Is the West African monetary zone an optimum currency area?

(1)

Faculty of Economics and Business

Department of Economics

Bachelor’s Thesis

Is the West African Monetary Zone an Optimum Currency Area?

Name: Annick Besançon

Student number: 10283366

Date: 1

st

of July 2014

(2)

cooperation. In 1975 the Economic Community of West African States (ECOWAS), was created to increase economic integration between the West-African countries Benin, Côte d'Ivoire, Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Mauritania (which left in 2002), Niger, Nigeria, Senegal, Sierra Leone, Togo, Burkina Faso and Cape Verde (who joined in 1977) by creating one monetary union. To achieve this the ECOWAS countries planned on first creating some sub-currency unions and eventually merging them together to get one monetary union. The first sub-currency union is the West African Economic

Monetary Union (WAEMU). The other one is the West African Monetary Zone (WAMZ), which is the subject of my thesis. This zone is created in 1993 and consists of six countries: Gambia, Ghana, Guinea, Liberia, Nigeria and Sierra Leone. The idea was to first implement a common currency in the West African Monetary Zone, the Eco, and eventually also the WAEMU, and therefore the whole ECOWAS, would adopt this currency, thus creating a West African Monetary Union. The creation of the monetary union in the WAMZ was scheduled to commence in 2003, but due to the disability of the WAMZ countries to satisfy the convergence criteria this was postponed until 2005, 2009 and now this is planned for 2015 (WAMI, 2014).

Since the creation is postponed because the countries differ so much, it could be questioned whether it is a good idea in the first place to create a monetary zone. One of the main theoretical ideas of establishing a monetary zone is Mundell’s Optimum Currency Area (OCA) theory. This theory describes under which conditions a region can be considered an OCA, meaning that it would be feasible for this region to form a monetary zone. Therefore, this thesis will try to answer the following research question: is the West African Monetary Zone an Optimum Currency Area?

Several studies have been done on currency unions within Africa. Banik and Yoonus (2010) investigate in their paper the possibility of forming an Optimum Currency Area among the countries of the ECOWAS. They find that ECOWAS indeed fulfils the requirements of being an Optimum Currency Area. This would mean that following this study the West African Monetary Zone would also be an OCA. This result is controversial to the findings of Bénassy-Quéré and Coupet (2005). In their study on monetary arrangements in Sub-Saharan

(4)

Africa they find that it would only be feasible for Ghana, Gambia and Sierra Leone to join the West African Economic Monetary Union (WAEMU), but that Nigeria better not be part of this arrangement. This would mean that the currently opted West African Monetary Zone that groups the four previous mentioned countries would not be an Optimum Currency Area. Likewise, the study of Debrun, Masson and Pattillo (2002) reveals that Nigeria would be the only gainer of establishing WAMZ, while Gambia, Ghana, Guinea and Sierra Leone would be the losers1. The main reason brought up for this is that Nigeria, which would be predominant in the WAMZ considering its size, has a high level of fiscal distortion which would put pressure on the monetary union’s central bank to produce monetary financing. This would lower the utility of the other countries. So, the research until now has not arrived to one conclusion on whether a monetary union would be beneficial for the West-African countries. Besides that, there is a limited amount of research on the West African Monetary Zone in particular and especially a limited amount of research on the WAMZ that also includes Liberia. Therefore, the question of whether it would be feasible for the West African

Monetary Zone to form a monetary union is of great significance and especially of relevance as the WAMZ countries are currently addressing the issue of creating a monetary zone.

To answer the research question stated before, Mundell’s OCA criteria were analysed for the WAMZ countries. The main criteria, asymmetric shocks, was analysed by conducting several Vector Autoregression (VAR) analyses using the variables GDP and inflation. Based on the VAR conducted using data from 1987 to 2012, I found that the West African Monetary Zone is not an Optimum Currency Area. This same conclusion was arrived using the data before and after the creation of the WAMZ. However, although the West African Monetary Zone is not an Optimum Currency Area yet, it might be when more convergence takes place between the West African Monetary Zone countries or when a fiscal federalism system is established and labour mobility is increased.

The structure of this thesis is as follows: 2) the literature review extends on the West African Monetary Zone itself, the costs and benefits of forming a monetary zone, the

Optimum Currency Area theory and on the Optimum Currency Area criteria applied to the West African Monetary Zone, 3) the method and data section explains the research method that was exploited in this thesis, 4) the results section elaborates on the results of the research done as described in chapter 3, 5) the conclusion concludes on the previous chapters and answers the research question.

1_{Note that Liberia is not mentioned in any of the previous studies, because Liberia has only been part of the}

West African Monetary Zone since 2010.

(5)

2. Literature Review

2.1 The West African Monetary Zone

As was already described in the introduction, the West African Monetary Zone consists of six countries, the Gambia, Ghana, Guinea, Liberia, Nigeria and Sierra Leone that are all members of the Economic Community of West African States (ECOWAS). The West African

Monetary Zone was created in 2000 with the signature of the Accra Declaration. The idea behind this Zone was to first let the six WAMZ countries form a monetary union, to eventually merge with the other monetary union within the ECOWAS, the West African Economic Monetary Union (WAEMU). So with the creation of the West African Monetary Zone a fast-track approach for eventually achieving an ECOWAS-wide monetary zone was created.

To prepare the WAMZ countries for forming a monetary zone, the West African Monetary Institute (WAMI) was established in 2001 to undertake policy and technical preparations for the launch of the monetary union for the West African Monetary Zone and the establishment of a West African Central Bank (WAMI, 2014). One of the preparations that the West African Monetary Institute took was setting convergence criteria that the WAMZ countries had to satisfy before the formation of the monetary zone in 2003. Those criteria were the following: 1) inflation target should be lower than ten percent, 2) target of fiscal balance (excluding grants) as per cent of GDP should be smaller than 4 percent, 3) target of Central Bank financing of fiscal deficit should be smaller or equal to 10 percent of the previous year’s tax revenue and 4) target of gross external reserves should be bigger than or equal to three months of imports cover (WAMI, 2014). The WAMI set these criteria, because when forming a monetary union the net benefits of this are the highest when the countries are as similar as possible, because this means that they experience symmetric shocks.

When looking at the WAMZ countries, it can be seen that the countries differ quite a lot. Nigeria and Ghana for example, are oil exporters while the other countries are oil

importers. This could be an indication of asymmetric shocks. Also, Nigeria has been fiscally undisciplined up to now. As Nigeria is the biggest of the WAMZ countries, this may have its effects on the other countries and it can therefore be questioned if it would be beneficial for the other countries to join the WAMZ (Masson and Pattillo, 2005). So, it is of significance that the countries satisfy the convergence criteria. As can be seen in appendix A, until 2012 still not all the convergence criteria had been satisfied. Therefore, the formation of the

(6)

monetary union has been postponed several times until 2005, 2009 and now it is planned for 2015.

2.2 Monetary union

A monetary union is formed when two or more countries abandon their own national currencies for a common currency managed by a common central bank with a common monetary policy. With forming a monetary union two or more countries irrevocably and irreversibly fix their exchange rates towards the common currency (Ishiyama, 1975). A well-known example of a monetary union is the European Monetary Union (EMU) that currently comprises 18 countries, which share the same currency, the Euro, and have a common central bank, the European Central Bank (ECB).

When forming a monetary union there are both costs and benefits that arise. The main cost is the loss of monetary independence. Countries experience shocks, which are

unpredictable events that effect aggregate supply and demand. These shocks can arise in different forms. For example a sudden change in oil prices or an earthquake can both result in a change in aggregate demand or supply. When countries have flexible exchange rates and therefore, are able to conduct their own monetary policy, they can alter the money supply or exchange rate to smoothen the shock and to bring demand and supply back in equilibrium. However, when countries form a monetary union they lose their monetary independence and can no longer use monetary policy to smoothen these shocks. Therefore, when countries form a monetary union other mechanisms will have to be present (Alesina and Barro, 2002). This will further be elaborated on in the next section.

The benefits arise usually at the microeconomic level and the main benefits that are often described by literature are increased price transparency, reduced transaction costs and reduced exchange rate risk. Because of the fact that a common currency is adopted when forming a monetary union, it is easier to compare the prices of a certain product in the different countries and therefore the price transparency increases (Alesina and Barro, 2002). As a result, consumers will buy their products at the cheapest country, which will increase competition and therefore market efficiency. The reduction in transaction costs arise, because there are no longer commissions or margins that have to be paid when exchanging currency. The benefits of the reduced transaction costs will be larger with the amount of trade among the countries that will join the monetary union. This argument has a component of

endogeneity in it, as the reduction of transaction costs will increase trade among the countries in the monetary union and as a result of this the benefits of the reduced transaction costs will

(7)

be even larger. The third benefit, reduced exchange rate risk, will also increase intra-trade. In general people are risk-averse and they will take into account the risk they face when doing business. Because the risk of a changing exchange rate is no longer present, the risk of doing business with the countries in the monetary union will decline and as a result of this intra-trade will increase (Alesina and Barro, 2002).

With respect to the West African Monetary Zone the benefits will initially not be caused by the reduction in transaction costs. This is because the trade among the WAMZ countries is low and the WAMZ countries mostly trade with countries outside Africa, mostly with countries in Europe. The exports of the West African Monetary Zone that stay in the West African Monetary Zone, accounts for only 3.6% of total exports. The imports account for 4.6% of total imports (Masson & Pattillo, 2005, p. 98). As was described before, the benefits of reduced transaction costs are highest when the intra-trade is high. As this is not the case for the West African Monetary Zone, this benefit is at this moment not significant. However, there are also several studies that have concluded that intra-trade will increase after forming a currency union. For instance, an often cited paper on this topic is the one of Rose (2000) where he investigates the effect of a common currency on trade and finds that two countries that share the same currency trade three times as much as they would with different currencies. So although the reduced transactions costs might not be of significance right now, it might be after forming a currency union. The other two described benefits, price

transparency and reduced exchange rate risk, will be of benefit to the West African Monetary Zone. Also, the creation of a monetary zone comes with strict rules and preparations. This together with the reduced exchange rate risk and price transparency will lead to a more stable economic environment and will lead to increased monetary and fiscal discipline. For the West African Monetary Zone this would be the main benefit (Karras, 2007).

2.3 Optimum Currency Area

Robert Mundell was the first one to write about Optimum Currency Areas in his Theory of Optimum Currency Areas (Mundell, 1961). In this theory he describes under which

circumstances it is best for a geographical region to apply a flexible exchange rate. He illustrates this by considering two countries that are initially in full employment and balance-of-payments equilibrium and that experience several asymmetric demand and supply shocks.

One of his examples is a shift of demand from the products of country A to the

products of country B. Because demand for the products of country A has declined, the output of country A will decline and therefore the unemployment will rise. In country B the opposite

(8)

happens. To restore equilibrium the central banks of country A and country B could

respectively lower the interest rate in country A to stimulate aggregate demand and raise the interest rate in country B to demotivate aggregate demand. The resulting depreciation of the currency of country A relative to the currency of country B would have further stimulated the aggregated demand in country A and reduced that of country B and the equilibrium would be restored. However, this would only be possible if the countries had flexible exchange rates. As has already been mentioned in the previous chapter, when the countries would be a monetary union this adjustment mechanism would not be present and other adjustment mechanisms are needed. The adjustment mechanisms described by Mundell are wage flexibility, labour mobility and fiscal policy. However, when the different countries in the monetary union experience symmetric shocks, e.g. shocks that need the same monetary policy to bring demand and supply back in equilibrium, then this adjustment mechanism can be used. Therefore, it can be deduced from Mundell’s theory that, if the regional shocks are symmetric in nature, fixed exchange rates would be appropriate, but if they are asymmetric than there should be wage flexibility and/or a high degree of labour mobility or a fiscal transfer

mechanism. The criteria for a region to be an Optimum Currency Area described by Mundell can be summarized as follows: 1) absence of asymmetric shocks, 2) flexibility of the labour market and 3) presence of a fiscal transfer mechanism.

Next to Mundell several others have written about Optimum Currency Areas, but the most well-known of these are McKinnon (McKinnon, 1963) and Kenen (Kenen, 1969). Both researchers extended the criteria set by Mundell. McKinnon added the degree of openness of a country and Kenen added the degree of commodity diversification. With the degree of

openness, he meant the ratio of tradable to non-tradable goods, where tradable goods are goods whose prices are largely determined on the world market and non-tradable goods are goods whose prices are determined on the domestic market. The higher the degree of

openness of a country, the lower will be the costs of the loss of the exchange rate tool. This is because in small open economies, which have a higher share of imports compared to

relatively more closed economies, the same exchange rate depreciation will lead to a higher increase in domestic prices. Therefore, the systematic use of the exchange rate tool in open economies will lead to systematically higher price variability, which is costly. Thus, the loss of an independent monetary policy is likely to be less costly in relatively more open

economies (McKinnon, 1963).

In Kenen’s ‘Theory of Optimum Currency Areas: an Eclectic view’(1969) he writes that there are three points why commodity diversification is so important. The first one is that

(9)

a well-diversified national economy will not have to undergo changes to its terms of trade as often as a single-product national economy and therefore also not as frequent changes in national exchange rates. The second one is that when there is a drop in demand for its principal exports, the rise in unemployment will not be as sharp as in a less-diversified national economy. The third one is that the links between external and domestic demand, especially the link between exports and investment, will be weaker in diversified national economies, so that variations in domestic employment “imported” from abroad will not be greatly aggravated by corresponding variations in capital formation (Kenen, 1969). Or in short, the more diversified a country is, the less vulnerable the country is to sector specific shocks and the lower the stabilization costs of joining a monetary union.

2.4 OCA and the West African Monetary Zone

Applying the Optimum Currency Area criteria to the West African Monetary zone, it can be concluded that they do not satisfy all the criteria.

2.4.1 Labour mobility

Labour mobility can function as an adjustment mechanism in the sense that when there is a shift in demand from the products in country A to the products in country B, which results in a decrease in output for country A, an increase in output of country B, a resulting decrease in employment in country A and an increase in employment in country B, demand and supply can be brought back to equilibrium by labour migration from country A to country B. The United States of America are often seen as a perfect Optimum Currency Area, while the European Union is not. One of the most frequently mentioned reasons for this is the lack of labour mobility in the European Union. Where in the US the labour mobility is high, the labour mobility within the European Union is quite low due to the big differences in language and culture (Nickell, 1997).

When looking at labour mobility in the West African Monetary Zone, it is difficult to decide on whether the West African Monetary Zone has a high level of labour mobility or not. The language in the West African Monetary Zone will at least not be an obstacle to labour mobility as almost all the countries are English-speaking countries. Of all the WAMZ

countries only Guinea is a francophone country. Also, the adoption of the ECOWAS passport in 2001 could stimulate labour mobility. However, up till now the effects of the adoption of the common passport have been limited. One of the reasons for the limited labour mobility mentioned by Masson and Pattillo, is the fact that not all the countries are geographically

(10)

adjacent. As can be seen on the map below, where the red countries are the WAMZ countries, only Guinea, Sierra Leone and Liberia have common borders. They also write that labour mobility is still limited because of ethnic conflict and a general lack of financial resources (Masson and Pattillo, 2005, p. 97). However, this contradicts the paper of Awumbila, Benneh, Teye and Atiim (2014). In their paper “Across artificial borders: an assessment of labour migration in the ECOWAS region” they state that labour migration has been and still is the dominant form of migration and that although the direction of migration has shifted, labour mobility is quite high. However, they also conclude that there is a paucity of reliable and up-to-date data on migration in the ECOWAS countries, due to amongst other things, a lack of harmonization of definitions. So, research has not lead to one conclusion on whether labour mobility is limited or not within the West African Monetary Zone.

(11)

2.4.2 Fiscal federalism system

Another adjustment mechanism is the presence of a fiscal federalism system, which is a risk-sharing or fiscal transfer system. When there is a shift in demand from the products of country A to country B, the output in country B will be higher and therefore GDP and tax income. This difference in output can be equalized by transferring part of the tax income from country B to country A. So the shift in demand is no longer a problem.

Currently there is no fiscal federalism system present in the West African Monetary Zone. Or at least, not a significant one. The amount of papers investigating this for West-Africa is limited, but one that does is the paper written by Tapsoba (2010). His paper estimates risk-sharing channels among West African countries from 1970 to 2004. In his paper he analyses several risk-sharing channels; factor markets, regional credit markets and regional fiscal transfer system. He concludes that risk-sharing among the West-African states is low and the channel through which some risk-sharing is present is through net savings and the adjustment of public net saving (Tapsoba, 2010). Though, the fiscal transfer system is not significant enough to function as an adjustment mechanism.

As it is not clear whether labour mobility could function as an adjustment system or not and a fiscal transfer mechanism is not present, I will concentrate on symmetry of shocks to decide whether the West African Monetary Zone is an Optimum Currency Area. Therefore, in the next section I will elaborate on the method to investigate to what extent the countries experience symmetric shocks.

3. Methods and data 3.1 Vector autoregression model

To investigate whether the countries experience symmetric shocks, I conducted a Vector autoregression (VAR) model. A Vector autoregression is a set of k time series regressions, in which the regressors are lagged values of all k series. This model differs from a basic

autoregression in the sense that it measures multiple equations at once that are all dependent of each other.

The VAR model that I conducted in this research consisted of two time series regressions with two variables: the percentage change in real GDP and the implicit GDP deflator. These two variables were used to measure output growth and inflation respectively. These two variables were used, because the VAR model that I conducted is based on the Aggregate Demand and Aggregate Supply Model (AD-AS model), which is shown in figure

(12)

1. This model explains output and price level through the relationship of aggregate demand and supply. The Aggregate Demand-curve is downward sloping. This is because when prices decline the demand will increase and therefore output will increase. The short run aggregate supply curve is upward sloping. Assuming that wages are sticky an increase in prices will lower the real wages and therefore it will be less costly for suppliers to produce, resulting in an increase in output. The long run aggregate supply curve is vertical, because wages are flexible in the long-run (Bayoumi and Eichengreen, 1992).

Figure 1: AD-AS model

Source: made by author

When the economy experiences a demand shock the aggregate demand curve goes up or down depending on whether the shock is positive or negative. In the example in figure 1, the demand curve goes down and the new equilibrium is in point B. Both output and prices decrease as a result of this. Because wages are sticky in the short-run it will take some time before the economy is back in a long-run equilibrium. However, as the process of decreasing output and prices continues the economy is finally in a long-run equilibrium, C, where output is the same as before the shock and prices are lower (Bayoumi and Eichengreen, 1992). How long this process takes depends on the amount of lags.

The VAR model can be depicted as follows: Xt≡       ∆ ∆ Pt Yt (1) where,

∆Yt = constant + a11 Yt-1 + a12 Yt-2 + a1i Yt-i + b11 Pt-1 + b12 Pt-2 + b1i Pt-i+ ɛ (2)

(13)

where Yt is the percentage change in real GDP, Pt the implicit GDP deflator, a and b are

coefficients of Yt and Pt, i is the amount of lags and ɛ are the shocks that the variables experience.

Before I could conduct the VAR model, I first had to investigate whether the variables were unit root, so that the vector Xt≡ 

     ∆ ∆ Pt Yt

would be stationary. When a vector is stationary it means that it reverts around a constant long-term mean or equilibrium and has a constant variance independent of time. When this is the case the VAR can be applied using the data as given. When this is not the case, it has to be investigated whether the data contains a trend, cycle or a drift. Before investigating this econometrically, based on economic theory I expected the variables to be unit root. As growth and inflation follow a business cycle that goes up and down, they move around a certain mean and therefore would be stationary. To investigate this even further I used the Augmented Dickey-Fuller test. The results, which can be found in Appendix B, showed that most of the data was stationary at a 5% significance level. The data that were not stationary at the 5% significance level, were at 10%. So, this in combination with economic theory, made me conclude that the variables were unit root and the vector Xt stationary.

As was mentioned before, the amount of lags show how long it takes for an economy to return to long-run equilibrium after it has experienced a shock. The amount of lags can be determined by either testing the coefficients of the model or by using information criteria such as the Bayes Information Criterion (BIC) and the Akaike information criterion (AIC). I used the latter option, because this way you have several information criteria that estimate the amount of lags in their own way and you can compare them to get the best fit. Which lag should be chosen depends on the values of the information criteria. In general the lag with the lowest BIC and AIC must be chosen, but when there is not a clear answer given by the BIC and AIC the decision should be based on economic theory. I determined the lags per country and the results of this can be found in Appendix C. As can be seen in this table the amount of lags differ per country. One lag means that the GDP for time t is effected by the GDP at time t-1, two lags means that the GDP for time t is effected by the GDP at time t-1 and t-2, etc. The higher the amount of lags, the more accurate the model for predicting GDP. However, the amount of lags is inversely related with the degrees of freedom. By adding another lag, another variable is added and therefore the degrees of freedom decrease, resulting in less accurate results.

(14)

After determining the lags and whether the variables were unit root, I conducted the VAR per country. Doing this I got the residuals per country that resembles the shocks that the countries experience. To estimate whether the countries experience symmetric shocks I correlated them. The correlation resulted in a value between -1 and 1. A correlation value of 1 would mean that when country A experienced a shock increasing GDP by a certain amount, country B would experience the exact same shock in size and direction, indicating a perfectly positive correlation and therefore symmetric shocks. As a result, monetary policy could be used to bring demand and supply back in equilibrium. A value of -1 would indicate the exact opposite of a correlation value of 1. The perfectly negative correlation would indicate

asymmetric shocks and therefore monetary policy could not be used as an adjustment mechanism. To get to the conclusion that the West African Monetary Zone experiences symmetric shocks and would be an OCA, the correlations must be positive and as high as possible.

3.2 Data

The data that I used for the estimation of the Vector autoregression model are at annual basis and span the period from 1960 to 2013. The data came from the website of the World Bank. Under macroeconomic indicators the real GDP and GDP deflator could be found for the six West African Monetary Zone countries; the Gambia, Ghana, Guinea, Liberia, Nigeria and Sierra Leone. The period of the World Bank data ranges from 1960 to 2013, but due to the unavailability of data for some of the countries I used the period from 1987 till 2012. To get growth in percentage from the real GDP data, I first subtracted Yt-Yt-1 and after that I took the

natural logarithm of this to get the change in percentages.

4. Results

4.1 Summary of the data

In this section I will describe the results of the Vector autoregression (VAR) model. To get an idea of the GDP growth and GDP deflator I summarized the two variables by

estimating the mean, the standard deviation and the minimum and maximum values. In table 1 the summary for the GDP growth is shown and in table 2 the one for the GDP deflator. The mean of the GDP growth does not show any striking results. The mean has the highest value for Ghana and Nigeria, which is not very surprising as these two countries are the biggest ones of the West African Monetary Zone and have experienced an increase in GDP growth the last ten years. The standard deviations however, do show some striking results.

(15)

The standard deviation of Liberia is 30.33 which is much higher than the standard deviations of the other countries. This means that the GDP growth of Liberia is much more volatile than the GDP growth of other countries. The unstable GDP growth is a sign of an unstable

economy, indicating that Liberia has experienced more shocks compared to the other countries. This could result in asymmetric shocks.

Table 1: summary of GDP growth (annual in %)

Looking at table 2, it is again Liberia that stands out with an inflation mean of 151.06 and a maximum value of 3789.209. Also Sierra Leone has an inflation mean that is significantly higher than the mean of the other countries. The higher inflation mean could indicate more changes in the price level in the AD-AS model, indicating more shocks, which could be a sign of asymmetric shocks.

Table 2: summary of the GDP deflator (annual in %)

The graphs below show a clearer image of the progress of the GDP growth and inflation. The high peak of inflation for Liberia can in part be explained by the increase in printing Liberian dollars without the aggregate output level to support this being present. This causes

(16)

Graph 1: GDP growth (annual in %)

Graph 2: Inflation

Graph 3 shows the GDP growth without Liberia included, to get a clearer image of the GDP growth of the other countries throughout the years. The graph of the GDP of Nigeria and Sierra Leone stands out immediately, because of the high volatility. Although the standard deviation of GDP growth is high for these countries, they move more or less in the same

(17)

direction. Hence, Nigeria and Sierra Leone will probably not experience asymmetric shocks compared to each other. Furthermore, Liberia, Sierra Leone and Nigeria all have GDP growth with a high range. Therefore, I expect that the correlation diagram of the residuals will show small correlations for Liberia, Sierra Leone and Nigeria with the other West African

Monetary Zone countries. Also, Gambia moves exactly the opposite way compared to Guinea, Sierra Leone and Ghana, especially from 2007 on. The opposite changes in GDP indicates opposite changes in aggregate supply and aggregate demand, indicating different shocks. Therefore, based on this graph I expect Gambia to have asymmetric shocks with Guinea, Sierra Leone and Ghana.

Graph 3: GDP growth in % without Liberia

Graph 4 shows the inflation in percentages for all the WAMZ countries, except for Liberia. Although there are some peaks in inflation for the countries Gambia, Sierra Leone, Ghana and Nigeria, it can be seen that throughout the years the inflation rate has converged. This started around 2000, which is the same year as the year in which the West African Monetary Zone was created. The converging of the inflation rate might therefore be caused by the

convergence criteria set with the creation of the WAMZ and could be a sign of further convergence.

(18)

Graph 4: inflation in %, without Liberia

4.2 Vector Autoregression Part l: 1987-2012

To investigate whether the countries experience symmetric shocks, I first conducted the VAR per country for the period ranging from 1987 till 2012. After that I conducted a VAR to investigate whether the creation of the West African Monetary Zone has had an effect on the level of symmetry of the residuals. The resulting residuals correlation diagrams are depicted in tables 3, 4 and 5 respectively.

(19)

Table 3 shows the correlation diagram of the residuals of the West African Monetary Zone countries; Ghana, Guinea, the Gambia, Liberia, Nigeria and Sierra Leone, for the period ranging from 1987 till 2012. The West African Monetary Zone would constitute an Optimum Currency Area based on symmetry of shocks, when the residual correlations between the countries are positive and close to 1. What immediately stands out when looking at the correlation diagram in table 3, is that the residual of Gambia has a negative correlation with all the other country’s residuals or in other words; Gambia has asymmetric shocks with every other WAMZ country. Not only are they all negative, they are also quite big except for the one with Nigeria. This would mean that based on the data ranging from 1987 till 2012, the West African Monetary Zone including all six countries is not an Optimum Currency Area, because of the negative correlations of Gambia with the other WAMZ countries. Besides the striking results of Gambia, there are no significant correlations that indicate asymmetric or symmetric shocks among all six WAMZ countries. Ghana has significant correlations with Nigeria (0.2275) and Gambia (-0.41), indicating symmetric shocks with Nigeria and asymmetric shocks with Gambia. Guinea has a significant positive correlation with Liberia (0.3995), indicating symmetric shocks, and a significant negative correlation with Gambia (-0.4154) and Nigeria (-0.2140), both indicating asymmetric shocks. Sierra Leone has

significant negative correlations with Liberia (-0.2340) and Gambia (-0.2152). Therefore, based on these correlations it would be feasible for Nigeria and Ghana to form a currency union and for Guinea and Liberia, but there is no evidence that forming a monetary zone with all six countries would be feasible.

The differences in shocks may be a result of the division of GDP among sectors. For example, as can be seen in table 4, where most of the WAMZ countries depend on agriculture, Gambia’s GDP per sector is highest for services. These are two totally different sectors and therefore they will experience different shocks that need a different monetary policy to be smoothen. The division in GDP could be an explanation for the negative correlations between Gambia and the other countries. The symmetry between the shocks of Nigeria and Ghana can be explained by the symmetry in the countries itself. Nigeria and Ghana are the biggest countries of the WAMZ and both have experienced an increase in growth the past years; Ghana because of the discovery of an oilfield just off Ghana’s coast in 2007 and Nigeria because of an increase in GDP growth caused by an increase in GDP in the non-oil sector. Another explanation for the symmetry in shocks of Ghana and Nigeria could be the fact that Nigeria and Ghana are both petroleum and petroleum products exporters, while the other WAMZ countries are all petroleum and petroleum products importers. As petroleum products

(20)

are the main export products of Nigeria, Nigeria experiences different shocks than the other petroleum-importing WAMZ countries. This can also be seen in the correlation diagram. All the correlations with Nigeria are negative or insignificant, except for the correlation with Ghana, another petroleum exporter.

Table 4: GDP division

Source: The World Factbook

4.3 Vector autoregression Part ll: before and after creation WAMZ

Two other VARs were conducted using data from before the creation of the West African Monetary Zone (1987-2000) and data from after the creation of the West African Monetary Zone (2001-2012). In 2000 several convergence criteria were set up to enhance integration, therefore I conducted those VARs to see whether this has had its effects on the correlation of the residuals. The results are summarized in table 5 and 6.

Table 5: correlations residuals before creation WAMZ

Table 5 shows the correlation diagram from the period before the creation of the West African Monetary Zone. The results suggest that there are some currency unions that would be

feasible; Nigeria in combination with Guinea and Ghana could be a feasible currency union, Ghana and Sierra Leone, and Guinea in combination with Gambia, Nigeria and Sierra Leone. Yet, none of the results support the creation of a monetary zone containing all six countries.

However, when looking at the results in table 6, more support can be found in favor of the creation of a monetary zone with all six countries; Ghana, Guinea, the Gambia, Liberia,

Ghana Guinea The Gambia Liberia Nigeria Sierra Leone

Services 49.8% 30.5% 67.7% 17.7% 26% 33.5%

Industry 28.7% 46.5% 12.6% 5.4% 43% 18.6%

(21)

Nigeria and Sierra Leone. Almost all the countries show a positive correlation with each other, except for Sierra Leone and Guinea. Guinea has asymmetric shocks with Gambia (-0.1225) and Nigeria (-0.2498) and positive insignificant shocks with Liberia (0.0967) and Sierra Leone (0.0828). For Sierra Leone the same story applies, it has asymmetric shocks with the other countries or insignificant ones. Therefore, based on these correlations Nigeria, Sierra Leone, Ghana, Gambia and Liberia would form a good currency union based on symmetric shocks and another feasible currency union would be a currency union between Ghana, Guinea, Gambia, Liberia and Nigeria. So, the only countries keeping the WAMZ from being an Optimum Currency Area based on symmetric shocks, are Guinea and Sierra Leone.

At first side this may seem odd, as they are already in the WAMZ and it would be more logical for Liberia to be an outcast since Liberia joined the WAMZ at a later point in time. However, when looking at the convergence criteria in appendix A, Guinea and Sierra Leone are the countries that have been lacking the most in satisfying the convergence criteria. This might be an indication that the West African Monetary Zone is an optimum currency area if further integration would be done concerning Sierra Leone and Guinea.

Table 6: correlations residuals after creation WAMZ

5. Conclusion

This thesis tried to answer the question whether the West African Monetary Zone is an Optimum Currency Area. This was done by applying Mundell’s OCA criteria for the West African Monetary Zone and by conducting several Vector Autoregression models using data of GDP and inflation for the period 1987 till 2012.

(22)

In the literature review section I discussed the West African Monetary Zone in general, the costs and benefits of forming a monetary union, Mundell’s Optimum Currency Area theory and the additional criteria of Kenen and McKinnon and the criteria of Mundell applied to the WAMZ. The main cost of forming a monetary union being the loss of a monetary policy instrument and the main benefit being the increase in monetary and fiscal discipline. The criteria from Mundell; 1) absence of asymmetric shocks, 2) flexibility of the labour market and 3) presence of a fiscal transfer mechanism, were the foundation for this research. However, as was concluded in the literature review, there is no significant fiscal transfer mechanism present in the West African Monetary Zone and the level of labour mobility within the West African Monetary Zone is unclear. For this reason, the symmetry of shocks emerged as the main criteria in this research. Hence, several Vector Autoregression models were conducted to analyse the symmetry of the shocks experienced by the different WAMZ countries.

The first VAR analysis that was conducted, used data of GDP and inflation for the period ranging from 1987 till 2012. This analysis found that there was no significant support for the West African Monetary Zone being an Optimum Currency Area. The second analysis, the VARs that were conducted using data before and after the formation of the WAMZ, investigated whether the convergence criteria had any effect on the symmetry of the shocks. The results showed that before the creation of the WAMZ there was no significant evidence for the WAMZ countries to be a feasible monetary zone, but that after the formation of the WAMZ and after several years of converging there was significant evidence of a feasible monetary zone between Nigeria, Ghana, Gambia and Liberia, indicating that the convergence had a positive effect on the symmetry of the shocks. From these results it may therefore be concluded that the West African Monetary Zone consisting of all six countries may not be an Optimum Currency Area at the moment, but that a monetary zone between Nigeria, Gambia, Liberia and Ghana would be and that when Sierra Leone and Guinea converge even further to the other countries, the West African Monetary Zone might be an OCA based on symmetry of shocks. Hence, the results indicate the importance of more integration. Aiming for increased labour mobility and establishing a significant fiscal transfer mechanism could of course also increase the feasibility of creating a monetary zone.

There are, however, some limitations to this research. First of all, in this research I did not make a difference between shocks that resulted from external shocks and shocks that resulted from policy implementations. My results indicate that the WAMZ is on its way to being an OCA, but when considering different shocks this might result in totally different

(23)

conclusions. Also, this research is economic-based and does not take into account political features. In developing countries the political will and political environment may be even more important than the economical environment. The north of Nigeria for example, is politically unstable at the moment. To get an exhaustive conclusion there should also be looked at the political characteristics of the countries. And finally, although symmetry of shocks is an important criterion, it is only one aspect of a monetary union. In my research I did not include openness or product diversification of the countries. These things could be areas for future research.

(24)

References

Alesina, A. and Barro, R.J. (2000) “Currency Unions” The Quarterly Journal of Economics Vol. 117, No. 2 pp. 409-436

Awumbila, M. Benneh, Y. Teye, J. and Atiim, G. (2014) “Across artificial borders: an assessment of labour migration in the ECOWAS region” ACP Research Report

Banik, N. and Yoonus, C.A. (2010) “Single Currency in the ECOWAS Region: Does it Make Sense?” African Journal of Economic Policy Vol. 17, No.1

Bayoumi, T. and Eichengreen, B. (1992) “Shocking aspects of European Monetary unification” NBER Working Paper Series, WP 3949, Cambridge, Massachusetts. Bayoumi, T. and Ostry, J.B (1997) “Macroeconomic shocks and trade flows within sub-Saharan Africa: Implications for optimum currency arrangements” Journal of African economies Vol. 6, No. 3, pp. 412-444

Bénassy-Quéré, A. and Coupet, M. (2005) “On the Adequacy of Monetary Arrangements in Sub-Saharan Africa” The World Economy Volume 28, Issue 3, pp. 349–373

Debrun, X. Masson, P. and Pattillo, C. (2002) “Monetary Union in West Africa: Who might gain, who might lose and why?” The Canadian Journal of Economics Vol. 38, No. 2, pp. 454-481

Frankel, J. and Rose, A. (1998) “The Endogenity of the Optimum Currency Area Criteria” The Economic Journal Vol. 108, Issue 449, pp. 1009–1025

Frankel, J. and Rose, A. (2002) “An Estimate of the Effect of Common Currencies on Trade and Income” The Quarterly Journal of Economics Vol. 117, No. 2

Ishiyama, I (1975) “The Theory of Optimum Currency Areas: A Survey” Staff Papers - International Monetary Fund Vol. 22, No. 2, pp. 344-383

Karras, G. (2007) “Is Africa an Optimum Currency Area? A comparison of Macroeconomic Costs and Benefits” Journal of African economies Vol. 16, pp. 234-258

Kenen, P.B (1969) “The theory of optimum currency areas: an eclectic view” Monetary Problems of the International Economy

Masson P. and Patillo, C. (2004), The Monetary Geography of Africa, Brookings Institution Press, Washington, D.C

McKinnon, R.I (1963) “Optimum Currency Areas” American Economic Review Vol. 53, pp. 717-725

Mundell, R. (1961) “A Theory of Optimum Currency Areas” The American Economic Review Vol. 60, No. 4 pages 657-665

(25)

Nickell, S. (1997) “Unemployment and Labor Market Rigidities: Europe versus North America” The Journal of Economic Perspectives Vol.11, No. 3

Rose, A. (2000) “One Money, One Market: Estimating the Effect of Common Currencies on Trade” Economic Policy Vol. 15, Issue 30, pp. 7–46,

Tapsoba, S.J (2010) “West African Monetary Integration and Interstates Risk-Sharing” West African Journal of Monetary and Economic Integration

West African Monetary Institute, WAMI (2012) Primary Convergence - Primary Criteria 2001-2012 in English & French

(26)

Appendix A: Convergence criteria

Appendix B: results Augmented Dickey-Fuller test

(27)

Guinea

The Gambia

Liberia and Nigeria

(28)

Inflation Ghana and inflation Guinea

Inflation Gambia and inflation Liberia

(29)

Appendix C: Lags per country, found by using BIC and AIC

Country Ghana Guinea Gambia Liberia Nigeria Sierra Leone