2.3.1. West and Central Africa

The gravity model specified in Appendix A is estimated for West and Central African countries as reporting countries and the rest of Africa as trading partner using the Pseudo Poisson Maximum Likelihood (PPML) technique that tackles the problems associated with OLS estimation (Silva and Tenreyro, 2006). In addition, an alternative panel-Tobit based estimation is also made, both for completeness and as a robustness check of the results. The sample covers all countries if the required data is available (see Appendix A for country samples used). The result of this estimation is given in Table 2a.

The result shows that most of the traditional gravity model variables have the expected signs. They are also statistically significant.² In addition, the remoteness index which is a proxy for multilateral resistance term (Baldwin and Taglioni, 2006; Anderson and van Wincoop, 2003; Rose, 2000; Rose and van Wincoop, 2001; Eicher and Henn, 2009) is found to be an important determinant of bilateral trade. The difference in per capita income between West African countries and their trading partners, a proxy that shows difference in preference or what is called ‘the Linder effect’, is found to affect bilateral trade negatively, as expected.

Having this result, a simulation exercise is conducted to examine the potential or theoretical level of intra-Africa trade for each West African country, given the parameters of the model as reported in Table 2a. This is then compared with actual trade of each country (as the ratio of that country’s total trade with all its major trading partners, both in West Africa and the rest of the world). The result of this exercise is given in Table 2b.

The result, as can be read from the last column of Table 2b, shows that for all countries in the region (except for Central African Republic, Guinea-Bissau and Liberia, where the model underestimated their level of trade) their actual trade is significantly below the potential given by the gravity model. This result varies across countries, however. Whereas some countries’ actual trade levels are near their potential (90% for Gabon), the majority are found to have a level of actual trade significantly below their potential. If the three countries are excluded, on the average, the actual trade is only 18% of the potential (or just 16% using the median value).

2.3.2. North, East and Southern Africa

A similar gravity model to that of Table 2a is also estimated for North, East & Southern African countries as a group. The result is given in Table 3a.

The result shows that most traditional gravity model variables have the expected signs. They are also statistically significant in both estimation techniques (except the common official language dummy which implausibly has a negative sign, though is not statistically significant and hence not important). In addition, the multilateral trade resistance proxy is found to have a statistically significant positive effect on bilateral trade in the region. The difference in per capita income is found to have a statistically significant negative coefficient, suggesting the existence of the ‘Linder effect’. Being landlocked is found to adversely affect bilateral trade. This is mainly because landlocked countries incur a higher logistic cost to accesses the sea which adversely affects their competitiveness and hence their level of exports and import bills.

Having the result in Table 3a, we have conducted similar simulation exercise to that of Table 2b. The result is given in Table 3b. Table 3b shows that, except for Djibouti which is heavily trading with Ethiopia; and Madagascar which is heavily trading with Mauritius, South Africa and Mozambique, for the rest of the countries in the region their actual trade is significantly below the potential. Excluding the three countries, on average, actual trade in the region is only 10% of the potential (or mere 3% using the median value). This is half the level we found for West and Central African countries in Table 2b.

In sum, the result from the simulation exercises in this sub-section shows the existence of significant potential for intra-Africa trade. This potential is found to vary both across countries and regions.³ Although the exercise suggests a great potential for intra-Africa trade, this does not mean that this potential could easily be realized. What the models suggest is that, given the geographic proximity, cultural affinity as well as the size of the economies, the level of intra-Africa trade could have been larger than what it is now. However, since the model is based on proxy variables noted above and uses aggregate value of exports, it fails to capture the structure of supply of exports and the demand for this in the importing country. Thus, the result indicates only the necessary, but not the sufficient condition, for enhancing intra-Africa trade. The sufficient condition for this trade to hold is that what is supplied by one of the African countries, is specified by commodity category, and needs to be demanded by the trading African partner country. This requires examining the pattern of demand and supply by commodity category. Moreover, even if one finds complementarities between these trading partners, it is imperative to examine the comparative advantage of the potentially exporting African countries in replacing the current trading partners of the importing African countries. These issues are discussed in the next section.

2.4.1. The structure of African exports and imports

Using disaggregated commodities from UNComtrade database, in this section, we have examined the structure of demand for and supply of exports in Africa. This is done by computing trade similarity indexes for all countries in the continent. Trade similarity (intensity) indices are generally defined as the ratio of the “share of a given country’s exports” to a partner country to “the share of the partner’s similar index” in the global imports of the commodity in question.

Using such indices for the Sub-Saharan African countries for the period 1989–1995 Yeats (1999) noted that the level of intra-Africa trade is not lower than what should be expected. Similarly, in an earlier study, Oramah and Abou-Lehaf (1998) computed two different measures of import–export similarity indices which are based on Van Beers and Linneman (1988) approach using the 1993 intra-Africa trade data. They have, then, estimated a gravity model that includes these indices as explanatory variables to elicit the relevance of these variables in explaining actual intra-Africa trade. They also computed a trade potential indicator (TPI) that shows the relative strength of each African country in exporting to other African countries. The results of the study indicate that the potential for improving intra-Africa trade is rather modest, with only a few countries having export structures strongly matching the import demands of other African countries.

When these measures of absolute trade potential were weighted by economic size of trading partners using the relationship obtained from estimated gravity equation, it was found that the countries that would stand to gain most from greater efforts to advance intra-Africa trade were Comoros, South Africa, Egypt, Sudan, Congo, Cote d’lvoire, Gabon, Algeria and Cameroon. Those which may not gain as much include Burundi, Namibia, Guinea and Ethiopia owing to the poor correspondence of their exports to the imports of economically strong African economies. However, the gains were found to be normally distributed without very large extremes (Oramah and Abou-Lehaf, 1998).

This is a story using a limited sample and data compiled 15 years ago. It is also the only study available. Updating the data to current period and expanding the coverage to all African countries offers empirical evidence about the current state of potential intra-Africa trade using a similar approach. The result of such an exercise is important to evaluate the feasibility of realizing the significant intra-Africa trade potential identified in the previous section. For this purpose, the export–import similarity index approach originally developed by Van Beers and Linneman (1988) and used successfully for Africa by Oramah and Abou-Lehaf (1998) is employed for the analysis in this section (see Appendix B for the methodology).

The import–export correspondence index, COS, value used in this study, varies between zero (no correspondence between the export of country i and the import of country j) and one (perfect similarity). COS is the cosine of the angle between the vectors of country i exports and the vectors of country j imports in an n-dimensional commodity space. Using top five import and export commodities of all African countries, for which data is available for the period 2004 to 2007, a COS measure is computed. This index may be interpreted as a variable reflecting the expected intensity of bilateral trade flows from exporting country i to importing country j. The similarity of exports and imports of all class of commodities is, however, one of the determinants of trade intensity between two countries. The term ‘intensity’ is used in Van Beers’ and Linneman’s (1988) interpretation to indicate that the analysis abstracts from the economic size of trading partners are reflected in the total volume or value of their exports and imports. The latter are seen as scale factors with which the ‘intensity’ has to be multiplied in order to arrive at observed or potential trade flow magnitudes. Moreover, apart from the degree of similarity between exporting and importing country commodities, other factors such as infrastructure, non-tariff barriers, political factors favoring or obstructing trade, as well as relative competitive position of exports do determine the level of bilateral trade between a pair of countries. Among all these factors, the role of the degree of similarity in trade structure is a limited yet an important one. “Without any commodity correspondence no trade will take place, and with perfect correspondence trade possibilities abound”, as Van Beers and Linneman (1988) noted.

Table 4 reports estimated trade a similarity profile of African countries using the export–import similarity index, COS. A few points need to be noted concerning the estimation techniques used to generate this result. First, in computing the similarity measures using aggregate data, aggregated under the Standard International Trade Classification (SITC) format, no information is required about the individual elements of the trade matrix at the commodity class level. Only total exports of country i in commodity class k and total imports of country j in commodity class k are needed to compute the values of the measures. As noted by Van Beers and Linneman (1988) the implication of this is that the measures indicate a trade potential or expected intensity of trade between a pair of countries. Thus, a non-zero value of COS_ij does not necessarily imply that in actual fact country i does export to country j. Rather, the estimated export–import similarity measure can be interpreted as reflecting the expected intensity of a bilateral trade flow from exporting country i to importing country j. Second, usually a country with a COS measure equal to or greater than 0.4 would be considered to have a reasonably matching import structure with the exporting country of reference in the literature (Van Beers and Linneman, 1988; Oramah and Abou-Lehaf, 1998). With this note, Table 4 shows a summary of COS index computed, using top five export and import commodities of each African country at three digit SITC classification.

Table 4 shows that the exports of South Africa, Nigeria, Algeria, Egypt and Cameroon fit well with the imports of most African countries. These countries are found to have the potential to supply exports to the trading partners listed from column two to six, in order of importance. Notwithstanding this potential, the actual pattern of trade shows neither the realization of this potential nor reciprocity from the potential importing countries. This may relate to the existence of significant challenges and lack of adequate incentive to engage in trade.

In Table 4 Algeria, Nigeria and Cameroon in particular have got a high value of import–export similarity index with most African countries. This is related to the nature of their major export commodity which constitutes mineral fuels, oils and related products. These are also major importable commodities by almost all African countries. South Africa and Egypt, on the other hand, are capable of exporting manufactured commodities that are demanded by the rest of Africa. Table 4 also shows the existence of countries such as Benin, Ethiopia, Niger, and Zambia with export commodities (cotton, edible fruit and nuts; peel of citrus fruit or melons, coffee, tea and spices, oil seeds and oleaginous fruits, copper) that have less or no demand at all in other African countries.

In general, the result shows that only few countries have the potential to supply/export commodities that match the demand of the majority of countries in the continent. These are Egypt, South Africa, Cameron, Nigeria, and Algeria — the latter three countries’ importance being related to their fuel exporting status. In addition, the composition of exports from these top suppliers is not well diversified and, hence, limited in matching the demand from the majority of countries in Africa. Even South Africa’s top exports, for instance, are dominated by mineral and precious metals, followed by Iron and steel and some manufactured goods. Similarly, Egypt’s export is dominated by mineral fuels with limited supply of manufactured goods. From this analysis it is not difficult to conclude that the potential for intra-Africa trade, which seems very large using the gravity model, is very modest from commodity composition perspective. The implication of this finding for policy relates to the importance of a conscious region or continent wide strategy of diversification and competitiveness aimed at enhancing intra-Africa trade.

2.4.2. Revealed comparative advantage of potential African-based export suppliers

Although the discussion in the above sub-section points out the limited scope available to exploit the potential for intra-Africa trade identified in Section 2.3, it has also shown that some African countries have, at least, the theoretical potential to supply export to other countries in the continent. It is not clear, however, whether these African countries are capable of competitively supplying to other countries in the continent. The majority of countries in the continent are currently importing from non-African suppliers such as the OECD countries and the emerging Asian economies such as China with which the Africa-based suppliers need to compete with.

Earlier similar empirical studies for other developing countries were not encouraging. Van Beers and Linneman (1988) used a wide range of data from developing (including Africa) and developed countries (OECD countries) to examine this issue in the late 1980s. They found that none of the Southern countries included in their sample are found to be competent enough to replace the Northern exporters. Heavy reliance of developing countries on developed countries’ exporters for the supply of manufactured products is at the core of the whole issue of raising South–South trade. This picture did not change fundamentally in the early 1990s either. A World Bank (1991) study, for instance, argued that African countries have similar production and export profiles and therefore have little or nothing to exchange. Other studies (Achy, 2006; Weeks and Subasat, 1998; Roelofsen, 1989; ECA, 1987), however, argued to the contrary. They pointed out that intra-Africa trade potentials are enormous. For instance, the gravity model based study of Achy (2006) revealed that compared to international trend the actual North African intra-regional trade is ten times lower than its potential (Achy, 2006). Achy (2006), using trade complementary index, further noted that a strong correlation exists between this trade complementarity index and the growth rate of intra-regional trade in the region. Thus, these studies had sought further trade liberalization, and speedier economic integration efforts to realize such trade potential. Examining the relative competitive position of African suppliers with those currently supplying countries in the continent could shed light on the challenge of realizing this potential. Such information is offered in Tables 5 and 6.

In Table 5, the relative value of the revealed comparative advantage indices of manufacture exports of African countries identified to have the potential to supply other countries in the continent (South Africa, Egypt, Cameroon, Nigeria, Algeria, and Kenya) is given. This is compared with the current major African suppliers (Western Europe, USA, Japan, China and India). Other related indicators of competitiveness of these two groups of suppliers are also given in Table 6.

The ‘revealed comparative advantage’ index used in Table 5 is based on Balassa’s (1965) approximation using post-trade data that manifests both post-trade relative prices and prevailing factor and product market distortions. According to Balassa (1965, 1979) and Mahmood (2001) comparative advantage is revealed in relatively high shares of export markets. However, to evaluate what is low or high, Balassa (1965, 1979) called for these shares to be compared to some average. Defined as such, the revealed comparative advantage index (RCAI) compares a country’s world export share of the commodity in question with the total world export share of the total exports of that country. If a country’s share of world export of a particular commodity is greater than the country’s share of world exports of all commodities, the RCAI will be greater than one — indicating that this country has RCA in that product (Lutz, 1987; Rana, 1990; Dowling and Cheang 2000; Mahmood, 2001; Geda and Meskel, 2008). Formally, the RCA index could be given by (Kwan, 2002; Geda, 2012):

Table 5 shows that African potential suppliers’ RCA is less than half the level registered for OECD countries as well as China and India. This would have been even less than 10% of the OECD and emerging South’s levels had it not been for Egypt’s and South Africa’s better performance. Thus, within Africa, South Africa, followed by Egypt and Kenya, is relatively better in terms of RCA, although none of them have a comparative advantage compared to the current manufactured good suppliers to the continent. This result, combined with the huge trade logistic and intra-Africa infrastructure problems noted, further shows the challenge of exploiting the theoretically significant trade potential suggested by the gravity model.

In addition to the RCAI shown above, Table 6 shows other related indicators of competitiveness between potential African suppliers and the current export suppliers to the continent. Consumer price index (CPI) ratio, as an indicator of general level of prices, is widely used in the literature for this purpose. In all African countries identified as potential suppliers, their domestic price is found to be larger than the domestic price of the current trading partners of Africa. The African based potential suppliers do not also seem to fare better either in the relative level of their real effective exchange rate, or export unit price and the composition of their exports (Table 6). This further shows how challenging the realization of the untapped potential trade of the continent is.

In sum, the analysis in this section shows that given the geographic proximity, economic size and cultural affinity African countries have a large potential for intra-Africa trade as we have shown in Section 2 of this study. However, its realization is challenged by the lack of diversification and low competitive position of potential African-based export suppliers which is making the short run potential of intra-Africa trade very modest at best. These are, thus, challenges of intra-Africa trade and regional integration that need to be examined. Understanding these challenges in turn requires understanding factors behind export supply constraint, low level of competitiveness as well as the record of poor diversification in the continent. The next section is devoted to that issue.

3.1.1. The nature of African economies and lack of progress in regional economic integration

The sizes of most African markets are small which entails high transaction cost in trading and the absence of benefits from economies of scale (Njinkeu and Fosso, 2006). These countries cannot generate adequate quantity and quality of competitive exports or attract significant amounts of foreign investment. Thus, active and conscious engagement on the global economy could help such African countries to overcome the small internal market constraints (Njinkeu and Fosso, 2006). However, exporting at world prices is challenging for many African countries as trade could still be costly even if one can source inputs from an efficient global supplier (Njinkeu and Fosso, 2006). This is owing to small consignment size, small-scale infrastructure and lack of competition (Njinkeu and Fosso, 2006). This is aggravated by severe competition from other relatively developed countries and emerging economies such as China and India (Geda, 2013). This is a challenge to most African countries. One policy direction to address this problem is to pursue regional integration vigorously. This is becoming difficult owing to the limiting structure of African production and trade as we have seen in the previous sections.

The relatively small weight of intra-regional trade in Africa, despite the existence of several (and frequently overlapping) RECs, is due largely to their structure of production and the composition of their exports. Many countries are still specialized in a small number of primary commodities (i.e.; there is lack of diversification) for their exports and most of their imports consist of manufacture goods (UNCTAD, 2007; Limão and Venables, 2000). For instance, the relatively low level of intra-regional trade in ECOWAS, according to Shams (2003), is explained by the high dependence of most member countries on exports of primary commodities, and by a trade liberalization scheme that has very strict rules of origin. Access to the regional market is especially difficult for those firms and sectors that are at an early stage of development. Such firms have to rely on imported inputs, and the content of domestic value added in their products is often too small to satisfy the rules of origin requirement. In early 2000, only 17 manufacturing firms in the region were able to comply with these rules (Shams, 2003).

Although it is difficult to establish whether intra-Africa trade causes progress in RECs or the other way around, the two could be re-enforcing each other to bring progress in economic integration in the region. Progress in regional integration, which would have been an invaluable instrument to expand intra-Africa trade is limited mainly due to unwillingness of governments to (i) surrender some sovereignty of macroeconomic policy making power to a regional authority; (ii) face potential consumption costs that may arise by importing from a high-cost member country; (iii) accept unequal distribution of gains and losses that may follow an integration agreement at least in the short run and (iv) unwillingness to discontinue existing economic ties with non-members (Johnson, 1995; Geda and Kibret, 2008).

3.1.2. Absence (or poor state) of trade related infrastructure

In many cases, formal trade liberalization is not successful partly because some fundamental aspects of trade logistics, such as infrastructure are limited. According to Longo and Sekkat (2001) a 1% increase in the stock of transportation and telecommunication infrastructure in the exporting country boosts its export towards other African countries by about 3%. Poor infrastructure, or its complete absence, makes trade physically difficult, quite independent of the trade regime. The performance of African countries with the potential to supply exports to the rest of the continent in terms of trade related macro indicators is less remarkable compared to their current competitors such as the OECD countries (Table 7). As can also be read from the last 3 columns of Table 7, the level and efficiency of infrastructure relevant for trading in African-based potential export suppliers is well below the level available in countries that are currently supplying exports to the African continent such as the OECD member countries. Such deficiencies have also tended to isolate African states from each other and thwarted the potential intra-Africa trade. Intra-Africa trade is further constrained by the absence of market information. In Africa where communication links among countries are few and indirect, relevant market information may be costly to obtain for both importers and exporters (Yeboah, 1993).

The existence of tariff barriers or quantitative constraints also poses formidable obstacles to trade, but they do not render trade exchanges completely impossible, as does the absence of an appropriate regional infrastructure (UNCTAD, 2007). According to the ECA (2004) study, transport costs are 63% higher in African countries compared to the average in developed countries. Similarly, according to UNCTAD (cited in Lisinge, 2005), the freight cost as a percentage of the imported value stood at 11% for North African countries. This is 111% more than those in industrialized countries and 25% more compared to the average in developing countries. Collier and O’Connell (2008) noted that globally coastal areas have a growth advantage of about 1.5% age points better compared with their landlocked counterparts. As about 35% of African population lives in landlocked countries, compared to the global average of 1%, this should have mattered a lot. However, neighborhood growth spillovers are much weaker, and indeed for Africa’s landlocked, resource-poor countries there are no significant neighborhood growth spillovers in Collier and O’Connell’s study. They have attributed this to Africa’s internal barriers to trade which is consistent with evidence on both the actual practice of trade policies and the very high level of transport costs.

Trade logistics, level of economic development and the size of the economy, for instance, is found to influence the bilateral trade potential in Southern African countries (Cassim, 2001). The study shows that transaction costs of trade, the growth paths of economies and changes in per-capita income should be the focus of economic policy in order to reap benefits from intra-Africa trade in the region. Thus, trade liberalization could be effective if it is combined with an overall reduction in trade costs (Cassim, 2001). Similarly Shams (2003) noted that some of the major obstacles to intra-regional trade in West Africa are bureaucratic and physical ones, such as road charges, transit fees and administrative delays at borders and ports (Shams, 2003). Using a modified gravity model, Nwabuzor (2002) found that trade growth among ECOWAS member nations has been thwarted by inadequate infrastructure that tended to increase transaction costs. There is also supporting evidence on this in the case studies discussed in Section 3.2 below.

In sum transport costs in Africa are found to be the highest in the world. The freight costs as percentage of total value of import were about 13% for the continent in 2000 while it was 8.8% for all developing countries and 5.2% for developed countries (UNCTAD, 2002). Ackah and Morrissey (2005) also noted that transport cost constitutes about 15% (this being about 20% for landlocked countries) of unit value of exports in Africa, which is considerably higher than other regions such as Asia, about 8% and Western Europe about 5%. The freight cost differs from region to region in Africa, however, with East and South Africa experiencing higher costs compared to other regions in the continent. This poor infrastructure has increased the trading costs and lowered competitiveness and, hence, contributed to the lower level of intra-Africa trade.

3.1.3. Lack of macroeconomic coordination and the multiplicity and inconvertibility of currencies

Macroeconomic coordination is important for enhancing intra-Africa trade and regional integration. This is because, as noted by O’Connell (1997), success on trade liberalization and growth could be undercut by problems of macro policy coordination. Since the early 2000 the continent as a whole has recorded one of the fastest economic growths in the world which was interrupted shortly in 2008–09 due to the global economic crisis. However other indicators of macro stability such as current account balance, exchange rate, and budget deficit were not that remarkable (Table 7). There is also variation across countries in meeting regionally agreed macroeconomic targets of convergence drawn for various RECs. This requires, among other things, building institutions with a mandate to realize macro-economic policy harmonization in each REC. Establishing such institutions that have a surveillance mechanism to enforce and monitor the convergence criteria drawn is a daunting task. It requires submission of a certain degree of autonomy in domestic policy making by member countries, establishing an elaborated and transparent institutional mechanism, and deploying skilled personnel as the history of European integrations shows. Studies on the issue show that there is an enormous variation across RECs in terms of their macroeconomic posture, asymmetry of shocks as well as the policy response to such shocks (Geda, 2001; Weeks, 2008; Geda and Kayizzi-Mugerwa, 2014). This underscores the need to design a macro policy harmonization scheme that suits the specific context of each REC with adequate safeguard measure for weaker members. This needs to be accompanied by building institutions for its implementation and a firm commitment by member countries.

A closely related macro policy harmonization agenda is the issue of currency. Establishment of monetary unions in Africa can play a vital role in promoting intra-Africa trade. It is estimated, for instance, that a pair of countries that start to use a common currency may see a near doubling of their bilateral trade. Thus, the effect of currency union on trade is positive, significant and large (Rose, 2000; Rose and Engel, 2002; cited in Njinkeu and Fosso, 2006). The drastic fall in trade following the issuing of separate currencies between Ethiopia and Eritrea, and North and South Sudan is a proof of this by negation (Geda and Kayizzi-Mugerwa, 2014). The multiplicity of currencies compounded by exchange rate problems encourages underground trade in the form of smuggling, and an expanding unrecorded trade (Yeboah, 1993). Multiplicity of currencies do also raise international trade costs as businessmen are confronted with the cost of changing from one currency to another as well as the uncertainties associated with the market for these currencies.

CEMAC and UEMOA are monetary unions, with CFA franc as the common currency. Although formally differentiated, the common currencies are exchangeable between the two communities one to one and are convertible into the euro at a fixed exchange rate. Southern Africa, Namibia and Swaziland are members of a Common Monetary Area, where the South African rand circulates freely as a common currency under a floating arrangement (Mutasa, 2003). In EAC, all three founding members (Kenya, Tanzania and Uganda) have floating currencies. Most of the non-UEMOA members of ECOWAS also have floating exchange rates. Although these are encouraging trends, the multiplicity of currencies and exchange rate arrangements (including uncertainties associated with them) that exist in Africa makes a case for the establishment of clearing mechanisms in the short run and some kind of monetary union in the long run. ECOWAS and COMESA have formally established clearing houses to promote intra-community trade with the use of local currencies against a background of exchange control dictated by the scarcity of hard currencies in most countries (Masson and Pattillo, 2001). Thus it is high time to build on this fragmented and less organized, yet important initiatives and create an enabling macroeconomic environment for intra-Africa trade.

3.1.4. Other supply side challenges: trade facilitation policy measures

Trade and industrial policy making in newly developed Asian countries featured prominently in their impressive growth record of the last five decades. Supply side measures that include export promotion policies, appropriate and timely incentive structure as well as efficient bureaucracy at the service of exporters are central for their trade growth. Failure to enhance trade facilitation measures hinders trade by reducing market access through delays and higher costs; it also raises input costs eroding the competitiveness of exporting firms (Biggs, 2007; Geda, 2012).

Customs authorities in many African countries are inefficient. Overall delays at African customs are on average longer than the rest of the world: 12 days in countries south of the Sahara, compared to 7 days in Latin America, 5.5 days in Central and East Asia, and slightly more than 4 days in Central and East Europe (ECA, 2004). Such delay increases the transaction cost of trading between African countries. Generally, each day lost in transport delays due to customs and related problems is equivalent to a tax of about 0.5% (World Bank, 2005). For example, in Southern Africa delays at the main border crossing between South Africa and Zimbabwe (Beit-Bridge) amounted to six days in February 2003, leading to an estimated loss in earnings per vehicle of US$1750, which is equivalent to the cost of a shipment from Durban in South Africa to US (World Bank, 2005). Another study (Alvis 2004, cited in World Bank, 2005) indicated that crossing a border in Africa can be equivalent to the cost of more than 1000 miles of inland transportation compared to its equivalent of 100 miles in Western Europe. Border crossing delays are also linked to other trade costs as well, especially corruption in customs.

Related trade facilitation challenges include problems with international payment mechanisms, insurance requirements, and customs guarantees. Documentary credit payment which is common in many African countries is characterized by cumbersome and complex procedures. The process is time consuming and is not well understood and badly managed by many users. Indeed it has been reported that half of all requests for payment are rejected on grounds of documentary inconsistencies (Njinkeu and Fosso, 2006; Mutasa, 2003). In addition, delays and complicated procedures related to insurance of goods and customs guarantee requirements are believed to raise the cost of exports from Africa and reduce its firms’ competitiveness. It is estimated that each day of delay reduces export earnings by about 1%. With this rate, if Uganda, for instance, reduces its factory-to-ship time from the current level of 58 days to 27 days, its exports would have increased by 30% (Njinkeu and Fosso, 2006). Table 8 also shows how competitiveness in African trade could be adversely affected by the cost of doing business that includes delays in customs clearing, business regulatory environment and related macroeconomic management, where African countries hardly register good ratings.

To give a continent wide quantifiable dimension of the issues briefly outlined in this section (infrastructure, trade related policies, the ease of ‘doing business’ and the role of regional economic communities), a gravity model of determinants of African trade is specified and estimated using the PPML method (see Appendix A for model specification). The result is reported in Table 9. The model is similar to the models reported in the previous sections except that it is estimated for all African countries using data from 1993–2010. In addition to traditional gravity model variables, indicators of membership to African RECs were also included as additional regressors in this model.

Table 9 shows that all traditional gravity model variables are found to be important determinants of African trade. Among these, distance from trading partners (a proxy for cost of international transport), the income of trading partners, sharing border, and landlockedness, in the order of importance, are found to be the most important determinants of African trade. In addition, the remoteness index which captures multilateral trade resistance term has the expected positive impact on bilateral trade among African countries. Bilateral real effective exchange rate is found to be another important determinant of bilateral trade. It has a positive coefficient suggesting that depreciation of home country’s currency makes this home country competitive and increases bilateral trade. Membership in RECs seems generally to enhance trade.⁴

To sum up, this section has shown that even if African countries have the potential to trade, the export supply constraint will act as a major hindrance to realize that potential. The quantifiable dimension of these issues is also summarized using a general gravity model estimated for Africa as a whole. The result confirms not only the importance of RECs to enhance African trade but also the importance of the challenges of exporting (infrastructure and competitiveness) in determining the level of African trade. Thus, progress in intra-Africa trade and regional integration requires addressing these major challenges. To substantiate the aggregate level finding in this section, recent country case study based evidence on the issue is examined in the next sub-section.

A.1. Data source

The trade data is taken from IMFs Direction of Trade Statistics; Macro indicators are taken from the World Bank’s World Development Indicators as well as from the African Development Indicators. Remoteness index is the researchers’ own computation. The relevant data for each African country and its trading partner is compiled for the sample of countries listed below.

Sample of countries used in the gravity model by regional economic communities.