Cambridge Journal of Regions, Economy and Society 2011, 4, 253267

doi:10.1093/cjres/rsr002 Advance Access publication 21 March 2011

Economic diplomacy, trade and developing countries

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Mina Yakopa and Peter A. G. van Bergeijkb

a

Faculty of Economics and Business, University of Amsterdam, Roetersstraat 11, 1018 WB Amsterdam, The Netherlands, m.yakop@student.uva.nl b The International Institute of Social Studies of Erasmus University, Kortenaerkade 12, 2518 AX The Hague, The Netherlands, bergeijk@iss.nl
Received on May 18, 2010; accepted on January 12, 2011

This article analyses the impact of economic diplomacy on the geography of international trade for the year 2006 and 63 importing and exporting countries. We demonstrate that diplomatic representation via embassies and consulates is not a relevant trade-enhancing factor for trade within the OECD, but that it is signicant in bilateral trade relationships of developing countries. We discuss implications for developing countries especially in view of SouthSouth trade. Keywords: gravity model, development, southsouth trade, economic diplomacy JEL Classications: F19, F55, F59, O19

Introduction
Ever since the rst application of the gravity model of international trade by Tinbergen (1962), empirical trade analysts have acknowledged the need to take political factors into account when they were trying to explain the geography of international trade. In a data set that related to 1959, about 70% of world trade and 42 countries (both developing nations and major OECD countries), Jan Tinbergen estimated trade-stimulating preferential treatment to yield colonial or ex-colonial multipliers of about 10, indicating that (ex-)colonial ties gave rise to 10 times the usual trade volume. The measurement of political factors in this seminal study, however, was rather crude because it only used a binary 1-or-0 variable.1 Recent empirical trade analyses link trade and trade-related diplomatic variables clearer and more unambiguously.

The focus is on the infrastructure of embassies and consulates (Rose, 2007), changes in this infrastructure (Afman and Maurel, 2010) and ofcial state visits supported by this infrastructure (Nitsch, 2007). Embassies and consulates are of interest for two distinct reasons. Firstly, good and stable political relationsby building on mutual trust provide the rst best instrument to reduce the risk of future distortions and trade disruptions (van Bergeijk, 2010; van Marrewijk and van Bergeijk, 1993). Secondly, embassies and consulates help to generate knowledge about (future) opportunities for trade increasing the stock of knowledge about foreign markets. This reduces the costs of exporting to and investing in these markets. The empirical multi-country investigations consistently nd signicant and empirically relevant indications for trade creation by means of economic diplomacy.

The Author 2011. Published by Oxford University Press on behalf of the Cambridge Political Economy Society. All rights reserved. For permissions, please email: journals.permissions@oup.com

Yakop and van Bergeijk Afman and Maurel (2010), for example, estimate that the opening of an embassy in an emerging market in Eastern Europe is equivalent to an ad valorem tariff reduction of 212 percentage points. Most studies follow Roses framework using the same or a comparable trade data set. The number of destinations in these studies is large (150220 countries) and coverage thus appears to be sufcient to enable useful generalization with respect to destinations, but the data sets are actually skewed.2 On the one hand, the 150200 import destinations imply large cross-country heterogeneity that may lead to biased and or inefcient results. This may be especially true because a very large share of these import destinations consists of very small economies, so that the results may have been inuenced by economically insignicant markets. On this account, the number of importing countries needs to be reduced. On the other hand, the sample of countries from which trade ows originate is also too restricted as the available studies focus on the exports of a select group of countries, namely the larger OECD countries and the BRIIC countries (Afman and Maurel, 2010; Rose, 2007; Segura-Cayela and Vilarrubia, 2008). Sometimes the focus is even narrower: Gil-Pareja et al. (2007), for example, investigate tourism originating in the G7 and the studies on state visits merely cover the USA, Germany, France (Nitsch, 2007) and Canada (Head and Ries, 2006). Thus, the available evidence is limited in terms of the number of exporting countries covered and, moreover, these studies offer little or no guidance on the question whether economic diplomacy can be useful for developing countries since these countries are not included as exporting nations. This article contributes to the literature in several ways. We add 22 low-income countries (LIC) and middle-income countries achieving a more balanced data set by increasing (reducing) the number of exporting (importing) countries to 63 (Table 1 compares the countries studied by Rose and the countries used in our analysis and Appendix Table A1 lists the countries). We show that the effects of economic diplomacy differ between different country groups according to different income levels, a point that so 254
Table 1. Comparison of the sample of Rose (2007) and the sample in this study. Rose (2007) No low-income countries 1 lower middle-income countries 2 upper middle-income countries 5 high-income countries 14 Total 22 % 5 9 23 64 100 This study No 8 14 14 27 63 % 13 22 22 43 100

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

far has not been recognized in the literature. This enables us to show that the general results in Rose (2007) need further classication regarding the level of development of both the exporting and the importing economies. In particular, Roses results do not hold for bilateral trade between industrialized countries. Our methodology also allows us to investigate the potential utility of economic diplomacy in the context of SouthSouth trade, an issue that has not been investigated so far. Finally, we do not only deal with the impact of diplomacy on exports (as all other studies happen to do): the symmetry of our 63 3 63 country matrix (that contrasts to the Rose 22 3 200, the Segura-Cayela and Vilarrubia 21 3 163, the GilPareja 7 3 156 and the Nitsch 3 3 200 matrices) allows us to asses the import facilitating role of embassies and consulates on the same footing as their export promoting function. The remainder of this article is organized as follows. The next section discusses the economic rationale for embassies and consulates as a means to provide market access and to reduce market failures. The market failures that can be repaired by means of economic diplomacy are especially relevant for developing countries. Then we introduce our empirical tool (the gravity model), and discuss the construction of our data set and present the general empirical results. Next we zoom in on the question of how the level of economic development inuences the impact of embassies and consulates on trade. The nal section concludes and discusses the policy relevance of our ndings.

Why does economic diplomacy matter?

In a nutshell, economic diplomacy is the use of international political tools (diplomacy) to obtain

Economic diplomacy, trade and developing countries economic objectives and as such has existed ever since ancient civilizations have engaged themselves in commerce and trade. Public ofcials from overseas missions such as embassies and consulates, Foreign Affairs and other government departments are involved in the promotion of international trade by supplying information and advice about trade opportunities and by organizing and helping to act as hosts to trade missions from the home country (Saner and Yu, 2003). The neo classical argument, however, is that rms can be expected to enter foreign markets on their own account and if they have to be aided by governmentsthen, perhaps their products are just not good enough. If so, the taxpayers money that ultimately nances economic diplomacy is wasted, just as in the case of a traditional export subsidy. Economic diplomacy implies a transfer (a free service) from the public sector to commercial activities while it is a priori unclear if the benets exceed the cost of providing the public service. Still economic diplomacy remains a relevant real-world phenomenon. Is this because the economic recipe is wrong, or is this just a case of economists that are right but are unable to get it right? At a very practical and concrete level, the public sectors involvement may be necessary for three reasons. Firstly, the type of product may require public sector involvement either on the demand side (for example, large infrastructural works) or on the supply side (for example, military or dual use goods). Secondly, trading may require public sector involvement with either the exporter or the importer for cultural reasons or because the public sector is dominant and/or (former) state enterprises are involved (typically relevant for relatively new entrants to the world economic system, that is, many developing countries). Thirdly, high ranking government ofcials may be needed to signal the importance that a country attaches to the commercial relationships that will be discussed offering an implicit guarantee that these relationships will be free from negative political disturbances. These arguments imply that public sector involvement is a necessary condition for some forms of market access and an instrument to reduce or eliminate cultural non-tariff barriers to trade and investment. A key insight of the literature is that diplomatic representations provide superior trade- and investment-related knowledge. This insight implies that market failure regarding the production of (private) knowledge provides a further argument for economic diplomacy. An exporter rm needs a lot of information about foreign markets before it can successfully attempt to trade with a rm in another country. Sometimes this information is acquired experimentally by trying to enter the country, but often consultants and business trips and information from export promotion agencies or colleagues will be tapped. As a general rule, the necessary information requires substantial investment and will be imperfect in nature. Often such information has an asymmetric character because rms in the importing country have a substantial advantage in both acquiring and developing country-specic knowledge and in assessing the reliability of such information. Typically, this manner of knowledge is built in networks of foreign rms (because proximity and existing business and personal networks are often a pre-condition); such networks may also act as a barrier to entree. Moreover, private investment in better knowledge is sub-optimal because of externalities (such as demonstration effects that can be followed suit by competitors), free rider problems or due to its public good character. According to Harris and Li (2005, 74), economic diplomacy provides: unique, reliable and impartial access to information, such as through the global embassy network and other government channels and contacts, which become available through the governments very long-term, and non-commercial attachment to overseas markets. The relevant rationale for government intervention, that is an active role in the generation of knowledge or the allocation of subsidies to investments in such knowledge, rests on the fact that the production of knowledge about foreign markets will be suboptimal and that access to such knowledge in some cases actually requires involvement of government ofcials. Indeed, the market does not supply the 255

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Yakop and van Bergeijk optimal investment of rms in international activities if such learning externalities exist. Governments, moreover, may have other roles to play when market failures occur (Alexander and Warwick, 2007), for example, in signalling the quality of its exporters. The authorities can clarify that their rms have to meet high standards in terms of product quality, environmental standards, corporate responsibility and, moreover, they may be able to communicate that their economy is a reliable partner in international trade. Such promotional information may be seen as an investment in the exporting nations trademark or trade capital and as such has a public good character. The mere existence of a market failure and the possibility that public intervention may x that problem, however, constitutes a necessary but never a sufcient condition (Hoekman and Smarzynska-Javoricik, 2004, 3). The upshot is that government intervention should generate more benets to society than it costs. Obviously, such a cost-benet analysis would go beyond the scope of this paper given the large number of countries involved. Note that we formally will not econometrically test the factors that we discussed in this section (our test is indirect) because an empirical analysis showing that the number of embassies and consulates has a positive impact on bilateral trade does not logically prove that this is due to the economic rationale that we discussed in the present section. We do, however, provide important basic information for cost-benet analysis by testing empirically whether embassies and consulates promote trade. Indeed, rejection of the hypothesized positive impact of embassies and consulates would imply a negative outcome for any cost-benet analysis. We now turn to this important empirical question. in regional economics. In the basic model, three explanatory variables appear: (i) the exporting countrys gross domestic product (GDP) Yi, (ii) the importing countrys GDP Yj and (iii) the distance Dij between the two countries. The basic model is known as the gravity equation, because of its similarity to the Newtonian law of gravity: the bilateral trade ow is assumed to be a function of the economic masses of the two trade partners and the inverted distance between the two countries. The intuition behind this formula is appealing. First, the supply of goods depends positively on the exporting countrys economic size and production capacity that is represented by its GDP. Second, the demand for these exports depends positively on the importing countrys market, which is also represented by its GDP. Third, transportation costs, transportation time and the economic horizon of the exporter (all assumed to correspond roughly with the geographic distance between the exporting and importing country) have a negative impact on trade. The gravity model in its simplest form is:
Eij =Yia Yjb Dc ; ij 1

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Empirical design: the gravity approach

A lot of models have been used to describe international trade ows. Models differ with respect to the degree of detail and the specic theoretical interest of the investigator. For the present study, we use a relatively simple empirical model. Tinbergen (1962, 262293) and Linnemann (1966) are traditional references on the early gravity model. The gravity model is an empirical workhorse for world trade analysis and

Here a > 0, b > 0 and c < 0, while Eij are the exports from country i to country j. Usually, the populations of the trade partners are added to this equation as are a number of other (often dummy) variables that represent trade-enhancing and trade-resistance factors that are typically relevant in bilateral exchanges. Examples are common borders, languages and currencies, (ex-)colonial relationships and individual country characteristics such as the area of the economy or the fact that it is an island or landlocked. The gravity model provides an empirical explanation for the geography and level of bilateral trade ows. Bergstrand (1985, 1989) was the rst to relate the gravity equation to its microeconomic foundations and to formally derive the gravity equation within the context of a general equilibrium model with imperfect competition and product differentiation (Bergstrand, 1989). Deardorff (1995), Evenett and Keller (2002), and Feenstra et al. (2001) are examples of contributions to the gravity literature that showed that the model can be derived from HeckscherOhlin, increasing returns to scale,

256

Economic diplomacy, trade and developing countries Ricardian models and so on. Interest in the gravity equation and its ability to track non-economic factors was also stimulated by the inuential article by Anderson and van Wincoop (2003), who show that policy barriers, information and enforcement costs and diverging rules and legal frameworks were shown to exercise the same impact as an ad valorem tariff equivalent of 44% (note well twice the impact of transportation costs). Indeed, it is this cost component that produces negative externalities and therefore it is the target for policy makers that want to minimize the cost of distance. Empirical research, moreover, uncovered distances truly multidimensional (economic, cultural, political and so on) nature (see Linders et al., 2004) and also these new dimensions proved to be very stimulating. We are concerned with the increased explanation of including instruments of economic diplomacy in a traditional trade model. As this investigation deals with the actual impact of economic diplomacy on the level and pattern of bilateral trade ows, the choice of the gravity model is almost unavoidable. In addition, robustness and general acceptance of the method are essential for our analysis. Therefore, it is relevant that the gravity model presently is accepted both in academic and in policy circles. In this article, we follow the research design of the studies in the eld as closely as possible (and particular Rose, 2007) so as to provide a useful basis for comparison. In particular, we will use his reduced form gravity equation and where possible we use the same, but often updated sources for the explanatory variables (as indicated in Table 2), but note that we construct a new data set for the key variable of interest, that is the number of embassies and consulates.3 The equation to be estimated is:
lnXij =b0 +b1 lnDij +b2 lnYi +b3 lnYj +b4 lnPopi +b5 lnPopj +b6 Langij +b7 Contij + b7 Landlij +b8 Islandij +b9 ln Areai Areaj +b10 Colij +b11 CUij +b12 FTAij + cEmbConij +eij ; 2

where i denotes the exporter and j denotes the importer. The relevant variables are as follows: Xij

is merchandise exports in dollars for the year 2006, from i to j; EmbConij is the number of embassies and consulates (that is ofcial foreign missions) that i has in j; Dij is the distance between i and j; Yk is GDP per capita in dollars in 2006, for k = {i, j}; Popk is average population (in millions of people in 2006), for k = {i, j} and Areak is the area of the country (in square kilometres) for k = {i, j}. The following are the dummy variables: Langij is 1 if i and j have a common language, 0 otherwise; Contij is 1 if i and j share a land border, 0 otherwise; Colij is 1 if i and j are colonies or ever shared a colonial relationship, 0 otherwise; CUij is 1 if i and j use the same currency, 0 otherwise; FTAij is 1 if i and j belong to the same regional trade agreement (RTA), 0 otherwise; Landlij is the number of landlocked countries in the country pair: 0 if both i and j are not landlocked, 1 if either i or j is landlocked and 2 if both i and j are landlocked and eij is the error term. Data on bilateral trade ows in millions of U.S. dollars at free-on-board prices were collected from the IMF Direction of Trade Statistics for 2006. GDP per capita and population are obtained from the IMF World Economic Outlook Database for 2006. Data on distance, area, common language, common border, colonial relationship and landlockedness were collected online from CEPII, which provides area of countries and geodesic distances between countries using the great circle formula. With respect to these data, we use the same sources as the other studies on economic diplomacy. We constructed the dummy variables for countries in the same currency union (CU) and the same RTA by inspecting the World Trade Organization website on all RTAs and CUs ratied until 2006 and the data set for the number of embassies and consulates by inspecting the websites of the 63 ministries of Foreign Affairs. Here our data also differ qualitatively since we count embassies and consulates irrespective of their location in a country and because we do not count so-called permanent representations because those diplomats work in international organizations (United Nation, International Labour Organization or North Atlantic Treaty Organization) and are not involved in bilateral trade and should thus not be considered for our topic. 257

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

258
Table 2. Empirical results of the OLS estimation of the gravity equation. Model Sample Year of observation N Number of foreign missions (export promoting) Number of foreign missions, j in i (import facilitating) Log distance Log exporter GDP p/c Log importer GDP p/c Log exporter population Log importer population RTA (dummy) CU (dummy) Log product area Common language (dummy) Land border (dummy) Landlocked (dummy) Islands (dummy) Colony (dummy) Adjusted R2 Data (1) Rose 22 3 200 2002 4123 D 0.10*** (0.02) D C U U U U D D C C C C C C 0.69*** (0.04) 0.86*** (0.03) 0.83*** (0.02) 0.96*** (0.03) 1.01*** (0.02) 0.86*** (0.08) 0.27 (0.18) 0.15*** (0.01) 0.57*** (0.07) 1.06*** (0.16) 0.75*** 0.05) 0.27*** (0.05) 3.25*** (0.38) 0.77 0.79*** (0.04) 0.86*** (0.04) 0.89*** (0.02) 0.98*** (0.03) 1.00*** (0.02) 0.29*** (0.07) 0.10 (0.15) 0.14*** (0.02) 0.34*** (0.10) 0.37** (0.16) 0.31*** (0.09) 0.05 (0.06) 0.14 (0.15) 0.77 0.74*** (0.04) 1.23*** (0.02) 0.99*** (0.02) 1.28*** (0.03) 1.18*** (0.03) 0.37*** (0.06) 0.34* (0.18) 0.20*** (0.02) 0.71*** (0.10) 0.90*** (0.17) 0.23*** (0.08) 0.07 (0.06) 0.04 (0.18) 0.67 0.05** (0.02) 0.75*** (0.04) 1.22*** (0.02) 1.01*** (0.02) 1.28*** (0.03) 1.21*** (0.03) 0.38*** (0.06) 0.33* (0.18) 0.19*** (0.02) 0.72*** (0.10) 0.95*** (0.17) 0.23*** (0.08) 0.08 (0.06) 0.07 (0.18) 0.67 0.03 (0.02) 0.74*** (0.04) 1.21*** (0.02) 1.99*** (0.02) 1.27*** (0.03) 1.17*** (0.03) 0.37*** (0.06) 0.35** (0.18) 0.20*** (0.02) 0.69*** (0.10) 0.86*** (0.17) 0.22*** (0.08) 0.07 (0.06) 0.02 (0.18) 0.67 0.06*** (0.01) 0.09*** (0.02) 0.09*** (0.02) (2) Replication 22 3 63 2006 1356 (3) Full sample 63 3 63 2006 3730 (4) (5) 63 3 63 2006 3730 63 3 63 2006 3730 Note: ***, **, * implies signicance at 99, 95 and 90% levels, respectively. D, empirical implementation of the variable is different from Rose (2007); U, same source as Rose (2007), but updated for 2006 and different country coverage; C, same source as Rose (2007) but with different country coverage. Standard errors reported in parentheses. Included in the regression but not reported is the constant.

Yakop and van Bergeijk

Economic diplomacy, trade and developing countries

Empirical results
Table 2 summarizes the results of the regression analyses. Table 2 starts with the results of Rose (2007) in Column 1 for a sample of 22 exporting countries and 200 importing countries in the year 2002. Next we provide for the year 2006 as close a replication as possible, starting with a sample that consists of Roses 22 exporting countries and our sample of 63 importing countries (Column 2). Column 3 reports the ndings for our full sample of 63 exporting and 63 importing countries. Before we move to the variable of interest, the number of foreign missions, we take a closer look at the general performance and the other variables of the gravity model. Comparing columns 1 and 2, we note that the results obtained are close to those of Rose with equal signs of the estimated coefcients, and a comparable explanatory power (see the adjusted R2). Roughly half of the coefcients estimated is about the same although we obtain estimates that are smaller in absolute terms for the coefcients of the dummy variables for RTA, CU, common language, land borders, landlockedness, islands and in particular for (ex-)colonial relationships. The latter is completely insignicant in our estimates, a result that conrms ndings by Bikker (1987, 2010) and Head et al. (2010) on the decreasing signicance of colonial relationships.1 We can now take a look at how the enlargement of the data set inuences the estimates as shown for the full sample 63 3 63 (Column 3). Typically, the absolute values of the coefcients for the core model (distance, per capita GDP and population) are comparable or larger. The coefcients for the dummy variables have a mixed record. Some coefcients converge to the estimates of Rose (2007) such as the common land border dummy and the CU dummy that is now both signicant and negative. Incidentally, we think that the latter result does not so much reect the impact from CUs as well as the fact that the major currency union (EMU) relates to OECD countries only and may thus be picking up the differences between developing and developed countries that we will discuss in the next section.4 Other coefcients increase such as product area and a common language illustrating how infor-

mation on a broader data set reects in the estimated coefcients. All in all, the estimated coefcient of interest (the number of embassies and consulates) appears to be rather stable. The coefcient estimates are positive and signicant ranging from 0.06 to 0.09. Therefore, based on a different sample of countries and sample year, our ndings would seem to agree with Rose (2007) when he argues that embassies and consulates are systematically linked with increasing exports. Admittedly, the actual replication (Table 2, Column 2) seems to indicate that this result is robust, but Roses conclusion is limited to a much smaller sample of mainly high-income exporting countries and, as will soon become clear, mainly driven by NorthSouth trade.

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Two econometric issues: endogeneity and zero ows

So far we assumed that an effect exists that foreign missions exert on trade, but it is not unreasonable to assume that diplomatic missions will be set up in countries with which one has already strong economic relationships. If so then endogeneity of the number of foreign missions is of econometric concern because the problem of reverse causality leads to biased and inconsistent estimates. We addressed this issue by introducing instrumental variables (IV) that correlate with the number of foreign missions and yet are uncorrelated with the error term of bilateral trade between these two countries. We collected two groups of instrumental variables that either measure the geo-political weight a country (proven gas reserves and proven oil reserves) or the quality of living and attractiveness as the desirability of residing in a particular country may be an incentive for diplomats to set up missions (number of Economist city guides, the number of Lonely planet guides and climate). Using two-stage least squares and generalized methods of moments, we nd that the previously obtained results continue to hold (the measured effect of foreign missions upon exports is even higher with IV estimation (although slightly less signicant): 0.16 versus 0.09 in Table 2, Column 3.5 259

Yakop and van Bergeijk

Table 3. Instrumental variables. (1) Two-stage least squares 63 3 63 (2) Generalized methods of moments 63 3 63 0.16** (0.06) 3730 0.67 (3) DurbinWu Hausman test (OLS) 63 3 63 0.13*** (0.05) 0.04 (0.05) 3730 0.67

Number of foreign missions, i in j ResidIV N Adjusted R2

0.16** (0.06) 3730 0.67

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Note: ***, ** implies signicance at 99 and 95% levels respectively. Standard errors reported in parentheses. Included but not reported are the explanatory variables of the gravity model as reported in Table 2. Full details are provided in Yakop and van Bergeijk (2009).

Column 3 of Table 3 in addition reports the DurbinWuHausman test for endogeneity, where we regress the number of foreign missions on the ve instrumental variables and then include the residuals of this auxiliary regression (ResidIV) as a new variable in the initial gravity equation. The condence level of ResidIV is 56% only so that the coefcient is not signicantly different from zero. Therefore, also on the basis of the DurbinWu Hausman test, we can safely and in line with the literature conclude that the number of foreign missions is exogenous and that our estimates do not appear to suffer from reverse causality. The second econometric issue is the occurrence and impact of zero trade ows. Out of a potential of 3906 bilateral trade ows, we have 176 cases (4.5%) in which no trade ow is recorded.6 For practical estimation purposes, zero-ow observations were left out. The most obvious practical problem is that ln(0) is not dened. The gravity model predicts zero ows only when at least one explanatory variable is zero and for the explanatory variables in the present model this is unlikely to be the case. Hence for (approximately) zero trade ows, the model may not give an appropriate description of the relationship between trade ows and its explanatory variables. Trade data sources, however, do generally only report annual bilateral trade ows that exceed some threshold. Consequently, relatively small transactions may occur when the data source reports a (rounded off) zero. The combination of actual trade and ofcially reported zero ows may also occur in the case of smuggling or politically sensitive goods such as military procure260

ments. So the trade registration system obviously generates zero or near-zero ows that do not require one of the explanatory variables to be zero. The relevant question is: does the existence of zero ows matter for our results. First we take a look at the characteristics of these trade ows. In our sample, zero exports occur rather frequently for Sudan (23 times), Vietnam (22), Algeria (16) and Uganda (16). Of the 176 zero trade ows, some 87 are zero in both directions (that is bilateral exports and imports between country A and country B are zero). This group mainly consist of small countries that are located in different continents. Vietnam has the largest share in these bi-directional zero trade ows (16 ows or 18.3%). In the other 89 cases, one ow (for example, export) is zero, but the other ow (for example, import) is strictly positive. There is no obvious pattern and one-sided zero ows occur in all continents and at levels of development although more often in the trade relations with lower and middle-income countries that are located in different continents. In order to check whether this problem inuences our results, we perform a linear transformation of the dependent variable and add a small constant v to all export data so that the zero observations become strictly positive and then re-estimated the equations (Table 4). Column 1 repeats for convenience (Table 2, Column 3). The other three columns show the results for different values of v since the choice of the constant is rather arbitrary and can inuence the results. Typically, however, the key result for the impact of embassies and consulates remains valid and appears to be even a bit stronger.

Economic diplomacy, trade and developing countries

Table 4. Zero trade ows. (1) Substituted value N Number of foreign missionsy (export promoting) Adjusted R2 (2) 0.5 3906 0.14*** (0.04) 0.49 (3) 500 3906 0.13*** (0.03) 0.62 (4) 5000 3906 0.12*** (0.02) 0.67

3730 0.09*** (0.02) 0.67

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Note: *** implies signicance at 99% level. Standard errors reported in parentheses. Included but not reported are the explanatory variables of the gravity model as reported in Table 2. Full details are provided in Yakop and van Bergeijk (2009).

All in all our estimates pass the econometric tests. We therefore conclude that the gravity equation provides a useful and robust tool to further investigate the relevance of our ndings for countries at different levels of development.

Import facilitation
The symmetric nature of our data matrices allows us to investigate import facilitation in addition to and in combination with the export facilitation that we have studied until now. The measurement of import facilitation is simply the inverse of export facilitation. Many of the functions of embassies and consulates such as providing information, building trust or public involvement matter as much for imports as for exports. Indeed attracting business and imports towards the domestic economy is an essential function of economic diplomacy. Thus, the exports of goods from, for example, Egypt to the Netherlands may be stimulated by both the export facilitating role of the Egypt Embassy in The Hague and the import facilitating role of the Netherlands Embassy in Cairo. This is, however, an empirical question: if import facilitation turns out to be insignicant, then outward economic diplomacy is more effective in ghting market failures than inward economic diplomacy. Two additional regressions provide information. Column 4 in Table 2 replaces the independent variable number of foreign missions with the import facilitating representation number of foreign missions, j in i and reports a signicant effect of import facilitation. An extra mission of the exporting county in the importing country leads to 5% more imports for the importing country if we do not take into

account of the export facilitating effect. Column 5 in Table 2, however, takes simultaneously the presence of the exporting countries in the importing countries (i in j) and also the presence of the importing countries in the exporting countries (j in i) into account. In this specication, the import facilitating embassies and consulates are insignicant. Export facilitating representations are still signicant increasing exports once more by 9%. This suggests that diplomatic representations do not focus on attracting business towards the domestic home markets as much as they support businesses abroad. Since this is a highly policy relevant nding, we provide additional tests in Table 5. Table 5 presents results of xed effects (FE) for all combinations of export facilitation, import facilitation and exporter FE and importer FE. For the sake of comparison, OLS coefcients are reported in the Column 1. Two forms of FE are used: exporter FE (Column 2) and importer FE (Column 3). For the equations that investigate the export promoting aspect of embassies and consulates, the coefcient estimates are in a range from 0.05 to 0.10 and are consistently signicant at a 99% condence level. The reported estimates for import facilitating representations range between an insignicant 0.03 (in the cases of simple OLS and exporter FE) to 0.06 in the other cases (signicant at a 95% condence level and better). These ndings imply that controlling for unobserved factors are important for the explanation of export facilitation and/or import facilitation. Note that the coefcient reported in Column 5, Table 2 suggested import facilitation to be of insignicant impact once 261

Yakop and van Bergeijk

Table 5. Coefcient estimates for the number of foreign missions with FE. Model OLS (1) Exporter FE (2) 0.10*** 0.06*** 0.09*** 0.03*** Importer FE (3) 0.06*** 0.06*** 0.05** 0.04**

1 2 3

Export promoting Import facilitating Export promoting Import facilitating

0.09*** 0.05** 0.09*** 0.03

Note: ***, ** implies signicance at 99 and 95% levels respectively. Standard errors reported in parentheses. Included but not reported are the explanatory variables of the gravity model as reported in Table 2. Full details are provided in Yakop and van Bergeijk (2009).

we controlled for the export promoting role of embassies and consulates. The ndings in Table 5 provide a different view. Although the effect of export promoting representations in general continues to be stronger than the effect of import facilitating, we nd that import facilitation matters once we take account of FE. One of the unobserved factors in our model is the level of development that we study in the next section.7

Development and effectiveness of economic diplomacy

The role of economic diplomacy is probably more important for the international economic relations of developing countries. First, asymmetries in the distribution of knowledge across countries are largest between the developed and the developing countries so that we would expect that public knowledge generation is especially relevant in the NorthSouth and SouthNorth relations. Second, trade-related institutions are less well developed outside the OECD and thus one expects that the government has a larger role to play in trade relations that start or end in the South. Third and related, transparent and easy accessible market information is available to a much larger extent for the developed countries so that we expect that economic diplomacy is a less relevant determinant of their mutual trade. Indeed Piermartini and Teh (2005, 37) are right when they point out that search costs will tend to be higher between countries with 262

different business practices and/or where competitiveness and reliability are not well known to one another. Fourth, the role of embassies and consulates may be of particular relevance in the starting up phase of international trade and investment relationships (Afman and Maurel, 2010) and on this account it may have a particularly high value added in trade between emerging markets. The assumption that embassies and consulates exert a stronger impact on trade for countries at lower levels of development is born out by the data at the sub-sample level. Theoretically, we could conceive of four different income levels (the wellknown analytical country classication of the IMF and World Bank): LIC, lower middle-income countries (LMIC), upper middle-income countries (UMIC) and nally high-income countries (HIC). With four categories, we have potentially 16 subregressions (4 3 4) that analyse the trade between the different country groupings.8 Due to the small amount of observations in many cross sub-sample trade ows, it makes sense to combine on the one hand, the sub-samples LIC and LMIC (8 + 14 = 22 countries) so that we have one group of lower income countries and on the other hand, UMIC and HIC (14 + 27 = 41 countries) so as to have one group of higher income countries. Thus four separate regressions can be run based on larger sub-samples, which provide more robust estimates. We estimate gravity equations that include both export facilitation and import facilitation (thus comparable to Table 2, Column 5). The results of these four regressions are found in Table 6 that shows that export facilitation is signicant at the 99% level for trade between developing and developed countries (and in both directions) and also within the group of lower income countries (LIC and LMIC). In contrast within the group of higher income countries (UMIC and HIC), the coefcient for embassies and consulates is not signicant. For import facilitation, we nd smaller coefcients that are often less signicant, much in line with the ndings reported in Tables 2 and 5. Import facilitation according to our estimates is in particular relevant in NorthSouth and SouthNorth trade.

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Economic diplomacy, trade and developing countries

Table 6. Impact export and import facilitation for trade between different income country groupings (b coefcients). (To) Importing countries LIC + LMIC (From) Exporting countries LIC + LMIC UMIC + HIC Export facilitation Import facilitation Export facilitation Import facilitation 0.12*** 0.04 0.10*** 0.10*** UMIC + HIC 0.11*** 0.08* 0.02 0.01

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Note: ***, * implies signicance at 99 and 90% levels respectively. Standard errors reported in parentheses. Included but not reported are the explanatory variables of the gravity model as reported in Table 2. Full details are provided in Yakop and van Bergeijk (2009).

Discussion and policy conclusions

Our ndings show that the positive and signicant estimates of the trade elasticities for the number of embassies and consulates in the literature need classication. Our estimates of this elasticity range from 0.06 to 0.16 and for trade by and with developing countries, we nd a larger and more signicant effect. At the same time, we nd that economic diplomacy does not enhance trade within the group of HIC and UMIC, which implies a negative outcome of a costbenet analysis for economic diplomacy aimed at the bilateral trade ows of these countries. So the generally reported positive effects of economic diplomacy relate only to trade ows that originate or end in developing countries where it adds value as it facilitates higher trade volumes. This is especially true not only for SouthSouth trade but also to a lesser extent between developing to developed countries and vice versa.9 Our ndings conrm earlier empirical studies on related topics. Recently, Lederman et al. (2006) took up the issue of the effectiveness of export promotion in a group of 104 countries and the year 2005. This study is especially quoted for having found a correlation between per capita export promotion and per capita exports. The elasticity in Lederman et al. (2006) at the mean of the sample explains about 8% of the median countrys export, which is more or less of the same order of magnitude as found in our study and in the empirical studies on the impact of embassies and consulates that we discussed in this article. However, and highly relevant from the perspective of the present article, these results are driven

by the developing countries suggesting substantial heterogeneity across levels of development. In particular, it is important that for OECD countries the relationship in Lederman et al. (2006) is insignicant like in our study; that is: the effect is zero. Our results therefore also offer indirect support for the emerging empirical literature on new and intangible barriers to trade such as a lack of trust, cultural differences and ineffective governance (a lack of an enforceable legal framework, accountability and stability). We have discussed a number of potential market failures that hamper international exchange and in our opinion our empirical results imply that embassies and consulates succeed in solving or reducing some of these market failures, the more so because of the fact that export facilitation between developed and developing countries is signicant whereas it is not within the separate group of higher income countries. In our opinion, this reects that markets in the developing countries tend to be more incomplete implying that market failures may be more of a problem in these countries then in the industrial countries. Therefore, it is relevant for our discussion that Dekker et al. (2006), den Butter and Mosch (2003) and Guiso et al. (2004) nd that a one standard deviation increase in trust increases bilateral trade on average by some 2438%. Highly relevant is the assessment of this by De Groot et al. (2004, 119) that good governance lowers the transaction costs for trade between HIC, while trade between LIC suffers from high insecurity and transaction costs. Typically, economic diplomacy can be useful in the context of such market and government failures. 263

Yakop and van Bergeijk Indeed, our nding of signicant and high estimates for embassies and consulates in the trade ows from higher income countries to lower income countries and vice versa is another corroboration of the fact that market failures do exist at least in international trade that justies the existence of economic diplomacy in order to solve problems related to market failures. We expect that this phenomenon will become more important over time. The emergence of new economies with very different institutions and cultural background will inuence global norms and values and this will undoubtedly have an impact on the rules of international trade. In particular, the historical, cultural and institutional background of China and India may in the long run exert an inuence on the ways the world denes and settles international conicts. From this perspective, a potentially interesting distinction could be to focus in future research on the extent of government involvement in the economic process.10 Typically, the available analyses of low volumes of SouthSouth trade are concerned with infrastructure and trade facilities. Our analysis adds a new element to that discussion, namely the need to establish good political relationship to breed trust as an important facilitator of mutually benecial trade.
effects is stronger for countries that are located near each other such as EU members (Head and Mayer, 2010; van Bergeijk and Brakman, 2010, 19, fn 22). One referee rightly points out that this explanation is not completely convincing in view of the fact that our empirical specication controls for GDP per capita. This coefcient is measured less precise in Table 3 as the standard error is three times the 0.02 reported earlier so that the level of signicance drops to 95%. Typically in other investigations, zero trade ows are a much more substantial nuisance. Rose, for example, has 277 cases or 6.3%, but even in his study the problem of zero trade ows is not really as important as in most applications of the gravity model, see Burger et al. (2009) and van Bergeijk and Brakman (2010).
7 6 5

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

One referee suggests to allow for interaction effects between different distance variables to see whether diplomacy has a different impact depending on the specic barrier to trade that is considered. Interaction terms could also provide an alternative method to analyse the impact of the level of development or to consider whether the impact of diplomacy varies depending on the type of good that is being traded. We have put these issues on the list of topics that we want to address in future work. A detailed analysis of these 16 sub-regressions is provided in Yakop and van Bergeijk (2009).

Endnotes
1

One of the sensitivity analyses in Rose (2007, Table 2, 31) implicitly foreshadows this results to some extent as he reports a smaller effect for a sub-sample of countries with per capita GDP above $1000. We owe this point to a referee.

Although preferences based on colonial ties from the past have become weaker in the course of time (Bikker, 1987, 330), signicance of political factors still is a consistent nding in alternative specications and data sets (see van Bergeijk, 2009, Table 1.2, 6).

10

References
Afman, E. and Maurel, M. (2010) Diplomatic relations and trade reorientation in transition countries. In P. A. G. van Bergeijk and S. Brakman (eds). The Gravity Model in International Trade: Advances and Applications, pp. 278295. Cambridge: Cambridge University Press. Alexander, C. and Warwick, K. (2007) Governments, exports and growth: responding to the challenges and opportunities of globalisation. World Economy, 17794. Anderson, J. E. and van Wincoop, E. (2003) Gravity with gravitas, a solution to the border puzzle. American Economic Review, 93: 170192.

In the discussion of our empirical results, we will often refer to Roses ndings as this has become the major reference on this topic, but it should be noted that the other studies that we discussed are equally vulnerable for a critique on the way their data sets have been constructed.

Note that we will to some extent take care of this issue as we replicate Roses research as a rst step in our procedure (for a matrix of his 22 exporters 3 our 63 import destinations).
4 Bikker (2010, Table 5.1) also reports a signicant negative coefcient for European Union (EU) integration in a gravity model for the year 2005. An alternative explanation for these ndings is that the extent of distance

264

Economic diplomacy, trade and developing countries

van Bergeijk, P. A. G. (2009) Economic Diplomacy and the Geography of International Trade. Cheltenham, UK: Edward Elgar. van Bergeijk, P. A. G. (2010) On the Brink of Deglobalization; an Alternative Perspective on the World Trade Collapse. Cheltenham, UK: Edward Elgar. van Bergeijk, P. A. G. and Brakman, S. (2010) The comeback of the gravity equation. In P. A. G. van Bergeijk and S. Brakman (eds). The Gravity Equation in International Trade: Advances and Applications, pp. 128. Cambridge: Cambridge University Press. Bergstrand, J. H. (1985) The gravity equation in international trade: some microeconomic foundations and empirical evidence. Review of Economics and Statistics, 67: 474481. Bergstrand, J. H. (1989) The generalized gravity equation, monopolistic competition, and the factor-proportions theory in international trade. Review of Economics and Statistics, 71: 143152. Bikker, J. A. (1987) An international trade ow model with substitution: an extension of the gravity model. Kyklos, 40: 315337. Bikker, J. A. (2010) An extended gravity model with substitution applied to international trade ows. In P. A. G. van Bergeijk and S. Brakman (eds). The Gravity Model in International Trade: Advances and Applications, pp. 135162. Cambridge: Cambridge University Press. Burger, M. J., van Oort, F. G. and Linders, G. J. M. (2009) On the specication of the gravity model of trade: zeros, excess zeros and zero-inated estimation. Spatial Economic Analysis, 4: 167187. den Butter, F. A. G. and Mosch, R. H. J. (2003) Trade, Trust and Transaction Costs. Tinbergen Institute Working Paper No. 2003-082/3. Amsterdam, The Netherlands: TI. Cheng, I. H. and Wall, H. J. (2005) Controlling for heterogeneity in gravity models of trade and integration. Federal Reserve Bank of St. Louis Review, 4963. Deardorff, A. V. (1995) Determinants of Bilateral Trade: Does Gravity Work in a Neoclassical World? NBER Working Paper W5377. Cambridge, MA: NBER. Dekker, P., Ederveen, S., de Groot, H., et al. (2006) Divers Europe, European Outlook 4. The Hague, The Netherlands: SDU. Evenett, S. and Keller, W. (2002) On theories explaining the success of the gravity equation. Journal of Political Economy, 110: 281316. Feenstra, R. C., Markusen, J. R. and Rose, A. K. (2001) Using the gravity equation to differentiate among alternative theories of trade. Canadian Journal of Economics, 34: 430437. Gil-Pareja, S., Llorca-Viveroa, R. and Martnez-Serrano, J. A. (2007) The impact of embassies and consulates on tourism. Tourism Management, 28: 355360. de Groot, H. L. F., Linders, G. J., Rietveld, P. et al. (2004) The institutional determinants of bilateral trade patterns. Kyklos, 57: 103124. Guiso, L., Sapienza, P. and Zingales, L. (2004) Cultural Biases in Economic Exchange. NBER Working Paper 11005. Cambridge, MA: NBER. Harris, R. and Li, Q. C. (2005) Review of the Literature: The Role of International Trade and Investment in Business Growth and Development. London: DTI. Head, K. and Mayer, T. (2010) Illusory border effects: distance mismeasurement inates estimates of home bias in trade. In P. A. G. van Bergeijk and S. Brakman (eds). The Gravity Equation in International Trade: Advances and Applications, pp. 165192. Cambridge: Cambridge University Press. Head, K., Mayer, T. and Ries, J. (2010) The erosion of colonial trade linkages after independence. Journal of International Economics, 81: 114. Head, K. and Ries, J. (2006) Do Trade Missions Increase Trade? Saunder School of Business Working Papers. University of British Columbia. Hoekman, B. and Smarzynska Javoricik, B. (2004) Policies Facilitating Firm Adjustment to Globalization. Policy Research Working Paper 3441. Washington, DC: World Bank. Kostecki, M. and Naray, O. (2007) Commercial Diplomacy and International Business. Discussion Papers in Diplomacy. The Hague, The Netherlands: Netherlands Institute of International Relations Clingendael. Lederman, D., Olarreaga, M. and Payton, L. (2006) Export Promotion Agencies: What Works and What Doesnt? World Bank Policy Research Working Paper 4044. Washington, DC: World Bank. Linders, G.-J., de Groot, H. L. F. and Nijkamp, P. (2004) Locality matters: myths and facts on the new economy. In J. Poot (ed.). On the Edge of the Global Economy, pp. 2748. Cheltenham, UK: Edward Elgar. Linders, G.-J., Slangen, A., de Groot, H. L. F. et al. (2005) Cultural and Institutional Determinants of Bilateral Trade Flows. Tinbergen Discussion Paper, TI2005-074/3. Amsterdam, The Netherlands: TI. Linnemann, H. (1966) An Econometric Study of International Trade Flows. Amsterdam, The Netherlands: North Holland. van Marrewijk, C. and van Bergeijk, P. A. G. (1993) Endogenous trade uncertainty: why countries may specialize against comparative advantage. Regional Science and Urban Economics, 23: 681694. Nitsch, V. (2007) State visits and international trade. World Economy, 30: 17971816.

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

265

Yakop and van Bergeijk

Piermartini, R. and Teh, R. (2005) Demystifying Modelling Methods for Trade Policy. WTO Discussion Papers 10. Geneva, Switzerland: WTO. Rose, A. (2007) The foreign service and foreign trade: embassies as export promotion. The World Economy, 30: 2238. Saner, R. and Yiu, L. (2003) International Economic Diplomacy: Mutations in Post-Modern Times. Discussion Papers in Diplomacy No. 84. The Hague, The Netherlands: Netherlands Institute of International Relations Clingendael. Segura-Cayuela, R. and Vilarrubia, J. M. (2008) The Effects of Foreign Service on Trade Volumes and Trade Partners. Working Paper 0808. Madrid, Spain: Banco de Espana. Tinbergen, J. (1962) Shaping the World EconomySuggestions for an International Economic Policy. New York: Twentieth Century Fund. Yakop, M. and van Bergeijk, P. A. G. (2009) The Weight of Economic and Commercial Diplomacy. ISS Working Paper 478, The Hague. Available online at: http:// www.iss.nl/News/ISS-Working-Paper-478.

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Appendix
Table A1. Import Destinations of Rose.
Afghanistan Angola Azerbaijan Bahamas Belize Botswana Burundi Bolivia Canada (HIC) Chile (UMIC) Comoros Croatia Congo, Rep Dominica El Salvador Ethiopia Finland (HIC) Gambia Greece (HIC) Guinea-Bissau Guinea Iceland Iraq Japan (HIC) Kenya (LIC) Kyrgyz Rep. Lesotho Lithuania Maldives Morocco (LMIC) Mauritius Nepal Niger Namibia Oman Albania Argentina (UMIC) Antigua & Barbuda Bahrain Benin Brunei Darussalam Belarus (LMIC) Brazil (UMIC) Cape Verde China (LMIC) Congo, Dem. Rep. Cuba Cyprus Dominican Rep. (LMIC) Equatorial Guinea Faeroe Islands France (HIC) Georgia Greenland Guyana Haiti India (LIC) Ireland (HIC) Jordan Kiribati Laos Liberia Macedonia Malta Myanmar Mexico (UMIC) Netherlands (HIC) Nigeria (LIC) Nauru Pakistan (LIC) Algeria (LMIC) Aruba Armenia Bangladesh (LIC) Bhutan Bulgaria (UMIC) Belgium (HIC) Cambodia Central African Rep. China, Hong Kong Costa Rica Czech Rep. (HIC) Denmark (HIC) Ecuador (LMIC) Eritrea Falkland Islands French Polynesia Germany (HIC) Grenada Ghana Honduras Indonesia (LMIC) Italy (HIC) Jamaica Korea (HIC) Latvia Libya Madagascar Mauritania Malawi Mongolia Netherlands Antilles North Korea New Zealand (HIC) Palau American Samoa Australia (HIC) Austria (HIC) Barbados Bosnia & Herz. Burkina Faso Bermuda Cameroon Chad China, Macao Cote DIvoire Colombia Djibouti Egypt (LMIC) Estonia Fiji Gabon Gibraltar Guam Guatemala Hungary (UMIC) Iran (LMIC) Israel (HIC) Kazakhstan Kuwait (HIC) Lebanon Luxembourg Malaysia (UMIC) Moldova Mali Mozambique New Caledonia Norway (HIC) Nicaragua Panama

266

Economic diplomacy, trade and developing countries

Table A1. Continued Afghanistan Angola Azerbaijan Bahamas Belize Botswana Burundi Bolivia Canada (HIC) Chile (UMIC) Comoros Croatia Congo, Rep Dominica El Salvador Ethiopia Finland (HIC) Gambia Greece (HIC) Guinea-Bissau Guinea Iceland Iraq Japan (HIC) Kenya (LIC) Kyrgyz Rep. Lesotho Lithuania Maldives Morocco (LMIC) Mauritius Nepal Niger Namibia Oman Papua New Guinea Poland (UMIC) Russia (UMIC) Saudi Arabia (HIC) Sierra Leone Somalia St. Helena Sweden (HIC) St. Vincent & Gren. Tajikistan Tunisia (LMIC) Tonga United Arab Emirates Uzbekistan Wallis-Futuna Albania Argentina (UMIC) Antigua & Barbuda Bahrain Benin Brunei Darussalam Belarus (LMIC) Brazil (UMIC) Cape Verde China (LMIC) Congo, Dem. Rep. Cuba Cyprus Dominican Rep. (LMIC) Equatorial Guinea Faeroe Islands France (HIC) Georgia Greenland Guyana Haiti India (LIC) Ireland (HIC) Jordan Kiribati Laos Liberia Macedonia Malta Myanmar Mexico (UMIC) Netherlands (HIC) Nigeria (LIC) Nauru Pakistan (LIC) Paraguay Portugal (HIC) Rwanda Senegal Singapore (HIC) South Africa (UMIC) St. Kitts & Nevis Switzerland (HIC) Sudan (LIC) Tanzania Turkmenistan Turkey (UMIC) UK (HIC) Vanuatu Yemen Algeria (LMIC) Aruba Armenia Bangladesh (LIC) Bhutan Bulgaria (UMIC) Belgium (HIC) Cambodia Central African Rep. China, Hong Kong Costa Rica Czech Rep. (HIC) Denmark (HIC) Ecuador (LMIC) Eritrea Falkland Islands French Polynesia Germany (HIC) Grenada Ghana Honduras Indonesia (LMIC) Italy (HIC) Jamaica Korea (HIC) Latvia Libya Madagascar Mauritania Malawi Mongolia Netherlands Antilles North Korea New Zealand (HIC) Palau Peru (LMIC) Qatar Samoa Serbia & Montenegro Slovakia Spain (HIC) St. Pierre-Miquelon Slovenia Swaziland Togo Tuvalu Uganda (LIC) USA (HIC) Vietnam (LIC) Zimbabwe American Samoa Australia (HIC) Austria (HIC) Barbados Bosnia & Herz. Burkina Faso Bermuda Cameroon Chad China, Macao Cote DIvoire Colombia Djibouti Egypt (LMIC) Estonia Fiji Gabon Gibraltar Guam Guatemala Hungary (UMIC) Iran (LMIC) Israel (HIC) Kazakhstan Kuwait (HIC) Lebanon Luxembourg Malaysia (UMIC) Moldova Mali Mozambique New Caledonia Norway (HIC) Nicaragua Panama Philippines (LMIC) Romania (UMIC) Sao Tome & Principe Seychelles Solomon Islands Sri Lanka Suriname St. Lucia Syria Trinidad & Tobago Thailand (LMIC) Ukraine (LMIC) Uruguay (UMIC) Venezuela (UMIC) Zambia

Downloaded from http://cjres.oxfordjournals.org/ at Biblioteca Centrala Universitara din Timisoara on May 2, 2012

Note: The exporting countries studied by Rose are shown in bold (Rose, 2007). The highlighted countries are the 63 exporting and importing countries in our analysis.

267