The impacts of the Euro on the real exchange rate and the growth in CFA zone countries

Transcription

1 DOCUMENT DE TRAVAIL N The impacts of the Euro on the real exchange rate and the growth in CFA zone countries Issiaka Coulibaly 1

2 The impacts of the Euro on the real exchange rate and the growth in CFA zone countries Issiaka Coulibaly Abstract: The aim of this article is to estimate the effects of the exchange rate policy on the competitiveness and the growth rate of the CFA zone economies. More particularly, we focus on the potential impact exerted by the move from the peg to the French Franc towards the peg to the Euro. We show that the appreciation of the euro in the last decade has been the main factor of the real appreciation of the CFA franc. However, this latter did not translate in a strong real overvaluation except in Central African Republic, Côte d Ivoire and Togo indicating the improvement of the macroeconomic fundamentals in the rest of the CFA zone countries. We also provide strong evidence of damaged effects of the nominal appreciation and the exchange rate misalignment on the growth in the CFA zone countries. However, the advent of euro didn t increase these damaged effects even if the monetary policy of the European Central Bank (ECB) is clearly different to that applied by the Bank of France. Key words: Growth, Exchange rate policy, CFA zone. JEL codes: O40, F01, C23. I m grateful to Cécile Couharde and Pierre Blanchard for thorough and constructive comments on earlier versions of this paper. ERUDITE, Université Paris-Est Créteil, Créteil 94010, France. issiaka1.coulibaly@etu.u-pec.fr 2

3 Introduction The CFA zone is an economic and monetary area consisting of two monetary unions: WAEMU 1 (West African Economic and Monetary Union) and CAEMC (Central Africa Economic and Monetary Community), plus the Comoros. The countries of this zone are mostly former French colonies which, after their independence, signed monetary cooperation agreements with France involving a guarantee of their currency CFA franc convertibility and the currency peg between the CFA franc and the French franc (the euro since 1999). In recent years, the CFA franc system (peg to the euro and the deposit of 65% at least of their foreign reserves in the French treasury s accounts called Operation Accounts ) has became a recurrent question in the CFA zone countries. In fact, some authors consider the peg to the Euro as the main source of recent difficulties faced by these countries, especially the growth instability they have been experiencing (e.g. Kuikeu, 2011). The creation of the European Monetary Union (EMU) in 1999 and the shift of the anchor currency of the CFA franc (the French Franc to the Euro) was an important transition with considerable issues for the CFA zone countries. It has also revived old questions about the future of those countries. The potential economic effect of the advent of the Euro on the CFA zone countries was first analyzed by Hadjimichael and Galy (1997). They investigate the likely ex-ante economic impacts of the EMU on the real effective exchange rate of the CFA franc and on the external competitiveness of the CFA zone economies. Indeed, the monetary and exchange rate policies of the anchor country influence both nominal and real effective 2 exchange rates of a pegged currency and therefore the pegged country s competitiveness and economic performance. For instance, Bolle (1997) emphasizes that, in the 1970s, the CFA zone gained competitiveness because their currency was pegged to a depreciating French Franc. On the contrary, the appreciation of the French currency (from mid 1980s to 1993), 1 The WAEMU unites Benin, Burkina Faso, Côte d Ivoire, Guinea Bissau, Mali, Niger, Senegal, and Togo and the CAEMC includes Cameroon, Gabon, Equatorial Guinea, Congo, Central African Republic, and Chad. The name CFA Franc means: Communauté Financière Africaine for the WAEMU member countries and Coopération Financière en Afrique Centrale for CAEMC member countries. 2 The effective or multilateral exchange rate of a country measures the evolution of its currency against all its main trading partners unlike the bilateral exchange rate which describes the evolution of the currency of one country against another. The real effective exchange rate (REER) corresponds to the Nominal Effective Exchange Rate (NEER) deflated by relative prices and represents the main global indicator of a country s competitiveness. Cottani et al. (1990), suggest that the REER is the main transition mechanism of economic policy on its performance. The interest of this variable is also mentioned by Eichengreen (2008) for which a competitive REER is at the heart of the Chinese authorities development strategy. 3

4 combined with inadequate internal economic policies, had dramatic effects leading to the devaluation of the CFA franc in Klau (1998) claims that the poor economic performance of the CFA zone observed during this period was caused by CFA overvaluation. The euro s peg to the CFA franc constitutes an exchange rate arrangement with EMU countries and was expected to have several positive effects on the CFA zone economies. The creation of the EMU was supposed to make the Euro a stable currency, to lead to low inflation and interest rates, and stronger-than-otherwise output growth in EMU countries. Hadjimichael and Galy (1997) state that the CFA countries should benefit from the stability of the euro with the reduction of uncertainty in exchange rate variability, in relative prices changes, in exports and thus in government revenues fluctuations expressed in CFA franc. Especially, the authors predicted that the CFA zone countries should register higher growth rates due to hypothetically strong future growth rates in the EMU countries because of the important trade between the two zones. Did the creation of the EMU improve the economic performance of CFA countries? As mentioned by Hadjimichael and Galy (1997), the main risk for CFA zone countries of the advent of the euro is the loss of external competitiveness that may be caused by a strong euro. A few years after its creation, the euro has continuously appreciated against the US dollar (38% between 2000 and 2008), bringing in its wake an appreciation of real exchange rates in CFA zone countries and thus a loss of their competitiveness. Indeed, a persistent real appreciation of a currency can generate a deviation to its equilibrium value and leads to a misalignment 3. This in turn can have negative impacts on economic growth through the loss of competitiveness. Particularly, the overvaluation of a currency diminishes the competitiveness of a country and then reduces its exports, production and investment, and increases its imports. Consequently, the current account is damaged and the GDP and the employment are deteriorated. This research is motivated by recent relatively low and unstable performance of CFA zone countries in terms of economic growth (see figure A and B in the appendix). More particularly, we are interested in the relationship between the real exchange rate and economic growth, and focus on the role of the advent of the euro. To this end, we compare the euro peg period and the French Franc peg period to highlight the effects of the anchor money shift on 3 The misalignment is defined as the gap between the observed real exchange rate and its equilibrium value. It determines whether the exchange rate of a country is under or overvalued. 4

5 the CFA economies 4. More than ten years after the creation of the EMU, it seems essential to determine whether the euro has exerted positive impacts on the CFA zone economies as expected and to analyze the differenced effects of the exchange rate policy between the two anchor periods. This issue, essential for the design of economic policies in the CFA zone countries, has not yet been studied to our knowledge. This research also could contribute to the ongoing and exciting debate on the status of the CFA franc, particularly for the WAEMU countries which are planning to form a new monetary union with others Economic Community of West African States (ECOWAS) countries. Empirically, we estimate an augmented growth equation by adding to traditional explanatory variables (investment, human capital, etc ), the currencies misalignment or the nominal effective exchange rate in order to capture the potential effects of the exchange rate policy. Since the exchange rate policy or more broadly the monetary policy in the CFA zone is based on the fixity of the CFA franc-euro parity, the NEER reflects the impact of the anchor currency variation vis-à-vis of others currencies. We also attempt to estimate the differenced potential effect of the anchor currencies (French Franc and Euro) on growth performance in the CFA zone. To do this, we estimate a dynamic model of growth on a small panel data by applying appropriate estimation methods. Finally, we test two alternative measures of the effects of the exchange rate policy (or of the euro effects) that are: Nominal Effective Exchange Rate (NEER) and the Real Effective Exchange Rate (REER) of the euro. Several results emerge from our study. Firstly, the consecutive appreciation of the euro in the last decade has been the main factor leading to the real appreciation in the CFA zone countries. However, this real appreciation has not lead to significant real overvaluation except in the Central African Republic, Côte d'ivoire and Togo to a lesser extent, indicating the improvement of the macroeconomic fundamentals in the rest of the CFA zone countries. Secondly, the advent of the euro did not cause a significant disruption in the relationship between growth and the exchange rate policy indicators in the CFA zone economies. The effects of the exchange rate policy remain the same for the two pegs periods (peg to the French franc and to the euro). This result holds whatever the variable used as proxy of the exchange rate policy: the NEER, currencies misalignments or Euro variables. Thus, the creation of the EMU with a more restrictive monetary policy did not significantly increase the 4 It should be noted that the European Central Bank (ECB) which is in charge of the management of the euro, applies a more restrictive monetary policy than the Bank of France. The ECB was founded on the model of the German central bank Bundesbank known for its liberal orthodoxy. 5

6 damaged effects of the anchor currency appreciation. However, taking into account the devaluation effects on the CFA economies growth rates seems to slightly mitigate the results. Finally, our results show that the more the countries are open, the more damaged are the effects of the nominal appreciation. The rest of this paper is organized as follows. The second section describes the relationship between the nominal and real effective exchange rates, the currency misalignments and the growth. Section 3 explains the methodology, describes the used data and analyzes the results of our estimations. Section 4 concludes the article. 2. NEER, REER, Currency misalignments and Growth 2.1. Literature review One of the first main empirical proofs of the adverse impact of the real exchange rate on the economic performance in the less developed countries (LDCs) has been developed by Cottani et al. (1990). These authors argue that economic performance of Latin American, East Asian and African countries are mainly explained by their exchange rates stability and their currencies misalignments. Indeed, even if they faced the same external conditions (terms of trade shocks, etc.), Asian developing countries have registered higher growth rates compared to African and Latin American countries. This discrepancy between the zones is explained by inappropriate domestic macroeconomic and trade policies that all are reflected in the evolution of real exchange rates. Consequently, the real exchange rate s instability and currencies misalignments are often considered as the main determinants of the growth rate in LDCs. Empirical studies, generally, found a negative relation between misalignment and growth in both developing and developed countries. Cottani et al. (1990) found a strong negative correlation between the economic growth and the RER instability and misalignment for a sample of developing countries. They show that persistent misalignments hamper the development of agriculture and reduce domestic food supply in African economies. Considering a large sample of developing, Toulaboe (2006) also finds that higher misalignments are associated to lower per capita growth. Gala and Lucinda (2006); Razin and Collins (1999) and Aguirre and Calderon (2005) obtain the same result although they include 6

7 some developed countries in their study. On the whole, empirical literature highlights that high overvaluations tend to harm growth by inducing some instability while moderate undervaluation enhances the growth rates. 5 Concerning Sub-Saharan African countries, an interesting contribution was provided by Ghura and Grennes (1991). They show a robust negative correlation between economic performance and REER misalignment on a sample of 33 SSA countries. They conclude that countries which have pursued more predictable macroeconomic policies and have registered lower levels of REER misalignment have experienced better economic performance. Klau (1998) and Elbadawi et al (2009) also obtain similar results for CFA zone and SSA countries respectively. Unlike these studies, we also consider the evolution of the nominal effective exchange rate to analyze the Growth-Exchange rate nexus. This is particularly important for the CFA zone countries which are in fixed exchange regime since their independences. Thus, the NEER would be an interesting proxy of the anchor currency effect on the pegged country. The determination of this effect becomes increasingly important since the creation of the EMU added to the continuous appreciation of the euro in the last decade Stylized Facts: REER, NEER, and Statistics The evolution of the real exchange rate depends on three main factors: the relative prices, the nominal exchange rate and the economic performance 6. For the CFA franc zone countries, it is important to note that the evolution of the nominal effective exchange rate is entirely related to that of the euro (the French franc before 1999) against the currencies of the CFA zone main trading partners. Thus, this variable can measure the cost and the benefit of the fixed exchange rate for CFA zone economies. Consequently, in this section, we focus the evolution of the REER and NEER of the two monetary unions of the CFA zone and provide some statistics about main variables of our study. We calculate nominal and real effective exchange rates (expressed as an index base 100 in 2000) 7 of the CFA zone countries over the period Nominal effective exchange 5 For recent development in that field, see Aguirre and Calderon (2005) or Béreau et al. (2009). These authors study a non-linearity in the relationship between misalignment and growth. 6 It depends on the economic performance when the exchange rate regime is de facto flexible and capital flows are not restricted 7 For the WAEMU and the CAEMC, the REER and NEER are calculated as a weighted means of their membership countries. The unweighted means data are available on request and don t affect significantly the 7

8 rates are calculated using bilateral exchange rates of the top ten trading partners of each country, weighted by their share in the external trade of the country over the period The weights used are from Couharde et al. (2011). Real effective exchange rates are the nominal corrected by the weighted relative prices. Then, the REER and NEER for the WAEMU and the CAEMC are calculated as a weighted means of their membership countries. 8 Figure 1: Evolution of the REER and the NEER, CAEMC countries Source: World Bank (WDI), IMF (DOTS) and author calculations. Figure 1 shows that the appreciation of real effective exchange rate in the CAEMC zone has been higher than the appreciation of the nominal effective exchange rate until the late 1980s. This implies that, the real appreciation of the exchange rate in the CAEMC countries was caused mainly by their higher inflation compared to their partners; the rest was due to the nominal appreciation. This trend was reversed between 1990 and 1993 when the real effective exchange rate depreciated while the nominal still appreciated. So, during this period, the CEMAC countries benefitted from competitiveness gains, compared to their partners, which were mitigated by the nominal appreciation. Since the CFA franc devaluation in 1994, the dynamics of the real exchange rate has been in line with the evolution of the nominal evolution of the variables. Guinea Bissau is not taken into account insofar as it is a member of WAEMU since Guinea Bissau is not taken into account insofar as it is a member of WAEMU since The unweighted means data are available on request and don t affect significantly the evolution of the variables. 8

9 exchange rate. Thus, real appreciation has been mainly caused by the nominal appreciation as relative prices have not changed during this period. Figure 2: Evolution of the REER and the NEER of WAEMU countries Source: World Bank (WDI), IMF (DOTS) and author calculations. The WAEMU s real exchange rate also appreciated strongly before the devaluation of the Franc CFA. From the devaluation to 2002, the evolution of the real effective exchange rate of the WAEMU has been in line with the nominal s one, suggesting that changes in the REER were caused by NEER evolutions. From 2002 to 2008, the nominal exchange rate has more appreciated than the real, showing the recent euro appreciation effects and the fairly good performance of these countries in terms of inflation. Table 1 provides a comparison of interest variables between the two peg periods. It shows that the growth rate in the CFA zone has been disparate. Less than half of the sample s countries have registered a positive average growth rate between denoting the weakness and instability of the latter. However, the CFA zone countries recorded positive and greater growth rate in the euro peg period except Côte d Ivoire, Togo, Gabon and Benin. In fact, the nominal appreciation was important in Côte d Ivoire, Togo and Central African Republic leading to the highest overvaluation between 1999 and The situation of Benin is slightly different since its growth rate was on average positive during the two periods. Concerning Gabon, the country experienced a severe economic crisis in 1999 and recovered only partially 9

10 in 2000 and 2001 thanks to the dynamic of the non-oil sector and a decline in oil production slower than expected. Globally, the exchange rate misalignments of the euro peg period had been lower than those of the French Franc peg period for most countries. Table 1: Evolutions of main variables (mean) Countries Growth Variation of TCEN Misalignment a 1998 All All Openness % Investment to GDP (%) Life Population Benin 0,6 0,3 1,0 1,4 1,6 1,1-0,03 0,038 44,4 17,1 56,7 3,2 Burkina Faso 2,0 2,0 2,0 0,3-0,2 0,9 0,05-0,072 35,4 19,2 49,3 2,9 Central African, -1,1-1,3-0,8-1,4-2,7 0,2-0,12 0,165 39,7 10,8 47,7 2,2 Rep. Côte d Ivoire -1,4-0,9-2,0 1,5 0,7 2,5-0,08 0,106 73,6 10,5 56,7 3,0 Cameroon -1,0-2,6 1,3 0,3-0,5 1,4 0,04-0,051 41,5 18,5 52,8 2,6 Congo, Rep. -0,6-2,0 1,4 0,6 0,1 1,2 0,05-0, ,3 26,6 56,1 2,4 Gabon -0,9-0,6-1,4-0,4-1,5 1,0 0,08-0,114 92,0 27,4 60,3 2,6 Equatorial Guinea 13,5 8,8 19,8-0,2-0,9 0,9 0,00-0, ,9 52,3 47,9 3,4 Mali 1,8 0,8 3,2-0,2-0,6 0,3 0,05-0,066 59,9 19,3 44,8 2,5 Niger -0,1-0,2 0,0 1,7 1,2 2,5 0,05-0,068 41,0 14,1 45,0 3,3 Senegal 0,6-0,2 1,7 1,1 0,7 1,6 0,04-0,059 62,3 18,8 53,3 2,7 Chad 2,6 1,0 4,9-0,2-1,8 1,8 0,05-0,074 64,1 20,5 49,9 3,2 Togo -0,3-0,1-0,7 1,8 1,7 2,0-0,04 0,053 83,2 14,6 59,2 2,9 Note : a By construction the mean of the misalignment in all simple is equal to zero. For details see the section 3.3. Source: World Bank (WDI), IMF (IFS), WEO and author calculations. Growth The CFA zone countries exhibit similar features in terms of life expectancy (low) and population growth (high) indicating widespread poverty in the region. The oil-exporting countries and those who improved their economic situations in recent years, devote a larger share of their income to investment (Equatorial Guinea, Senegal, and Mali, among others). Note also that the most open economies are generally coastal countries and/or oil exporters. This preliminary analysis shows that changes in the real effective exchange rate were more linked to those of the nominal effective exchange rate since the devaluation. This finding suggests the growing importance of the exchange rate policy or more generally the monetary policy of the anchor currency for these countries. A thorough analysis seems necessary in order to specify the effects of exchange rate policy on growth. 10

11 3. Methodology 3.1. Augmented growth equation Empirical studies provide several potential determinants of growth referring to both the neoclassical and the endogenous theory of growth. 9 Sala-I Martin (1997) founds 60 potential variables that could explain the growth performance of a country. These variables can be gathered in several groups: regional variables, political variables, religious variables, economic variables (human capital, type of investment, share of primary sector, inflation, initial level of GDP, government consumption, financial development ) and market distortions (exchange rate misalignment, etc ). On the whole, there is no consensus on the key explanatory variables of growth rates. Thus, on the basis of previous studies, especially on the SSA or CFA countries 10, we consider the following growth equation: (1) With as the logarithm of the real GDP. is our interest variable corresponding to the nominal effective exchange rate or to the real exchange rate misalignment. represents the convergence term of traditional neoclassical models. It supposes that richer countries tend to grow more slowly than poorer countries, so must be negative and significant to confirm the convergence hypothesis between CFA zone countries. is a set of standard control variables that are robustly associated with growth performance. We separate these variables into two groups. On the one hand, we consider investment rate which represents the physical capital, the human capital and the population growth. These variables are considered as the principal control variables (see Levine and Renelt, 1992) and are maintained in almost all regression. Due to the data limitation, we use the life expectancy at birth as a proxy of human capital as proposed by Ghura and Hadjimichael (1994) and Sala-I Martin (1997), even if some studies use the gross secondary school enrollment. On the other hand, we take into account other potential determinants of economic growth related to the 9 For more detail on the growth framework, and the various theoretical models that underlie it, see Barro and Sala-i-Martin (1995) or Dramani (2010) about CFA zone countries. 10 See among others: Ghura and Grennes (1991); Klau (1998); Esterly and Levine (1997); Sachs and Warner (1997); Elbadawi et al. (2009); Diop et al. (2010); Dramani (2010). 11

12 macroeconomic policy: the inflation rate to capture the effect of the monetary policy; the government expenditure as an indicator of fiscal policy and the degree of openness which reflects the effect of the globalization (see Sachs and Warner, 1997; Dollar and Kraay 2004; Dufrenot et al., 2009). With this last indicator, we examine whether the damaged effects of NEER appreciation increase when the economies are most open. We may expect that the most open economies of the CFA zone (Congo, Equatorial Guinea, Côte d Ivoire, Gabon and Togo) will be more impacted by the nominal appreciation. To simplify the estimation, we rewrite equation (1) as follow: (2) In practice, the estimation of growth equations has evolved in conjunction with the improvement of econometric techniques. First empirical studies used time series (single country) and cross country estimations to analyze economic growth. Thereafter, panel data and the Least Squares Dummy Variable (LSDV or fixed effect or within) estimator have been used in order to take into account the joint occurrence of dynamics and unobserved individual heterogeneity (see Islam, 1995 for instance). This method was criticized by Kevin et al. (1998) because of the inconsistence of the fixed effect estimator in the context of dynamic panel-data models as growth equations. In fact, since Nickell s (1981) paper, it is clear that the fixed effect estimator is not consistent in dynamic panel-data models. This has led to a development of consistent estimators: instrumental variable (IV) proposed by Anderson and Hsiao (1982) and generalized method of moments (GMM) proposed by Arellano and Bond (1991) and Blundell and Bond (1998). These estimators have been used almost systematically in recent studies on linear 11 dynamic panel-data models. However, the IV and GMM estimators must be used with caution since they can be severely biased and imprecise in panel data with a small number of individual dimension, which is the case with our sample s countries (see Kiviet, 1995; Bruno, 2004; Buddelmeyer et al. 2008; etc.). Using Monte Carlo experiments, these authors show that IV-GMM estimators are the most biased and the less efficient estimators when the individual dimension of the panel is small. According to Kiviet 11 Concerning the non linear models in panel, we can mention the Gonzalez et al. (2005) Panel Smooth Transition Regression (PSTR) models used by Bereau and al. (2009); Dufrénot et al. (2009) s quantile regressions or the Hansen (1999) Panel Transition Regression (PTR) models used by Combey and Nubukpo (2010). 12

13 (1995) and others, they are outperformed by the corrected LSDV estimator (LSDVC) in term of bias and root mean squared error (RMSE). Consequently, as our panel in this study has a relatively small individual dimension (13 CFA countries), we use the LSDVC estimator. This estimator is obtained by subtracting an estimated value of a bias to the fixed effect coefficients. Bun and Kiviet (2003) provide three types of bias that depend on the true values of the coefficients. Since the latter are unknown, an IV-GMM estimator could be chosen as an approximation of the true values of the coefficient (see quoted papers for more details). However, for robustness and comparison, we also use Arellano and Bond (1991) 12 estimator for estimating our growth equation. Arellano and Bond (1991) (AB) estimator is a GMM estimator for the first-differenced model. It is more efficient than the AH estimator, since it relies on a greater number of internal instruments and takes into account the structure of the idiosyncratic term Data sources and description Our study covers 13 countries of the CFA zone: Benin, Burkina Faso, Côte d Ivoire, Guinea Bissau, Mali, Niger, Senegal, Togo, Cameroon, Gabon, Equatorial Guinea, Congo, Central African Republic, and Chad. 13 The data are annual and cover the period from 1985 to To compute the misalignments, we refer to our previous study on this subject for the CFA zone countries (cf. Couharde et al., 2011). They correspond to the difference between observed real effective exchange rates and their equilibrium levels. Those are derived from a reduced equation where real exchange rates are explained by the following variables: terms of trade, the Balassa-Samuelson effect, the government spending, the net foreign assets and the degree of openness. Data have been updated until Regarding the growth regression, investment rates and inflation rates, based on Consumption Price Index, are extracted from the World Economic Outlook (WEO) database of the International Monetary Fund. The other variables namely real GDP per capita and its growth rate, population growth, life expectancy at birth or degree of openness come from the database World Development Indicators (WDI) of the World Bank. 12 Firstly, we consider the Anderson and Hsiao (1982) (AH) method based on two simple IV estimators that, upon transforming the model in first differences to eliminate the unobserved individual heterogeneity, use the second lags of the dependent variable, either differenced or in levels, as an instrument for the differenced onetime lagged dependent variable. However the AH estimator suffers from the weakness of the instrument, relying on their weak correlation with the endogenous variable. So, the model may be under-identified i.e. the rank of the instrument matrix is lower than that of the explanatory. 13 Guinea Bissau is not taken into account insofar as it has been a member of WAEMU since

14 3.3. Misalignment Calculations The literature considers several ways of assessing the equilibrium exchange and thereafter the misalignment: from old well-known purchasing power parity (PPP) to macroeconomic approaches and econometric approaches. The latter are generally used in the context of developing countries and have shown their relevance in several empirical studies about these countries (see table C in the appendix). Thus, basing on the BEER (Behavioral Equilibrium Exchange Rate) approach developed by Clark and McDonald (1999), we have assessed the equilibrium exchange rate of the CFA zone countries as in our previous study (Couharde et al., 2011). According to the BEER framework, the equilibrium exchange rate is the solution of a cointegration relationship between real exchange rate and a set of macroeconomic variables called macroeconomic fundamentals. Since data are updated, we re-estimate this cointegration relation using the same macroeconomic fundamentals as proposed by Couharde et al. (2011): the terms of trade (TOT), the relative per capita GDP corresponding to the famous Balassa-Samuelson effect (PROD), the Net Foreign Asset (NFA) and the government spending (GOV). All variables are in logarithm form, except the net foreign position. They are all expected to appreciate the equilibrium exchange rate, thus their expected signs are positive since the rise of the real exchange rate stand for a real appreciation. The following equation sums up the estimation, by PDOLS (Panel Dynamic OLS) developed by Kao and Chiang (2000) and Mark and Sul (2003), of the equilibrium exchange rate regression: (8.24) (17.09) (23.15) (2.24) (3) The coefficients are statistically significant 14 and have all the expected signs. We use these coefficients to assess the exchange rate misalignments by using the method developed by Elbadawi et al. (2008) and used by Couharde et al. (2011) Empirical results Table 2 reports the results of the GMM and LSDVC estimations of the growth equation for CFA zone economies. They show that an appreciation of the NEER of 10 points leads to a 14 Values in brackets are the associated t statistics. They are greater than Couharde et al. (2011) t statistics implying that these results are more robust. 15 For more details, see the quoted papers. The calculated misalignments are almost the same as those reported in Couharde et al. (2011). 14

15 decrease of the real GDP per capita of 0.01%. These negative effects tend to be slightly more important when countries are more open to foreign trade as evidenced by the negative and significant coefficient 16 of the variable (NEER*OPEN). It also reveals that exchange rate s misalignments exert a negative impact on the CFA zone economies as shown by previous studies. But, unlike these studies, the results also reveal that the advent of the euro did not significantly increase the negative effects of the exchange rate policy indicators (NEER and misalignment). In addition, tests show that there is no convergence process between the CFA zone countries since is greater than 0 and non significant. 17. To test the robustness of our findings, we firstly take into account other determinants of the growth rate; secondly we use two alternative measures of the euro effects (NEER and REER of the euro); and finally we take into account the effects of the CFA devaluation (see table A and table B in the appendix). On the whole, despite their sensitivities, our findings are robust to these considerations. Table A shows the sensitivity of the coefficients of the misalignment to the other growth determinants since they considerably diminish (see LSDVC estimation, column 5). This table also highlights the importance of external factors in the determination of growth rate in the CFA zone since the main significant variables are exchange rate policy variable and degree of openness variable. On the contrary, except the investment, the other internal variables are insignificant. Concerning the table B, it highlights the importance of the 1994 devaluation for the CFA zone countries. Indeed, considering the devaluation dummy, the effects of the NEER appreciation and the misalignment are significantly lower in the euro peg period relative to the period before. This can be explained by the fact that during the most of the French Franc peg period considering in this study (exactly ); the CFA zone countries registered a nominal appreciation greater than that observed in the last decade. Table B also shows that the effects of the CFA countries nominal effective exchange rates on their growth rates are greater than those of the NEER and REER of the euro. In fact, the variables related to the euro influence indirectly the growth rates of CFA economies through their nominal effective exchange rates reinforcing the choice of the latter as a proxy of the exchange rate policy. 16 The coefficient is negative for the AB estimation and counter intuitively positive for the LSDVC estimation. However the robustness tests (table B in the appendix) show that the coefficient is negative and significant for the two estimators. 17 is not significantly different from 1 thus is not significantly different to 0. This result is in line with Dramani s one (2010) who, in a cross section data study, concludes about the lack of convergence for all CFA zone countries. This result indicates the heterogeneity often noted in this zone. 15

16 Table 2: GMM and LSDVC estimation results of CFA zone Growth equation (1) (2) (3) (4) (5) (6) VARIABLES LSDVC 1 AB LSDVC AB LSDVC AB L.lnGDP 1.008*** 1.003*** 1.025*** 1.027*** 1.051*** 1.047*** (0.0263) (0.0700) (0.0185) (0.0364) (0.0143) (0.0886) NEER *** *** *** *** ( ) ( ) ( ) ( ) NEER*DUM 3.85e e-05 (9.50e-05) ( ) NEER*OPEN 8.53e-06*** -3.88e-06** (2.15e-06) (1.34e-06) MES *** *** (0.0369) (0.0450) MES*DUM (0.0722) (0.498) No. Observations No. Countries No. Instruments Hansen AR (2) Note: ***, **, * mean that the variable is respectively significant at 1%, 5% and 10%. The values in brackets are the robust standard errors. 1 The LSDVC standard errors are estimated by bootstrap. We use 100 iterations (see Bruno, 2005). 2 DUM corresponds to 1 in the euro period ( ) and 0 otherwise. 3 The reported values are the associated probabilities. Source: Author calculations. Conclusion In this article we determine the effects of the euro on the CFA zone countries competitiveness and its impacts on their economic growth. We have firstly calculated nominal and real effective exchange rates and showed that recent real appreciations in the CFA zone have been mainly caused by a strong euro. Then, we have used non stationary panel data econometrics in order to estimate equilibrium exchange rates and to assess misalignments. It appears that recent appreciations of the CFA did not translate in a strong real overvaluation except in Central African Republic, Côte d Ivoire and Togo indicating the improvement of the macroeconomic fundamentals in the rest of the CFA zone countries. Our estimations highlight strong evidence of damaged effects of the nominal effective exchange rate appreciation and the exchange rate misalignment on the growth in the CFA zone countries. These negative effects seem to be slightly more important in more open countries. We also show that the advent of euro didn t increase these damaged effects of the exchange rate policy even if the monetary policy of the European Central Bank (ECB) is clearly different to that applied by the Bank of France. However, a more flexible monetary 16

17 policy by the ECB may allow these countries to record more economic growth through gain in competitiveness. References Aguirre A. and Calderon C. (2005) Real exchange rate misalignments and economic performance, Central Bank of Chile Working Paper No. 315, April. Anderson T. W. and C. Hsiao (1982) Formulation and estimation of dynamic models using panel data, Journal of Econometrics, No. 18, p Arellano M. and S. Bond (1991) Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations, Review of Economic Studies, No. 58, p Barro R. and X. Sala-i-Martin (1995) Economic Growth, McGraw-Hill. Béreau S., Lopez-Vallenciano A. and Mignon V. (2009) Currency Misalignment and Growth: A New Look Using Nonlinear Panel Data Methods, CEPII Working Paper No. 17. Bolle P. (1997) Devaluation of the CFA franc four years on: Economic Integration and Employment Agenda, International Labour Reviews, vol. 3 (136). Bruno G. S. F. (2005) Estimation and inference in dynamic unbalanced panel-data models with a small number of individuals, The Stata Journal 5, No. 4, p Buddelmeyer H., Jensen H. P., Oguzoglu U. and Webster E (2008) Fixed Effects Bias in Panel Data Estimators, IZA Discussion Paper No Clark P. and MacDonald R. (1998) Exchange rates and economic fundamentals: A methodological comparison of BERRs and FEERs, IMF Working Paper No. 98/67. Combey A. and Nubukpo Kako (2010) Effets Non Linéaires de l'inflation sur la Croissance dans l'uemoa, MPRA paper No , July. Cottani J. A., Cavallo D. F. and M. S. Khan (1990) Real Exchange Rate Behavior and Economic Performance in LDCs, Economic Development and Cultural Change, vol. 39, No. 1, p Couharde C., Coulibaly I. and O. Damette (2011) Misalignments and Dynamics of Real Exchange Rates in CFA countries, EconomiX Working Paper No Diop A., Dufrenot G. and G. Sanon (2010) Is per-capita growth in Africa hampered by poor governance and weak institutions? An empirical study on the ECOWAS countries, African Development Review, 22(2), p

18 Dollar D. (1992) Outward-oriented developing economies really do grow more rapidly: Evidence from 95 LDCs, Economic Development and Cultural Change, vol. 40, No. 3. Dollar D. and A. Kraay (2004) Trade, Growth and Poverty, The Economic journal, vol. 114, No Dramani L. A. G. (2010) Convergence and Economic Integration in Africa: the Case of the Franc Zone Countries, AERC Research Paper No. 200, August. Dufrénot G., Mignon V. and C. Tsangarides (2010) The trade growth nexus in the developing countries: A quantile regression approach, Review of World Economics, vol. 146, No. 4, p Easterly W. and R. Levine (1997) Africa s growth tragedy: policies and ethnic divisions, Quarterly Journal of Economics, vol. 112, No. 4, p Elbadawi I., L. Kaltami and R. Soto (2009) Aid, Real Exchange Rate Misalignment and Economic Performance in Sub-Saharan Africa, Universidad Católica de Chile: Instituto de Economia, Working Paper No. 368, Pontificia. Gala P. and C. R. Lucinda (2006) Exchange Rate Misalignments and Growth, EconomiA, vol. 7, No. 4, p Ghura D. and T. Grennes (1993) The real exchange rate and macroeconomic performance in Sub-Saharan Africa, Journal of Development Economics, vol. 42. Ghura D. and T. M. Hadjimichael (1996) Growth in Sub-Saharan Africa, IMF staff papers, Vol. 43, No 3. Hadjimichael M. T. and M. Galy (1997) The CFA Franc Zone and the EMU, IMF Working Paper No. 97/156, November. Islam N. (1995) Growth Empirics: A Panel Data Approach, the Quarterly Journal of Economics, vol. 110, No. 4. Kevin L., Pesaran M. H. and R. Smith (1998) Growth Empirics: A Panel Data Approach A comment, the Quarterly Journal of Economics, vol. 113, No. 1. Kiviet J. F. (1995) On bias, inconsistency, and e_ciency of various estimators in dynamic panel data models, Journal of Econometrics, No. 68, p Klau M. (1998) Exchange Rate Regimes and Inflation and Output in Sub-Saharan Countries, BIS Working Paper No. 53, March. Kuikeu O. (2011) Against the CFA Franc zone, MPRA paper No , September. Mark C. N. and D. Sul (2003) Cointegration Vector Estimation by Panel DOLS and Long-run Money Demand, Oxford Bulletin of Economics and Statistics, 65:

19 Razin O. and S. M. Collins (1997) Real Exchange Rate Misalignments and Growth, Georgetown University and The Brookings Institution, manuscript. Rodrik D. (2008) The real exchange rate and economic growth, John F. Kennedy School of Government, Harvard University, October. Sachs J. D. and M. A. Warner (1997) Sources of Slow Growth in African Economies, Journal of African Economies, vol. 6, No. 3. Sala-i-Martin X. (1997) I Just Ran Two Million Regressions, American Economic Review, Vol. 87, No. 2. Toulaboe D. (2002) Real Exchange Rate Misalignment and Economic Growth in Developing Countries, Southwestern Economic Review, 33(1),

20 Appendix Table A: Robustness of the GMM and LSDVC estimation results (1) (2) (3) (4) (5) (6) VARIABLES LSDVC 1 AB LSDVC AB LSDVC AB L.lnGDP 1.006*** 1.018*** 1.017*** 1.033*** 1.023*** 1.103*** (0.0347) (0.199) (0.0218) (0.148) (0.0214) (0.0762) NEER ** ** ( ) ( ) ( ) ( ) NEER*DUM -6.02e e-05 ( ) ( ) OPEN ** *** *** *** ** ( ) ( ) ( ) ( ) ( ) ( ) Investment * * ( ) ( ) ( ) ( ) ( ) ( ) Life expectancy ( ) ( ) ( ) ( ) ( ) ( ) Population Growth (0.0171) (0.0222) (0.0111) (0.0240) (0.0126) (0.0190) NEER*OPEN -1.98e-05** -1.21e-05 (8.01e-06) (1.20e-05) MES * ** (0.0485) (0.0574) MES*DUM (0.0905) (0.0914) No. Observations No. Countries No. Instruments Hansen AR (2) Note: ***, **, * mean that the variable is respectively significant at 1%, 5% and 10%. The values in brackets are the robust standard errors. 1 The LSDVC standard errors are estimated by bootstrap. We use 100 iterations (see Bruno, 2005). 2 DUM corresponds to 1 in the euro period ( ) and 0 otherwise. 3 The reported values are the associated probabilities. Source: Author calculations. 20

21 Table B: Robustness of the GMM and LSDVC estimation results, effects of the devaluation and testing of alternative explanatory variables (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES LSDVC 1 AB LSDVC AB LSDVC AB LSDVC AB L.lnGDP 1.004*** *** 1.083*** 1.013*** 1.136*** 1.012*** 1.066*** (0.0343) (0.483) (0.0214) (0.0697) (0.0405) (0.277) (0.0397) (0.219) NEER ** *** ( ) ( ) NEER*DUM e * ( ) ( ) OPEN ** *** ** * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Investment * * ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Life expendancy ** ( ) (0.0142) ( ) ( ) ( ) (0.0136) ( ) (0.0117) Population Growth (0.0161) (0.0295) (0.0126) (0.0163) (0.0288) (0.0235) (0.0282) (0.0169) Devaluation *** ** ** (0.0263) (0.0551) (0.0223) (0.0217) (0.0438) (0.171) (0.0430) (0.114) MES * *** (0.0531) (0.0645) MES*DUM ** (0.0916) (0.0846) Euro REER ** ( ) ( ) Euro REER *DUM -7.07e e-05 ( ) ( ) Euro NEER *** ( ) ( ) Euro NEER *DUM -1.80e e-05 ( ) ( ) No. Observations No. Countries No. Instruments Hansen AR (2) Note: ***, **, * mean that the variable is respectively significant at 1%, 5% and 10%. The values in brackets are the robust standard errors. 1 The LSDVC standard errors are estimated by bootstrap. We use 100 iterations (see Bruno, 2005). 2 DUM corresponds to 1 in the euro period ( ) and 0 otherwise. 3 Devaluation corresponds to 1 in 1994 and 0 otherwise 4 The reported values are the associated probabilities. Source: Author calculations. 21

22 Table C: Literature review Authors Sample Economic Performance Indicators Measures of misalignment Estimation methods Results Cottani, Cavallo and Khan (1990) 24 developing countries Growth - Investment to GDP - ICOR - Export - Agricultural - PPP - Their model based measure - Cross-section OLS - Non significant for PPP - Negative for model Elbadawi, Kaltami and Soto (2009) 83 (36 SSA) countries Growth - Export diversification - Permanent Equilibrium Exchange Rate (PEER) - GMM-IV system - Negative Gala and Lucinda (2006) 58 (23 African ) countries Growth - Adjusted Real Exchange Rate - Pooled OLS, Within and IV; GMM system - GMM difference - Negative - Non significant for difference GMM Ghura and Grennes (1993) 33 SSA countries Growth - Investment - Saving - Exports and Imports - BEER - PPP - Black market premium - Cross-section estimation - Panel OLS - Negative Klau (1998) 22 SSA countries - Growth - Variation of the RER - Panel OLS - Negative for growth Inflation - Panel IV - Positive for inflation Razins and Collins (1997) 93 (23 SSA) countries Growth - BEER method including short-run variables shocks - Panel OLS - Negative Toulaboe (2006) 33 developing countries - Growth - PEER - Panel OLS - Negative Source: Author 22

23 Figure A: NEER, REER, Misalignment and Growth in the CAEMC countries Note: A positive (resp. negative) value of the misalignment corresponds to an overvaluation (resp. undervaluation). NEER and REER are read from left axis and the others from right axis. 23

24 Figure B: NEER, REER, Misalignment and Growth in the WAEMU countries 24

25 Note: A positive (resp. negative) value of the misalignment corresponds to an overvaluation (resp. undervaluation). NEER and REER are read from left axis and the others from right axis. 25