POLICY RESEARCH WORKING PAPER 3 13 9 Contracting Models of the Phillips Curve Empirical Estimates for Middle-Income Countries Pierre-Richard Aggnor Nihal Bayraktar The World Bank World Bank Institute Poverty Reduction and Economic Management Division September 2003 | POLICY RESEARCH WORKING PAPER 3139 Abstract This paper provides empirical estimates of contracting The results indicate that for both Chile and Turkey past models of the Phillips curve for four middle-income and future inflation are of about the same magnitude in developing economies-Chile, the Republic of Korea, affecting current inflation. In Korea past inflation has a the Philippines, and Turkey. Following an analytical larger impact on inflation, whereas in the Philippines it is review, models with both one lead and one lag, and two future inflation that plays a larger role. Homogeneity lags and three leads, are then estimated using restrictions are satisfied for Korea and Turkey but not Generalized Method of Moments (GMM) techniques. for Chile and the Philippines. This paper-a product of the Poverty Reduction and Economic Management Division, World Bank Institute-is part of a larger effort in the institute to understand the dynamics of inflation in developing countries. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contactMaria Gosiengfiao, room J4-280, telephone 202-473-3363, fax 202-676-9810, email address mgosiengfiao@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at pagenor @worldbank.org or nbayraktar@worldbank.org. September 2003. (24 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about aeveiopment issues. Ain ojecwive of te series is to get tbe fnaings ouL qluciz;y, euen if ebe presentations are less tban suliy pomlsnca. Tbe papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper a.rA ontire, tt-o of tbe auth-rs The¢o. not n4rcenrii- rflrcnrt tbe, iJ of th W-orJd Rank, ;ts F-eCutibe flrenonrc -r th, countries they represent. Produced by the Research Support Team Contracting Models of the Phillips Curve: Empirical Estimates for Middle-Income Countries Pierre-Richard Ag6nor and Nihal Bayraktar The World Bank Washington DC 20433 JEL Classification Numbers: E44, F32, F34 I Introduction The degree of wage and price stickiness plays an important role in the trans- mission of macroeconomic shocks. In traditional backward-looking Phillips curves, inertia in the wage- and price-contracting process is generally cap- tured by introducing measures of past inflation.' By contrast, in models of overlapping contracts with forward-looking agents, inflation is represented as a function of its expected future realizations, based on all available informa- tion about the state of the economy. Indeed, several of these models, such as the Taylor-Calvo staggered contracts approach (see Taylor (1979, 1980) and Calvo (1983)) and the quadratic price adjustment cost approach of Rotem- berg (1982), have a common formulation that is similar to an expectations- augmented Philipps curve, with current prices depend on future prices.2 Several recent studies have provided estimates of forward-looking con- tracting models (also called New Keynesian models) of the Phillips curve, most notably Fuhrer (1997), Galf and Gertler (1999), Guerrieri (2002), Roberts (1995, 2001), and Rudd and Whelan (2001). However, all of these studies fo- cus on the United states; given the conflicting nature of some of the results, it is difficult to draw general inference from them for other industrial countries, or, for that matter, developing countries. In this paper, we provide empirical 'It is also argued that past inflation is a proxy for expectations of future inflation. 2See Roberts (1995). These two approaches are often referred to as "time-dependent" models (see Ball, Mankin and Romer (1988)), because firms set their prices for fixed periods of time. An alternative approach to sticky prices relies on "state-contingent" adjustment rules, in which firms change price when underlying determinants, such as demand and costs, reach some pre-specified upper or lower bounds. See, Romer (2000) and Sutherland (1995) for a discussion, and Mankiw and Reis (2001) for an effort to move away from the sticky-price models that underlie contracting models of the Phillips curve, in understanding inflation dynamics. 2 evidence on contracting models of inflation for four middle-income develop- ing countries: Chile, Korea, the Philippines, and Turkey. These countries represent a fairly diverse experience in terms of inflation, with countries like Korea and the Philippines experiencing relatively low and stable inflation during the sample period, whereas Chile and Turkey experienced higher and more persistent episodes of inflation. Section II examines various types of backward- and forward-looking models of inflation, including specifically the model of Fuhrer and Moore (1995). Section III presents our econometric methodology, which is based on GMM estimation. We use in our empirical framework a weighted average of both past and expected future inflation, to reflect elements of both the backward- and forward-looking approaches to the Phillips curve, with the importance of each determined empirically. Section IV presents the empirical results, whereas Section V concludes. 2 Contracting Models of the Phillips Curve Much of the recent research on the dynamics of price adjustment in Keynesian models dwells on the staggered contracts models of Taylor (1979, 1980), Calvo (1983), Rotemberg (1982) and Fuhrer and Moore (1995). These models are based on the assumption that wages are set in nominal terms at discrete periods of time in an asynchronous fashion (because they are set by different agents at different points in time) and, as a result, contracts overlap. Agents are assumed to contract a wage that reflects their anticipations of future price and output levels for the expected duration of the contract. These models typically assume that prices are a constant markup over wages and focus on the persistence induced in the aggregate price (average wage) level due to 3 the asynchronous and overlapping nature of wage contracts.3 The main difference between Taylor-type contracts and Calvo and Rotem- berg models is that the individual firm's price-setting decision is derived from an explicit optimization problem. In the latter group of models, the start- ing point is an environment with monopolistically competitive firms: when it has the opportunity, each firm chooses its nominal price to maximize prof- its subject to constraints on the frequency of future price adjustments. For instance, in the quadratic price adjustment model developed by Rotemberg (1982), firms are assumed to minimize the total costs of changing prices. Nevertheless, as shown by Roberts (1995), and as discussed by Clarida, Galf, and Gertler (1999), all of these models can be formulated in the form of a forward-looking Philipps curve, of the form 7rt = alyt + a2Et7rt+1 + Et, a,i > 0, where 0 < a2 < 1, 7rt is the inflation rate, Et7rt+l denotes the one-period ahead expected inflation rate, conditional on period t information, Yt the output gap, and et a random shock with zero mean and constant variance. In this setting, it can also be shown that cal is inversely related to the degree of price rigidity; the longer prices are held fixed (on average), the less responsive is inflation to cyclical fluctuations in output. Various criticisms have been addressed to the Taylor-Calvo-Rotemberg (TCR) approach. Ireland (2001) has argued that the approach is flawed in the sense that it focused on the costs (or impossibility) of changing prices in relation to the previous period's level, instead of focusing on changes 3Note that this approach does not postulate that formal contracts are actually written, but rather that nominal prices (wages) are preset for some period of time. 4 relative to the previous period's level plus the average one-period inflation rate. Andersen (1998) has argued that focusing on price staggering, instead of wage staggering, does matter for output and price dynamics (namely, the degree of persistence of shocks), in contrast to the TCR approach in which implicitly it is assumed that there is no qualitative difference between the two cases. Most importantly for our purpose here, it has been pointed out that the TCR approach implies no persistence to the inflation rate (see Estrella and Fuhrer (2000) for a detailed discussion). It has the implausible implication that a credible disinflation program can be implemented at no output cost. Fuhrer and Moore (1995) have developed a price formation equation that can indeed generate inflation inertia. Their model can be summarized as follows. Suppose that agents negotiate nominal wage contracts that remain in effect for four quarters. Unlike Taylor (1980), however, there is no fixed markup from wages to prices. This difference is essential, because it allows a meaningful distinction between prices and wages. The aggregate log price index in quarter t, Pt, is a weighted average of the log of contract prices, Xt-i, that were negotiated in the current and the previous three quarters and are still in effect. The weights, 6i, are the proportions of the outstanding contracts that were negotiated in quarter t -i, 3 Pt = 6ixt-i, (1) i=o where bi > 0 and i = 1. Fuhrer and Moore (1995) assume that the distribution of contract prices can be characterized by a downward-sloping 5 linear function of contract length, 6i = 0.25 + (1.5 - i)s, with 0 < s < 1/6 and i = 0, ...3. This distribution depends on a single slope parameter, s, and it is invertible. When s = 0, it is the rectangular distrib- ution of Taylor (1980), and when s = 1/6, it is the triangular distribution. Let vt be the index of real contract prices that were negotiated on the contracts currently in effect, 3 Vt = bi(Xt-i - Pt-i), (2) i=o where bi is the fraction of wage contracts negotiated in period t - i that are still in effect at period t. Agents set nominal contract prices so that the current real contract price equals the average real contract price index expected to prevail over the life of the contract, adjusted for expected excess demand, measured by the output gap, yt: 3 Xt - Pt = S &iEt(vt+i + 7Yt+i) + et, (3) i=O where etis an error term. Substituting (2) in (3) yields the "relative" (or real) version of Taylor's (1980) contracting equation: 3 3 Xt - Pt = 5i (xt-i - Pt-i) + EI3iEt(xt+i - Pt+i) i=o i=o 3 +y* E 6iEt(yt+i) + (t, i=O where >2j63i+j ___ p3i = I__ Ejb a -Ej 6 Letting 7rt = Pt - Pt-1, the Phillips curve derived from the model is a two-sided curve defined as 7rt = 6(L)6(L-1)[7rt - yg9-(L)yt], where 6(L) = o0 + 61L + 62L2... is the lag polynominal that describes the distribution of price contracts in the model. In the Fuhrer-Moore model, agents in their contracting decisions care about the relative real contract price in effect during the life of their contracts. They therefore compare the current real contract price with an average of the real contract prices that were negotiated in the recent past and those that are expected to be negotiated in the near future; the weights in the average measure the extent to which the past and future contracts overlap the current one. When output is expected to be high, the current real contract price is high relative to the real contract prices on overlapping contracts.4 In contrast, the Taylor (1980) specification assumes that agents care about relative nominal contract wages (and prices) in effect during the life of their contracts. The attractive feature of the Fuhrer-Moore contracting specification is that it helps to explain the high degree of persistence in inflation that is typically found in the data, whereas the conventional Taylor specification does not. This high degree of persistence is well illustrated in Figure 1 for the case of Chile and Turkey. While the Taylor specification can be shown to imply that prices depend symmetrically on past and expected future prices, thus imparting significant inertia to the price level, it implies that the 4Note that y measures the impact of the output gap on the log real contract price, not on inflation or on the price index. The inflation rate is related to the real contract price via a complex lag/lead polynomial. 7 inflation rate is highly flexible-that is, it can jump in response to news. In contrast, the Fuhrer-Moore relative contracting specification implies that inflation depends symmetrically on past and expected future inflation, thus imparting significant inertia to both inflation and the price level. In addition, the relative contracting model, because it implies a link between the inflation rate and excess demand, can account for a correlation between demand and inflation; the Taylor model, by contrast, links the price level and excess demand and is thus not able to do so.5 With two periods, the Fuhrer-Moore contracting equation is Xt -Pt = 2x[t- --Pt-i + Et(xt+l - pt+i)] + YYt, where Yt is (the log of) the output gap. If prices are a simple average of the nominal contract wage negotiated at t and t - 1, Pt = 1(xt + Xt-1), and defining inflation as 7rt = Pt - Pt-1, we have lrt 2(rt-, + Etirt+i) + 9yt, where 9t is a moving average of current and past output. Inflation thus depends on its past value (which imparts inertia to both inflation and the price level) as well as its future value.6 5However, as shown in Figure 2, the correlation between inflation and the (lagged) output gap is neither high nor persistent in middle-income countries. 6By contrast, a two-period contracting equation of the Taylor type would imply that the contract wage is given by an average of the lagged and expected future wage contracts, 8 3 Econometric -Methodology Some early tests of New Keynesian models have attempted to take into ac- count both backward- and forward-looking elements in price setting by esti- mating an equation of the -form 7rt = W7rt-1 + (1 - IL)Et7rt+l + C!lYt- + Et, where all variables are as defined above. The relative importance of backward- and forward-looking components in inflation are thus measured by I and 1 - , respectively. Chadha, Masson and Meredith (1992), for instance, strongly reject values of 0 and lfor IL in their estimation for major industrial countries (excluding the United Kingdom), whereas Fuhrer (1997) cannot reject the assumption /s = 1 for the United States. In this study, we use a more general approach and estimate two alternative equations, in an attempt to distinguish between Taylor-type price equations and the Fuhrer-Moore specification. The Fuhrer-Moore equation is specified as follows: 7rt = 1v7rt-I + 627rt+l + alYt + ... + c&mYt_m - 771Zt-1 + et, (4) where the output gap Yt is calculated as the log difference of the actual output level and its trend component, calculated using the generalized Baxter-King adjusted- for excess demand: xt =(xt-, + Etxt+i) + yyt. This specification implies that-inflation is given by 7rt = Etxt+l + YYt. Thus, inflation persistence does not -result from the contracting specification per se but rather from the persistence of the excess demand term, yt. 9 method (see below). The optimal number of lags for y is determined by Akaike statistics. The regression equation that produces the lowest Akaike statistics is used in the analysis. Zt is the rate of change of the real effective exchange rate, defined as the difference between inflation and the sum of the rate of nominal exchange rate depreciation and the rate of imported inflation, so that a positive value of zt indicates that the real exchange rate is appreciating. We include the real exchange rate to account for the fact that our sample of countries consists of small open economies. The negative sign captures an "error correction" adjustment, in the sense that high inflation in period t - 1 (relative to foreign inflation) tends to be followed by a reduction in domestic inflation. In addition, this specification means that in the steady state, with 61 + 62 = 1, and given that Yt = Yt-, = ... = Yt-m-i = 0 in the long run, domestic inflation will be equal to the sum of the rate of nominal exchange rate depreciation and the rate of foreign inflation, that is, relative purchasing power parity will hold.7 The second price equation is a four-period, Taylor-type equation, as dis- cussed in Fuhrer (1997): It = 317rt-2 + 327rt-1 + 337rt+l + 347rt+2 + 357rt+3 (5) +0iYt + ... + 9mYt-m-i - Ozt-i, where the optimal number of lags for yt is again determined by using the Akaike criterion. If /31 + ,B2 + /33 + /34 + /53 = 1, and the output gap is zero in the long run, the inflation rate in the steady state will again be equal to 7The price equation leading to (4) could be derived by assuming that inflation in consumer prices is a weighted average of inflation in domestic prices (which are determined on the basis of a contracting framework) and foreign inflation measured in domestic- currency terms. We use the lagged value of z to reduce simultaneity bias. 10 the nominal rate of depreciation plus foreign inflation. The trend component of output is estimated by using a modified version of the "ideal" band pass filter of Baxter and King (1999). The Baxter-King filter is a linear transformation of the data, which leaves intact the components of the data within a specified band of frequencies and eliminates all other components. But this methodology has a limitation: its application requires a large amount of data. Christiano and Fitzgerald (1999) approximate the Baxter-King filter using an optimal linear approximation. Their method can be explained as follows. Let yt be the data created by applying the ideal band pass filter to the raw data, xt. yt is approximated by &, which is a filter of xt. The filter weights are chosen to minimize the mean square error: E [(yt- gt)2Ix] . Yt can be computed as Yt = Boxt + B1xt+1 + ... + BT-1-tXT-1 + BT-tXT + B1xt-1 +.+Bt-2X2 + Bt-lxl, fort=1,2,4,...,T, where Bj sin(jb) - sin(ja) j> 1 7Tj b-a 2wr 27r Bo = -, a=-, b=-, ir Pu Pi and BT-t and Bt_1 are linear functions of Bj's: 1 T-t-1 BT-t =- -B- E Bj j=1 and Bt-1 solves = Bo + B, +..+ BT-1-t +BT-t +. + Bt_2+ Bt_l, with Pu = 24 and pL = 2 in our case. Equations (4) and (5) contain future price expectations. We substitute these expected values by actual future inflation and use a Generalized Method of Moments (GMM) technique for estimation. In order to apply GMM, the moment condition that the parameter set, y, needs to satisfy is E[m(y, 7rt)] = 0, where ir is the dependent variable (inflation), and E(.) stands for the esti- mated value. The GMM estimator is obtained by minimizing the following equation, which is defined as the distance between m(.) and 0: min E m(y, 7rt, Xt, Zt)ft-'m(Y, 7rt, Xt, Zt) where X is the set of independent variables and Z the instrumental variables. Q is the weighting matrix. Here we use the lagged values of X as instrumental variables. The moment condition is written as an orthogonality condition between the residuals of the regression equation, -, and Z: m(y, rt, Xt, Zt) = Z'U(Y, Irt, Xt). The weighting matrix is estimated as the heteroskedasticity and autocor- relation consistent covariance matrix: T-1\ Q= r(o) + E(k(i,q)(I(i) - 1'(i))) 1= 12 where 1(O) T- k (E =zt-iut-izt t=i+l and k is the number of coefficients, q the bandwidth, and k(-) the Kernel function, which is included to ensure that Q is a positive semi-definite. The Bartlett Kernel is used in this study: k(i, q) { 1-(i/q) O < i < q The bandwidth is the Newey-West fixed bandwidth, which is based on the number of observations: q = integer(4(T/100)(2/9)), which is equal to 3 here. 4 Empirical Results We estimate equations (4) or (5) using quarterly data for four countries: Chile, Korea, the Philippines, and Turkey.8 The regression results are pre- sented in Table 1. The rate of change of the real exchange rate is not included in the regressions for Korea and Philippines because its coefficient had a wrong sign and was statistically insignificant. In order to decide on which price formation equation is suitable for each country, we use an F test. Equation (4) is accepted if one cannot reject the null hypothesis that ,31 = ,34 = ,35 = 0 and that the coefficients of the lagged 8See the Appendix for data sources and results of unit root tests. 9Because the data are quarterly, seasonal effects are captured by including seasonal dummies, in all regression equations. The coefficients of these variables are not reported to save space. 13 output gap (which do not appear in equation (4)) are zero. The results indicate that specification (4) is accepted for Chile and Turkey, whereas specification (5) is the correct one for Korea and Philippines. Thus, the duration of contracts appears to be relatively short in Chile and Turkey, compared to Korea and the Philippines. Interestingly enough, the first two countries are precisely those where inflation was relatively high on average during the sample period. For both Chile and Turkey, the estimated coefficients for the lead and lagged values of inflation are of about the same magnitude (as predicted by the Fuhrer-Moore specification). The sum of these coefficients is not significantly different from unity for Turkey, but not for Chile. In addition, for Chile, the estimated coefficient of the output gap measure has the right sign and is significant at the 10 percent level, whereas the rate of change of the real exchange rate has the correct sign and is highly significant. These results imply indeed that (relative) purchasing power parity holds in the long run. For Turkey, by contrast, the rate of depreciation of the real exchange is not statistically significant (despite having the expected sign), and the estimated coefficient of the output gap is wrongly sign and insignificant. For Korea, the total effect of lagged inflation on current inflation is much higher than the total effect of future inflation, whereas the reverse holds in the case of the Philippines. In both countries, the sum of coefficients on the output gap variable is positive and statistically significant, as expected. In all of these regression equations, we do not restrict the sum of the estimated coefficients for past and future inflation to be equal to unity. Us- ing the unrestricted estimates, however, we do test for the validity of the 14 constraint that weights sum to unity, in order to ensure that no long-run trade-off exists between inflation and excess demand pressures. Specifically, we use a Wald test. The results, which are also reported in Table 1, indicate that the null hypothesis that the sum of the estimated coefficients of past and future inflation is equal to unity is accepted for Korea and Turkey, but is rejected for Chile and the Philippines. 5 Concluding Remarks The purpose of this paper has been to provide empirical estimates of con- tracting models of the Phillips curve with backward- and forward-looking expectations for four middle-income developing economies-Chile, Korea, the Philippines, and Turkey. The first part reviewed various New Keyne- sian models of the Phillips curve, including those of Taylor (1979, 1980) and Fuhrer and Moore (1995). The second part presented the econometric methodology and the third the estimation results. We found that contract duration appears to be shorter in Chile and Turkey, compared to Korea and the Philippines. The output gap has a positive effect in all countries except Turkey, and the rate of change of the real exchange rate is significant only in Chile. For both Chile and Turkey, past and future inflation are of about the same magnitude, whereas in Korea (the Philippines) past inflation plays a larger (smaller) role. Homogeneity restrictions are satisfied for Korea and Turkey, but not for Chile and the Philippines. The analysis in this paper can be extended in various directions. In par- ticular, there is growing evidence for industrial countries suggesting that the inflation process may be asymmetric, in that excess demand tends to have 15 a larger effect on inflation than an equivalent degree of excess supply. Early contributions for industrial countries include Chadha, Masson and Meredith (1992), and Laxton, Meredith, and Rose (1995), with more recent studies in- cluding Dupasquier and Ricketts (1998) and Clark, Laxton, and Rose (2001). Although some preliminary work on nonlinearities in the Phillips curve has been done for developing countries (see Ag6nor (2002)), the scope for further research for these countries is significant. 16 Data Surces andi uJnt Root Taests The data used are obtained by from the quarterly database compiled by Agenor, McDermott and Prasad (2000). The quarterly dataset covers the following years for the countries in the sample: 1978:01-2001:01 for Chile and Korea, 1981:01-2000:04 for Philippines, and 1981:01-2001-01 for Turkey. The variables and the sources of them axe as follows o 7r is the percentage change in the consumner price index. Source: Inter- national Monetary Fund (IMF). o y is the log difference of output to the trend component of output, where output is the ir.dustrial production index. The trend component is calculated using a generalized version of the Baxter-King filtering method, as explained in the text. Source: IMF. o z is the rate of change of the real effective exchange rate, defined as the difference between the inflation rate and the sum of the growth rate of nominal effect exchange rate and the growth rate of import prices. A rise is a depreciation for the nominal effective exchange rate. Source: IMF. Each series used in the regressions is tested for stationarity. Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root tests are employed. The null hypothesis is that a unit root exists and the alternative hypothesis is that the series is trend stationary. Results of these tests are reported in Table Al. In general, both test statistics give similar results. The presence 17 of a unit root is rejected for all series, with most series being stationary at a 1 percent significance level. 18 Table Al Order of Integration: Unit Root Test St.atistics Country Variable ADlF test PIP test k test statistic Chile ?r 1 -2.63 -334* y 3 -3.22** -8.21*** z 1 -.3.43** -2.95** Korea 7r 0 -3.81*" y 3 -3.83***' -1.64* z G -4.37'>* -4.12*** Philippines 7r 1 4512***.99** y 0 -4.35"6 -4.2*** z 0 44*':* * .56* Turkey 7r 1 -3.14** -2.S4* y 0 -8.24";*:c* -8.24*** z C -3.5*** -3.83*** Notes: Variables and estimated period are as defined in the text. 1c denotes the number of lags in the ADF test. Asterisks *, ** and ''** denote rejection of the null hypothesis of a unit root at the 10%, 5%, and 1% significance levels. Critical values are from McKinnon (1991). References Agenor, Pierre-Richard, "Monetary Policy under Flexible Exchange Rates: An Introduction to Inflation Targeting," in A Decade of Inflation Targeting in the World, ed. by Norman Loayza and Raimundo Soto, Central Bank of Chile (Santiago: 2002). 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, 2nd. ed., Princeton University Press (Princeton, New Jersey: 1999). Andersen, Torben M., "Persistency in Sticky Price Models," European Economic Review, 42 (May 1998), 593-603. Ball, Laurence, and David Romer, "Real Rigidities and the Non-Neutrality of Money," Review of Economic Studies, 57 (April 1990), 183-204. Baxter, Marianne, and Robert G. King, "Measuring Business Cycles: Approx- imate Band-Pass Filters for Economic Time Series," Review of Economics and Statistics, 81 (November 1999), 575-93. Calvo, Guillermo A., "Staggered Contracts in a Utility-Maximizing Framework," Journal of Monetary Economics, 12 (September 1983), 383-98. Chadha, Bankim, Paul R. Masson, and Guy Meredith, "Models of Inflation and the Costs of Disinflation," IMF Staff Papers, 39 (June 1992), 395-431. Christiano, Lawrence J., and Terry J. Fitzgerald, "The Band Pass Filter," Fed- eral Reserve Bank of Cleveland, Working Paper No 9906 (July 1999). Clarida, Richard, Jordi Galf, and Mark Gertler, "The Science of Monetary Pol- icy: A New Keynesian Perspective," Journal of Economic Literature, 37 (De- cember 1999), 1661-707. Clark, Peter, Douglas Laxton, and David Rose, "An Evaluation of Alternative Monetary Policy Rules in a Model with Capacity Constraints," Journal of Money, Credit, and Banking, 33 (February 2001), 42-64. 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). Erceg, Christopher J., and Andrew T. Levin, "Imperfect Credibility and Inflation Persistence," unpublished, Board of Governors of the Federal Reserve System (June 2001). Fuhrer, Jeffrey C. "The (Un)Importance of Forward-Looking Behavior in Price Specifications," Journal of Money, Credit, and Banking, 29 (August 1997), 20 338-50. Fuhrer, Jeffrey C., and Geoff-rey R. Moore, "Inflation Persistence," Quarterly Journal of Economics, 110 (February 1995), 127-59. Gali, Jordi, and Mark Gertler, "Inflation Dynamics: A Structural Econometric Analysis," Journal of Monetary Economics, 44 (October 1999), 195-222. Guerrieri, Luca, "The Inflation Persistence of Staggered Contracts," Interna- tional Finance Discussion Paper No. 734, Board of Governors of the Federal Reserve System (August 2002). Ireland, Peter N., "Money's IRole in the Monetary Business Cycle," Working Paper No. 8115, National Bureau of Economic Reseazch (February 2001). Laxton, Douglas, Guy Meredith, and D. Rose, "Asymmetric Effects of Economic Activity on Inflation," I/iF Staff Papers, 2 (June 1995), 344-74. Mankiw, N. Gregory, and Ricardo Reis, "Sticky Information versus Sticky Prices: a Proposal to Replace the New Keynesian Phillips C'urve," Working Paper No. 8290, National Bureau of Economic Research (May 2001). Roberts, John M., "New Keynesian Economics and the Phillips Curve," Journal of Money, Credit, and Banking, 27 (November 1995), 975-84. , "How Well does the New Keynesian Sticrky-Price Model Fit the Data?," unpublished, Board of Governors of the Federal Reserve System (February 2001). Romer, David, Advanced Macroeconomics, 2nd ed., MscGraw Hill (New York: 2000). Rotemberg, Julio J., "Sticky Prices in the United States," Journal of Political Economy, 60 (November 1982), 1187-211. , "Prices, Output and Hours: An Empirical Analysis based on a Sticky Price Model," Journal of Monetary Economics, 37 (June 1996), 505-33. Rudd, Jeremy and Karl Whelan, "New Tests of the New Keynesian Phillips Curve," unpublished, Board of Governors of the Federal Reserve System (June 2001). Sutherland, Alan, "Menu Costs and Aggregate Price Dynamics," in The new macroeconomics: Imperfect Markets and Policy Effectiveness, ed. by Huw David Dixon and Neil Rakin, Cambridge University Press (Cambridge: 1995). Taylor, John B., "Staggered Contracts in a Macro Model," American Economic Review, 69 (May 1979), 108-13. , "Aggregate Dynamics and Staggered Contracts," Journal of Political Econ- omy, 88 (February 1980), 1-23. 21 Figure 1 Autocorrelation Function: inflation, Lagged Inflation 1.2 - 1.2 Korea Chile Turey 1 0.8 0.80. 060.6 1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19 20. 0.8~~~~~~~~~~~~~~~~~~. 0.4~~~~~~~~~~~~~~~~~~~. 0. 0.2 -~~~~ ~~~~~~~~~~~~0.5 JUU[ 0 -~~~~~~~~~~~~~~~~~~0.4 1 3 5 7 9 1 1 13 15 17 19 1 3 5 7 9 1 1 13 15 17 19 1.2 Philippines 1 Turke22 Figure 2 Cross-correlation Function: Inflation, Lagged Output Gap 0.1 Chile 0.15 Korea 0.1 0.05 0.05 0 Fn -0.05 -0.05 -0.1~~~~~~~~~~~~~~~~~~~~-. -0.1~~~~~~~~~~~~~~~~~01 -0. 15 -0.2 l l l 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0.4 Philippines 0.15 Turkey 0.3 0.1 0.2 0.1~~~~~~~~~~~~~~~~00 0- -0.05 -0.1 -0.2 0.1i 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 20 0 2 4 6 8 1 0 1 2 14 16 18 20, 2 3 Table I Determinants of Inflation Chile Korea Philippines Turkey 7C t- 10.529 ... ... 0.497 (13.961) (4.254) 7t t+ 1 0.528 ... ... 0.489 (13.464) (9.422) S7r ; ... 0.642 0.361 ... (8.789) (16.245) a + i ... 0.393 0.552 ... (4.643) (17.992) 7-y ti X0.093 0.055 0.086 -0.429 (1.620) (2.344) (2.571) (-0.923) Zt -I -0.028 ... ... -0.022 (-3.741) (-0.774) Adjusted R2 0.959 0.816 0.935 0.898 see 1.695 1.097 3.411 7.237 Number of observations 80 75 50 73 Number of lags for yt 3 6 7 1 (calculated using Akaike criterion) F test for 1.49 4.18 37.15 0.05 Ho: use only i t and i t +1 H 1: use 7r t 1 n t -2 7l t +1, n t +2' It t +3 Accept Ho Reject Ho Reject Ho Accept Ho Wald test for 81+52 = I Reject ... ... Accept Wald test for 1+02+0 = I ... Accept Reject ... The estimation technique is the generalized method of moments. The instrumental variables are the lagged values of the independent values. a ,., and n t+1 stand for a lag of inflation and a lead of inflation, where inflation is defined as the percentage change in the consumer price index. En t stands for the sum of coefficients of n .1 and n t2a En t I represents the sum of the estimated coefficients of n ,1,, n1 +2. and 7 t13 Yt is the deviation of actual capital from its trend component calculated using Baxter and King generalized filtering method. Eyt., represents the sum of coefficients of y, and its lagged values, where the number of lags presented in the 'number of lags for y, row. z,.1 is the lagged value of the rate of change of the real effective exchange rate and it is equal to the difference between n 1 and the sum of the growth rate of nominal effect exchange rate and import price. A rise is a depreciation for nominal effective exchange rate. see is the standard error of regression. &1 is the estimated coefficient of n I . I and 82 Is the estimated coefficient of a, + 1. Test for 81+82=1 reports the test result for Ho: 1+82=1. fr, P2. 3, 04, and j5 are the estimated coefficients of n t-2, 71 .1, n 1, ,7 t+2, and a1,+3, successively. Test for PI+02+03+P4+05 = 1 reports the test result for Ho: P1+02+P3+P4+P5 = 1 24 Policy Research Working Paper Series Contact Title Author Date for paper WPS3116 Dollarization of the Banking System: Gianni De Nicol6 August 2003 A. Yaptenco Good or Bad? Patrick Honohan 38526 Alain Ize WPS3117 Policy Research on Migration and David Ellerman August 2003 B. Mekuria Development 82756 WPS3118 To Share or Not to Share: Does Local Beata Smarzynska August 2003 P. Flewitt Participation Matter for Spillovers from Javorcik 32724 Foreign Direct Investment? Mariana Spatareanu WPS3119 Evaluating the Impact of Conditional Laura B. Rawlings August 2003 M. Colchao Cash Transfer Programs: Lessons Gloria M. Rubio 38048 from Latin America WPS3120 Land Rights and Economic Quy-Toan Do August 2003 P. Sader Development: Evidence from Vietnam Lakshmi lyer 33902 WPS3121 Do Bilateral Investment Treaties Mary Hallward-Driemeier August 2003 A. Bonfield Attract Foreign Direct Investment? 31248 Only a Bit ... and They Could Bite WPS3122 Individual Attitudes Toward Roberta Gatti August 2003 N. Obias Corruption: Do Social Effects Matter? Stefano Paternostro 31986 Jamele Rigolini WPS3123 Production and Cost Functions and Beatriz Tovar August 2003 G. Chenet-Smith Their Application to the Port Sector: Sergio Jara-Diaz 36370 A Literature Survey Lourdes Trujillo WPS3124 The Impact of Structural Reforms on Neil McCulloch August 2003 M. Faltas Poverty: A Simple Methodology with 82323 Extensions WPS3125 Economic Analysis of Health Care Vicente B. Paqueo August 2003 R. Guzman Utilization and Perceived Illness: Christian Y. Gonzalez 32993 Ethnicity and Other Factors WPS3126 Public Disclosure of Environmental Jong Ho Hong August 2003 Y. D'Souza Violations in the Republic of Korea Benoit Laplante 31449 Craig Meisner WPS3127 Small and Medium Enterprises Meghana Ayyagari August 2003 A. Yaptenco Across the Globe: A New Database Thorsten Beck 31823 Asli Demirgu9-Kunt WPS3128 Child Growth, Shocks, and Food Aid Takashi Yamano August 2003 H. Sladovich in Rural Ethiopia Harold Alderman 37698 Luc Christiaensen WPS3129 Price Caps, Efficiency Payoffs, and Antonio Estache August 2003 A. Estache Infrastructure Contract Renegotiation Jose-Luis Guasch 81442 in Latin America Lourdes Trujillo WPS3130 The Role of Advocacy in Competition Tomas Serebrisky September 2003 G. Chenet-Smith Policy: The Case of the Argentine 36370 Gasoline Market policy lesearch Working Papa sefiCa Contact Title Author Date for paper WPS3131 Social Sector Expenditures and Christian Y. Gonzalez September 2003 P. Holt Rainy-Day Funds Vicente B. Paqueo 37707 WPS3132 Regional Integration and Technology Maurice Schiff September 2003 P. Flewitt Diffusion: The Case of the North Yanling Wang 32724 America Free Trade Agreement WPS3133 Emerging Trends in WTO Dispute Peter Holmes September 2003 P. Flewitt Settlement: Back to the GATT? Jim Rollo 32724 Alasdair R. Young WPS3134 Institutional Reform and the Roumeen Islam September 2003 R. Islam Judiciary: Which Way Forward? 32628 WPS3135 Trade Reforms, Market Access, Guido G. Porto September 2003 P. Flewitt and Poverty in Argentina 32724 WPS3136 Legal Institutions and Financial Thorsten Beck September 2003 A. Yaptenco Development Ross Levine 31823 WPS3137 Using Survey Data to Assess the Guido G. Porto September 2003 P. Flewitt Distributional Effects of Trade Policy 32724 WPS3138 Fiscal Federalism and Regional Raj M. Desai September 2003 V. Sapinoso Growth: Evidence from the Russian Lev M. Freinkman 81105 Federation in the 1990s Itzhak Goldberg