A New Trade Agreement in Africa:
Predicting the AfCFTA Impact on Trade
Bas de Wit
S2958910
B.de.wit.1@student.rug.nl 06-01-2020
Master’s Thesis ED&G
University of Groningen, Faculty of Economics and Business
1
A New Trade Agreement in Africa:
Predicting the AfCFTA Impact on Trade
Abstract
In this paper the impact of the recently created African Continental Free Trade Area
(AfCFTA) is predicted. This is done by turning the AfCFTA impact on for African countries, using the estimated average impact of trade agreements on exports for other developing countries. The results indicate a predicted average increase in trade for African countries of 2.45%, although highly varying between countries. For individual countries the magnitude of the AfCFTA impact is explained by their share of within Africa exports and the share of these exports that are not going to African countries within a common already existing African TA.
2 Introduction
On May 30th, 2019 the African Continental Free Trade Area (AfCFTA) went into force. The AfCFTA has received a large support from the African countries. At this point Eritrea is the only African Union member that has not signed the agreement. 28 of the countries have already ratified it1. If it is successful in reducing barriers to trade it could be a stimulus for economic development on the continent. There is not much previous literature examining the possible impacts of trade agreements (TAs) on trade in Africa, let alone of the AfCFTA. The few studies that have been done conclude that the AfCFTA could have a positive effect on trade and welfare for the African countries (Abrego, Amado, Gursoy, Nicholls and Perez-Saiz, 2019; Luke and Macleod, 2019)
There are currently already regional TAs in force on the African continent. Almost every African country is a member of at least one of the African TAs. Although they have been quite successful in reducing tariffs, within Africa trade remains limited. Tariffs between regions are high and the infrastructure is of poor quality (Abrego et al., 2019). The AfCFTA is expected to lower the tariff and non-tariff barriers on a continental scale on an increasing amount of product types. The AfCFTA is intended to become a FTA which eventually should eliminate all internal tariffs on goods and services (Abrego et al., 2019; Luke and Macleod, 2019).
It is expected that the AfCFTA will stimulate economic development if it is successful in increasing within Africa trade. A reduction in trade costs will increase economies of scale and make it easier for countries to join international value chains. This stimulates the
development of the manufacturing sector and diversifies the economy away from a sole focus on agriculture or natural resources (Luke and Macleod, 2019; UNCTAD, 2019). The aim of this paper, however, is to examine the AfCFTA impact on trade and not on economic development.
Although, there is a lot of theoretical and empirical literature on the effect of TAs (See for example Baier and Bergstrand, 2007; Carrère, 2006; Ghosh and Yamarik 2004a/b; Kohl, 2014), relatively few of it focuses on developing countries only (Ekanayake, Mukherjee, Veeramacheneni, 2010; MacPhee and Sattanayuwat, 2014), and very little focuses on Africa (Carrère, 2004; Yang and Gupta, 2008). This paper can, thus, contribute to the little
knowledge there is about the impact of TA s on Africa specifically, and developing countries more generally.
In this paper it is predicted what the impact of the AfCFTA will be on the amount of trade of the African countries. To do this, first the gravity equation is used to estimate the average impact on exports of other TAs in the world involving developing countries. Then, the estimated impact is added to the amount of exports of African countries. The estimates capture the average TA impact, including both the potential reduction of tariff and non-tariff barriers. The predictions, thus, also take these two aspects of lowering trade costs into account. The results show that while the average impact is only an increase in exports of 2.45%, for some countries exports are predicted to increase by more than 20%. The predicted impact differs greatly per country based on how much of its exports are within Africa and whether it has already joined TAs together with its main African trade partners.
3
The gravity model is widely used in the trade literature to estimate the impacts of TAs. Predicting the impact of a newly created TA, especially with developing countries, has not been done often before (Kepaptsoglou, Karlaftis and Tsamboulas, 2010). Two papers that are close to this paper in their research question and used methods are: Foroutan and Pritchett (1993), who have used estimated values of the gravity model to predict the amount of trade between African countries to see whether trade in Africa is actually lower than it should be. Another paper by Brakman, Garretsen and Kohl (2017) estimates the impact of leaving the European Union on trade for the United Kingdom. They have used a gravity model switching the estimated effect of the EU as TA off. I do the opposite, switching the estimated AfCFTA impact on for African countries.
The paper builds up as follows: First, I will discuss the theoretical mechanisms behind TAs, focussing especially on developing countries, and the working of the gravity model. Second, I discuss the data that is used for the empirical analysis. Third, the empirical method and the results are discussed. Finally, I end with the conclusion.
Literature Review
This section consists of a short review of the current state of the literature on the effect of trade agreements, focusing on developing countries, and the gravity model. The concepts theorized by the previous literature form the basis for the set-up of this study and the empirical model that is used. The first section consists of a discussion of the mechanisms in which TAs affect the trade of member states as theorized by the previous literature.
Thereafter, it will be discussed how the gravity model is used in previous literature to estimate the impact of TAs for developing countries, including some previous empirical findings. The chapter is concluded by discussing the contributions of this paper.
How do TAs affect trade?
This paper focuses on the impact TAs have on the amount of bilateral trade of the member states. For this study it goes too far to also look at the impact TAs have on the outside countries or the relationship between member countries and outside countries. The impact of TAs on countries that are not involved in them is not explicitly considered.
Most of the literature concludes that the overall impacts of TAs on member’s bilateral trade are positive (Baier and Bergstrand, 2007; Carrère, 2006; Kohl, 2014; Ghosh and
Yamari, 2004a). The simple reasoning is that by concluding TAs members agree to reduce or remove tariffs and other trade barriers, which lowers the costs for trade. Lower trade costs result in more trade. There are, however, complicating theoretical mechanisms that may diminish or possibly nullify the positive impact of the reduction in tariffs.
4
products they trade with outside countries and other developing TA member countries are very different. African countries mostly trade manufacturing products within Africa, while they export primary products to the rest of the world (Abrego et al., 2019). Moreover, I make the assumption that the AfCFTA will not lead to an increase in external tariffs.
Consequently, the assumption is made that the trade diversion effect of the AfCFTA will be limited. This allows me to use a partial equilibrium model. This model only considers the increase in trade between member countries on the basis of the current trade patterns and does not take into account the changes in trade patterns with the outside world or within the TA. Furthermore, the predicted impact of the AfCFTA is based on the estimates of the impact of TA membership on the total average amount of trade. Because these capture the total increase in trade the trade diversion effect is also accounted for. So, the predicted impact of the AfCFTA accounts for it as well.
Most of the literature on TAs focuses on developed countries. There can, however, be significant differences between the impact of TAs on developed and developing countries (Brada and Mendez, 1983). First, developing countries may have similar trade patterns, and may not be specialized in manufacturing (Collier, 1995). Developing countries often have similar factor endowments, but while developed countries with similar endowments often engage in intra-industry trade, this is less common for developing countries (Havrylyshyn and Civan, 1985). This means that developing countries often produce similar products and are less likely to trade these with each other. Consequently, a TA may have a smaller impact on the trade levels of developing countries. Second, institutions in developing countries may not be strong enough to enforce the provisions in the agreement (Luke and Macleod, 2019). If there is little enforcement member countries may be tempted to keep trade barriers high to protect their industries from competition or raise revenues from tariffs. Third, besides tariffs, infrastructure is a large barrier to trade in many developing countries, especially in Africa (Collier, 1995). The impact of a TA may be limited if other barriers are still in place. Other barriers of trade in developing countries can also consist of violent conflicts, non-convertible currencies or large cultural differences (Foroutan and Pritchett, 1993; Longo and Sekkat, 2004).
Another factor that can complicate the effect of TAs, and that should ideally be taken into account when studying them, is that they can differ in substance. There is an obvious difference between Free Trade Agreements (FTA) Customs Unions (CU) and Internal Markets (IM). There are also hybrid forms that would be placed somewhere in between and TAs that do not go so far as to be a FTA. Different types of TAs are expected to impact trade levels differently (Kohl, 2014; Kohl, Brakman and Garretsen, 2015; Dür, Baccini and Elsig, 2013). The studies by Kohl (2014) and Kohl et al. (2015) use an index to account for the TA heterogeneity. However, at this point there is not enough data available to account for TA heterogeneity when studying the AfCFTA. This paper will, therefore, not account for TA heterogeneity and use a general TA dummy.
5
forms of trade protection themselves (Krueger, 1993). When rules of origin are heavily used as means of trade protection, the impact of the TA will be limited.
Whether TAs are also welfare increasing is a different story. The welfare effect also depends on the relative importance of trade creation and diversion, as first described by Viner (1950). Trade diversion has negative welfare implications when trade shifts from more
efficient producers in outside countries to less efficient producers with the TA that have become relatively cheaper because of the lower tariffs. See Panagariya (2000) for a discussion of these mechanisms and their welfare effects. In this paper the focus will, however, be on trade and not on welfare. On the other hand, because welfare is what matters for development and not trade per se, it is important to keep these concepts in mind in every discussion about TAs.
The gravity model for developing countries
The gravity model, developed by Tinbergen (1962), has been the most widely used and successful model to estimate and predict trade flows. The gravity model has often been used to estimate the effects of TAs on the amount of trade of members (Baier and Bergstrand, 2007; Carrère, 2006; Soloaga and Winters, 2001). The model has been used in many different forms and estimated with different empirical methods. The basis, however, is a function with the amount of trade as the dependent variable and the size of the trading partners, often GDP for countries, and distance measures to proxy trade costs as the explanatory variables. Dummy variables that equal one when two countries are in a TA together are often used to study the TA impact (see Bergeijk and Brakman, 2010).
The gravity model has been regularly used to estimate the impact of TAs involving developing countries. Developing country TAs are often studied separately as they are thought to have a different impact than TAs with only developed countries. Also, it is argued by some that the gravity model may work differently for developing and developed countries (Foroutan and Pritchett, 2003). For example, distance may be a more important determinant for developing countries because of their lower levels of infrastructure. Trade may also increase non-linearly with GDP because richer countries have more intra-industry trade than poor countries. Zero trade values are also more common for developing countries, reducing the options to estimate the gravity model without bias. There is no literature that suggests that there are large differences in the working of the gravity model for Africa compared to other developing countries. Some scholars stress an even greater importance of infrastructure for Africa. They argue that infrastructure variables should be added to the gravity model as an additional measure for distance (Carrère, 2004).
Different forms of extended gravity models have been used to study the impacts of developing country TAs, such as: the Asian TAs (Ekanayake and Mukherjee, 2010), the TAs in Central and South America (Croce, Juan-Ramón, and Zhu, 2004), and African TAs
(Carrère, 2004; Musila, 2005). Almost all literature finds that developing country TAs also increase member’s internal trade. The magnitude of the impacts is so different for each study, however, that it would not make sense to discuss individual coefficients here.
6
One of the dummies indicates the common TA membership, and one or more are added for the trade between a member and an outside country. If the dummy for trade with outside countries has a negative coefficient there is trade diversion taking place and, depending on its relative size, it may or may not be more important than the trade creation. There is no real consensus whether trade creation or trade diversion is larger in developing country TAs (Carrère, 2004; Ghosh and Yamarik, 2004b; MacPhee and Sattanayuwat, 2014; Musila, 2005; Ekanayake, Mukherjee, Veeramacheneni, 2010; Freund and Ornela, 2010).
I will now shortly discuss some empirical results of papers that are related to this paper. Foroutan and Pritchett (1993) estimate a gravity model for non-African developing countries and use these estimates to predict the amount of trade for African countries. They compare these predictions with the actual amount of trade and conclude that the share of within Africa trade is actually higher than predicted. The authors do not control for econometric issues, such as zero trade values and the MRT. Brakman et al. (2017) also
estimate the impact of a TA, although for developed countries, and turn this impact off for the United Kingdom to simulate the Brexit. They conclude that the Brexit has large negative consequences on trade for the UK. I use the same methods in this paper to turn the impact of a TA on. Moreover, Brakman et al. also control for the MRT and use PPML estimations.
MacPhee and Sattayanuwat (2014) use a similar gravity model as in this paper, including the MRT and estimating it with PPML to account for zero trade values and other econometric issues. They use dummies for the within TA effect and the rest of the world effect and find large differences in impacts between developing country TAs. Some TAs do not have a significant impact and others even a negative impact. They do find that most of the already existing African TAs increase trade. However, the average developing country TA impact is not estimated. Cernat (2001) finds that the African TAs are mainly trade creating, and argues that this could be caused by low enough external tariffs or by the fact that the TAs are not sufficiently implemented, or by the elimination of non-tariff trade costs associated with the TA. Cernat does, however, not account for either the MRT or zero trade values. Abrego et al. (2019) estimate an average welfare increase from the AfCFTA between 2 and 4 percent using a general equilibrium model. They do not estimate the increase in trade.
While most of the previous literature concludes that developing country TAs have a positive effect on trade of the member countries, it is apparent that the estimated magnitude of the impact differs considerably. Therefore, the estimations of the TA impacts done in this paper will contribute to this issue. Moreover, the paper contributes to the literature by
predicting the impact of a newly created TA between developing countries, and by examining the impact of the AfCFTA, which have both not extensively been done before.
The Data
In this section the data that is used to predict the impact of the AfCFTA on trade in Africa is discussed.
7
country, only data from entirely sovereign countries is used. I look at the average bilateral exports from all developing countries, as defined by the IMF, to all other countries in the world. The developing countries are indicated in the list with included countries in Appendix A. The countries that export in the bilateral trade flow are called the “exporter”, while the receiving country is called the “importer”. This gives a set of 143 developing exporting countries and 192 importing countries, both developing and developed. Empirically, there is no real distinction between export or import data. One country’s exports are another country’s imports. In this paper the change of the exports is analysed because higher levels of exports are thought to stimulate development, while this may not necessarily be the case with higher levels of imports. Stimulating economic development is of great importance for African countries.
For most countries there is data available from the 1990s to 2017. 2017 is the most recent year with a wide coverage and is, therefore, used in most estimates. For different reasons there is export data lacking for some African countries for 2017. These countries are Chad, Democratic Republic of the Congo, Djibouti, Equatorial Guinea, eSwatini, Ethiopia, Eritrea, Gabon Guinea, Guinea-Bissau, Liberia, Libya, Niger, Rwanda, Somalia and South Sudan. Because there is data available from 2016 for Ethiopia, Niger and Rwanda these countries are included using the data from that year.
The COMTRADE dataset does not include zero trade values. Not including zero trade values will bias the result (Bergeijk and Brakman, 2010). Therefore, the zero trade values were restored for all missing values with the assumption that there was zero trade when the data from a country pair was missing. This makes it unnecessary to use other trade databases that explicitly include zero trade values, such as the BACI by the CEPII. While that database is also thought to be more accurate on trade for African countries (Mitaritonna and Traoré, 2017), it groups Botswana, Lesotho, Namibia, South Africa and eSwatini together as the Southern Africa Customs Union. Because there is only data on the distance variables for the individual countries and because it would be more informative to predict the AfCFTA impact on these countries individually, the BACI database is not suitable for this study.
COMTRADE data is also used by some other similar studies, such as Foroutan and Pritchett (1993) and Carrère (2004).
Second, the trade agreements that are covered in this study are plurilateral Free Trade Agreements (FTAs) or Custom Unions (CU), as defined by the World Trade Organization (WTO), that include at least one developing country2. Some of these agreements also cover services making them Economic Integration Agreements (EIA). Not covered are agreements that are only EIAs or partial scope agreements, as their scope is too limited to be relevant in predicting the AfCFTA impact. TAs between a country and a TA are included as well because for this study signing an agreement with a TA is similar to joining that TA. A total of 31 TAs are covered. Appendix C includes a list with the covered TAs and the years they went into force or stopped being active. TAs are only included in the dataset the year after they went into force. This means that if the TA went into force in May 2017, it will only be included
2 The list with trade agreements was taken from the WTO website:
8
from 2018 onwards. Only TAs that went into force on the 1st of January will be included from the same year onwards.
In fact, a very large share of all countries in the world is a member of at least one plurilateral TA. Some of the largest TAs in terms of member states including mainly developing countries are the ASEAN Free Trade Area, the Caribbean Community and
Common Market (CARICOM), The Pacific Island Countries Trade Agreement (PICTA), and The Pan-Arab Free Trade Area (PAFTA). With all countries that have signed, or intend to do so, the AfCFTA will have 55 member countries, which would make it much larger than any other TA. The list in the appendix makes clear that there are currently already nine regional African TAs in force. These are the Agadir Agreement (AA) and the Pan-Arab Free Trade Area (PAFTA) in Northern Africa, the Common Market for Eastern and Southern Africa (COMESA), the East African Community (EAC), the Economic and Monetary Community of Central Africa (CEMAC), the Economic Community of West African States (ECOWAS), the Southern African Customs Union (SACU), the Southern African Development Community (SADC), and the West African Economic and Monetary Union (UEMOA). A large share of the TAs entered into force in the 1990s or early 2000s. Because trade barriers have already been removed substantially within the current African TAs, they have to be taken into account when estimating the impact of the AfCFTA.
The dataset indicates for each included country its TA memberships. To analyse the impact of TAs on bilateral trade, I create a dummy variable per country pair for each year. This equals one if both the exporter and importer country are a member of the same TA in the particular year. The common TA membership dummy is expected to have a positive effect on exports.
Third, distance measures from the CEPII are used to account for the distance within the gravity model (Mayer and Zignago, 2011). The distance measure that is weighted by the distribution of the population and the border contingency dummy are used to account for geographical distance. Furthermore, the indicators for a colonial relationship, and a common language, as spoken by the population, are included to control for cultural distance. Distance is expected to negatively impact trade. In this case this means that the geographical distance is expected to be negative in the results, while a common border, language or colonial
relationship are expected to have a positive impact on trade.
Finally, including country fixed effects to account for the multilateral resistance terms (MRT) will control for all country specific variables that do not depend on the country pair. Including the MRT makes it, thus, unnecessary, and in fact impossible, to include country specific variables such as infrastructure, GDP, or GDP per capita that have been included in other papers. I will discuss the use of the MRT in the next section. The GDP variables with data from the World Bank Development Indicators are included in some of the estimates to compare the different models.
9
Table 1: Summary Statistics (2017)
Variable Description Observation Mean St. Dev. Min Max
Distanceij Population weighted distance from CEPII
20,304 8012.369 4467.823 94.27333 19735.32
Exportsij Exports from the exporting to the
importing country from COMTRADE
20,304 2.68e+08 4.43e+09 0 4.30e+11
GDPi GDP of the
exporting country from WDI
20,304 2.71e+11 1.04e+12 2.19e+08 1.01e+13
GDPj GDP of the
importing country from WDI
19,548 4.33e+11 1.62e+12 4.08e+07 1.73e+13
Borderij Common border dummy from CEPII 20,304 0.018814 0.1358711 0 1 Languageij Common language dummy from CEPII 20,304 0.1449468 0.3520559 0 1 Colonyij Common colonial relationship dummy from CEPII 20,304 0.0077817 0.0878723 0 1
Summary statistics for the variables used in the paper with the data of 2017.
Empirical method and results
10
common border and a colonial relationship, and the GDP of both the exporting and importing country. Including the error term, this gives the following model:
1. ln(export)ij = β0 + β1lnGDPi + β2lnGDPj + β3TAij + β4lnDistanceij + β5Borderij + β6Languageij + β7Colony + eij
Many recent trade studies include multilateral resistance terms (MRT) to account for the differences in trade of the rest of the world caused by the trade agreement. I follow Redding and Venalbes (2004) and estimate the MRT using exporter and importer fixed effects. Adding MRT gives the following model:
2. ln(export)ij = DiX + DjM+ β3TAij + β4lnDistanceij + β5Borderij + β6Languageij + β7Colony + eij
In model 2 the GDP variables drop out because of collinearity. Because GDP is static for the countries it is also captured by the country fixed effects. Others (Anderson and Wincoop, 2003; and Bergeijk and Brakman, 2010 for a discussion) use a more complicated estimation method in which the static country specific variables can be kept. For the predictions I do not need the static variables, however, since I would only need the impact of the TA dummies. Therefore, using the simple method of the country fixed effects to account for MRT is
sufficient. I will estimate the model both with and without the MRTs in the next section to see which one suits best.
Estimation
When estimating the models above (1 and 2), there is choice between using OLS or PPML as the estimation method. With OLS it is possible to estimate the gravity model using panel data. The advantage is that using panel data will account for TA endogeneity (Baier and
Bergstrand, 2007). The PPML method is used by many recent papers (Brakman et al., 2017; MacPhee and Sattayanuwat, 2014). This method accounts for the zero trade values that occur in the data. To estimate with OLS the logarithmic values of the exports have to be taken which drops out all the zero values. With PPML the zero values will be used for the
estimation. Not taking into account zero trade values will bias the estimates, especially when they are common. On the other hand, PPML estimation is not compatible with time series, which means I will have to do the estimation using cross-sectional data or pool the data over more years. Because using the most recent data gives the relevant information on what the impact would be now, and because I want to compare the impact in different years, I estimate the model using cross-sections.
11
different impacts, including them as exporter countries would bias the results. As explained above, to get even closer to the expected AfCFTA impact, I divide the common TA dummy in a poor-rich TA and a poor-poor TA dummy.
To see which method and model suits the data best, I will estimate the different scenarios as described above. I first estimate the models, with and without the MRT, with OLS and PPML for all developing countries for the latest year there is sufficient data available, which is 2017.
I estimate the model in different ways to see which method would be the best to use for the prediction. The results are presented in table 2. In the first column I estimate the model using OLS without the MRT. The MRT are included in the second column. As a result, the GDP variables are only estimated in the first column. All variables are significant with both models, and all coefficients have the expected sign. However, the coefficients for the common TA dummy and the distance dummies are somewhat smaller when the MRT are included, whereas the geographical distance coefficient is larger. The R-squared is also larger when the MRT are included.
Table 2: Comparing the models and estimation methods
(1) (2) (3) (4) (5) (6) OLS OLS PPML PPML PPML PPML Common TAij 1.0206*** 0.9484*** 0.6331*** 0.6038*** 0.0689 0.0654 0.0945 0.0811 Poor-Rich TAij 0.5767*** 0.7358*** 0.1139 0.0953 Poor-Poor TAij 0.7164*** 0.4693*** 0.1291 0.1102 Distanceij −1.3361*** −1.8386*** −0.6615*** −0.9012*** −0.6627*** −0.9127*** 0.0366 0.0360 0.0730 0.0528 0.0735 0.0515 Borderij 1.1302*** 0.7780*** 0.6130*** 0.3087*** 0.6040*** 0.2960*** 0.1577 0.1338 0.1478 0.0984 0.1498 0.0983 Languageij 1.0410*** 0.9959*** 0.4346*** 0.2396** 0.4250*** 0.2706** 0.0705 0.0703 0.0926 0.1052 0.1005 0.1077 Colonyij 1.2489*** 0.8918*** 0.1140 0.5448*** 0.1335 0.5323*** 0.2307 0.2088 0.1355 0.1036 0.1433 0.1032 GDPi 1.3095*** 1.0338*** 1.03110*** 0.0135 0.0415 0.0403 GDPj 0.8851*** 0.8314*** 0.8387*** 0.0117 0.0433 0.0460 Observations 11,623 11,959 19,548 20,116 19,548 20,116 R² 0.5567 0.7033 0.7120 0.9271 0.7109 0.9283
The dependent variable is the exports from the exporting country (i) to the importing country (j). The
12
In Column three and four the model is estimated using PPML without and with the MRT respectively. All variables are significant again, except for the colony dummy in column three. The coefficients are much smaller than with the OLS estimation, and again the coefficients of the common TA dummy and distance dummies are even smaller when the MRT are included and geographical distance is larger. Also here, the R-squared is much larger when the MRT are included.
Moreover, it stands out that the amount of observations is much larger when PPML is used compared to the OLS estimates in column 1 and 2 (11,959 compared with 20,116 when MRT is added). This indicates that there are indeed many zero trade values in the data and that omitting them would bias the results to a large extent. This means that PPML is the best method of estimation.
As pointed out, using the PPML method means the models have to be estimated using cross-sectional data. Therefore, I cannot use panel data to account for TA endogeneity. Using panel data with country pair fixed effects as Baier and Bergstrand (2007) did to account for TA endogeneity, would either way have been problematic with this dataset. Many TAs are already in force for a long time. This means that the common TA dummy variables do not change over time for many country pairs and the TA impact will also be captured by the country pair fixed effects. As a result, only the TAs that have entered into force after the first included year can be estimated, which would only leave a small sub sample of countries. Moreover, estimating with MRT in a panel setting is problematic because the MRT are expected to change over time (Brakman et al., 2017). Finally, using cross-section allows me to compare the impact of TAs across time.
In column 5 and 6 I estimate the poor-rich and poor-poor TA dummies without and with the MRT respectively. The same differences between the two models are apparent again. The TA and distance dummies are smaller, while geographical distance is larger. Again, the R-squared is much larger when including the MRT. That the R-squared is larger when the MRT is included indicates that this model better explains the data. Moreover, I do not need the GDP coefficients that are omitted for the prediction. Therefore, I use the model including MRT for further estimates, including the estimates that are used for predicting the impact of the AfCFTA.
13
Table 3: Estimation for African and non-African developing
countries (1) (2) (3) All Countries African Non-African Poor-Rich TAij 0.7358*** 0.5120* 0.8043*** 0.0953 0.2694 0.0984 Poor-Poor TAij 0.4693*** 0.9523*** 0.4810*** 0.1102 0.1936 0.1172 Distanceij −0.9127*** −1.0391*** −0.8878*** 0.0515 0.1271 0.0544 Borderij 0.2960*** 0.5866*** 0.3133*** 0.0983 0.2261 0.0947 Languageij 0.2706** 0.2756 0.2466** 0.1077 0.1982 0.1148 Colonyij 0.5323*** 0.5881*** 0.5151*** 0.1032 0.2134 0.1098 Observations 20,116 6,882 13,160 R² 0.9283 0.5928 0.9413
The dependent variable is the exports from the exporting country (i) to the importing country (j). The independent variables are listed in the rows. All columns are estimated using PML and include the MRT. Column 1 is the same as column 6 in table 2. The table reports the estimated beta coefficients and the corresponding standard errors. *, ** and *** indicate significance at the 10%, 5% and 1% level respectively.
14
Figure 1: The graph shows the estimated beta coefficient for the poor-poor and poor-rich TA dummies from 2000 to 2017 with data from all developing countries. The results corresponding with the graph are presented in Appendix B.
Figure 2: The graph shows the estimated beta coefficient for the poor-poor and poor-rich TA dummies from 2000 to 2017 with data from the African countries. The results corresponding with the graph are presented in Appendix B.
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 20 17 20 16 20 15 20 14 20 13 20 12 20 11 20 10 20 09 20 08 20 07 20 06 20 05 20 04 20 03 20 02 20 01 20 00
Estimated TA impact: All developing countries
Poor-Poor TAs Poor-Rich TAs -0,5 0 0,5 1 1,5 2 20 17 20 16 20 15 20 14 20 13 20 12 20 11 20 10 20 09 20 08 20 07 20 06 20 05 20 04 20 03 20 02 20 01 20 00
Estimated TA impact: African countries
15
Figure 3: The graph shows the estimated beta coefficient for the poor-poor and poor-rich TA dummies from 2000 to 2017 with data from all non-African developing countries. The results corresponding with the graph are presented in Appendix B.
Prediction
In this section the method to predict the impact of the AfCFTA on trade for African countries will be discussed and the results will be presented. In Table 3 the estimated average increase in exports for poor-poor TA country pairs was presented. Because the AfCFTA will only involve African developing countries the estimated impact of the TAs only involving non-African developing countries (column 3 in table 3) is used to turn the AfCFTA impact on. It can be argued that it is more relevant to predict the AfCFTA impact based only on developing country TAs in other regions than using the estimations with the already existing African TAs included as well. Moreover, as we see in the table, the difference between the coefficients including and excluding the African countries is rather small. For some countries of which there is no data from 2017 there is data available from 2016. For these countries I will use the data from 2016 for the predictions and this will be clearly indicated in the results.3
The estimated poor-poor TA impact is added to the bilateral exports of African
countries that will become both a member of the AfCFTA. The estimated impact is not added when the countries are already both a member of the same already existing African TA. For now I assume that all African countries will become a member of the AfCFTA. Later I will also predict the AfCFTA impact when only the countries that currently ratified the agreement are a member and compare the difference. The current average amount of exports together with the estimated average TA impact gives the predicted average amount of exports for the African countries when the AfCFTA is in force.
Because the estimated TA impact is a logarithmic value it is added to the logarithmic value of the actual export. In the logarithmic values the zero trade values are dropped, however. This means that the increase in exports is only added when the countries already
3 As discussed in the data section, these countries are Ethiopia, Niger and Rwanda. 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 20 17 20 16 20 15 20 14 20 13 20 12 20 11 20 10 20 09 20 08 20 07 20 06 20 05 20 04 20 03 20 02 20 01 20 00
Estimated TA impact: Non-African developing countries
16
trade with each other. This is not an unrealistic assumption, because it may be argued that countries pairs that have no trade at all will probably not suddenly start trading when they are both in the same TA. Moreover, adding the estimated TA impact to the zero trade values would highly bias the predictions because there will be no zero trade values in Africa anymore as these will now all have the value of the TA impact. This would result in unrealistically high average increases in exports.
The results of this prediction are presented in table 4 and Figure 4. The predicted export values with the AfCFTA impact are compared with the actual average exports. To have a relevant comparison the zero trade values are not taken into account in the actual average export values either.
Figure 4: The Map shows the AfCFTA impact in percentages on the
17
Table 4: Predicted AfCFTA impact (2017)
Country Actual mean
Exports Predicted TA Exports TA impact TA impact (%) Algeria 292.64 293.43 0.79 0.27% Angola 266.06 267.29 1.24 0.47% Benin 7.70 7.93 0.24 3.08% Botswana 61.91 61.93 0.01 0.02% Burkina Faso 37.20 37.49 0.29 0.79% Burundi 2.67 2.67 0.01 0.28% Cameroon 27.65 28.41 0.76 2.74% Cape Verde 3.56 3.58 0.01 0.34%
Central African Republic 5.60 5.95 0.35 6.30%
Congo 80.66 89.87 9.21 11.41% Côte d'Ivoire 87.65 90.40 2.75 3.13% Egypt 165.18 168.26 3.08 1.87% Ethiopia (2016) 12.90 13.14 0.24 1.87% Gambia 0.54 0.55 0.00 0.88% Ghana 107.95 112.42 4.46 4.14% Kenya 35.82 36.17 0.35 0.97% Lesotho 11.20 11.21 0.01 0.09% Madagascar 22.47 23.00 0.53 2.36% Malawi 8.71 8.73 0.03 0.33% Mali 22.64 28.57 5.93 26.18% Mauritania 25.14 26.19 1.04 4.15% Mauritius 15.81 15.84 0.03 0.18% Morocco 162.99 170.21 7.22 4.43% Mozambique 37.81 37.93 0.12 0.31% Namibia 36.38 36.60 0.21 0.59% Niger (2016) 11.43 11.46 0.03 0.28% Nigeria 404.99 420.76 15.77 3.89% Rwanda (2016) 5.68 5.70 0.02 0.28%
São Tomé and Principe 0.48 0.49 0.01 1.73%
Senegal 22.26 23.15 0.89 3.99% Seychelles 7.59 7.60 0.02 0.22% Sierra Leone 2.35 2.39 0.04 1.68% South Africa 430.13 440.95 10.83 2.52% Sudan 48.99 49.04 0.05 0.11% Tanzania 30.60 31.14 0.54 1.75% Togo 9.55 9.70 0.15 1.52% Tunisia 98.41 99.74 1.32 1.34% Uganda 20.01 20.33 0.32 1.58% Zambia 77.92 77.96 0.04 0.05% Zimbabwe 35.16 35.16 0.00 0.01% Averages 68.61 70.33 1.72 2.45% Total 2744.39 2813.32 68.93 2.51%
18
Table 5: Within Africa exports shares
Country Share of within Africa exports Share of exports to African countries not in a common TA Share of exports to South Africa Share of exports to Nigeria Share of exports to Egypt Algeria 5.27% 8.25% 0.02% 0.02% 24.69% Angola 4.86% 15.49% 83.72% 0.41% 1.19% Benin 24.69% 20.22% 2.76% 39.91% 12.27% Botswana 13.42% 0.27% 70.80% 0.01% 0.04% Burkina Faso 13.11% 9.76% 7.71% 0.02% 0.00% Burundi 16.58% 2.76% 0.62% 0.00% 28.17% Cameroon 11.38% 39.03% 1.59% 10.20% 0.38% Cape Verde 0.63% 87.84% 0.00% 0.00% 0.00%
Central African Republic 15.77% 64.72% 0.00% 0.16% 0.00%
Congo 26.20% 70.53% 0.03% 6.13% 1.57% Côte d'Ivoire 22.92% 22.12% 13.61% 5.13% 0.39% Egypt 14.23% 21.24% 4.64% 2.68% 0.00% Ethiopia (2016) 12.47% 31.00% 6.30% 2.40% 2.04% Gambia 40.75% 3.48% 1.31% 0.09% 0.00% Ghana 14.29% 46.85% 44.71% 3.10% 0.55% Kenya 33.37% 4.69% 1.47% 1.41% 10.12% Lesotho 51.99% 0.28% 94.80% 0.02% 0.00% Madagascar 7.68% 49.81% 37.05% 0.27% 0.71% Malawi 33.43% 1.58% 23.47% 0.04% 16.68% Mali 60.10% 70.51% 68.26% 0.03% 0.00% Mauritania 6.71% 100.00% 2.86% 28.93% 0.76% Mauritius 21.35% 1.39% 42.48% 0.07% 0.04% Morocco 8.94% 80.21% 1.80% 7.37% 2.67% Mozambique 22.62% 2.23% 84.70% 0.07% 0.01% Namibia 44.56% 2.14% 53.66% 0.66% 0.01% Niger (2016) 20.32% 2.89% 0.01% 46.65% 0.79% Nigeria 12.15% 51.87% 37.29% 0.00% 4.21% Rwanda (2016) 27.51% 2.08% 1.58% 0.01% 0.01%
São Tomé and Principe 2.80% 100.00% 1.84% 20.44% 0.00%
Senegal 47.17% 13.70% 0.35% 1.67% 0.03% Seychelles 5.11% 7.05% 36.81% 0.00% 0.02% Sierra Leone 34.02% 7.98% 3.15% 1.68% 0.63% South Africa 28.22% 14.44% 0.00% 1.92% 0.80% Sudan 15.80% 1.11% 0.06% 0.47% 64.25% Tanzania 33.84% 8.39% 50.07% 0.27% 0.10% Togo 68.08% 3.62% 0.05% 7.08% 0.04% Tunisia 10.79% 20.17% 0.99% 0.59% 2.90% Uganda 39.61% 6.47% 0.88% 0.09% 0.25% Zambia 12.89% 0.61% 45.18% 0.10% 0.00% Zimbabwe 90.10% 0.03% 81.97% 0.00% 0.02% Averages 24.39% 24.92% 22.72% 4.75% 4.41%
19
As expected, because the estimated TA impact is positive, average bilateral exports increase for all African countries. The average increase in exports for African countries is predicted to be 2.45%4. There are, however, large differences between countries. The increase in export ranges from 26% for Mali to only 0.01% for Zimbabwe. African countries that have a relatively large share of intercontinental trade without being already in a TA with many of their main trade partners are expected to benefit the most from the AfCFTA. This is also what we see in the results. To be able to compare the predicted trade increase from the AfCFTA with the share of within Africa trade, these shares are calculated for every African country, presented in table 5 and illustrated in Figure 5.
Figure 5: The map illustrates the exports within Africa as a share of total exports for each African country. The map is based on the export shares in column 1 of table 5. The Islands states of Mauritius, São Tomé and Principe, and the Seychelles are not included on the map because of their small size.
20
Figure 6: The map illustrates the exports that are going to African countries not in a common TA as a share of the total exports within Africa for that country. The map is based on column 2 of table 4. The Islands states of Mauritius, São Tomé and Principe, and the Seychelles are not included on the map because of their small size.
A country such as Mali has a very large share of within Africa trade, and is thus benefiting greatly from a new TA encompassing all the African countries. Congo and Senegal are further examples of countries that are trading relatively much within Africa and would benefit above average from the AfCFTA. On the other hand, countries such as Zimbabwe, Mozambique and Lesotho are benefiting very little, while having relatively high shares of within Africa trade. These countries are already in a TA with their main trade partner South Africa and will not benefit much from tariff reductions with other African countries. Therefore, to see which countries would benefit from the AfCFTA it is also necessary to look at the shares of exports going to African countries that are not already in a common TA. These shares are presented in table 5 in column 2 and illustrated in Figure 6. The shares of within Africa exports and within Africa exports to countries not in a common TA together explain the predicted impact of the AfCFTA. For example, Togo has a very large share of within Africa exports but is already in a TA with most of its main trade partners. As a result, Togo’s predicted increase in exports is below average. The same is true for Gambia, Kenya, Mauritius, Namibia, Rwanda, Sierra Leone, Tanzania, Uganda, and the earlier mentioned Lesotho, Mozambique and Zimbabwe.
21
exports. Countries that are mostly only exporting primary products export relatively much outside of Africa, while countries that export a combination of primary and manufacturing products, export more within Africa (Luke and MacLeod, 2019; UNCTAD, 2019). For
example, a country as Angola, of which a large share of the exports is taken up by oil exports, has a very small share of within Africa exports of only 4.86%. The UNCTAD (2019) finds that Tunisia, Egypt, South Africa, Mauritius, Uganda, Namibia, Mali, Morocco, Senegal and Kenya have the most diversified exports. With the exception of the North African countries5, these countries all have indeed relatively high shares of within Africa exports.
It has to be kept in mind that the predictions are based on a proportional increase in the current export patterns of the African countries. It may, however, be expected that there will be a structural change in the export patterns. Countries that currently export relatively little within Africa may develop new sectors and join production chains that involve more within Africa trade (Luke and MacLeod, 2019; UNCTAD, 2019). If these changes are taken into account the AfCFTA impact may be rather different and much larger, but also very hard to predict.
Different scenarios
In this section some different scenarios are performed to see whether other choices in the assumptions made would greatly alter the results. The first other scenario is that not all African countries will join the AfCFTA. Currently all African countries have signed the agreement, or in the case of Eritrea intend to do so. However, only half of them have ratified it so far. Therefore, I predict the impact of the AfCFTA in the scenario that only the countries that have currently ratified the agreement join the TA (Figure 7). This gives the results in column 1 of table 6, illustrated in Figure 8. It goes without saying that exports are only predicted to increase for the countries that have ratified the agreement. The increases are, however, lower than when all African countries join, because these countries have fewer countries they can export to with the lower trade barriers provided by the AfCFTA. The average increase in exports for the countries that have ratified is 2.88% when the AfCFTA will only involve these countries. The same countries would increase their exports on average by 4.31% when all African countries would join. Total within African exports are predicted to increase by 1.07%, and when all African join by 2.51%.
22
Figure 7: The map indicates which countries have ratified, only signed, or not yet signed the AfCFTA. As of January 1st, 2020 these are: Ratified: Burkina Faso, Chad, Republic of the Congo, Ivory Coast, Djibouti, Egypt, Equatorial Guinea, eSwatini, Ethiopia, Gabon, Gambia, Ghana, Guinea, Kenya, Mali Mauritania, Mauritius, Namibia, Niger, Rwanda, São Tomé and Principe, Senegal, Sierra Leone, South Africa, Togo, Uganda, Zimbabwe. Signed: Algeria, Angola, Benin, Botswana, Burundi, Cameroon, Cape Verde, Central African Republic, Comoros, Democratic Republic of the Congo, Guinea-Bissau, Lesotho, Libya, Liberia, Madagascar, Malawi, Morocco, Mozambique, Nigeria, Seychelles, Somalia, South Sudan, Sudan, Tanzania, Tunisia, Zambia. Not Signed: Eritrea. The Islands states of Mauritius, São Tomé and Principe, and the Seychelles are not included on the map because of their small size.
23
Table 6: Other scenarios
Country Ratified countries 2016 2000
Algeria 0.00% 0.15% 1.29% Angola 0.00% 0.18% n/a Benin 0.00% 2.75% 4.27% Botswana 0.00% 0.02% 10.15% Burkina Faso 0.73% 1.90% 0.32% Burundi 0.00% 0.62% 0.11% Cameroon 0.00% 3.08% 2.91% Cape Verde 0.00% 0.21% 1.63%
Central African Republic 0.00% 0.38% 0.36%
Comoros n/a n/a 0.11%
Congo 5.94% n/a n/a
Côte d'Ivoire 2.44% 2.54% 4.44%
Egypt 1.37% 1.40% 0.84%
Ethiopia n/a 1.87% 0.28%
Gabon n/a n/a 1.70%
Gambia 0.52% 0.08% 9.76%
Ghana 4.02% 1.77% 1.91%
Guinea n/a n/a 4.12%
Kenya 0.44% n/a 1.21% Lesotho 0.00% n/a 24.79% Madagascar 0.00% 2.32% 0.60% Malawi 0.00% 0.32% 11.21% Mali 25.82% 23.22% 34.37% Mauritania 2.66% 4.95% 11.12% Mauritius 0.07% 0.13% 0.86% Morocco 0.00% 3.33% 1.56% Mozambique 0.00% 0.25% 41.04% Namibia 0.34% 0.22% 30.71% Niger n/a 0.28% 0.90% Nigeria 0.00% 3.86% 1.77%
Rwanda n/a 0.28% n/a
São Tomé and Principe 0.12% 6.15% 5.45%
Senegal 3.09% 3.81% 5.89% Seychelles 0.00% 0.05% 1.71% Sierra Leone 1.22% 0.12% 0.00% South Africa 1.76% 2.11% 16.16% Sudan 0.00% n/a 1.97% Tanzania 0.00% 2.08% 8.32% Togo 0.97% 1.79% 1.64% Tunisia 0.00% 0.85% 7.64% Uganda 0.27% 0.81% 18.19% Zambia 0.00% n/a 5.45% Zimbabwe 0.01% 0.01% n/a Averages 1.40% 2.11% 7.14%
24
Figure 8: The Map shows the AfCFTA impact in percentages on the
individual African countries, based on the results in column 1 of Table 6. The Islands states of Mauritius, São Tomé and Principe, and the Seychelles are not included on the map because of their small size.
25
Figure 10: The Map shows the AfCFTA impact in percentages on the individual African countries, based on the results in column 3 of Table 6. The Islands states of Mauritius, São Tomé and Principe, and the Seychelles are not included on the map because of their small size.
Conclusion
In this paper I have predicted the impact the entry into force of the AfCFTA will have on the exports of the member countries. I have first estimated the average developing country TA impact with the gravity model using PPML estimation, to account for the zero trade values, and including the MRT. These estimates were added to the current amount of exports of African countries to predict the impact of the AfCFTA.
The results showed that the average increase of exports generated by the AfCFTA is predicted to be 2.45%. This is the increase of exports compared to the current situation with the regional African TAs taken into account. It has to be considered, however, that there is a large variance in the impact on countries. Some countries will be able to increase their exports by more than 25%, while others hardly benefit at all. The extent to which exports are
predicted to increase depends on the share of a country’s exports within Africa and what share of these is exported to countries not within the same already existing African TA.
26
into account the structural changes can be considered a limitation of the paper and would be an interesting topic for future research.
A further issue that has to be considered is that the prediction is based on the average of all other developing country TAs, and does not take TA heterogeneity into account. If the AfCFTA would be more successful than the average developing country TA, it can be expected that trade will be increased more, and vice versa. I have also assumed that all already existing African TAs have the same impact, equal to the average estimated TA impact. In reality some are more successful than others. If an African TAs is actually less successful in promoting trade than the AfCFTA will be, the AfCFTA may also have a larger impact than predicted.
A final limitation is that I do not account for TA endogeneity. TAs may also be a result of already existing large trade flows between countries instead of a cause of more trade. Some of the literature accounts for endogeneity problems using country pair fixed effects with panel data, but this is not possible using PPML estimation. Therefore, endogeneity could potentially be an issue.
For future research it would be interesting to see whether the predictions in this paper have become reality ad how large the role of structural changes will be. This research can only be done after the AfCFTA has entered into force and the trade data has become
27 Bibliography
Abrego, L., Amado, M. A., Gursoy, T., Nicholls, G. P., and Perez-Saiz, H. 2019. The African Continental Free Trade Agreement: Welfare Gain Estimates from a General Equilibrium Model. IMF Working Paper, WP/19/124.
Anderson, J. E., and Van Wincoop, E. 2003. Gravity with gravitas: A solution to the border puzzle. American Economic Review, 93(1), 170–192.
Baier, S. L., and Bergstrand, J. H. 2007. Do free trade agreements actually increase members' international trade? Journal of International Economics, 71(1): 72-95.
Bergeijk, P., & Brakman, S. (2010). Introduction. In P. Van Bergeijk & S. Brakman (Eds.),
The Gravity Model in International Trade: Advances and Applications:1-26. Cambridge:
Cambridge University Press.
Brada, J. C., and Mendez, J. 1983. Regional economic integration and the volume of intra-regional trade: a comparison of developed and developing country experience. Kyklos, 36(4): 589-603.
Brakman, S., Garretsen, H., and Kohl, T. 2017. Consequences of Brexit and options for a ‘Global Britain’. Regional Science, 97(1): 55-72.
Carrère, C. 2004. African Regional Agreements: Impact on Trade with or without Currency Unions. Journal of African Economies, 13(2): 199–239.
Carrère, C. 2006. Revisiting the effects of regional trade agreements on trade flows with proper specification of the gravity model. European Economic Review, 50(2): 223-247. Cernat, L. 2001. Assessing Regional Trade Arrangements: Are South-South RTAs More
Trade Diverting? Global Economy Quartely, 2: 235-260.
Collier, P. 1995. The Marginalization of Africa. International Labour Review, 134(4/5): 541-557.
Croce, E., Juan-Ramón, V. H., and Zhu, F. 2004. Performance of Western Hemisphere Trading Blocs: A Cost-Corrected Gravity Approach. IMF Working Paper, WP/04/109. Dür, A., Baccini, L., and Elsig, Manfred. 2014. The Design of International Trade
28
Ekanayake, E. M., Mukherjee, A., Veeramacheneni, B. 2010. Trade Blocks and the Gravity Model: A Study of Economic Integration among Asian Developing Countries. Journal of
Economic Integration, 25(4): 627-643.
Foroutan, F., and Pritchett, L. 1993. Intra-Sub-Saharan African Trade: is it too Little? Journal
of African Economies, 2(1): 74–105.
Freund, C. and Ornelas, E. 2010. Regional Trade Agreements. Annual Review of Economics, 2: 139-166.
Ghosh, S., and Yamarik, S. 2004a. Does trade creation measure up? A reexamination of the effects of regional trading arrangements. Economics Letters, 82(2): 213-219.
Ghosh, S., and Yamarik, S. 2004b. Are regional trading arrangements trade creating?: An application of extreme bounds analysis. Journal of International Economics, 63(2): 369-395.
Havrylyshyn, O., and Civan, E. 1985. Intra-industry trade among developing countries.
Journal of Development Economics, 18(23): 253-271.
Kepaptsoglou, K., Karlaftis, M. G., and Tsamboulas, D. 2010. The Gravity Model
Specification for Modelling International Trade Flows and Free Trade Agreement Effects: A 10-Year Review of Empirical Studies. The Open Economics Journal, 3: 1-13.
Kohl, T. 2014. Do we really know that trade agreements increase trade? Review of World
Economics, 150(3): 443-469.
Kohl, T., Brakman, S., and Garretsen, H. 2015. Do Trade Agreements Stimulate International Trade Differently? Evidence from 296 Trade Agreements. The World Economy, 39(1): 97-131.
Krueger, A. O. 1993. Free Trade Agreements as Protectionist Devices: Rules of Origin.
Working Paper No. 4352.
Longo, R., and Sekkat, K. 2004. Economic Obstacles to Expanding Intra-African Trade.
World Development, 32(8): 1309-1321.
Luke, D. and MacLeod, J. (Eds.). 2019. Inclusive Trade in Africa: The African Continental
Free Trade Area in Comparative Perspective. Oxon: Routledge.
29
Mayer, T. and Zignago, S. 2011. Notes on CEPII’s Distances Measures: The GeoDist Database, CEPII Working Paper 2011-25.
Mitaritonna, C. and Traoré, F. 2017. Existing Data to Measure African Trade. International
Food Policy Research Institute, Discussion Paper 01618.
Musila, J. W. 2005. The Intensity of Trade Creation and Trade Diversion in COMESA, ECCAS and ECOWAS: A comparative Analysis. Journal of African Economies, 14(1): 117-141.
Panagariya, A. 2002. Preferential Trade Liberalization: The Traditional Theory and New Developments. Journal of Economic Literature, 38(2): 287-331.
Redding, S., and Venables, A. J. 2004. Economic Geography and International Inequality.
Journal of International Economics, 62:53-82.
Soloaga, I., and Winters, L. A. 2001. Regionalism in the Nineties: What Effect on Trade?
North American Journal of Economics and Finance, 12: 1-29.
Tinbergen, J. 1962. Shaping the World Economy: Suggesteions for an International
Economic Policy. New York: Twentieth Century Fund.
Viner, J. 1950. The Customs Union Issue. New York: Carnegie Endowment for International Peace.
United Nations Conference on Trade and Development (UNCTAD). 2019. Economic
Development in Africa Report 2019: Made in Africa - Rules of Origin for Enhanced Intra-African Trade.
Yang, Y., and Gupta, S. 2008. Regional Trade Arrangements in Africa: Past Performance and the Way Forward. African Development Review, 19(3): 399-431.
WTO. Regional Trade Agreements Database.
30 Appendix A: Included countries.
In bold: exporter/developing countries
All: importer countries, developing and developed
Afghanistan, Albania, Algeria, Andorra, Angola, Antigua and Barbuda, Argentina, Armenia,
Australia, Austria, Azerbaijan, Bahamas, Bahrain, Bangladesh, Barbados, Belarus, Belgium,
Belize, Benin, Bhutan, Bolivia, Bosnia and Herzegovina, Botswana, Brazil, Brunei Darussalam, Bulgaria, Burkina Faso, Burundi, Cambodia, Cameroon, Canada, Cape Verde, Central African Republic, Chad, Chile, China, Colombia, Comoros, Congo, Costa Rica, Croatia, Cuba, Cyprus, Czechia, Côte d'Ivoire, Democratic People's Republic of Korea, Democratic Republic of the Congo, Denmark, Djibouti, Dominica, Dominican Republic, East Timor, Ecuador, Egypt, El Salvador, Equatorial Guinea, Eritrea, Estonia, Ethiopia, FYR Micronesia, Fiji, Finland, France, Gabon, Gambia, Georgia, Germany, Ghana, Greece, Grenada, Guatemala, Guinea, Guinea-Bissau, Guyana, Haiti, Honduras, Hungary, Iceland, India, Indonesia, Iran, Iraq, Ireland, Israel,
Italy, Jamaica, Japan, Jordan, Kazakhstan, Kenya, Kiribati, Kuwait, Kyrgyzstan, Lao People's
Democratic Republic, Latvia, Lebanon, Lesotho, Liberia, Libya, Lithuania, Luxembourg,
Madagascar, Malawi, Malaysia, Maldives, Mali, Malta, Marshall Islands, Mauritania, Mauritius, Mexico, Moldova, Mongolia, Morocco, Mozambique, Myanmar, Namibia, Nauru, Nepal,
Netherlands, New Zealand, Nicaragua, Niger, Nigeria, North Macedonia, Norway, Oman,
Pakistan, Palau, Panama, Papua New Guinea, Paraguay, Peru, Philippines, Poland, Portugal, Qatar, Republic of Korea, Romania, Russian Federation, Rwanda, Saint Kitts and Nevis, Saint Lucia, Saint Vincent and the Grenadines, Samoa, San Marino, São Tomé and Principe, Saudi
Arabia, Senegal, Serbia, Seychelles, Sierra Leone, Singapore, Slovakia, Slovenia, Solomon Islands,
Somalia, South Africa, Spain, Sri Lanka, Sudan, Suriname, eSwatini, Sweden, Switzerland, Syria, Tajikistan, Tanzania, Thailand, Togo, Tonga, Trinidad and Tobago, Tunisia, Turkey,
31
Appendix B: TA estimates for many years, the basis for Figures 1, 2 and 3.
Table B1: Estimates for many years
All Countries Non-African African
Year Poor -Poor Poor - Rich Poor - Poor Poor - Rich Poor - Poor Poor - Rich 2017 0.4693 0.7358 0.4810 0.8043 0.9523 0.5120 2016 0.4144 0.7048 0.3946 0.7611 0.9183 0.4004 2015 0.4811 0.7348 0.4684 0.7881 1.1718 0.2500 2014 0.3137 0.6826 0.2973 0.7432 0.9790 0.1554 2013 0.3727 0.7319 0.3739 0.8192 0.7739 0.2715 2012 0.3788 0.7240 0.3989 0.8400 0.7670 0.0436 2011 0.3512 0.7223 0.3698 0.8816 0.6842 0.0932 2010 0.3554 0.6815 0.3808 0.8431 0.5903 −0.0220 2009 0.4048 0.6428 0.4630 0.8217 0.7344 −0.1136 2008 0.3601 0.3906 0.3653 0.5628 1.0920 −0.1507 2007 0.4109 0.4812 0.4522 0.6590 0.9214 −0.0179 2006 0.4040 0.4961 0.3476 0.6050 1.2570 0.3948 2005 0.4917 0.5741 0.4227 0.6338 1.4416 0.2884 2004 0.5448 0.5546 0.4602 0.5682 1.3979 0.2070 2003 0.5631 0.6017 0.4844 0.6433 1.3103 0.5202 2002 0.5057 0.6754 0.4008 0.7068 1.6111 0.4724 2001 0.5437 0.6524 0.4710 0.7482 1.4153 0.4128 2000 0.6239 0.6271 0.6389 0.7853 1.0225 0.5605
32 Appendix C: List of included trade agreements
TA Year of entry into force Countries Other information African Common Market July 1963 - December 1998
Algeria; Ghana; Guinea; Mali; Morocco CU
Agadir Agreement March 2017
Jordan; Morocco; Tunisia; Egypt FTA
Andean Community
May 1988 Bolivia, Plurinational State of; Colombia; Ecuador; Peru; Venezuela
CU Arab Common Market January 1965 - December 1998
Iraq; Jordan; Libya; Mauritania; Syrian Arab Republic; Egypt; Yemen
CU
ASEAN Free Trade Area
January 1993
Brunei Darussalam; Myanmar; Cambodia; Indonesia; Lao People's Democratic Republic; Malaysia; Philippines; Singapore; Viet Nam; Thailand
FTAs with Australia and New Zealand (2010), China (2005), India (2010), Japan (2008), Republic of Korea (2010) FTA Caribbean Community and Common Market (CARICOM) August 1973
Antigua and Barbuda; Bahamas; Barbados; Belize; Dominica; Grenada; Guyana; Haiti; Jamaica; Saint Kitts and Nevis; Saint Lucia; Saint Vincent and the Grenadines; Suriname; Trinidad and Tobago
CU & EIA, a revised version of the TA went into force in 2002 Central American Common Market
June 1961 Costa Rica; El Salvador; Guatemala; Honduras; Nicaragua; Panama (May 2013)
CU Central European Free Trade Agreement March 1993 - May 2004
Bulgaria; Czech Republic; Hungary; Poland; Romania; Slovak Republic; Slovenia (-, 2004)
FTA
Central European Free Trade Agreement (2006)
May 2007 Albania; Bosnia and Herzegovina; Moldova, Republic of; North Macedonia;
UNMIK/Kosovo; Serbia; Croatia( -, 2013);
FTA
Common Economic Zone
May 2004 Belarus; Kazakhstan; Russian Federation; Ukraine
FTA
Common Market for Eastern and Southern Africa
December 1994
Angola (-, 2007); Burundi; Comoros; DR Congo; Djibouti; Egypt (1999,-); Eritrea; ESwatini; Ethiopia; Kenya; Lesotho (-, 1997);
33
(COMESA) Libyan Arab Jamahiriya (2006, -); Madagascar;
Malawi; Mauritius; Rwanda; Seychelles; Somalia; Sudan; Tanzania (-, 2000) Tunisia; Uganda; Zambia; Zimbabwe.
Commonwealth of Independent States
January 2012
Armenia; Belarus; Kazakhstan; Kyrgyz Republic; Moldova, Republic of; Russian Federation; Tajikistan; Ukraine
FTA
East African Community
July 2000 Burundi; Kenya; Rwanda; Uganda; Tanzania; South Sudan; Burundi (2007, -); Rwanda (2007, -) CU & EIA Eurasian Economic Community October 1997 - January 2015
Belarus; Kazakhstan; Kyrgyz Republic; Russian Federation; Tajikistan
CU
European Union January 1958
Belgium; France; Germany; Italy; Luxembourg; Netherlands
Austria(1995, -); Bulgaria(2007, -); Croatia (2013, -); Cyprus(2004, -); Czech
Republic(2004, -); Denmark; Estonia(2004, -); Finland(1995, -); Greece; Hungary(2004, -); Ireland; Latvia(2004, -); Lithuania(2004, -); Malta(2004, -); Poland(2004, -); Portugal; Romania(2007, -); Slovak Republic(2004, -); Slovenia(2004, -); Spain; Sweden(1995, -); United Kingdom
FTAs with: Albania (2006), Algeria (2005), Andorra (1991), Armenia (2018), Bosnia and Herzegovina (2008), Cameroon (2014), Canada (2017), Cariforum states (2008), Costa Rica, El Salvador, Guatemala, Honduras, Nicaragua and Panama (2013), Chile (2003), Colombia, Peru and Ecuador (2013), Côte d’Ivoire (2016), Madagascar, Mauritius, Seychelles and Zimbabwe (2012), Egypt (2004), Georgia (2014), Ghana (2016), Iceland (1973), Israel (2000), Japan (2019), Jordan (2002), Lebanon (2003), Mexico (2000), Moldova (2014), Morocco (2000), North Macedonia (2001), Norway (1973), Papua New Guinea and Fiji (2009), Botswana, Lesotho, Mozambique, Namibia, South Africa and eSwatini (2016), Serbia (2010), South Africa (2000), Switzerland (1973), Syria (1977), Tunisia (1998), Turkey (1996), Ukraine (2014),
34 Economic and
Monetary Community of Central Africa
June 1999 Cameroon; Central African Republic; Chad; Democratic Republic of the Congo; Republic of the Congo; Equatorial Guinea; Gabon
CU Economic Community of West African States (ECOWAS) August 1995
Cabo Verde; Benin; The Gambia; Ghana; Guinea; Côte d'Ivoire; Liberia; Mali; Niger; Nigeria; Guinea-Bissau; Senegal; Sierra Leone; Togo; Burkina Faso
CU
Eurasian
Economic Union
January 2015
Belarus; Kazakhstan; Russian Federation Armenia; Kyrgyz Republic
CU & EIA
European Free Trade Association (EFTA)
May 1960 Iceland; Liechtenstein; Norway; Switzerland Austria; Denmark; Finland; Iceland;
Liechtenstein; Portugal; Sweden; United Kingdom
FTAs with Albania (2010), Bosnia and
Herzegovina (2015), Canada (2009), Costa Rica and Panama (2014), Chile (2004), Colombia (2011), Egypt (2007), Georgia (2017), Israel (1993), Jordan (2002), Republic of Korea (2006), Lebanon (2007), Mexico (2001), Morocco (1999), North Macedonia (2002), Peru (2011), SACU (2008), Serbia (2010), Singapore (2003), Tunisia (2005), Turkey (1992), Ukraine (2012),
FTA & EIA, from 2002 onwards in in its current form GUAM December 2003
Azerbaijan; Georgia; Moldova; Ukraine FTA & EIA
Gulf Cooperation Council
January 1982
Bahrain, Kingdom of; Kuwait, the State of; Oman; Qatar; Saudi Arabia, Kingdom of; United Arab Emirates
FTA & EIA, CU before 2003 North American Free Trade Agreement (NAFTA) January 1994
Canada; Mexico; United States of America FTA & EIA
Pacific Alliance May 2016 Chile; Colombia; Mexico; Peru FTA & EIA Pacific Island
Countries Trade Agreement (PICTA)
April 2003 Solomon Islands; Cook Islands; Fiji; Kiribati; Nauru; Vanuatu; Niue; Micronesia, Federated States of; Papua New Guinea; Tonga; Tuvalu; Samoa
Solomon Islands; Kiribati; Vanuatu (2005; -); Papua New Guinea; Tuvalu (2008, -)
FTA
Pan-Arab Free Trade Area
January 1998
Algeria (2009, -); Bahrain, Kingdom of; Iraq; Jordan; Kuwait, the State of; Lebanese
Republic; Libya; Morocco; Oman; Qatar; Saudi Arabia, Kingdom of; Sudan; Syrian Arab Republic; United Arab Emirates; Tunisia;
35 Egypt; Yemen
South Asian Free
Trade Agreement
January 2006
Afghanistan (2006. -); Bangladesh; Bhutan; Sri Lanka; India; Maldives; Nepal; Pakistan
FTA
Southern African Customs Union
July 2004 Botswana; Lesotho; Namibia; South Africa; eSwatini CU Southern African Development Community September 2000
Angola; Botswana; Lesotho; Malawi; Mauritius; Mozambique; Namibia; Seychelles (2015, -); South Africa; Zimbabwe; eSwatini; Tanzania; Zambia FTA Southern Common Market (MERCOSUR) November 1991
Argentina; Brazil; Paraguay; Uruguay Venezuela (2012,-)
FTAs with Egypt (2017), Israel (2009), SACU (2014) CU & EIA, since 2005 in its current form Trans-Pacific Strategic Economic Partnership
May 2006 Brunei Darussalam; Chile; New Zealand; Singapore
FTA & EIA
West African Economic and Monetary Union
January 2000
Benin; Côte d'Ivoire; Guinea Bissau (1997, -); Mali; Niger; Senegal; Togo; Burkina Faso
