afghan trade policy

7/30/2019 Afghan Trade policy

1/33

1

Trade Facilitation: What Is

It and How Does It Help?

ByDominique Njinkeu,

John Wilsonand Bruno Powo Fosso

Paper prepared for AERC collaborative research project on Supply ResponseComments [email protected]

We most grateful to comments from xxMatthais Helbleas well as participants at the AERCcollaborative research workshop on Export Supply Response, Capacity Constraints inAfrica, April 23-24, 2007in Dar es Salaam (Tanzania)

July 18, 2007


2/33

2

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---from3.5 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 morethan five times the $13 billion in annual aid flows to Africa (World Bank, 2003). Thedismal performances of African trade can be attributed to several factors traditionallyassociated with trade facilitation such as complex customs requirements, lengthy andnon-transparent bureaucratic procedures associated with the movement of goods andservices across international borders. These trade impediments could be compoundedwhen countries are parties to several non-functioning regional and bilateral tradeagreements, leading to significantly high cost of doing business and competitiveness.

African countries, from colonial times attached great importance to tradefacilitation. Autonomous trade liberalization comprised major customs reforms. Assummarized by a recent COMESA-EAC-SADC task force (2007)1 The RegionalEconomic Communities (RECs) of COMESA, SADC and EAC have long recognised theimportance of improving trade facilitation (amongst other issues) in the context of deepening regional integration and in reducing the costs of cross-border transactions andso improving economic livelihoods. As such, the RECs have supported a number of tradefacilitation instruments such as regional customs bond guarantee systems, a regional 3rd party vehicle insurance scheme, harmonised axle loads and vehicle dimensions, a singlecustoms document, harmonised customs procedures, regional carriers license, etc.

A central question is however, whether the benefits outweigh the cost. Even whena government can afford some key trade facilitation reform elements it may be requiredto forgo other more pressing needs. The choices can be politically and socially difficult,even when on economic ground things may be evident. As a result, despite significantadvances in key areas of the trade facilitation reform agenda, countries have beenreluctant to enter into a process of negotiated binding agreements whether at themultilateral or the bilateral levels. Such reluctance has three motivations. First, tradefacilitation measures are very costly and require very efficient institutional setting. Anegotiated trade facilitation agreement, especially in a multilateral setting, would meanmajor legal and institutional reforms and some of these are beyond the capabilities of African governments. Second, given the weak legal and institutional framework, even inthe event that the required reforms are undertaken the process could take a long time inmost countries. Moreover, the implementation of trade facilitation agreements will bringtogether private sector interests from various legal regimes and contractual arrangements.As such, unless adequate safeguards are introduced in a negotiated and binding

1 COMESA-EAC-SADC Task Force (2007) Deepening Regional Integration: Developingand Upgrading of Transport Systems and Infrastructure on the North-South and Dar-es-Salaam Corridors.


3/33

3

agreement on trade facilitation, the likelihood of governments being taken to the disputesettlement is high; yet the human and financial resource for adequate participation in atrade dispute settlement is far beyond what African and other low-income countries canafford. Third, trade facilitation is a fast evolving and complex area and African policymakers have very limited understanding of a typical trade facilitation agenda in

bilateral or multilateral agreement; they do not have proper understanding of the link toongoing reform process, primarily as part of regional integration programs. Compoundedwith this, research in the trade facilitation field on Africa or by African scholars is verylimited, which further increases the uneasiness of African governments to engage innegotiated trade facilitation agreements.

Against this backdrop, the objective of this paper is to characterize a tradefacilitation agenda and illustrate its benefits. A range of questions need to be answeredsatisfactorily, including the following. What are the institutional reforms that should precede commitments and what are the associated political economy implications? Whatare the effects of trade facilitation procedures on trade, GDP? Do the benefits outweighthe costs? For a particular country, what is the extent of reform necessary to generate thelevel of trade expansion that can lead to the desired welfare impacts? The next section provides alternative definitions of trade facilitation, followed by a review of the relevanttheoretical and empirical literature on trade facilitation. The empirical results, because of space and time constraints do not address all the above issues, rather we focus on theextension of the gravity model based on a two-year panel data spanning 2003 and 2004on a sample of 100 countries which includes 25 African countries. We conclude with proposals of an agenda on the research and policy fronts.

I: Characterizing a Trade Facilitation Agenda for Africa:Trade facilitation could be considered broadly, encompassing the environment in

which trade transactions take place and therefore trade facilitation reform will payattention to the whole process aiming at streamlining trade procedures in order to reducethe risk and transaction cost, measured in time and money, associated with internationaltrade. Trade facilitation would entail availability and cost of essential services requiredfor import and export of goods and services, import and export procedures (e.g. customsor licensing procedures); transport formalities; and insurance, and payment requirements.As such trade facilitation is a comprehensive and integrated approach to reducing thecomplexity and cost of the trade transactions process, and enhancing the efficiency,transparency and predictability of international trade. Trade facilitation brings around thesame theme various stakeholders in the public and private sectors, domestic andinternational; many of which not traditionally associated with international tradenegotiations. With the interplay of these stakeholders, a successful trade facilitation program can help define a consistent, transparent and predictable trading environment based on internationally accepted norms and practices characterized by four features:

Rationalization of procedures and regulations leading to review of existingmeasures to ensure outdated procedures are dropped or modernized.

Simplification of formalities and procedures leading to easier but effectivecontrols in trade transactions.


4/33

4

Harmonization of applicable laws and regulations to introduce consistency andclarity.

Automation of processes and improvement of physical infrastructure and facilitiesthrough effective use of information and communication technology.

These elements are differently considered in multilateral, bilateral and regionalnegotiations ; with the agenda in the negotiations context usually more focused.

I.1 The multilateral agendaThe trade facilitation agenda at the multilateral level is limited. The July 2004 framework (WTO (2004) starting the negotiation towards a potential multilateral trade facilitationagreement is limited to three Articles of GATT, including Article V on freedom of transit; Article VIII dealing with fees and formalities connected with importation andexportation; and Article X publication and administration of trade regulations. Each of these is briefly reviewed.

Article V has three main obligations. Firstly, that traffic in transit should not besubject to unnecessary delays or restrictions and should be exempt from customs duties,transit duties and other charges. Secondly, charges such as those associated withtransportation or administrative expenses need to be reasonable (transportation chargesapply to products being transported on the same route under like conditions). Thirdly products in transit should be given most favored nation (MFN) treatment. This wouldrequire each member to designate roads and routes for goods in transit as well as providefor the necessary facilities such as customs and excise warehouses and container terminals where containers may be landed for transit, for carriage or for delivery to acontainer depot. An associated requirement is to endeavor to minimize delays. For mostAfrican governments to be able to move to world standards the pressing needs are

connected to infrastructure development and improvement of services, the creation of training centres and the training of personnel.Article VIII of GATT deals with fees and formalities for importation and

exportation. Members are required to keep, at a reasonable level, fees for governmentregulatory activities performed in connection with the importation and customs entry processes , such as the processing and clearing of documents and goods, and inspections.A second requirement is the need to reduce the number and range of fees and charges, aswell as minimizing the incidence and complexity of import and export formalities and todecreasing and simplifying import and export documentation requirements. A necessarycondition for compliance is the ability to distinguish between import, export and exciseduties and anti-dumping duties and countervailing duties on the one hand, and customsrelated service charges, such as import and export entry transaction fees, inward andoutward cargo transaction fees, licensing fees, stamp fees etc on the other. Despite progress in most African countries the region lacks such ability. Numerous documentsare still required. Many of the issues covered in Article VIII have been dealt with in someAfrican countries e.g through the 1999 Revised Kyoto Convention of the World CustomsOrganization (WCO) covering pre-arrival clearance, post audit systems, risk managementand single administration document (SAD). Most countries and regional integrationschemes have adopted the ASYCUDA system to upgrade their customs administration.


5/33

5

As a result the only impediment to making binding commitment in the negotiations is dueto limited human, institutional and financial resources.

Article X (Publication and Administration of Trade Regulations) imposes oncontracting parties four main obligations. First, is the obligation to promptly publish andin the relevant media all decisions affecting international trade policy such as laws,

regulations, judicial decisions and administrative rulings of general application that affectimports and exports. A second obligation is for this official publication to precedeenforcement. Third obligation is the need to administer laws, regulations, decisions andrulings related to imports and exports in a uniform, impartial and reasonable manner.Fourth is the obligation to maintain or institute independent judicial, arbitral or administrative tribunals or procedures for review and correction of administrative actionrelating to customs operations. Overall the main intention of Article X is to provide to the business community the basic rule of law and administrative justice procedures andguarantee predictability and fairness. All African governments have moved with somesort of reform in this area. Negotiations in the World Trade Organization (WTO)framework would aim at limiting policy reversals. It would be expected that theadministration of trade laws and regulation do not hinder the activities of traders anyfurther than it needs to in order to achieve their stated purpose. Predictability requiresindependent judicial, arbitral or administrative tribunals or procedures for review andcorrection of administrative actions in custom transactions. Relevant independentagencies need to be created and provided with the means for delivering on their mandates. Such agencies cannot be given extensive discretionary powers.

Other GATT agreements also deal with trade facilitation related issues such as theAgreement on customs valuation; the Agreement on Import Licensing, Agreement onPreshipment Inspection; the Agreement on Rules of Origin; the Agreement on TechnicalBarriers to Trade; or the Agreement on the Application of Sanitary and PhytosanitaryMeasures.

I.2 Trade Facilitation in Bilateral NegotiationsTrade facilitation has been at the center of most African regional trade integration

schemes, however although it is an important element of the trade arrangements theseregions enter into, especially with their main trading partner the European Union (EU)the provisions on trade facilitation are not in binding terms. The only existing agreement between an African region and the EU is the Trade, Development and CooperationAgreement (TDCA) with South Africa. This Agreement does not have explicit provisionon trade facilitation. Trade facilitation-related aspects of this agreement pertain tocooperation between the customs administrations, particularly with respect to customsduties and mutual administrative assistance in customs matters. The same approach islikely to apply to the ongoing Economic Partnership (EPA) negotiations.

The draft EPA between the EU and the Eastern and Southern Africa (ESA) andSouthern African Development Community (SADC) does not have a chapter per se ontrade facilitation but it is subsumed under trade-related areas. These regions seem driven by the concern of using the negotiations not to enter into binding obligations but rather amechanism for securing technical assistance to help put together efficient tradefacilitation services. The emphasis is on cooperation aiming at strengthening the capacityof ESA countries and region to deal with all aspects of trade facilitation such as customs


6/33

6

valuation, procedures and documentation, reforms and modernization; cooperation between customs authorities and border agencies; regional transit arrangement; efficienttransport systems and networks. The envisaged cooperation in the EPA context isexpected to enhance customs reform programs already initiated as part of ongoingnational or regional programs. The intended objective appears to be capacity building in

core areas such as information and communication technology infrastructure along transitcorridors; assistance for the implementation of an enhanced regional trade facilitation program; assistance for clearing and forwarding activities; or more broadly humanresource development in trade facilitation.

In the case of West Africa, trade facilitation is to be aligned to the regionalintegration process where trade facilitation is the centerpiece of efforts towards a WestAfrican free trade area. The proposed strategy for trade facilitation includessimplification and harmonization of procedures and regulations, starting with aninventory of the facilities available to the ports authorities, the transport system, thecustoms authorities and the national guarantor for the Inter-State Road Transit System.This will be achieved through harmonization of clearing and other customs formalities inthe Economic Community of West African States (ECOWAS). The EPA focuses on thesimplification of procedures; administrative cooperation and information exchange;implementation of rules decided bilaterally or multilaterally; cooperation in customsvaluation, among other things.

In Central Africa, trade facilitation was identified right from the beginning as themain objective of the region in the EPA negotiations. A broad concept guides thisgrouping where, in addition to issues of interest to West African above, sanitary and phyto-sanitary measures (SPS) and technical barriers to trade (TBT) are also included.

Overall both in the EPA and WTO negotiations the mandate of the negotiationscan help African countries address enhance their supply response capabilities (WTO2004). With the decision taken at the Fifth WTO ministerial conference in Hong-Kong tolaunch aid for trade the main building blocks for a lasting solution have been put in place.The challenge is now proactively determine an adequate work program which itself needto be supported by institutional and quantitative examination of options. The next sectionreviews selected methodologies for addressing the issues at the center of the tradefacilitation agenda.

II. Methodological Approaches to Addressing Trade FacilitationChallenges

II.1 Institutional Analysis:

To ensure trade facilitation reforms are consistent with long-term developmentaspirations of African countries, it is essential to put in place an adequate institutionalframework that enables implementation of a negotiated trade facilitation agreement. Thiswould require determining the readiness of the implementing agencies; exploring thedifficulties and challenges associated with implementation, and propose and cost


7/33

7

additional reforms that are required for the smooth implementation; and identifyingacceptable elements of the trade facilitation agreement (TFA).2

An exploratory exercise, undertaken for Kenya (CETFRO, 2006), provides indication of what can be done3.. It was perceived right from the beginning of the Doha Development

Agenda (DDA) that full and effective implementation of trade facilitation (TF) measureswas likely to play a major role in improving Kenya's trade logistics. For example theinfrastructure along the Northern Corridor is poor due to very low levels of maintenanceand investment over a long period; regulations and procedures are complicated and notsuitable for the modern trading environment. This weakness slows down the movementof goods. Several agencies organize working committee meetings involving other agencies and stakeholders to discuss matters concerning their operations. However, thereis no adequate consultation and co-ordination at the national level. Since theimplementation of any TF measure involves more than one implementing agencies,without adequate consultation and co-ordination at the national level there is likely to beoverlap, duplication of effort, conflicts, wastage in the application of resources, poor working relationships: all resulting in slow movement of goods and therefore increasedtransaction cost. Overall the institutional reform can be structured around the following:

Road transport , customs interf ace and tr ansit r egulation

Actions in this area could cover four areas (a) harmonization of road transportregulation; (b) streamlining of transit procedures including the implementation of aregional bond guarantee system, the single custom document, and the creation of one-stop windows for customs and other formalities for transit to landlocked countries; (c)capacity building to implement the program, and the establishment of national andregional facilitation committees; and (d) modernization of the computerized custominformation systems, focusing on the provision of interfaces between the differentsystems existing in the region.

Corri dor eff iciency

Related to the above is the need to identify key corridors in each region and put together adequate coordination mechanism, a monitoring scheme to, on a timely basis, identifyand address obstacles to transit along corridors and corridor performance. An importantelement of this process would be the securitization of transit through electronicmonitoring of truck movements and sealing of trucks used for international transit; the

2 The Trade and Finance and Trade Facilitation Division of the WTO has developed auseful instrument for such an analysis (see WTO (2007)).

3 See also Diagne (2005) and Ongolo and Berets (2005) who undertook similar analysis for the ECOWASand CEMAC regions respectively.


8/33

8

improvement of facilities at borders and of juxtaposed border posts; and monitoring of regional ports efficiency.

I mprovement of r egional and coordination and implementati on

This could include rehabilitation as well as an efficient monitoring of regional roadscondition. Regional economic community (REC) will have to play a central role and as aresult there is need for strengthening their role and capacity in policy implementation,including setting up of sanction mechanisms. REC could also consider the establishmentof a Regional Agency for Trade Facilitation and Transit that would be charged primarilywith implementation and monitoring of TF-related decisions. A particular emphasis willneed to be attached to investment in information and communication technologies (ICT).To ensure the process is sustainable there would be a need for a network of TF expertswith an appropriate input mechanism to high-level decision-makers. In addition, such anetwork should seek to improve cooperation between public and private stakeholdersinterested in efficient international transit by creating regional and national facilitationcommittees or corridor-based coordination bodies

I denti fi cation of TF Needs and Prior iti es Pri or to Commitments

Technical assistance to support the process of identifying needs and priorities along, for example, the WCO Self-Assessment Checklist. African countries can assess needs and priorities through integrated regional and country-specific projects such as theestablishment of National Trade Facilitation Bodies and National Trade and TransportCommittees with adequate representation of public and private sectors.


9/33

9

II.2 Empirical work

II.2.1 Measuring trade facilitationA centerpiece of any empirical work on trade facilitation is the definition and

compilation of relevant indicators. Wilson, Mann and Otsuki (2003a, 2003b, and 2004) present four indicators that simultaneously capture the main policy concerns: portefficiency, customs environments, regulatory environment, and quality of informationand communication technology infrastructure. Various approaches have been followed inmeasuring each of these.

The port efficiency (PE) measure is constructed in accordance with GATT articleV (freedom of transit) that regulates the free flow of goods across borders. PE is a proxyfor the impediments to transit in port operations. Poorly-performing ports have beenshown to strongly reduce trade volumes and having a greater dampening effect on trade(Wilson et al, 2003a, 2003b, 2004; Clark et al, 2004); hence we may expect thatimprovement of port efficiency affects positively trade flows. The compilation of PEindicators has been driven by data availability. One approach, followed by Clark and al(2004), relies on two proxies (i) the total square number of largest seaports by country,normalized by the product between foreign countrys population and area; (ii) the ratio of foreign countries GDP per capita on the ground that countries GDP are correlated withtheir level of infrastructure. A second approach relies on firm surveys such as the GlobalCompetitiveness Report4(GCR) (Wilson et al, 2003a, 2003b, 2004; and Clark et al,2004). The GRC provides indexes based on micro-data from annual surveys at firm level,on a representative group of enterprises. PE in the GRC can be appraised by port facilitiesand inland waterways, and air transport indexes. A third approach is econometrically based. Blonigen and Wilson (2006) provide a new statistical method of uncovering portefficiency measures using U.S. Census data on imports to U.S. ports districts. Thisapproach, provides standards errors of estimated port efficiency measure, and usesreadily available and high quality data to estimate port efficiencies across hundred of ports over a number of years. It also has the flexibility to provide port efficiencycomparisons on a commodity-by-commodity basis. A fourth approach, used by Blonigenand Wilson (2006), relies on the port-to-port nautical miles and 6-digit HS product codereleased by National Data Center and Army Corps of Engineers, both in the UnitedStates. The set of fixed-effects parameters that estimate the separate impact of each U.S. port on imports charges, holding all other factors constant, represent the estimatedmeasures of U.S. ports efficiencies.

Custom environment is the main concern of GATT article VIII calling for streamlining customs procedures, limiting fees charged by customs officials. TheCustoms environment (CE) indicator measures direct customs costs as well asadministrative transparency of customs and border crossings. Customs is a mandatoryelement in the movement of goods across borders and the procedures applied to these

4 The Global Competitiveness Report is an annual publication of the World Economic Forum that enhancesglobal understanding of the factors influencing private-sector led economic growth and explains why somecountries are much more successful than others at creating new employment opportunities and raising theincome level of their respective population.


10/33

10

goods significantly influence the role of national industry in international trade and their contribution to the national economy. Effective and efficient clearance of goods increasesthe participation of national industry in the world marketplace and contributes toeconomic competitiveness of nations, encourages investment and development of industry. CE can be derived from surveys such as: i) GCR (Wilson et al, 2003a, 2003b,

2004), ii) Doing Business (OECD, 2007), and iii) Investment Climate Survey (Clarke,2005). Wilson et al (2003a, 2003b, 2004) use an average of four indexes from GCR. Thefirst index assesses the extent to which irregular, additional payments are paid in order toobtain import or export permit, business licences, exchange controls, tax assessment, police protection, or loan application. The second focuses on whether import fees are perceived to be low; while the third dimension is whether hidden import barriers other than published tariffs and quotas are perceived as an important impediment; and thefourth index is the perception of prevailing level of corruption. OECD (2007) uses ametrics of customs and administrative procedure from the World Bank "Doing Business"survey (2005). Clark (2005) uses customs and trade regulation from the World Bank Investment Climate Surveys.

Regulatory environment issues are dealt with in GATT article X that requires prompt publication of laws and regulations affecting imports and exports so that foreigngovernments and traders may clearly understand them. Well developed institutions or regulatory environment (RE) that characterizes the economys approach to regulationsare likely to decrease the transaction costs for market participants and thus increase theefficiency of markets. They can do this through three channels (World Bank, 2002)(i)they decrease information asymmetries as they channel information about marketconditions, participants and goods; (ii) they reduce risk as they define and enforce property rights and contracts, determining who gets what and when, and (iii) theyincrease competition in markets or decrease it. In GCR, the regulatory environment hasthree components. The first is the transparency and stability of the regulatoryenvironment. The second component is regulatory standards, with focus on whether product, energy, safety and environment standards are perceived to be among the worldsmost stringent. The third is the extent to which compliance with internationalenvironment agreements is a high priority in countrys government. De Groot et al (2003)use the database on governance indicators developed by Kaufman et al (2002) to examinethe link between trade and institutions. This database is derived from firm survey on threeaspects: (i) Governance effectiveness which refers to quality of public service provision, the quality of the bureaucracy, the competence of civil service officials, theindependence of civil service from political pressures, and the credibility of thegovernments commitment to policies. It is therefore the measure of the quality of government inputs; (ii) Rule of Law which is based on several indicators whichmeasure the extent to which agents have confidence in and abide by the rules of society.These include perceptions of the incidence of crime, the effectiveness and predictabilityof the judiciary, and the enforceability of contracts; (iii) Control of corruption whichmeasures perceptions of corruption, conventionally defined as the exercise of publicservice for private gain. Wilson et al (2003a, 2003b, 2004) use the GCR survey indexes.Franois and Manchin (2006) use the Fraser Institute data5.

5 See Gwartney, James and Robert Lawson (2006). Economic Freedom of the World: 2006 Annual Report. Vancouver: The Fraser Institute. Data retrieved from www.freetheworld.com.


11/33

11

The last measure of trade facilitation is the extent of information andcommunication technology or E-business. E-business (EB) is designed to measure theextent to which an economy has the necessary domestic infrastructure (e.g.,telecommunications, financial intermediaries, and logistic firms) and is using networkedinformation to improve efficiency and to transform activities to enhance economic

activity. For example, the Internet can improve the productivity in three ways. (i) Internetcan lower prices by lowering search costs; (ii) Internet use can cut the cost of holdinginventories 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 itcomfortable to do business. Wilson et al (2003a, 2003b, 2004) use the percentage of companies that use the internet for e-commerce, the effect of internet on business, and thespeed and cost of internet access taken from CGR.

DataOur empirical work uses the Global Competitiveness Report (GCR), an annual publication of the World Economic Forum (WEF). The GCR aims to enhance globalunderstanding of the factors influencing private-sector led economic growth and explainswhy some countries are much more successful than others at creating new employmentopportunities and raising the income level of their respective populations. For participating countries, the GCR reports on performance and policy conditions affectingthe ability of private sector firms to be globally competitive -- able to create and addvalue within the global marketplace. WEF has expanded its worldwide country coveragefrom 75 countries in 2001, to 125 countries in its 2006 edition with increased coverage of Africa6. The raw data are opinions ranking the extent of progress from 0 to 7, with ahigher value an indication of good performance.7 See table 1 and A1 respectively for thefour indicators of trade facilitation.

Trade facilitation indicators are lower in Africa than elsewhere. Using our samesample configuration services infrastructure indicators are highly correlated with portefficiency (0.822), customs environment (0.737) and regulatory environment (0.721).Port, customs and regulatory are strongly correlated with each other. Port efficiency ishighly correlated with per capita GNI. Moreover, GNI is highly correlated with per capitaGNI. When we limit ourselves to the African sample, the correlations between tradefacilitation indicators are low (see table A2) compared to the correlations in the wholesample.

To be able to derive policy implication it is necessary to link these tradefacilitation indicators to trade performance. We do that after a summary of the alternativemethodological approaches.

Table 1 : Trade Facilitation summary statisticsVariables Mean Std. Deviation Min MaxWhole samplePort 4.548 1.195 2.350 6.850

6 For example, the number of African countries has increased from 5 countries (Egypt, Mauritius, Nigeria,South Africa, and Zimbabwe) in 2001, to 25 countries in 2006.7 This applies for Cameroon and Senegal where we use the data of the previous year for the year 2004.


12/33

12

Customs 4.757 1.115 2.500 6.800Regulatory 3.541 1.151 1.700 6.067Services 4.568 0.881 2.900 6.500African samplePort 3.414 0.992 1.450 5.450Customs 3.853 0.695 2.200 5.350Regulatory 3.003 0.671 1.867 4.633Services 3.482 0.714 1.550 4.950Source: Authors computations based on World Economic ForumGlobal Competitiveness Report .

II.2.2 MethodologiesSeveral methodologies lend themselves to empirical work to address the central questionsof trade facilitations considered earlier, such as the following: What are the costs and benefits of trade facilitation reform to trade expansion and GDP? What is the extent of reform in TF necessary to generate x% of trade and what are the net welfare impacts? Inthis section, we will describe these methods with emphasis on the econometric method8.We consider in turn the gravity model estimation of the Tobit model.

II.2.2.1 Gravity modelsThe gravity model of international trade has been developed independently by Tinbergen(1962) and Pyhnen (1963). The gravity model is a kind of 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 thecost of transportation between them (as measured by the distance between their economiccenters) as follows.

ij

jiij D

Y Y G F

**= , (1)

where i and j represent the exporter and importer respectively, Fij is the trade flow between exporting country i and importing country j, Yi , Y j are their national incomes,Dij is the distance between them, and G is an intercept term. Using logarithmictransformation, equation (1) can be converted to a linear form for econometric analysis asfollow:

ijij jiij DY Y F +++= lnlnlnln . (2)

8 Whereby computable general equilibrium models, including the Global Trade Analysis Project (GTAP)are best candidates for counterfactual issues they are not covered because of data requirements.


13/33

13

In the so called augmented gravity equations, most authors add9 a few variablessuch as income per capita, adjacency, common language, or colonial links. Such anaugmented gravity equation has been widely used as a baseline model for estimating theimpact of a variety of policy issues including those connected to trade facilitation (seeLimao and Venable (2001); Wilson et al (2003a, 2003b, 2004))10. The estimated equation

can therefore be written as follow:

ijijk jiij DTF Y Y F k k ++++= lnlnlnln , (3)

where TF = {PE, CE, RE, EB}, represents the main trade facilitation indicators. Thereduced form for the gravity equation (3) can be expressed as follow:

ijijijij X F ++= '0ln , (4)

where Xij = [Yi, Y j, TFk , Dij, ] is the 1q vectors of gravity variables (gross domestic product, trade facilitation indicators, distance, and other variables); = [, ?, ?,?, ]is the q1 vector of coefficients to estimate,a 0 = a is the intercept; ?ij is the disturbanceterm assumed to be normally distributed with mean zero and constant variance for allobservations.

Estimation methods/issuesTo estimate (4), given the ambiguous causal relationship arising from the high

correlations between trade facilitation measures and income and accordingly not all righthand variables are exogenous and cross-section regression with OLS would not beappropriate. Also the use of a single-year set of trade facilitation measures would limitthe interpretation of the coefficients as elasticities. To solve the problem two approachescould be considered. The first one addressed the endogeneity problem and requiresinstrumental variables (IV) that are exogenous or pre-determined to the trade facilitationmeasures. To use the IV estimator one must first find an instrument for each regressor that is contemporaneously correlated with the error (e.g. the lagged value of theindependent variable in question) and hence must have two characteristics. First, it must be contemporaneously uncorrelated with the error term; and second, it must be correlatedwith the regressors from which it is to serve as an instrument. Such instruments aredifficult to find given the very large number of countries and the power of suchinstruments is always an issue. The second approach is to reconstruct a time-series of TF

9 See for example Carrere (2004); Musila (2005); Frankel et al (1995); Glick and Rose (2002); Frankel andRose (2002); Rose ( 2001); Rose and Engel (2002); Longo and Sekkat (2004); Wilson et al (2003a, 2003b,2004).10 Many other issues have been considered with this formulation including the effect of regional tradinggroups (Carrere, 2004; Musila, 2005; Longo and Sekkat 2004); political blocs (Frankel et al, 1995); andcurrency unions (Glick and Rose, 2002; Frankel and Rose, 2002; Rose, 2001; Rose and Engel, 2002;Carrere, 2004).


14/33

14

indicators and to re-estimate the model using these two time-series indicators both withand without the other indicators of trade facilitation, and with lagged values.

Heterogeneity can be addressed through the use of the panel data method and acountry-pair fixed effect. The fixed effects model is express by the equation (5):

ijijijijij X F ++= 'ln , (5)

where aij is the country-pair effect between the trading partners. The country-pair intercept includes the effect of all omitted variables that are cross-sectionally specific butremain constant over time, such as distance, border, language, contiguity, or culture. Tocorrect the assumption that the error term is identically distributed, the Huber/Whiterobust standard errors could be computed. The white test examines whether the error variance is affected by any of the regressors, the squares of their cross-products. TheStrength of this test is that it tests specifically for whether or not any heteroskedasticity present causes the variance-covariance matrix of the OLS estimator to differ from itsusual formula. The simple approach would be to use the weighted least squares (WLS) bycorrecting the error term, using the estimated variance of some regressors (see Kennedy, p. 118).

II.2.2 Sample selection modelThe sample selection model refers to the fact that in estimating equation (2) by OLS wemay get inconsistent estimators because of zero or missing trade. Our empirical analysisemploys a sample selection model developed by Heckman (1976) as used the literatureon trade facilitation (Hummels (2001); Franois and Manchin (2006)). The model isspecified as follows11:

F1 =p X + u1 , (6)

F2 =? Z + u2 , (7)

where F1 is the value of exports, F2 is a dummy variable taking the value one if tradeoccurs as zero otherwise; X = {Yi, Y j, TF, D, etc.}, Z = {Yi, Y j, TF, D, etc.};p and? arethe coefficients to estimate; and u1 and u2 are the error terms which are jointly normallydistributed, independently of X and Z, with zero expectation, and correlated. The variable

F1 is only observed if F2 > 0. Equation (6) shows how the value of exports is affected bydifferent factors, and the equation (7) gives some information into why trade occurs between two partner countries. To work with Heckmans method, we must estimate a probit model on trade occurring or not, and then estimate the level of trade using OLS.Thus, for our primary equation (2), we have the following specification:

11 OCDE (2006) focuses on trade barrier and logistic services by using probit model to measure the probability that a firm in country i will export to country j. After that, they use the gravity model to find thedeterminants of bilateral trade.


15/33

15

u DTF Y Y F ijk jiij k k 1lnlnlnln ++++= , (8)

and for the probit estimation for non-zero flows we assume that Fij is observed when we

have

u DTF Y Y ijk ji k k 2lnlnln ++++ > 0. (9)

Equation (8) assesses the determinants of bilateral trade and shows the main factorsinfluencing the trade flows, given trade occurred between the two trading partners.Equation (9) sets out the selection criteria and provides information on the factors thatdetermine whether or not we observe trade between country-pairs.

II.2.3.3 Tobit modelThe issue of zero or missing trade observations is important especially in an Africancontext. In this case, we may use a semi-log form for the gravity model. Thus, bilateraltrade is expressed in levels while regressors appear in logs. But, while such aspecification allows us to use all the information, OLS is no longer an appropriateestimator. Since the dependant variable is truncated at zero12, estimation by OLS may produce biased and inefficient estimators. The Tobit is the appropriate estimator and isdefined as follow:

ijijijij X F ++= '0 if RHS > 0(10)

ij= 0 otherwise,

where RHS represents the right hand side of equation (10); ij are the residuals that areindependently and normally distributed, with zero mean and common variance s2 . The problem is to estimate the coefficients of the equation and s2 on the basis of the number of observations on Fij and RHS variables. In the case of the Tobit model, because thelikelihood function is well behaved and the computation of maximum likelihoodestimates is easy, there is no need for the two-stage Heckman method (Maddala, p. 222).The extended gravity model to take account the trade facilitation indicators used in this paper is described as:

12 To include the observations of zero or missing trade, Musila (2005) defines the dependent variable asln(1+ Fij) rather than ln Fij .


16/33

16

( ) ( )

( )

,39

)3)

))

100

2003363534333231

30292827262524

23222120191817

1615141212

11109876

543210

lnln(ln(

ln(ln(lnlnlnlnlnlnlnln1ln

t

jiij

RUS PORGMN ARB FRC

SPN CHN ENG EU MERCOSUR AUNZ LAIA

ASEAN NAF TAUEMOASADC ECOWA S COMESACEAMC

AMU ADJ

t

j

t

i

t

j

t

i j j j j

iiii

t

ij

t

ij

b Db Db Db Db Db Db Db

Db Db Db Db Db Db Db Db Db Db Db Db Db Db

Db Db DIST

bGNIPC bGNIPC b

GNI bGNI bSI b RE bCE b PE b

SI b RE bCE b PE bTARIFF bb F

Colony

ij

++

+++++++

+++++++

+++++++

+++++

++++++

++++++=

+

+

where i and j stand for exporter and importer respectively, and t is trading year (t=2003,2004). Ftij denotes the value of manufactures exports from country j to i at year t.

TARIFFtij is the applied tariff rate in the percent ad valorem term that is specific totrading 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 serviceinfrastructures. 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 arrangementsdummies (see the appendix for the definitions of different trade arrangements) includeAMU (DAMU), CEMAC (DCEMAC), COMESA (DCOMESA), ECOWAS (DECOWAS), SADC(DSADC), UEMOA (DUEMOA), ASEAN (DASEAN), NAFTA (D NAFTA), LAIA (DLAIA),AUNZ (DAUNZ), MERCOSUR (DMERCOSUR ), and EU (DEU). The language dummiesinclude English (DENG), French (DFRC), Spanish (DSPN), Chinese (DCHN), Arabic (DARB),

German (DGMN), Portuguese (DPOR ), and Russian language (DRUS). The adjacencydummy 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 controlfor time-specific shocks. DColony is a dummy that takes a value one if country i andcountry j have been colonized by the same foreign country, and zero otherwise.


17/33

17

III. Empirical results: Application with TobitIn 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 paneldata, such heterogeneity can be controlled for by means of a country-pair fixed effect. Inthis paper, we use country-specific effect when a country is an exporter. With thisspecification, distance, adjacency, and language are eliminated because they are fixedover time13. Moreover, in our sample, zero or missing bilateral trade observations reaches43.53% of the total. Since the dependant variable is truncated at zero, estimation withOLS would produce biased results accordingly we used a Tobit estimator . We use threesub-samples. The first concerns bilateral trade flows between 100 countries. The secondis, a sample comprising 25 African countries (see table A7 in annex). The last oneconcerns intra-African trade and trade with the rest of the world14.

III.1 Whole sample

In this section, we extended the Wilson, Mann and Otsuki (WMO) (2003a, 2003b,2004)15 sample to include 25 countries primarily from Africa. We have also captured theeffect of African regional integration schemes. The gravity model was run using andordinary least squares (OLS) with Huber/White robust standard error.

The coefficients of ports efficiency of both the importer and the exporter is 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.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 coefficientof the exporter is not significant. This implies that global trade flows get a bigger boostwhen the importer's customs environment improves. The regulatory environment of theexporter is significant and negatively associated to trade in all regressions. The regulatoryenvironment of the importer is insignificant. The services infrastructure coefficient of theimporter is significant and positively associated with trade in all regressions, whileservices 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 positivelyrelated to trade.

13 See Cheng and Wall (2005) for the various specifications to controlling for heterogeneity in the gravitymodels of trade.14 Due to endogeneity problem and space, we dont 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 onrequest to the authors.15 The results obtained with WMOs sample are very similar with the previous one. These results areavailable on request to the authors.


18/33

18

We applied the Tobit model (table A4) on the whole sample countries to take intoaccount missing or zero trade. We also added a dummy to isolate the effect of intra-African trade. The results (see third column of table A4) show that when Africancountries 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 tradefacilitation indicators, port efficiency of exporter is positive and significant whereas portefficiency of importer country is negative and not significant in pooled cross-sectionregression and year 2003 fixed effects. This means that importer country must improvetheir port infrastructure to be competitive in the global market. The customs environmentvariable of both the importer and the exporter is positively associated with trade in thefixed effect regressions. Comparing the effect of customs environment on import vs.exports, we note that the coefficient is higher for importer than exporter. The regulatoryenvironment of the importer has significant and positive effect on trade whereasregulatory environment of the exporter has significant and negative effect. The servicesinfrastructure variable of importer is significant and positive in all the regression whereasit is negative and significant for exporter country in the fixed effect regressions.

When using pooled cross-section regression, CEMAC, COMESA, ECOWAS andSADC regional integration affects trade positively. The other African regionalagreements (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.

III.2 Intra-African TradeWe use only the sample of 25 African countries as reporter (exporter) and partner

(importer). Using a Tobit model, table A5 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 coefficientis higher for importer than exporter. Customs environment of the importer havesignificant and negative effect on trade. The policy implication of this result is thatAfrican countries must improve their customs environment to boost their economy. Theregulatory environment of the importer has significant and positive effect on tradewhereas the regulatory environment of the exporter has significant and negative effect.Only services infrastructure of the importer is positively associated with trade. Theservices infrastructure of the exporter is insignificant.

A comparison of the results of the fixed effects and pooled cross-section showsthat allowing for year 2003 fixed effect and exporter heterogeneity (see second and thirdcolumn of table A5) lowers the absolute value of the trade facilitation indicators, andincrease the estimated income elasticities of trade.

The CEMAC, ECOWAS and UEMOA agreements are not significant andnegatively related to trade in the pooled cross-section regression whereas AMU issignificant and negatively related to trade. COMESA and SADC are significant and positively associated with trade. When we use the fixed effect regressions, all the Africanregional agreements become significant and positively related to trade.

Tariff is positive and significant associated with trade. This means that for African countries, tariff is not a constraint for trade contrary to non tariffs barriers thatappear to be the main challenge for African countries.


19/33

19

.

III.3 Intra-African Trade with the Rest of the WorldWe use a sample of African countries as reporter (exporter) and partner (importer)

and other countries as partner (importer) only. Using a Tobit model, table A6 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 portefficiency on import vs. exports, we note that the coefficient is higher for importer thanexporter. 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 customsenvironment to boost their economy. Regulatory variable of both the importer and theexporter is positively associated with trade in exporter fixed effect regression. Comparingthe effect of regulatory environment on import vs. exports, we note that the coefficient ishigher for exporter than importer.

A comparison of the results of the fixed effects and pooled cross-section showsthat allowing for year 2003 fixed effect and exporter heterogeneity (see second and thirdcolumn of table A6) lowers the absolute value of port efficiency, customs environmentand services infrastructure coefficients, and increases the absolute value of regulatoryenvironment coefficient. We also observe an increase of the estimated income elasticitiesof trade.

For regional agreements, we have the same results as described in the previoussection.

IV. The Way forward

The objective of this study was to examine the role of trade facilitation on Africaneconomies. The results show that i) the port efficiency have a positive impact on trade; ii)

customs environment of importer country decreases trade; iii) regulatory environment of exporter (importer) country have a negative (positive) effect on trade; iv) the servicesinfrastructure of importer country have a positive impact on trade.

African countries and regions should give priority to harmonization of ongoingfacilitation efforts, with special emphasis on improved efficiency and competitivenessalong transport corridors. Along each corridor there would be a need to streamline andharmonize transit and customs regulations, improve coordination of corridor operations,monitor performance, and fight corruption. This would in turn imply greater cooperationin view of harmonizing border crossing and visa procedures as well as transport modalstandards and regulation. A program along a corridor could therefore focus onstrengthening regional regulatory framework; harmonizing customs operations, andfacilitating stakeholders consultations. For example, on the transport side lack of harmonization implies that transit vehicles can comply with transport regulation in onecountry and not comply in another country, which in turn triggers additional controls of vehicles and hinders the flows of goods and vehicles. Lack of trust among nationalauthorities implies redundant inspections and poor facilitation of transit cargo clearanceleading to long port clearance process, costly escort system to prevent excessive controlsor smuggling, border crossing delays.


20/33

20

Infrastructure in good condition is a prerequisite for stimulating trade andsustaining competitiveness, especially for landlocked countries. Improved coordination of stakeholders is essential. Each trade facilitation project should provide room for forumsinvolving the public and private sectors to discuss and address operational issues. Negotiation both at bilateral and multilateral levels provide a golden opportunity for

entering into a longer-term process of capacity building and smooth movement towardsinternational trading norms and practices. A trade facilitation network, as discussedabove, would provide a framework for that. Part of the work program for such a network is research and analysis, with urgency on database of trade facilitation measures buildingcurrent work e.g. the Doing Business Report data, the GCR.


21/33

21

Appendix: Data sources and definitionsThe yearly data are constructed for 100 countries and 11 regional trade agreements

spanning from 2003 to 2004. These data are computed from the Global CompetitivenessReport (GCR) released. by The World Economic Forum, the World DevelopmentIndicators (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 bythe United Nations Conference on Trade and Development (UNCTAD). The individualcountries and regional trade agreements are presented in table A7.

The 75 countries included in the Wilson, Mann and Otsuki (2003a, 2003b, 2004)are: Argentina, Australia, Austria, Bangladesh, Belgium, Bolivia, Brazil, Bulgaria,Canada, Chile, China, Colombia, Costa Rica, Czech Republic, Denmark, DominicanRepublic, Ecuador, Egypt, El Salvador, Estonia, Finland, France, Germany, Greece,Guatemala, Honduras, Hong Kong, Hungary, Iceland, India, Indonesia, Ireland, Israel,Italy, Jamaica, Japan, Jordan, Korea, Latvia, Lithuania, Malaysia, Mauritius, Mexico, Netherlands, New Zealand, Nicaragua, Nigeria, Norway, Panama, Paraguay, Peru,

Philippines, Poland, Portugal, Romania, Russia, Singapore, Slovak Republic, Slovenia,South Africa, Spain, Sri Lanka, Sweden, Switzerland, Taiwan, Thailand, Trinidad andTobago, Turkey, Ukraine, United Kingdom, United States, Uruguay, Venezuela,Vietnam, Zimbabwe.

A.1 Real GNIFor all countries, we deflate Gross National Income (GNI) (source: WDI) by the

all- item consumer price index (CPI) for the base-year 2000 (source: WDI) except for Serbia and Zimbabwe. For these two countries, we use GDP deflator (source: WDI) for the base-year 2000.

A.2 Real GNI per capitaFor 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 efficiencyThe Port efficiency for each country is the average of two indexed inputs fromGCR:

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 environmentThe customs environment for each country is the average of two indexed inputsfrom GCR:


22/33

22

Irregular Payments in Imports and Exports (How common do firms inyour industry give irregular extra payments or bribes connected withimport 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 threeindexed input from GCR:

Irregular Payment in Public Contracts (How commonly do firms inyour 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 andindividuals)

Public Trust of Politicians (Public trust in the honesty of the politiciansis: 1 = very low; 7 = very high)

A.3.4 Services sector infrastructureServices sector infrastructure for each country is the average of two indexed inputsfrom GCR:

Quality of competition in Internet service providers (ISP) sector (Iscompetition 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 FlowsThe bilateral trade flows are bilateral trade in manufactured goods (source:

COMTRADE), defined as commodities in categories 5 to 8 in SITC 1 digit industryexcept those in category 68 (non-ferrous metals). The Trade flow data aggregate the tradeflows over the manufactured goods for a given importer-exporter. We deflate the tradeflow using the world import index (source: IFS).

A.5 Tariffs

We use weighted average of applied tariff rates (source: TRAINS) for themanufactured goods under the above definition where bilateral trade valuescorresponding to each tariff line are used as the weight.

A.6 Adjacency, language, colony, and distanceGeographic data, together language and colonial links are extracted from the Centre

d'tudes Prospectives et d'Informations Internationales (CEPII) database (see


23/33

23

www.cepii.fr). Distance data are calculated following the great circle formula, which useslatitudes and longitudes of the relevant capital cities.


24/33

24

References

Blonigen, B.A and W.W. Wilson (2006). "New Measures of Port Efficiency UsingInternational Data." NBER Working Paper No. 12052.

Carrere, C. (2004). "African Regional Agreements: Impact on Trade with or withoutCurrency Union." Journal of African Economies , 13(2): 199-239.

CETFRO (2006) AUDIT Trade Facilitation Measures in Kenya Paper prepared for ILEAP by the Centre for Trade Facilitation, Kenya,available at www.ileap-jeicp.org

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.

Clarke, G.R.G (2005). "Beyond Tariffs and Quotas: Why Dont African ManufacturersExport More?" World Bank Working Paper No. 3617, Washington D.C.

Clark, X., D. Dollar, and A. Micco (2004). "Port Efficiency, Maritime Transport Cost andBilateral Trade." NBER Working Paper No. 10353.

COMESA-EAC-SADC Task Force (2007). Deepening Regional Integration: Developingand Upgrading of Transport Systems and Infrastructure on the North-South and Dar-es-Salaam Corridors.

De Groot, H.L.F., G.J., Linders, P. Rietveld and U. Subramanian (2003). "TheInstitutional Determinants of Bilateral Trade Patterns" Tinbergen Institute DiscussionPaper, TI 2003-044/3.

Diagne, I. (2005) Etude sur la Facilitation des Echanges dans la Rgion Ouest-Africaine , Paper prepared for the ILEAP pre -Hong-Kong Ministerial ConferenceCotonou October 17-18, 2005. available at www.ileap-jeicp.org

Franois, J. and M. Manchin (2006). "Institutional Quality, Infrastructure and thePropensity to Export." Mimeo.

Frankel, J.A and A.K. Rose (2002). "An Estimate of the Effect of Common Currencies onTrade 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.

Glick, R. and A.K. Rose (2002). "Does a Currency Unions Affect Trade? The TimeSeries Evidence." European Economic Review , 46: 1125-1151.


25/33

25

Gwartney, James and Robert Lawson (2006). Economic Freedom of the World: 2006 Annual Report. Vancouver: The Fraser Institute. Data retrieved fromwww.freetheworld.com.

Heckman, J.J. (1976). "The Common Structure of Statistical Models of Truncation,

Sample Selection and Limited Dependent Variables and a Simple Estimator for SuchModels." Annal of Economics and Social Measurement , 5: 475-492.

Hummels, D. (2001). "Time as a Trade Barrier." Mimeo. Purdue University.

Kaufman, D., A. Kraay and P. Zoido-Lobaton (2002). "Governance Matters: UpdatedIndicators for 2000-2001." World Bank Working Paper No. 2772, Washington D.C.

Kennedy, P (1992). A Guide to Econometrics . 3rd Edition. Cambridge, Massachusetts:MIT Press.

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. Serkkat (2004). "Economic Obstacles to Expanding Intra-AfricanTrade." World Development, 32(8): 1309-1321.

Maddala, G.S. (1983). Limited-Dependent and Qualitative Variables in Econometrics .Cambridge: Cambridge University Press.

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'sDevelopment Five Years into the Twenty-first Century: Tunis, 22-24 November.

OECDE (2007). "Examing the Trade Effect of Certain Customs and AdministrativeProcedures." OECD Trade Policy Working Paper No. 42.

--------- (2006). "Logistics and Time as a Trade Barrier." OECD Trade Policy WorkingPaper No. 35.

Ongolo Z. V. and Beret, S. (2005) Un Programme Positif Pour Les Ngociations Sur LaFacilitation Des changes En Afrique Centrale Paper prepared for the ILEAP pre -Hong-Kong Ministerial Conference Yaounde October 31- November 2, 2005.available at www.ileap-jeicp.org


26/33

26

Pyhnen, P. (1963). "A Tentative Model for the Volume of Trade between Countries."Weltwirtschaftliches Archiv , 90(1): 93-100.

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 EconomicPolicy. New York: The Twentieth Century Fund, 1962.

Wilson, J.S., C.L. Mann and T. Otsuki (2004). "Assessing the Potential Benefit of TradeFacilitation: A Global Perspective." World Bank Working Paper No. 3224, Washington,D.C.

--------- (2003a). "Trade Facilitation and Economic Development: Measuring the Impact."World Bank Working Paper No. 2988, Washington, D.C.

--------- (2003b). "Trade Facilitation and Economic Development: A New Approach toQuantifying the Impact."World Bank Economic Review , 17(3): 367-389.

World Bank (2002). "Building Institutions for Market." World Development Report2002, Washington, D.C.: World Bank.

--------- (2003). "Africa Region Trade Progress Note." World Bank, Washington D.C.

World Customs Organization. Role of Customs in Trade Facilitation (seewww.wcoomd.org).

WTO (2004). Doha Work Programme: Decision Adopted by the General Council on 1August 2004, Annex D, WT/L/579 of 2 August 2004.

WTO (2007) Trade Facilitation National Needs Assessment Project, WTO Secretariat:Trade and Finance and Trade Facilitation Division


27/33

27

AnnexTable A1 : Trade Facilitation Indicators for Africa in 2003

PortEfficiency

CustomsEnvironment

RegulatoryEnvironment

ServicesInfrastructure

Countries PortsAir Transport

Irregular Payment

Costs of Corruption Bribes Favoritism

PublicTrust Internet Marke

Algeria 3.00 3.00 4.50 3.70 3.50 3.80 2.30 2.70 2.5Angola 2.00 2.70 3.40 2.80 2.80 2.10 1.50 2.70 2.2Botswana 3.00 4.30 5.10 4.50 4.50 4.40 4.70 2.90 4.0Cameroon 2.60 2.90 3.30 3.10 2.70 3.20 2.30 3.80 3.6Chad 1.30 1.70 2.30 2.10 2.20 1.90 1.50 1.60 1.5Egypt 3.90 4.00 4.40 4.00 4.50 3.80 2.90 4.70 3.8Ethiopia 1.40 5.00 4.30 3.10 3.40 3.00 2.20 1.40 2.1Gambia 4.10 4.90 4.60 3.80 3.80 4.10 3.30 4.30 2.7Ghana 3.20 3.50 3.90 3.60 3.50 3.20 2.80 3.80 3.7Kenya 2.90 4.70 3.40 2.50 2.40 2.60 2.00 3.70 4.1Madagascar 2.10 2.90 3.50 2.40 2.50 2.60 1.70 3.60 3.3Malawi 1.80 3.50 5.30 3.70 3.60 3.00 2.00 4.10 3.2Mali 1.20 2.80 3.40 2.80 2.70 3.10 2.00 3.40 2.2Mauritius 5.20 5.40 3.70 3.20 3.20 3.50 1.80 1.90 4.6Morocco 3.70 4.30 3.60 3.20 3.00 3.20 3.10 3.70 4.4Mozambique 2.20 3.20 3.30 2.90 2.90 2.40 1.80 3.10 2.4 Namibia 4.90 4.80 4.40 3.50 3.60 3.20 3.10 4.00 3.9 Nigeria 2.50 3.50 2.80 2.80 2.20 2.10 1.50 3.60 3.9Senegal 3.50 4.30 3.80 3.00 2.80 3.20 2.00 3.30 3.2South Africa 4.60 6.00 4.60 4.70 3.80 3.40 3.10 4.20 5.2Tanzania 3.20 3.40 4.20 3.00 3.60 3.40 3.20 4.10 3.3Tunisia 4.50 4.90 5.30 4.50 4.60 4.70 4.60 4.00 4.2Uganda 2.10 3.20 3.20 3.00 2.60 2.60 2.00 4.10 2.8

Zambia 1.70 3.50 4.00 3.10 3.20 3.10 1.70 3.60 3.3Zimbabwe 1.80 3.10 3.50 3.40 2.60 2.00 1.20 3.20 3.7

AMU 3.73 4.07 4.47 3.80 3.70 3.90 3.33 3.47 3.7CEMAC 1.95 2.30 2.80 2.60 2.45 2.55 1.90 2.70 2.5COMESA 2.49 3.80 3.87 3.12 3.08 2.83 1.92 3.30 3.3ECOWAS 2.90 3.80 3.70 3.20 3.00 3.14 2.32 3.68 3.1SADC 2.95 3.89 4.09 3.38 3.30 3.01 2.35 3.40 3.5UEMOA 2.35 3.55 3.60 2.90 2.75 3.15 2.00 3.35 2.7AFRICA 2.90 3.82 3.91 3.30 3.21 3.10 2.42 3.42 3.3

ASEAN 4.43 5.10 4.28 4.12 3.97 3.67 3.53 4.35 4.3AUNZ 5.95 6.25 6.55 6.20 6.25 5.20 4.80 5.30 5.6EU 5.27 5.71 5.95 5.27 5.21 4.27 3.93 4.98 5.5LAIA 3.04 3.83 4.56 3.25 3.45 2.61 1.77 4.20 4.0MERCOSUR 3.33 3.68 4.48 3.50 3.55 2.65 1.88 4.33 4. NAFTA 5.07 5.70 5.43 4.67 4.77 3.70 3.43 5.33 5.6Source: Authors computations based on World Economic ForumGlobal Competitiveness Report .


28/33

28

Table A2: Correlations: WMO SampleCorrelations: Whole Sample

Trade Tariff PE CE RE SI GNI GNIPC DistanceTrade 1Tariff 0.021 1PE 0.409 -0.085 1

CE 0.337 -0.056 0.758 1RE 0.319 -0.077 0.774 0.880 1SI 0.462 -0.028 0.808 0.727 0.676 1GNI 0.613 0.068 0.552 0.419 0.400 0.655 1GNIPC 0.447 -0.09 0.764 0.766 0.684 0.745 0.654 1Distance -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 SampleTrade Tariff PE CE RE SI GNI GNIPC Distance

Trade 1Tariff 0.116 1PE 0.338 0.106 1CE 0.249 0.118 0.500 1RE 0.238 0.244 0.594 0.813 1SI 0.285 0.106 0.513 0.410 0.445 1GNI 0.331 0.199 0.455 0.119 0.376 0.365 1GNIPC 0.2693 0.155 0.683 0.226 0.501 0.358 0.781 1Distance -0.209 -0.11 0.072 -0.022 -0.066 0.054 -0.038 0.012 1

Correlations: Intra-Africa SampleTrade Tariff PE CE RE SI GNI GNIPC Distance

Trade 1Tariff 0.162 1PE 0.331 0.071 1CE 0.221 0.061 0.459 1RE 0.197 0.172 0.572 0.794 1SI 0.290 0.052 0.510 0.421 0.457 1GNI 0.298 0.159 0.456 0.036 0.352 0.380 1GNIPC 0.224 0.113 0.675 0.159 0.479 0.339 0.796 1Distance -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 ForumGlobal Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, andWorld Bank World Development Indicators for GNI.


29/33

29

Table A3: Whole sample: OLSPooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect

Coeff.Std.Error Coeff.

Std.Error Coeff. Std. Error

Constant -16.687*** 1.853 -24.333*** 2.127 -25.703*** 2.101Tariff -1.431*** 0.391 -1.886*** 0.446 -1.740*** 0.443

Ports Exporter 1.081*** 0.136 1.089*** 0.156 1.160*** 0.155Ports Importer 0.767*** 0.150 0.725*** 0.167 0.513*** 0.167Custom Exporter 0.403** 0.206 0.918*** 0.256 0.358 0.231Custom Importer 0.173 0.224 -0.067 0.254 -0.146 0.245Regulatory Exporter -0.841*** 0.155 -1.009*** 0.183 -0.736*** 0.173RegulatoryImporter 0.154 0.158 0.476** 0.177 0.569*** 0.175Services Exporter 0.205 0.175 -0.043 0.199 0.062 0.198Services Importer 1.064*** 0.215 0.844*** 0.228 1.062*** 0.226GNI Exporter 0.694*** 0.014 0.666*** 0.015 0.666*** 0.015GNI Importer 1.144*** 0.021 1.101*** 0.021 1.124*** 0.020Distance -1.327*** 0.024 - - - -

AMU 0.332 0.474 3.190*** 0.254 3.177*** 0.207CEMAC 3.614** 1.331 6.376*** 1.153 6.723*** 1.283COMESA 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** 1.331ASEAN 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*** 0.129AUNZ 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*** 0.063English 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 12525 12548 12548 R 2 0.693 0.601 0.605

Note: All non-dummy variables are in logs. The notations *, **, and *** denotesignificance at 10, 5 and 1 percent levels, respectively.Source: Authors computations based on World Economic ForumGlobal Competitiveness Report for trade facilitation indicators, COMTRADE for trade flows, UNCTAD TRAINS for tariffs, andWorld Bank World Development Indicators for GNI.


30/33

30

Table A4: Whole Sample: Tobit Model

Pooled Cross-SectionYear 2003 and Exporter

Fixed Effects

Intra Africa TradeDummy and 2003

Fixed Effect

Coeff. Std. Error Coeff. Std. Error Coeff.Std.Error

Constant -54.465*** 4.770 -80.233*** 4.983 -80.914*** 4.994Tariff -2.168** 1.021 -1.248 1.069 -1.432 1.072Ports 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* 0.421Services 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*** 0.036Distance -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 2.076LAIA 2.335*** 0.568 7.310*** 0.528 7.203*** 0.530AUNZ 0.497 4.824 3.325 5.049 3.314 5.072

MERCOSUR 1.160

0.999 2.227** 1.026 3.216*** 1.026EU -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 17555 17579 17579 Log-likelihood -46159.607 -46871.893 -46940.51



31/33

31

Table A5: Intra Africa: Tobit ModelPooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect

Coeff.Std.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 RegulatoryExporter -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



32/33

32

Table A6: Intra Africa and Rest of world (only as partner): Tobit ModelPooled Cross-Section Year 2003 Fixed Effect Exporter Fixed Effect

Coeff.Std.Error Coeff.

Std.Error Coeff.

Std.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



33/33

Table A7: Sample and Definition of the Regional Trade AgreementsRegional Trade Blocs Countries Members Main dates*CEMAC (Economic and MonetaryCommunity of Central Africa)

Cameroon, Chad 1962, creation of the UDE(Equatorial Customs Union); 1964,revision and creation of the UDEAC(Central African Customs andEconomic Union); 1973, revision of

the UDEAC; end of the 1980s, failureof the compensation funds; 1994,revision and creation of the CEMAC

ECOWAS (Economic Community of West African States)

Gambia, Ghana, Mali, Senegal, Nigeria

1975, creation of the ECOWAS;1990, start of a more general schemeof intra-regional liberalization

COMESA (Common Market for Eastern and Southern Africa)

Angola, Egypt, Ethiopia, Kenya,Madagascar, Malawi, Mauritius,Uganda, Zambia, Zimbabwe

1981, creation of the PTA (Easternand Southern African PreferentialTrade Area); end of 1980s, firstreduction in the customs tariffs onintra-regional trade; 1993, revisionand creation of the COMESA

SADC (Southern AfricanDevelopment Community)

Angola, Botswana, Madagascar,Malawi, Mauritius, Mozambique, Namibia, South Africa, Tanzania,Zambia, Zimbabwe

1980, creation of the SADCC(Southern African DevelopmentCoordination Conference); 1992,revision and creation of the SADC

AMU (Arab Maghreb Union) Algeria, Morocco, Tunisia 1989, creationUEMOA (West African Economicand Monetary Union)

Mali, Senegal 1994, creation

NAFTA (North American Free TradeAgreement)

Canada, Mexico, United States of America

1992, creation

LAIA (Latin American IntegrationAssociation)

Argentina, Bolivia, Brazil, Chile,Colombia, Ecuador, Mexico,Paraguay, Peru, Uruguay, Venezuela

1980, creation

MERCOSUR (Southern CommonMarket)

Argentina, Brazil, Paraguay, Uruguay 1991, creation

AUNZ Australia, New-ZealandASEAN (Association of SoutheastAsian Nations)

Indonesia, Malaysia, Philippines,Singapore, Thailand, Vietnam

1967, creation

EU (European Union) Austria, Belgium, Czech Republic(since 2004), Denmark, Estonia(since 2004), Germany, Greece,Hungary (since 2004), Italia, Ireland,Finland, France, Latvia (since 2004),Lithuania (since 2004), Luxembourg, Netherlands, Poland (since 2004),Portugal, Slovak Republic (since2004), Slovenia (since 2004), Spain,Sweden, United Kingdom

1951, creation of the European Coaland Steel Community; 1957,establishment of the EuropeanCommunity; 1994, revision andcreation of the European Union

Other countries Bangladesh, Bulgaria, China, CostaRica, Dominican Republic, ElSalvador, Guatemala, Honduras,Hong Kong, Iceland, India, Israel,Jamaica, Japan, Jordan, Korea,Macedonia, Malta, Nicaragua, Norway, Pakistan, Panama, Romania,Russian Federation, Serbia, SriLanka, Switzerland, Taiwan,Trinidad and Tobago, Turkey,Ukraine

*The dates for CEMAC, ECOWAS, COMESA and SADC are drawn from Carrere (2004).

afghan trade policy

Documents