Monetary integration in the Southern Cone

North American Journal of Economics and Finance 13 (2002) 323–349

Monetary integration in the Southern Cone Ansgar Belke a,∗ , Daniel Gros b a

Department of Economics, University of Hohenheim, Museumsflügel, D-70599 Stuttgart, Germany b Centre for European Policy Studies (CEPS), Brussels, Belgium

Received 14 April 2002; received in revised form 9 September 2002; accepted 10 September 2002

Abstract The effect of exchange-rate volatility on the domestic economy depends in part on the importance of trade in total economic activity. Unlike the European Union (EU), trade among the Mercosur countries is less highly integrated, so that movements in intra-area exchange rates are less important than exchange rates vis-à-vis the dollar and the euro. This paper analyzes the impact of exchange-rate and interest-rate volatility on investment and labor markets in the Southern Cone and finds that both volatility against the dollar and the euro and variability of interest rates have significant dampening effects on employment and investment. © 2002 Elsevier Science Inc. All rights reserved. JEL classification: E42; F36; F42 Keywords: Currency union; Exchange-rate and interest-rate variability; Job creation; Mercosur; Option value effects

1. Introduction Flexible exchange-rate regimes are often criticized for excessive volatility and its destabilizing effect on the economy. This paper explores the effects of exchange-rate and interest-rate volatility on investment and employment. Our previous research has shown that exchange-rate variability among European countries can have a significant impact on the economy generally and on labor markets in particular (Belke & Gros, 2001a). It is important, of course, to distinguish between exogenous and endogenous exchange-rate movements. If high exchange-rate volatility is the result of domestic monetary policy shocks, then it would be inappropriate to blame the exchange rate for observed variations in investment ∗ Corresponding author. Tel.: +49-711-459-3246; fax: +49-711-459-3815. E-mail address: [email protected] (A. Belke).

1062-9408/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 1 0 6 2 - 9 4 0 8 ( 0 2 ) 0 0 1 0 0 - 6

324

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

and employment. In such cases, variability of the exchange rate and of the other two variables is jointly dependent on monetary instability. Analogous considerations hold for interest-rate volatility. Where the sources of exchange-rate volatility are exogenous, on the other hand, the impact on the domestic economy is likely to depend on the importance of trade in overall economic activity. Here, an important difference between Mercosur and the EU is that regional trade is far more important in Europe than in the Southern Cone. We document the degree of trade integration in Section 2. In Section 3, we develop a model which enables us to consider certain analytical aspects of the relationship between exchange-rate and interest-rate volatility, on the one hand, and investment and labor markets, on the other. The empirical results are presented in Section 4, with robustness tests discussed in Section 5. Section 6 considers the implications of the results for the debate on the design of intra-Mercosur monetary relations and concludes. 2. Trade integration: the EU and the Southern Cone The focus of this section is on Argentina and Brazil, the two large members of Mercosur.1 Together, they represent 95% of the total population of the Mercosur and produce 97% of its GDP. The “peripheral” countries, Paraguay and Uruguay, are closely tied to Argentina and Brazil via trade, have very small internal markets and possess limited access to international capital markets. The importance of trade for Argentina, Brazil and Chile, relative to Spain (whose figures are not far from the EU average), is displayed in Table 1. It is clear at once that trade with countries outside the region dominates intra-regional trade. Indeed, the relatively small size of Brazil’s trade with Argentina is noteworthy. It is difficult to call this grouping of countries a trade “bloc.” In assessing the role of financial volatility on trade, therefore, it is not mainly intra-regional trade that is of primary interest here, but extra-regional trade. Contrary to a widely held view, the EU is slightly more important to Latin America as a trading partner than the U.S. and NAFTA, especially for Argentina. Any evidence suggesting that exchange-rate volatility affects trade may then help make the case for fixed rates with the U.S. dollar, the euro or some combination of the two. That is, extra-regional fixes would offer more protection than intra-regional fixes. Table 1 Trade integration within the Southern Cone Exports as % of GDP

Argentina Brazil Chile Spain

Intra-regional/extra-regional

Total

Intra-regional

8.9 7.6 26.5 26.6

2.7 0.9 2.8 16.4

Sources: Center for Global Trade Analysis (2001), own calculations.

0.44 0.13 0.11 1.61

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

325

3. Model The key question concerns the effect of financial volatility on investment and employment in emerging markets. The answer will be influenced by two considerations, accepting the fact that the link between exchange-rate variability and the volume of trade is known to be weak (Cˆoté, 1994; Lafrance & Tessier, 2000). The first involves the pattern of trade invoicing, which is different in emerging markets. Primary commodities, for example, are mainly dollar-invoiced. Since Mercosur exports have high primary-commodity content,2 exchange-rate volatility should have a significant impact on trade. This is especially valid for Argentina with its primary product share of 48.2% of total domestic value-added induced by exports.3 Second, capital markets in emerging economies are relatively weak and incomplete. Futures markets tend to be illiquid or nonexistent, so that the tools for hedging exchange- rate risk are simply not available. Emerging market authorities are more concerned about large exchange rate fluctuations, because the pass-through from exchange rate to inflation is higher than in industrialized countries (Calvo & Reinhart, 2000). The relevant theoretical literature (Baldwin & Krugman, 1989; Krugman, 1989 to start with) focuses on the idea that in order to export firms need to sustain sunk costs. This applies to all types of production, and perhaps especially to primary products, which require large capital investments. In view of the relatively low trade linkages among Mercosur countries and the importance of primary commodities priced in dollars, it is likely that intra-Mercosur exchange-rate variability is of less concern than variability with respect to the dollar, yen and euro. As the devaluation of the real against the dollar in the late 1990s made clear, however, what affects the Argentine economy is not merely the revaluation of the peso against the real, but the relative appreciation of the peso against the G-3 currencies brought about by the devaluation of the Brazilian currency. Our interest is in the relationship between exchange-rate uncertainty, caused by high volatility, and job creation. In the following we develop a complete model of this relationship, which goes beyond the focus in Reinhart and Reinhart (2001) on the spending channel. This model is based on the idea that uncertainty of future earnings raises the “option value of waiting” with respect to decisions governing investment and the hiring of workers.4 Applied to the labor market, this model works as follows. When firms create a job, they incur hiring and capital costs. Once a worker is hired, wage payments may also become sunk costs if there are restrictions on firing in the employment contract. If future exchange rates are uncertain and firms have reason to fear appreciation of the domestic currency and the losses that come with it, they may delay job creation. The option value of waiting is likely to increase with the bargaining strength of workers and their unions. This, in turn, raises the impact of uncertainty on employment. The literature provides other mechanisms through which uncertainty may have an adverse impact on employment. First, in unionized labor markets in which contract wages are set in advance, uncertainty in labor demand (coming from uncertainty in productivity or in the exchange rate) may cause rational unions to set a higher wage than would otherwise be the case. Uncertainty results in a “risk premium” in the wage, and thus in higher unemployment (Andersen & Sorensen, 1988; Sorensen, 1992). Another channel by which uncertainty might

326

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

affect employment is via its impact on investment. Our theoretical arguments are equally valid for firms who decide to postpone an investment project because of excessive uncertainty (see Belke & Gros, 2001a).5 Unemployment can be expected to rise if investment falls, because investment is an important component of demand. Moreover, technological complementarities between labor and capital imply that a capital slowdown entails a fall in employment (see, e.g., Rowthorn, 1999). Let us now spell out more fully and clearly the relevant responses of firms due to a rise in uncertainty. Consider a set-up in which there are three periods and a single firm active in an export-oriented industry decides about job creation. During the first two periods (called zero and one) the firm can open a job, hire a worker and produce output that is sold in a foreign market during the following periods. If the job is created during period zero, the worker is hired for two periods (zero and one) to produce output to be sold in periods one and two. If the job is created in period one, the worker is hired only for period one and output is sold in period two. To create a job, the firm pays a start-up cost, c, which reflects the cost of hiring, and training and the provision of job-specific capital. After a job is created, a worker is hired and is paid a wage, w, above the worker’s fallback (or reservation) wage, w, during every period the worker is employed. The fallback wage measures (besides disutility of work) all opportunity income that the worker has to give up by accepting the job. In particular, it includes unemployment benefits, but it might also be positively related to a collective wage set by a trade union or to a minimum wage, both of which should raise the worker’s fallback position. In general, we would argue that the fallback wage should be higher in countries that are characterized by generous unemployment benefit systems, by strong trade unions or by minimum wage legislation. In every period in which the worker is employed, he produces output to be sold in the following period in a foreign market at domestic price, p, which has a certain component, p∗ (the foreign price), plus a stochastic component, e (the exchange rate). We assume that the foreign price is fixed (“pricing-to-market”), and that the exchange rate follows a random walk.6 The latter assumption is crucial, but seems justifiable in light of the monetary policy regimes of Mercosur countries during the time period investigated here. In period one, the exchange rate, e1 , is uniformly distributed between −σ 1 and +σ 1 . The exchange rate in period two, e2 , is uniformly distributed between e1 − σ2 and e1 + σ2 . An increase in σ i means an increase in uncertainty, or an increase in the mean-preserving spread in period i = 1, 2 (σ i is proportional to the standard deviation of ei ). Uncertainty can be temporary (e.g., if σ1 > 0 and σ2 = 0) or persistent (if also σ2 > 0). As will become apparent soon, however, the variability of the exchange rate during the second period has no influence on the result. The wage rate, w, for the job is determined by the (generalized) Nash bargaining solution that maximizes a weighted product of the worker’s and the firm’s expected net return from the job. We assume that both the firm and the worker are risk-neutral. This assumption implies that risk-sharing issues are of no importance for our analysis. Thus, we may assume realistically (but without loss of generality) that the worker and the firm bargain about a fixed wage rate, w (which is independent of realizations of the exchange rate), when the worker is hired, so that the firm bears all the exchange-rate risk. A wage contract which shifts some exchange-rate risk to the worker would leave the (unconditional) expected net returns unaffected, and has therefore no effect on the job creation decision. Of course, if the firm were risk-averse, the

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

327

assumption that it bears all exchange-rate risk would make a postponement of job creation in the presence of uncertainty even more likely. Consider first the wage-bargaining problem for a job created in period zero, in which case the worker is hired for two periods. After the job is created (and the job creation cost is sunk), the (unconditional) expected net return of this job is equal to E0 (S0 ) = 2p∗ − 2w = 2π , where π = p∗ − w denotes the expected return of a filled job per period (we abstract from discounting). Denoting the bargaining power of the worker by 0 < β < 1, the firm’s net return from the job created in period zero is7 E0 (Π0 ) = (1 − β)E0 (S0 ) − c = 2(1 − β)π − c.

(1)

In order to make the problem non-trivial, the expected return from job creation in period zero must be positive, i.e., we assume that 2(1−β)π –c > 0. Implicit in our model is the assumption that the firm and the worker sign a binding employment contract for two periods (zero and one). Hence, they cannot sign a contract that allows for the possibility of job termination in the first period whenever the exchange rate turns out to be unfavorable. In period one (after realization of the exchange rate), the conditional expected surplus from job continuation is E1 (S1 ) = π +e1 , which may be negative if the exchange rate falls in period one below −π < 0. In such circumstances, both the worker and the firm would benefit from termination. If a contract allowing for termination in period one could be signed, the unconditional expected surplus in period zero would be larger (consequently both the worker and the firm would prefer to sign such a contract).8 However, having in mind the interpretation of a rather short period length (a month, to be compatible with our empirical analysis), the assumption of a binding contract for two periods seems to be more appropriate. Of course, once a binding contract for two periods is signed, the worker always prefers continuation (since the contract wage exceeds the fallback wage), and the firm would incur losses if the exchange rate turns out to be unfavorable. In an earlier version of this paper (Belke & Gros, 2002b) we consider an alternative set-up which allows for the possibility of job destruction. It turns out that in this case uncertainty does not delay job creation, but job destruction becomes more likely if uncertainty increases. However, in this scenario, the firm keeps the option to terminate the work relationship whenever it becomes unprofitable. Hence, the negative relationship between exchange-rate variability and employment is robust to this variation. If the firm waits until period one, it keeps the option of whether or not to open a job. It will create a job only if the exchange rate realized during period one (and so expected for period two) is above a certain threshold level, or barrier, denoted by b. Given that an employment relationship in period one yields a return only during period two, the barrier to make the creation of the job just worthwhile is given by the condition that the (conditional) expected net return to the firm is zero: c c (1 − β)(p ∗ + β − w) − c = 0 or b = + w − p∗ = − π. (2) 1−β 1−β Whenever e1 ≥ b, the firm creates a job in period one, and the conditional expected net return to the firm is E1 (Π1 ) = (1 − β)(π + e1 ) − c ≥ 0. Whenever e1 < b, the firm does not create a job in period one, and its return is zero. Hence, whenever both events occur with

328

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

positive probabilities (i.e., whenever σ1 > b > −σ1 ),9 the unconditional expected return of waiting in period zero is given by   


σ1 − b σ1 + b σ1 + b −c , (3) 0+ (1 − β) π + E0 (Π1 ) = 2σ1 2σ1 2 where the first element is the probability that it will not be worthwhile to open a job (in this case the return is zero). The second term represents the product of the probability that it will be worthwhile to open the job (because the exchange rate is above the barrier) and the average expected value of the net return to the firm under this outcome. Given condition (2), this can be rewritten as E0 (Π1 ) =

(1 − β)(σ1 − b)2 . 4σ1

(4)

This is the key result, since it implies that an increase in uncertainty increases the value of the waiting strategy, since Eq. (4) is an increasing function of σ 1 .10 As σ 1 increases, it becomes more likely that it is worthwhile to wait until more information is available on the expected return during period two. At that point, the firm can avoid the losses that arise if the exchange rate is unfavorable by not opening a job. This option not to open the job becomes more valuable with more uncertainty. The intuitive explanation is that waiting implies that the firm foregoes the expected return during period one, but it keeps the option not to open the job, which is valuable if the exchange rate turns out to be unfavorable. The higher the variance, the higher the potential losses the firm can avoid and the higher the potential for a very favorable realization of the exchange rate, with consequent very high profits. It is now clear from (1) and (4), that a firm prefers to wait if and only if (1 − β)(σ1 − b)2 > 2(1 − β)π − c. 4σ1

(5)

As the left-hand side is increasing in σ 1 , the firm delays job creation if exchange-rate uncertainty is large enough. The critical value at which (5) is satisfied with equality can be solved as 11 
 c c ∗ σ1 = 3π − + 2 π 2π − . (6) 1−β 1−β Whenever σ1 > σ1∗ , firms decide to postpone job creation in period zero. Since σ1∗ is increasing in π (and thereby decreasing in the fallback wage, w), decreasing in the cost of job creation, c, and decreasing in the worker’s bargaining power, β, we conclude that a strong position of workers in the wage bargain (reflected in a high fallback wage or in the bargaining power parameter) and higher costs of hiring raise the option value of waiting and make a postponement of job creation more likely. Thus, the adverse impact of exchange-rate uncertainty on job creation and employment should be stronger if the labor market is characterized by generous unemployment benefit systems, powerful trade unions, minimum wage restrictions or large hiring costs. That such features of the labor market are detrimental to employment is, of course, not surprising. The adverse impact of these features on employment has been

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

329

confirmed empirically in various studies, and there are many other theoretical mechanisms explaining it (see, e.g., Layard, Nickell, & Jackman, 1991; Nickell, 1997). What our simple model shows, is that these features also reinforce the negative employment effects of exchange-rate uncertainty. Another important implication of the model is that only the current, short term uncertainty σ 1 has an impact on the decision to wait. Future uncertainty, represented here by σ 2 , does not enter in the decision under risk neutrality. If one takes a fixed period, e.g., 1 month, the likelihood that job creation will be postponed to the end of that period depends only on the uncertainty during that period and not on future uncertainty. This implies that even short spikes in uncertainty as, e.g., measured by a contemporaneous uncertainty proxy in empirical investigations of the real option effect, can have a strong impact on employment. Our crude model has abstracted from risk aversion. However, we would argue that the basic conclusion, that even a temporary increase in uncertainty can make a postponement of job creation optimal, does not change and is robust because a prolonged period of high uncertainty means that expected returns beyond the next period would be discounted more heavily. Moreover, the additional impact of risk aversion on job creation should be stronger under the realistic assumption that firms bear all the exchange-rate risk. In sum, we obtain two conclusions from the model. First, even a temporary “spike” in exchange-rate variability can induce firms to wait with their creation of jobs (of course and for exactly this reason, the level of the exchange rate at the same time loses explanatory power). Second, the relationship between exchange-rate variability and (un-) employment should be particularly strong if the labor market is characterized by rigidities that improve the bargaining position of workers. A stronger fallback position of workers raises the contract wage, lowers the net returns to firms and induces firms to delay job creation in the face of uncertainty. Our argument rests on the assumption that workers cannot be fired immediately if the exchange rate turns out to be unfavorable. Hence, sunk wage payments are associated with the decision to hire a worker. These sunk costs and, consequently, the impact of uncertainty on job creation become more important if there are high firing costs. However, as we argued above, even if there are no firing costs and if workers can be laid off at any point in time, exchange-rate uncertainty should have a direct impact on job destruction. A more elaborate labor market model of job creation and job destruction (e.g., following the model of Pissarides, 2000, Chapter 3), might further clarify these issues, but we would expect that uncertainty has a negative effect on both job creation and destruction. In the empirical analysis, we therefore prefer to employ aggregate labor market indicators, rather than more disaggregate job flow data.12 Interest-rate volatility should have effects similar to exchange-rate volatility in the context of our model. A weaker domestic exchange rate increases the profits of an exporter (or the profits on domestic sales for producers competing with imports). Lower interest rates have the same effect, for all types of producers (as all production involves some investment). Uncertainty about future interest rates will be particularly important for longer-term investments in the Mercosur countries in which long-term financing was simply not available for decades, thus forcing producers to rely on rolling over short-term credits over long time periods. Since generous unemployment compensation schemes, union power, and firing restrictions generally improve workers’ bargaining positions, we would expect the link between exchange-rate uncertainty and employment to be relatively high in the Mercosur countries.

330

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

While Latin American governments have implemented economic reforms in almost every sector, in most countries labor markets remain highly regulated. As of the late 1990s, only a handful of Latin American nations had reformed their labor markets in any significant way (Edwards & Cox Edwards, 2000; Eichengreen, 1998; Hopenhayn, 2001). Existing laws tend to protect employment, while taxing employers heavily. It is widely agreed that the social protection provided through labor-market regulation limits the market’s ability to adjust wages and unemployment. Many of the rules governing labor markets in Latin America raise labor costs, create barriers to entry and exit, and, hence, introduce rigidities into the employment structure. As in continental Europe, these rigidities include exceedingly restrictive regulations on hiring and firing practices, as well as burdensome social insurance schemes. Mandated severance payments and other regulations penalizing employment termination are stricter than in the majority of OECD countries. All in all, it appears as if rigidities are tighter in Argentina than in Brazil (Heckman & Pagés, 2000; Márquez & Pagés, 1998).13 4. Empirical analysis For each year of our sample from 1970 to 2001, we calculate the standard deviation of 12 monthly observations of the first differences of exchange rates and interest rates.14 Exchange rates include the euro exchange rates of the Argentinean peso and of the Brazilian real.15 We include nominal and real euro-dollar exchange-rate volatility in order to test for possible impacts of G-3 exchange-rate volatilities on the Mercosur countries, as conjectured by Reinhart and Reinhart (2001).16 For each of our variables, we conducted unit root tests. While the standard ADF-tests do not reject the stationarity of the first differences of all variables, the stationarity of exchange-rate volatility (which is itself composed of changes) was at times borderline dependent on sample length. The idea of stationarity of a volatility measure is controversial from a theoretical point of view, given the exchange-rate crises that have plagued the Mercosur countries.17 In cases of doubt, we use differences since the disadvantages of differencing when it is not needed appear to be less severe than those of failing to do so when it is appropriate. In the first instance, the worst outcome is that the disturbances are moving average, but the estimators would still be consistent, whereas in the second instance, the usual properties of the OLS test statistics would be invalidated. We begin with some simple specifications, in which the change in unemployment and the growth of employment are related to their own past and to lagged values of our volatility measures. The results are summarized in Tables 2a and 2b. These tables summarize regression results from bivariate VARs on annual data (1970–2001 or shorter depending on data availability). The hypothesis tested is that exchange-rate and interest-rate variability have no effect on the real variables investigated here.18 The results are implicitly based on comparison of two regression equations, as exemplified in Eqs. (7) and (8) with respect to the effect of exchange-rate variability on unemployment19 N  DUEt = α0 + αi · DUEt−i + ut i=1

(7)

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

331

Table 2a Regression results for Argentina (1970–1990)

VOLNER ARPUSD VOLRER ARPUSD VOLNER ARPEUR VOLRER ARPEUR VOLNER USDEUR VOLRER USDEUR VOLREER ARG VOLNEERINTRAMERC ARG VOLREERINTRAMERC ARG VOLINTEREST ARG VOLREALINTEREST ARG

DUNEMPRATE ARG

DEMPRATE ARG

GROWTH REALINVEST ARG

0.06∗∗∗ (0) 0.07∗∗∗ (0) 0.04∗∗ (0) 0.05∗ (0) 1.38∗∗∗ (0) 1.41∗∗∗ (0) 0.05∗ (0) 0.06∗∗∗ (0) 0.07∗∗∗ (0) 0.01∗∗∗ (0) 0.01∗∗∗ (0)

−0.02∗∗ (−1) −0.03∗∗∗ (−1) −0.02∗∗ (−1) −0.03∗∗ (−1) −0.52∗∗∗ (−1) −0.53∗∗∗ (−1) −0.03∗∗ (−1) −0.02∗∗ (−1) −0.03∗∗∗ (−1) −0.003∗ (−1) −0.003∗ (−1)

−0.44∗ (0) −0.51∗ (0) −0.65∗∗ (0) −0.78∗∗ (0) −11.33∗∗ (0) −10.57∗ (0) −0.80∗∗ (0) −0.44∗ (0) −0.48∗ (0) −0.11∗∗∗ (0) −0.10∗∗∗ (0)

Note: Point estimates for the impact of exchange-rate volatility are displayed together with their significance levels (***: 1%; **: 5%; *: 10%). Numbers in parentheses refer to the lags of the volatility variable.

Table 2b Regression results for Brazil (1970–1993)

VOLNER BRRUSD VOLRER BRRUSD VOLNER BRREUR VOLRER BRREUR VOLNER USDEUR VOLRER USDEUR VOLREER BRA VOLNEERINTRAMERC BRA VOLREERINTRAMERC BRA VOLINTEREST BRA VOLREALINTEREST BRA

DUNEMPRATE BRA

GROWTHEMP BRA

GROWTH REALINVEST BRA

0.11∗ (−1) 0.28∗∗∗ (0) 0.12∗∗ (−1) 0.26∗ (0) – – 0.28∗ (0) 0.39∗ (−2) 0.04∗ (−1) 0.05∗∗ (−1) – –

−0.50∗∗∗ (−1) −0.92∗∗∗ (−1) −0.65∗∗∗ (−2) −0.82∗ (−1) −1.78∗∗ (−2) −1.93∗∗ (−2) −1.37∗∗∗ (−1)

−2.03∗∗∗ (−1) −4.46∗∗∗ (0) −2.19∗∗ (−1) −5.59∗∗∗ (−0) – – −7.13∗∗∗ (0) −4.50∗ (−2) −0.72∗∗∗ (−1) −0.87∗∗∗ (−1) −0.16∗∗ (−1) −0.13∗∗ (−1)

−0.13∗∗∗ (−2) −0.12∗ (−2) −0.03∗∗ (−1) −0.03∗∗ (−1)

Note: Point estimates for the impact of exchange-rate volatility are displayed together with their significance levels (***: 1%; **: 5%; *: 10%). Numbers in parentheses refer to the lags of the volatility variable; (–): means “not significant.”

and DUEt = α0 +

N N   αi · DUEt−i + βi · EXVt−i + ut , i=1

(8)

i=0

where DUEt stands for the change in the unemployment rate between period t and t −1, EXVt−i specifies the level of exchange-rate variability between period t − i and period t − i − 1, ut represents the usual i.i.d. error term and N is the maximum number of lags (2 in this case).

332

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

Exchange-rate variability is said to “cause” a change in unemployment if at least one β, i.e., one of the coefficients on past and contemporaneous exchange-rate variability is significantly different from zero. In other words, these tests measure the impact of exchange-rate variability on changes in national unemployment rates once the autonomous movements in unemployment have been taken into account by including lagged unemployment rates among the explanatory variables. Thus, a significant effect implies that one can reject the hypothesis that the change in exchange-rate variability does not affect unemployment. Use of the standard t-distribution for the purpose of model selection necessitates the use of first differences in the unemployment rate, given that its level is clearly non-stationary. The specification of the underlying equations is based on the usual diagnostics combined with the Schwarz Bayesian Information Criterion (SCH). The latter is chosen as our primary model selection criterion, since it asymptotically leads to the correct model choice (if the true model is among the models under investigation, Lütkepohl, 1991). The regression which reveals the lowest SCH value and at the same time fulfills the usual diagnostic residual criteria is chosen.20 Tables 2a and 2b show the results for Argentina and Brazil, using the 11 volatility measures and the three economic activity variables. The tables present the results for limited sample estimates.21 The results for full sample estimates for Argentina and Brazil are provided in Appendix A. While the low overall trade flows of the Mercosur might suggest to some that the impact of exchange-rate volatility may not significantly affect the region, it may nonetheless be very relevant for the two small countries of the region. This heterogeneity is de-emphasized in this paper, which mostly analyzes Argentina and Brazil. However, we report the striking results for the “small” country, Uruguay, in Appendix A. For each of the activity variables, we first used as explanatory variables only their own past values and lags of the exchange-rate and interest-rate variability measures. The results reported in the first row of Table 2a, for example, suggest that exchange-rate variability, as measured by the standard deviation of the nominal exchange rate of the peso against the U.S. dollar, has a significant impact on labor markets and investment in Argentina. Only the coefficient estimate, its significance level and the lag order of exchange-rate variability are displayed in the tables. The numbers in parentheses correspond to the lag order of exchange-rate variability. According to our priors, the expected sign of exchange-rate and interest-rate variability is positive for the unemployment rate and negative for employment and investment. As stated above, the full sample covers the period 1970–2001. However, in the case of Argentina, the model is estimated for the years 1970–1990 in order to exclude the currency-board period. Inclusion of the latter would have introduced structural breaks in the relationships, because the correlation between exchange-rate volatility as a variable that does not move and a real sector variable is nil per se. In a strict sense, however, this is true only for the nominal peso-dollar rate, but not (directly) for the others and not for the real peso-dollar rate if there are differences in inflation rates. Hence, the full sample results for the latter variables are listed in Appendix A. We add country-specific dummies from time to time in order to account for possible breaks in the VAR relations. For instance, in the regressions for Argentina, a dummy for the year 1981, the beginning of a floating exchange-rate regime, often proves to be significant. The dummies

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

333

for the years 1974 and 1975 approximate the expansionary fiscal and monetary policies with which the government of Isabel Peron tried to rekindle economic growth (D´ıaz-Bonilla & Schamis, 2001, p. 76). However, as the labor market figures reveal, these measures were only successful in the short run. In the medium run, they only fueled fiscal deficits and inflation and the economy entered a recession shortly afterwards in 1975. Correspondingly, we find the sign of these dummies for 1974 and 1975 to be negative in unemployment equations, as expected. However, only 4 out of 11 cases for the change in the unemployment rate displayed in Table 2a, namely, those for the peso-dollar rate and those for the intra-Mercosur exchange-rate volatilities, are based on regressions incorporating these dummies. These dummies play an even smaller role in explaining the results for the change in employment (0 out of 11 cases) and for the growth of real investment (2 out of 11 cases, namely, those for intra-Mercosur volatilities). We also conducted some robustness tests to test whether the dummy for the exchange-rate crisis in 1989 completely determines the results.22 4.1. Interpreting the results For Argentina (Table 2a), the point estimate obtained from the first specification implies that an increase of one percentage point in the variability (standard deviation) of the nominal bilateral exchange rate of the peso vis-à-vis the U.S. dollar is associated with an increase in the unemployment rate of 0.06 percentage points. This is a very small response, but a limited contemporaneous reaction is not completely surprising. Apart from the impact coefficients, the longer-run effects also depend on the coefficients of the lagged endogenous variable. Since the coefficients of the unemployment rate, lagged one, two, or more periods, turn out to be positive, the long-term impact on the natural rate of unemployment is larger (see Belke & Gros, 2001a, 2002a, for examples). For Argentina, the results cover the period from 1970 to 1990, i.e., the period prior to inauguration of the currency board. It is apparent that one could no longer expect nominal dollar variability of the peso to have any influence on macroeconomic variables after the installation of the currency board.23 We see that the volatility variables have statistically significant effects on labor markets and investment. As predicted, the effect of volatility is negative for employment and investment and positive for unemployment. Further, both types of volatility have contemporaneous effects on unemployment and investment, but affect employment with a lag of one period. Firms can react immediately by investing less in machinery and in the workforce (no new hiring). However, they might take some time to see how things work out before they actually start hiring (provided the labor market does not allow them quick firing as well). Our theory predicts that increased uncertainty delays investment and hiring. Hence, unemployment and investment should react immediately, where the rise in unemployment cannot be caused, according to our theory, by firing (if firms do not maintain the option to terminate the work relationship whenever it becomes unprofitable), but only by non-hiring. So, as long as the labor force is growing, unemployment rises. Then, employment declines with a lag, because immediate firing is ruled out by our theory. Real and nominal measures of volatility typically have similar point estimates and significance levels. This is not surprising in view of the well-known fact that in the very short run

334

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

(monthly data for the volatility measures) changes in nominal and real exchange rates are highly correlated.24 It is also not surprising that dollar/euro exchange-rate variability has a larger point estimate than that of the volatility of the home currency against the dollar, because the former is much less variable than the latter. Alternatively, the larger coefficient might simply be explained with reference to the importance of the European market for Argentina. We observe similar patterns for Brazil, but the results are much stronger when we limit the sample to the period before the real plan, i.e., up to 1993.25 For this sample period, we find again that the volatility coefficients are significant and have the expected sign. The typical lag period appears to be 1 or 2 years, which supports the argument that the volatility variables are exogenous relative to the activity variables. It is striking that the lag structure is exactly the same for the unemployment rate and real capital formation. As noted above, in constrained labor markets unemployment and investment can react more quickly because it is easier to stop new hiring and new investment projects. The main difference with respect to Argentina is that for Brazil the dollar/euro exchange rate does not seem to be as important and interest-rate volatility is not always significant. The former might be due to the slight difference in the geographical distribution of exports (Belke & Gros, 2002b). This result is consistent with Reinhart and Reinhart (2001), who argue that only the volatility of the own currency should matter. The weakness of the interest-rate effect may be due to the widespread use of clauses indexing mortgage payments to wage inflation in Brazil prior to the real plan period (Amadeo, Gill, & Neri, 2000). The point estimates are generally higher for Brazil than for Argentina. This may be caused by the fact that volatilities are generally higher for Argentina,26 so that firms have found other ways of coping in this environment. Whereas this explanation may sound pretty speculative to some without further evidence, the argument raised in chapter 3, that rigidities are tighter in Argentina so that reactions are more constrained seems to be more persuasive in this respect. When it comes to volatility of the exchange rate between the national currency and the dollar or the euro, the results are relatively similar with respect to the size of the coefficients. This was to be expected, as the average volatility of the dollar/euro rate VOLRER USDEUR is at 2.37% (sample 1978–1990) so much lower than, for example, the average volatility of the Argentinian currency in real terms against either of these two major currencies (9.63% against the euro, for example). As Belke and Gros (2002b) show empirically, interest rate and exchange-rate volatility are highly correlated (with values around 0.85–0.9% for most measures). Hence, it is not surprising that the coefficients of both volatility variables reveal similar signs and significance levels. What drives interest-rate volatility? In an OECD country with a flexible exchange rate, one would consider short-term domestic interest rates to constitute a measure of monetary policy. In emerging market economies this might not be the case, whatever the exchange-rate regime. Especially for highly indebted countries like Argentina and Brazil, developments in international financial markets might be much more important. Both exchange and interest rates can shoot up if foreign financing is no longer available (contagion after the Asian and Russian crisis) or the perception in international financial markets of the country’s political and economic future changes (witness the 30% depreciation of the real when a leftist candidate had a lead in the opinion polls for the presidential elections).

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

335

It can by now be considered a stylized fact that exchange rates are “disconnected” from fundamentals (e.g., Obstfeld & Rogoff, 2000). Section 5 below finds additional support using the second moment. But the constant threat of speculative attack hanging over emerging market economies can actually cause a co-movement, which does not exist for developed economies. Belke and Gros (2002b) report that the correlation between the volatilities of the bilateral dollar/euro exchange rate and the respective interest-rate differential is essentially zero (about 0.1). However, we cannot rule out that variability in the exchange rate and the interest rate are jointly caused by variability in monetary policy.27 If this were the case, the cost of exchange-rate volatility reported here should be considered the cost of erratic monetary policy. We are confident that for Argentina and Brazil, the general “disconnect” between exchange rates and fundamentals holds in the short run and extends to (domestic) interest rates, which for emerging markets are determined by shocks coming from international financial markets.

5. Robustness tests: exogeneity of volatility? Reverse causation and missing third variables are possible objections to the simple test results presented so far. Whenever exchange-rate variability influences real variables with a lag, reverse causation appears less plausible. But even in cases of a contemporaneous relationship, reverse causation appears not to be a problem as suggested by additional pairwise Granger causality tests which are applied to exchange-rate and interest-rate variability and the real sector variables. In Tables 3a and 3b we report the results from numerous pairwise Granger causality tests. For the data for Argentina and Brazil, we do not reject the hypothesis that the real-sector variables do not Granger-cause our volatility measures in 63 out of 66 cases. However, based on our estimates displayed in Tables 2a and 2b, we do in the overwhelming majority of cases reject the hypothesis that our volatility measures do not “cause” the three real-sector variables. Therefore, it appears that “causality” runs from volatility to the real sector. Without further extensive testing, we are not able to show that interest- and exchange-rate volatilities are also not caused by erratic monetary policy. The exchange-rate disconnect referred to above provides an additional general argument in favor of our exogeneity hypothesis for the volatility variables. We are skeptical in general about the possibility that exchange-rate and interest-rate variability at our high frequency were caused by slow-moving variables such as labor market rigidities or unemployment and investment. A further argument validating our methodology and our results comes from the work of Canzoneri, Vallés, and Viñals (1996) and others, who show for a different sample of countries that exchange rates reacted mainly to financial shocks like changes in money supply rather than real fundamentals. Rose (1996) and Flood and Rose (1995) also emphasize that exchange-rate volatility is largely noise.

6. Concluding remarks The results suggest that exchange-rate variability (whether extra- or intra-Mercosur) and interest-rate variability have had statistically significant negative impacts on employment and

336

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

Table 3a Pairwise Granger causality tests for exogeneity, Argentina (sample: 1970–1990) Lags: 2; null hypothesis

Observed

F-statistic

Probability

DUNEMPRATE ARG does not Granger Cause VOLNER ARPUSD DEMPRATE ARG does not Granger Cause VOLNER ARPUSD GROWTHREALINVEST ARG does not Granger Cause VOLNER ARPUSD DUNEMPRATE ARG does not Granger Cause VOLRER ARPUSD DEMPRATE ARG does not Granger Cause VOLRER ARPUSD GROWTHREALINVEST ARG does not Granger Cause VOLRER ARPUSD DUNEMPRATE ARG does not Granger Cause VOLNER ARPEUR DEMPRATE ARG does not Granger Cause VOLNER ARPEUR GROWTHREALINVEST ARG does not Granger Cause VOLNER ARPEUR DUNEMPRATE ARG does not Granger Cause VOLRER ARPEUR DEMPRATE ARG does not Granger Cause VOLRER ARPEUR GROWTHREALINVEST ARG does not Granger Cause VOLRER ARPEUR DUNEMPRATE ARG does not Granger Cause VOLNER USDEUR DEMPRATE ARG does not Granger Cause VOLNER USDEUR GROWTHREALINVEST ARG does not Granger Cause VOLNER USDEUR DUNEMPRATE ARG does not Granger Cause VOLRER USDEUR DEMPRATE ARG does not Granger Cause VOLRER USDEUR GROWTHREALINVEST ARG does not Granger Cause VOLRER USDEUR DUNEMPRATE ARG does not Granger Cause VOLREER ARG DEMPRATE ARG does not Granger Cause VOLREER ARG GROWTHREALINVEST ARG does not Granger Cause VOLREER ARG DUNEMPRATE ARG does not Granger Cause VOLNEERINTRAMERC ARG DEMPRATE ARG does not Granger Cause VOLNEERINTRAMERC ARG GROWTHREALINVEST ARG does not Granger Cause VOLNEERINTRAMERC ARG DUNEMPRATE ARG does not Granger Cause VOLREERINTRAMERC ARG DEMPRATE ARG does not Granger Cause VOLREERINTRAMERC ARG GROWTHREALINVEST ARG does not Granger Cause VOLREERINTRAMERC ARG DUNEMPRATE ARG does not Granger Cause VOLINTEREST EUR DEMPRATE ARG does not Granger Cause VOLINTEREST EUR GROWTHREALINVEST ARG does not Granger Cause VOLINTEREST EUR DUNEMPRATE ARG does not Granger Cause VOLREALINTEREST ARG DEMPRATE ARG does not Granger Cause VOLREALINTEREST ARG GROWTHREALINVEST ARG does not Granger Cause VOLREALINTEREST ARG

18 14 18

0.12724 0.11310 1.41721

0.88160 0.89431 0.27747

18 14 18

0.12798 0.77229 1.87850

0.88096 0.49030 0.19202

10 10 10

0.41030 0.15039 2.97760

0.68392 0.86412 0.14073

10 10 10

0.56661 0.87667 1.66061

0.60007 0.47166 0.27987

11 11 11

0.13773 0.23971 3.46332

0.87401 0.79405 0.10000

11 11 11

0.01988 0.07952 2.43737

0.98038 0.92452 0.16796

10 10 10

0.69747 0.93030 1.87016

0.54055 0.45344 0.24752

18

0.03811

0.96272

14

0.35379

0.71137

18

1.19252

0.33457

18

0.10576

0.90041

14

0.77773

0.48803

18

1.64289

0.23114

18

2.60862

0.11156

14 18

4.35821 0.46633

0.04747 0.63740

12

0.10970

0.89762

12

0.17749

0.84102

12

4.20507

0.06317

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

337

Table 3b Pairwise Granger causality tests for exogeneity, Brazil (sample: 1970–1993) Lags: 2; null hypothesis

Observed

F-statistic

Probability

DUNEMPRATE BRA does not Granger Cause VOLNER BRRUSD GROWTHEMP BRA does not Granger Cause VOLNER BRRUSD GROWTHREALINVEST BRA does not Granger Cause VOLNER BRRUSD DUNEMPRATE BRA does not Granger Cause VOLRER BRRUSD GROWTHEMP BRA does not Granger Cause VOLRER BRRUSD GROWTHREALINVEST BRA does not Granger Cause VOLRER BRRUSD DUNEMPRATE BRA does not Granger Cause VOLNER BRREUR GROWTHEMP BRA does not Granger Cause VOLNER BRREUR GROWTHREALINVEST BRA does not Granger Cause VOLNER BRREUR DUNEMPRATE BRA does not Granger Cause VOLRER BRREUR GROWTHEMP BRA does not Granger Cause VOLRER BRREUR GROWTHREALINVEST BRA does not Granger Cause VOLRER BRREUR DUNEMPRATE BRA does not Granger Cause VOLNER USDEUR GROWTHEMP BRA does not Granger Cause VOLNER USDEUR GROWTHREALINVEST BRA does not Granger Cause VOLNER USDEUR DUNEMPRATE BRA does not Granger Cause VOLRER USDEUR GROWTHEMP BRA does not Granger Cause VOLRER USDEUR GROWTHREALINVEST BRA does not Granger Cause VOLRER USDEUR DUNEMPRATE BRA does not Granger Cause VOLREER BRA GROWTHEMP BRA does not Granger Cause VOLREER BRA GROWTHREALINVEST BRA does not Granger Cause VOLREER BRA DUNEMPRATE BRA does not Granger Cause VOLNEERINTRAMERC ARG GROWTHEMP BRA does not Granger Cause VOLNEERINTRAMERC ARG GROWTHREALINVEST BRA does not Granger Cause VOLNEERINTRAMERC ARG

11

0.57322

0.59181

11

0.15391

0.86063

20

0.34319

0.71493

11

0.02013

0.98014

11

0.42746

0.67057

20

0.31563

0.73406

10

0.06160

0.94096

10

0.34632

0.72301

13

0.76528

0.49646

10

0.14623

0.86753

10

0.33368

0.73109

13

0.05090

0.95068

10

1.94686

0.23698

10

3.15206

0.13012

14

1.50949

0.27207

10

1.41545

0.32576

10

2.59733

0.16846

14

1.42910

0.28907

10 10 13

0.27249 1.51958 0.65453

0.77210 0.30507 0.54543

11

0.00899

0.99106

11

0.03552

0.96531

20

1.38112

0.28148

338

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

Table 3b (Continued ) Lags: 2; null hypothesis

Observed

F-statistic

Probability

DUNEMPRATE BRA does not Granger Cause VOLREERINTRAMERC BRA GROWTHEMP BRA does not Granger Cause VOLREERINTRAMERC BRA GROWTHREALINVEST BRA does not Granger Cause VOLREERINTRAMERC BRA DUNEMPRATE BRA does not Granger Cause VOLINTEREST BRA GROWTHEMP BRA does not Granger Cause VOLINTEREST BRA GROWTHREALINVEST BRA does not Granger Cause VOLINTEREST BRA DUNEMPRATE BRA does not Granger Cause VOLREALINTEREST BRA GROWTHEMP BRA does not Granger Cause VOLREALINTEREST BRA GROWTHREALINVEST BRA does not Granger Cause VOLREALINTEREST BRA

11

0.06423

0.93842

11

0.14094

0.87134

20

1.61251

0.23210

11

1.79351

0.24513

11

0.05188

0.94986

20

0.74368

0.49210

11

1.95062

0.22253

11

0.08504

0.91956

20

0.49117

0.62141

investment in Argentina and Brazil. We have argued that this result is due to the fact that all employment and investment decisions have some degree of irreversibility. This is not a test of the effect of exchange-rate and interest-rate variability on trade, but rather of their effects on investment and employment. Irreversibility of set-up costs plays a key role in our analytical framework and has crucial implications for long-run decisions, such as decisions to invest or to hire workers. In general, we find that exchange-rate and interest-rate variability have statistically significant impacts on investment and employment. Moreover, one would expect economies with relatively closer ties to the U.S., like Brazil, to show a stronger response to dollar-exchange-rate variability, a result that is confirmed by the data. The estimated impact coefficients for Argentina were typically smaller than for Brazil. But we also acknowledge that some aspects of the results remain unsatisfactory. The prior that intra-Mercosur exchange-rate volatility should have a greater impact on Argentina’s real sector because of greater trade exposure is only partially corroborated by the estimations. This is a general feature also of our earlier work, in the sense that for Europe we also did not find a systematic correlation between openness and the strength of the impact of exchange-rate volatility on trade. In addition, one has to be cautious in view of the rather weak quality of the Mercosur labor market data and in view of the as yet limited samples underlying our regressions which prevent serious tests for structural stability. What are the implications of the results for the debate on exchange-rate policy in Mercosur and on the design of intra-Mercosur monetary relations? By accepting our main result, one could jump to the policy conclusion that fixing exchange rates either within the Mercosur or

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

339

against G-3 currencies should bring about significant benefits. Our estimates are not precise enough to decide which option would yield larger benefits. Whether there are benefits, depends essentially on whether the gains from suppressing exchange-rate variability are lost if the volatility reappears elsewhere, for example, in a higher dollar-variability or higher interest-rate variability or the slow build-up of large disequilibria. We would argue that fixing the exchange rate might be beneficial if underlying policies are compatible with this choice.28 This is a big if as the experience of Argentina shows; if fiscal policy is out of control, then fixing the exchange rate might just suppress the appearance of the problem temporarily. In the case of Argentina, one might even argue that the currency board worked too well for too long, thus allowing a considerable disequilibrium to accumulate under the surface. The explosion that occurred in the end might have such high costs that it can easily offset the benefits of a stable exchange rate accumulated in the preceding 10 years. In sum, we maintain that the high degree of exchange-rate variability observed in Mercosur has tangible economic costs, but that fixing exchange rates does not necessarily offer a free lunch. Notes 1. All data come from Center for Global Trade Analysis (2001) at Purdue University. 2. See Belke and Gros (2002b, Table 3). 3. However, even the lower values for Brazil (25.8%) and Uruguay (28.5%) are high compared with the EU trade bloc (5.5%). 4. See also Dixit (1989), Belke and Kaas (2002). 5. Aizenman and Marion (1999) provide further empirical evidence on a negative relation between various volatility measures and private investment. They argue that increasing volatility has a negative impact on investment if investors are disappointment-averse. Moreover, in the presence of credit constraints, realized investment is on average lower when investment demand is more volatile, since credit constraints bind more often. Real impacts of volatility are also confirmed by Ramey and Ramey (1995). 6. For a detailed exploration of the relationship between monetary policy and exchange-rate volatility in a small open economy, see Gal´ı and Monacelli (2002). 7. Formally, the wage bargain leads to a wage rate maximizing the Nash product (2w − 2w)β (2p∗ − 2w)1−β , whose solution is w = (1 − β)w + βp ∗ , and hence the expected net return for the firm is 2p ∗ − 2w − c = (1 − β)(2p ∗ − 2w) − c. 8. Of course, such a flexible contract implies that some exchange-rate risk is shared between the worker and the firm. However, the reason why they both benefit is not the risk-sharing aspect, but the fact that the flexible contract excludes continuation of unprofitable work relationships. 9. We do not a priori restrict the sign of the barrier b. Hence, one of these conditions is automatically satisfied, whereas the other is satisfied only if uncertainty is large enough. 10. Formally, this results from the fact that Eq. (4) is only valid whenever σ 1 exceeds b (otherwise the exchange rate could never exceed the barrier and the firm never

340

11. 12.

13.

14. 15. 16.

17.

18. 19. 20.

21.

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

creates a job in period one) and whenever −σ 1 is lower than b (otherwise the exchange rate could never fall below the barrier and the firm always creates a job in period one). The other (smaller) solution to this equation is less than |b| and is therefore not feasible. Klein, Schuh, and Triest (2000) investigate the impact of exchange-rate movements on job flows in the U.S. They find a response of job destruction to dollar appreciation, whereas job creation does not respond significantly to depreciations. This result reflects the asymmetric responses of job creation and destruction to aggregate shocks that have been detected in other studies. It does not contradict our conclusions, however, since job creation might just respond to exchange-rate volatility rather than to actual appreciations or depreciations. See also Eichengreen (1998, pp. 30 ff.) and Levy Yeyati and Sturzenegger (2000, pp. 73 ff.) for a survey on this issue. The foregoing ignores the potentially large grey or underground economy. The focus on the official labor market is, however, entirely appropriate. In the grey economy the costs of firing are presumably lower, because official employment regulations do not apply. This implies that our model of firing costs applies mainly to official employment and we would expect volatility to be mainly a deterrent to official employment. Further details are given in Belke and Gros (2002b). As might have been expected, the correlation between dollar and euro volatilities of the two Mercosur currencies is near 99%. In the variable names, VOL stands for volatility, NER for the nominal exchange rate, RER for the real exchange rate, ARP for the Argentinian peso, BRR for the Brazilian real, EUR for the euro, and USD for the U.S. dollar. INTRAMERC ARG indicates exchange rates between the peso and other Mercosur currencies. INTEREST and REALINTEREST stand for nominal and real interest rates, respectively. DUNEMPRATE defines changes in the unemployment rate, EMPGROWTH refers to employment growth, and GROWTHREALINVEST defines changes in real gross fixed capital formation. Details on data sources and construction are provided in Appendix B. Since the results of our unit root tests clearly support the standard result in the literature that exchange-rate volatility is integrated of order 0, i.e., stationary, we only display the results based on the levels of this variable. However, as shown by our discussants, the conclusions are throughout the same as in the I(1) case. The significance of the coefficient estimates of the lags is based on the Student-t-distribution. See Belke and Gros (2001a, 2002a) and Haldrup (1990). See Belke and Gros (2001a, 2002a) for details. However, one important precondition for their application is the same number of observations for the alternative specifications. See Banerjee et al. (1993, p. 286), Mills (1990, p. 139), and Schwarz (1978). In the case of Argentina, we ended the sample in 1990, thereby excluding the transition from different attempts to fix or to control the exchange rate under Alfons´ın and Menem to the currency board. In the case of Brazil, we let the sample run until the year 1993.

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

341

22. The results, which are available on request, do not confirm this possible caveat. The concern that such tests would be necessary was raised by Campos e Cunha and Alves (2002). 23. For Argentina, the results are more significant when the nineties are excluded from the sample. Even a dummy variable for the currency board does not help. Our results with respect to the significance of the volatility variables are corroborated by Campos e Cunha and Alves (2002) with slightly different data. 24. See Belke and Gros (2002b). 25. The full-sample results for Brazil and also Uruguay can be found in Appendix A. 26. See Belke and Gros (2002b). 27. In order to test for robustness, i.e., for missing variables, Belke and Gros (2002b) try to take into account the two most plausible ways in which our measures of exchange-rate and interest-rate variability could stand for some other variables. The two hypotheses we consider are: (i) exchange-rate variability is just a sign of a misalignment (i.e., a wrong level of the exchange rate), and (ii) interest-rate variability just reflects the financial stress defined as high real (short-term) interest rates. In Section 2, we saw that for both Argentina and Brazil, the EU is a more important trading partner than NAFTA. However, we do not find in our robustness tests that exchange-rate variability vis-à-vis the euro is more important than that vis-à-vis the dollar, as the point estimates are in most cases virtually indistinguishable. This is again indirect confirmation that it is not trade that matters here at all, but destabilizing domestic monetary policies that cause volatility in interest rates and exchange rates. Hence, future work should check whether the relationship between exchange-rate volatility and investment and/or the labor market is driven by volatility in, e.g., the money supply. 28. Recent work by Frankel and Rose (2002) suggests that the move to a currency union (permanent and credible exchange-rate stability) has a large effect on trade flows and income per capita. This raises the possibility that even though trade flows are low, as in Mercosur, the benefits from exchange-rate stability might be significant in Mercosur. We are grateful to Lu´ıs Campos e Cunha and Nuno Alves for drawing our attention to this aspect.

Acknowledgments We are grateful to Kai Geisslreither, Ralph Setzer (University of Hohenheim), and Oliver Kreh (Stuttgart Chamber of Commerce) for excellent research assistance and to Roberto Duncan (Central Bank of Chile) for the delivery of valuable data. We also profited very much from comments by two anonymous referees, and by Sven Arndt, Eduard Hochreiter, Lu´ıs Campos e Cunha and Nuno Alves, and other participants in the Conference “Towards Regional Currency Areas,” Economic Commission for Latin America and the Caribbean (ECLAC), Santiago de Chile, March 26–27, 2002, and in the Conference “Monetary Union: Theory, EMU Experience, and Prospects for Latin America” (Österreichische Nationalbank, Vienna-University, and Banco Central de Chile), April 14–16, 2002.

342

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

343

344

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

Appendix B. Data sources CPI ARG: Consumer Price Index Argentina (1995 = 100). Source: Instituto Nacional de Estad´ıstica y Censos (http://www.indec.mecon.gov.ar). CPI BRA: Consumer Price Index Bra‘zil (1995 = 100). Source: IFS (IMF) series CPI (22364 . . . ZF . . . ) + IMF − Statistical Yearbook and various Monthly Reports. CPI EUR: Consumer Price Index (1995 = 100). Source: Until December 1994 Bundesbank, from January 1995 on ECB. CPI US: Consumer Price Index (1995 = 100). Source: IFS (IMF) series CPI (11164 . . . ZF . . . ) + IMF − Statistical Yearbook and various Monthly Reports. CPI UY: Consumer Price Index Uruguay (1995 = 100). Source: IFS (IMF) series CPI + IMF − Statistical Yearbook and various Monthly Reports. DNER USDEUR: = D(LOG(NER USDEUR)) × 100; growth rate of the nominal dollar exchange rate of the euro; the remaining exchange rate growth rates are constructed analogously. EMP BRA: Employment general level Brazil (in thousands): Persons aged 10 years and over. Excl. rural population of Rondˆonia, Acre, Amazonas, Roraima, Par´a and Amap´a. Sep. of each year. Prior to 1979: excl. rural areas of Northern Region, Mato Grosso, Goi´as and Tocantins. 1992 methodology revised; data not strictly comparable. Source: LABORSTA (http://laborsta.ilo.org/), IFS (IMF) and http://www4.bcb.gov.br/series-i/default.asp. EMPRATE ARG: Evoluci´on de la las principales variables ocupacionales (en % of employed population to total population), Empleo, Tasa de Empleo en Aglomerados Urbanos. Source: Encuesta Permanente de Hogares, INDEC (http://www2.mecon.gov.ar/infoeco/). EMP UY: Employment general level (in thousands), urban areas, incl. Professional army; excl. compulsory military service, persons aged 14 years and over. 1984 and 1986 first ´ IMPORTANTE: Hasta el año 1997 la encuesta cubr´ıa a las semester, ACLARACION localidades de 900 y m´as habitantes y a partie del año 1998 cubre de 5.000 o m´as habitantes. Source: IFS (IMF), LABORSTA (http://laborsta.ilo.org/), Instituto Nacional de Estad´ıstica (http://www.ine.gub.uy/), Principales Resultados Encuesta Continua de Hogares. INTEREST ARG: Deposit Rate (in home currency). Source: IFS (IMF) series 21360L. . . ZF. . . . INTEREST BRA: Money Market Rate (in home currency). Source: IFS (IMF) series 22360B. . . ZF. . . . INTEREST EUR: Until December 1994: German money market rate. Source: Bundesbank; from January 1995 on: 3-month rate. Source: ECB, Monthly Reports. INTEREST US: Treasury bill rate. Source: Federal Reserve Bank. INTEREST UY: Deposit Rate (in home currency). Soure: IFS (IMF) series. INVEST ARG: Gross Fixed Capital Formation Argentina (millions of Arg. peso). Soure: IMF Statistical Yearbook, IFS (IMF). INVEST BRA: Gross Fixed Capital Formation Brazil (millions of real). Soure: IMF Statistical Yearbook, IFS (IMF).

345

346

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

INVEST UY: Gross Fixed Capital Formation Uruguay (millions of Urug. peso). Soure: IMF Statistical Yearbook, IFS (IMF). NER ARPUSD: IMF—Statistical Yearbook and various Monthly Reports. NER BRRUSD: IMF—Statistical Yearbook and various Monthly Reports. NER PYGUSD: IMF—Statistical Yearbook and various Monthly Reports. NER UYPUSD: Banco Central del Uruguay (until June 1973) and IMF—Statistical Yearbook and various Monthly Reports (from July 19973 on). NER USDEUR: Bilateral nominal U.S. $/ECU exchange rate period average. Soure: IMF—Statistical Yearbook and various Monthly Reports, IFS (IMF) series 111. . . EB.ZF. . . . The remaining bilateral nominal exchange rate series were created via cross-rates. NEER EUR: Nominal effective exchange rate euro zone. Soure: IFS (IMF) series 163. . . NEUZF. . . . NEER US: Nominal effective exchange rate of the U.S.-dollar based on unit labor costs. Source: IFS (IMF) series 111. . . NEUZF. . . . NEER UY. Source: IFS (IMF) series. REALINTEREST ARG: Real short-term interest rate of Argentina; INTEREST ARG deflated by the consumer price index. REALINTEREST BRA: Real short-term interest rate of Brasil; INTEREST BRA deflated by the consumer price index. REALINTEREST EUR: Euroland real short-term interest rate; INTEREST EUR deflated by the consumer price index. REALINTEREST US: U.S. real short-term interest rate of Argentina; INTEREST US deflated by the consumer price index. REALINTEREST UY: Real short-term interest rate of Uruguay; INTEREST UY deflated by the consumer price index. REER US: Real effective exchange rate of the U.S.-dollar based on unit labor costs. Soure: IFS (IMF) series 111. . . REUZF. . . . REER EUR: Real effective exchange rate Euro area based on unit labor costs. Soure: IFS (IMF), series 163. . . REUZF. . . . REER ARG: Annual data: Real effective exchange rate Argentina in terms of import prices. Soure: Comisi´on Econ´omica para Am´erica Latina y el Caribe (http://www.eclac.org/publicaciones/DesarrolloEconomico). Monthly data: REER ARG = 4.739 × RER ARPJPY + 22.058 × RER ARPUSD + 35.402 × RER ARPEUR + 35.004 × RER ARPBRR + 2.797 × RER ARPUYP (weights from Center for Global Trade Analysis, 2001: exports + imports). REER BRA: Annual data: Real effective exchange rate Brazil in terms of import prices. Soure: Comisi´on Econ´omica para Am´erica Latina y el Caribe (http://www.eclac.org//publicaciones/DesarrolloEconomico). Monthly data: REER BRA = 8.258 × RER BRRJPY + 31.974 × RER BRRUSD + 41.362 × RER BRREUR + 16.431 × (1/RER ARPBRR) + 1.974 × RER BRRUYP (weights from Center for Global Trade Analysis, 2001: GTAP 5: exports + imports). REER UY: Real effective exchange rate based on relative CPI. Soure: IMF—Statistical Yearbook and various Monthly Reports.

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

347

UNEMPRATE ARG: Evoluci´on de la las principales variables ocupacionales (en %), Desocupaci´on (in percent). Soures: Encuesta Permanente de Hogares, INDEC (http://www.indec.mecon.gov.ar/DEFAULT.HTM). Series Hist´oricas, Tasas de actividad, empleo, desocupaci´on y subocupaci´on, total de Aglomerados Urbanos, años 1974/2000 (semi-annual data). UNEMPRATE BRA: Unemployment rate Brazil (in percent), Taxa de Desemprego Aberto—Original e Dessazonalizada—Taxas Medias 30 dias. Soure: http://www.ibge.gov.br. UNEMPRATE URU: Unemployment rate Uruguay (in percent). Soure: Instituto Nacional de Estadistica INE, TASA DE DESEMPLEO ANUAL—Total Pa´ıs urbano y Por Departamento (http://www.ine.gub.uy/bancodedatos/ECH/ECH%20TOT%20Des%20A.xls). VOLNEER EUR: Exchange rate variability from NEER EUR. VOLNEER US: Exchange rate variability from NEER US. VOLREER EUR: Exchange rate variability from REER EUR. VOLREER US: Exchange rate variability from REER US. VOL USDEUR: Exchange rate variability from NERDOLLECU. VOLREERINTRAMERC ARG = 0.926 × volrer arpbrr + 0.074 × volrer arpuyp. VOLREERINTRAMERC BRA = 0.8927 × volrer arpbrr + 0.1073 × volrer brruyp. VOLREERINTRAMERC UY = 0.60 × volrer brruyp + 0.40 × volrer arpuyp. VOLNEERINTRAMERC BRA = 0.8927 × volner arpbrr + 0.1073 × volner brruyp. VOLNEERINTRAMERC ARG = 0.926 × volner arpbrr + 0.074 × volner arpuyp. VOLNEERINTRAMERC UY = 0.60 × volner brruyp + 0.40 × volner arpuyp. (weights = exports + imports weights from Center for Global Trade Analysis, 2001 for consistency reasons) References Aizenman, J., & Marion, N. P. (1999). Volatility and investment: Interpreting evidence from developing countries. Economica, 66, 157–179. Amadeo, E. J., Gill, I., & Neri, M. (2000). Brazil: The pressure points in labor legislation. Ensaios Econˆomicos da EPGE, No. 395, Rio de Janeiro. Andersen, T. M., & Sorensen, J. R. (1988). Exchange rate variability and wage formation in open economies. Economics Letters, 28, 263–268. Baldwin, R., & Krugman, P. (1989). Persistent trade effects of large exchange rate shocks. Quarterly Journal of Economics, 104, 635–654. Banerjee, A. et al. (1993). Co-integration, error correction, and the econometric analysis of non-stationary data, Oxford. Belke, A., & Gros, D. (2001a). Real impacts of intra-European exchange rate variability: A case for EMU? Open Economies Review, 12, 231–264. Belke, A., & Gros, D. (2002a). Designing EU–U.S. Atlantic monetary relations: Exchange rate variability and labor markets. The World Economy, 25, 789–813. Belke, A., & Gros, D. (2002b). Monetary integration in the Southern Cone: Mercosur is not like the EU? Paper presented at the conference “Towards Regional Currency Areas,” Economic Commission for Latin America and the Caribbean (ECLAC), Santiago de Chile, March 26–27, 2002. Belke, A., & Kaas, L. (2002). The impact of exchange rate volatility on labor markets: Europe versus United States. Paper presented at the 2002 Meeting of the European Economic Association, Venice.

348

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

Calvo, G. A., & Reinhart, C. M. (2000). Fixing for your life. NBER Working Paper No. 8006. Cambridge, MA: NBER. Campos e Cunha, L., & Alves, N. (2002). Comments on Belke, A. & Gros, D., monetary integration in the Southern Cone: Mercosur is not like the EU? OENB Working Paper 72, Österreichische Nationalbank, Vienna. Canzoneri, M. B., Vallés, J., & Viñals, J. (1996). Do exchange rates move to address national imbalances? CEPR Discussion Paper No. 1498. London: CEPR. Center for Global Trade Analysis. (2001). GTAP version 5 database. Purdue University, USA. Cˆoté, A. (1994). Exchange-rate volatility and trade—A survey. Working Paper 94-5. Ottawa, Canada.: Bank of Canada. D´ıaz-Bonilla, E., & Schamis, H. E. (2001). From redistribution to stability: The evolution of exchange rate policies in Argentina, 1950–98. In Frieden, J., & Stein, E. (2001), The currency game—Exchange rate politics in Latin America (pp. 65–118). Washington, DC: Inter-American Development Bank. Dixit, A. (1989). Entry and exit decisions under uncertainty. Journal of Political Economy, 97, 620–638. Edwards, E., & Cox Edwards, A. (2000, April). Economic reforms and labor markets: Policy issues and lessons from Chile. Paper presented at the 29th Economic Policy Panel, Frankfurt, Germany. Eichengreen, B. (1998). Does the Mercosur need a single currency? Keynote address for the IPEA Conference “ALCA and MERCOSUL—The Brazilian Economy and the Process of Subregional and Hemispheric Integration,” Brasilia, October 5–6, 1998. Flood, R. P., & Rose, A. K. (1995). Fixing exchange rates—A virtual quest for fundamentals. Journal of Monetary Economics, 36, 3–37. Frankel, J., & Rose, A. (2002). Estimating the effect of currency unions on trade and output. CEPR Discussion Paper No. 2631. London: CEPR. Gal´ı, J., & Monacelli, T. (2002). Monetary policy and exchange-rate volatility in a small open economy. NBER Working Paper No. 8905. Cambridge, MA: NBER. Haldrup, N. (1990). Testing for unit roots with a maintained trend when the true data generating process is a random walk with drift. Institute of Economics Memo 1990–22, University of Aarhus. Heckman, J., & Pagés, C. (2000). The cost of job security regulation: Evidence from Latin American labor markets. Research Department Working Paper No. 430. Washington, DC: Inter-American Development Bank. Hopenhayn, H. A. (2001). Labor market policies and employment duration: The effects of labor market reform in Argentina. Research Department Working Paper No. 407. Washington, DC: Inter-American Development Bank. Klein, M. W., Schuh, S., & Triest, R. K. (2000). Job creation, job destruction, and the real exchange rate. NBER Working Paper No. 7466. Cambridge, MA: NBER. Krugman, P. (1989). Exchange rate instability. Cambridge, MA: MIT Press. Lafrance, R., & Tessier, D. (2000, November). Exchange rate variability and investment in Canada. Bank of Canada Conference on Revisiting the Case for Flexible Exchange Rates, Ottawa, Canada. Levy Yeyati, E., & Sturzenegger, F. (2000). Is EMU a blueprint for Mercosur? Cuardernos de Econom´ıa, 37, 63–99. Lütkepohl, H. (1991). Introduction to multiple time series analysis. New York: Springer. Layard, R., Nickell, S., & Jackman, R. (1991). Unemployment: Macroeconomic performance and the labor market. Oxford: Oxford University Press. Márquez, G., & Pagés, C. (1998). Ties that bind: Employment protection and labor market outcomes in Latin America. Research Department Working Paper No. 373. Washington, DC: Inter-American Development Bank. Mills, T. C. (1990). Time series techniques for economists. Cambridge: Cambridge University Press. Nickell, S. (1997). Unemployment and labor market rigidities: Europe versus North America. Journal of Economic Perspectives, 11, 55–74. Obstfeld, M., & Rogoff, K. (2000). The six major puzzles in international macroeconomics: Is there a common cause? (pp. 339–390). In B. Bernanke & K. Rogoff (Eds.), NBER Macroeconomics Annual, 2000. Cambridge: MIT Press. Pissarides, C. (2000). Equilibrium unemployment theory (2nd ed.). Cambridge, MA: MIT Press. Ramey, G., & Ramey, V. A. (1995). Cross-country evidence on the link between volatility and growth. American Economic Review, 85, 1138–1151.

A. Belke, D. Gros / North American Journal of Economics and Finance 13 (2002) 323–349

349

Reinhart, C. M., & Reinhart, V. R. (2001). What hurts most? G-3 exchange rate or interest rate volatility? NBER Working Paper No. 8535. Cambridge, MA: NBER. Rose, A. K. (1996). After the deluge: Do fixed exchange rates allow inter-temporal volatility trade-offs? International Journal of Finance and Economics, 1, 47–54. Rowthorn, R. (1999). Unemployment, wage bargaining and capital–labour substitution. Cambridge Journal of Economics, 23, 413–425. Schwarz, G. (1978). Estimating the dimension of a model. Annals of statistics, 6, 461–464. Sorensen, J. R. (1992). Uncertainty and monetary policy in a wage bargaining model. Scandinavian Journal of Economics, 94, 443–455.