Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA

Mensah, Lord Kwaku Article Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA Cogent Economics & Finance Provided in Cooperation with: Taylor & Francis Group Suggested Citation: Mensah, Lord Kwaku (2023) : Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA, Cogent Economics & Finance, ISSN 2332-2039, Taylor & Francis, Abingdon, Vol. 11, Iss. 2, pp. 1-15, https://doi.org/10.1080/23322039.2023.2259738 This Version is available at: https://hdl.handle.net/10419/304216 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Cogent Economics & Finance ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/oaef20 Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA Lord Mensah To cite this article: Lord Mensah (2023) Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA, Cogent Economics & Finance, 11:2, 2259738, DOI: 10.1080/23322039.2023.2259738 To link to this article: https://doi.org/10.1080/23322039.2023.2259738 © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. Published online: 28 Sep 2023. Submit your article to this journal Article views: 505 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=oaef20 GENERAL & APPLIED ECONOMICS | RESEARCH ARTICLE Intra-Africa regional trade comovements and shock transmission: A baseline for AfCFTA Lord Mensah 1 * Abstract: This paper examines the trade co-movements and shock spillover across four African geographic regions. Specifically, we were motivated by the very low intra-trade activities in Africa, despite increased regionalism to study the possibility of a country’s trade shock being transferred to its trading partners on the continent. Knowing the trade connectedness and shock transmission among African countries will serve as a baseline for the AfCFTA implementation. In our analysis, we considered the four African regional quarterly data between 2005 and 2021 from UNCTAD. The continent was divided into four regions namely, Western, Middle, Eastern, and Southern Africa. We divided the time into preand post-AfCFTA periods. The Dynamic Conditional Correlation (DCC) and the Diebold and Yilmaz (2012) models were adopted to determine the trade co-movement and the shock spillover respectively. The results show different trade co-movement and trade shock spillovers among the regions at different times. The trade co-movement seems to be dominant between the Middle and Southern African regions. Further analysis shows the presence of trade shock transmission across all four regions. The Western African region exhibits a sign of the biggest trade shock receiver from the other regions, while the Southern African regions turn out to be the largest contributor of trade shocks to the other regions both in the post and preAfCFTA period. The study contributes by sending a signal to AfCFTA implementers that trading on the African continents behaves differently among the various geographic regions. It also provides early warning signal for AfCFTA policy implementation. Subjects: International Trade (incl; Econometrics; Development Economics; Finance Keywords: African Intra-trade; co-movement; trade shock spillovers; AfCFTA 1. Introduction The African Trade Unions agreed to establish the African Continental Free Trade Area (AfCFTA) in a bid to speed up the implementation of the 1991 Abuja Treaty. The treaty entreats Member States of the Organization for African Unity to establish a common African Market AfCFTA since its launch in 2015 became a flagship program of the AU. But for the COVID-19 shock, African economic growth was forecasted to be around 3.6% between 2019 and 2020 with the world’s fastest growing economies found on the continent (African Economic Outlook 2018). The continent still relies on the imports of capital goods outside Africa with heavy dependence on commodity and agriculture export. According to World Bank report published in February 2022, Africa has a global trade share of only 3 percent with export diversification yet to be achieved. Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 1 of 15 Received: 16 October 2022 Accepted: 08 September 2023 *Corresponding author: Lord Mensah, Department of Finance, University of Ghana Business School, Legon Accra Ghana E-mail: [email protected] Reviewing editor: Christian Nsiah, School of Business, Baldwin Wallace University, United States Additional information is available at the end of the article © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. Many countries on the African continent depend heavily on rent from extractive exports and are slow in industrialization efforts. Africa has integrated very fast with the rest of the world in terms of trade to the detriment of itself. As shown in Figure 1, Intra-African trade remained below its potential reflecting about 15 percent of total African trade volumes. This is far below intracontinental trade in other parts of the world. For instance, European intracontinental trade accounts for about 67% of total trade, 60 percent in Asia, and 46 percent in America. Relatively African markets have remained disintegrated while other Regional Economic Communities have seen improvement in trade integration through a reduction in tariffs. In Africa, non-tariff barriers such as uncoordinated bureaucratic procedures, long waiting time at borders or lengthy and crooked export procedures have raised trade costs on the continent. To some extent some papers have found that country to country bilateral trade agreements have been more beneficial than multiple membership trade agreements. For example, Sunge and Mapfumo (2014) found that Zimbabwe’s bilateral trade agreement has created more trades and growing. They also find that the multi-membership trade agreements have limits on bilateral trade. Their results therefore encouraged Zimbabwe to negotiate for more bilateral trade agreements especially with its border countries. The baseline analysis of trade connectedness for AfCFTA implementation appears to be of central importance in determining how trade shocks from a particular country will affect other countries. A better understanding of existing trade connectedness among African countries will enhance policy direction in AfCFTA execution. It will provide the baseline as to which countries to focus on trade among the various countries on the African continent. It will also provide information to investors on how to diversify their investment in relation to intra-trade to avoid trade risk spillovers. Several papers have researched on Trade integration but failed to investigate trade connectedness. For instance, Bonga-Bonga and Mabe (2020) used the DCC -GARCH model to assess the degree of financial integration between three African trading blocs, namely, the Common Market of Eastern and Southern Africa (COMESA), the East African Community (EAC) and the Southern African Development Community (SADC). They find that not all the countries within each of the three trading blocs are integrated into their regional factors. Bailey et al. (2021) used de-identified from Facebook to construct a new and publicly available measure of pairwise social connectedness between 170 countries and 332 European regions. They find that two countries trade more when they are more socially connected, especially for goods where information frictions may be large. Figure 1. Intraregional trade (imports + exports) as a percentage of total trade. Source: https://www.brookings. edu/blog/africa-in-focus/2021/ 05/19/why-the-extent-of-intraafrican-trade-is-much-higherthan-commonly-believed-andwhat-this-means-for-theafcfta/ Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 2 of 15 Having recorded lower intra-African trade provides a baseline for trading among the countries on the continent. This tells us that, before the AfCFTA, African countries were trading among themselves but not enough to project the continent to the industrialization pedigree expected of it. With the level of intra-trade among the countries on the continent, it is expected that there would be a certain level of cyclical synchronization across Africa. As indicated by De Grauwe and Ji (2017) that, for common shocks, synchronization takes place through trade integration and financial integration. The fascinating issue here is to know how existing impact of a country-specific trade shock is transferred to its trading partners on the continent. A similar issue was investigated by Pham and Sala (2022) by looking at the connectedness of inflation and unemployment among G7 countries. They find that connectedness is larger for inflation than to unemployment. Investigating the level of intra-trade connectedness among African countries will provide baseline information for the countries to determine which trade partners they should prioritize in AfCFTA implementation. The objective of this paper is to unravel the extent of trade connectedness and shock transmission among African regional blocs. Understanding how the economies are trade connected and shocks transmitted is crucial and will be a guide for AfCTA policy makers. The paper takes a step forward with respect to the extant literature on trade integration in Africa by considering the asymmetries in connectedness across a key trade variable, country specific export as a percentage of total volume of merchandize trade. Understanding how trade among African economies are connected and shocks transmitted is crucial for AfCFTA policy makers. For example, trade shock from an African region can spread a complex effect to other regions on the continent. Therefore, studying the trade connectedness of African countries is an important and early action for AfCFTA policy decisions. The paper also adds to the existing literature by contributing methodologically and empirically by utilizing the Diebold-Yilmaz Connectedness Index (DYCI) to investigate the baseline connectedness among all the sub-Saharan African countries. The DYCI methodology is adaptable and direct to implement approach to study intra-trade connected among African countries. This method has been used extensively to study the connectedness among several economic variables but not on trade and this is what makes this research unique. Bonga-Bonga and Mabe (2020) in their paper measured trade connectedness as how current distribution of exports is close to a theoretical one or trade frictionless full potential integration. Aside the DYCI, several authors have tried to use correlation-related measures for spillover and connectedness (Gasper, 2012). The classical correlation as a measure of connectedness may be misleading and has led to the development of alternative models like the dynamic conditional correlation of Mink et al. (2007) and Adrian and Brunnermeier (2008). Less direct measures of connectedness have also been advanced in literature by Corsetti et al. (2005) and Dungey et al. (2005). The underlying concept here is to evaluate the number of unexpected moves in some economies may be explained by moves in other economies. The DYCI models fit into this category. The DYCI method extends the earlier models by introducing directional spillovers. The rest of the paper is organized as follows: In the next section we review the literature related to the topic understudy. In section 3, we discuss the methodology and the various tests needed to unravel our hypothesis. Empirical results and their economic implications are discussed in section 4. In section 5, we discussed the conclusion and the policy implications. 2. Literature Review Recent economic context has shown a strong market interdependence across borders. The literature on modelling and measuring this economic connectedness and systematic shocks has been growing. The literature in this area has increased because of the transmission of idiosyncratic economic or financial shocks from one market to the other. This was clearly exhibited in the 2008– 2009 financial and economic crisis, and it has triggered research in market connectedness and systemic risk. Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 3 of 15 Several papers have used network models to capture a group of financial institutions and the bilateral assets or liabilities exposures between them. For example, Acemoglu et al. (2015) studied the direct and spillover effect of state capacity on Colombia. They model the determination of the state capacity as a network game between municipalities and the national government. The model specifically explores the municipality network and the roots of local state capacity related to the presence of the colonial state and royal roads. They find that local state capacity decisions are strategic complement. They also found that the spillover effects are significant, constituting about 50% of the quantitative impact of an expansion in local state capacity, however, network effect driven by equilibrium reactions of other municipalities is bigger. Another set of methods papers use statistical models to derive the financial and economic connectedness and systemic risk that are typically non-structural in nature. An example is the CoVar approach adopted by Adrian and Brunnermeier (2011) and the volatility spillover structure by Diebold and Yilmaz (2009, 2012, 2014). Recent paper by Costa et al. (2022) used social network models to position the Italian industrial sectors within their trade networks to analyze their ability to transmit economic shocks within the economic system. Their results show a mismatch between the industries which are pivotal within the international trade network and industries which play a central role for domestic shock propagation. With their results, they conclude that the capacity of transmitting stimuli from abroad is limited, weakening the chances of benefiting from positive shocks, though it could be a partial neutralizer to negative shocks that emanate from business cycles. Ekeocha and Ogbuabor (2019) in their working paper investigated trade shock transmission between Africa, the BRIC and the rest of the global economy with to understand Africa’s exposure to trade shocks. They find that Africa economies are predominantly net receivers of trade shocks. The network models have been applied empirically since the 2008–2009 financial and economic crisis. Uluceviz and Yilmaz (2021) applied the model to analyze connectedness between the real and financial sectors of the US economy and find that during times of financial distress or business cycle turning points, the direction of connectedness runs from the real sector to financial market However, the direction connected reverses when they add real activity index to the real sector measure. The financial markets generate positive net connectedness to the real side the economy. Bostanci and Yilmaz (2020) use network models to estimate the global network structure of sovereign credit risk. They used sovereign credit default swaps to establish that the level of credit risk connectedness among sovereigns is high and it is comparable to the connectedness among stock markets and foreign exchange markets. They further establish that emerging market countries have played a major role in the transmission of sovereign credit risk, whiles developed countries and the debt-ridden developed country have played a marginal role after the global financial crises in the 2008. On trade integration, Arribas et al. (2020) present several indicators of trade integration by focusing on European unions. They suggest measures such as openness, connectedness, and integration to establish that the process of trade integration as grown among European union members, but the integration among non-member states has been increasing slowly. Anderson and Yotov (2016) in their paper estimated the effect of bilateral trade volume in two digits manufacturing goods from 1990–2002, using the panel data gravity methods to resolve two-way causality. They find that some countries gain over 5% of real manufacturing income, while some lose less than 0.3%. They also estimated that global efficiency of manufacturers trade rises 0.9% depending on the distance function of measure of iceberg melting. Baier et al. (2019) indicate that estimation of the effects of regional trade agreements can be varied within the same trade agreement. They show in their paper that Romania, Bulgaria and Hungary exhibit positive effect of regional trade agreement, whiles Cyprus show negative effect. In effect, they established that the effect of trade agreement can be heterogenous within the same trade area. Bailey et al. (2021) use the de-identified data from Facebook to construct a new publicly available measure of the pairwise social connectedness between 180 countries and 332 European regions. They find trade to be prevalent in pairs of countries which are socially connected, especially for goods with larger information friction. They indicate that after controlling social connectedness, the estimated Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 4 of 15 effects of geographic distance and country borders on trade decline substantially. On the intraregional economic trading effect on economic growth, Hazman et al. (2021) used data spanning 1995 to 2019 across five regional economies, namely ASEAN, EU, NAFTA, MERCOSUR, and COMESA to establish this relationship. They adopted the GMM and the 2SLS panel approach to establish that most increases in export may reverse economic growth in intra-regional trading. They concluded that trading in commonality is doubtful in promoting long run economic growth. The above papers did not consider the co-movement and transmission of trade shocks among the Sub-Saharan African countries. This paper is unique because it brings the notion of connectedness in testing the bilateral transmission of trade shocks and co-movement of trades among SubSaharan geographical blocs. 3. Data and methodology The study seeks to unravel the extent of co-movement and trade shock spillovers of intra trade in Africa. We sourced quarterly four regional bloc trade data from the United Nations Conference on Trade and Development (UNCTAD). UNCTAD categorizes the sub-Saharan Africa into four trade blocs, that is Eastern, Middle, Southern and Western African trade groups. Countries in each category and their corresponding commodities traded are found in Appendix I & II. Each bloc export time series trade volumes in Africa are used to proxy intra-trade in Africa. We analyze the co-movement and spillovers of trade shocks in different periods between 2005 and 2021. We also split the data into two subperiods, that is pre and post AfCFTA announcement (event) of 2005– 2017 and 2018–2021 respectively. We understudy the comovement and the spills overs of the trade shocks in the two event periods. The addition of event study on trade shock transmission is supported by Kočenda and Moravcová (2019) who find a strong comovement and spillover among economic variables on the European market during financial and economic distress. In this paper, we are looking at the shock spillovers and co-movement before and after the announcement of the AfCFTA. The addition of the event study is also supporting the evidence in the empirical literature that economic series may exhibit structural breaks in times of economic crisis, Andreou and Ghysels (2009) or changes in policy (Pesaran et al., 2006). Quaterly percentage change in trade @t ð Þ is computed from the quarterly trades from each trade zone as @t¼ ðln dt=dt1Þ�100, where dt is the year t0s to Africa. The Augmented Dickey-Fuller (ADF) generalized least square test is used to test the stationarity of the of the percentage change series. Following Kočenda and Moravcová (2019), the paper adopted the DCC model by Engle (2002) to evaluate the dynamics of co-movement between the African countries’ trades. Here, we are trying to ascertain whether dynamic correlation exists among the African geographical trade zones, and if its exist, does it increase, decrease, or remain the same over the period of our study. The advantage of the DCC model is that it is parsimonious compared to other multivariate GARCH models. The DCC model is also flexible because it allows for the estimation of time varying volatilities, covariances, and correlations of various series over time. The two-stage approach is adopted in the estimation of the DCC model. The univariate GARCH models are estimated for each trade series @t. The residuals transformed by their standard deviations from the first stage GARCH models are used to construct conditional correlation matrix. Based on the assumption of no serial correlation, the series@t in the mean equation follows a random walk and the composition of the conditional covariance matrix is given by: Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 5 of 15 where δii;tis the univariate GARCH model, that is the AR(1)-GARCH (1,1) is model if serial correlation is found in the trade series of the GARCH(1,1) model. Φt¼γij;t � �is the N�N symmetric positive definite matrix given by where εt¼ε1t;ε2t;���;εNt ð Þ0is the N�1 vector of standardized residual; � Φ is the N�N of the unconditional variance of εt; and αand β are non-negative scaler parameters satisfying the condition αþβ<1. The log maximum likelihood function under a heavy tailed multivariate generalized error distribution (GED) is used to estimate the DCC model. The DCC model helps us to test the hypothesis that the dynamic conditional correlation between the African trade zones do not change pattern and magnitude across the sub-periods under study. Finally, estimate the trade shock spillovers based on variance decompositions as proposed and created in earlier papers by Diebold and Yilmaz (2009, 2012, 2014). The variance decomposition spill-over measures are very intuitively appealing for several reasons. Firstly, it permits different horizon volatility spillovers, enabling examination of variety of horizons and selection of the preferred horizon. This is very necessary because 1-year volatility spill-over may be different from 10 or 30-year spillovers. Second, the model is intuitively appealing in a sense that it answers key practical questions. It answers question at the primary pairwise level, as to how much of trade zone i0s trade uncertainty (at horizon H) is caused by the shocks due to another trade zone j0strade but not trade of zone i itself. The model is also strongly related to modern network theory, specifically, the degree of distribution and mean degree. Finally, the model is intimately linked to the recent proposed measures of systemic risk, such as the marginal expected shortfall developed by Acharya et al. (2017) and CoVar by Tobias and Brunnermeier (2016). The variance decomposition of Nvariable VAR pð Þ; @t¼∑p i¼1Θi@tiþvt;where vt~ 0;ð Þ. The moving average representation is @t¼∑p i¼1Bivti, where the N�N coefficient matrices Biobey the recursion Bi¼Θ1Bi1þΘ2Bi2þ���þΘpBip, with B0 an N�N identity matrix and Bi¼0fori<0. We follow Diebold and Yilmaz (2012) and Balcılar et al. (2016) in using the generalized identification framework of Koop et al. (1996) and Pesaran and Shin (1998) which allows for invariant ordering of variables, though it will produce the variance decomposition. The generalized approach instead of orthogonalizing the trade shocks allow for its correlation and appropriately accounts for it. The j0s contribution to trade i0s H-step-ahead generalized forecast error variance, dg ij Hð Þ;is dg ij Hð Þ ¼ θ1 jj ∑H1 h¼0P0 i ð ÞBh�Pj ð Þ2 ∑H1 h¼0P0 iBh�B0 hPi ð Þ , H¼1;2;���; where �is the covariance matrix of the disturbance term vt, θjj is the standard deviation of the disturbance of the jthequation, and Pi is the selection vector with one as the ithelement and zeros otherwise. From the Koop-Pesaran-Potter-Shin generalized VAR framework, the variance weights do not add up to 1, that is ∑N j¼1dg ij Hð Þ�1:Therefore, normalizing each entry of the generalized variance decomposition matrix by the row sum to obtain pairwise directional connectedness from country j to country i gives: Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 6 of 15 therefore, ∑N j¼1~ dg ij Hð Þ ¼ 1 and ∑N i;j¼1~ dg ij Hð Þ ¼ N. We now denote the ~ dg ij Hð Þ as VH i j (V is the trade shock), this is less burdensome and more clearly informative. Having obtained the pairwise shock spillover measure, VH i j, we move on to the total directional trade shock spillover measures. The total directional trade shock spillover to zone i from all other zone j is: on the other hand, total directional measure from zone i to all other zone j is: Now the system-wide trade shock spillover measure is obtained as: The equation above estimates the total system-wide trade shock spillover, and it is the sum of all total trade shock spillovers whether to or from a particular country. 4. Empirical results and discussions Table 1 reports the descriptive statistics for the various Africa regions trade growth. Overall, the largest trade growth across the regions over the period of our study can be found in the Western part of Africa. This is followed by the Middle African countries. And the least of all the trade flows are Eastern African countries. Specifically, the Western African region recorded an average trade growth of about 1.08, representing 108% growth. The second is about 100% growth, which is the Middle African countries. South Africa was third in terms of the rankings in the average trade growth. The standard deviation measuring the variations in the trade growth across the various regions on the African continent is also recorded in Table 1 row 5. From the Table, the Middle African countries record the highest volatile trade growth. The is expected because the countries with the region recorded quite several wars which may have impact on trade flows among themselves with the period of our study. 1 Typical of these wars are the ongoing 2006 Bakassi Conflict and 2014 Boko Haram insurgency in Cameroon, the present Central African Republic Civil War, 2010 Civil War, 2008 Battle of N’Djamena in Chad, 2016–2017 Pool War in Congo, 2012–2013 M23 Rebellion and 2013–2018 Batwa-Luba clashes, all in Congo. This is followed by Eastern Africa. The third volatile region in terms of trade growth is Western Africa and the least is Southern Africa. It is not surprising to see Southern Africa recording the least trade growth variability. The main exports in these regions are minerals and other precious metals for which the demand on the global market is consistent. The Middle African countries may experience the high variability in trade growth because of the persistent war in the largest country in this region. The Jarque-Bera test of Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 7 of 15 Appendices Appendix I: UNCTAD categorization of Intraregional trade in Africa Northern Eastern Africa Middle African Southern Western Africa Algeria British Indian Ocean Territory Angola Botswana Benin Egypt Burundi Cameroon Eswatini Burkina Faso Libya Comoros Central African Republic Lesotho Cabo Verde Morocco Djibouti Chad Namibia Côte d'Ivoire Sudan Eritrea Congo South Africa South Africa Sudan Ethiopia Congo, Dem. Rep. Ghana Tunisia Ethiopia Equatorial Guinea Guinea Western French Southern Territories Gabon Guinea-Bissau Kenya Sao Tome and Principe Liberia Madagascar Mali Malawi Mauritania Mauritius Niger Mozambique Nigeria Rwanda Saint Helena Seychelles Senegal Somalia Sierra Leone South Sudan Togo Tanzania Uganda Zambia Zimbabwe Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 14 of 15 Appendix II Northern Eastern Africa Middle African Southern Western Africa Petroleum and petroleum products Coffee Crude oil Diamonds Petroleum and petroleum products Natural gas Tea Timber Platinum Cocoa Textiles and clothing Cut flowers Minerals (iron ore, cobalt, copper, gold) Gold Gold Fertilizers Horticultural products Cocoa Coal Timber Phosphates Tobacco Coffee Copper Rubber Wheat Sesame seeds Rubber Agricultural products (maize, wheat, sugarcane, citrus fruits) Cashew nuts Olive oil Cashew nuts Palm oil Wine Cotton Dates Maize (corn) Diamonds Livestock and meat products Palm oil Citrus fruits Beans Cassava Fish and seafood Shea butter Fish and seafood Livestock and meat products Fish and seafood Tourism Fish and seafood Mensah, Cogent Economics & Finance (2023), 11: 2259738 https://doi.org/10.1080/23322039.2023.2259738 Page 15 of 15