7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 1/11

Trade Policy, Trade Costs, and Developing Country Trade

BERNARD HOEKMANWorld Bank and CEPR, Washington DC, USA

and

ALESSANDRO NICITA *

UNCTAD, Geneva, Switzerland 

Summary.  —  This paper reviews some indices of trade restrictiveness and trade facilitation and compares the trade impact of differenttypes of trade restrictions applied at the border with the effects of domestic policies that affect trade costs. Based on a gravity regressionframework, the analysis suggests that tariffs and non-tariff measures continue to be a significant source of trade restrictiveness forlow-income countries despite preferential access programs. The results also suggest that behind-the-border measures to improve logisticsperformance and facilitate trade are likely to have a comparable, if not larger, effect in expanding developing country trade, especiallyexports.  2011 Elsevier Ltd. All rights reserved.

Key words  — TARIFFS, nontariff measures, trade facilitation, logistics, economic development, DOHA Round

1. INTRODUCTION

The trade policy literature has for many years emphasizedthe importance of taking into account the impact of a varietyof sources of trade costs in addition to import tariffs—see Forexample,  Deardorff and Stern (1998)  and Anderson and vanWincoop (2004). Recent research on trade and developmenthas emphasized the magnitude of the trade costs associatedwith administrative red tape and entry barriers, informed by

the emergence of new datasets, such as the OECD’s ProductMarket regulation database, the World Bank’s   “Doing Busi-ness”   indicators and Logistics Performance Index (LPI), aswell as firm-level surveys of the investment climate, and busi-ness environment that prevails in countries.

In this paper we assess the impacts of different sources of trade costs on international trade. The objective is to comparethe effect of border barriers (import tariffs, adjusted for bilat-eral preferences, and non-tariff measures) with other sourcesof trade costs. Our interest is to explore the relative impactson trade volumes of different sources of policy-induced tradecosts.

The plan of the paper is as follows. Section 2 summarizes thecurrent pattern of tariff protection and the aggregate of all

non-tariff measures (NTMs) captured in the UNCTAD data-base on NTMs (WITS). Section 3 discusses how we take intoaccount the extensive system of preferential trade that hasbeen put in place by OECD countries as well as the numerouspreferential trade agreements concluded between subsets of WTO members. Section 4 discusses some of the componentsof the aggregate NTM measure, as well as indicators of tradefacilitation performance: the   “trading across borders” compo-nent of the World Bank’s Doing Business report, and theLogistics Performance Index. These indicators reflect regula-tory policies that directly affect trade costs but are not cap-tured by the tariff and NTM databases commonly used byanalysts. Section 5 presents the results of an empirical assess-ment of the relative trade impacts of these different sets of pol-icies and the possible trade effects of convergence by

developing countries to the average levels of border protection

and trade facilitation performance levels prevailing in middle-income countries. We find that the latter will have a substan-tially larger positive impact on trade volumes than the former.Section 6 concludes.

2. TRADE POLICIES

Trade policies can be broadly divided into tariffs (ad-valorem

and specific) and non-tariff measures. Although tariffs are stillthe most widely used policy instrument to restrict trade, theirrelative importance has been declining. Trade liberalization,whether unilateral, the result of agreements negotiated underthe auspices of the World Trade Organization, or the conse-quence of preferential trade agreements (PTAs), has greatly re-duced the average level of applied tariffs. Conversely, the use of NTMs has been increasing both in terms of the number of prod-ucts covered and the number of countries utilizing them (WorldBank and IMF, 2008). The use of tariffs, specific duties andNTMs in 2006 is illustrated in Figure 1. In general, the use of non-tariff measures increases with the level of economic devel-opment of countries. This is particularly true for NTMs, whichareincreasingly used to regulate trade, especially in high income

countries. Similarly, specific duties, although affecting only arelatively small share of total imports, are more prominent inhigh income countries.

The type of NTMs included in the analysis of this paper islimited by the availability of data. In particular, as a measureof NTMs we use the ad-valorem equivalent of NTMs esti-mated in (Kee, Nicita, & Olarreaga, 2009). This ad-valorem

* We are grateful to Alan Deardorff, Simon Evenett, Sheila Page, and Ben

Shepherd for comments on an earlier draft, and to three referees of this

 journal for comments that greatly improved the paper. The authors accept

sole responsibility for any errors remaining. The views expressed are pe-

rsonal and should not be attributed to the World Bank or the United

Nations Conference on Trade and Development. Final revision accepted:

March 28, 2011.

World Development Vol. 39, No. 12, pp. 2069–2079, 2011  2011 Elsevier Ltd. All rights reserved

0305-750X/$ - see front matter

www.elsevier.com/locate/worlddevdoi:10.1016/j.worlddev.2011.05.013

2069

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 2/11

equivalent of NTMs captures the effect of quantitative restric-tions, technical product regulations, anti-dumping and coun-tervailing measures,   and any monopolistic measure ordiscretionary licensing. 1

The ad-valorem equivalent of NTMs can be combined withdata on tariffs into an overall trade restrictiveness index(OTRI) to capture the effect of both types of measures ( Keeet al., 2009). To isolate the effect of tariffs from the overall in-dex we also calculate the tariff trade restrictiveness index(TTRI). The difference between the TTRI and OTRI is thatthe OTRI includes the effect of both tariff and NTMs, whilethe TTRI captures only tariffs, both   ad valorem and the advalorem equivalents of specific tariffs. 2 These indices are cal-culated on a bilateral basis, using the effectively applied tariff and taking into account all preferential regimes. Both theOTRI and the TTRI are a measure of the uniform tariff equiv-alent implied by observed trade policies affecting a country’simports—that is, they represent the ad-valorem tariff   thatwould be needed to generate the observed level of trade. 3

The prevailing average TTRI and OTRI across countries isplotted in Figures 2 and 3. Trade policies are generally morerestrictive in lower-income countries, reflecting both lower tar-

iffs in higher-income economies and the fact that their importsare highly skewed toward manufactures, which face relativelylow barriers.

Agricultural trade is much more restricted than manufac-tured products, both in terms of the TTRI and the OTRI,especially in high income countries. This reflects both highertariffs and greater use of NTMs in agricultural trade. A com-parison of  Figures 2 and 3 reveals that NTMs contribute sub-stantially to the set of policies restricting global trade,especially in agriculture. Differences in the OTRI and TTRIare also evident across geographic regions (Table 1). In gen-eral, East Asian, Central Asian, and East European countriesare less restrictive, while countries in South Asia and the Mid-dle East and North Africa are more restrictive. This pattern issimilar for the TTRI and the OTRI, and for agriculture andmanufacturing.

As a result of unilateral reforms and bilateral and regionalagreements, global trade has been substantially liberalized inrecent years.  Figure 4 presents scatter plots of the TTRI forthe years 2000 and 2006. While liberalization has been sub-stantial in most countries, tariff reduction has centered moreon manufacturing than agricultural products. Agricultural

simple average tariff 

specific duties

non-tariff measures

   0

 .   1

 .   2

 .   3

    %   l   i  n  e  s  s  u   b   j  e  c   t   t  o  :   N   T   M  a  n   d  s  p  e  c   i   f   i  c   d  u   t   i  e  s

 .   0   4

 .   0   6

 .   0   8

 .   1

 .   1   2

  s   i  m  p   l  e  a  v  e  r  a  g  e

   t  a  r   i   f   f

4 6 8 10 12

log gdp per capita

Figure 1.  Use of Tariffs, Specific Duties, and NTMs (% of HS six digit lines), 2006.

 Agriculture

Manufacturing

 All Goods

     0

 .     0     5

 .     1

 .     1     5

     T     T     R     I

4 6 8 10 12

log gdp per capita

Figure 2.  TTRI and GDP per capita, 2006.

2070 WORLD DEVELOPMENT

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 3/11

trade restrictiveness increased for some countries between2000 and 2006. 4 The TTRI has declined both for all countrygroups. Middle income economies have seen the largest de-cline, including in agriculture. By region, countries in EastAsia and Latin America (sub-Saharan Africa) have reducedtariffs the most (least).

3. MARKET ACCESS AND THE EFFECT OF TRADEPREFERENCES

The effect of trade policies on exporters’ access to marketsdiffers across trading partners and geographic regions. Tomeasure the average restrictiveness that exporters face in aparticular market the OTRI and TTRI discussed above arecomputed in terms of exports. Those are reported in  Table 2where restrictiveness is provided by a market access perspec-tive. We label these measures MA-TTRI and MA-OTRI.Those can be thought as the equivalent uniform tariff that ap-plied to exports of a given country would result in the ob-served level of exports. These measure the restrictiveness of policies confronting exporters from in each geographic regionand country group by taking into account preferential traderegimes. Upper middle income countries generally enjoy better

market access in both developing and developed countries.

This is largely due to the composition of exports from thesecountries, which are skewed toward manufacturing. Low in-come countries face more restrictive market access conditionsbecause their exports are more biased toward agriculture.

Across developing country regions, South Asia faces themost restrictive market access, due to export composition(agriculture, textiles and apparel) and because it has relativelylimited preferential access. Sub-Saharan countries have thebest market access, especially in high income countries, reflect-ing again export composition (minerals, primary products,

plantation agriculture), as well as low or zero (preferential)tariffs in many high income countries. By far the highest levelsof market access barriers apply to South-South trade flows.Sub-Saharan African countries confront TTRIs and OTRIsin low-income countries that are 3 to 4 times higher than thosethat apply in middle- and high-income markets.

Comparing the MA-TTRI and the MA-OTRI suggests thatNTMs are generally more important in restricting trade thantariffs: their measured ad-valorem equivalent is much higherthan existing tariffs. Standards, licensing requirements andsimilar regulatory instruments typically affect all productsentering a market regardless of their origin, so that the impactof NTMs is relatively similar across trading partners. Tariffs,conversely, are generally negotiated on a bilateral basis, thus

giving some trading partners a substantial advantage in

 Agriculture

Manufacturing

 All Goods

     0

 .     1

 .     2

 .     3

 .     4

     O

     T     R     I

4 6 8 10 12

log gdp per capita

Figure 3.  OTRI and GDP per capita, 2006.

Table 1.  OTRI and TTRI by developing country region, 2006 

Region (developing countries only) Total trade (%) Agriculture (%) Manufacturing (%)East Asia   11.3 26.6 10.4%

5.0 8.7 4.8

Europe and Central Asia   10.1 25.9 9.0

4.5 10.3 4.0

Latin America   15.0 28.1 13.8

5.4 6.6 5.3

Middle East and N Africa   21.6 32.3 19.4

11.9% 12.1 11.8

South Asia   19.5 46.4 18.2

14.0 31.4 13.2

Sub-Saharan Africa   14.4 24.9 12.9

8.4 13.8 7.6  Note: OTRI figures in  bold and TTRI figures in  italics.

TRADE POLICY, TRADE COSTS, AND DEVELOPING COUNTRY TRADE 2071

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 4/11

market access. With the increase in reciprocal and nonrecipro-cal preferential trade agreements, almost all trade flows today

are affected by some sort of tariff preference. This is particu-larly true for high income countries, where market access is af-fected by an increasing number of such agreements.

The proliferation of preferential trade arrangements makesit important to properly measure preferential margins inassessing the relative market access conditions confrontingexporters. This is done in the TTRI and OTRI in a directway, as the calculations take into account the bilateral marketaccess conditions that apply. But what matters for a givencountry, however, is the   relative  preference (the relative mar-ket access conditions), not just the absolute level of prevailingbarriers at the border.

Commonly used measures of preference margins comparethe preferential tariff to the most-favored-nation (MFN) rate.This will overestimate the relative preference enjoyed by coun-

tries as in most instances other countries will also have prefer-ential access. In practice it is possible that preferential rates

granted to a particular country, although lower than MFN,still penalize it relative to other countries that benefit fromeven lower or zero tariffs. To calculate the relative preferentialmargin the focus needs to be on the average advantage – intariff percentage points – that a given basket of goods enjoyswhen exported from country A as compared to when it origi-nates in other countries.

To clarify with an example, in what follows we calculate therelative preferential margin that Mexico enjoys in the US byusing as the counterfactual the average tariff for Mexico’s ex-port bundle if this were to originate in other countries. The rel-ative preferential margin is the difference between the bilateraltrade-weighted preferential tariff imposed by the US onMexico and that counterfactual. There are two sets of weightswhen calculating this margin: first, the counterfactual, which is

   0

 .   1

 .   2

 .   3

 .   4

   T   T   R   I   2   0   0   0

0 .1 .2 .3 .4

TTRI 2006

   0

 .   1

 .   2

 .   3

 .   4

   A  g  r   i  c  u   l   t  u  r  e   T   T   R   I   2   0   0   0

0 .1 .2 .3 .4

 Agriculture TTRI 2006

   0

 .   1

 .   2

 .   3

 .   4

   M

  a  n  u   f  a  c   t  u  r   i  n  g   T   T   R   I   2   0   0   0

0 .1 .2 .3 .4

Manufacturing TTRI 2006

Figure 4.  TTRI 2000 and 2006.

Table 2.  Market access TTRI and OTRI, 2006 

Importing countries Exporting countries

Highincome

Upper middleincome

Lower middleincome

Lowincome

EastAsia

E. EuropeCent. Asia

LatinAmerica

Mid. EastN. Africa

South Asia Sub-SaharanAfrica

High income   6.3 5.7 7.9 9.1 8.3 5.1 7.0 4.3 10.4 4.4

 2.4 1.2 2.5 2.4 2.6 1.1 1.5 0.8 3.1 0.7 

QUAD   6.3 5.2 8.6 10.6 8.9 5.2 6.9 4.4 13.6 4.5 2.1 0.9 2.5 2.5 2.7 0.8 1.2 0.5 3.3 0.5

Upper middle   15.6 11.8 15.8 14.7 19.2 10.2 13.6 6.0 14.3 5.9

5.6 3.8 5.6 5.7 7.2 4.4 2.6 2.5 6.6 3.5

Lower middle   12.4 11.1 12.9 9.4 13.6 11.2 12.6 6.7 9.9 4.0

7.1 4.8 6.7 5.1 6.6 6.2 5.1 2.8 6.2 2.7  

Low income   18.2 14.3 19.5 25.4 22.2 17.7 15.9 16.3 16.2 16.3

10.9 8.1 12.2 12.9 13.8 6.2 9.0 10.0 10.4 12.2

Note: MA-OTRI in  bold; MA-TTRI in  italics. QUAD countries are Canada, European Union, Japan, and USA.

2072 WORLD DEVELOPMENT

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 5/11

a weighted average of tariffs imposed on all other (potential)exporters to the US; and second, the preferential margin,which is an average constructed across many tariff lines.

To measure the counterfactual, we first calculate the trade-weighted average tariff at the tariff line level that an importer(the US) imposes on all other countries except the country forwhich the preferential margin is calculated (Mexico). This is

done by using (US) bilateral imports as weights, so as to takeinto account the supply capacity of (US) trading partners. Wethen aggregate across tariff lines using (Mexican) exports (tothe US) to take   care of differences in product compositionacross partners. 5

A simpler alternative measure would be to compare the (US)import weighted average tariff imposed on a country (Mexico)with that imposed on all other countries. 6 This approach usestotal imports (by the US) at the HS six digit level as weights. Aproblem with this method is that it disregards product compo-sition: if Mexico’s export bundle to the US is not representa-tive of the composition of US imports (e.g., Mexican exportsto the US are mainly agricultural, while US imports aremainly manufactured goods) using exclusively US imports asweights in the calculation of the counterfactual would likely

lead to biased results.7

A further complication arises in the aggregation across tariff lines. A proper aggregation would take into account that im-ports of some goods are more responsive to changes in pricesthan others. In theory, imports that are less sensitive to prices(inelastic) should be given less weight as tariffs change   –   asthey would have little effect on overall volumes of trade. 8

To correct for this, HS six digit product lines are aggregatedusing the import demand elasticities. 9

The relative preferential margin (RPM) for exports fromcountry  j   is then calculated as:

 RPM  j  ¼Pk  Phs

imp  jk ;hsk ;hs

Pmimp k 

m;hst k m;hs

P mimp k m;hs

!Pk 

Phs

imp  jk ;hsk ;hs

Pk 

Phs

imp  jk ;hsk ;hst 

k m;hsP

k 

Phs

imp  jk ;hsk ;hs

; m– j   ð1Þ

where,  imp are imports,   e is the import demand elasticity,   t isthe tariff, k  indexes importers, hs are HS 6 digit categories, andv are exporters competing with country j  in exporting to coun-try k . The second term in the equation is simply the MA-TTRIwhile the first term is the MA-TTRI but calculated on the ba-sis of tariffs applied to competitors of country j . This index canbe calculated bilaterally (i.e., Mexico’s relative preferences inthe US market) or at the country level (i.e., Mexico’s overall

level of relative preferences for its exports relative to all its ex-port markets). In the latter case all bilateral trade flows of aparticular country are considered.

This measure of preference margin can be positive or nega-tive, depending on the advantage or disadvantage of the coun-try with respect to other exporters.   Table 3   reports relativepreferential margins averaged by region. All regions have po-sitive relative preferential margins with themselves. This indi-cates the importance of regional trade agreements. The most“effective”  regional agreements in terms of preferences are inLatin America, where countries enjoy a relative preferentialmargin of about 3%. Latin America both enjoys and providesa substantial preferential margin to the USA and Canada,reflecting trade agreements with the US and within the region(MERCOSUR, etc.). This is mirrored by the negative prefer-

ence that countries outside Latin America face when they ex-port to that region. Relative preference margins, whetherpositive or negative, are much smaller for other regions. Coun-tries in Sub-Saharan Africa, for example, enjoy a relative pref-erential margin of only about 0.5% in the EU, as they competeboth among themselves and other countries to which the EUprovides preferences (Eastern Europe, North Africa and Latin

America). Relative preferential margins are mostly negativefor East Asian states.Country-specific estimates of the relative preference margin

for a number of Sub-Saharan countries are reported in Table4. Bilateral preferences are substantial in only a few cases, andin a few instances they are actually negative, putting the coun-tries concerned in a situation similar to that applying to EastAsian economies. Only Madagascar has relatively large prefer-ential margins in more than two markets, while most countrieshave meaningful preferential margins in only one or two mar-kets, if any.

4. OTHER TRADE COSTS

The foregoing discussion illustrates that NTMs as well astariffs are a significant source of barriers to trade. A questionto which we return below concerns the relative importance of various barriers, especially from a developing country perspec-tive. In particular, given the still high tariffs in agriculturalproducts, remaining tariffs may still have large effects relativeto NTMs, especially in the case of South-South trade flows.On the other hand, the average impact of NTMs in regulatingimports into higher-income countries clearly suggests that ac-tion to reduce their trade-impeding effects could have highpayoffs.

Besides tariffs and NTMs, we are also interested in the im-pact of other trade related costs. Internal trade and transac-tions costs may be of equal if not greater importance in

reducing volumes of trade. Many of these trade costs reflectthe domestic economic environment: the legal and regulatoryframework, the efficiency of infrastructure services and re-lated regulation, customs clearance procedures, administra-tive red tape, etc. Data on domestic trade costs are oftenlacking. However, the World Bank has recently initiatedthe collection of data for a large number of developing coun-tries on the performance of logistics services and on the inter-nal costs associated with shipping goods from the factorygate to the port, and from ports to retail outlets. The firstis captured by the Logistics Performance Index (World  Bank2007); the second is covered by the Doing Business database(World   Bank, 2008). All of these indicators capture dimen-sions of prevailing domestic regulatory regimes that affecttrade.

The Doing Business   “cost of trading” indicator measures thefees associated with complying with the procedures to exportor import a 20-foot container, measured in US dollars(Djankov, Freund, & Pham, 2010). These include costs fordocuments, administrative fees for customs clearance andtechnical control, terminal handling charges and inland trans-port. The cost measure does not include tariffs or trade taxesand only official costs are recorded. The indicator is part of the Doing Business trading across borders index. The method-ology, survey instruments and data are available at   http://www.doingbusiness.org.

The  Logistics Performance Index (LPI) provides a snapshotof the supply chain performance of countries. Based on aworldwide survey of global freight forwarders and expresscarriers, the LPI measures the logistics friendliness of the

TRADE POLICY, TRADE COSTS, AND DEVELOPING COUNTRY TRADE 2073

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 6/11

countries surveyed. Feedback from the survey is supple-

mented with data on the performance of key componentsof the logistics chain. For the analysis in this paper we usethe overall LPI score. This score summarizes six sub-indica-tors (efficiency of clearance process, quality of trade andtransport infrastructure, ease of arranging competitivelypriced shipments, competence and quality of logistic services,ability to track and trade consignments, and timeliness of shipments within the expected delivery time. The LPI indica-tors range from 1 (worst) to 5 (best). In practice, the LPIscore varies from 1.2 (Afghanistan) to 4.2 (Singapore). Theunderlying methodology and data on the LPI are availableat   http://www.worldbank.org/lpi.

Table 5 reports the average of these indices by income coun-try groups. Developing countries generally have weaker tradefacilitation performance than higher-income economies.

5. AN EMPIRICAL ASSESSMENT

The literature on trade costs has recently devoted someattention to the role that domestic costs play in affecting inter-national trade. The impact of different sources of trade costs isfrequently assessed through the inclusion of specific variables(generally measures of infrastructure availability such asroads, railways, phone lines, etc.) in gravity type models. Ingeneral, the literature supports the hypothesis that domestictrade costs and the economic business environment are signif-icant determinants of the volume of trade between countries(Limao & Venables, 2001; Wilson, Mann, & Otsuki, 2003;Anderson & Marcouiller, 2002;   Francois and Manchin,2007). What follows builds on the existing gravity model liter-ature to investigate the importance of trade and related regu-latory policies on trade flows.

Table 3.  Relative preference margins, 2006 (percentage points)

Importers EastAsia

East EuropeCentral Asia

LatinAmerica

Middle East andNorth Africa

South Asia Sub-SaharanAfrica

High incomecountries

East Asia 0.22   0.06   0.09   0.02   0.03 0.01   0.03East Europe Central Asia   0.01 0.45   0.37 0.39   0.20 0.04   0.15Latin America   2.54   1.88 2.98   0.51   2.13   1.22 1.69Middle East and North Africa   0.29   0.24   0.25 0.91   0.22 0.10   0.03

South Asia   0.21   0.08   0.04   0.26 2.03   0.15   0.05Sub-Saharan Africa   0.10   0.03   0.06   0.02   0.12 0.30   0.06High Income Countries   0.46 0.42 0.71 0.19   0.46 0.13 0.08Australia and New Zealand   0.18   0.61   0.28   0.08   0.23 0.11 0.10Canada   1.00   0.85 1.75 0.01 1.79 0.02 1.01European Union 0.05 1.07 0.98 0.64   0.70 0.51   0.50Japan 0.34 0.02 0.07 0.00 0.70 0.08   0.13USA   0.67   0.03 1.01   0.08 0.22 0.11   0.03

Table 4.  Relative preference margins for selected african countries in high income markets, 2006 (percentage points)

Exporter Australia and New Zealand Canada European Union Japan USA Other High Income Countries

Angola 4.70 0.00 0.04 0.03 0.08 0.02Benin 0.00 2.72 0.02 0.23 0.18 0.02Burkina Faso 4.23 1.02 0.60 0.02 0.40 0.27Cent.African. Rep 1.93 0.65 0.06 0.07 0.03 0.03Chad 4.61 0.78 0.08 0.00 0.17 0.01Cote D’Ivoire   0.04   0.04 0.36 0.04   0.02 0.24Cameroon   0.02   0.16 0.40 0.01 1.16 0.27Congo 0.00 0.01 0.07 0.08 0.27 0.01Ethiopia 0.11 0.34 0.50 0.06 0.01 0.76Ghana   0.15 0.06 0.92 0.02 0.38 0.59Kenya   0.04   0.48 1.25 0.08   0.92 0.68Madagascar 1.43 7.53 3.89 0.83   0.97 2.92Mali 1.89 1.63 0.43 2.84 0.24 0.29Mozambique 0.54 0.17 4.46 0.53   0.18 2.76Mauritania 0.37 4.37 0.37 6.95 0.00 0.22Malawi 0.04 0.35 0.02 0.05

  2.96 0.01

Niger 2.27 0.52 0.04 0.43 2.01 0.02Nigeria   0.29 0.03 0.06 0.00 0.13 0.05Rwanda 4.20 0.14 0.02 0.20 0.05 0.01Sudan 1.33 0.00 0.06 0.00 0.00   0.03Senegal 1.41 0.60 2.29 2.99 0.22 0.98Togo 0.00 0.06 0.32 0.28 1.42 0.12Tanzania 0.15 0.00 1.13 0.05 0.32 2.33Uganda 0.03 0.16 1.45 0.22 -0.01 1.22Zambia 0.37 0.82 0.13 0.82 0.56 -0.32Zimbabwe   0.10 0.04 0.60 0.01 0.02 0.25

2074 WORLD DEVELOPMENT

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 7/11

To capture the effect of traditional trade policies as opposedto other trade-related costs we use a traditional cross-sectiongravity model that includes time invariant trade impediments(distance, adjacency, common language, access to the sea)augmented with a set of trade policy variables. Border barriersare captured by the TTRI, the NTM component of the OTRI(defined as the difference between the OTRI and the TTRI)and the relative preferential margin (RPM). Deeper economicintegration is controlled for by adding a dummy for the exis-tence of a bilateral trade agreement. Domestic trade costs arecaptured by the two trade facilitation indicators discussedabove: the LPI and the Doing Business cost of trading. 10

The dataset covers 105 countries. All data are for 2006.Trade flow data are from WITS (i.e., United NationsCOMTRADE), GDP data are from the World Bank’s WorldDevelopment Indicators, the standard gravity variables areobtained from the Trade Production and Protection database(Nicita & Olarreaga, 2007), and trade policy data are obtainedfrom the OTRI database (Kee et al. 2008; Kee et al. 2009). Fi-nally, data on the LPI and domestic trade costs were sourcedfrom the LPI and the Doing Business websites.

The approach used to estimate the gravity model follows theexisting literature. However, we are subject to the limitation of 

a cross-sectional dataset. Thus the presence of unobserved rel-ative trade impediments that a country has with all its tradingpartners is controlled for by using multilateral resistance termsrather than the country-time fixed effect commonly used in pa-nel settings. Thus, multilateral resistance (Anderson & VanWincoop, 2003) is proxied by adding multilateral resistancevariables as in Baier & Bergstrand, 2009 and Baier, Bergstrandand Mariutto, 2010. 11 This methodology produces consistentestimates and, contrary to using country fixed effects, allowsthe estimation of the impact of country specific factors suchas the LPI score and the cost of trading as measured in DoingBusiness.

A recurring problem with gravity estimation is the presenceof zero trade observations. As the gravity model is estimatedin a log-normal specification, it will discard observationswhere there is no trade. Recent procedures to take intoaccount zero trade flows are the Poisson estimation (pseudo-Poisson maximum likelihood—PPML), or a two-stage estima-tion procedure. Although the incidence of zeros in our sampleis limited (trade flows occur across about 90% of countries),our preferred model is PPML. This model, besides beingrobust to truncation, also controls for heteroskedasticity byusing robust standard errors. 12

Table 6   reports the trade related PPML estimated coeffi-cients for a series of specifications with and without thedomestic trade costs terms. We start with the simpler specifica-tion consisting in a traditional cross section gravity model,augmented with two trade policy variables: the TTRI dis-cussed above and the OTRI purged by the TTRI component

so as to isolate the marginal effect of NTMs. This specification

controls   for multilateral resistance with the a-theoreticalterms. 13 Specification 2 is identical to 1 but multilateral resis-tance is controlled for by adding  multilateral resistance termsas in (Baier & Bergstrand, 2009). 14 Specification 3 is identicalto 2 but adds the RPM term, while specifications 4 and 5 addthe Doing business and the LPI terms, respectively. Finally,specification 6 includes all trade policy and trade cost vari-ables.

Results are generally typical of those of gravity equationmodels. Distance is an important determinant of bilateraltrade, as is a common border and common language. Land-locked countries tend to trade less and larger and more popu-lous countries tend to trade more. With regard to trade policyvariables the coefficients are of the expected sign and are sig-nificant across all specifications. In particular, countries thatare partners in a reciprocal trade agreement are found to haveabout 20% more trade. Both tariffs and NTMs are statisticallysignificant determinants of bilateral trade. On average, areduction in tariffs (as measured by the TTRI) of 10% wouldincrease trade volumes a little more than 2%, while a similarreduction in NTMs would add another 1.7% (as in specifica-tion 2). 15

Relative preferences matter a lot. Countries with an average

1% point advantage over competitors are estimated to tradeabout 2.3% more with their counterpart. This suggests that,even controlling for deeper integration (the bilateral agree-ment), tariff preferences are significant not only in and of themselves (in terms of affecting the TTRI), but also in regardto the relative advantage that is provided versus other coun-tries (the RPM). Moreover, the inclusion of the relative pref-erence margin results in somewhat lower coefficient estimatesfor the effect of a bilateral trade agreement. This suggests thatbilateral trade agreements have impacts that go beyond thatprovided by nonreciprocal tariff preferences.

Besides trade policy, the impact of trade costs on trade flowsis also estimated to be sizeable. Internal trade costs, as cap-tured by the Doing Business indicator, are found to be similarin magnitude for imports and exports. The elasticity of im-ports to the domestic cost of importing is about 0.50, and thatof exports to the domestic cost of exporting is about 0.48. Thatis, a 10% reduction in the cost associated with importing(exporting) would increase imports (exports) by about 5%(4.8%). Significant and qualitatively similar results are alsofound in regard to the LPI score. However, the larger coeffi-cient on the LPI for exporters indicates that the quality of logistics matters relatively more for them. One possible expla-nation could be that competition among exporters is generallymore intense, so that better logistics services can provide moreof an edge. The last specification (6) includes both the DoingBusiness indicators and the LPI score. The results suggestsome overlap in the costs captured by these two indices asthe coefficient estimates for specification 6 are lower than in

specifications 4 and 5. Nonetheless, the fact that all coefficients

Table 5.  Measures of trade costs (averages by country group)

Policy/indicator High income Middle income Low income

Trade policy tariff (imposed) (%) 1.98 6.86 10.22Trade policy NTB (imposed) (%) 4.45 4.45 6.71Trade policy tariff (faced) (%) 2.77 2.79 6.02Trade policy NTB (faced) (%) 4.56 4.72 5.07Trade policy RPM (%) 0.04 0.44 0.23

Doing business costs of exports (USD) 774.40 867.20 949.30Doing business costs of import (USD) 813.60 1024.20 1212.00Logistic performance index (higher better) 3.90 3.00 2.80

TRADE POLICY, TRADE COSTS, AND DEVELOPING COUNTRY TRADE 2075

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 8/11

maintain their significance suggests that each is capturing dis-

tinct dimensions of the trade facilitation environment.Although PPML is often used in gravity model estimation,

other models may fit the data better. As a robustness check,we also estimate the model in standard OLS (by adding 1 tothe trade values and taking the log), zero inflated Poisson(ZIP), negative binomial regression (NBREG), and zero in-flated negative binomial regression (ZINBR) as in   Burger,van Oort, and Linders (2009). 16 Table 7 reports the relevantcoefficients for specification 6 of  table 6 estimated with thesealternative models. 17 In general, the coefficients are similaracross the various models. This is partly due to the limited pres-ence of zero observations in our dataset (around 10%). How-ever, while the results of the PPLM and the ZIP are quitesimilar, some differences arise with respect to OLS, NBREGand ZINB. The alternative models indicate that tariffs play a

relatively more important role than NTMs. The NBREG and

ZINB models, but not OLS, also suggest lower impacts of domestic trade costs (Doing Business and LPI). All in all, how-ever, the models suggest qualitatively similar results: domestictrade costs are a significant determinant of trade flows. 18

From a policy perspective a relevant question is how muchdifferent sources of trade costs matter for low income coun-tries. To get a sense of this we can assess the magnitude of the trade-impeding effect of border trade policies and internaltrade costs. Table 8 reports some simple calculations of the im-pacts of trade policy on trade flows. The calculations give theincrease in trade for low income countries if they were to con-verge to a set of policies that would generate the observedaverage levels of the various policy indicators in middle-income countries (as reported in   Table 5). The calculationsare based on the results of specification 6 in  Table 6. 19

Table 6.  Gravity model results.

Variable Poisson Maximum Likelihood Estimation(Dependednt variable, volume of trade)

Expected Sign (1) (2) (3) (4) (5) (6)

GDP Importer + 0.837*** 0.722*** 0.724*** 0.701*** 0.552*** 0.606***

(log) + (0.041) (0.045) (0.045) (0.041) (0.041) (0.069)GDP Exporter + 0.739*** 0.620*** 0.622*** 0.617*** 0.318*** 0.430***

(log) + (0.030) (0.024) (0.023) (0.022) (0.058) (0.053)Population Importer + 0.056 0.089** 0.084** 0.077** 0.209*** 0.149***

(log) + (0.040)   (0.042)   (0.042)   (0.036)   (0.049)   (0.051)Population Export + 0.146*** 0.191*** 0.187*** 0.168** 0.354*** 0.312***

(log) (0.059)   (0.048)   (0.047)   (0.033)   (0.061)   (0.051)Multilateral Resistance + 1.246*** 0.007*** 0.007*** 0.006** 0.008*** 0.007***

(Distance) (0.163)   (0.001)   (0.001)   (0.001)   (0.001)   (0.001)Multilateral Resistance – 1.413*** 0.005*** 0.004*** 0.001*** 0.044** 0.029*

(Board) (0.131) (0.015) (0.015) (0.014) (0.018) (0.015)Land Importer   0.019   0.030   0.037 0.012   0.080   0.032(dummy) (0.131) (0.139) (0.138) (0.131) (0.132) (0.129)Land Exporter   0.078   0.109   0.106   0.061   0.168   0.126(dummy) (0.100)   (0.104)   (0.104)   (0.103)   (0.104)   (0.105)Distance (log)   0.784**** 0.786*** 0.784*** 0.748*** 0.701*** 0.705

(0.054)   (0.054) (0.054) (0.050)   (0.052)   (0.050)

Common Board + 0.428

***

0.216 0.184 0.331

**

0.374

**

0.401(dummy) (0.143) (0.159) (0.159) (0.156) (0.152) (0.150)Common language +   0.038** 0.228   0.210   0.228   0.270* 0.261(dummy) (0.140) (0.171) (0.166) (0.165) (0.162) (0.161)Bilateral trade agreement + 0.199** 0.227** 0.199** 0.222*** 0.203*** 0.213***

(dummy) (0.092) (0.088) (0.087) (0.081) (0.085) (0.083)Trade policy tariff (log)   0.209*** 0.211*** 0.204*** 0.176*** 0.221*** 0.198**

(0.041) (0.041) (0.041) (0.044) (0.042) (0.044)Trade policy NTM(log)   0.185*** 0.172*** 0.173*** 0.134*** 0.167*** 0.146***

(0.053) (0.054) (0.054) (0.047) (0.050) (0.046)Trade policy RPM + 0.023** 0.023** 0.030*** 0.027***

(0.011) (0.010) (0.008) (0.008)DB import costs(log)   0.496*** 0.324***

(0.094)   (0.094)DB import costs (log)   0.485*** 0.222

(0.108) (0.096)LPI importer(index) + 0.671*** 0.408***

(0.142)   (0.149)LPI exporter(index) + 0.905*** 0.701***

(0.168) (0.150)Pseudo  R-squared 0.872 0.874 0.879 0.888 0.885 0.897Observations 10920 10920 10920 10920 10920 10920

Note: Robust standard errors are in parentheses. Multilateral resistance terms in the first specification are a-theoretical: partner GDP-weighted averages of each importer and exporter distance from all of its trading partners.* Significance level of 10%.** Significance level of 5%.

*** Significance level of 1%.

2076 WORLD DEVELOPMENT

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 9/11

These results suggest that traditional trade policies still havea substantial trade-restricting effect for low income countries.

A reduction of their TTRI to the level observed in middle-income countries would increase their imports by about6.5%. Improving market access through reductions in tariffsin export markets also remains an important policy objective.If low income countries were to improve their market access tothat which confronts middle income countries this would in-crease their exports by more than 10%. NTMs are also astrong determinant of trade flows, but our simulation suggeststhat their impact is generally less than that of tariffs. In thecase of exports in particular the limited impact is due to thefact that the level of NTMs faced by low- and middle-incomecountries is similar. Thus, this result should not be interpretedas NTMs not being an important factor affecting trade flows.The effects of relative preferences, for low income countries are

on average very small. This is driven by the small difference inthe average relative preference margin for middle- and low-in-come countries (0.2%). Relative preferences are important ona country by country basis, not across the averages that applyto each of broad grouping by income.

Domestic costs, measured by the Doing business and theLogistic Performance Index are quantitatively important. Of the two, the LPI has the largest effect on trade. The domestictrade costs captured by the Doing Business variable have lessof an impact on trade. However, given that the Doing Businessand the LPI indicators capture, at least in part, similar domes-tic trade costs, their relative importance is difficult to assess. Amore proper interpretation is to use their sum as the total ef-fect of domestic costs. This suggests that trade costs representan important bottleneck for low income countries, both in re-

gard to imports (13.5%) and to exports (17%).

As a caveat, it should be borne in mind that these resultsoriginate from a regression based on cross sectional data. Thisgenerally does not allow unambiguous conclusions regardingthe direction of causation, as issues of endogeneity may notbe completely controlled for. However, the results clearly indi-cate that there is strong correlation between trade costs, bothdomestic and at the border, and trade flows. In general terms,these results indicate that administrative and regulatory poli-cies are at least as important as traditional border policies inimpeding trade. This supports the recent focus of many devel-oping countries on taking action to facilitate trade by reducingdomestic trade costs.

6. CONCLUDING REMARKS

The focus of policy debates and international cooperation ismore and more on non-tariff measures, regulatory policies,and on efforts to facilitate trade. In this paper we investigatethe impact on trade of a subset of such policies, and comparethese with the effects of traditional trade policies. Our findingsindicate that while traditional trade policies continue to beimportant in developing countries as well as for some sectorsin high-income countries (agriculture in particular), non-tariff measures and domestic trade costs are also of great impor-tance. In particular, the analysis in this paper suggests thatthe impact of reducing the costs associated with policies thatreduce the behind-the-border domestic cost of trade could

have a greater payoff than further reductions in border barri-ers such as tariffs and NTMs, or seeking additional trade pref-erences. Our results imply that focusing attention on thepolicies that affect logistics performance and the internal costof trading is likely to generate large trade gains, especially interms of exports. Given that the LPI and the Doing Businesscost of trading capture the effect of a multiplicity of specificpolicies, from a practical perspective this suggests that apriority for both research and policy advice is to further“unpack” these findings.

The analysis also makes clear that there are still large tradegains to be had from traditional trade liberalization. Whilebilateral and regional trade agreements have proliferated, pro-gress at the multilateral level has unfortunately been slow.Bringing the Doha Round to a successful conclusion and thus

capturing some of these potential trade gains is, therefore,

Table 7.  Robustness checks: alternative models

0LS PPML NBREG ZIP ZINB

Trade policy tariff (log)   0.315*** 0.198** 0.312*** 0.197*** 0.299***

(0.025) (0.044) (0.035) (0.044) (0.032)Trade policy NTB (log)   0.030   0.146*** 0.053* 0.146*** 0.054**

(0.025) (0.046) (0.028) (0.046) (0.026)Trade policy RPM (index) 0.016* 0.027*** 0.023** 0.027*** 0.025**

(0.010)   (0.008)   (0.010)   (0.008)   (0.010)DB Import Costs (log)   0.098* 0.324*** 0.240*** 0.326*** 0.245***

(0.057)   (0.094)   (0.078)   (0.094)   (0.070)DB Export Costs (log)   0.394*** 0.222** 0.201*** 0.224** 0.168***

(0.057)   (0.096)   (0.070)   (0.096)   (0.065)LPI importer (index) 0.357*** 0.408*** 0.279** 0.403*** 0.300***

(0.083)   (0.149)   (0.109) (0.149)   (0.100)LPI exporter (index) 1.182*** 0.701*** 0.135 0.695*** 0.150

(0.087) (0.150) (0.108) (0.150) (0.101)

Standard errors are in parentheses.* Significance level of 10%.

** Significance level of 5%.***Significance level of 1%.

Table 8.  Effects of policy convergence by low income countries to middleincome average (percent increase)

Policy/Indicator Increase inImports (%)

Increase inExports (%)

Trade Policy Tariff 6.5 10.6Trade Policy NTB 4.9 1.0Trade Policy RPM 0.6 0.6Doing Business 5.0 1.9Logistic Performance Index 8.5 15.1

TRADE POLICY, TRADE COSTS, AND DEVELOPING COUNTRY TRADE 2077

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 10/11

important. Making progress to reduce internal trade costsdoes not require or hinge on multilateral (or bilateral) negoti-ations – the costs that are incurred by traders in developingcountries can and should be reduced through unilateral

actions. The analysis in this paper strongly supports theargument made by   Ikenson (2008)   that there is great scopeto enhance economic growth opportunities   “while Dohasleeps.”

NOTES

1. Although important, other types of NTMs such as rules of origin andgovernment procurement could not be included in the analysis of thispaper as comprehensive and suitable data are lacking.

2. The inclusion of NTMs in the OTRI is done through estimation of ad-valorem tariff equivalents. Both the TTRI and the OTRI provide ameasure of the uniform tariff equivalent of observed policies that is neededto generate the observed level of trade for a country. See Kee et al. (2009)for details.

3. These indices are superior to more commonly used indicators such asaverage tariffs or NTM frequency and coverage ratios as they take intoaccount the elasticity of import demand with respect to prices. Incalculating the indices more weight is given to products for which demandis more responsive to changes in prices (so that smaller movements inprices produce larger shifts in imports).

4. As NTM data have not been updated recently, the change in the OTRIis not reported.

5. As trade flows are generally reported at the six digit level of theHarmonized System (HS), this indicator is constructed at the HS 6 digitlevel instead of the tariff line level.

6. This is the approach taken by   Low, Piermartini, and Richtering(2008).

7. This simpler methodology would be consistent with a framework

where export composition is a function of the structure of the tariff of theimporting country (everything else equal, exports would concentrate inproduct lines with lower tariffs), however, as the intent is to assess thepreferential margin applied on the existing structure of trade, the measurecontrolling for product composition is to be preferred. Also, exportcomposition is more likely to be determined by other factors affectingcomparative advantage, such as level of development, land quality,climate, geography and labor force skills.

8. Complicating the framework even more, one should take into accountsubstitution possibilities across similar products originating in differentcountries. For simplicity, we abstract from this and assume thatsubstitution elasticities are equal to one.

9. See Kee, Nicita, and Olarreaga (2008)   for the methodology used toestimate import demand elasticities.

10. This approach allows us to assess the impact of border and domestictrade impediments on international trade. However, it does not allowdisentangling how these impediments affect the fixed and variable costs of 

trading. Although this is an important question that is the subject of recentresearch (Lawless, 2010; Bernard et al., 2009), it is beyond the scope of thispaper.

11. Eac h of the se bilate ral variables (var) is c onstruc te d as MRvarij  ¼

Pk varik  þ

Pmvarm j

Pk 

Pmvarmk  .

12. See Silva and Tenreyro (2006) for the rationale behind Poisson MLE.

13. That is, by adding   ad hoc  remoteness variables constructed as theGDP-weighted averages of each country’s distance from all of its tradingpartners (Wei, 1996;  Baier & Bergstrand, 2004.). We also estimated the

model by alternatively using importer and exporter fixed effects. Thisresulted in qualitatively similar coefficients.

14. Following the existing literature, we add two terms for multilateralresistance in our preferred specification: distance and border. Resultsobtained by adding two additional terms (language and trade agreements)were similar.

15. The effect of NTMs is captured at the margin, that is given the effectof the existing tariff structure.

16. For ZIP and ZINBR the presence of a zero trade flow is determinedby the standard gravity variables.

17. Results from other specifications are qualitatively similar.

18. To identify the most appropriate model for our analysis we base thechoice on visual inspection of how the various functional forms fit thedata. Given the limited presence of zeros in the dataset, the choice isrestricted between the PPML and the NBREG. A plot of the mean-variance relationship of the PPML and NBREG did not indicate a clearwinner. The Poisson model has a better fit for most of the data, whilenegative binomial appears to fit better the few extreme high values such asintra-NAFTA trade and China-USA trade. In the absence of a clearlysuperior model and to be consistent with the mainstream literature on thissubject, we choose to report the main results of  Table 6   based on thePPML model.

19. For the logged variables the change in trade is simply the estimatedelasticities multiplied by the change in the policy variable. For the indexvariables (RPM and LPI) the change in trade is calculated as theexponential of the difference in the index multiplied by the coefficient,minus 1.

REFERENCES

Anderson, J., & Marcouiller, D. (2002). Insecurity and the Pattern of Trade: An Empirical Investigation.  Review of Economics and Statistics,84(2), 342–352.

Anderson, J., & Van Wincoop, E. (2003). Gravity with gravitas.  A solutionto the border puzzle. American Economic Review, 93 , 170–192.

Anderson, J., & Van Wincoop, E. (2004). Trade Costs.   Journal of Economic Literature, 42(3), 691–751.

Baier, S., & Bergstrand, J. (2004). Economic determinants of free trade agreements.   Journal of International Economics, 64,29–63.

2078 WORLD DEVELOPMENT

7/18/2019 Hockman and Nicita (2011)_Trade Policy, Trade Costs, And Developing Country Trade

http://slidepdf.com/reader/full/hockman-and-nicita-2011trade-policy-trade-costs-and-developing-country 11/11

Baier, S., & Bergstrand, J. (2009). Bonus vetus OLS: A simple method forapproximating international trade-cost effects using the gravity equa-tion.  Journal of International Economics, 77 (1), 77–85.

Baier, S., Bergstrand J & R. Mariutto. (2010). The Growth of Bilateralism.CAGE Online Working Paper Series 12, Competitive Advantage in theGlobal Economy (CAGE).

Bernard, A. B., Bradford, J. J., Redding, S. J., & Schott, P. K. (2009).The Margins of US Trade.   American Economic Review, 99(2), 487– 

493.Burger, M. J., van Oort, F. G., & Linders, G. M. (2009). On thespecification of the gravity model of trade: zeros, excess zeros and zero-inflated estimation.  Spatial Economic Analysis, 4, 167–187.

Deardorff, A. V. & Stern R. M. (1998).   Measurement of Nontariff Barriers:Studies in International Economics. Ann Arbor MI: Universityof Michigan Press.

Djankov, S., Freund, C., & Pham, C. (2010). Trading on Time.  Review of Economics and Statistics, 92(1), 166–174.

Francois, J. & Manchin M. (2007). Institutions, Infrastructure, and Trade.Policy Research Working Paper  No. 4152. The World Bank.

Ikenson, D. (2008). While Doha Sleeps: Securing Economic Growththrough Trade Facilitation.  Cato Institute Trade Policy Analysis  No.37.

Kee, H. L., Nicita, A., & Olarreaga, M. (2009). Estimating traderestrictiveness indices.  The Economic Journal, 119(534), 172–199.

Kee, H. L., Nicita, A., & Olarreaga, M. (2008). Import demand elasticitiesand tradedistortions. Review of Economicsand Statistics, 90(4),666–682.

Lawless, M. (2010). Deconstructing gravity: trade costs and extensive andintensive margins.  Canadian Journal of Economics, 43(4), 1149–1172.

Limao, N., & Venables, A. J. (2001). Infrastructure, GeographicalDisadvantage, Transport Costs, and Trade.   World Bank EconomicReview, 15(3), 451–479.

Low, P., Piermartini, R., & Richtering, J. (2008). Multilateral Solutions to

the Erosion of Non-Reciprocal Preferences in NAMA. In B. Hoek-man, W. Martin, & C. Primo Braga (Eds.),  Trade Preference Erosion.Palgrave-McMillan.

Nicita, A., & Olarreaga, M. (2007). Trade, Production, and ProtectionDatabase 1976–2004.  World Bank Economic Review, 21(1), 165–171.

Silva, S., & Tenreyro, S. (2006). The Log of Gravity.   The Review of Economics and Statistics, 88(4), 641–658.

Wilson, J. S., Mann, C., & Otsuki, T. (2003). Trade Facilitation andEconomic Development: A New Approach to Measuring the Impact.World Bank Economic Review, 17 (3), 367–389.

World Bank and IMF. (2008).   Global Monitoring Report, 2008. Wash-ington DC: World Bank.

Bank, World. (2007). Connecting to Compete. Trade Logistics in the localEconomy. Washington DC: World Bank.

Bank, World. (2008). Doing Business Report 2008. Washington DC: WorldBank.

 Available online at www.sciencedirect.com

TRADE POLICY, TRADE COSTS, AND DEVELOPING COUNTRY TRADE 2079