, Nigeria, Panama, Paraguay, Philippines, Saudi Arabia, Senegal, South Africa, Sri Lanka, Syria, Tanzania, Thailand, Trinidad and Tobago, Tunisia, Turkey, Uganda, Uruguay, Venezuela, Zimbabwe. Source: World Bank. Figure 6 Developing Countries: Consumer Price Inflation and Inflation Volatility (Averages over 1981-98) Correlation coefficient: 0.9.. 40 0 0 w- 0 _c 30 c) E * 20 0 C., > 0% so.4 10~~~~~~~~~~ o * 4~~~o 2 00~~% C: ~~00 0 U3 1 * 0 1 00 0 10 20 30 40 50 60 Average consumer price inflation Note: For the list of countries represented in this figure, see the note to Figure 5. Source: World Bank. Figure 7 Brazil: Consumer Price Inflation and Inflation Targets, 1995-2002 (in percent per annum) 30 20 10 LO Irt cD 0 - v- o a0 0) o 0 0 N N 0) a) 0 ) 0 ) 0) 0) 0) CY) 0 00 0 0 0 a 0 C iD C Ci i °) ° °i C U C o Suc 0 at 0 rnational Monetar 0 Source: International Monetary Fund and Banco Central do Brasil. Co o C 0~~~~~~~~~~~~~ 0~~~~~~~~~~~~~~~~~~ CD~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C CD~~~~~~~~~~~~~~~~~~~~~~~~~rC CD~~~~~~~~~~~~~~~~~~~~~~~~~~C -o H CD -~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~-- \~~~~~~~~~~~~ o ,, \ (D~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, D~~~~~~C 0 I-. a) Table I Industrial Countries: Basic Features of Inflation Targeting Regimes Separate Publishes Country Inflation Excluded Items from Inflation Inflation Target Contingencies for Breaches Targeting Adoption Date Inflation Who Sets Inflation Measure Measure Calculation Period of the Inflation Target Horizon? (mmnVddyy) Report? Target? Forecast? Mortgage interest 2 to 3 percent on average over the business Over the cycle None None 01/01/93 No Govemment No Australia CPI Government-controlled pnices Nn oe 0119 o oenet N Energy prices cycle 2 to 4 percent by the end of 1992 Canade cpll lIndirect taxes 1.5 to 3.5 per-cent by June 1994 (original) Annual None Yes 02126191 Yes Joint No Food and energy prices (revised) I to 3 percent Dec 1993 to Feb 20013 Housing capital costs Finland CPI Indirect taxes 2 percent annual average from 1995 Annual None No 02/02/93 No" Central bank No Govemment subsidies Short-term influences embodied 14 to 15 percent (1991-92) in certaincopntsfth Israel CPI None less than 4 percent (1999) Annual prce index, e.g. prices of fruits Yes 01/01/91 Yes Central bank Yes 3 to 4 percent (2000-1) and vegetables, housing, and 3 to 4 percent (2000-1) ~~~~~~~imports. 3 to 5 percent (Dec 1990); Commodity prices 2.5 to 4.5 percent (Dec 1991) 'Unusual' events that do not New Zealand Cptl Govemment-controlled pOces 1.5 to 3.5 percent (1992:Q1-04) Annual generate general inflationary Yes 03102190' Yes Joint Yes Interest, credit charges 0 to 2 percent (1993-96) pressures. 0 to 3 percent since December 19969 3.5 to 4 percent (June 1996) 3 to 3.25 percent (1997:Q1) Spain CPI Mortgage interest costs 3 percent upper limit for 1997 Annual None Yes 01/01/95 Yes Central bank No 2.5 to 2.75 percent upper limit for late 1997 2 percent for 1998' Sweden CPI Nominally none but condifional ito3percentsince1995 Annual None No 01/15/93 Yes Centralbank Yes on indirect taxes, subsidies United Kingdom RPI Mortgage Interest payments I to 4 percent until June 1997 elections Annual None No 10/08/92 Yes9 Govemment Yes 2.5 percent since June 1997 Source: based on Allen (1999), Archer (2000). Leiderman and Bar-Or (2000), Schaechter, Stone and Zelmer (2000), and national central bank reports. 'The Govemors, however, 'available' to report on the conduct of monetary policy twice a year to the House of Representatives Standing Committee on Financial Institutions and Public Administration. 2Although the target is officially specified in terms of overall CPI, the Bank of Canada focuses on the CPI excluding food, energy, and the effect of indirect taxes. 'Renewed in February 1998. 4Finland reports quartedy on the inflation outlook in its Monthly Bulletin. 'Since December 1997, the CPI excluding credit services is targeted. Before that date, overall CPI was targeted. In late 1999, mortgage interest rates were removed from the index. 'The term of the new Policy Targets Agreement (PTA) coincides with the current term of the Govemor, which expires August 31, 2003. The PTAs were agreed to in December. 'The Reserve Bank's announced policy included a specific target for inflation and a specific date for that target to be achieved as eady as mid 1989 (Archer (2000)). "The Law of Autonomy was put in place in June 1994 and, although the inflation target was announced in December 1994, it was formally adopted only as of January 1,1995. Between 1995 and 1997 the aim was to reduce inflation to the 2 percent range. In 1998, the aim was to keep the annual inflation rate 'close to 2 percent" during the year. 9Only since 12 June 1997. Table 2 Developing Countries: Basic Features of Inflation Targeting Regimes Country Brazil Chile South Africa Thailand Inflation Measure CPI CPI CPI CPI Excluded Items from Inflation None None Mortgage interest costs Raw food and energy Measure prices Inflation Target 1999: 8 percent±2 percent 1991: 18 percent 2002: 3 to 6 percent 2000-01: 0-3.5 percent 2000: 6 percent t 2 percent 2000: 3.5 percent 2001: 4 percent ± 2 percent 2001 onwards: 3 percent ± 1 percent 2002: 3.5 percent i 2 peroent Calculation Period Annual Annual Annual Annual Contingencies for Breaches None None Major unforeseen events None of the Inflation Target outside central bank control Targeting Horizon? Yes No Yes Yes Adoption Date (mmfddlyy) 06/21/99 09/01190 2/1/00 05/01/00 Separate Inflation Report? Yes Yes' Yes3 Yes2 Who Sets Target? Joint Central Bank Central Bank Central Bank Publishes Inflation Forecast? Yes Yes No Yes Source: based on Landerretche, Morande and Schmidt-Hebbel (1999), Bogdanski. Tombini and Weriang (2000), Schaechter, Stone and Zelmer (2000), and national central bank reports. 'The first inflaton report was published in May 2000 and will be released three imes a year (January, May, Septernber). 2The first inflabon report was published in July 2000. 3An inflation report has been prepared since the end of 1996 by the Reserve Bank. At this stage, the report is not distributed outside the Bank. Table 3 Estimation Results: Linear Phillips Curve Dependent variable: inflation rate (Ordinary Least Squares) Colombia Korea Mexico Nigeria Philippines Turkey HP QT HP QT HP QT HP QT HP QT UP QT C 0.026 0.026 0.007 0.008 0.021 0.017 0.021 0.025 0.015 0.020 0.012 0.022 (2.237) (2.289) (3.216) (3.232) (2.956) (2.454) (1.951) (2.218) (2.618) (3.707) (0.802) (1.165) INFMj 1.097 1.096 0.822 0.794 1.312 1.284 1.461 1.454 1.207 1.119 0.872 0.866 (9.490) (9.474) (27.641) (19.011) (14.429) (12.962) (15.514) (15.371) (11.325) (11.454) (18.009) (17.145) INF,.2 -0.226 -0.266 -0.465 -0.436 -0.566 -0.572 -0.443 -0.389 (-1.955) (-1.956) (-6.401) (-5.534) (-5.958) (-6.012) (-5.048) (-4.781) GAP, 0.019 -0.002 -0.019 0.021 -0.018 0.002 (0.465) (-0.283) (-0.473) (1.009) (-0.283) (0.154) GAP,., -0.136 0.021 0.107 -0.040 (-0.619) (1.380) (1.356) (-0.815) GAP,. 0.101 -0.011 (0.873) (-0.649) NEERGR, 0.107 0.109 0.180 0.190 0.118 0.097 0.387 0.374 (6.582) (6.834) (5.687) (6.401) (5.385) (4.530) (8.825) (8.152) NEERGR,., -0.052 -0.048 -0.058 -0.072 -0.275 -0.276 (-2.900) (-2.658) (-1.408) (-1.830) (-5.333) (-5.290) GAP,*NEERGR, 0.807 0.733 (1.832) (3.388) GAP,.,*NEERGR,, 0.753 -0.030 (1.935) (-0.632) PMGR, 0.074 0.045 (3.418) (2.017) PMGR, l 0.094 0.086 0.058 0.047 (5.026) (4.457) (1.190) (0.925) PXGR_DEVED, 0.063 0.070 (0.761) (0.844) Adj. R2 0.792 0.792 0.958 0.958 0.982 0.981 0.897 0.896 0.946 0.953 0.925 0.923 Standard error ofregression 0.015 0.015 0.011 0.011 0.031 0.032 0.055 0.056 0.021 0.020 0.038 0.038 Total Number of 76 76 79 79 79 79 81 81 64 64 75 75 Observations Period 1980:1 - 1998:4 1980:2 - 1999:4 1980:2 - 1999:4 1979:3 - 1999:3 1981:1 - 1999:3 1981:2- 1999:4 Note: t-statistics are in parentheses. C is a constant term. INF is the inflation rate. GAP is the log difference of output and trend component of output which is calculated using either Hodrick and Prescott method (11) or quadratic trend method (QT). NEERGR is the growth rate of nominal effective exchange rate. PMGR is the growth rate of the import price index. PXGKRDEVED is the growth rate of the export price index of industrial countries. Table 4 Estimation Results: Non-Linear Philips Curve Dependent variable: inflatfon rate (Non-Linear Least Squares) Colombia Korea Mexico Nigeria Philippines Turkey HP QT HP QT HP QT HP QT HP QT HP QT C 0.028 0.027 0.007 0.007 0.022 0.016 0.023 0.022 0.015 0.026 0.013 0.016 (2.380) (2.291) (3.183) (3.296) (3.119) (2.205) (2.137) (1.878) (2.621) (4.395) (0.847) (1.010) INF, , 1.096 1.093 0.823 0.821 1.289 1.332 1.489 1.455 1.209 1.125 0.880 0.873 (9.531) (9.351) (27.486) (27.228) (14.253) (14.205) (15.881) (15.251) (11.375) (11.472) (17.784) (18.010) INF.-2 -0.234 -0.225 -0.446 -0.473 -0.599 -0.566 -0.448 -0.430 (-2.029) (-1.923) (-6.176) (-6.284) (-6.323) (-5.877) (-5.105) (-5.043) AGAPE) -0.001 0.002 -0.004 0.002 0.017 0.067 (-0.195) (0.105) (-0.748) (0.232) (0.724) (1.721) AGAP,.1) -0.073 -0.026 -0.022 0.026 (-0.458) (-0.774) (-0.542) (0.548) flGAPs2) 0.087 0.015 (1.605) (0.146) co 0.049 0.052 0.047 0.055 0.081 0.060 0.051 0.043 0.045 0.042 0.048 0.044 (48.498) (2.782) (19.978) (13.177) (5.713) (10.175) (27.717) (4.806) (13.894) (19.390) (15.601) (26.268) NEERGR, 0.110 0.109 0.207 0.197 0.119 0.095 0.389 0.382 (6.819) (6.795) (5.855) (6.365) (5.340) (4.564) (8.601) (8.719) NEERGR,., -0.052 -0.050 -0.080 -0.091 -0.289 -0.281 (-2.922) (-2.796) (-1.840) (-2.308) (-5.464) (-5.458) GAP,*NEERGR, 0.763 0.835 (2.120) (4.146) GAP,.,*NEERGR,, 0.642 -0.008 (2.102) (-0.189) PMGR, 0.072 0.039 (3.305) (1.804) PMGR,j 0.092 0.096 0.053 0.069 (5.019) (5.150) (1.076) (IA01) PXGR_DEVED, 0.028 0.070 (0.343) (0.829) Adj. R2 0.794 0.789 0.958 0.958 0.983 0.981 0.900 0.895 0.946 0.956 0.923 0.924 Standarderrorofregression 0.015 0.015 0.011 0.011 0.031 0.032 0.054 0.056 0.021 0.019 0.039 0.038 Total Numberof 76 76 79 79 79 79 81 81 64 64 75 75 Observations Period 1980:1 - 1998:4 1980:2 - 1999:4 1980:2 - 1999:4 1979:3 - 1999:3 1981:1 - 1999:3 1981:2 - 1999:4 Note: t-statistics are in parentheses. C is a constant term. INF is inflation rate. GAP is the log difference of output and trend component of output which is calculated using either Hodrick and Prescott method (HP) or quadratic trend method (QT).AGAP) is a finction of GAP such that (oa/(nD-GAP)}-os. ot is apsrameter identifying maximum short-run level of GDP. Its starting value is 0.049. NEERGR is the growth rate of nominal effective exchange rate. PMGR is the growth rate of the import price index. PXGR_DEVED is the growth rate of the export price index of industrial countries. Policy Research Working Paper Series Contact Title Author Date for paper WPS2493 Annuity Markets in Comparative Estelle James November 2000 A. Yaptenco Perspective: Do Consumers Get Dimitri Vittas 31823 Their Money's Worth? WPS2494 The Relevance of Index Funds for Ajay Shah November 2000 A. Yaptenco Pension Investment in Equities Kshama Fernandes 31823 WPS2495 The Australian Annuity Market David M. Knox November 2000 A. Yaptenco 31823 WPS2496 Perspectives on the Sources of Somik V. Lall November 2000 R. Yazigi Heterogeneity in Indian Industry G. Chris Rodrigo 37176 WPS2497 State Policies and Women's Monica Das Gupta November 2000 M. Das Gupta Autonomy in China, India, and the Sunhwa Lee 31983 Republic of Korea, 1950-2000: Patricia Uberoi Lessons from Contrasting Danning Wang Experiences Lihong Wang Xiaodan Zhang WPS2498 Sustaining Economic Welfare: Kirk Hamilton November 2000 L. Rivera Estimating Changes in Per Capita 82819 Wealth WPS2499 The Treatment of Non-Essential Isidro Soloaga December 2000 L. Tabada Inputs in a Cobb-Douglas Technology: 36896 An Application to Mexican Rural Household-Level Data WPS2500 Investigating Corruption Canice Prendergast December 2000 H. Sladovich 37698 WPS2501 Anti-Corruption Policies and Jeff Huther December 2000 A. Santos Programs: A Framework for Anwar Shah 31675 Evaluation WPS2502 Implications of the Currency Crisis Masahiro Kawai December 2000 A. Azarcc,n for Exchange Rate Arrangements Shigeru Akiyama 36049 in Emerging East Asia WPS2503 Proposed Strategy for a Regional Masahiro Kawai December 2000 A. AzarcoGn Exchange Rate Arrangement in Shinji Takagi 36049 Post-Crisis East Asia WPS2504 Social Capital and the Quality of Stephen Knack December 2000 P. Sintim-Abotg,. Government: Evidence from the 38526 United States Policy Research Working Paper Series Contact Title Author Date for paper WPS2505 Family Altruism and Incentives Roberta Gatti December 2000 R. Bonfield 31248 WPS2506 Ethnicity and Wage Determination Abigail Barr December 2000 R. Bonfield in Ghana Abena Oduro 31248 WPS2507 Public Expenditures and Gunnar S. Eskeland December 2000 H. Sladovich Environmental Protection: When Is 37698 the Cost of Funds Irrelevant? WPS2508 Sources of Financial Assistance for Mattias Lundberg December 2000 V. Soukhanov Households Suffering an Adult Mead Over 35271 Death in Kagera, Tanzania Phare Mujinja WPS2509 How Tax Policy and Incentives Jacques Morisset December 2000 N. Busjeet Affect Foreign Direct Investment: Neda Pirnia 33997 A Review WPS251 0 Environmental Protection and Gunnar S. Eskeland December 2000 H. Sladovich Optimal Taxation 37698