Exchange rate arrangements in the accession to the EMU.

Coricelli, Fabrizio ; Jazbec, Bostjan

INTRODUCTION

In this paper, we investigate the dynamics of real exchange rates in candidate countries and the implications for inflation dynamics. We argue that the observed real appreciation of the last few years can be partially ascribed in most cases to the workings of the Balassa-Samuelson effect.

We also found that the real appreciation trend results from higher domestic inflation rather than a nominal appreciation of the exchange rate. Interestingly, inflationary pressures seem to arise irrespective of the exchange rate regime and could be partially attributed to rigidities in the non-tradable sectors in transition economies. However, neither the Balassa-Samuelson effect nor these rigidities can be considered a justification for flexible exchange rates. Indeed, wage and price setters internalise the exchange rate accommodation rule and a higher inflation rate would arise. A non-accommodating stance of the exchange rate policy--for instance, an early adoption of the euro--would avoid this outcome, although the resulting benefits may not be the same for every country (Nuti, 2000). Output in the non-tradable sector would decline temporarily, but the adverse effects of monopoly power on welfare would be contained as the rate of inflation declines. The effects on price setting in the non-tradable sector are likely to more than compensate for the Balassa-Samuelson effects. Nevertheless, if the Balassa-Samuelson effects were predominant, a change in the Maastricht criteria would be advisable (Buiter and Grafe, 2002).

The paper proceeds as follows. Next section contains a short overview of stylised facts and different interpretations of the real exchange rate appreciation in transition economies, including evidence on the Balassa-Samuelson effect in transition economies. Section on A simple analytical framework and econometric results presents a simple analytical framework and econometric results on determinants of real exchange rate behaviour. It is argued that structural reforms implemented in transition economies have indeed determined the level of real exchange rate during the transition process. These results are used to draw conclusions on the implications for Maastricht criteria and the choice of an exchange rate regime. Finally, the conclusions are presented.

STYLISED FACTS AND INTERPRETATIONS

All transition economies have undergone major reforms, all of which have had appreciable consequences for their real exchange rate. Changes in production and productivity, trade liberalisation and removal of state subsidies, restrictive monetary policy accompanied by tax reform, the underlying process of financial innovations and bank restructuring, are among the factors that played an important role in determining the real exchange rate in transition economies.

As reform strategies differed across countries, one would have expected that real exchange rate development also should have differed across transition countries. However, in the early days of transition, real exchange rate paths were similar in all transition economies. In general, transition started with the abrupt depreciation of local currencies that accompanied the end of a command economy and the dismantling of previously prevailing multiple exchange rates. Despite differences in monetary and real shocks that these countries have experienced, real exchange rate movements in all transition economies have followed the same time path.

Following the initial undervaluation, the real exchange rate subsequently appreciated. The appreciation of domestic currencies was associated with two phenomena (Roubini and Wachtel, 1999). First, the appreciation was a response to the initial undervaluation of the real exchange rate. And second, it reflected an appreciation of the equilibrium real exchange rate. Among the determinants of the equilibrium real appreciation, the so-called Balassa-Samuelson effect has been singled out as the most important.

Balassa-Samuelson Effect in Transition Economies

Studies on the behaviour of real exchange rates in transition economies emphasise the productivity approach to explain the trend appreciation of the real exchange rate (Egert, 2002; Halpern and Wyplosz, 1996). There is vast potential for gains in productivity in transition economies both through more efficient use of existing resources and technologies and through upgrading technology. However, this approach should also take into account the initial conditions in transition economies at the beginning of reforms (Coricelli and Jazbec, 2001). Decades of central planning have resulted in distorted structures of these economies. Industries were favoured by the emphasis of central planners on material production, while services were largely neglected. The structure of the economy was reflected in distorted price levels, as empirical studies on price development in transition economies indicate. Transition and the introduction of market-determined prices along the other market-enhanced reforms have brought about massive changes in output, employment, and last but not least, relative prices.

To explain the price differential used to measure the real exchange rate, let us assume that there is an economy-wide wage that is equal to the marginal product of labour in each sector. To the extent that there are differences in productivity between countries, wages will differ as well. In less-developed countries, productivity is generally lower than in more developed countries. While this applies to both sectors of the economy, there is evidence that the productivity gap is larger for tradables than it is for non-tradables. Also, the scope for productivity gain is more limited in non-tradables than in tradables. Because of this, the price of non-tradables will typically be lower in less-developed countries than in industrial countries. Since the overall price level is a weighted average of the price levels of tradable and non-tradable goods, the general price level will be lower in less-developed countries, with the difference being a function of the proportion of goods that are non-tradable, and the price differential for non-tradables (Richards and Tersman, 1996). As an increase in tradable productivity is the main determinant of economic growth--assuming that non-tradable productivity is more or less the same across countries--higher relative growth is reflected in a more appreciated real exchange rate.

Generally, countries that have grown faster during the transition process have experienced stronger real exchange rate appreciation. Also, poor performers--in our case, Bulgaria and Romania--have experienced strong appreciation owing to larger distortions and poor initial conditions at the beginning of the transition process. Figure 1 presents the cumulative change in GDP from 1995 to 2001 plotted against the cumulative change in the real exchange rate index.

[FIGURE 1 OMITTED]

Visual inspection does not reveal a clear relationship between real exchange rate changes and output growth. However, this cannot be taken as a proof of the absence of the Balassa-Samuelson effect, as many other factors affect real exchange rate movements, and possibly with varying effects on different countries. In what follows, the extent of the Balassa-Samuelson effect is estimated in a framework, which enables one to disentangle the effect of structural reforms at earlier stages of the transition process and the pure Balassa-Samuelson effect in recent years on the level of the real exchange rate in transition economies. The results broadly coincide with other studies on transition economies (Egert, 2002; Halpern and Wyplosz, 1996; Krajnyak and Zettelmeyer, 1997; De Broek and Slok, 2001; and various IMF transition country studies) with respect to the existence of the Balassa-Samuelson effect in transition economies. However, the extent of its effect is rather lower than in comparable studies, as the effect of structural reforms on the level of the real exchange rate is separated from the pure productivity differential effect. To explain the separation of structural reforms from the working of the Balassa-Samuelson effect, a simple model is introduced in the next section.

A SIMPLE ANALYTICAL FRAMEWORK AND ECONOMETRIC RESULTS (1)

It can be shown that in a transition economy, the labour market adjustment owing to structural changes in the economy may affect the real exchange rate determination. The flow of labour from one sector to another is one of the indicators of structural reforms, and as such, the determinant of the real exchange rate in transition economies. It follows that the tradable sector employs relatively more labour than the non-tradable production. The bias in the production of tradables was widely observed in formerly centralised economies. These economies were inclined to favour heavy industry and industrial production at the expense of private housing, consumer goods and services (see Melitz and Waysand, 1996). (2) There exists a level of tradable production for which the wages in both sectors are the same. The relative price of a tradable good in terms of non-tradables is determined on the basis of labour employment decisions in both sectors of the economy. For this reason, the relative price of tradables in terms of non-tradables in the pretransition period is greater than 1. If the measure of the real exchange rate is taken to be the relative price of tradables in terms of non-tradables, then one could say that the value of the domestic currency is overvalued.

The price of tradable goods is determined in the world market and considered to be given by its pre-transition level. The average price in the economy is, therefore, a function of central planners' preference for the tradable goods production. Since technology is given by the same production functions in both sectors of the economy, the real equilibrium wage can easily be defined solely as a function of the central plan and thus abstracted from market forces. Consequently, the preferences of central planners stand as a proxy for the initial conditions in transition economies. The higher the required volume of the tradable goods production, the greater the initial price discrepancy between the relative price tradables in terms of non-tradables. For this reason, transition economies inherited pressure for real exchange rate appreciation.

Figure 2 depicts the initial conditions in transition economies in a simple two-sector factor-specific model (capital is specific to the sector, while labour is mobile). On the vertical axis are reported wages, measured in terms of tradable goods, and the value of marginal product of labour in the two sectors: non-tradables on the left and tradables on the right. The two lines depict the value of the marginal product of labour in the two sectors in relation to levels of employment. The real pre-transition wage measured in terms of tradable goods is below the market equilibrium, forcing the economy to employ at E where there is more labour in tradables than in non-tradables production.

[FIGURE 2 OMITTED]

Price liberalisation and the beginning of the transition process pushes labour demand in non-tradable production to (MP[L.sub.NT][P.sub.NT])'. The value of the marginal product of labour in the non-tradable sector jumps to a new steady state. Monopoly power in the non-tradable sectors implies that higher wages can be accommodated through higher prices. Consequently, the real exchange rate appreciates and labour shifts from the tradable to the non-tradable sector. If we assume the Law of One Price for tradable goods, devaluation of the nominal exchange rate may counteract the increase of real wages, if measured in terms of tradable goods, thereby allowing a higher level of employment in the non-tradable sector. Nonetheless, wages in the non-tradable sector could still be higher than in the tradable sector after the initial price liberalisation. Wages in the tradable sector cannot adjust instantaneously, as their value crucially depends on productivity in that sector and on international competition. As long as there are forces of monopolistic competition at work in the non-tradable sector or increased trade union pressure in that sector, wages in the non-tradable sector will be higher than in the tradable sector, implying unemployment in the economy.

As long as we observe a large reallocation of resources across sectors, Figure 2 may well serve for an explanation of the transition process in the first few years following the initial price liberalisation and cuts in subsidies to the state-owned sector. Models of transition (Aghion and Blanchard, 1993; Chadha and Coricelli, 1997) pay special attention to labour market dynamics once the transition process has begun. The observed productivity increase in tradable sectors at the beginning of transition is, therefore, due mainly to labour shedding in that sector. Once the initial labour reallocation is settled, productivity increase in the tradable sector may occur mainly because of technology advances in that sector. It is only then that the pure Balassa-Samuelson effect takes place. Figure 3 shows the working of the Balassa-Samuelson effect once market forces correct for an initially distorted labour market. Labour productivity in tradables increases from MP[L.sub.T] to MP[L.sub.T]' causing wages and employment in the tradable sector to increase correspondingly. Monopolistic competition in the non-tradable sector or powerful trade unions bid up wages in that sector trying to maintain the difference between wages in both sectors that has been established after price liberalisation is intact. Wages in the non-tradable sector can increase only through price increases in that sector causing real exchange rate appreciation and confirming the working of the Balassa-Samuelson effect. Correspondingly, the value of the marginal product of labour in the non-tradable sector shifts from (MP[L.sub.NT] [P.sub.NT])' to (MP[L.sub.NT] [P.sub.NT])". Employment in the non-tradable sector can either slightly increase or stay the same. A shift in employment in the non-tradable sector, therefore, crucially depends on the competition in that sector or on trade union power.

[FIGURE 3 OMITTED]

The simple model suggests that the real exchange rate appreciated during the first few years of the transition process mainly because of structural reforms taking place. Price liberalisation was accompanied by a large reallocation of resources and real exchange rate appreciation. Monopolistic competition in non-tradable sectors, which was almost implied by the nonexistence of that sector before transition and possibly the increasing power of trade unions, helps to explain the relatively higher wages in the non-tradable sector in transition economies.

Summing up, one would expect an increase in non-tradable wages relative to tradable wages at the beginning of transition. After initial price liberalisation and labour reallocation, relative wages should fall. The fall in relative wages should correspond to an increase in labour productivity differential between the tradable and non-tradable sector as wages in the tradable sector start to grow because of higher productivity. However, the correction of relative wages, which were established during the first years of transition, depends on the degree of competition and trade union power in the non-tradable sector. As long as firms in the non-tradable sector can take advantage of monopolistic power and trade unions can successfully negotiate wage increases, relative wages in the non-tradable sector measured in terms of wages in the tradable sector may stay constant or even increase. Data for CEE countries, Bulgaria, Romania, and the Baltics seem to confirm this line of reasoning. Figure A1 in Appendix A depicts wages in the non-tradable sector relative to wages in manufacturing for ten transition economies. Data are from the ILO database. Wages in the non-tradable sector are presented by unweighted average over sectors. The non-tradable sector consists of sectors from E to O in NACE classification of economic activity. In addition to the entire non-tradable sector (E-O), relative wages in market and public sector services measured in terms of manufacturing wages are shown separately. Market services are represented by sectors E to K, while public sector services are represented by sectors L to O. Wages in the manufacturing sector are represented by sector D. (3)

In all transition economies, relative wages in the non-tradable sector increased in the first few years of the transition process. After that, the relative wages in the non-tradable sector declined perhaps as a result of further labour reallocation or/and productivity increase in the tradable sector. Although the magnitude of the relative wage swing is different across countries, they all mimic the hump shape of the relative wage path in the first few years of the transition process. Furthermore, all countries thereafter resume a slight increase in the relative wage in the non-tradable sector. However, the magnitude of the relative wage increase in the later years of transition was much higher in Latvia, Romania, and Bulgaria than in other countries. Also, in Slovenia, the relative wage in the non-tradable sector is much higher than that in other countries. Slovenia also distinguishes itself from other countries in the behaviour of the relative wage in the public sector--the relative wage in the public sector is much higher than in any other country. (4)

Although similar patterns of the relative wage of non-tradables in terms of tradables are observed in transition economies, it is instructive to look at labour market developments during the transition process in order to disentangle the working of the Balassa-Samuelson effect. Figure A2 in Appendix A displays labour reallocation together with productivity differential between industry and services in selected transition economies. The criterion for the period of observation was the year after which the relative price of tradables in terms of non-tradables began to decline monotonically. The time series ends in 1998. (5) Data correspond to empirical results presented in the next section. (6)

Labour reallocation in Figure A2 is represented by the ratio between labour employed in industry and services, which encompass both market and public services, respectively. Both measures--labour ratio and productivity differential--are indexed to the base year that corresponds to the beginning of the transition process. Although the statistical properties of the corresponding time series are not thoroughly examined due to the short time interval, the polynomial trend of order 3 is added to Figure A2 to ease the explanation. In all cases, labour reallocation took place at the beginning of transition. Except in Romania, the labour reallocation process stabilised into the fifth or sixth year of transition. In this respect, conclusions from Figure 2 seem to explain the extent of structural reforms presented by the labour reallocation process. One would expect that productivity differential between labour productivity in industry and services would begin to increase. However, in the case of Estonia and Slovak Republic, the productivity differential even declined in the first few years, while in Lithuania, Latvia, and Romania, it increased only by a modest amount. Moreover, one can see from Figure A2 that only more advanced transition economies experienced an increase in productivity differential once the labour reallocation ended. Surprisingly, Bulgaria experienced the most dramatic increase in productivity differential prior to 1998. For this reason, we can distinguish cases where the Balassa-Samuelson effect has indeed taken place after the initial period of labour reallocation (Poland, Czech Republic, Hungary, and Slovenia) and cases where the increase in productivity differential was mainly due to labour reallocation. The dynamics of the ratio of labour employed in industry and services together with productivity differential broadly corresponds to the dynamics of the relative wage in the non-tradable sector measured in terms of tradables presented in Figure A1.

Empirical results

In this section, we give a quantitative assessment of the impact of different factors on the dynamics of real exchange rates in transition economies.

As shown in Coricelli and Jazbec (2001), the real exchange rate measured as the relative price of tradables in terms of non-tradable goods negatively depends on the productivity differential, the share of non-tradable consumption in total private consumption, and real government consumption. The regression equation is as follows:

(1) log[([P.sub.T]/[P.sub.N]).sub.i,t] = [[alpha].sub.oi] - [[alpha.sub.1] log[([a.sub.T] - [a.sub.N]).sub.i,t] - [[alpha].sub.2] [share.sub.i,t] - [[alpha.sub.3][govreal.sub.i,t] + [[alpha].sub.4][lab.sub.i,t] + [[epsilon].sub.i,t],

where [([P.sub.T]/[P.sub.N]).sub.i,t] is the relative price of tradables in terms of non-tradable goods; [([a.sub.T]-[a.sub.N]).sub.i,t] is the productivity differential between tradable and non-tradable goods production and is measured in terms of labour productivity in both sectors; [share.sub.i,t] represents the share of non-tradable consumption in total private consumption; [govreal.sub.i,t] is the share of government consumption in GDP measured in constant prices; and [lab.sub.i,t] represents the structural misalignment variable. It is proxied for by the ratio between labour employed in the tradable sector versus labour employed in the non-tradable sector. All coefficients have a negative sign except for the structural variable, which enters the equation with a positive sign. This constitutes the positive correlation between the real exchange rate and the labour employed in the tradable sector relative to the non-tradable sector. For this reason, the structural variable proxied for by the labour ratio represents the parameter that measures the rigidity of the labour market to structural changes in the economy. (7)

Data used to construct price indices, productivity measures, demand variables, and structural parameters cover 19 transition economies. (8) Each transition economy is observed from the beginning of its most serious stabilisation attempt as defined by Fisher et al. (1996). This implies that the relative price of tradables in terms of non-tradables is set to 1 in the year of the most serious stabilization attempt. The implicit GDP deflator for industry in each country represents the price of tradables. Analogously, the implicit GDP deflator for services defines the price of non-tradables. The criterion for the period of observation was the year after which the relative price of tradables in terms of non-tradables started to decline consistently. However, this criterion has not been followed in all cases. (9) Different periods of observation were examined and compared to each other. For all countries, the period of observation ends in 1998. The longest series runs from 1990 to 1998, while the shortest covers the period from 1995 to 1998. The entire sample includes 122 observations. (10)

The independent variable is the relative price of tradables in terms of the price of non-tradable goods. The implicit sectoral GDP deflators for industry and services are used to proxy for the price indices in these two sectors. The relative price takes value 1 at the beginning of transition and enters the regressions in logarithms.

Regression equation (2) reproduces the results for the full sample of 19 economies, each observed in time since the beginning of the transition process. (11)

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The results are fully consistent with the view that structural reforms in transition economies contributed to the real appreciation trend observed in the region since the beginning of transition. Since all regressions (12) are run in transition time, the results indicate that we can still expect further appreciation of the real exchange rate in those economies that began transition at a later time. As indicated in regression equation (2), the productivity differential used to measure the Balassa-Samuelson effect had a pronounced effect on the appreciation of the real exchange rate in transition economies in the period prior to 1999. A 1 percent increase in productivity differential has on average contributed to almost a 0.9 percent appreciation of the real exchange rate measured in terms of relative prices. This result is in line according to the Balassa-Samuelson effect, which states that prices of tradables are determined in the world market and therefore equalised across countries. Prices of non-tradables are assumed to be determined domestically based on the domestic wage and productivity levels. To the extent that productivity in the two sectors within the country grows at different rates, it is likely that there will be offsetting movements in the relative price of tradables in terms of non-tradables. If the growth of productivity trend in the tradable goods sector exceeds that of the non-tradable goods sector, there will be a tendency for the relative price of tradables to decline over time.

Once we take into account the initial labour reallocation as the main source of an increase in productivity differential in the first few years of transition, the estimates of the Balassa-Samuelson effect following equation (2) are rather lower than in comparable studies (see Halpern and Wyplosz, 1996; De Broek and Slok, 2001; and Deutsche Bundesbank, 2001). If we take into account only the period after the labour reallocation has been completed, the estimate of the Balassa-Samuelson effect should increase for CEE countries. Coricelli and Jazbec (2002) show that in the case of Slovenia, a 1 percent increase in productivity differential translates into about 1.5-1.7 percent real appreciation, considering only the period after initial labour reallocation. This figure is almost twice as high as the estimate on productivity differential effect in equation (2). As Slovenia is the most developed transition economy with respect to GDP per capita, then the Balassa-Samuelson effect in other transition economies should be even higher. A visual inspection of Figure A2 where labour reallocation and productivity differential are shown could provide a tentative conclusion that in Hungary, Poland, and the Czech Republic, the Balassa-Samuelson effect is even stronger than in Slovenia. One should additionally take into account the approach to measure the Balassa-Samuelson effect proposed by MacDonald and Ricci (2001). They believe that independent of the productivity differential between productivity in manufacturing and services, the size and efficiency of the distribution sector of the economy could substantially contribute to the real exchange rate appreciation. As the size and efficiency of the distribution sector in transition economies grew since the beginning of transition, this channel to the real exchange rate appreciation should be taken into account when explaining the consequences of higher growth in transition economies. For those reasons, one could conclude that the working of the Balassa-Samuelson effect would nevertheless imply higher inflation rates even after accession to the European Union. Thus, transition economies are likely to face problems in meeting the Maastricht criterion on inflation, unless growth is proportionally suppressed to meet the criterion of 1.5 percentage points above the best three EU inflation performers. Again, this implies that after accession, inflation differential will remain a serious problem, irrespective of the exchange rate regime.

EXCHANGE RATE PASS-THROUGH

With respect to exchange rate regimes, most CEE countries moved to more flexible exchange rate regimes during transition. The main reason for the move was believed to be from pressure caused by a surge in foreign capital inflows (Corker et al., 2000). In so doing, CEE countries added a potential new source to higher inflation rates in addition to the working of the Balassa-Samuelson effect. Namely, the move from a fixed to a more flexible exchange rate regime could backfire attempts to lower inflation as the exchange rate pass-through could add to inflationary pressure instead of suppressing it. The evidence on selected transition economies could partially support this kind of argument, although the extent of the pass-through cannot be firmly established (Darvas, 2001; Campa and Goldberg, 2002). Taking into account caution in explaining econometric results, Darvas (2001) finds short-run estimates of pass-through of nominal exchange rate to fundamental prices (food, energy, and administered prices were excluded from CPI) in 2000 higher in Hungary and Slovenia than in Poland and the Czech Republic. He tentatively concludes that part of the difference in the pass-through estimates could be attributed to the exchange rate regime, as Hungary and Slovenia had a managed exchange rate regime opposed to Poland and the Czech Republic, which had a floating regime in 2000. Although Darvas (2001) takes into account the change of the exchange rate regime in Hungary, the Czech Republic, and Poland during the transition process, the main concern explaining results for pass-through in transition economies is still the shortness of time series for the exchange rate and prices if one seriously considers the importance of the initial period of the labour reallocation process as explained above. Interestingly, the timing of the change of exchange rate regimes in Hungary, the Czech Republic, and Poland vaguely corresponds to the periods when the process of structural reforms proxied by labour reallocation settled down.

More thorough analysis of the exchange rate pass-through in four acceding countries (Czech Republic, Hungary, Poland, and Slovenia) is provided in Coricelli et al. (2003), who approach the empirical analysis within the framework of cointegrated VAR model. The paper finds a strong pass-through from nominal exchange rates to domestic inflation. In such a context, the dichotomy between inflation targeting and exchange rate targeting is more apparent than real. Moreover, in many instances, flexibility of exchange rates turns out to be a policy of accommodation of inefficiencies and monopoly power in non-tradable sectors.

Although the existence of the Balassa-Samuelson effect and potential exchange rate pass-through could provide an explanation for the real exchange dynamics in CEE countries on average, it is in the Baltics where both effects had a rather modest occurrence. On the one hand, all Baltic countries have currency boards, which offsets the exchange rate pass-through, while on the other, it seems that the increase in productivity differential was rather small after the initial labour reallocation. For these reasons, real exchange rate appreciation in the Baltics could mostly be attributed to demand factors. The dynamics of relative wages in Latvia shown in Figure A1 could provide justification for this kind of reasoning. Also, wages in the public sector have been increasing more in the Baltics than in other transition economies, with the exception of Slovenia and Romania.

As real appreciation in transition economies resulted in higher inflationary pressure rather than nominal appreciation, part of the inflationary pressure could derive from labour market rigidities as pointed out in the discussion of Figure 2 above. For this reason, it is not surprising that countries with a higher relative non-tradable wage growth--either growth of wages in market or public services--on average face higher inflation rates. This brings up the issue of the relationship between exchange rate policy and disinflation in an economy with price-wage and inflation inertia in the non-tradable sectors.

Exchange rate policy and disinflation

The discussion in previous sections was generally related to real models. However, the analysis of the relationship between exchange rate policy and inflation requires consideration of the monetary sector. Candidate countries face the challenging decision on the speed of convergence to the euro zone rate on inflation. Before entering the ERM2, the Maastricht criterion on inflation will not apply. However, a fundamental policy question is whether candidate countries should aim for rapid convergence and whether this will hurt their growth performance. We already stated that the Balassa-Samuelson effect is going to be relevant and that Maastricht criterion on inflation should be revised to take into account the equilibrium phenomenon of increasing prices of non-tradable goods. However, inflation performance in candidate countries is not only related to the Balassa-Samuelson effect but also to the policy stance of the different countries. Exchange rate policy is one of the key aspects of this policy stance. A useful reference framework for discussing the costs and benefits of different speeds of disinflation is a two-sector model with monopolistic power in the non-tradable sector. In the context of perfect capital mobility, interest rates in candidate countries would be determined by foreign interest rates and expected depreciation of the exchange rate. In the staggered price model of Calvo (1983) with price level inertia in the non-tradable sector, it is easy to show that by reducing the rate of depreciation of the exchange rate, a country can reduce the overall rate of inflation inducing a temporary fall in output in the non-tradable sector, as in standard new-Keynesian models. A more interesting model is a recent extension of staggered price models by Calvo et al. (2002) that takes into account the average rate of inflation for the price setting of firms in a monopolistically competitive market. The intuition of the model is that firms choose a price rule that includes a revision of price schedule depending on the rate of inflation in the economy. This implies that firms internalise the effects of policies such as a policy where the central bank targets the real rate though persistent rate of depreciation of the nominal rate. In this version of the model, there is inflation inertia in addition to price-level inertia. The implication is that a disinflation policy implemented through a reduction of the rate of depreciation of the exchange rate induces, as in the previous model, a temporary decline in output in the non-tradable sector. However, in this model, disinflation brings welfare gains as it reduces the welfare losses associated with monopolistic power in the non-tradable sector. A disinflation policy can thus be seen as a way of reducing the welfare losses of monopolistic price setting. This line of reasoning seems very relevant for an exchange rate policy in candidate countries.

CONCLUSIONS

In the paper, we analysed the behaviour of the real exchange rate in candidate transition countries. We examined both the relevance of the Balassa-Samuelson effect and of monopolistic price setting in non-tradable sectors. We argued that while significant, the Balassa-Samuelson effect cannot be an argument for flexibility of the exchange rate, as it is a long-term and fully anticipated phenomenon. Exposing countries to the high-frequency oscillations of the exchange rate determined in the short term by capital flows seems hard to justify. Regarding the view that exchange rate flexibility is a substitute for rigidities in the goods and labour market, the paper concludes that an accommodating exchange rate policy can protect output in the non-tradable sector in the short run, but at high welfare costs. For the reasons stated, the best policy for acceding countries would be the adoption of the euro as early as possible.

APPENDIX A: FIGURES A1 AND A2

[ILLUSTRATION OMITTED]

(1) See Coricelli and Jazbec (2001) for a full derivation of the model of real exchange rate determination in transition economies.

(2) The non-tradable sector employs fewer workers relative to the tradable goods sector. If state-owned firms in both sectors of the planned economy still follow some kind of profit maximisation objective, then they would employ labour up, to a point where the marginal product of labour equals the real wage paid in that sector. Any other employment decision would not substantially alter the conclusions. It follows that the higher the preferences of central planners for tradable goods production are, the lower the nominal wage attained in that sector.

(3) It should be mentioned that not all transition economies began the transition process at the same time. According to Fisher et al. (1996), the transition process begins with the year of the most serious stabilisation attempt. In this respect, the transition process, which in most cases started with price liberalisation, began in 1990 for Poland and Hungary; in 1991 for Bulgaria, Czech Republic and Slovak Republic; in 1992 for the Baltics and Slovenia; and in 1993 for Romania.

(4) For an insightful analysis of the case of Slovenia, see Bole (2002).

(5) Exceptions are Romania--where the relative price of tradables has indeed increased--and Estonia--where the relative price of tradables started to decline already in 1990.

(6) A detailed description of the data is presented in Coricelli and Jazbec (2001).

(7) As for the rest of the story, this rigidity is assumed to be exogenously determined in the economy and thus independent of all other right-hand side variables in equation (1). This is a relatively stringent assumption on the structure of a transition economy, and its validity can be seriously questioned.

(8) Armenia, Azerbaijan, Belarus, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Poland, Romania, Russia, Slovak Republic, Slovenia, Ukraine, and Uzbekistan.

(9) Exceptions are Belarus, Romania, and Russia where the relative price of tradables has indeed increased. For these cases, the beginning of the observed period starts after the initial depreciation.

(10) While theoretical literature on real exchange rates relies upon the division of commodities into tradables and non-tradables, it is almost impossible to construct these two groups of commodities in reality. An obvious benchmark for tradability should be the extent to which the particular good is actually traded. For example, the sector is defined as tradable if more than 10 percent of total production is exported. In general, one would label manufactures as tradables and services as non-tradables. However, this is quite impossible at this stage in transition economies. In what follows, the tradable sector is represented by the industry sector, which includes manufacturing; gas, electricity, and water; mining and quarrying; and construction. The reason that all other sub-sectors besides manufacturing were included in the measure for the tradable sector was that for some countries, sectoral data and data on international trade flows were not available. To ensure consistency, all tradable sectors in different countries include gas, electricity, water, mining and quarrying, and the construction sector, although one could doubt their tradability. A more substantial problem arises from the inclusion of non-market services into the variable representing the non-tradable sector. However, the reasons for the inclusion of non-market services into the total services sector are the same as for the construction of the tradable sector variable.

(11) Coefficient estimates are reported with standard errors adjusted for heteroscedasticity in parenthesis. Superscripts indicate their possible insignificance at a 5 percent level of confidence. Country-specific dummies (not reported) are significant in most of the specifications.

(12) Several regressions were run by adding region-specific dummies to distinguish possible effects across transition economies included in the sample. The results confirm those presented by equation (2). For the full description of econometric results, see Coricelli and Jazbec (2001).

REFERENCES

Aghion, P and Blanchard, OJ. 1993: On the speed of transition in central Europe. EBRD working paper no. 6, July.

Bole, V. 2002: Disinflation and labur market distortions: Lessons from Slovenia. EIPF working paper no. 2.

Buiter, WH and Grafe, C. 2002: Anchor, float or abandon ship: exchange rate regimes for accession countries. CEPR discussion paper no. 3184, London, January.

Calvo, G. 1983: Staggered prices in a utility-maximizing framework. Journal of Monetary Economics 12: 383-398.

Calvo, G, Celasun, O and Kumhof, M. 2002: Nominal exchange rate anchoring under inflation inertia. IMF working paper, no. 30, February.

Campa, MJ and Goldberg, LS. 2002: Exchange rate pass-through into import prices: a macro or micro phenomenon? NBER working paper series no. 8934, May.

Chadha, B and Coricelli, F. 1997: Fiscal constraints and the speed of transition. Journal of Development Economics 52: 221-249.

Coricelli, F and Jazbec, B. 2001: Real exchange rate dynamics in transition economies. CEPR discussion paper no. 2869, London, July.

Coricelli, F and Jazbec, B. 2002: Accession to the European union: real exchange rate dynamics for candidate countries. Rivista di Politica Economica 3rd Series: 109-138.

Coricelli, F, Jazbec, B and Masten, I. 2003: Exchange rate pass-through in candidate countries. CEPR discussion paper no. 3894, London, May.

Corker, R, Beaumont, C, van Elkan, R and Iakova, D. 2000: Exchange rate regimes in selected advanced transition economies: Coping with transition, capital inflows, and EU accession. IMF policy discussion paper no. 3, April.

Darvas, Z. 2001: Exchange rate pass-through and real exchange rate in EU candidate countries. Deutsche Bundesbank discussion paper no. 10, May.

De Broek, M and Slok, T. 2001: Interpreting real exchange rate movements in transition economies. IMF working paper no. 56, May.

Deutsche Bundesbank, 2001: Monetary aspects of the enlargements of the ED. in Monthly Report, October pp. 15-30.

Egert, B. 2002: Investigating the Balassa-Samuelson hypothesis in the transition: Do we understand what we see? A panel study. Economics of Transition 10: 273-309.

Fisher, S, Sahay, R and Vegh, CA. 1996: Stabilization and growth in transition economies: The early experience. Journal of Economic Perspectives 10: 45-66.

Halpern, L and Wyplosz, C. 1996: Equilibrium exchange rates in transition economies. IMF working paper no. 125.

Krajnyak, K and Zettelmeyer, J. 1997: Competitiveness in transition economies: What scope for real appreciation? IMF working paper no. 147, November.

MacDonald, R and Ricci, L. 2001: PPP and the Balassa-Samuelson effect: The role of the distribution sector. IMF working paper no. 38, March.

Melitz, J and Waysand, C. 1996: The role of government aid to firms during the transition to a market economy: Russia 1992-1994. The Journal of Policy Reform 1: 299-534.

Nuti, M. 2000: The costs and benefits of Euro-isation in central-eastern Europe before or instead of EMU membership. University of Rome and London Business School, mimeo.

Richards, AJ and Tersman, GHR. 1996: Growth, nontradables, and price convergence in the Baltics. Journal of Comparative Economics 23: 121-145.

Roubini, N and Wachtel, P. 1999: Current account sustainability in transition economies, in Blejer, M and Skreb, M (eds). Balance of Payments, Exchange Rates and Competitiveness in Transition Economies. Kluwer: Dordrecht.

FABRIZIO CORICELLI (1,2) & BOSTJAN JAZBEC (3,4)

(1) Department of Economics, University of Siena, Central European University, Budapest, Hungary,

(2) CEPR, London;

(3) Faculty of Economics, University of Ljubljana, Slovenia;

(4) Bank of Slovenia, Slovenia