Inflation and Interest Rate Differentials Between Germany and its EMS Partners

Helmenstein, Christian; Rünstler, Gerhard Article Inflation and Interest Rate Differentials Between Germany and its EMS Partners Zeitschrift für Wirtschafts- und Sozialwissenschaften (ZWS) - Vierteljahresschrift der Gesellschaft für Wirtschafts- und Sozialwissenschaften, Verein für Socialpolitik Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Helmenstein, Christian; Rünstler, Gerhard (1996) : Inflation and Interest Rate Differentials Between Germany and its EMS Partners, Zeitschrift für Wirtschafts- und Sozialwissenschaften (ZWS) - Vierteljahresschrift der Gesellschaft für Wirtschafts- und Sozialwissenschaften, Verein für Socialpolitik, ISSN 0342-1783, Duncker & Humblot, Berlin, Vol. 116, Iss. 4, pp. 593-613, https://doi.org/10.3790/schm.116.4.593 This Version is available at: https://hdl.handle.net/10419/291876 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Zeitschrift für Wirtschafts- u. Sozialwissenschaften (ZWS) 116 (1996), S. 593 - 613 Duncker & Humblot, Berlin Inflation and Interest Rate Differentials Between Germany and its EMS Partners* By Christian Helmenstein and Gerhard Runstler** 1. Introduction Most observers agree that the European Exchange Rate Mechanism (ERM) has been a major factor in prompting its inflation-prone member countries to disinflate. In a series of papers Giavazzi and Giovannini (1989, 1985) forcefully argued that the ERM is best described as an asymmetric exchange rate system. Germany as the center country of this system sets its monetary policy independent of other ERM members who are bound to accommodate the German policy stance or, if unwilling to do so, see their exchange rates depreciate relative to the Deutsche Mark (DM). Expressing this notion, the ERM has frequently been called a "greater Deutschmark area" (Herz and Roger, 1992; Giavazzi and Giovannini, 1987). Recently, several contributions have approached the issue of asymmetries empirically The majority of the empirical work suggests that the monetary policy adopted by the German Bundesbank plays a predominant role in the European Monetary System (EMS) (Funke and Hall, 1994; Kirchgassner and Wolters, 1993; von Hagen and Fratianni, 1990). Hoping to borrow monetary credibility from the Bundesbank, countries with higher inflation rates than Germany pegged the respective exchange rate to the Deutsche Mark (Giavazzi and Pagano, 1988).1 Persistent inflation differentials in combination with fixed exchange rates, however, led to large swings in competitiveness across Europe. These were at least partially offset by a series of realignments between 1980 and 1987 but the 1992 EMS turbulence, following a period of apparently stable parities, virtually abandoned the concept of a comprehensive fixed exchange rate regime with temporary realignments. In the present paper we address the expectations of market participants on EMS realignments. Inflation differentials cause market participants to * Verantwortlicher Herausgeber / editor in charge: J. W. ** Institute for Advanced Studies, Vienna. The authors gratefully acknowledge suggestions by Bernhard Felderer, Andreas Wörgötter, and two anonymous referees. 1 For a critical comment see Obstfeld (1988). ZWS 116 (1996)4 38 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 594 Christian Helmenstein and Gerhard Rünstler entertain expectations on corrective exchange rate adjustments. While a floating exchange rate regime allows for such an adjustment to occur instantaneously, a fixed exchange rate system may in the short-run prevent the expected exchange rate adjustments to materialize. Anticipating the impending realignment, rational market participants will therefore demand a compensating return on foreign-currency assets which should cause the emergence of a new, or the widening of an existing interest rate differential. We suppose long-term interest rate differentials between Germany and its EMS partners to serve as a proxy for exchange rate expectations and examine their link with inflation differentials. If market participants are suspicious about the sustainability of a narrowed inflation differential, it may be expected that for a country with lower credibility a decrease in the inflation differential entails a smaller effect on interest rate differentials than a similar decrease for a country with higher credibility. Subsequently we consider three different aspects of the relationship between interest rate and inflation differentials. First, both interest rate and inflation differentials appear to be non-stationary for the majority of the countries under study. However, if realignments are intended to offset changes in competitiveness, one should find a stationary long-run relationship between the two. Second, by affecting market expectations on future realignments, inflation differentials should Granger-cause interest rate differentials. Third, asymmetries within the EMS suggest an asymmetric response of realignment expectations to deviations from the equilibrium exchange rate. If inflation differentials vis-à-vis Germany have driven the exchange rate of an EMS member country currency below its equilibrium level, market participants would nevertheless expect a nominal appreciation to be unlikely. This notion translates into a slower adjustment of interest rate differentials to long-run equilibrium in response to a negative deviation. Observations pertaining to the 1990-1993 period back these considerations. In the aftermath of unification, Germany found itself in the company of countries that were traditionally underperforming in terms of price level stability such as Italy, whereas France and Denmark performed remarkably better. Despite the narrowing and even the reversion of the inflation differential, the nominal interest rate differentials to Germany were, at best, eliminated but did not turn negative. The spread between ex-post real interest rates across countries thus widened. The paper is organized as follows. The subsequent chapter 2 discusses the theoretical link between inflation differentials and interest rate differentials. In particular, we draw upon the concepts of uncovered interest parity and purchasing power parity to derive the principal equation of our econometric model. Chapter 3 introduces the econometric model. We conduct a cointegration and Granger causality analysis for long-term interest rate and ZWS 116 (1996)4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 595 inflation differentials of EMS members vis-à-vis Germany using a bivariate vector autoregression (VAR) framework. The empirical results are presented in chapter 4. Chapter 5 concludes the paper. 2. Inflation and Interest Rate Differentials: A Theoretical Link Uncovered interest parity (UIP) as well as real interest parity (RIP) are concepts which link financial markets in different countries.2 Uncovered interest parity holds if the nominal interest rate of currency A, if, plus its expected appreciation (or depreciation, respectively) equals the nominal interest rate of currency B, if, (1) UIP: if-if =Et(St+1-St) or (2) UIP : dit = Se t+1 - St with Et denoting the conditional expectations operator with respect to the information set available at period t, dit the nominal interest rate differential in period t, St the spot rate of currency B in terms of currency A (in logs), and the value of St+1 expected as of period t. Short-term deviations from UIP (nominal interest parity) depend on three different factors. First, capital controls together with the risk of future controls - political risk according to Aliber (1973) - may lead to the emergence of a country premium on returns across national markets. Second, as the ex-post return of portfolio investments and of direct investments strongly depends on future exchange rates, the volatility of exchange rates may drive wedges between ex-ante returns in different markets and may therefore give rise to an exchange risk premium in excess of the expected appreciation (or depreciation) (Marston, 1993).3 Post-tax yield differences constitute a third reason for deviations from nominal interest parity, and thus the average uncovered interest differential should equal the sum of country and exchange risk premiums as well as post-tax yield differences (Isard, 1991). Since expected future exchange rates are not directly observable, in testing the empirical validity of UIP the most important strand of studies draws 2 See Hodrick (1987) and Cumby and Obstfeld (1984) for reviews. An extensive individual study on the degree of financial market integration as measured by deviations from UIP has been provided by Frankel and Mac Arthur (1988). 3 In terms of the EMS before enlarging the bands the primary risk was that of future realignments. ZWS 116 (1996)4 38! OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 596 Christian Helmenstein and Gerhard Rünstler upon the hypothesis that forward exchange rates should serve as unbiased predictors of future spot exchange rates (Froot and Frankel, 1989; Fama, 1984). Disregarding transaction costs, the absence of riskless arbitrage possibilities implies that where Ft denotes the forward exchange rate.4 Tests of the joint hypothesis that both this covered interest parity (CIP) and the unbiasedness hypothesis hold, however, have regularly been rejected (MacDonald and Taylor, 1989; Meese, 1989). Reexamining the uncovered interest parity relationship, McCallum (1992) confirms the earlier results on forward rate biasedness but demonstrates that this evidence does not necessarily entail rejection of UIP. If monetary authorities are assumed to manage interest rate differentials so as to avoid rapid changes in exchange rates and in the related differentials, the failure of unbiasedness is nevertheless consistent with UIP. The results of empirical studies on Euromarkets indeed suggest that deviations from uncovered interest parity are mainly due to a country premium associated with capital controls or political risk and post-tax yield differences (Marston, 1993; Dooley and Isard, 1980) and that UIP holds in the long run. Long-run exchange rate expectations, in turn, are related to cross-coun- try inflation differentials. In its weak form purchasing power parity (PPP) states that inflation differentials are equal to the change of the nominal exchange rate, with pA (pB) as the inflation rate of country A (B) and dp as the inflation differential between countries A and B. While tests of PPP for floating exchange rate regimes provide mixed evidence (Juselius, 1995; MacDonald, 1995),5 for a managed exchange rate regime like the EMS the validity of PPP in its weak form is subject to discretionary policy intervention. Subtracting the expected inflation differential Et(dpt+i) from both sides of equation (1), we obtain an expression for ex-ante real interest parity, 4 Empirical support for this relationship is surveyed in MacDonald (1988) and Levich (1985). 5 Tests of the strong form of purchasing power parity regularly fail to give evidence in favor of its validity. Several reasons have been advanced to explain this finding: Consumer price indices might differ due to the inclusion of non-traded goods or the application of different weighing schemes across countries (Dornbusch, 1985). A reason other than such measurement problems is differences in the elasticity of substitution in consumption across countries which allow firms to exert some degree of price discrimination. If a firm can at least partially control the cross-border trade in its good(s), the real exchange rate can deviate from one (Tootell, 1992). (3) CIP : dit = Ft-St (4) PPP : p?+1 - p?+1 = dpt+1 = St+i - St , ZWS 116 (1996)4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 597 (5) RIP : drt = Et(Sm - St) - Et(dpt+1) . If we assume that the risk of capital controls and post-tax yield differences is either negligible or constant over time, the hypothesized direct link between nominal interest rate differentials and inflation differentials may be caused by deviations from PPP, time-varying exchange risk premia or both (Capitelli and Schlegel, 1991). It seems reasonable to suppose that market participants expect cumulated inflation differentials in the long run to be offset by successive realignments which counteract the deterioration of competitiveness. Therefore time-varying exchange risk premia should have no effect on uncovered interest parity over long sample periods which is tantamount to stating that the right hand side of equation (5) equals zero. We thus have (6) dtt-Et(dpt+i) = 0 which represents the principal equation for our econometric model. Since PPP is a long-term concept, equation (6), as it stands, stipulates a long-run relationship between inflation and interest rate differentials. Deviations from PPP induce an adjustment process to equilibrium. To capture the associated short-run dynamics, we draw upon a dynamic model specification that will be introduced subsequently. 3. Methodology Consider a 2-dimensional Vector Autoregressive (VAR) representation of order m for interest rate and inflation differentials (dity dpt), m fdit\ = Z/xA (an{L) a12{L)\ (dit-A felt\ { ) \dptj UJ Wi W a22(W \dpt-i) + W where n = (¿¿i,/^)' is a vector of constant drift, and et = (en,¿21)' is white noise.6 Furthermore it is assumed that the roots of the characteristic polynomial are strictly outside the unit circle or equal to one, |I-A(z)| = 0 => \z\ > 1. Crucial for the behavior of the system is the number of unit roots in A(L). If there are no unit roots, that is, |I - A(z)| = 0 \z\ > 1, the VAR (7) is stable and (dit, dpt) follow a stationary process. If there are two unit roots, the system is nonstationary while in the intermediate case of one unit root the system can be rewritten in the vector error correction (VEC) 6 In order to derive (7) from (6), it is necessary to assume static expectations, Et(dpf+i) = dpt. ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 598 Christian Helmenstein and Gerhard Rünstler representation with one error-correction term (see Engle and Granger, 1987), • (£)-©•ess';;£)«a+CSK-—• C;::) where bij(L) are lag-polynomials of order m- 2. This is equivalent to the case that both dpt and dit, taken individually, have a unit root and are thus integrated processes while there exists a stationary linear combination dit ~ 0dpt, that is the series are cointegrated. The adjustment coefficents a* describe the speed of adjustment of the particular series to the long-run equilibrium while the lag polynomial b(L) represents additional short-run dynamics. Engle and Granger (1987) proposed a two-step estimator of (8). The first step involves a regression of dpt on dit and a subsequent test for the stationarity of the residual. In the case of cointegration the estimated equilibrium error dit - ¡3dpt is inserted into equation (8), and b(L) and a = (ai,a2)/ are estimated by OLS. This procedure is less efficient than the maximum likelihood estimator proposed by Johansen (1991). In order to analyze the time-related interaction between the two series, we perform Granger-causality tests (Granger; 1969). Granger-causality from dpt to dit means that the conditional forecast for dit can be significantly improved by adding lagged dpt to the information set. This amounts to a nonzero polynomial ai2(L) in the bivariate VAE (7). The feasibility of Granger causality tests depends on the stationarity features of the system. If the series are stationary, the null hypothesis of no causality can be tested by standard Wald tests (e.g. Liltkepohl, 1991). With integrated processes the situation generally becomes intractable as the distribution of the Wald statistic is nonstandard (Toda and Philips, 1993). However, for the case of a bivariate cointegrated system the standard distribution theory applies and conventional test procedures can be used. Toda and Philips (1991) have pointed out that the null hypothesis of dpt not Granger-causing dit is equivalent to bu (L) = 0 and a\ = 0 in equation (8). They proposed to test each subhypothesis sequentially using the same significance level. Their simulations show that this method is in many cases superior to the straightforward VAR test. As a side-effect this procedure allows a distinction between the shortrun and the long-run dynamics. As b(L) describes only the short-run dynamics, ai = 0 implies that dit is not Granger-caused by dpt in the long-run while there might be dependencies in the short-run. Note that the Engle- Granger representation theorem implies that at least one au is non-zero in the bivariate system (8). Granger and Lee (1989) propose the nonsymmetric error-correction model ZWS 116 (1996)4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 599 » (¡a-C)•«<£)+•(:;>*- - *(:;:) where (dit-/3dpt)+ = max(dit-(3dpt, 0) and (dit-f3dpt)~ = mm(dit-(3dpt, 0). This model allows for different adjustment coefficients for positive and negative deviations from the cointegration relationship. A higher absolute value of the coefficient for positive deviations would imply that the corresponding series responds more quickly to a positive equilibrium error compared to a negative one. Since the estimator for ¡3 is super-consistent, testing for a+ = a~ can be done by estimating ¡3 in the context of a linear model, inserting ¡3 into (9), and applying the usual Wald test. For positive and negative deviations to be defined, the mean of the cointegrating relationship must be identified. The problem relates to the constant /i in (9). If = amo (i = 1,2), equation (9) can be rewritten as <io> O=b(L)(AAtO+(;>+- ^++(:;;) • Only in this case the mean of the cointegration relationship is well defined and so are positive and negative deviations. Note that the rejection of (10) would imply that a linear trend component is present in inflation differentials and/or in interest rate differentials.7 4. Estimation and Results Using monthly data from 1980:1 to 1994:98, in the present paper we investigate the 10-year government bond yield differential and the inflation differential vis-à-vis Germany for Austria, Belgium, Denmark, France, Italy, and the Netherlands.9 Inflation is measured by the consumer price index, all data are taken in first differences of logs. With the exception of Austria, this selection of countries draws upon their membership in the EMS which was formed in 1979. Instead of joining the EMS, Austria opted for a rigorous pegging of the Austrian Schilling to the Deutsche Mark during the same 7 Generally the constant term |i can be decomposed into two parts, |i0, which represents the intercept in the cointegrating relationship, and the remainder, |i-|io, which determines a linear trend in dit and dpt (Johansen, 1991). 8 For Denmark: 1983:6 to 1994:9. The estimation period ranges from 1981:6 to 1994:9 (for Denmark: 1984:10 to 1994:9). Data source: International Financial Statistics of the International Monetary Fund. 9 Luxembourg has been excluded from consideration as it forms a currency union with Belgium. ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 600 Christian Helmenstein and Gerhard Rünstler period which we assume economically equivalent to being a formal EMS member country. Figures 1-6 display strikingly close comovements between inflation differentials and government bond yield differentials between Germany and the other European countries which joined the EMS before 1981. Except for Italy, these countries are characterized by a comparatively high degree of integration as measured by the correlation of demand and supply disturbances with Germany (Helmenstein and Url, 1993). The graphs suggest the division of the countries into two groups. Belgium, Denmark, France, and Italy (Figures 1-4) faced high inflation differentials vis-a-vis Germany at the beginning of the eighties which entailed several devaluations of their currencies between 1980 and 1987. The reduction of the inflation differentials was accompanied by a parallel decrease of the interest rate differentials. Nevertheless, financial markets seem to have retained some suspicion about the stability of the EMS parities after 1987. This conclusion can be inferred from the observation that an increase of the inflation rate relative to Germany continued to induce an increase in the interest rate differentials. By contrast, a reversal in the inflation differentials towards Germany for the first time since 1980 paired with still positive (or at best zero) interest rate differentials led to a remarkable gap between real interest rates. Supposing that real interest parity holds, lower inflation rates in the neighboring countries to Germany imply that the German inflation rate is expected to return to the conventional level of inflation in the future, that is, the Bundesbank is expected to exert a comparable degree of monetary stability as before unification. Austria and the Netherlands (Figures 5 and 6), which have maintained a stable exchange rate to Germany during the eighties, form the second group. Temporary inflation differentials seem to have exerted little impact on interest rate differentials. Figure 6 demonstrates that in the case of the Netherlands positive inflation differentials have always been quickly eliminated. We test the null hypothesis of a unit root by Augmented Dickey-Fuller tests (ADF-tests) (Dickey and Fuller, 1981) with a constant but without a trend. The inclusion of a trend is economically implausible as this would imply continuously widening interest rate differentials. The outcome of this test may depend on the order of the approximating autoregression. Taking up the advice of Ng and Perron (1993), we select the lag length by a sequence of t-tests starting with an autoregression of order 4. Table 1 presents the results of ADF-tests for interest rate differentials and inflation differentials with Germany. For interest rate differentials the hypothesis of a unit root cannot be rejected at the 10% level for any country under study, while for inflation differentials this holds true for all countries except for Austria and ZWS 116 (1996)4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 607 Table 3 Granger-Causality From Inflation Differentials to Interest Rate Differentials (F-tests) Country Number of lags Error-correction terms H0: a, = 0 Short-run dynamics Ho:ML) = 0 Both H0: a12 = ML) = 0 Results for the VEC (8): Belgium (81:9-94:9) (85:1-94:9) 4 3.41* 3.09* 1.74 1.38 1.72 1.37 Denmark 1 11.97*** 1.10 6.34*** France (81:9-94:9) (85:1-94:9) 2 7.91*** 5.11** 0.37 0.93 2.91** 2.26* Results for the VAR (7) in levels: Austria 1 n.a. n.a. 4.04*** Netherlands 4 n.a. n.a. 0.98 *** significant at the 1 %-level. ** significant at the 5 %-level. * significant at the 10 %-level. Table 4 Granger-Causality From Interest Rate Differentials to Inflation Differentials (F-tests) Country Number of lags Error-correction terms H0: a2 = 0 Short-run dynamics H0: ML) = 0 Both H0: b2, = ML) = 0 Results for the VEC (8): Belgium (81:9-94:9) (85:1-94:9) 4 7.01*** 0.95 2.99** 1.52 3.30*** 1.28 Denmark 1 0.04 5.37** 3.15** France (81:9-94:9) (85:1-94:9) 2 5.70** 0.03 0.94 2.04 3.46** 1.58 Results for the VAR (7) in levels: Austria 1 n.a. n.a. 0.25 Netherlands 4 n.a. n.a. 1.55 *** significant at the 1 %-level. ** significant at the 5 %-level. * significant at the 10 %-level. With Belgium and Denmark only two countries fulfill the preconditions for an application of the non-symmetric error-correction model (10). Wald tests for non-symmetric adjustment (H0: a+ = a"), however, do not indicate asymmetries (Table 5).11 ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 608 Christian Helmenstein and Gerhard Rünstler Table 5 Wald Tests for Asymmetries in the Response to Equilibrium Errors Country Interest rate differentials H0: a| = a, Inflation differentials H0: a+2 = a2 Belgium 0.15 0.45 Denmark 0.66 2.01 The Wald statistic is x2 distributed with one degree of freedom. *** significant at the 1 %-level; ** significant at the 5 %-level; * significant at the 10 %-level. 5. Conclusions The results of the foregoing analysis largely confirm the hypothesis that inflation differentials and interest rate differentials are indeed cointegrated. The relationship between inflation and interest rate differentials seems to be affected by the country-specific degree of credibility of monetary and fiscal policies. Recent literature stresses the importance of credibility in the impact of temporary imbalances in fiscal and monetary policies on the exchange rate, pointing out that the build-up of credibility is a long-lasting process. While it is difficult to quantify the concept of credibility, most observers will probably consent to an indicative ordering of EMS countries. The Netherlands and Austria are well-known for a comparatively high degree of fiscal and monetary stability that is reflected by relatively sound fiscal balances, low inflation rates, and a lasting exchange rate peg to the DM that is also rationalized by their close trade links with Germany. By contrast, the other countries in the sample were more inflation-prone for the most part of the period under consideration. Belgium, Denmark, and Italy also carried high fiscal deficits during the early eighties. While Denmark and, to a minor extent, Belgium were able to reduce their net lending significantly during the eighties, public debt decreased in Denmark and stabilized in Belgium from 1987 on. In Italy, fiscal deficits have remained at 11 The above results may not be correct if there is more than one cointegrating relation within the four-dimensional system of the inflation and interest rates of both countries. We checked for this possibility by applying the Johansen procedure to the four-dimensional system. Several authors pointed out (e.g. Runstler, Jumah, and Karbuz, 1995; Urbain, 1993) that the efficiency of the Johansen method substantially declines with the dimensionality of the system which justifies the usage of the differentials themselves. In estimating the four-dimensional system for Denmark, the Netherlands, and Italy we find no cointegrating vector at all while there appears to be one for Austria, Belgium, and France. For Austria see also Wôrgotter (1992). ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 609 high levels with public debt rising throughout. Taken together, the credibility of the Danish, French, Belgian, and Italian exchange rate peg would thus be considered weaker than the Austrian or Dutch. We find that countries with highly credible policies have both stationary inflation and interest rate differentials with inflation Granger-causing interest rate differentials only for Austria. The returns of the Dutch and the Austrian hard-currency policy accrue in times of temporarily higher inflation rates vis-à-vis Germany (1991 and 1990 respectively) during which the interest rate differentials remain unaffected. The findings of the causality analysis back this conclusion. For EMS members with less credibility, by contrast, the respective differentials seem to be better characterized as nonstationary, with a cointegrating relationship arising from expected devaluations that are in line with cumulated inflation differentials. For Italy, the country with the least stable policies, a cointegrating relationship cannot be established, however. This may be due to a widely fluctuating risk premium component apart from inflation differentials. The hypothesis of real interest parity was rejected for Belgium and France at the 5% level and for Denmark at the 10% level. In all three cases the coefficient in the cointegrating relationship, B, was found to be smaller than one. It thus appears that interest rate differentials did not fully respond to inflation differentials or, in other words, real interest parity does not hold as stipulated by (6). Given that market participants expected devaluations to be equal to inflation differentials, nominal interest rate differentials did not fully offset expected devaluations. This finding suggests that within the EMS capital was not perfectly mobile during the estimation period. Two more findings deserve further discussion. First, our estimates indicate that for Belgium, Denmark, and France from the mid-eighties on inflation differentials have Granger-caused inflation differentials but not vice versa while in the earlier period there was bidirectional Granger-causality. This change of the causality pattern coincides with the beginning of effective fiscal consolidation around 1984. Other factors than inflation differentials might therefore have played a major role in the formation of interest rate differentials, that is, large budget deficits may strongly influence the conduct of capital market participants. Second, despite higher inflation rates in Germany than in most of the neighboring countries, interest rate differentials did not become negative in the early nineties. That is, Germany could draw on its „credibility capital" during the phase of comparatively high inflation rates due to extensive government spending in the course of German unification. Applying a nonsymmetric error-correction model, however, we did not find evidence for an ZWS 116 (1996) 4 39 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 610 Christian Helmenstein and Gerhard Riinstler asymmetric response of interest rates to positive and negative deviations from the long-run equilibrium. References Aliber, R. (1973): The Interest Rate Parity Theorem: A Reinterpretation, Journal of Political Economy, vol. 81, pp. 1451-1459. Capitelli, R. / Schlegel, A. (1991): Ist die Schweiz eine Zinsinsel? Eine multivariate Untersuchung über den langfristigen internationalen Realzinsausgleich, Schweizerische Zeitschrift für Volkswirtschaft und Statistik, vol. 127, pp. 647-664. Cumby, R. / Obstfeld, M. (1984): International Interest Rate and Price Level Linkages Under Flexible Exchange Rates: A Review of Recent Evidence, in: J. Bilson and R.C. Marston (eds.): Exchange Rate Theory and Practice, University of Chicago Press: Chicago, pp. 121-151. Dickey, D. A. / Fuller, W. A. (1981): The Likelihood Ratio Statistics for Autoregressive Time Series with a Unit Root, Econometrica, vol. 49, pp. 1057-1072. Dooley, M. P. / Isard, P. (1980): Capital Controls, Political Risk, and Deviations from Interest-Rate Parity, Journal of Political Economy, vol. 88, pp. 370-384. Dornbusch, R. (1985): Purchasing Power Parity, NBER Working Paper no. 1591, National Bureau of Economic Research: Cambridge (MA.). Engle, R. E / Granger, C. W. J. (1987): Co-integration and Error Correction: Representation, Estimation and Testing, Econometrica, vol. 55, pp. 251-271. Fama, E. E (1984): Forward and Spot Exchange Rates, Journal of Monetary Economics, vol. 14, pp. 319-338. Frankel, J. A. / Mac Arthur, A. (1988): Political vs. Currency Premia in International Real Interest Rate Differentials: A Study of Forward Rates for 24 Countries, European Economic Review, vol. 32, pp. 1083-1114. Froot, K. A. / Frankel, J. A. (1989): Forward Discount Bias: Is it an Exchange Rate Premium?, Quarterly Journal of Economics, vol. 104, pp. 139-161. Funke, M. / Hall, S. (1994): Is the Bundesbank Different from Other Central Banks: A Study Based on P*, Empirical Economics, vol. 19, pp. 691-707. Giavazzi, F. / Giovannini, A. (1985): Asymmetries in Europe, the Dollar and the European Monetary System, in: Giovannini, A. and R. Dornbusch (eds.): Europe and the Dollar, Istituto Bancario San Paolo di Torino: Turin. - (1987): Models of the EMS: Is Europe a Greater Deutsche-Mark Area?, in: Bryant, R. and R. Portes (eds.): Global Macroeconomics: Policy Conflict and Cooperation, St. Martin's Press: New York. - (1989): Limiting Exchange Rate Flexibility, MIT Press: Cambridge (MA.). Giavazzi, F. / Pagano, M. (1988): The Advantage of Tying One's Hands: EMS Discipline and Central Bank Credibility, European Economic Review, vol. 32, pp. 1055- 1075. Granger, C. W. J. (1969): Investigating Causal Relations by Econometric Models and Cross-Spectral Methods, Econometrica, vol. 37, pp. 424-438. ZWS 116(1996)4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 611 Granger, C. W. J. / Lee, T. H. (1989): Investigation of Production, Sales and Inventory Relationships Using Multicointegration and Non-symmetric Error Correction Models, Journal of Applied Econometrics, vol. 4, pp. 145-159. Hagen, J. von / Fratianni, M. (1990): German Dominance in the EMS: Evidence from Interest Rates, Journal of International Money and Finance, vol. 9, pp. 358-375. Helmenstein, C. / Url, T. (1993): Kleine Wahrungsunion in Europa, Konjunkturpolitik, vol. 39, pp. 251-263. Herz, B. / Roger, W. (1992): The EMS Is a Greater Deutschmark Area, European Economic Review, vol. 36, pp. 1413-25. Hodrick, R. (1987): The Empirical Evidence on the Efficiency of Forward and Futures Foreign Exchange Markets, Harwood Academic Publishers: New York. Isard, P. (1991): Uncovered Interest Parity, IMF Working Paper No. 91/51, International Monetary Fund: Washington, D.C. Johansen, S. (1991): Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vectorautoregressive Models, Econometrica, vol. 59, pp. 1551-1580. Johansen, S. / Juselius, K. (1990): Maximum Likelihood Estimation and Inference on Cointegration - With Applications to the Demand for Money, Oxford Bulletin of Economics and Statistics, vol. 52, pp. 169-210. Juselius, K. (1995): Do Purchasing Power Parity and Uncovered Interest Rate Parity Hold in the Long-run? - An Example of Likelihood Inference in a Multivariate Time-Series Model, mimeo, Institute of Economics, University of Copenhagen. Kirchgassner, G. / Wolters, J. (1993): Does the DM Dominate the Euro Market? An Empirical Investigation, Review of Economics and Statistics, vol. 75, pp. 773-778. Levich, R. M. (1985): Empirical Studies of Exchange Rates: Price Behavior, Rate Determination, and Market Efficiency, in: Jones, R.W. and P.B. Kenen (eds.): Handbook of International Economics, vol. II, North Holland: Amsterdam. Lutkepohl, H. (1991): Multivariate Time Series Analysis, Springer Verlag: Berlin/Heidelberg/New York. MacDonald, R. (1988): Floating Exchange Rates: Theories and Evidence, Unwin Hyman: London. - (1995): Long-run Exchange Rate Modeling: A Survey of the Recent Evidence, IMF Working Paper No. 95/14, International Monetary Fund: Washington, D.C. MacDonald, R. / Taylor, M. P. (1989): International Parity Conditions, Greek Economic Review, vol. 11, pp. 257-290. Marston, R. C. (1993): Three Parity Conditions in International Finance, in: Frisch, H. and A. Worgotter (eds.): Open-Economy Macroeconomics, Macmillan Press: Houndmills and London. McCallum, B. T. (1992): A Reconsideration of the Uncovered Interest Parity Relationship, NBER Working Paper no. 4113, National Bureau of Economic Research: Cambridge (MA.). McKinnon, J. G. (1991): Critical Values for Cointegration Tests, in: Engle, R.F. and C. W. J. Granger (eds.): Long-Run Economic Relationships, Oxford University Press: New York. ZWS 116 (1996) 4 39* OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 612 Christian Helmenstein and Gerhard Riinstler Meese, R. (1989): Empirical Assessment of Foreign Currency Risk Premiums, in: Stone, C.C. (ed.): Financial Risk: Theory, Evidence, and Implications, Kluwer Academic Publishers: Norwell (MA.), Dordrecht, and Lancaster, pp. 157-180. Ng, S. / Perron, P. (1995): Unit Root Tests in ARMA Models with Data-Dependent Methods for the Selection of the Truncation Lag, Journal of the American Statistical Association, vol. 90, pp. 268-281. Obstfeld, M. (1988): The Advantage of Tying One's Hands: EMS Discipline and Central Bank Credibility: Comments, European Economic Review, vol. 32, pp. 1077- 1082. Riinstler, G. / Jumah, A. / Karbuz, S. (1995): Arbitrage in Commodity Markets: A Full Systems Cointegration Analysis, Economics Working Paper No. 4, Institute for Advanced Studies, Vienna. Toda, H. J. / Phillips, P. C. B. (1991): Vector Autoregression and Causality: A Theoretical Overview and Simulation Study, Discussion Paper No. 1001, Cowles Foundation. - (1993): Vector Autoregression and Causality, Econometrica, vol. 61, pp. 1367-1393. Tootell, G. M. B. (1992): Purchasing Power Parity within the United States, New England Economic Review, July/August 1992, Federal Reserve Bank of Boston: Boston (MA.). Urbain, J.-P. (1993): Partial Versus Full System Modelling of Cointegrated Systems: An Empirical Illustration, Research Memorandum 93-044, Faculty of Economics, Limburg University, Maastricht. Wörgötter, A. (1992): Das österreichisch-deutsche Zinsdifferential und der österreichische Kapitalverkehrssaldo, Schriftenreihe Volkswirtschaft, no. 12/1992, Österreichische Postsparkasse, Wien. Zusammenfassung Die Theorie der ungedeckten Zinsparität sowie die Kaufkraftparitätentheorie implizieren das Bestehen einer langfristigen Gleichgewichtsbeziehung zwischen Inflations- und Zinsdifferentialen. In der vorliegenden Arbeit untersuchen wir die Trendeigenschaften dieser Differentiale für die EMS-Mitgliedsländer im Vergleich zu Deutschland. Den Ergebnissen von Einheitswurzeltests zufolge sind die niederländischen und österreichischen Differentiale jeweils stationär, während Kointegrationstests für Belgien und Frankreich die Existenz eines kointegrierenden Vektors anzeigen. Kausalitätstests zufolge verhalten sich Inflationsdifferentiale im allgemeinen Granger-kausal zu Zinssatzdifferentialen. Bei einzelnen Ländern läßt sich für die Zeit vor 1985 zusätzlich eine Granger-Kausalität von Zinsdifferentialen für Inflationsdifferentiale feststellen. Diese Resultate deuten darauf hin, daß der Glaubwürdigkeit der Geld- und Fiskalpolitik eine wesentliche Bedeutung für die Reaktion der Anleihemärkte auf die Preisdynamik zukommt. ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40 Inflation and Interest Rate Differentials 613 Abstract Uncovered interest parity and purchasing power parity suggest an equilibrium relationship between inflation and long-term interest rate differentials. In this paper we investigate the time trend properties of these differentials for the EMS members vis-à-vis Germany. The results of unit root tests indicate that for the Netherlands and for Austria both differentials are stationary while for Belgium and France we find a cointegration vector that represents the hypothesized relationship. Causality tests provide empirical evidence that inflation differentials generally Granger-cause interest rate differentials. For specific countries we also find Granger causality from interest rate to inflation differentials for the period before 1985. These results support the view that the credibility of monetary and fiscal policies plays an important role in determining the reaction of bond markets to inflation. JEL Klassifikation: F15, C32 Keywords: Economic integration, cointegration, uncovered interest parity, European Monetary System, long-run causality ZWS 116 (1996) 4 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.116.4.593 | Generated on 2023-04-04 12:22:40