Vi Digital Library - Text Preview
On the Importance of Market Access for Trade
Working paper by Fugazza, Marco, Nicita, Alessandro, 2011
Download original document (English)
UNITED NATIONS CONFERENCE ON TRADE AND DEVELOPMENT
POLICY ISSUES IN INTERNATIONAL TRADE AND COMMODITIES
STUDY SERIES No. 50
ON THE IMPORTANCE OF MARKET ACCESS
New York and Geneva, 2011
The purpose of this series of studies is to analyse policy issues and to stimulate discussions
in the area of international trade and development. The series includes studies by UNCTAD staff
and by distinguished researchers from academia. This paper represents the personal views of the
authors only, and not the views of the UNCTAD secretariat or its member States.
The designations employed and the presentation of the material do not imply the
expression of any opinion whatsoever on the part of the United Nations Secretariat concerning the
legal status of any country, territory, city or area, or of its authorities, or concerning the
delimitation of its frontiers or boundaries.
Material in this publication may be freely quoted or reprinted, but acknowledgement is
requested, together with a reference to the document number. It would be appreciated if a copy of
the publication containing the quotation or reprint could be sent to the UNCTAD secretariat at the
Trade Analysis Branch
Division on International Trade in Goods and Services, and Commodities
United Nations Conference on Trade and Development
Palais des Nations
Trade Analysis Branch
UNITED NATIONS PUBLICATION
© Copyright United Nations 2011
All rights reserved
One of the consequences of the proliferation of preferential trade agreements is that an
increasing share of international trade is not subject to the most favoured nation (MFN) tariff, but
enters markets through preferential access. Preferential access affects trade because, by providing
some countries with a relative advantage, it is essentially a discriminatory practice. This paper
examines the extent to which preferential access affects bilateral trade flows. The empirical
approach consists first in providing two indices: one summarizing direct market access conditions
(the overall tariff faced by exports) and one measuring relative market access conditions (the
overall tariff faced by exports relative to that faced by competitors). Then, the indices are used in a
gravity model in order to estimate how changes in market access conditions affect international
trade. Although those conditions are generally more important, the results indicate that the relative
advantage provided by the structure of preferences also affects the magnitude of bilateral trade
flows. That is, bilateral trade flows depend on the advantage provided by the system of preferences
over other competitors.
Keywords: Gravity model, trade policy, international trade flows, market access; tariffs.
JEL Classification: F10, F15
We thank those who participated in the Geneva Workshop Series on Trade and
Development and the European Trade Study Group 2010 Conference for their valuable
comments and suggestions.
Any mistakes or errors remain the authors' own, who may be contacted at the following
e-mail address: firstname.lastname@example.org.
2. Market access and trade flows ............................................................................................3
2.1 Market access...............................................................................................................3
2.2 Estimating the effect of preferences on trade flows.....................................................5
2.3 The relative preferential margin and the theoretically based gravity model ................7
3. Data ......................................................................................................................................9
4. Results ...................................................................................................................................9
4.1 Tariff trade restrictiveness and relative preferences ....................................................9
4.2 Econometric results....................................................................................................14
4.2.1 Core specifications ............................................................................................14
4.2.2 Robustness checks.............................................................................................16
4.3 Impact of preferential access on trade flows..............................................................18
5 Conclusions .........................................................................................................................21
List of figures
Figure 1. Distribution of direct and relative market access indices ..............................................12
Figure 2. Correlation in the changes in tariff trade restrictiveness index
and the relative preferential margin (2000–2007) .........................................................13
List of tables
Table 1. Indices of direct and relative market access conditions, by country .............................10
Table 2. Gravity model estimation ..............................................................................................15
Table 3. Gravity model estimation: robustness checks ...............................................................16
Table 4. Gravity model results with standard measure of preferential margin............................18
Table 5. Effects of changes in market access conditions, by country .........................................19
In the past twenty years, trade liberalization has been used as an effective development
tool, based on the evidence that there are many benefits that a country can gain through a more
active participation in world trade. While tariff liberalization was initially pursued through trade
agreements under the auspices of the World Trade Organization (WTO), preferential trade
agreements (PTAs) are the basis behind the recent trade liberalization process. The proliferation of
PTAs in the recent past has been impressive. At the launch of WTO in 1994, only 37 such
agreements were in place. By 2008, more than 200 had been implemented with more in the
implementation stage. Participation in regional and bilateral trade agreements is widespread, as
virtually all members of WTO have notified participation in one or more PTAs. One consequence
of the large number of PTAs is that an increasing share of international trade is not subject to the
most favoured nation (MFN) tariff, but enters markets through preferential access.1 Preferential
access affects trade because, by providing some countries with a relative advantage, it is essentially
a discriminatory practice.
In general, the existing literature on the effects of trade liberalization suggests that
preferential access has a large impact on trade flows. For example, Baier and Bergstrand (2007)
find that the average impact of free trade agreements is to double the bilateral trade after a 10-year
period. Similarly, Magee (2008), in estimating the anticipatory and long-run impact of PTAs, finds
that PTAs have increased trade among members by an average of 90 per cent. Baier and
Bergstrand (2009), using a non-parametric estimation, find substantially similar results in the long-
run effects of the European Community and the Central American Common Market. Carrère
(2006) examines the effects of seven regional trade agreements and finds that they have resulted
both in trade creation and trade diversion, often at the expense of non-member countries. Lee and
Shin (2006) analyse East Asian free trade agreements and find an increase in trade among members
as well as trade diversion, depending on certain characteristics of member countries. Clausing
(2001), and Calvo-Pardo et al. (2009) find trade creation but no trade diversion effects with regard
to the free trade agreement concluded between the United States of America and Canada, and the
regional trade agreement of the Association of Southeast Asian Nations (ASEAN).2 Most of the
literature has generally examined the overall impact of PTAs as a discrete event rather than
focusing on tariff liberalization.3 Such an approach captures not only tariff changes but also any
advantage that PTAs can carry, such as customs harmonization, a reduction in non-tariff measures,
trade facilitation and a decline in other trade costs. Although informative, an approach that
examines the overall impact of PTAs is not suitable for isolating the effect of tariff preferences on
This paper adds to the existing literature by specifically examining the effects of changes
in market access conditions in terms of tariffs, whether caused by PTAs, or by multilateral or
unilateral liberalization. In doing so, the paper provides two contributions. The first is two indices
measuring market access conditions that take into account the complex structure of tariff
preferences. Those indices are calculated at the bilateral level to summarize market access
conditions affecting each bilateral trade relationship. One index summarizes the tariffs faced by
exports and is related to the work on trade restrictiveness (Kee et al., 2008, 2009). The other index
1 Although 40 per cent of world trade is free under MFN regimes, an additional 30 per cent is exempted from
tariffs because of preferential access.
2 One dissenting study is Ghosh and Yamarik (2004). In their analysis of 12 regional trade agreements, they
are sceptical about the results of the previous literature on a positive trade-creation effect. The use of fixed-
effect estimation in the subsequent literature has somewhat alleviated their criticism.
3 One exception is a study by Robertson and Estevadeordal (2009). Their findings suggest that the tariff
liberalization of Latin American countries between 1985 and 1997 caused trade-diverting effects.
measures the relative tariff advantage or disadvantage that the tariffs provide compared with other
competitors and builds on the work on preferential margins (Hoekman and Nicita, 2008; Carrère et
al., 2008; and Low et al., 2009). Second, this paper contributes to the literature by examining
whether trade between two countries depends not only on changes in the tariff applied to trade, but
also on market access conditions applied to third countries. The empirical approach used to
measure the extent to which changes in market access conditions affect international trade consists
of a gravity model augmented by the two indices.
The two indices show that that bilateral market access conditions have generally improved
during the period of analysis and that the system of preferences is on average not very
discriminatory. However, both indices have a large variance indicating that in some cases
preferential access does provide a substantial advantage. The econometric results show that,
although the impacts of tariffs applied to trade are the most important, relative market access
conditions also affect trade flows. That is, bilateral trade flows are found to be larger, the greater
the relative advantage provided by the system of preferences. On average, the results indicate that
the effect of change in market access conditions on trade has had a modest effect on trade values,
largely because market access conditions have not changed dramatically. Still, looking beyond
averages, the results show a large variance, indicating that for some countries, changes in market
access conditions have produced substantial effects on trade.
This paper is organized as follows. Chapter 2 illustrates the empirical approach for
assessing the impact of preferential access on trade flows. Chapter 3 briefly summarizes the data.
Chapter 4 provides some statistics on market access measures and discusses their impact on trade
flows, and chapter 5, conclusions.
2. Market access and trade flows
In the last decade, market access conditions have increasingly been affected by bilateral
trade agreements. Trade agreements generally provide trading partners with lower tariffs and as a
result, different tariff rates are applied to the same product, depending on its origin. As of 2007,
only 40 per cent of international trade is discrimination-free, as each given country applies the
same tariff to all trading partners at the harmonized system (HS) six-digit product. About 30 per
cent of trade is in products where there are two different tariff rates, and about 15 per cent, in
products with three different tariff rates. The remaining 15 per cent consists of products for which
there are four or more different tariff rates.
From an exporter's perspective, market access depends not only on the disadvantages that
exporters face versus domestic producers, but also on the relative advantages or disadvantages that
exporters have over other external competitors. Tariffs affect both of these components. In tariff
terms, the disadvantage versus domestic competitors is simply given by the tariff applied to the
specific good, while the advantage or disadvantage versus other competitors is given by the
The preferential margin is directly related to preferential access. For example, high-income
countries often grant non-reciprocal preferential access to least developed countries in order to
facilitate the latter's economic growth by providing an incentive to their exports. Likewise, regional
trade agreements are a common form of reciprocal preferential access in which lower or zero tariffs
are applied to products originating among members so as to foster trade cooperation. Such
agreements, by providing some trading partners with a lower tariff, inevitably discriminate against
those trading partners outside the trade agreement (Hoekman et al., 2009). Preferential access also
implies that trade agreements can vary in strength, depending on the provided relative advantage.
Some trade agreements may give great advantage because of high external tariffs; others may have
more muted effects because preferential treatment is granted to a large number of countries. The
above discussion argues that the entry into force of a trade agreement affects not only the signing
countries, but all other trading partners.
The following two chapters illustrate the steps to be taken to measure the effect of market
access on trade flows. The first chapter illustrates two indices measuring market access conditions
faced by exports. One index summarizes the tariffs faced by exports; another index measures the
preferential margin at the bilateral level. The second chapter lays down the estimating framework
utilized in assessing the contribution of the two indices to explain bilateral trade flows.
2.1 Market access
To measure market access conditions, we provide two trade policy variables: the first
measure captures direct market access conditions (the overall tariff faced by exports), the second
measure captures relative market access conditions (the overall tariff faced by exports relative to
that faced by competitors). Both measures are calculated at the bilateral level.
The variable capturing the tariff restrictiveness faced by exports is an index based on the
work of Kee et al. (2008, 2009). In the construction of this index, the aggregation across products
takes into account the fact that imports of some goods can be more responsive than others to price
changes. In aggregating across tariff lines, products where imports are less sensitive to prices
(inelastic) should be given less weight because preferential access (a lower tariff) would have less
effect on the overall volumes of trade. We call this variable the tariff trade restrictiveness index
(TTRI). In formal terms, the TTRI faced by country j in exporting to country k is:
where, exp are exports, ε is the import demand elasticity, T is the applied tariff and hs are
HS six-digit categories.4 In practice, this index provides the equivalent uniform tariff that will
maintain exports from country j to country k constant.5
The variable capturing the stance of the system of preferences relative to competitors is
provided by another index measuring the preferential margin. The commonly used measure of
preference margins is simply the difference between the preferential tariff and the MFN rate. As in
most instances, other countries will also have some form of preferential access: this measure
generally overestimates the actual preferential margin. In practice, preferential rates granted to a
particular country, although lower than MFN, may still penalize it, compared with other countries
that benefit from an even lower, or no tariffs at all. The above discussion implies that the
advantage, in terms of tariffs, of a free trade agreement depends on the pre-existing structure of
To measure the effect of the advantage provided by the preferential tariff compared with
that faced by other competitors, this paper builds on the arguments of Hoekman and Nicita, 2008;
Carrère et al., 2008; and Low et al., 2009. These studies recognize that a better measure of the
preferential margin is one that takes into account the preferences accorded to third-party countries.
In this regard, the strength of a preferential trade agreement is measured by the relative preferential
margin (RPM), as it is relative to any preferential access provided to other competitors. In this
study, the RPM that a country grants to a given country is calculated as the difference in tariff
percentage points that a determined basket of goods faces when imported from the given country as
opposed to being imported from any other.6
There are two set of weights when calculating such a preferential margin. First, the
counterfactual is a weighted average of the tariff imposed on all other partners. Second, the overall
tariff and the preferential margin are averages constructed across many tariff lines. To calculate the
counterfactual, the first step is to calculate the trade-weighted average tariff at the tariff-line level
that one country (i.e. the United States) imposes on all other countries except the country for which
the preferential margin is calculated (i.e. Mexico). This is done by using (United States) bilateral
imports as weights, so as to take into account the supply capacity of (United States) trading
partners. The second step is to aggregate across tariff lines. This is done by using (Mexico) exports
(to the United States) so as to take into consideration the different product composition across
partners. As in the TTRI case, a further complication arises in the aggregation across products.7 To
correct this, in the calculation of the RPM, the HS six-digit product lines are aggregated using
import demand elasticities.
4 As a cross-country comparison of trade flows is only possible at the HS six-digit level, this indicator is
constructed on the basis of the 6 digits (HS).
5 See Feenstra (1995).
6 For example, in a proper measure of the preferential market access that Mexico enjoys in the United States,
the counterfactual is the average tariff for Mexico’s bundle of exports to the United States if this were to
originate in other countries. The RPM is the difference between the counterfactual and the bilateral trade-
weighted preferential tariff imposed by the United States on Mexico.
7 When aggregating across product lines, the overall RPM should be higher if the exporting country has a
higher preferential margin in products for which demand is more elastic to small movements in prices.
In more formal terms, the RPM measuring the advantage that exports of country j have in
exporting its goods to country k can be calculated as:
where exp are exports, ε is the import demand elasticity, T is the tariff, hs are HS six-digit
categories, and v are exporters competing with country j in exporting to country k. And where
hskT , , is the trade-weighted average of the tariffs applied by country k to imports originating from
each country v (for each HS six-digit product). Note that any measure of preference margin can be
positive or negative, depending on the advantage or disadvantage of a given country with respect to
other competing exporters. In summary, the RPM provides a measure of the tariff advantage (or
disadvantage) provided to the actual exports of country j in country k, given the existing structure
of tariff preferences.8
2.2 Estimating the effect of preferences on trade flows
The standard approach for measuring the impact of policy variables on trade flows is the
gravity model. This model relates bilateral trade to economic sizes and transport and transaction
costs, controlling for variables such as common languages and shared borders.9 The effect of a
specific trade policy is often estimated by including dummy variables for the presence of policy
factors affecting trade, such as trade agreements. Although the econometric estimation of this paper
follows that of the recent literature on gravity models, the approach used to identify the effect of
trade liberalization is different. The difference is in the fact that analysis is not based on discrete
events – that is, the effects resulting from the implementation of a PTA – but examines market
access conditions in terms of tariff changes, whether caused by PTAs or not.10 In this set-up, we
expect the impact of PTAs on trade due to any other trade-agreement-related factor to be absorbed
to a large extent by the inclusion of importer- and exporter-time-specific dummies, as well as
country-pair fixed effects.
In summary, the estimating framework consists of a panel gravity model where a set of
fixed effects addresses heterogeneity, multilateral resistance and control for all the determinants of
trade flows normally included in gravity model specifications. To capture the effect of change in
preferential access, the estimation includes the two trade policy variables discussed above: the
TTRI, which captures the absolute stance of the bilateral tariff between two countries, and the
RPM, which captures the stance of the bilateral tariff relative to all the other partners. By including
the RPM in the estimation, we determine whether the effect of the change in tariff is stronger the
greater the advantage it provides over other competitors. In practice, one should expect a negative
sign for the TTRI coefficient, as a higher tariff would hinder trade, and a positive sign for the RPM
8 A substantial advantage of this policy variable is that it also captures the strength of preferential access,
something that the use of dummy variables cannot control. For example, some trade agreements may offer a
great advantage because of a high external tariff; others may have more muted effects because of a low
external tariff or because preferential treatment is granted to a large number of countries.
9 Linder, 1961; Linnemann, 1966; Anderson and van Wincoop, 2003.
10 Owing to the lack of comprehensive data, the issue related to the utilization rates of trade preferences is not
taken into account. It can thus be assumed that all trade is subjected to the lowest available tariff.
coefficient, as a relatively higher preferential margin would provide the country with an advantage.
Note also that this specification provides the estimates necessary to measure the strength of trade
agreements in different countries, including third parties.
With regard to other issues related to the estimation of the gravity model, the inclusion of
gravity-type variables alone does not take into account all the factors impeding bilateral trade
flows. Well-specified gravity models consider not only frictions between pairs of countries, but
also frictions relative to the rest of the world. In particular, one needs to check for the presence of
unobserved relative trade impediments that a country has with all its trading partners (Anderson
and van Wincoop, 2003). Therefore, bilateral trade depends not only on bilateral trade costs but
also on the average trade costs faced or imposed by these countries. Multilateral resistance can be
verified by adding multilateral price terms or more commonly with country-fixed effects. In our
case, multilateral price terms are likely to be time varying and therefore it is necessary to estimate
the model by using country-time fixed effects. Moreover, in our setting, country-time fixed effects
would also capture any importer- and exporter-specific effects of the tariff regime. As the RPM
provides the relative advantage not with respect to the average, but to each trading partner, it also
captures the discriminatory effects of the system of preferences.
Trade models dealing with policy variables often suffer from endogeneity. That is,
countries choose to enter trade agreements with partners where trade flows are larger and thus
reduce tariffs. In cross-section models, such endogeneity is generally treated with the use of
instrumental variables. However, instrumental variable estimation may not be fully satisfactory for
treating policy variables because the endogeneity bias may be due to unobserved time-invariant
heterogeneity (Baier and Bergstrand, 2009). In a panel setting, such endogeneity bias is treated by
adding country-pair fixed effects (Baier and Bergstrand, 2007). Besides checking for gravity-type
variables such as distance and shared borders, country-pair fixed effects check for any unobserved
variable simultaneously affecting the change in the tariff and the level of trade.
In summary, the estimation of the effect on trade from changes in market access conditions
is based on a gravity model according to the following specification:
jktkjktjtjktjktjkt εθψωRPMβ)TTRIln(ββXln +++++++= 210 1 (3)
where the subscript j denotes exporters, k denotes importers and t denotes year; and where
X is the value of trade from country j to country k, TTRI is the tariff trade restrictiveness index as
in equation (1), RPM is the real preferential margin as in equation (2); jtω is importer-time fixed
effects, ktψ is exporter-time fixed effects, kjθ is importer-exporter pair fixed effects and jktε is an
independent and identically distributed term with mean zero and variance λ .11
11 Note also that country-pair dummies also soak up any variance due to the presence of time-invariant
preferential trade agreements. Thus our coefficients capture the effect of the change in tariffs, also checking
for implementation periods.
2.3 The relative preferential margin and the theoretically
based gravity model
The empirical framework discussed above can be reconciled with the theoretically based
gravity model as follows. In the standard Dixit-Stiglitz-Krugman set-up, country-k’s imports from
country j is given by:
where jkτ reflects all trade costs including tariffs, jY is nation-i’s output, kE is the
destination nation expenditure on tradable goods, σ is the elasticity of substitution (σ >1) among
all varieties from all nations (varieties are usually assumed to be symmetric for simplicity), kP is
nation-k’s ideal constant elasticity of substitution (CES) index (all goods are assumed to be traded)
and jΩ measures the real market potential of nation-j’ s exports.12
Trade costs can be redefined as jkjkjk ft=τ where jkt is the tariff component of trade
costs and jkf incorporates other trade costs such as freight costs, the latter being mostly a function
of geographical features. This definition of trade costs makes the price index that prevails in the
destination country an explicit function of tariffs applied to varieties coming from different
exporting countries. The properties of the price index do not allow separating tariffs from other
components of the various landed prices. This means that it is not possible to derive the RPM index
from the standard Dixit-Stiglitz-Krugman approach and, or from any approach using a CES utility
function as representative of consumer preferences. However, the aim of the paper is not to offer
an alternative theoretical modelling strategy. In order to reconcile our measure with standard
theory, we simply include it in equation (4) and assess the consequences in terms of empirical
strategy. By adding both to the numerator and the denominator, the two components of the RPM
index (the tariff applied to competitors and the tariff applied to country j), equation (4) becomes:
where wkt is the average tariff faced by all exporters to country k other than those from
country j.13 Then, using standard proxies and measures defined in the previous chapter, equation
(4’) can be rewritten as:
12 kP and jΩ are given respectively by ( )( )∑
1 σσ and ∑
στ , where
ikp is the landed price in nation k of goods produced in nation i and in is the number of varieties exported
from nation i. The landed price is made of the producer price in the country of origin augmented by trade
costs that are destination-specific and takes the standard iceberg form.
13 This average tariff is specific to the country of origin as it is computed using the number of varieties
exported by country j to country k and not all varieties imported by j. This makes it different from the tariff
component of price index prevailing in k. In this context, only the number of varieties matters not the variety
itself as they are all charged the same producer price within the same country.
( ) ⎟⎟⎠
++= −−−− w
)1( 1121 σσσ (4’’)
jk ≠= ∑
Note that the measure of average tariff ( wjkT ) does not reflect the tariff component of the
Anderson and van Wincoop resistance term unless exports to country k from any trade partner all
share the same composition in terms of products exported. Hence wjkT should not be absorbed by
the importer and time-fixed effects and should therefore be treated explicitly.
In the standard estimation approach, bilateral exports are weighted by the product of trade
partners’ gross domestic product (GDP). In our context, this would mean treating the (1+ wjkT )
variable independently. As this may raise some additional statistical issues because of the possible
correlation of the latter term with (1+ jkTTRI ), another implementation strategy may be necessary.
Instead of imposing a unity coefficient on the product of GDP as done with the standard weighting
procedure, we keep the product on the right-hand side and normalize it by (1+ wjkT ).
The empirical specification corresponding to (4') that we consider for estimation is thus
TTRIX εθψωββββ ++++⎟⎟⎠
where the notation is as before. Specification (5) provides a theoretically based robustness
for assessing the impact of trade preferences on exports. In practice, we do not expect considerably
different results in estimating equation (5) compared with equation (3).
The data utilized in this paper are comprehensive of trade flows and tariffs. Trade data
originate from the United Nations Commodity Trade Statistics Database, commonly known as UN
COMTRADE; tariff data (MFN and preferential rates) originate from the UNCTAD Trade
Analysis and Information System database, or TRAINS. Trade and tariff data are available through
the software, the World Integrated Trade Solution (WITS).14 Import demand elasticities are from
Kee et al. (2008); GDP data are derived from the World Bank World Development Indicators
database. Tariff, trade, and import demand elasticities data follow the HS at the six-digit level. The
underlining data to compute the bilateral TTRI and the RPM cover about 5,000 different products
for 85 countries, over 8 years (2000–2007).15 Another contribution of this paper is the provision of
a dataset on the bilateral TTRI and RPM indices for each year of the analysis. These data are
available from the authors on request.
This chapter first illustrates some descriptive statistics on the policy variables, and then it
discusses the estimation results from the gravity model. Finally, it summarizes the impact of trade
policy variables on bilateral trade.
4.1 Tariff trade restrictiveness and relative preferences
The first step of the analysis is to calculate the two policy variables measuring the market
access conditions that each country faces in exporting its products. The TTRI measures the tariff
restrictiveness that exports of a given country face in a determined market. The larger the given
country's TTRI, the higher the tariff restrictiveness faced by the country’s exports. The RPM
indices provide the average tariff advantage or disadvantage that the country has in exporting to
each trading partner relative to other competitors. A negative value of the RPM measure implies
that the country’s exports, on average, are relatively disadvantaged compared with its competitors.
The TTRI and the RPM indices are calculated at the bilateral level for every year of the analysis.
Some descriptive statistics for the first year (2000) and last year (2007) of the data are provided in
table 1 and figure 1.
14 WITS preferential data are not always complete for the earlier years of the analysis. We further validate the
data on tariff preferences by using some online databases (McGill Faculty of Law Preferential Trade
Agreements Database, the Tuck Trade Agreements Database and the WTO Regional Trade Agreements
15 The lack of reliable time series data on preferential tariff precludes the inclusion of a number of countries
in our sample. Still, the sample includes all major countries and covers more than 90 per cent of world trade.
Table 1 provides the list of the countries covered by the data.
Table 1. Indices of direct and relative market access conditions, by country
Absolute market access Relative market access
Country name TTRI 2000
Algeria 0.006 (0.018) 0.022 (0.028) -0.001 (0.009) -0.001 (0.003)
Argentina 0.071 (0.094) 0.034 (0.056) 0.000 (0.062) 0.018 (0.039)
Australia 0.038 (0.034) 0.035 (0.028) -0.001 (0.009) -0.001 (0.007)
Austria 0.009 (0.023) 0.006 (0.019) 0.010 (0.009) 0.003 (0.003)
Azerbaijan 0.004 (0.014) 0.010 (0.023) -0.001 (0.007) 0.000 (0.005)
Bangladesh 0.050 (0.051) 0.035 (0.055) 0.025 (0.045) 0.027 (0.040)
Belgium 0.031 (0.028) 0.009 (0.027) -0.029 (0.019) 0.004 (0.004)
Benin 0.096 (0.097) 0.188 (0.150) 0.015 (0.103) 0.007 (0.038)
State of 0.031 (0.032) 0.003 (0.010) 0.032 (0.035) 0.023 (0.059)
Brazil 0.046 (0.048) 0.037 (0.069) -0.002 (0.025) 0.004 (0.029)
Bulgaria 0.028 (0.037) 0.015 (0.037) -0.014 (0.016) 0.003 (0.013)
Cameroon 0.010 (0.024) 0.016 (0.049) -0.002 (0.007) 0.004 (0.007)
Canada 0.007 (0.024) 0.006 (0.020) 0.010 (0.008) 0.006 (0.009)
Chile 0.022 (0.028) 0.009 (0.016) 0.005 (0.019) 0.006 (0.012)
China 0.032 (0.032) 0.035 (0.030) -0.005 (0.011) -0.011 (0.014)
China, Hong Kong 0.055 (0.046) 0.042 (0.034) -0.012 (0.014) -0.001 (0.024)
Taiwan Province of 0.040 (0.036) 0.034 (0.024) -0.006 (0.012) -0.006 (0.009)
Colombia 0.046 (0.061) 0.007 (0.019) -0.002 (0.012) 0.023 (0.042)
Costa Rica 0.034 (0.044) 0.007 (0.033) -0.003 (0.012) 0.008 (0.016)
Côte d'Ivoire 0.049 (0.082) 0.029 (0.063) -0.006 (0.011) 0.003 (0.011)
Croatia 0.018 (0.033) 0.008 (0.025) -0.006 (0.014) 0.005 (0.007)
Czech Republic 0.022 (0.039) 0.004 (0.017) -0.021 (0.019) 0.004 (0.003)
Denmark 0.016 (0.031) 0.009 (0.023) 0.006 (0.011) 0.002 (0.004)
Ecuador 0.078 (0.131) 0.021 (0.098) 0.003 (0.029) 0.014 (0.028)
Egypt 0.053 (0.052) 0.021 (0.032) -0.008 (0.018) 0.007 (0.032)
El Salvador 0.094 (0.036) 0.005 (0.027) -0.009 (0.005) 0.051 (0.031)
Estonia 0.015 (0.017) 0.007 (0.019) -0.004 (0.007) 0.002 (0.005)
Ethiopia 0.008 (0.028) 0.019 (0.104) 0.001 (0.009) 0.008 (0.015)
Finland 0.013 (0.028) 0.015 (0.027) 0.003 (0.008) 0.000 (0.004)
France 0.016 (0.036) 0.014 (0.029) 0.009 (0.011) 0.004 (0.005)
Ghana 0.029 (0.057) 0.050 (0.084) -0.007 (0.009) 0.002 (0.010)
Greece 0.020 (0.039) 0.010 (0.028) 0.017 (0.016) 0.004 (0.006)
Guatemala 0.069 (0.042) 0.007 (0.029) -0.004 (0.008) 0.036 (0.027)
Germany 0.015 (0.034) 0.014 (0.032) 0.008 (0.015) 0.004 (0.006)
Honduras 0.084 (0.045) 0.010 (0.054) -0.008 (0.007) 0.043 (0.039)
Hungary 0.023 (0.029) 0.005 (0.016) -0.015 (0.012) 0.005 (0.004)
Iceland 0.076 (0.043) 0.016 (0.033) -0.024 (0.021) 0.013 (0.017)
India 0.042 (0.034) 0.039 (0.030) -0.004 (0.011) -0.011 (0.015)
Indonesia 0.086 (0.185) 0.034 (0.053) -0.003 (0.014) 0.001 (0.013)
Ireland 0.010 (0.026) 0.005 (0.014) 0.007 (0.007) 0.002 (0.003)
Israel 0.019 (0.029) 0.011 (0.021) -0.002 (0.011) 0.001 (0.005)
Italy 0.022 (0.037) 0.014 (0.028) 0.008 (0.014) 0.004 (0.006)
Japan 0.036 (0.035) 0.042 (0.033) -0.010 (0.013) -0.012 (0.014)
Jordan 0.073 (0.081) 0.059 (0.047) -0.012 (0.035) 0.022 (0.022)
Kazakhstan 0.010 (0.021) 0.007 (0.012) -0.002 (0.007) 0.002 (0.008)
Kenya 0.053 (0.069) 0.013 (0.040) 0.004 (0.042) 0.038 (0.053)
Latvia 0.013 (0.016) 0.006 (0.022) -0.003 (0.012) 0.002 (0.007)
Lebanon 0.062 (0.075) 0.012 (0.031) 0.001 (0.021) 0.018 (0.032)
Lithuania 0.025 (0.031) 0.009 (0.025) -0.008 (0.018) 0.003 (0.006)
Absolute market access Relative market access
Country name TTRI 2000
Malaysia 0.020 (0.043) 0.021 (0.025) 0.000 (0.008) 0.002 (0.008)
Mauritius 0.069 (0.042) 0.006 (0.030) -0.019 (0.015) 0.032 (0.038)
Mexico 0.008 (0.012) 0.004 (0.015) 0.015 (0.009) 0.011 (0.013)
Morocco 0.067 (0.032) 0.014 (0.031) -0.020 (0.013) 0.011 (0.015)
Netherlands 0.011 (0.027) 0.008 (0.021) 0.011 (0.011) 0.003 (0.004)
New Zealand 0.065 (0.068) 0.048 (0.049) -0.007 (0.016) -0.007 (0.021)
Nicaragua 0.066 (0.048) 0.005 (0.016) -0.005 (0.007) 0.039 (0.021)
Nigeria 0.008 (0.023) 0.008 (0.020) -0.001 (0.008) 0.000 (0.009)
Norway 0.020 (0.014) 0.009 (0.021) -0.009 (0.007) 0.000 (0.006)
Oman 0.007 (0.019) 0.013 (0.019) 0.000 (0.003) -0.001 (0.006)
Peru 0.035 (0.035) 0.007 (0.016) -0.003 (0.019) 0.008 (0.015)
Philippines 0.025 (0.024) 0.015 (0.019) -0.001 (0.008) 0.006 (0.025)
Poland 0.021 (0.031) 0.006 (0.022) -0.018 (0.012) 0.004 (0.004)
Portugal 0.009 (0.029) 0.010 (0.026) 0.022 (0.015) 0.007 (0.008)
Republic of Korea 0.045 (0.042) 0.046 (0.031) -0.011 (0.021) -0.011 (0.016)
Romania 0.023 (0.041) 0.008 (0.023) -0.017 (0.023) 0.009 (0.008)
Russian Federation 0.021 (0.023) 0.013 (0.017) -0.003 (0.005) -0.002 (0.005)
South Africa 0.028 (0.037) 0.023 (0.034) -0.003 (0.006) -0.002 (0.008)
Saudi Arabia 0.018 (0.024) 0.022 (0.023) 0.000 (0.006) 0.000 (0.009)
Senegal 0.034 (0.065) 0.030 (0.066) 0.019 (0.019) 0.019 (0.037)
Singapore 0.020 (0.022) 0.016 (0.023) 0.000 (0.007) 0.003 (0.009)
Slovakia 0.013 (0.036) 0.005 (0.030) 0.009 (0.021) 0.008 (0.012)
Slovenia 0.022 (0.025) 0.005 (0.020) -0.020 (0.015) 0.004 (0.003)
Spain 0.018 (0.047) 0.011 (0.027) 0.011 (0.019) 0.005 (0.006)
Sri Lanka 0.083 (0.044) 0.049 (0.055) -0.010 (0.019) 0.006 (0.025)
Sweden 0.016 (0.033) 0.011 (0.024) 0.003 (0.015) 0.002 (0.004)
Thailand 0.036 (0.040) 0.034 (0.038) -0.001 (0.010) 0.004 (0.022)
Trinidad and Tobago 0.008 (0.035) 0.003 (0.011) -0.001 (0.012) 0.008 (0.007)
Tunisia 0.044 (0.026) 0.008 (0.029) -0.016 (0.015) 0.016 (0.014)
Turkey 0.059 (0.049) 0.015 (0.035) -0.029 (0.021) 0.009 (0.014)
Uganda 0.015 (0.085) 0.018 (0.087) 0.004 (0.025) 0.010 (0.026)
United Kingdom 0.018 (0.041) 0.015 (0.031) 0.004 (0.011) 0.002 (0.006)
of Tanzania 0.093 (0.202) 0.058 (0.111) 0.005 (0.040) 0.004 (0.012)
United States 0.023 (0.032) 0.026 (0.038) 0.019 (0.043) 0.005 (0.022)
Uruguay 0.088 (0.060) 0.034 (0.049) -0.002 (0.016) 0.010 (0.030)
Republic of 0.032 (0.027) 0.002 (0.008) -0.001 (0.006) 0.008 (0.012)
Zambia 0.033 (0.080) 0.016 (0.028) 0.006 (0.023) 0.008 (0.012)
Simple average 0.035 (0.044) 0.020 (0.035) -0.002 (0.016) 0.007 (0.015)
Table 1 (continued)...
Figure 1. Distribution of direct and relative market access indices
Note: Direct market access index is expressed as TTRI; relative market access index is expressed as RPM.
Tariff restrictiveness varies considerably across countries. The variation is due to both
product composition and preferential access. Countries whose exports are largely concentrated on
energy, minerals and raw materials are those facing a lower TTRI. Similarly, countries that belong
to free trade areas generally face a lower TTRI. However, countries whose major exports are
agricultural products or countries that are not part of preferential trade agreements tend to have a
larger TTRI. In value terms, average export restrictiveness is not large. The 2007, TTRI was less
than 1 per cent for about 34 countries of the sample and above 4 per cent in only 9 countries. Due
to the widespread tariff liberalization during the period of analysis, the TTRI declined steadily
between 2000 and 2007. The simple average TTRI across the countries in our sample fell from
about 3.5 per cent in 2000 to about 2 per cent in 2007. This decline is also reflected in figure 1,
which reports the distribution of the simple and import-weighted average bilateral TTRI. The
import-weighted bilateral TTRI is generally much lower than the simple bilateral TTRI, suggesting
that trade is concentrated where tariff restrictions are lower.
Regarding the RPM, for 2007 it varies from about minus 1 per cent for Japan, China, the
Republic of Korea and India, to about plus 3 per cent or more for El Salvador, Guatemala,
Honduras, Kenya, Mauritius and Nicaragua. In general, low-income countries benefit most from
the existing structure of preferences as a substantial number of them benefits from preferential
access to the United States and European Union markets. However, high-income countries and
countries with limited participation in trade agreements are found to be those with a negative RPM.
As of 2007, the existing structure of preferences discriminates against imports from about 13
-.05 0 .05
Bil. RPM 2000 (simple) --- Mean = -0.0066
Bil. RPM 2007 (simple) --- Mean = -0.0037
-.05 0 .05
Bil. RPM 2000 (weighted) --- Mean = 0.0028
Bil. RPM 2007 (weighted) --- Mean = 0.0004
0 .02 .04 .06 .08 .1
Bil. TTRI 2000 (weighted) --- Mean = 0.0252
Bil. TTRI 2007 (weighted) --- Mean = 0.0217
0 .02 .04 .06 .08 .1
Bil. TTRI 2000 (simple) --- Mean = 0.0689
Bil. TTRI 2007 (simple) --- Mean = 0.0499
countries (negative RPM). Those countries face a tariff on their exports higher than that of their
competitors. The variance of the bilateral RPM depends both on export composition and on the
number and strength of existing PTAs. The Plurinational State of Bolivia, Kenya, Colombia, and
Bangladesh are the countries with the highest dispersion.
The RPM has increased for the majority of countries. The simple average across our
sample of 86 countries indicates that the RPM has grown from -0.2 per cent in 2000 to about 0.7
per cent in 2007. This increase is largely due to the proliferation of PTAs, which has produced an
increase in the number of trade relationships with a positive RPM. This point is more evident in
figure 1, which indicates that while the simple mean of bilateral RPM has increased, the weighted
mean has decreased during our period of analysis.16 The proliferation of PTAs has also reduced the
relative tariff advantage of pre-existing PTAs. Countries that enjoyed preferential access before the
year 2000 find their preferential margin eroded by any new PTA involving their trading partners.
In practice, the structure of relative preferences has moved from a situation where few bilateral
trade relationships enjoyed relatively large preferential access, to a situation where there is a higher
number of positive, but smaller relative preferential access trade relationships. In particular, most
of the countries where the RPM has decreased are found to be high-income countries. As high-
income countries were early adopters of PTAs, their preferential advantage was eroded by the
proliferation of PTAs from 2000–2007. In addition, relative preferences have declined for
countries that did not actively engage in bilateral agreements, such as China and India.
In the majority of cases, improvement in direct market access is accompanied by better
relative market access conditions and vice-versa. Figure 2 illustrates the correlation between
changes in bilateral RPM and TTRI between 2000 and 2007. Still, there are a number of cases
(24.5 per cent) where a tariff reduction is not accompanied by an amelioration of relative market
access. That is, some of the advantage provided by the improvement in direct market access
conditions is lost by the reduction in the RPM. Similarly, in a limited number of cases (8.4 per
cent), some of the amelioration of relative market access conditions is offset by an increase in trade
Figure 2. Correlation in the changes in the tariff trade restrictiveness index
and the relative preferential margin (2000–2007)
16 The simple average RPM is generally negative, as most bilateral trade relationships are not subject to
13.7 % 8.4 %
24.5 % 53.4 %
-.2 -.1 0 .1 .2
Change in RPM 2000-2007 (positive = higher advantage)
4.2 Econometric results
Estimates from the gravity model indicate that TTRI and RPM have a significant effect on
trade flows. This implies that the magnitude of trade flows depends not only on the tariffs faced,
but also on the relative advantage provided by the structure of preferences. That is, bilateral trade
flows are found to be larger the greater the relative advantage provided by the system of
Before discussing the results in detail, some interpretation of the two coefficients of
interest is in order. The TTRI and the RPM index are trade-weighted measures, implying that their
magnitude could change even in cases when tariffs are kept fixed.17 For example, the TTRI would
decline when a country's export basket shifts towards goods facing lower tariffs. Similarly, the
RPM will increase when exports shift towards goods which enjoy a larger relative advantage. In an
extreme case, assuming that tariffs are kept fixed – changes in TTRI and RPM are thus driven by a
shift in exports – a positive TTRI coefficient would imply that exports are shifting towards goods
facing lower tariffs. It can be argued that the country is reacting to previous changes in trade policy
by taking better advantage of pre-existing favourable market access conditions. In this regard, our
estimated coefficients capture the effect of actual and previous tariff changes on trade flows.18 To
ascertain whether the changes in TTRI and RPM are related to export composition or a change in
tariff, we calculate the TTRI and RPM with time-unvarying weights.19 The original TTRI and RPM
result highly correlated (>0.9) with time-unvarying TTRI and RPM, indicating that that most of the
changes in the TTRI and RPM are driven by changes in tariffs rather than by adjustments in the
export basket. Moreover, estimating the gravity model using TTRI and RPM constructed with
time-unvarying weights produced very similar results.
4.2.1 Core specifications
Table 2 reports the estimated coefficients for a series of specifications of the gravity
model. Columns (1) and (2) report results obtained with a naïve gravity specification. All
coefficients are in line with standard predictions. Specifications reported in columns (3) to (7)
include country-pair fixed effects; hence time invariant variables are dropped. In all of these
specifications, the coefficients of interest are significant and of the correct sign. Column (3)
estimated the coefficient according to the simple specification without the RPM term. In this
specification, the TTRI variable coefficient of -0.679 implies that bilateral trade flows are
estimated to decrease by 0.72 per cent for a one percentage point increase in the TTRI at its mean.
Specification (4) adds the RPM variables. The RPM coefficient indicates that for each percentage
point of increase in the RPM, bilateral trade is expected to expand by about 0.34 per cent. The
TTRI coefficient is not substantially affected by including the RPM variable.20 A decrease of one
percentage point in the TTRI at its mean value increases trade by about 0.61 per cent. Without
country-time fixed effects (jt and kt), that is, ignoring the time-varying multilateral price terms, the
TTRI and RPM coefficients increase significantly in absolute value indicating clearly the incidence
of an omitted-variable bias. Specification (5) is similar to (4) but scale the dependent variable by
the product of trade partners’ GDP, as suggested by standard theoretical derivation. Imposing
17 Note that this issue would not raise problems of reverse causality, as TTRI and RPM are aggregated across
18 In this paper we do not aim to disentangle these effects, we leave this to future research.
19 Weights are given by average export over time at the product level.
20 For robustness, we also estimated all specifications using the two policy variables, RPM and TTRI,
computed without import demand elasticities. Coefficients keep their significance and are similar in
unitary income elasticities restrictions does not influence the coefficients on the two variables of
We also estimate the gravity model in first differences. This provides an alternative
approach in order to account for a possible endogeneity bias caused by an omitted variable bias and
time-invariant unobserved heterogeneity. Specifications (6) and (7) correspond to (4) and (5)
respectively. Although the coefficients indicate a smaller effect of the TTRI and a larger effect of
the RPM indices, they maintain their significance and sign.
Table 2. Gravity model estimation
(1) (2) (3) (4) (5) (6) (7)
RPM 0.339** 0.350**
Ln (1 + TTRI) -2.348*** -0.679*** -0.609*** -0.618***
(0.1391) (0.0912) (0.0967) (0.0975)
Ln distance -0.961*** -0.934***
Language 0.986*** 1.007***
Colonial ties 0.201*** 0.183***
Border 0.791*** 0.818***
Ln GDP j 0.992*** 0.978***
Ln GDP k 1.194*** 1.196***
Diff_RPM 0.630** 0.644***
Diff_TTRI -0.478** -0.472**
Importer-year No No Yes Yes Yes Yes Yes
Exporter-year No No Yes Yes Yes Yes Yes
Country-pair No No Yes Yes Yes Yes Yes
Weighted No No No No Yes No Yes
Observations 50031 50031 50031 50031 50031 50031 50031
Adjusted R2 0.716 0.718 0.956 0.956 0.885 0.087 0.085
Within R2 - - 0.303 0.303 0.303 0.303 0.303
Notes: Dependent variable – natural log of export
Robust standard errors in parentheses (* p < 0.1, ** p < 0.05, *** p < 0.01)
4.2.2 Robustness checks
Table 3 reports the estimated coefficients for a series of robustness-check specifications.
The estimation of equation (5) is shown in column (8). Signs and significance of parameters of
interest are maintained. However, the interpretation of the RPM-ratio coefficient in terms of trade
effects is not straightforward. Assuming that changes in the RPM ration are due to changes in the
TTRI, a one percentage point decrease in the TTRI taken at its mean value translates into a 0.63
per cent change in bilateral trade via the RPM-ratio impact. As far as the direct impact of TTRI is
concerned, the corresponding figure is 0.65 per cent, again computed at the mean value of TTRI
and assuming that only TTRI varies. The trade-weighted version of (8) is presented in column (9).
As in the core specification, results are not significantly affected. Columns (10) and (11) report
results obtained by estimating equation (5) in first differences. Significance and signs are
maintained, although the coefficients indicate a smaller effect of the TTRI and a larger effect of the
RPM indices. Overall specifications with ln (RPM ratio) generate comparable results in terms of
sign and significance of the coefficients of interest and in terms of goodness of fit.
Table 3. Gravity model estimation: robustness checks
(8) (9) (10) (11) (8') (10')
Ln(RPMratio) 0.598*** 0.581***
Ln (1+TTRI) -0.555*** -0.598*** -0.581***
(0.1032) (0.0970) (0.1042)
Ln GDPprod 0.0337 0.0282
Diff_lnRPMratio 0.964*** 0.944***
Diff_TTRI -0.425** -0.458** -0.443**
(0.2027) (0.2021) (0.2051)
Diff_lnGDPprod 0.0386 0.0349
Importer-year Yes Yes Yes Yes Yes Yes
Exporter-year Yes Yes Yes Yes Yes Yes
Country pair Yes Yes - - Yes -
Weighted No Yes No Yes No No
Observations 50031 50031 42140 42140 50031 42140
Adjusted R2 0.956 0.885 0.087 0.085 0.956 0.087
Within R2 0.303 0.308 - - 0.303 -
Notes: Dependent variable – natural log of export
Robust standard errors in parentheses (* p < 0.1, ** p < 0.05, *** p < 0.01)
Finally, columns (8') and (10') refer to specifications where ⎟⎟⎠
ln has been
approximated by ( )jktwjkt TTRIT − . The approximation is satisfactory for small values of both wjktT
and jktTTRI . A direct comparison can be made with results obtained in columns (4) and (6) of
table 2. As coefficients keep the same sign and significance level, and vary only marginally in
absolute value, the inclusion of our theoretically predicted control variable does not entail the
validity of core results.
As a further test, we are interested in knowing whether the RPM variable of equation (2)
provides a better fit in explaining bilateral trade than the standard measure of preferential margin
(that is, not considering preferential access granted to third country exporters but only the MFN
rate).21 Results are reported in table 4, where the specifications reflect those of table 2. All
coefficients for the preferences margin measure lose their significance. The TTRI coefficients are
substantially equal to those obtained by using the RPM index. The lack of significance may reflect
two things. First it may reflect a lower degree of variation across country-time of the standard
preferential margin. But it may also reflect the fact that the preferential margin (PM) is much less
bilateral that TTRI, and its impact is likely to be absorbed by country-and-time dummies. In both
cases, if the PM variable were affecting bilateral trade significantly, this effect would appear with
the removal of country-and-time dummies. Results obtained for the non-weighted specification in
levels are reported in column (11). The estimated coefficient increases in absolute value but does
not remain significant at any reasonable level. Similar remarks apply to results obtained with other
core specifications. As a result, the issue may not be solely statistical. From an economic point of
view the latter result could indicate that the commonly used measure of preferential margin does
not properly reflect the advantage provided by the system of preferences.
A recurring problem with gravity estimation is the presence of zero trade. As the gravity
model is generally estimated in a log-normal specification, it will discard observations where there
is no trade. Recent procedures to take into account zero trade flows are the Poisson estimation
(Santos Silva and Tenreyro, 2006), or a two-stage estimation procedure (Helpman et al., 2008).
Our preferred specification does not check for the presence of zeros for a number of reasons. First,
the incidence of zero trade observations remains relatively limited in our sample. All considered,
the matrix of bilateral trade has about 14 per cent of zero-trade observations. Second, in our
sample, inexistent bilateral trade is generally confined to cases of small and distant countries
(Frankel, 1997) or between countries lacking cultural and historical links (Rauch, 1999). In our
specification, these are captured by country-pair fixed effects.22 Most importantly, our main
variables of interest, the RPM indices and TTRI, utilize trade values at the product level as weight.
Thus, these variables cannot be properly computed when bilateral trade is zero.
21 The preferential margin is given by
22 The country-pair composition of zero trade varies only marginally from year to year. The contribution of
the new bilateral trade flows to the variation in total trade is on average equal to 0.2 per cent.
Table 4. Gravity model results with standard measure of preferential margin
(4') (5') (6') (7') (11)
Pref Margin 0.0251 0.0121 0.145
(0.1089) (0.1104) (0.1252)
Ln (1+TTRI) -0.744*** -0.754*** -1.756***
(0.0934) (0.0941) (0.1021)
Diff_ Pref Margin -0.0277 -0.0301
Diff_TTRI -0.665*** -0.658***
Importer-year Yes Yes Yes Yes No
Exporter-year Yes Yes Yes Yes No
Country-pair Yes Yes - - Yes
Weighted No Yes No Yes No
Observations 50031 50031 42140 42140 50031
Adjusted R2 0.956 0.885 0.086 0.085 0.939
Within R2 0.303 0.308 - - 0.302
Notes: Dependent variable – natural log of export
Robust standard errors in parentheses (* p < 0.1, ** p < 0.05, *** p < 0.01)
4.3 Impact of preferential access on trade flows
In this chapter we estimate the effect of the system of preferences on trade flows. As
mentioned above, results are based on the coefficients obtained in specification (4). Nevertheless,
similar qualitative and quantitative conclusions would be reached using coefficients obtained in
specification (8). The impact on exports for every country is calculated as:
( ) ( )∑ ++=
jkjkj RPMΔβTTRIlnΔβexplnΔ 21 1 (6)
We report two sets of results. The first set of results is based on changes in the differences
between the first to the last year of the analysis. This provides the estimated percentage change in
trade since 2000 due to the change in market access conditions between 2000 and 2007. In a
second set of results, equation (6) is calculated relative to a hypothetical no-preference scenario
where all bilateral flows are subject to MFN rates.23 This provides an indication of the overall
effect of preferential access on trade.
23 That is: ( ) )TTRIln()MFNln(TTRIlnΔ jkjjk +−+=+ 111 and jkjk RPMRPMΔ = .
Results for the effect of the change in market access conditions on bilateral trade are
provided in table 5. For the average country, trade gains from the current system of preferences are
estimated to be about 1.6 per cent with respect to market access conditions in 2000, and about 2.3
per cent with respect to a hypothetical scenario based on the prevailing 2007 MFN rate. In both
cases, most of the effects originated from an amelioration in direct market access conditions.
Improvements in relative market access conditions are less significant and account for about a 0.3
per cent increase in trade. These small average effects can be explained by the marginal impact of
tariff liberalization on the larger trade flows during the period of analysis because a large share of
trade had previously been liberalized on a preferential basis or under MFN tariffs (about 40 per
cent of world trade is free under MFN tariffs).
Table 5. Effects of changes in market access conditions, by country
Percentage change in trade,
2007 versus 2000
Percentage change in trade,
2007 compared with MFN
Country name Total TTRI RPM Total TTRI RPM
Algeria 0.34 0.28 0.07 0.39 0.37 0.02
Argentina 3.64 2.69 0.96 4.09 3.14 0.95
Australia 0.56 0.58 -0.02 0.27 0.32 -0.05
Austria -0.01 0.23 -0.24 2.21 2.10 0.11
Azerbaijan 0.26 0.10 0.16 0.94 0.82 0.12
Bangladesh -0.25 0.07 -0.32 4.02 3.46 0.56
Belgium 2.35 1.21 1.14 2.45 2.33 0.12
Benin -0.06 0.31 -0.38 2.04 1.90 0.14
Bolivia, Plurinational State of 2.29 1.75 0.54 6.80 5.15 1.65
Brazil 1.18 0.87 0.31 2.34 2.10 0.24
Bulgaria 1.17 0.56 0.61 2.20 2.09 0.10
Cameroon -0.12 -0.31 0.18 1.22 1.11 0.11
Canada -0.05 0.08 -0.13 1.84 1.64 0.21
Chile 0.98 0.92 0.06 1.92 1.67 0.25
China 0.21 0.34 -0.13 -0.12 0.17 -0.29
China, Hong Kong 1.21 0.94 0.26 0.48 0.63 -0.15
China, Taiwan Province of 0.64 0.68 -0.03 -0.20 0.06 -0.25
Colombia 3.21 2.43 0.78 3.62 2.92 0.70
Costa Rica 2.48 1.97 0.51 3.78 3.37 0.41
Côte d'Ivoire 1.56 1.20 0.37 1.94 1.78 0.16
Croatia 1.35 0.93 0.42 2.16 1.96 0.20
Czech Republic 2.12 1.22 0.90 2.48 2.33 0.15
Denmark 0.26 0.40 -0.14 1.76 1.68 0.08
Ecuador 3.88 3.47 0.41 4.88 4.37 0.51
Egypt 2.67 2.16 0.51 2.46 2.22 0.24
El Salvador 7.67 5.54 2.13 7.41 5.58 1.83
Estonia 0.93 0.69 0.24 1.74 1.63 0.10
Ethiopia 0.02 -0.04 0.06 0.92 0.84 0.08
Finland -0.06 0.02 -0.08 0.87 0.85 0.02
France -0.04 0.15 -0.19 2.04 1.91 0.13
Ghana 0.70 0.46 0.24 1.60 1.59 0.01
Greece 0.08 0.57 -0.49 2.96 2.84 0.12
Guatemala 5.55 4.02 1.53 6.26 4.88 1.38
Germany 0.03 0.18 -0.15 2.10 1.96 0.14
Honduras 7.02 5.06 1.96 6.87 5.18 1.69
Hungary 1.92 1.24 0.69 2.05 1.89 0.17
Iceland 5.26 4.05 1.21 3.24 2.87 0.37
Percentage change in trade,
2007 versus 2000
Percentage change in trade,
2007 compared with MFN
Country name Total TTRI RPM Total TTRI RPM
India 0.25 0.42 -0.17 0.10 0.43 -0.33
Indonesia 2.77 2.69 0.09 0.76 0.78 -0.03
Ireland 0.15 0.33 -0.18 1.02 0.95 0.07
Italy 0.23 0.39 -0.16 2.00 1.87 0.13
Israel 0.76 0.66 0.10 1.00 0.95 0.05
Japan 0.02 0.09 -0.07 -0.30 0.10 -0.41
Jordan 2.78 2.03 0.75 2.00 1.68 0.32
Kazakhstan 0.38 0.12 0.26 1.25 1.06 0.19
Kenya 3.27 2.41 0.86 4.59 3.58 1.00
Latvia 0.65 0.48 0.17 2.80 2.72 0.08
Lebanon 3.12 2.63 0.49 3.04 2.50 0.54
Lithuania 1.42 1.01 0.41 2.77 2.63 0.14
Malaysia 0.39 0.35 0.04 0.47 0.43 0.04
Mauritius 5.57 3.83 1.74 7.53 6.48 1.06
Mexico 0.18 0.33 -0.15 2.40 2.04 0.37
Morocco 4.38 3.27 1.11 4.13 3.71 0.42
Netherlands -0.18 0.10 -0.28 2.28 2.18 0.10
New Zealand 0.90 0.95 -0.06 0.23 0.52 -0.29
Nicaragua 5.34 3.90 1.44 5.92 4.65 1.27
Nigeria 0.33 0.32 0.01 1.54 1.57 -0.03
Norway 0.95 0.65 0.31 0.77 0.78 -0.01
Oman -0.20 -0.18 -0.02 0.09 0.12 -0.03
Peru 2.19 1.74 0.45 2.89 2.55 0.35
Philippines 0.74 0.56 0.18 1.04 0.88 0.16
Poland 1.68 0.92 0.77 3.07 2.92 0.15
Portugal -0.44 0.04 -0.48 3.42 3.13 0.29
Republic of Korea 0.36 0.37 0.00 -0.31 0.09 -0.39
Romania 1.94 1.02 0.92 3.07 2.75 0.32
Russian Federation 0.47 0.44 0.03 0.61 0.67 -0.06
South Africa 0.51 0.48 0.03 0.62 0.68 -0.06
Saudi Arabia 0.02 0.01 0.01 0.24 0.24 0.00
Senegal 0.47 0.65 -0.18 4.42 3.95 0.47
Singapore 0.56 0.45 0.11 0.60 0.50 0.10
Slovakia 0.52 0.60 -0.07 2.89 2.63 0.26
Slovenia 2.02 1.16 0.86 2.49 2.34 0.15
Spain 0.19 0.41 -0.22 3.23 3.05 0.18
Sri Lanka 2.03 1.66 0.38 1.71 1.70 0.01
Sweden 0.26 0.32 -0.06 1.36 1.33 0.04
Thailand 0.73 0.61 0.12 1.22 1.14 0.08
Trinidad and Tobago 0.66 0.32 0.34 1.71 1.42 0.29
Tunisia 3.45 2.29 1.16 3.87 3.28 0.60
Turkey 4.00 2.65 1.36 3.47 3.14 0.33
Uganda 0.27 0.14 0.13 1.99 1.72 0.27
United Kingdom 0.02 0.09 -0.07 1.43 1.36 0.07
United Republic of Tanzania 1.60 1.62 -0.02 2.03 1.89 0.15
United States -0.35 0.05 -0.40 2.16 1.88 0.28
Uruguay 4.87 4.07 0.79 4.54 3.83 0.71
Republic of 2.32 1.97 0.35 3.46 3.13 0.33
Zambia 0.62 0.68 -0.06 0.88 0.74 0.14
Simple average 1.5 1.2 0.3 2.3 2.1 0.3
Table 5 (continued)...
Although the results are on average small, they show substantial variance. Changes in
market access conditions between 2000 and 2007 have produced an increase in trade of 5 per cent
or more for six countries: El Salvador, Honduras, Mauritius, Guatemala, Nicaragua and Iceland. In
contrast, however, because of the deterioration in relative market access conditions, about 11
countries have seen their trade negatively affected by changes in market access conditions between
2000 and 2007. Exports from these countries have lost competitiveness to countries benefiting
from better preferential access. In general, the countries where gains have been larger are those
which were the most active in implementing PTAs during the period of analysis. Countries that
were already parties to PTAs have gained less or lost, as a portion of any gains from the overall
trend in tariff reduction is somewhat offset by preference erosion (a negative RPM). With regard to
changes in the MFN scenario, the results are qualitatively similar. Most countries consider that
improvement in direct market matters the most. However, for China, the Republic of Korea, Japan
and Taiwan Province of China the change in relative market access conditions produces a
relatively larger negative impact.
The purpose of this paper is twofold. The first objective is to provide two indices of market
access conditions that take into account the complex structure of tariff preferences. The second
objective is to determine how changes in market access conditions impact international trade. In
particular, we explore whether the effect of change in tariff is stronger the larger the advantage
provided by preferential access over competitors.
In relation to the title of this paper, the overall results indicate that preferential market
access is valuable in terms of export performance. Its value depends not only on the direct
advantage provided by preferential access but also on the advantage provided with respect to other
competitors. In numbers, a decline of one percentage point in terms of the overall tariff faced by
exports is reflected in an average increase by almost 0.7 per cent in bilateral trade. Similarly, for
every percentage point increase in relative market access, trade increases by slightly more than 0.3
Although preferential access matters, its effect on trade is not large on average. The reason
is that changes both in direct and relative preferential access, as measured by the two indices, are
often small. Given the already low MFN tariffs and the large number of PTAs, the improvement in
preferential access is often marginal in terms of tariff advantages, and consequently has marginal
effects on trade. Still, looking beyond averages, the results show a large variance. This indicates
that for some countries changes in market access conditions have produced substantial effects on
trade. More importantly, we do not investigate changes at the product level where improvements in
market access conditions are likely to produce much larger effects.
Regarding the two indices of market access conditions, the results show that overall tariff
restrictiveness is on average not large and has constantly declined during the period of analysis.
Similarly, the RPM is also small in most cases; however, it has increased because the structure of
preferences shifted from a situation in which few bilateral trade relationships benefited from
relatively large preferential access to a situation in which a higher number of bilateral trade
relationships benefited from smaller relative preferential access. While the proliferation of PTAs
has led to an increase in the number of bilateral trade relationships with a positive RPM, it has also
eroded some of the tariff advantages of pre-existing PTAs.
Anderson JE and van Wincoop E (2003). Gravity with gravitas: a solution to the border puzzle.
American Economic Review. 93 (1): 70–192.
Baier SL and Bergstrand JH (2009). Estimating the effects of free trade agreements on
international trade flows using matching econometrics. Journal of International
Economics. 77(1): 63-76.
Baier SL and Bergstrand JH (2007). Do free trade agreements actually increase members'
international trade? Journal of International Economics. 71(1): 72-95.
Calvo-Pardo H et al (2009). The ASEAN free trade agreement: impact on trade flows and external
trade barriers. In: Barro R and Jong-Wha L, eds. Costs and Benefits of Regional Economic
Integration. Oxford University Press. Oxford.
Carrère C et al. (2008). Disentangling market access effects for ASEAN members under an
ASEAN-EU FTA. CEPR Discussion Papers No. 6762.
Carrère C (2006). Revisiting the effects of regional trade agreements on trade flows with proper
specification of the gravity model. European Economic Review. 50(2): 223-247.
Clausing K (2001). Trade creation and trade diversion in the Canada-United States Free Trade
Agreement. Canadian Journal of Economics. 34: 678-96.
Feenstra R (1995). Estimating the effects of trade policy. In: Grossman G and Rogoff K, eds.
Handbook of International Economics, volume 3. Elsevier. Amsterdam.
Frankel JA (1997). Regional Trading Blocs. Institute for International Economics. Washington,
Ghosh S and Yamarik S (2004). Are regional trading arrangements trade creating?: An application
of extreme bounds analysis. Journal of International Economics. 63(2): 369-395.
Helpman E et al. (2008). Estimating trade flows: Trading partners and trading volumes. The
Quarterly Journal of Economics. 123(2): 441-487.
Hoekman B et al. (2009). Quantifying the value of preferences and potential erosion losses. In:
Hoekman B et al., eds. Trade Preference Erosion Measurement and Policy Response.
Hoekman B and Nicita A (2008). Trade policy, trade costs and developing country trade. Policy
Research Working Paper Series 4797. The World Bank. Washington, D.C.
Kee HL et al. (2009). Estimating trade restrictiveness indices. Economic Journal. 119(534): 172-
Kee HL et al. (2008). Import demand elasticities and trade distortions. The Review of Economics
and Statistics. 90(4): 666-682.
Lee JW and Shin K (2006). Does regionalism lead to more global trade integration in East Asia?
North American Journal of Economics and Finance. 17: 283-301.
Linder SB (1961). An Essay on Trade and Transformation. Wiley and Sons. New York.
Linnemann, H (1966). An Econometric Study of International Trade Flows. North-Holland.
Low P et al. (2009). Multilateral solutions to the erosion of non-reciprocal preferences in NAMA.
In: Hoekman B et al., eds. Trade Preference Erosion Measurement and Policy Response.
Magee C (2008). New measures of trade creation and trade diversion. Journal of International
Economics. 75(2): 340-62.
Rauch J (1999). Networks versus markets in international trade. Journal of International
Economics. 48(1): 7-35.
Robertson R and Estevadeordal A (2009). Asymmetric tariffs and multilateral resistance.
Cuadernos de Economía (Latin American Journal of Economics). 46: 3-31.
Santos Silva J and Tenreyro S (2006). The log of gravity. The Review of Economics and Statistics.
UNCTAD Study Series on
POLICY ISSUES IN INTERNATIONAL TRADE
No. 1 Erich Supper, Is there effectively a level playing field for developing country
exports?, 2001, 138 p. Sales No. E.00.II.D.22.
No. 2 Arvind Panagariya, E-commerce, WTO and developing countries, 2000, 24 p. Sales
No. 3 Joseph Francois, Assessing the results of general equilibrium studies of multilateral
trade negotiations, 2000, 26 p. Sales No. E.00.II.D.24.
No. 4 John Whalley, What can the developing countries infer from the Uruguay Round
models for future negotiations?, 2000, 29 p. Sales No. E.00.II.D.25.
No. 5 Susan Teltscher, Tariffs, taxes and electronic commerce: Revenue implications for
developing countries, 2000, 57 p. Sales No. E.00.II.D.36.
No. 6 Bijit Bora, Peter J. Lloyd, Mari Pangestu, Industrial policy and the WTO, 2000, 47 p.
Sales No. E.00.II.D.26.
No. 7 Emilio J. Medina-Smith, Is the export-led growth hypothesis valid for developing
countries? A case study of Costa Rica, 2001, 49 p. Sales No. E.01.II.D.8.
No. 8 Christopher Findlay, Service sector reform and development strategies: Issues and
research priorities, 2001, 24 p. Sales No. E.01.II.D.7.
No. 9 Inge Nora Neufeld, Anti-dumping and countervailing procedures – Use or abuse?
Implications for developing countries, 2001, 33 p. Sales No. E.01.II.D.6.
No. 10 Robert Scollay, Regional trade agreements and developing countries: The case of
the Pacific Islands’ proposed free trade agreement, 2001, 45 p. Sales No.
No. 11 Robert Scollay and John Gilbert, An integrated approach to agricultural trade and
development issues: Exploring the welfare and distribution issues, 2001, 43 p. Sales
No. 12 Marc Bacchetta and Bijit Bora, Post-Uruguay round market access barriers for
industrial products, 2001, 50 p. Sales No. E.01.II.D.23.
No. 13 Bijit Bora and Inge Nora Neufeld, Tariffs and the East Asian financial crisis, 2001,
30 p. Sales No. E.01.II.D.27.
No. 14 Bijit Bora, Lucian Cernat, Alessandro Turrini, Duty and quota-free access for LDCs:
Further evidence from CGE modelling, 2002, 130 p. Sales No. E.01.II.D.22.
No. 15 Bijit Bora, John Gilbert, Robert Scollay, Assessing regional trading arrangements in
the Asia-Pacific, 2001, 29 p. Sales No. E.01.II.D.21.
No. 16 Lucian Cernat, Assessing regional trade arrangements: Are South-South RTAs more
trade diverting?, 2001, 24 p. Sales No. E.01.II.D.32.
No. 17 Bijit Bora, Trade related investment measures and the WTO: 1995-2001, 2002.
No. 18 Bijit Bora, Aki Kuwahara, Sam Laird, Quantification of non-tariff measures, 2002,
42 p. Sales No. E.02.II.D.8.
No. 19 Greg McGuire, Trade in services – Market access opportunities and the benefits of
liberalization for developing economies, 2002, 45 p. Sales No. E.02.II.D.9.
No. 20 Alessandro Turrini, International trade and labour market performance: Major
findings and open questions, 2002, 30 p. Sales No. E.02.II.D.10.
No. 21 Lucian Cernat, Assessing south-south regional integration: Same issues, many
metrics, 2003, 32 p. Sales No. E.02.II.D.11.
No. 22 Kym Anderson, Agriculture, trade reform and poverty reduction: Implications for
Sub-Saharan Africa, 2004, 30 p. Sales No. E.04.II.D.5.
No. 23 Ralf Peters and David Vanzetti, Shifting sands: Searching for a compromise in the
WTO negotiations on agriculture, 2004, 46 p. Sales No. E.04.II.D.4.
No. 24 Ralf Peters and David Vanzetti, User manual and handbook on Agricultural Trade
Policy Simulation Model (ATPSM), 2004, 45 p. Sales No. E.04.II.D.3.
No. 25 Khalil Rahman, Crawling out of snake pit: Special and differential treatment and
post-Cancun imperatives, 2004.
No. 26 Marco Fugazza, Export performance and its determinants: Supply and demand
constraints, 2004, 57 p. Sales No. E.04.II.D.20.
No. 27 Luis Abugattas, Swimming in the spaghetti bowl: Challenges for developing
countries under the “New Regionalism”, 2004, 30 p. Sales No. E.04.II.D.38.
No. 28 David Vanzetti, Greg McGuire and Prabowo, Trade policy at the crossroads – The
Indonesian story, 2005, 40 p. Sales No. E.04.II.D.40.
No. 29 Simonetta Zarrilli, International trade in GMOs and GM products: National and
multilateral legal frameworks, 2005, 57 p. Sales No. E.04.II.D.41.
No. 30 Sam Laird, David Vanzetti and Santiago Fernández de Córdoba, Smoke and mirrors:
Making sense of the WTO industrial tariff negotiations, 2006, Sales No.
No. 31 David Vanzetti, Santiago Fernandez de Córdoba and Veronica Chau, Banana split:
How EU policies divide global producers, 2005, 27 p. Sales No. E.05.II.D.17.
No. 32 Ralf Peters, Roadblock to reform: The persistence of agricultural export subsidies,
2006, 43 p. Sales No. E.05.II.D.18.
No. 33 Marco Fugazza and David Vanzetti, A South-South survival strategy: The potential
for trade among developing countries, 2006, 25 p.
No. 34 Andrew Cornford, The global implementation of Basel II: Prospects and outstanding
problems, 2006, 30 p.
No. 35 Lakshmi Puri, IBSA: An emerging trinity in the new geography of international
trade, 2007, 50 p.
No. 36 Craig VanGrasstek, The challenges of trade policymaking: Analysis, communication
and representation, 2008, 45 p.
No. 37 Sudip Ranjan Basu, A new way to link development to institutions, policies and
geography, 2008, 50 p.
No. 38 Marco Fugazza and Jean-Christophe Maur, Non-tariff barriers in computable general
equilibrium modelling, 2008, 25 p.
No. 39 Alberto Portugal-Perez, The costs of rules of origin in apparel: African preferential
exports to the United States and the European Union, 2008, 35 p.
No. 40 Bailey Klinger, Is South-South trade a testing ground for structural
transformation?, 2009, 30 p.
No. 41 Sudip Ranjan Basu, Victor Ognivtsev and Miho Shirotori, Building trade-relating
institutions and WTO accession, 2009, 50 p.
No. 42 Sudip Ranjan Basu and Monica Das, Institution and development revisited: A
nonparametric approach, 2010, 26 p.
No. 43 Marco Fugazza and Norbert Fiess, Trade liberalization and informality: New stylized
facts, 2010, 45 p.
No. 44 Miho Shirotori, Bolormaa Tumurchudur and Olivier Cadot, Revealed factor intensity
indices at the product level, 2010, 55 p.
No. 45 Marco Fugazza and Patrick Conway, The impact of removal of ATC Quotas on
international trade in textiles and apparel, 2010, 50 p.
No. 46 Marco Fugazza and Ana Cristina Molina, On the determinants of exports survival,
2011, 40 p.
No. 47 Alessandro Nicita, Measuring the relative strength of preferential market access,
2011, 30 p.
No. 48 Sudip Ranjan Basu and Monica Das, Export structure and economic performance in
developing countries: Evidence from nonparametric methodology, 2011, 58 p.
No. 49 Alessandro Nicita and Bolormaa Tumurchudur-Klok, New and traditional trade flows
and the economic crisis, 2011, 22 p.
No. 50 Marco Fugazza and Alessandro Nicita, On the importance of market access for trade,
2011, 35 p.
Copies of UNCTAD Study series on Policy Issues in International Trade and Commodities may be
obtained from the Publications Assistant, Trade Analysis Branch (TAB), Division on International
Trade in Goods and Services and Commodities (DITC), United Nations Conference on Trade and
Development, Palais des Nations, CH-1211 Geneva 10, Switzerland (Tel: +41 22 917 4644).
These studies are accessible on the website at http://www.unctad.org/tab.
Since 1999, the Trade Analysis Branch of the Division on International Trade in Goods and
Services, and Commodities of UNCTAD has been carrying out policy-oriented analytical work
aimed at improving the understanding of current and emerging issues in international trade and
development. In order to improve the quality of the work of the Branch, it would be useful to
receive the views of readers on this and other similar publications. It would therefore be greatly
appreciated if you could complete the following questionnaire and return to:
Trade Analysis Branch, DITC
United Nations Conference on Trade and Development
Palais des Nations, CH-1211 Geneva 10, Switzerland
(Fax: +41 22 917 0044; E-mail: email@example.com)
1. Name and address of respondent (optional):
2. Which of the following describes your area of work?
Government Public enterprise
Private enterprise institution Academic or research
International organization Media
Not-for-profi t organization Other (specify) _________________
3. In which country do you work? _________________________________________
4. Did you fi nd this publication Very useful Of some use Litt le use
to your work?
5. What is your assessment of the contents of this publication?
Excellent Good Adequate Poor
6. Other comments:
UNCTAD Study series on
POLICY ISSUES IN INTERNATIONAL TRADE
(Study series no. 50: On the importance of market access for trade)