WPS4790 Policy ReseaRch WoRking PaPeR 4790 Expanding Trade within Africa The Impact of Trade Facilitation Dominique Njinkeu John S. Wilson Bruno Powo Fosso The World Bank Development Research Group Trade Team December 2008 Policy ReseaRch WoRking PaPeR 4790 Abstract This paper examines the impact of trade facilitation on agreements relate to intra-African trade flows. Using trade intra-African trade. The authors examine the role of trade data from 2003 to 2004, they find that improvement in facilitation reforms, such as increased port efficiency, ports and services infrastructure promise relatively more improved customs, and regulatory environments, and expansion in intra-African trade than other measures. upgrading services infrastructure on trade between They also show that, almost all regional trade agreements African countries. They also consider how regional trade have a positive effect on trade flows. This paper--a product of the Trade Team, Development Research Group--is part of a larger effort in the department to explore the linkages between trade facilitation, transparency, and trade costs as they affect development. The paper is part of a project funded through the Multi Donor Trust Fund on Trade and Development at the Bank on "Trade Costs and Facilitation" which received financial support from the governments of Norway, Sweden and the United Kingdom. The trade costs project website is accessible at: http://econ.worldbank.org/projects/trade_costs. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at jswilson@worldbank. org, njinkeu@ileap-jeicp.org, and bfosso@gmail.com. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Expanding Trade within Africa: The Impact of Trade Facilitation1 Dominique Njinkeu2, John S. Wilson3 and Bruno Powo Fosso4 JEL Classification Codes: F10; F15. Keywords: Trade, Africa, Regional Integration, Trade Facilitation 1This paper is part of a project at the World Bank on Trade Costs and Facilitation with support from the governments of Norway, Sweden and the United Kingdom through the Multidonor Trust Fund for Trade and Development. It is also aligned with work under a Trust Fund of the U.K. Department for International Development. The views expressed here are entirely those of the authors, they do not necessarily represent those of the World Bank, its Executive Directors, or the countries they represent. We are grateful to comments by the World Bank working paper series anonymous commentator on an earlier version of this paper 2Executive Director, International Lawyers and Economists Against Poverty (ILEAP), Toronto, Canada. Dominique.njinkeu@ileap-jeicp.org ; www.ileap-jeicp.org 3Lead Economist, Development Research Group, World Bank. 4 Policy Analyst, Human Resources and Social Development Canada (HRSDC), Gatineau (Canada). 4 1. Introduction World trade has expanded faster than economic growth and has been a vehicle for raising the standard of living in OECD countries. African countries have not seen similar trade expansion. Africa's share of world exports has dropped by nearly 60 percent---from 3.5 percent in 1970 to 1.5 percent in 1999, representing a staggering income loss of $70 billion annually, an amount equivalent to 21 percent of the region's GDP and to more than five times the $13 billion in annual aid flows to Africa (World Bank, 2003). The dismal performance of African trade can be attributed to several factors traditionally associated with growth, institutions, customs environment and infrastructure. Rodrik (1998) finds that the trade/GDP ratios of Sub-Saharan African countries are comparable to those of countries of similar size and income, and that Africa's marginalization in world trade is mainly due to low income growth. Fosu (2003) shows that coups d'état have an adverse effect on African export growth that is higher than their effect on GDP. Clarke (2005) finds that African manufacturing enterprises are less likely to export in countries with restrictive trade and customs regulation and poor customs administration. Longo and Sekkat (2004) examine the possibility of expanding intra-African trade with a gravity model, but also pay attention to obstacles to intra-regional trade. They show that insufficient infrastructure, mismanagement of economic policies and internal political tensions are the main obstacles to trade in African countries. Limao and Venables (2001) also show that poor infrastructure accounts for 40 percent of predicted transport costs for coastal countries and up to 60 percent for landlocked countries. In the case of Sub-Saharan Africa (SSA), Limao and Venables conclude that intra-SSA trade costs are substantially higher and trade volumes substantially lower than those for non-SSA countries. Also, in Africa, relatively high trading costs at the border, low resource complementarities between member countries, small market size, and poor transport infrastructure limit both intra- and inter-regional trade (Yang and Gupta, 2007; Foroutan and Pritchett,1993; Njinkeu and Powo Fosso, 2006). While several of the above fall in the overall trade facilitation debate, none of the studies has explicit focus on the competitiveness agenda of African countries as pursued in various trade facilitation options including those considered in the autonomous policy debate or ongoing multilateral (WTO) or bilateral (the Economic Partnership Agreements (EPA)) trade negotiations. This paper addresses the role of the trade facilitation in promoting intra-African trade. To do this, we extend the gravity model considered in the literature (see Wilson et al 2003a, 2003b, 2004; Clark et al 2004; and OCDE, 2007), with focus on the role of port efficiency, customs environment, regulatory environment, and services infrastructure. Using a panel data that cover years 2003 and 2004 on a sample of 100 countries which includes 25 African countries, we find that port efficiency and services infrastructure have a positive impact on African trade but that customs and regulatory environments are the factors that lower intra-African trade. The rest of the paper is organized as follows. Section 2 presents the review of literature. Section 3 describes the data and estimation framework. Section 4 discusses the results while section 5 concludes. 2. Previous studies Empirical studies on trade facilitation are very recent. Following Fox et al (2003) and OECD (2003), Denis (2006) uses the Global Trade Analysis Project (GTAP) Computable General Equilibrium model to investigate the impact of trade facilitation and regional integration in the Middle East North Africa (MENA) region. He shows that both intra-regional and integration with European Union (EU) have a favourable effect on welfare in the MENA region. With the addition of trade facilitation improvements to trade liberalization, the welfare gains from integrating with the EU increase from US $ 1.8 billion to US $ 7.2 billion (0.82 percent increase to base GDP) whereas the welfare gains from intra-regional integration increase from US $ 913 million to US $ 3 billion (0.1 percent increase to base GDP). APEC (1999) uses the same GTAP model and finds that trade liberalization and trade facilitation could increase APEC countries' income by US 75 billion (at 1997 prices, 0.4% of GDP). UNCTAD (2001) finds that a one percent reduction in the cost of maritime and air transport services in developing countries could increase global GDP by US 7 billion (1997 value). Clark et al. (2004) find that improving port efficiency from the 25th to 75th percentiles reduces shipping costs by more than 12%. Wilson et al. (2003a, 2003b, 2004) show that port efficiency of both the importer and the exporter is positively associated with trade. Moreover, comparing the effect of port efficiency on imports versus exports, they find that the coefficient is higher for exporter than importer, which implies that global trade flows get a bigger boost when the exporter's port efficiency improve. For landlocked countries, Wilson, Mann and Otsuki (2003a) show that the ports are as important for both import and export as in non-landlocked countries. For island countries, it appears that ports are more important for their import and less important for their export compared to non-island countries. Similar result was found by Limao and Venables (2001). Wilson et al. (2003a, 2003b, 2004) show that the customs environment has a significantly positive effect on trade of the importing country. Moreover, they argue that trade facilitation is a possible avenue for reducing the cost of imports through customs improvements. OECD (2007) uses a metrics of customs and administrative procedure from the World Bank "Doing Business" survey (2005) to estimate a gravity model. The study concludes that all countries can benefit from more efficient customs and administrative procedures, with the greatest benefits accruing to those reforming countries with the least efficient customs and administrative procedures. Wilson et al (2003a, 2003b, 2004) find that improving the regulatory environment of the importer and exporter has a positive and significant association with trade. De Groot et al. (2003) also show that a better quality of formal institutions tend to increase trade. Rodrik et al (2002) conclude that the quality of institutions has a significant and positive effect on country's total trade flows. Otsuki et al. (2001) find that African export of cereals, nuts and dried fruits will decline by 4.3 (cereals) and 11 (nuts and dried fruits) percents with a 10 percent tighter European Union standard on aflatoxin contamination levels of these products. Francois and Manchin (2007) also find that exports performance depend on institutional quality. Wilson et al (2003a, 2003b, 2004) use the percentage of companies that use the internet for e- commerce to assess the effect of internet. They conclude that E-business usage has a positive and significant effect on trade. Freund and Weinhold (2004) show that an increase by 10 percent in the relative number of web hosts could increase by one percent the country's trade flow. Choi (2003) shows that when the number of the internet hosts or users in a host country increased by 10 percent, the foreign direct investment inflows increased by more than 2 percent. Fink et al. (2005) suggest that a 10 percent decrease in the bilateral calling price is associated with an 8 percent increase in bilateral trade. Francois and Manchin (2007) also find that exports performance depend on communications infrastructure. In general relatively high trading costs at the border, low resources complementary between member countries, small market size, and poor transport infrastructure limit both intra- and inter-regional African trade (Yang and Gupta, 2007; Foroutan and Pritchett,1993; Njinkeu and Powo Fosso, 2006). Studies on trade facilitation in African countries are scarce. Using estimates from a gravity model, Portugal-Perez and Wilson (2008) estimate ad-valorem equivalents of improvements in national trade-cost indicators, measured by the Logistic Performance Indicators and Doing Business, for African countries. They found that the gains for African exporters from improving trade logistics half-way to the level in South Africa are more important than a substantive cut in tariffs. For example, improving trade logistics to lower costs in Ethiopia half-way to the level in South Africa is approximately equivalent to a 7.5 percent cut in tariffs faced by Ethiopian exporters. Njinkeu, Wilson and Powo Fosso (2007) present the trade facilitation agenda for Africa, and discuss alternative methodologies for measuring the impact of trade facilitation on trade. Longo and Sekkat (2004) suggest that improvements in the transport infrastructure indicators (length road per capita and number of telephone per capita) by 1 percent in one country can boost intra-African trade by about 2 percent. Using a gravity model, Limao and Venables show that poor infrastructure accounts for 40 percent of predicted transport costs for coastal countries and up to 60 percent for landlocked countries. In the case of Sub-Saharan Africa (SSA), Limao and Venables conclude that intra-SSA trade costs are substantially higher and trade volumes substantially lower than those for non-SSA countries. Clarke (2005) shows that African firms engaged in manufacturing are less likely to export in countries with restrictive trade, cumbersome customs regulation, and poor customs administration. In this paper, we use the gravity model of bilateral trade in the African countries and the Rest of the world, and include a set of trade facilitation indicators (port efficiency, customs and regulatory environment, and services infrastructure) as well as incorporate tariffs and regional trade agreements to see which of these factors might have a greater effect on intra-African trade flows. 3. Methodology 3.1 Rationale for trade facilitation measures The main impediment to quantitative analysis of trade facilitation is availability of properly defined and measured development of trade facilitation indicators. Wilson, Mann and Otsuki (2003a, 2003b, 2004) present four distinct approaches that meet policymakers' needs. They are: (i) port efficiency, (ii) customs environment, (iii) regulatory environment, and (iv) services sector infrastructure. Port efficiency (PE) is designed to measure the quality of infrastructure of maritime and air ports. Customs environment (CE) is aimed at measuring direct customs costs as well as administrative transparency of customs and border crossings. Regulatory environment (RE) is designed to measure the economy's approach to regulations. Services infrastructure sector (SI) is designed to measure the extent to which an economy has the necessary domestic infrastructure (e.g., telecommunications, financial intermediaries, and logistic firms) and is using networked information to improve efficiency and to transform activities to enhance economic activity. These trade facilitation measures have parallel with GATT articles covered in the trade facilitation negotiations of the Doha Development Agenda (DDA), and the bilateral negotiations of Economic Partnership Agreements (EPA) between African regions and the European Union. The negotiation mandate in the DDA for example is provided by Annex D of the July 2004 Framework as follows: "Negotiations shall aim to clarify and improve relevant aspects of Articles V, VIII and X of the GATT 1994 with a view to further expediting the movement, release and clearance of goods, including goods in transit.5 Negotiations shall also aim at enhancing technical assistance and support for capacity building in this area. The negotiations shall further aim at provisions for effective cooperation between customs or any other appropriate authorities on trade facilitation and customs compliance issues." WTO (2004) The port efficiency measure has been constructed in accordance with GATT article V (freedom of transit). This article deals with free movement of goods, which should be allowed 5It is understood that this is without prejudice to the possible format of the final result of the negotiations and would allow consideration of various forms of outcomes. to move via most convenient route, should be exempt from customs or transit duties, and should be free from unnecessary delays or restrictions. The definition of custom environment used in this paper addresses the concerns of GATT article VIII. GATT article VIII states that in order to minimize impediments to trade due to customs procedures, fees charged by customs officials must be limited to the approximate cost of customs services. Also, there should not be substantial penalties for minor breaches of customs regulations such as clerical errors. Regulatory environment issues are addressed in GATT article X which discusses Publication and Administration of Trade Regulations. This article comes from the basic transparency obligation that requires prompt publication of laws and regulations affecting imports and exports so that foreign governments and traders may clearly understand them. 3.2 Defining and measuring trade facilitation indicators Trade facilitation indicators, in this paper, are constructed from the Global Competitiveness Report (GCR) produces by the World Economic Forum (WEF).6 The GCR is an annual publication that aims to enhance global understanding of the factors influencing private-sector led economic growth and explains why some countries are much more successful than others at creating new employment opportunities and raising the income level of their respective populations. For participating countries, the GCR reports on performance and policy conditions affecting the ability of private sector firms to be globally competitive -- able to create and add value within the global marketplace. WEF has expanded its worldwide country coverage from 75 countries in 2001, to 125 countries in 2006's GCR. Particular attention has been paid to including more countries from Africa, as well as several from Central Asia, given the increasing importance of this region on the global stage7. This database offers then the opportunity for a focus on Africa in a manner that was not possible with previous data sets. After developing the trade facilitation measures, the next stage in creating the trade facilitation indicators involves collecting the GCR's indexed inputs into the four specific 6 Our approach is therefore different from that followed by Wilson, Mann and Otsuki (2004) who used three survey (Kaufmann, Kraay,and Zoido-Lobaton (2002), World Economic Forum Global Competitiveness Report 2001-2002, and IMD Lausanne, World Competitiveness Yearbook 2002). 7For example, the number of African countries has increased from 5 countries (Egypt, Mauritius, Nigeria, South Africa, and Zimbabwe) in 2001 to 25 countries in 2006. facilitation measures. Each index ranges from 0 to 7 with a low value meaning poor, and a higher value good performance. We use interpolation for the years where no data are available8. Therefore: Port efficiency for each country is the average of two indices namely port infrastructure quality and air transport infrastructure quality. Customs environment is the average of two indices representing irregular payments in imports and exports and the business cost of corruption. The regulatory environment is the average of three indices representing their regular payment in public contracts, favouritism in decisions of government officials, and public trust of politicians. The services sector infrastructure for each country is the average of the quality of competition in internet service providers (ISP) sector and the extent of marketing. Tables 1 present, for the trade facilitation indicators, the mean, standard deviation, minimum and maximum values for the different samples. As expected, the statistics are very low in Africa compared to other regions. The comparison is interesting if we contrast individual countries or regional integration agreements. Tables 2 and 3 show that for our sample services infrastructure indicators are highly correlated with port efficiency (0.822), customs environment (0.737) and regulatory environment (0.721). Port, customs and regulatory environment are strongly correlated with each other. Port efficiency is highly correlated with gross national income (GNI) per capita. Moreover, GNI is highly correlated with per capita GNI. When we limit ourselves to the African sample, the correlations between trade facilitation indicators are low compared to the correlations in the whole. Two reasons can explain this (Wilson, Mann, and Otsuki, 2004; therefore WMO). First explanation stems from the fact that trade facilitation indicators are different facets of overall trade facilitation and secondly some elements of trade facilitation (administrative transparency, available resources to build quality ports, and so on) are more prevalent in higher income economies than in developing countries. Port efficiency, customs environment, services infrastructure and regulatory environment may induce reforms that improve with a country's import and export flows and the estimated 8This applies for Cameroon and Senegal where we use the data of the previous year for the year 2004. coefficients for these variables would be biased particularly due to possible endogeneity associated with high correlations between the trade facilitation and income variables. WMO (2003a, 2003b) however find weak evidence of endogeneity. Other variables not explicitly associated with trade facilitation are also defined in the database. These include for example the bilateral trade flows in manufactured goods. The trade flow data aggregate the trade flows over the manufactured goods for a given importer- exporter pair. We use import data as it is likely to be more reliable than export data since imports constitute a tax base and governments have an incentive to track import data. We deflate the trade flow with the world import index taken from IFS. We use weighted average of applied tariff rates obtained from UNCTAD TRAINS for the manufactured goods under the above definition where bilateral trade values corresponding to each tariff line are used as the weighted gross national income (GNI) and GNI per capita data. Consumer price index and GDP deflator are taken from the World Development Indicators database. Geographic data and the dummies for same language and colonial links are extracted from the Centre d'Études Prospectives et d'Informations Internationales (CEPII) database9. The distance data are calculated following the great circle formula, which uses latitudes and longitudes of the relevant capital cities. 3.3. The empirical model The gravity model, is a short-hand representation of supply and demand forces in which the amount of trade between countries is assumed to be increasing in their sizes (as measured by their national incomes), and decreasing in the cost of transportation between them (as measured by the distance between their economic centers). By assuming first Cobb-Douglas preferences and then CES preferences, Anderson (1979) was the first to derive the gravity equations from models that assumed product differentiation. In both cases, he made the Armington assumption that products were differentiated by country of origin. Anderson and Wincoop (2003) developed a method that consistently and efficiently estimated a theoretical gravity equation, and correctly calculates the comparative statistics of trade frictions. Anderson and Wincoop (2004) introduced the border costs as premium on the export prices. In the so called augmented gravity models, most authors (e.g. Carrere, 2004; Musila, 2005; 9http://cepii.fr/anglaisgraph/bdd/distance.htm Frankel and al., 1995; Glick and Rose, 2002; Frankel and Rose, 2002; Rose, 2001; Rose and Engel, 2002; Longo and Sekkat 2004; Wilson et al., 2003a, 2003b, 2004) add other variables such as income per capita, adjacency, common language, common currency, or colonial links. The augmented gravity model of trade has also been used widely as a baseline model for estimating the impact of a variety of policy issues including, regional trading groups (Carrere, 2004; Musila, 2005, Longo and Sekkat 2004), political blocs (Frankel et al, 1995), currency unions (Glick and Rose, 2002; Frankel and Rose, 2002; Rose, 2001; Rose and Engel, 2002; Carrere, 2004), and trade facilitation (Limao and Venable, 2001; Wilson et al, 2003a, 2003b, 2004; Clark et al, 2004). We follow this approach. In our sample, zero or missing bilateral trade observations reaches 43.53% of the total. Since the dependant variable is truncated at zero, estimation with OLS will produce biased results. We therefore use a Tobit model and our augmented gravity is therefore as follows: ln(1+F t )= + b b ln (100+TARIFF )+b ln PE +b t CE b lnRE +b lnSI + ij 0 1 ij 2 i 3 i 4 i 5 i t + b lnPE b lnCE b lnRE b lnSI b + + + + t ) 6 j 7 j 8 j 9 j 10ln(GNI )+b ln(GNI i 11 j + b t t (DISTij)+b + 12ln(GNIPC)+b ln(GNIPC)+b ln D b D i 13 j 14 15 ADJ 16 AMU + b D + + + + + + 17 CEMAC b D 18 COMESA b D 19 ECOWAS b D 20 SADC b D 21 UEMOA b D 22 NAFTA b D 23 ASEAN + b D + + + + + + 24 LAIA b D 25 AUNZ b D 26 MERCOSUR b D 27 EU b D 28 ENG b D 29 CHN b D 30 SPN + b D + + + + + + ,t 31 FRC b D 32 ARB b D 33 GMN b D 34 POR b D 35 RUS b D +b 36 2003 i ji where i and j stand for exporter and importer respectively, and t is trading year (t=2003, 2004). Ftij denotes the value of manufacture exports from country j to i at year t. TARIFFtij is the applied tariff rate in the percent ad valorem term that is specific to trading partners i and j and year t. The terms PE, CE, RE, and SI denote country's indicators of port efficiency, customs environment, regulatory environment, and service infrastructures. The term GNI denotes the gross national income and GNIPC denotes per capita GNI. Dummy variables are included in the model to capture the effect of preferential trade agreements, language similarity and adjacency. The trade arrangements dummies (see the appendix for the definitions of different trade arrangements) include AMU (DAMU), CEMAC (DCEMAC), COMESA (DCOMESA), ECOWAS (DECOWAS), SADC (DSADC), UEMOA (DUEMOA), ASEAN (DASEAN), NAFTA (DNAFTA), LAIA (DLAIA), AUNZ (DAUNZ), MERCOSUR (DMERCOSUR), and EU (DEU). The language dummies include English (DENG), French (DFRC), Spanish (DSPN), Chinese (DCHN), Arabic (DARB), German (DGMN), Portuguese (DPOR), and Russian language (DRUS). The adjacency dummy DADJ takes the value one if country i is adjacent to country j and zero otherwise. Geographical distance between capital cities i and j is denoted DiSTii. b0 is the intercept, D2003 is a dummy for year t (t = 2003). This dummy is included in the model to control for time-specific shocks. bi is a country-specific effect when a country is an exporter. Parameters b's are the coefficients. ij(t) is the error term that is assumed to be normally distributed with mean zero. In the literature, it has been found that poorly-performing ports can strongly reduce trade volumes and may have a greater dampening effect on trade for small, less-developed countries than many other trade frictions (Wilson et al, 2003a, 2003b, 2004; Clark et al, 2004). Thus we may expect that improvement of port infrastructure affects positively the trade flows. Customs is a mandatory element in the movement of goods across borders and the procedures apply to these goods significantly influence the role of national industry in international trade and their contribution to the national economy. Effective and efficient clearance of goods increases the participation of national industry in the world marketplace and contributes to economic competitiveness of nations, encourages investment and development of industry, and increases the small and medium enterprises in international trade (World Customs Organization). Well developed institutions are likely to decrease the transaction costs for market participants and thus increase the efficiency of markets. They can do this through three channels (World Bank, 2002). (i) They decrease information asymmetries as they channel information about market conditions, participants and goods; (ii) They reduce risk as they define and enforce property rights and contracts, determining who gets what and when, and (iii) They increase competition in markets or decrease it. The improvements of the regulatory environment will affect trade positively. Services infrastructure sector (SI) is designed to measure the extent to which an economy has the necessary domestic infrastructure (e.g., telecommunications, financial intermediaries, and logistic firms) and is using networked information to improve efficiency and to transform activities to enhance economic activity. For example, the Internet can improve the productivity in three ways. (i) Internet can lower prices by lowering search costs; (ii) Internet use can cut the cost of holding inventories by allowing large suppliers to bypass retailers and contact customers directly; and (iii) Internet usage can improve the transparency of the host countries and make it comfortable to do business. It is expected that the improvement of SI will have a positive impact on the trade. Because trade flows are expected to be positively related to national incomes, and negatively related to distance, b10 and b11 are expected to be positive and b14 is expected to be negative. Trade flows are also expected to be positively associated with regional integration, language, and adjacency. In contrast to cross-section data, panel data permit more general types of heterogeneity. For a single cross-section, these controls can only depend on observed country-pair attributes such as common language, and estimates can thus be biased if there is additionally an observed component to the country-pair propensity to trade. With panel data, such heterogeneity can be controlled through a country-pair fixed effect. In this paper, we use country-specific effect when a country is an exporter. With this specification, distance, adjacency, and language are eliminated because they are fixed over time10. We use three sub-samples. The first concerns bilateral trade flows between 100 countries. The second is, a sample comprising 25 African countries (see table 8 in appendix). The last one concerns intra-African trade and trade with the rest of the world11. The coefficients of the three sub-samples are assumed stable over the period 2003-2004. 10See Cheng and Wall (2005) for the various specifications to controlling for heterogeneity in the gravity models of trade. 11Due to endogeneity problem and space, we don't present the results of the estimations with GNI per capita as regressor. However, the results with all the variables, including GNI per capita, are available on request to the authors 4. The results (i) Whole sample In this section, we extended the Wilson, Mann and Otsuki (WMO) (2003a, 2003b, 2004)12 sample to include 25 countries primarily from Africa. We have also captured the effect of African regional integration schemes. The gravity model was run using an ordinary least squares (OLS) with the Huber/White robust standard error. The coefficients of ports efficiency of both the importer and the exporter are positively associated with trade. Comparing the effect of port efficiency on import vs. exports, we find, as did WMO, that the coefficient is higher for exporter than importer (see table 4). Customs environment of the importer is significant and positively related to trade in pooled cross- section and year 2003 fixed effect regressions, while the customs coefficient of the exporter is not significant. This implies that global trade flows get a bigger boost when the importer's customs environment improves. The regulatory environment of the exporter is significant and negatively associated to trade in all regressions. The regulatory environment of the importer is insignificant. The services infrastructure coefficient of the importer is significant and positively associated with trade in all regressions, while services infrastructure of the exporter is not significant. African regional agreements, namely, CEMAC, COMESA, SADC, and UEMOA, have a positive and significant effect on trade. When we use the year 2003 and exporter fixed effects, coefficients of regional trade agreements become significant and positively related to trade. We applied the Tobit model (table 5) on the whole sample countries to take into account missing or zero trade. We also added a dummy to isolate the effect of intra-African trade. The results (see third column of table 5) show that when African countries trade among them, the world trade increases significantly. In the pooled cross-section regression, tariff is negative and significant. In the fixed effects regressions, it becomes insignificant but remains negative. For the trade facilitation indicators, port efficiency of exporter is positive and significant whereas port efficiency of importer 12The results obtained with WMO's sample are very similar with the previous one. These results are available on request to the authors. country is negative and not significant in pooled cross-section regression and year 2003 fixed effects. This means that importer country must improve their port infrastructure to be competitive in the global market. The customs environment variable of both the importer and the exporter is positively associated with trade in the fixed effect regressions. Comparing the effect of customs environment on import vs. exports, we note that the coefficient is higher for the importer than the exporter. The regulatory environment of the importer has significant and positive effect on trade whereas regulatory environment of the exporter has significant and negative effect. The services infrastructure variable of importer is significant and positive in all the regression whereas it is negative and significant for exporter country in the fixed effect regressions. When using pooled cross-section regression, CEMAC, COMESA, ECOWAS and SADC regional integration affects trade positively. The other African regional agreements (AMU, UEMOA) have not effect on trade. In the year fixed effect regression, all the African regional agreements are significant and positively associated with trade. (ii) Intra-African Trade We use only the sample of 25 African countries as reporter (exporter) and partner (importer). Using a Tobit model, table 6 shows that port efficiency of both the importer and the exporter is positively associated with trade in the pooled cross-section regression. Comparing the effect of port efficiency on import vs. exports, we note that the coefficient is higher for importer than exporter. Because the dependant variable is the value of manufacture exports, this result can be explained by the fact that African countries are net importer of manufactured goods. The customs environment of the importer has a significant and negative effect on trade. The policy implication of this result is that African countries must improve their customs environment to boost their economy. The regulatory environment of the importer has significant and positive effect on trade whereas the regulatory environment of the exporter has significant and negative effect. Only services infrastructure of the importer is positively associated with trade. The services infrastructure of the exporter is insignificant. A comparison of the results of the fixed effects and pooled cross-section shows that allowing for year 2003 fixed effect and exporter heterogeneity (see second and third column of table 6) lowers the absolute value of the trade facilitation indicators, and increase the estimated income elasticity of trade. The CEMAC, ECOWAS and UEMOA agreements are not significant and negatively related to trade in the pooled cross-section regression whereas AMU is significant and negatively related to trade. COMESA and SADC are significant and positively associated with trade. When we use the fixed effect regressions, all the African regional agreements become significant and positively related to trade. Tariff is positive and significantly associated with trade. This means that for African countries, tariff is not a constraint for trade contrary to non tariff barriers that appear to be the main challenge for African countries. (iii) Intra-African Trade with the Rest of the World We use a sample of African countries as reporter (exporter) and partner (importer) and other countries as partner (importer) only. Using a Tobit model, table 7 shows that tariff is still significant and positively related to trade. Port efficiency of both the importer and the exporter is positively associated with trade. Comparing the effect of port efficiency on import vs. exports, we note that the coefficient is higher for importer than exporter. In contrast to the previous case, customs environment of the exporter is now positive (negative) and significant in pooled cross-section (fixed effect regression). The policy implication of this result is that African countries must improve their customs environment to boost their economy. Regulatory variable of both the importer and the exporter is positively associated with trade in exporter fixed effect regression. Comparing the effect of regulatory environment on import vs. exports, we note that the coefficient is higher for exporter than importer. A comparison of the results of the fixed effects and pooled cross-section shows that allowing for year 2003 fixed effect and exporter heterogeneity (see second and third column of table 7) lowers the absolute value of port efficiency, customs environment and services infrastructure coefficients, and increases the absolute value of regulatory environment coefficient. We also observe an increase of the estimated income elasticity of trade. For regional agreements, we have the same results as described in the previous section. 5. Conclusion The aim of this paper was to examine the impact of trade facilitation on intra-African trade. Port efficiency, customs environment, regulatory environment and services infrastructure are the measures of trade facilitation used. The results show that besides the traditional determinants of bilateral trade, port efficiency and services infrastructures are the factors that have a positive effect on African trade. On the other hand, customs and regulatory environments are the main obstacles to intra-African trade. Overall a coherent and comprehensive trade facilitation agenda should be an essential element of domestic reform. Negotiations at the WTO and Economic Partnership Agreements (EPA), especially because they explicitly recognize the need for technical assistance and capacity building, offer an opportunity. The further consensus about the relevance of trade facilitation to the aid for trade initiative confirms the centrality of the above conclusion for African countries (see Mugadza, 2008). First, the customs environment is a key factor for accelerating trade among countries. Improvement of African countries' customs can generate positive spill-overs. Second, a good regulatory environment attracts foreign investment and facilitates the trade. In our sample, 43.53% of importer-exporter pairings had zero bilateral trade. Thus apart from analyzing the effects of different variables on trade, we must also focus on factors that may explain why trade does not occur at all among countries. Employing a selection model allowed us to take account of this censoring process (Francois and Manchin, 2007). Appendix: Data sources and definitions The yearly data are constructed for 100 countries and 11 regional trade agreements spanning from 2003 to 2004. These data are computed from the Global Competitiveness Report (GCR) released. by The World Economic Forum, the World Development Indicators (WDI) published by the World Bank, the International Financial Statistics (IFS) released by the International Monetary Funds, as well as the Commodity Trade (COMTRADE), and the Trade Analysis and Information System (TRAINS) published by the United Nations Conference on Trade and Development (UNCTAD). The individual countries and regional trade agreements are presented in table 8. A.1 Real GNI For all countries, we deflate Gross National Income (GNI) (source: WDI) by the all-item consumer price index (CPI) for the baseyear 2000 (source: WDI) except for Serbia and Zimbabwe. For these two countries, we use GDP deflator (source: WDI) for the baseyear 2000. A.2 Real GNI per capita For all countries, we deflate GNI per capita (source: WDI) with the CPI except for Serbia and Zimbabwe. For these countries, we use GDP deflator (source: WDI). A.3 Trade facilitation Indicators A.3.1 Port efficiency The Port efficiency for each country is the average of two indexed inputs from GCR: Port Infrastructure Quality (Port facilities and inland waterways in your country are: 1 = underdeveloped; 7 = as developed as the world's best) Air Transport Infrastructure Quality (Air Transport in your country is: 1 = infrequent and inefficient; 7 = as extensive and efficient as the world's best) A.3.2 Customs environment The customs environment for each country is the average of two indexed inputs from GCR: Irregular Payments in Imports and Exports (How common do firms in your industry give irregular extra payments or bribes connected with import and export permits: 1 = common; 7 = never) Business Cost of Corruption (Do unfair or corrupt activities of other firms impose costs on your firm? 1 = impose large costs; 7 = impose no costs/not relevant) A.3.3 Regulatory environment The Regulatory environment for each country is constructed as the average of three indexed input from GCR: Irregular Payment in Public Contracts (How commonly do firms in your industry give irregular extra payments or bribes connected with public contracts/investment projects: 1 = common; 7 = never) Favoritism in Decisions of Government Officials (When deciding upon policies and contracts, government officials: 1 = usually favor well-connected firms and individuals; 7 = are neutral among firms and individuals) Public Trust of Politicians (Public trust in the honesty of the politicians is: 1 = very low; 7 = very high) A.3.4 Services sector infrastructure Services sector infrastructure for each country is the average of two indexed inputs from GCR: Quality of competition in Internet service providers (ISP) sector (Is competition among your country's ISP sufficient to ensure high quality, infrequent interruptions and low prices? 1 = no; 7 = yes, equal to world's best) Extend of marketing (the extend of marketing in your country is: 1 = limited or primitive; 7 = high and among the world's best sophisticated) A.4 Trade Flows The bilateral trade flows are bilateral trade in manufactured goods (source: COMTRADE), defined as commodities in categories 5 to 8 in SITC 1 digit industry except those in category 68 (non-ferrous metals). The Trade flow data aggregate the trade flows over the manufactured goods for a given importer-exporter. We deflate the trade flow using the world import index (source: IFS). A.5 Tariffs We use weighted average of applied tariff rates (source: TRAINS) for the manufactured goods under the above definition where bilateral trade values corresponding to each tariff line are used as the weight. A.6 Adjacency, language, colony, and distance Geographic data, together language and colonial links are extracted from the Centre d'Études Prospectives et d'Informations Internationales (CEPII) database (see www.cepii.fr). Distance data are calculated following the geat circle formula, which uses latitudes and longitudes of the relevant capital cities. References Anderson, J.E. (1979). "A Theoretical Foundation for the Gravity Equation." American Economic Review, Vol. 69 (1): 106-116. Anderson, J.E and E.V. Wincoop (2004). "Trade Costs." Journal of Economic Literature, Vol. 42, 691-751. Anderson, J.E and E.V. Wincoop (2003). "Gravity with Gravitas: A Solution to the Border Puzzle." American Economic Review, Vol. 91 (1): 170-192. Asia Pacific Economic Cooperation (APEC) (1999). Assessing APEC Trade Liberalization and Facilitation: 1999 Update. Economic Committee, September 1999. APEC: Singapore. Bersgtrand, J.H. (1985). "The Gravity Equation in International Trade: Some Microeconomic Foundations and Empirical Evidence." Review of Economics and Statistics, Vol. 67 (3): 474- 481. Carrere, C. (2004). "African Regional Agreements: Impact on Trade with or without Currency Union." Journal of African Economies, 13(2): 199-239. Cheng, I.H.and H.J. Wall (2005). "Controlling for Heterogeneity in Gravity Models of Trade and Integration" Federal Reserve Bank of St-Louis Review, 87(1): 49-63. Choi, C. (2003). "Does the Internet Stimulate Inward Foreign Direct Investment?" Journal of Policy Modeling, 25(2003): 319-326. Clarke, G.R.G (2005). "Beyond Tariffs and Quotas: Why Don't African Manufacturers Export More?" World Bank Working Paper No. 3617, Washington D.C. Clark, X., D. Dollar, and A. Micco (2004). "Port Efficiency, Maritime Transport Cost and Bilateral Trade." NBER Working Paper, No. 10353. Coe, D.T. and A.W. Hoffmaister (1998). "North-South Trade: Is Africa Unusual?" Working Paper of the International Monetary Fund, Washington, D.C. Deardorff, A. V. (1995). "Determinants of Bilateral Trade: Does Gravity Works in a Classical World?" NBER Working Paper No. 5377. De Groot, H.L.F., G.J., Linders, P. Rietveld and U. Subramanian (2003). "The Institutional Determinants of Bilateral Trade Patterns" Tinbergen Institute Discussion Paper, TI 2003- 044/3. Denis, A. (2006). "The Impact of Regional Trade Agreements and Trade Facilitation in the Middle East North Africa Region." World Bank Working Paper No 3837, Washington, D.C. Fink, C., A. Mattoo, and H. C. Neagu (2005)." Assessing the Impact of Communication Costs of International Trade." Journal of International Economics, Vol. 67: 428-445. Foroutan, F. and L. Pritchett (1993). "Intra-Sub-Saharan African Trade: Is It Too Little?" Journal of African Economies, 2(1): 74-105. Fosu, A.K. (2003). "Political Instability and Export Performance in Sub-Saharan Africa." Journal of Development Studies, 39(4): 68-82. Fox, A., J. François, and P. Londono-Kent (2003). "Measuring Border Crossing Costs and their Impact on Trade Flows: The United-States-Mexico Trucking Car." GTAP Resources No. 1282, Purdue University. François, J. and M. Manchin (2007). "Institutional Quality, Infrastructure and the Propensity to Export." World Bank Policy Research Working Paper No. 4152, Washington, D.C. Frankel, J.A and A.K. Rose (2002). "An Estimate of the Effect of Common Currencies on Trade and Income." Quarterly Journal of Economic, 117(2): 437-66. Frankel, J., E. Stein and S-J. Wei (1995). "Trading Blocs and the Americas: The Natural, the Unnatural, and the Super-Natural." Journal of Development Economics, 47: 61-95. Freund, C. and D. Weinhold (2004). "The Effect of the Internet on International Trade." Journal of International Economics, 62(1): 171-189. Glick, R. and A.K. Rose (2002). "Does a Currency Unions Affect Trade? The Time Series Evidence." European Economic Review, 46: 1125-1151. Greenaway, D. and C. Milner (1990). "South-South Trade Theory, Evidence, and Policy." World Bank Research Observer, 5(1): 47-68. Greene, W.H. (2000). Econometric Analysis. 4th ed. New Jersey: Prentice-Hall. Limao, N. and A.J. Venables (2001). "Infrastructure, Geographical Disadvantage, Transport Costs, and Trade." World Bank Economic Review, 15(3):451-479. Longo R. and K. Sekkat (2004). "Economic Obstacles to Expanding Intra-African Trade." World Development, 32(8): 1309-1321. Mugadza, T. (2008) "Measuring The Opportunity Costs Of Trade-Related Capacity Development In Sub-Saharan Africa1" Paper presented at joint World Bank and African Economic Research Consortium (AERC) workshop on Trade Costs and the Business Environment: A Focus on Africa, Entebbe (Uganda) May 31, 2008, 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. Njinkeu, D. and B. Powo Fosso (2006). "Intra-African Trade and Regional Integration." Prepared for the ADB/AERC International Conference on Accelerating Africa's Development Five Years into the Twenty-first Century: Tunis, 22-24 November. Njinkeu, D., J.S.Wilson and B. Powo Fosso (2007). "Trade Facilitation: What Is It and How Do It Help." AERC Collaborative Research on Export Supply Response, Working Paper No. ESWP_08. OCDE (2007). "Examing the Trade Effect of Certain Customs and Administrative Procedures." OCDE Trade Policy Working paper No. 42. --------- (2003). "Quantitative Assessment of Benefits of Trade Facilitation." OCDE Document TD/TC/WP(2003)31/Final. Otsuki, T., J.S. Wilson, and M. Sewadeh (2001). "A Race to the Top? A Case Study of Food Safety Standards and African Exports." World Bank Working Paper No. 2563, Washington D.C. Portugal-Perez, Alberto and John S. Wilson (2008) "Trade Costs in Africa: Barriers and Opportunities for Reform" World Bank Policy Research Working Paper 4619 Pöyhönen, P. (1963). "A Tentative Model for the Volume of Trade between Countries." Weltwirtschaftliches Archiv, 90(1): 93-100. Rodrik, D. (1998). "Trade Policy and Economic Performance in Sub-Saharan Africa." NBER Working Paper No. 6562. Rodrik, D., A. Subramanian, and F. Trebbi (2002). "Institutions Rule: The Primacy of Institutions on Integration and Geography in Development Economics." IMF Working Paper, WP/02/189. Rose, A.K (2001). "Currency Unions and Trade: The Effect is Large." Economic Policy, 16(33): 433-62. Rose, A.K and C. Engel (2002). "Currency Unions and International Integration", Journal of Money, Credit, and Banking, 34(4): 1067-1089. Tinbergen, Jan. Shaping the World Economy: Suggestions for an International Economic Policy. New York: The Twentieth Century Fund, 1962. UNCTAD. E-commerce and Development Report. UNCTAD: Geneva, 2001. Wilson, J.S., C.L. Mann and T. Otsuki (2004). "Assessing the Potential Benefit of Trade Facilitation: A Global Perspective." World Bank Working Paper 3224, Washington, D.C. --------- (2003a). "Trade Facilitation and Economic Development: Measuring the Impact." World Bank Working Paper 2988, Washington, D.C. --------- (2003b). "Trade Facilitation and Economic Development: A New Approach to Quantifying the Impact." World Bank Economic Review, 17(3): 367-389. World Bank (2002). "Building Institutions for Market." World Development Report 2002, Washington, D.C.: The World Bank. --------- (2003). "Africa Region Trade Progress Note." Washington, D.C.: The World Bank. WTO (2004). "Doha Work Programme." World Trade Organization, WT/L/579. Yang, Y. and S. Gupta (2007). "Regional Trade Arrangements in Africa: Past Performance and the Way Forward", African Development Review, 19(3), 399-431. Table 1: Trade Facilitation summary statistics Variables Mean Std. Deviation Min Max Whole sample Port 4.548 1.195 2.350 6.850 Customs 4.757 1.115 2.500 6.800 Regulatory 3.541 1.151 1.700 6.067 Services 4.568 0.881 2.900 6.500 African sample Port 3.414 0.992 1.450 5.450 Customs 3.853 0.695 2.200 5.350 Regulatory 3.003 0.671 1.867 4.633 Services 3.482 0.714 1.550 4.950 Source: Authors' computations based on World Economic Forum Global Competitiveness Report. Table 2: Trade Facilitation Indicators for Africa in 2003 Port Customs Regulatory Services Efficiency Environment Environment Infrastructure Air Irregular Costs of Public Countries Ports Transport Payment Corruption Bribes Favoritism Trust Internet Marketing Algeria 3.00 3.00 4.50 3.70 3.50 3.80 2.30 2.70 2.50 Angola 2.00 2.70 3.40 2.80 2.80 2.10 1.50 2.70 2.20 Botswana 3.00 4.30 5.10 4.50 4.50 4.40 4.70 2.90 4.00 Cameroon 2.60 2.90 3.30 3.10 2.70 3.20 2.30 3.80 3.60 Chad 1.30 1.70 2.30 2.10 2.20 1.90 1.50 1.60 1.50 Egypt 3.90 4.00 4.40 4.00 4.50 3.80 2.90 4.70 3.80 Ethiopia 1.40 5.00 4.30 3.10 3.40 3.00 2.20 1.40 2.10 Gambia 4.10 4.90 4.60 3.80 3.80 4.10 3.30 4.30 2.70 Ghana 3.20 3.50 3.90 3.60 3.50 3.20 2.80 3.80 3.70 Kenya 2.90 4.70 3.40 2.50 2.40 2.60 2.00 3.70 4.10 Madagascar 2.10 2.90 3.50 2.40 2.50 2.60 1.70 3.60 3.30 Malawi 1.80 3.50 5.30 3.70 3.60 3.00 2.00 4.10 3.20 Mali 1.20 2.80 3.40 2.80 2.70 3.10 2.00 3.40 2.20 Mauritius 5.20 5.40 3.70 3.20 3.20 3.50 1.80 1.90 4.60 Morocco 3.70 4.30 3.60 3.20 3.00 3.20 3.10 3.70 4.40 Mozambique 2.20 3.20 3.30 2.90 2.90 2.40 1.80 3.10 2.40 Namibia 4.90 4.80 4.40 3.50 3.60 3.20 3.10 4.00 3.90 Nigeria 2.50 3.50 2.80 2.80 2.20 2.10 1.50 3.60 3.90 Senegal 3.50 4.30 3.80 3.00 2.80 3.20 2.00 3.30 3.20 South Africa 4.60 6.00 4.60 4.70 3.80 3.40 3.10 4.20 5.20 Tanzania 3.20 3.40 4.20 3.00 3.60 3.40 3.20 4.10 3.30 Tunisia 4.50 4.90 5.30 4.50 4.60 4.70 4.60 4.00 4.20 Uganda 2.10 3.20 3.20 3.00 2.60 2.60 2.00 4.10 2.80 Zambia 1.70 3.50 4.00 3.10 3.20 3.10 1.70 3.60 3.30 Zimbabwe 1.80 3.10 3.50 3.40 2.60 2.00 1.20 3.20 3.70 AMU 3.73 4.07 4.47 3.80 3.70 3.90 3.33 3.47 3.70 CEMAC 1.95 2.30 2.80 2.60 2.45 2.55 1.90 2.70 2.55 COMESA 2.49 3.80 3.87 3.12 3.08 2.83 1.92 3.30 3.31 ECOWAS 2.90 3.80 3.70 3.20 3.00 3.14 2.32 3.68 3.14 SADC 2.95 3.89 4.09 3.38 3.30 3.01 2.35 3.40 3.55 UEMOA 2.35 3.55 3.60 2.90 2.75 3.15 2.00 3.35 2.70 AFRICA 2.90 3.82 3.91 3.30 3.21 3.10 2.42 3.42 3.35 ASEAN 4.43 5.10 4.28 4.12 3.97 3.67 3.53 4.35 4.38 AUNZ 5.95 6.25 6.55 6.20 6.25 5.20 4.80 5.30 5.65 EU 5.27 5.71 5.95 5.27 5.21 4.27 3.93 4.98 5.55 LAIA 3.04 3.83 4.56 3.25 3.45 2.61 1.77 4.20 4.04 MERCOSUR 3.33 3.68 4.48 3.50 3.55 2.65 1.88 4.33 4.10 NAFTA 5.07 5.70 5.43 4.67 4.77 3.70 3.43 5.33 5.63 Source: Authors' computations based on World Economic Forum Global Competitiveness Report. Table 3: Simple Correlations Correlations: Whole Sample Trade Tariff PE CE RE SI GNI GNIPC Distance Trade 1 Tariff 0.021 1 PE 0.409 -0.085 1 CE 0.337 -0.056 0.758 1 RE 0.319 -0.077 0.774 0.880 1 SI 0.462 -0.028 0.808 0.727 0.676 1 GNI 0.613 0.068 0.552 0.419 0.400 0.655 1 GNIPC 0.447 -0.09 0.764 0.766 0.684 0.745 0.654 1 Distance -0.33 0.026 0.003 -0.034 -0.036 0.011 0.033 -0.071 1 Correlations: Intra-Africa and Rest of the World Sample Trade Tariff PE CE RE SI GNI GNIPC Distance Trade 1 Tariff 0.116 1 PE 0.338 0.106 1 CE 0.249 0.118 0.500 1 RE 0.238 0.244 0.594 0.813 1 SI 0.285 0.106 0.513 0.410 0.445 1 GNI 0.331 0.199 0.455 0.119 0.376 0.365 1 GNIPC 0.2693 0.155 0.683 0.226 0.501 0.358 0.781 1 Distance -0.209 -0.11 0.072 -0.022 -0.066 0.054 -0.038 0.012 1 Correlations: Intra-Africa Sample Trade Tariff PE CE RE SI GNI GNIPC Distance Trade 1 Tariff 0.162 1 PE 0.331 0.071 1 CE 0.221 0.061 0.459 1 RE 0.197 0.172 0.572 0.794 1 SI 0.290 0.052 0.510 0.421 0.457 1 GNI 0.298 0.159 0.456 0.036 0.352 0.380 1 GNIPC 0.224 0.113 0.675 0.159 0.479 0.339 0.796 1 Distance -0.36 0.014 0.240 0.041 0.127 0.110 0.193 0.202 1 Note: All variables are in logs. Source: Authors' computations based on World Economic Forum Global Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, and World Bank World Development Indicators for GNI. Table 4: Whole sample: OLS Pooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect Std. Std. Coeff. Error Coeff. Error Coeff. Std. Error Constant -16.687*** 1.853 -24.333*** 2.127 -25.703*** 2.101 Tariff -1.431*** 0.391 -1.886*** 0.446 -1.740*** 0.443 Ports Exporter 1.081*** 0.136 1.089*** 0.156 1.160*** 0.155 Ports Importer 0.767*** 0.150 0.725*** 0.167 0.513*** 0.167 Custom Exporter 0.403** 0.206 0.918*** 0.256 0.358 0.231 Custom Importer 0.173 0.224 -0.067 0.254 -0.146 0.245 Regulatory Exporter -0.841*** 0.155 -1.009*** 0.183 -0.736*** 0.173 Regulatory Importer 0.154 0.158 0.476** 0.177 0.569*** 0.175 Services Exporter 0.205 0.175 -0.043 0.199 0.062 0.198 Services Importer 1.064*** 0.215 0.844*** 0.228 1.062*** 0.226 GNI Exporter 0.694*** 0.014 0.666*** 0.015 0.666*** 0.015 GNI Importer 1.144*** 0.021 1.101*** 0.021 1.124*** 0.020 Distance -1.327*** 0.024 - - - - AMU 0.332 0.474 3.190*** 0.254 3.177*** 0.207 CEMAC 3.614** 1.331 6.376*** 1.153 6.723*** 1.283 COMESA 0.995*** 0.313 1.967*** 0.359 1.934*** 0.352 ECOWAS 0.201 0.548 2.454*** 0.698 2.515*** 0.702 SADC 1.956*** 0.294 3.675*** 0.321 3.517*** 0.316 UEMOA 3.614** 1.266 4.141** 1.404 3.875** 1.331 ASEAN 2.175*** 0.172 4.333*** 0.155 4.141*** 0.153 NAFTA -0.261 0.505 2.163*** 0.374 2.032*** 0.388 LAIA 0.970*** 0.118 2.740*** 0.124 2.833*** 0.129 AUNZ 1.610*** 0.262 3.528*** 0.124 3.416*** 0.251 MERCOSUR 0.322 0.250 0.911** 0.307 0.986*** 0.301 EU -0.039 0.064 2.369*** 0.062 2.351*** 0.063 English 0.679*** 0.080 - - - - French -0.068 0.156 - - - - Spanish 0.848*** 0.102 - - - - Arab 0.173 0.366 - - - - Chinese 2.731*** 0.207 - - - - German -0.536** 0203 - - - - Portuguese 1.443*** 0.283 - - - - Russian 2.006*** 0.204 - - - - Adjacency 0.720*** 0.115 - - - - Observations 19800 19800 19800 R2 0.693 0.601 0.605 Note: All non-dummy variables are in logs. The notations "*", "**", and "***" denote significance at 10, 5 and 1 percent levels, respectively. Source: Authors' computations based on World Economic Forum Global Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, and World Bank World Development Indicators for GNI. Table 5: Whole Sample: Tobit Model Intra Africa Trade Year 2003 and Exporter Dummy and 2003 Pooled Cross-Section Fixed Effects Fixed Effect Std. Coeff. Std. Error Coeff. Std. Error Coeff. Error Constant -54.465*** 4.770 -80.233*** 4.983 -80.914*** 4.994 Tariff -2.168** 1.021 -1.248 1.069 -1.432 1.072 Ports Exporter 3.367*** 0.372 3.154*** 0.388 3.152*** Ports Importer -0.413 0.361 -0.551 0.373 -0.687* 0.374 Custom Exporter -1.115** 0.551 2.489*** 0.626 3.237*** 0.634 Custom Importer 1.905*** 0.536 3.690*** 0.587 4.375*** Regulatory Exporter -1.022** 0.419 -2.793*** 0.450 -3.399*** 0.457 Regulatory Importer 1.973*** 0.394 1.300*** 0.416 0.779* 0.421 Services Exporter -0.226 0.453 -1.175** 0.475 -0.812* 0.477 Services Importer 5.591*** 0.448 4.819*** 0.466 5.195*** 0.468 GNI Exporter 1.323*** 0.033 1.321*** 0.034 1.310*** 0.034 GNI Importer 1.947*** 0.035 1.908*** 0.036 1.907*** 0.036 Distance -2.227*** 0.079 - - - - AMU -0.625 2.268 8.241*** 2.068 - - CEMAC 9.660* 4.857 16.313*** 5.054 - - COMESA 3.411*** 0.587 4.959*** 0.607 - - ECOWAS 1.969* 1.197 5.991*** 1.247 - - SADC 4.363*** 0.551 7.391*** 0.562 - - UEMOA 3.316 3.713 5.543 3.841 - - ASEAN -2.368** 1.120 1.518 1.158 1.808 1.164 NAFTA -3.757* 1.985 -0.506 2.067 -0.604 2.076 LAIA 2.335*** 0.568 7.310*** 0.528 7.203*** 0.530 AUNZ 0.497 4.824 3.325 5.049 3.314 5.072 MERCOSUR 1.160 0.999 2.227** 1.026 3.216*** 1.026 EU -0.748** 0.306 3.288*** 0.283 3.248*** 0.285 English 0.854*** 0.191 - - - - French 1.692*** 0.494 - - - - Spanish 4.397*** 0.346 - - - - Arab 3.739** 1.257 - - - - Chinese 1.795 1.366 - - - - German -2.612 1.630 - - - - Portuguese 5.865*** 0.809 - - - - Russian 5.592*** 1.267 - - - - Adjacency 0.446 0.355 - - - - Intra Africa trade dummy - - - - 3.770*** 0.257 Observations 19800 19800 19800 Log-likelihood -46159.607 -46871.893 -46940.51 Note: All non-dummy variables are in logs. The notations "*", "**", and "***" denote significance at 10, 5 and 1 percent levels, respectively. Source: Authors' computations based on World Economic Forum Global Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, and World Bank World Development Indicators for GNI. Table 6: Intra Africa: Tobit Model Pooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect Std. Coeff. Error Coeff. Std. Error Coeff. Std. Error Constant -66.081*** 20.523 -115.210*** 20.988 -98.439*** 21.256 Tariff 10.362** 4.052 12.268** 4.312 8.256* 4.457 Ports Exporter 4.542*** 1.458 1.216 1.516 1.386 1.487 Ports Importer 7.194*** 1.547 6.352*** 1.515 5.815*** 1.503 Custom Exporter 3.153 2.941 3.146 3.439 2.365 3.114 Custom Importer -6.430** 3.02 -6.367* 3.537 -7.136** 3.212 Regulatory Exporter -10.885*** 2.656 -9.506*** 2.887 -9.139*** 2.801 Regulatory Importer 5.461** 2.702 4.738 2.957 6.768** 2.904 Services Exporter 0.711 1.454 1.368 1.550 1.512 1.541 Services Importer 17.971*** 1.818 17.899*** 1.870 15.121*** 1.990 GNI Exporter 0.850*** 0.177 1.058*** 0.187 1.044*** 0.185 GNI Importer 0.398** 0.174 0.588*** 0.181 0.738*** 0.186 Distance -4.163*** 0.642 - - - - AMU -5.577* 3.258 9.901*** 2.696 10.009*** 2.679 CEMAC -2.470 4.999 8.970* 5.231 9.764* 5.135 COMESA 2.628*** 0.844 4.186*** 0.866 4.368*** 0.860 ECOWAS -1.691 1.700 3.292** 1.618 3.482** 1.607 SADC 1.519* 0.873 5.603*** 0.805 5.412*** 0.801 UEMOA -0.122 4.700 7.318 4.821 6.714 4.793 English -1.211* 0.667 - - - - French 5.342*** 1.109 - - - - Arab 4.853** 2.318 - - - - Portuguese 5.301*** 1.700 - - - - Adjacency 2.223* 1.167 - - - - Observations 1200 1200 1200 Log-likelihood -2776.061 -2768.998 -2770.627 Note: All non-dummy variables are in logs. The notations "*", "**", and "***" denote significance at 10, 5 and 1 percent levels, respectively. Source: Authors' computations based on World Economic Forum Global Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, and World Bank World Development Indicators for GNI. Table 7: Intra Africa and Rest of world (only as partner): Tobit Model Pooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect Std. Std. Std. Coeff. Error Coeff. Error Coeff. Error Constant -119.909*** 10.678 -162.977*** 10.540 -149.413*** 10.695 Tariff 11.058*** 2.160 14.477*** 2.234 11.486*** 2.303 Ports Exporter 2.461** 0.937 1.255 0.970 1.436 0.965 Ports Importer 8.615*** 0.747 8.453*** 0.754 8.107*** 0.750 Custom Exporter 5.164*** 1.379 -5.863*** 1.535 -7.039*** 1.428 Custom Importer 1.415 1.511 2.091 1.725 0.631 1.572 Regulatory Exporter 4.118*** 1.049 6.485*** 1.115 7.033*** 1.078 Regulatory Importer 0.309 1.350 0.863 1.441 2.871** 1.420 Services Exporter -0.9078 1.125 -1.494 1.182 -1.252 1.175 Services Importer 14.055*** 0.855 13.514*** 0.879 11.262*** 0.945 GNI Exporter 1.984*** 0.089 1.900*** 0.091 1.896*** 0.091 GNI Importer 0.828*** 0.085 0.973 0.088 1.087*** 0.091 Distance -2.978*** 0.265 - - - - AMU -5.527** 2.928 8.173*** 2.539 8.214*** 2.529 CEMAC 3.172 4.826 15.193*** 4.966 15.568*** 4.900 COMESA 6.296*** 0.763 8.921*** 0.774 9.075*** 0.773 ECOWAS 3.373** 1.525 8.249*** 1.547 8.310*** 1.539 SADC 5.219*** 0.753 9.490*** 0.716 9.310*** 0.715 UEMOA -0.114 4.601 7.075 4.680 6.551 4.663 English 1.239*** 0.381 - - - - French 5.001*** 0.745 - - - - Arab 4.528** 1.805 - - - - Portuguese 6.060*** 1.055 - - - - Colony dummy 1.075 1.146 - - - Adjacency 2.731** 0.968 - - - - Observations 4950 4950 4950 Log-likelihood -10283.827 -10436.04 -10424.171 Note: All non-dummy variables are in logs. The notations "*", "**", and "***" denote significance at 10, 5 and 1 percent levels, respectively. Source: Authors' computations based on World Economic Forum Global Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, and World Bank World Development Indicators for GNI. Table 8: Sample and Definition of the Regional Trade Agreements Regional Trade Blocs Countries Members Main dates* CEMAC (Economic and Monetary Cameroon, Chad 1962, creation of the UDE (Equatorial Community of Central Africa) Customs Union); 1964, revision and creation of the UDEAC (Central African Customs and Economic Union); 1973, revision of the UDEAC; end of the 1980s, failure of the compensation funds; 1994, revision and creation of the CEMAC ECOWAS (Economic Community of Gambia, Ghana, Mali, Senegal, Nigeria 1975, creation of the ECOWAS; 1990, West African States) start of a more general scheme of intra- regional liberalization COMESA (Common Market for Angola, Egypt, Ethiopia, Kenya, 1981, creation of the PTA (Eastern and Eastern and Southern Africa) Madagascar, Malawi, Mauritius, Southern African Preferential Trade Uganda, Zambia, Zimbabwe Area); end of 1980s, first reduction in the customs tariffs on intra-regional trade; 1993, revision and creation of the COMESA SADC (Southern African Development Angola, Botswana, Madagascar, 1980, creation of the SADCC Community) Malawi, Mauritius, Mozambique, (Southern African Development Namibia, South Africa, Tanzania, Coordination Conference); 1992, Zambia, Zimbabwe revision and creation of the SADC AMU (Arab Maghreb Union) Algeria, Morocco, Tunisia 1989, creation UEMOA (West African Economic and Mali, Senegal 1994, creation Monetary Union) NAFTA (North American Free Trade Canada, Mexico, United States of 1992, creation Agreement) America LAIA (Latin American Integration Argentina, Bolivia, Brazil, Chile, 1980, creation Association) Colombia, Ecuador, Mexico, Paraguay, Peru, Uruguay, Venezuela MERCOSUR (Southern Common Argentina, Brazil, Paraguay, Uruguay 1991, creation Market) AUNZ Australia, New-Zealand ASEAN (Association of Southeast Indonesia, Malaysia, Philippines, 1967, creation Asian Nations) Singapore, Thailand, Vietnam EU (European Union) Austria, Belgium, Czech Republic 1951, creation of the European Coal (since 2004), Denmark, Estonia (since and Steel Community; 1957, 2004), Germany, Greece, Hungary establishment of the European (since 2004), Italia, Ireland, Finland, Community; 1994, revision and France, Latvia (since 2004), Lithuania creation of the European Union (since 2004), Luxembourg, Netherlands, Poland (since 2004), Portugal, Slovak Republic (since 2004), Slovenia (since 2004), Spain, Sweden, United Kingdom Other countries Bangladesh, Bulgaria, China, Costa Rica, Dominican Republic, El Salvador, Guatemala, Honduras, Hong Kong, Iceland, India, Israel, Jamaica, Japan, Jordan, Korea, Macedonia, Malta, Nicaragua, Norway, Pakistan, Panama, Romania, Russian Federation, Serbia, Sri Lanka, Switzerland, Taiwan, Trinidad and Tobago, Turkey, Ukraine *The dates for CEMAC, ECOWAS, COMESA and SADC are drawn from Carrere (2004).