ECONOMICS THE GRAVITY OF RESOURCES AND THE TYRANNY OF DISTANCE. Peter E Robertson Business School University of Western Australia.

ECONOMICS THE GRAVITY OF RESOURCES AND THE TYRANNY OF DISTANCE by Peter E Robertson Business School University of Western Australia and Marie-Claire Robitaille The University of Nottingham Ningbo China DISCUSSION PAPER 15.01

The Gravity of Resources and the Tyranny of Distance Peter E. Robertson The University of Western Australia Marie-Claire Robitaille The University of Nottingham Ningbo China November 2014 DISCUSSION PAPER 15.01 Abstract Falling transport costs and the rise of global production networks have reshaped world trade. But endowments still determine production locations for fuels and minerals. Moreover, because they are often bulky or difficult to store, unit transport costs for natural resources may still be very large. To what extent, therefore, does geography remain an important determinant of comparative and absolute advantage in these markets? We estimate gravity models and show that some minerals and fuels, particularly Iron Ore and Gas, have very high elasticities of trade with respect to distance. We then consider counterfactuals, how trade would differ if location advantages were eliminated. We find that for a few resource intensive countries, distance barriers have a large impact of their market share and are equivalent to a 70-100% ad valorem export tax relative to an average country. Similar implicit subsidies apply for a large number of well located countries. Keywords: International Trade, Energy, Resources, Gravity Model, Geography. JEL: F14, Q02 Robertson acknowledges the hospitality of St. Antony s College Oxford and The Centre for the Study of African Economies (CSAE) in the Department of Economics at The University of Oxford. We are grateful for comments from Peter Hartley and the participants at the Asia Pacific Trade Seminars (APTS) at Southeast University, Nanjing, China, 2013 and the International Workshop on Distance and Border Effects in Economics, Loughborough University, January 2014. Corresponding Author; Peter E. Robertson, Economics, School of Business, University of Western Australia, Perth. Email: peter.robertson@uwa.edu.au. Phone: +61 6488 5633.

1 Introduction Historically only the most precious commodities, such as silk and spices, were traded over very long distances, due to high transport costs. Remote antipodean colonies thus suffered from a tyranny of distance. 1 Technological changes, however, have allowed dramatic reductions in transport costs. Consequently, world manufacturing is now characterized by global production networks and intra-industry trade in product varieties. The importance of endowments as a source of comparative advantage is thus believed to have been eroded and preferences and institutions are argued to now be critical in understanding trade patterns (Romalis 2004, Levchenko 2007, Chor 2010). But for resource exporting countries, supply is indelibly linked to that country s endowments. Moreover many resources, such as iron ore and coal, are bulky and have high unit transport costs. Others, such as as fresh food and gas, have high storage costs. These two facts the link between endowments and supply and the high transport costs suggest that geography remains a very important factor in explaining the pattern of resources trade as well as the sources of absolute advantage, or competitiveness, in resource sectors. Understanding these issues is important since the impact of distance on each country s absolute advantage, will influence the degree of market power wielded by that country. Likewise it will influence the impact of tax and other policies aimed at resource companies in terms of their competitiveness and ability to seek alternative host countries. It is also important for understanding how future changes in technology that effect transport costs might affect world trade patterns and country s competitiveness in international markets. The aim of this paper, therefore, is to quantify the impact of distance on world food, minerals and fuels trade patterns. Specifically we show how world commodity trade patterns would differ if the natural barriers due to country location did not exist. We estimate gravity models for several energy and resource commodities and then consider a counterfactual set of distance values that removes any advantage or disadvantage of distance for any single exporting country. Thus we consider what would trade patterns in food, minerals and fuels would look like, if no country had an advantage or disadvantage in terms of geographical location. We find that several resource intensive economies exports are significantly suppressed 1 The term is due to Blainey (2001). He notes that the net inflow of diggers during the gold rush in Australia provided empty ships in Melbourne which could be used for wool exports. Blainey argues that the availability of empty ships reduced transport costs enough to make wool exports, which also had low perishability and a high value to weight ratio relative to wheat, a feasible export. 2

by their geographical disadvantage. Specifically there are five countries in which exports would be 35-50% higher if their location were an average distance from their markets. For these countries transport costs are the equivalent of a 70-100% ad valorem resource export tax relative to the average country in terms of geographical location. Moreover there are many countries that enjoy similar, or larger, implicit subsidy due to their location. The difference between being well located or remote is therefore very substantial. These differences are therefore very important in explaining the global pattern of resources trade, tax decisions of governments and location decision of firms. The remainder of this paper is organized as follows. In Section 2 we discuss the geography of world demand for different commodities. In Section 3 we consider gravity models for food, different raw materials and fuels. Sections 4 and 5 then consider the implications of a counterfactual experiment where distance barriers are equalized across countries and Section 6 concludes. 2 The Geography of World Resources Demand Our first task is to characterize the geography of world demand patterns for resources. This differs substantially by commodity. For example, although the USA is the world s largest importing country overall, China is the world s largest importer of Minerals followed by Japan. Likewise, with respect to Coal, Japan is by far the largest importer accounting for almost a quarter of world import demand while China is only the 10 th largest country, behind countries like Italy and India. Thus measures of the distance of a country from their export market will differ by commodity. Table 1 summarizes the import shares for selected resource commodities based on COMTRADE/WITS data, following the Standard International Trade Classification 1 (SITC-1), for 2006 while Figures 1 to 7 (Panel A) map the major importers for selected commodity groups. It can be seen that overall world import demand is largest in Europe-Central Asia (44%) followed by roughly equal shares from East Asia (23%) and North America (22%). 2 However, with respect to Minerals, world import demand is roughly split between Europe- Central Asia (38%) and East Asia (46%). Within this category world demand for Iron Ore is predominantly from East Asia (70%), of which China alone accounts for two thirds (48%). Thus East Asia s world share of Minerals imports is more than twice as large as 2 The list of countries included in each region is provided in Table A.1 while the definitions of the commodity groups are provided in Table B.1. 3

its overall world import share while in North America s Minerals import share is a mere third of its overall world import share. Likewise world Coal demand is mostly split geographically between East Asia and Europe- Central Asia, with relatively little demand from North America. However, for Gas and Petroleum, world demand is more evenly divided between Europe-Central Asia (35 and 36%), North America (27 and 25%) and East Asia (29 and 29%). Thus the world import demand is very biased toward East Asia for Minerals, especially Iron Ore, and for Coal. Conversely North American demand for these same commodities is much lower than its total trade shares. With respect to fuels other than Coal, however, world import demand shifts toward North America. Europe and Central Asia, in contrast, has a very large demand for Manufactured goods and Food, but relatively little demand for Minerals, except Coal. Finally as can be seen in Table 1, Food and Raw materials import patterns are similar to Manufacturing patterns and the overall average, though Europe can be seen to be a relatively large importer of Food and Raw Materials, while North America and Asia tend to import less food relative to other commodities. To what extent, therefore, does geography account for world trade patterns in these resources? One way of addressing this is to consider what the world pattern of trade would be if there was no location advantage or disadvantage for any country. This is our aim in the rest this paper. [Table 1 and Figures 1-8 about here] 3 The Gravity Model 3.1 Estimation Strategy and Data To estimate the impact of distance on the patterns of trade and specialization, we use the gravity model. Given our focus on commodities, we adopt Santos Silva and Tenreyro (2006) specification of the gravity model as our preferred specification. This has the advantage of handling the numerous zero trade flow observations, which are highly prevalent in the Minerals and Fuel sectors. Specifically in our sample, 61% of the country-pair-year do not trade Minerals, but only 18% do not trade Manufactured Goods. Our data set is constructed from COMTRADE/WITS yearly data for the 104 economies 4

for which all variables are available, following the SITC-1 classification for the period 1998-2003 and 2006. 3, 4 We use import data to measure trade flows between country-pair as imports data are believed to be less at risk of double-counting and misreporting of the country of origin/destination than exports data (Athukorala 2009). There are two problems with the traditional log-linearization of the gravity model equation. The first is that the data usually contain many zero values. This may arise from missing values or represent genuine instances of zero trade between country-pairs. The common solution of omitting the zero trade observations leads to selection bias (Santos Silva and Tenreyro 2006, Disdier and Head 2008). Santos Silva and Tenreyro (2006) also point out that the log linearization of the gravity equation leads to biased coefficient estimates in the presence of heteroscedasticity. They propose the use of a Poisson Pseudo-Maximum-Likelihood (PPML) model which, by avoiding log-linearization, thus avoids the problem of zeroes and bias. 5 Thus following Santos Silva and Tenreyro (2006), we estimate the gravity model using the PPML estimator. X c ijt = αgrav γg g ɛ F E f δ f (1) with Xijt c the trade volume in million USD (constant 2000 USD) between exporting country i and importing country j for commodity c in year t. GRAV is a matrix including the standard gravity variables. From the World Development Indicators (WDI), we include the log of GDP of the exporting country (ln gdp) and of the importing country (p ln gdp), the log of the population of the exporting country (ln pop) and of the importing country (ln p pop) as well as the log of the land area of the country-pair (ln land paire). From the Centre d Etudes Prospectives et d Informations 3 As in Greenaway, Mahabir and Milner (2008), we include Hong Kong s exports with China s exports as the two economies are closely integrated. As pointed out by Greenaway et al. (2008) many exports originating from those two countries combine management and distribution skills from Hong Kong and labour from China. 4 Those years and classification system have been selected so as to include, as far as possible, all large natural resources exporters. 5 An alternative approach is to use the Tobit model. However, as pointed out by Linders and Groot (2006) this approach relies on assumptions on the data generating process that do not hold in the gravity model of trade. More precisely, the Tobit model assumes that the data suffer from rounding, which is highly uncommon for trade data, or that the desired outcome may not be measured by the actual outcome, for example, negative value, again a characteristic not applicable to trade data. Helpman, Melitz and Rubinstein (2008) similarly consider a selection model where only the most productive firms export. We find that their model is compelling for understanding manufacturing trade but less convincing for resources trade. 5

Internationales (CEPII) data set, we take the log of the weighted great-circle distance between the country-pair, with the weight depending on the population distribution within both partner countries (ln distwces), a set of dummy variables classifying the pair of country as none is landlocked (reference category), one country is landlocked (landlocked 1) and two countries are landlocked (landlocked 2), a dummy variable taking the value of one if the two countries in the country-pair are contiguous (contig), a set of dummy variables classifying the country-pair into none of the countries is an island (reference category), one country in the country-pair is an island (island 1), and the two countries in the country-pair are islands (island 2), a dummy variable taking the value of one if at least one language is spoken by at least 9% of the population in both countries (comlang ethno) and, finally, a dummy variable taking the value of one if the two countries in the country-pair have ever been in a colonial relationship (colony). We also include the log of human capital level in the exporting country (hum k) and of the importing country (p hum k). 6 Finally, FE is a set of fixed effects for exporting countries (COUNTRY ), for importing countries (PCOUNTRY ) and for years (YEAR). The standard errors are adjusted for clustering at the country-pair level. The commodities have been classified into broad categories, that is, All, Food, Raw Materials, Minerals, Coal, Petrol, Gas and Manufactured Goods. Given its importance in world trade, we also estimate a model for Iron Ore (24% of Minerals exports in 2006). The exact SITC-1 categories included into each element of this classification are presented in Table B.1. 3.2 Trade-Distance Elasticities by Commodity The results are given in Table 2. We find that an increase of 1% in distance leads to a drop of approximately 0.77% in total exports and a 0.75% fall in Manufacturing trade. These results are consistent with the existing literature (Disdier and Head 2008). It can be seen, moreover, that distance is highly significant for all of the different commodity groups. Nevertheless it can also be seen that the elasticities for some commodities differ substantially from Manufacturing. They range from -0.86 for Raw Materials to -1.96 for Iron Ore and -2.56 for Gas. In general the trade elasticities for Iron Ore and fuels are substantially larger than the elasticities of other commodities and Manufacturing. Interestingly we do not find much evidence than geography matters in other respects 6 Human capital is measured using the Mincerian relationship e 0.15s where s is the average years of schooling in the labour force (Barro and Lee 2010). 6

than distance and the contiguous border dummy. Sharing a border increases the volume of trade by 26%. Moreover the impact of sharing a border is particularly large in the case of Minerals and Gas, increasing the volume of trade by approximately 60%. [Table 2 about here] 3.3 Quantifying the Impact of Distance To quantify the implications of geographical location in determining the pattern of world trade we consider a counterfactual experiment, where each export market is at the same distance from every country. In designing this experiment, we keep the average distance to importing country the same so as to keep the geography of demand constant. More precisely, let denote the actual distance from exporting country i to import country j as d ij. We then replace all of the d ij with a common value d j where d j is the average distance for all exporters to destination market j, that is d j = i d ij /(n 1), where n is the number of countries. One way to think of this is that we impose a destination specific export tax, or subsidy, on all countries in proportion to their distance from the export market, such that that no country has any transport cost advantage or disadvantage due to the distance from its export markets. Thus we calculate the average distance of all countries in our sample to each destination country. We hence have an average value specific to each importing country. These values are presented in Table C.1. We then use these counterfactual values d j to recalculate total world trade for each commodity group using the coefficients from (1). 4 Geography and Global Resource and Energy Export Shares The aggregate change in trade implied by this experiment for each commodity group is shown in Table 3. It can be seen that although average trade weighted distance only increases by 5%, total world trade in the experiment falls by 34%. 7 7 The average trade weighted distance is calculated for the year 2006 using the following formula. Let J be the set of world markets and J i be the set of export markets for country i J. We define the average distance to market for country i for commodity c as D i,c = j d ijs jc, j J i, where s jc is country j s share of imports for all countries j J i. Likewise the average distance to market implied by the counterfactual distances is D i,c = j d j s jc, where it will be recalled that d j is the counterfactual common distance between all exporters and the destination market j. 7

Thus as we redistribute distance equally across countries, but preserve total distance between all country pairs, trade volumes fall substantially. Hence, despite keeping the total average distance constant, the total trade costs have increased. This reflects the fact that neighboring countries trade much more with each other than with distant countries, and that the relationship between trade and distance is non-linear. In particular Europe consists of many large countries with a lot of trade, while many remote southern hemisphere countries are small. Effectively by breaking up Europe, we reduce world trade in the counterfactual. Nevertheless there are significant differences by commodity. For example it can be seen that equalizing export distance across countries causes Gas trade to collapse falling by 78%. In contrast Iron Ore trade increases by 48%. A detailed visual description of the changes in each market is given in panel B of Figures 1-8. Table 4 summarizes the actual and counterfactual exports share by broad region while Table 5 presents the actual and counterfactual export shares by country. [Tables 3 to 5 about here] 4.1 Food First we consider the pattern of change in world Food supply. As Europe and Central Asia alone accounts for over 50% of world Food imports with the two other major regions, East Asia and North America, have, respectively, a world share of 16% and 19% (Figure 2, Panel A), European and Central Asian countries have a clear locational advantage when it comes to food exports. Thus, in our experiment, European countries lose their world share of Food exports, which falls from 48% to 25% (Table 4 and Figure 2, Panel B). Most other regions gain exports shares but particularly South America and East Asia and the Pacific. Thus in this market there is a very clear locational advantage for the European and Central Asian region. As we shall see, this last result is very important when we come to look at how distance affects individual country s overall advantage or disadvantage in world natural resources markets. 8

4.2 Minerals Recall from the preceding discussion that Minerals imports are focused geographically on East Asia, particularly China, Japan and South Korea (Figure 4, Panel A). Panel B of Figure 4 shows the actual and counterfactual export patterns. It can be seen that Australia and the Americas (Brazil, Chile, Canada and the USA) dominate the world supply of Minerals, accounting for 42% of world Minerals exports. In the counterfactual there are large increases in exports in the southern hemisphere South America, Australia and South Africa roughly neutral impact in North America and a collapse of exports in Europe-Central Asia and East Asia. The shares of Australia, Brazil and Chile increase dramatically with the collective share doubling from 27% to 44% of world exports of Minerals (see Tables 4 and 5). Hence in the world Minerals market there is a relatively large dispersion between producers and their main destination markets and this dispersion has a very significant cost on the southern Minerals giants such as Australia and Brazil. This also suggests that these regions still stand to make substantial gains in terms of world market shares with new transport technologies and reduction in transport costs. 4.3 Iron Ore Iron Ore is a subset of Minerals that, as noted above, is mainly imported by East Asia, especially China. It is dominated on the supply side by Brazil and Australia, which mutually account for 61% of world exports (Figure 5). Because of this concentration of demand in East Asia, Brazil is much farther from the world s largest Iron Ore importers than Australia. So in relative terms Australia is close to the Iron Ore market. 8 In the counterfactual, because of its relative proximity to East Asia, Australia market share falls by approximately one third to 23% of world trade. Likewise European based exporters such as Sweden and Ukraine are driven out from the market. Brazil however almost doubles its export share from 31% to 59% of world exports. Similar results are found for the other Latin American countries Peru, Venezuela and Chile though their shares of world trade are very small. Moreover total Iron Ore trade flows increase in the counterfactual, suggesting that Brazil s distance from China is an important impediment to world Iron Ore trade. 8 In absolute terms the Iron Ore market is the most dispersed with trade weighed distance falling by 19% in our counterfactual experiment, as shown in Table 3. 9

The large changes in world Iron Ore trade shares reflect not only Brazil s rather large distance from its market but also the very large elasticity of Iron Ore trade with respect to distance of -1.96. The experiments thus endorse the popular view that Australia has benefited from its locational advantage to China and that Brazil is particularly disadvantaged, at least in the Iron Ore market. 4.4 Coal Whereas the world Coal market is relatively dispersed across northern hemisphere countries in Europe-Central Asia and East Asia, Panel B of Figure 6 shows that supply is very concentrated in the South with one country, Australia, accounting for 28% of the world exports. Removing any distance disadvantage increases Australia s share of world exports substantially to 47%. Hence with respect to Coal the popular notion that Australian resource exports have benefited from its proximity to Asia can be seen in a somewhat different light. For Coal, Australia s relative proximity to Japan and China does not offset the cost of remoteness from Europe. Moreover, unlike Iron Ore, there are apparently no major South American suppliers that would be in a position to expand substantially supply if distance disadvantages were removed. The results also show that South Africa faces a similar geographical disadvantage as Australia - though its Coal exports are substantially smaller. The two second largest exporters, however, China and Russia, have significant geographical advantages being relatively close to both East Asian and European-Central Asian markets. 4.5 Petroleum The demand for imported Petroleum is almost equally split between Europe-Central Asia (36%), North America (25%) and East Asia (29%) while the supply of Petroleum is essentially split equally between the Middle-East (32%) and Europe-Central Asia, including Russia, (34%) (Figure 7). The European-Central Asian exporters are therefore strongly advantaged by their geographical location within the largest Petroleum import market. This is confirmed by the counterfactual results which show that Saudi Arabia s share of world Petroleum exports increases from 15 to 25% and Kuwait s share increases from 4 to 7% in the counterfactual case. Overall the Middle-East share rises to 46%, while Europe and Central Asia s share 10

falls to 25% (Table 4). Thus the European-Central Asian exporters have a significant advantage relative to the Middle-East. The exporters on the American continent, though small on the world market, nevertheless are also shown to be gaining significantly from their proximity to the USA. 4.6 Gas It can be seen immediately in Figure 8 that the Gas market lacks the South-North pattern seen in the preceding commodity markets due to significant Gas exporters in the northern hemisphere. The exception is the Australia-Indonesia-Malaysia East Asia LNG corridor which accounts for 17% of world Gas exports. Central to the Gas market is the issue of the delivery mode. Over a short distance pipelines are the cheapest mode of transport, while over long distances, shipping liquified gas (LNG) is the only viable solution. Indeed, estimating the model separately for LNG and for Gaseous gas, we find that the elasticity of distance for Gaseous Gas (-5.7) is approximately double the LNG elasticity (-2.7) (Table 6). 9 In either case however the elasticity of trade to distance is much higher than for other commodities. Thus countries that share a border with a large Gas importer benefit significantly from their geographical location, with notably Canada and the USA market and, to a smaller extent, Algeria and the European market. At the other extreme the Middle-Eastern countries, in particular Saudi Arabia and Qatar, but also Australia, suffer from their incapacity to be connected via pipelines to large import markets. Our counterfactual of removing distance advantages or disadvantages would lead to major changes in the Gas market, with current big players, such as Canada and Algeria, to be almost completely eliminated from the market. Conversely Australia s world exports share rise from 3% to 13% and Saudi Arabia and Qatar s world share would more than double. This is of interest given that there have been recent technological advances that have lowered the cost of transporting LNG (Ruester 2010). Consequently there is also evidence that the elasticity of trade to distance has started dropping significantly in the LNG sector (see Table 7). [Tables 6 and 7 about here] 9 To maximise the sample size and to ensure that as many major natural resource exporters are included in our sample, we use SITC-1 classification throughout this paper. However, for Gas, as the SITC-1 classification does not distinguish between LNG and Gaseous Natural Gas, we use SITC-3 classification for this set of results. 11

5 The Tyranny of Distance Given the preceding discussion it is clear that a country s location has an important effect on its volume of resource trade. Brazil probably exports much more Petroleum and much less Minerals than it would if it were located elsewhere. Australia is close to Asia and this is typically regarded as being very advantageous in terms of resource exports. We have found that this is true for Iron Ore but Australia still suffers a tyranny of distance in terms of Coal. How then do these costs and benefits of location add up for each country? Specifically, which countries face the greatest disadvantage of location in resources? Table 5 reports the total percentage change in resources trade for each country as a result of removing the relative distance barriers across countries. The countries with the largest gains are thus the countries that currently suffer the greatest competitive disadvantage from their geographical location. For each country we also report the composition of this change by commodity, so we can see which particular markets contribute to that country s remoteness. Figure 9 shows this information for all the countries where the total increase in trade exceeds 10%. There are 11 such countries out of a sample total of 104. It can be seen that the countries at the greatest disadvantage are the antipodean countries: Australia, New Zealand and the South American and South African resource exporters. For the 5 most disadvantaged countries the losses are very significant, exceeding 30%. Chile and New Zealand are particularly affected with a loss of around 50%. That is Chile and new Zealand s resources trade flows are predicted to be around 50% larger if their distance to exports markets were at the world average for each commodity. 10 To convert these quantity changes to ad valorem tax equivalents, a useful rule of thumb is that the elasticity of demand is approximately -1/2 across a wide range of commodities (Deaton 1974, Clements 2008). Thus the change in export quantities of around 50% for Chile and New Zealand implies that the cost disadvantage of distance, relative to the average country, is approximately the equivalent of a 100% export tax. Likewise the magnitude for South Africa, Brazil and Australia is similar, with a handicap equivalent to approximately a 70% export tax across their resources trade as a whole. 10 These values are all percentage changes. In terms of absolute changes the countries that stand the most are Australia in terms of Minerals, Brazil in terms of Coal and the USA in terms of everything else. Thus the absolute gains and losses tend to reflect the country s size in the market. In general Brazil, Australia and the USA are the biggest remote countries. 12

These values however are relative to a country that suffers no particular advantage or disadvantage. Kenya is an example of such a country (Table 5). The disadvantage of these antipodean countries relative to Kenya is much smaller than their disadvantage relative to the least remote countries. For convenience Figure 10 thus shows the 20 countries that experience the greatest fall in their resource trade volumes in the counterfactual. These countries have the opposite of a tyranny of distance, that is, the benevolence of proximity. It can be seen that, in our counterfactual, trade volumes fall by around 70% in these countries. Thus we can reconsider the disadvantage of distance implied by our experiments by comparing for example New Zealand with a well-located economy such as France. In France the total percentage loss in trade is 70% whereas for New Zealand the total gain is 46%. This suggests that the combination of implicit tax on New Zealand producers and an implicity subsidy to French producers adds up to a 106 percent change in volumes and a 212% ad valorem tax disadvantage for New Zealand producers relative to France. In general the results show that many country, mostly European, benefit from an enormous locational advantage over more remote countries. This advantage is often in the dimension of a subsidy of over 100%. It can also be seen that the reasons behind these changes differs by country. For New Zealand the key factor is its distance from European Food markets. For Chile and Brazil the distance from world Minerals demand is the key component. As we have seen above this is due to their distance from Asia relative to other countries. Likewise Australia and South Africa have a very similar pattern in terms of the composition of their distance disadvantage with Food, Minerals and also Coal being similarly important for each country. It can be seen further that Argentina is not as remote overall as, for example, Chile and Brazil. This is in part because it exports Petroleum and Gas to North America, and in part because minerals is a relative small component of its exports, so it is not affected by the distance to Asia in the same way Brazil is. Finally for the least remote countries we can see in, Figure 9, that their advantage is mostly in Food and Petroleum. Thus the Slovak Republic, Korea, Algeria, Norway and Mexico are shown to have particular luck in terms of having both Petroleum endowments and proximity to large markets. In each case the benevolence of proximity adds around 30-50% to their Petroleum export sales. Two exceptions from the general pattern are Poland and the Czech Republic which also benefit considerably from having coal deposits and being within Europe. Similarly Canada and Algeria have significant advantages in terms of the world Gas market. 13

[Figures 9 and 10 about here] 6 Conclusion While a broad literature exists on the importance of geography in explaining the volume of trade across countries, little is known on the importance of geography in explaining the predominance of some countries in resource exports. In contrast to manufacturing trade we have find that geography plays a very important role in explaining trade in commodities. This is due to both the relatively high transports costs and the fact that resource export supply is limited by natural endowments. We show first that many resources, particularly Iron-Ore and Gas, have very large elasticities of trade with respect to distance. However the impact of geography also depends the geographic separation of the export sources and the geographic distribution of demand. Thus we also show show that equalizing distances to markets would have large effects in some markets and on a country shares of world resource exports. In particular we found that the several southern resource exporting countries are significantly disadvantaged by their location. However Southern Food producers such as New Zealand were also found to have a large disadvantage, while many Latin American petroleum exporters were found to have a smaller total disadvantage due to the proximity to the USA. Finally, we also show that the impact of distance on exports is very large. The counterfactual resulted in an increase in export sales in the range of 35-50% for several countries. This implies the equivalent of an export tax in the order of 70-100%. Likewise for many countries the location advantage is very large. Especially the changes in trade volumes in the counterfactual for many countries are also consistent with subsidies exceeding 100%. Consequently, we conclude that the ability of many countries to be competitive in the world resource markets derives, to a large extent, from their location. This suggests that the responsiveness of extraction firms to national policies, for example with respect to resource taxation and regulation, may be quite low, particularly in Europe. Nevertheless the results also show that this advantage differs by sector, due to the differing geographical distribution of world demand by commodity. Finally the results also suggest that future technological changes that reduce transport costs have the potential to substantially alter country market shares in world resource markets. 14

Appendix A: Country Coverage by Region [Table A.1 about here] Appendix B: Definition of Variables [Table B.1 about here] Appendix C: Average Distance [Table C.1 about here]

Figure 1: Trade Shares: All Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 2: Trade Shares: Food Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 3: Trade Shares: Raw Materials Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 4: Trade Shares: Minerals Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 5: Trade Shares: Iron Ore Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 6: Trade Shares: Coal Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 7: Trade Shares: Petroleum Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 8: Trade Shares: Gas Import Shares Export Shares Source: WITS/COMTRADE data for selected countries in 2006, SITC-1 classification. Authors calculation.

Figure 9: The Tyranny of Distance: Percentage Change of Export Values when Location Advantages/Disadvantages Are Removed Chile New Zealand South Africa Brazil Australia Paraguay Uruguay Mauritius Peru Zambai Argentina -20 0 20 40 60 % Change of Export Values Food Raw Minerals Coal Petrol Gas

Figure 10: The Benevolence of Proximity: Percentage Change of Export Values when Location Advantages/Disadvantages Are Removed Italy Malaysia United Kingdom Mexico Norway Canada Ireland France Croatia Slovenia Poland Denmark Hungary Algeria Germany Korea Austria Netherlands Czech Rep. Switzerland Slovak Rep. -80-60 -40-20 0 % Change of Export Values Food Raw Minerals Coal Petrol Gas

Table 1: Regional Share of World Imports in 2006 Europe and North America South America Middle-East East Asia and South Asia Sub-Saharan Central Asia the Pacific Africa All 44.23 21.78 5.57 2.51 23.26 1.56 1.09 Food 52.96 16.15 5.42 4.65 18.80 0.58 1.44 Raw 41.92 13.61 5.25 3.10 31.95 3.02 1.16 Min. 38.30 6.72 3.52 1.12 46.30 3.76 0.28 Iron Ore 22.51 2.89 2.35 1.03 70.91 0.20 0.11 Coal 40.62 6.83 5.03 2.14 37.38 7.44 0.55 Petrol 35.88 25.33 3.61 1.48 29.22 3.21 1.27 Gas 34.57 26.85 5.64 1.96 29.06 1.81 0.11 Manuf. 44.89 22.13 6.02 2.54 22.14 1.26 1.03 Source: COMTRADE/WITS data. Authors calculation.

Table 2: PPML Results (1) (2) (3) (4) (5) (6) (7) (8) (9) All Food Raw Min. Iron Ore Coal Petrol Gas Manuf. ln gdp 1.502*** 0.479*** 0.458** -0.406* -0.008 0.792** 0.904*** 0.212 1.638*** (0.000) (0.000) (0.025) (0.093) (0.991) (0.016) (0.000) (0.796) (0.000) p ln gdp 1.289*** 0.803*** 1.553*** 2.384*** 3.032*** 1.089*** 1.006*** -0.240 1.342*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.006) (0.001) (0.746) (0.000) ln pop 0.330-1.274*** -1.214*** 2.109*** 2.117 0.115-0.925*** 1.029-0.828** (0.249) (0.000) (0.008) (0.005) (0.268) (0.942) (0.002) (0.150) (0.028) p ln pop -0.846*** -0.126-0.863* -1.714** -1.754 0.176 0.926 6.935** -1.004*** (0.002) (0.703) (0.083) (0.021) (0.119) (0.863) (0.264) (0.022) (0.000) ln distwces -0.767*** -0.968*** -0.864*** -0.959*** -1.964*** -1.220*** -1.372*** -2.557*** -0.752*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) ln land paire -0.090-0.027-0.141** 0.043-0.135 0.082 0.095 0.110-0.139*** (0.112) (0.620) (0.018) (0.613) (0.572) (0.742) (0.397) (0.540) (0.004) landlocked 1 0.790-0.410 3.827* 8.455*** 8.077* 2.707 3.806 9.772 0.679 (0.554) (0.802) (0.083) (0.007) (0.065) (0.553) (0.255) (0.202) (0.602) landlocked 2 1.918-0.178 8.302* 17.431*** 15.205* 7.225 9.652 19.490 1.571 (0.474) (0.957) (0.061) (0.005) (0.081) (0.430) (0.149) (0.205) (0.547) contig 0.264*** 0.248*** 0.358*** 0.590*** 0.839** 0.361 0.158 0.596* 0.194** (0.000) (0.006) (0.000) (0.000) (0.013) (0.139) (0.348) (0.051) (0.011) island 1-1.519-0.378 0.437 3.789 8.866 1.848 6.708 25.882* -2.306 (0.359) (0.843) (0.876) (0.381) (0.163) (0.757) (0.139) (0.058) (0.158) island 2-2.717-0.559 0.790 7.965 17.876 4.215 13.254 52.465* -4.346 (0.408) (0.882) (0.887) (0.358) (0.159) (0.723) (0.146) (0.055) (0.179) comlang ethno 0.305*** 0.362*** 0.115 0.236* 0.169 0.006 0.549*** -0.087 0.306*** (0.000) (0.001) (0.298) (0.055) (0.556) (0.979) (0.004) (0.746) (0.000) colony -0.101 0.179 0.101 0.411*** 1.734*** 0.382* 0.228 0.419-0.142 (0.324) (0.147) (0.430) (0.005) (0.000) (0.073) (0.241) (0.201) (0.162) hum k 0.209*** 0.177*** 0.096* 0.041 0.203 0.189* 0.029 0.350 0.197*** (0.000) (0.000) (0.088) (0.456) (0.306) (0.089) (0.716) (0.266) (0.000) p hum k 0.093*** 0.015 0.068 0.180*** -0.016 0.004-0.038-0.117 0.118*** (0.000) (0.487) (0.244) (0.000) (0.854) (0.968) (0.417) (0.400) (0.000) Constant -61.765*** -13.704** -31.129*** -49.403*** -73.869** -45.247* -39.501*** -19.879-59.360*** (0.000) (0.017) (0.000) (0.000) (0.022) (0.052) (0.000) (0.473) (0.000) Observations 74,984 74,984 74,984 74,984 74,984 74,984 74,984 74,984 74,984 Pseudo R2 0.952 0.908 0.897 0.870 0.932 0.892 0.859 0.902 0.963 Notes: Country, partner country and year fixed effects are controlled for. Standard-errors are adjusted for clustering at the country-pair level. ***, p-value< 0.01; **, p-value< 0.05; *, p-value< 0.10.

Table 3: Distance and Total Trade Unweighted Distance Weighted Distance Trade Value Actual Count. Change in % Actual Count. Change in % Actual Count. Change in % All 7726 7730 0.05 7310 7703 5.37 8245530 5445163-33.96 Food 7374 7381 0.09 7216 7385 2.34 493876 310746-37.08 Raw 7762 7768 0.07 7943 7748-2.46 193021 140350-27.29 Min. 7927 7939 0.15 8858 7932-10.45 156250 133808-14.36 Iron Ore 8594 8617 0.26 10589 8618-18.61 37322 55200 47.90 Coal 7859 7866 0.08 8896 7856-11.69 52327 51711-1.18 Petrol 8086 8087 0.00 7462 8084 8.33 867712 444992-48.72 Gas 8061 8062 0.01 7475 8054 7.75 113589 24685-78.27 Manuf. 7714 7717 0.04 7251 7687 6.00 6191161 4122514-33.41

Table 4: Regional Share of World Exports in 2006: Actual and Counterfactual Europe and North America South America Middle-East East Asia and South Asia Sub-Saharan Central Asia the Pacific Africa Actual Count. Actual Count. Actual Count. Actual Count. Actual Count. Actual Count. Actual Count. All 41.40 28.98 14.93 16.41 6.16 7.58 4.21 4.91 31.14 38.78 1.28 1.77 0.88 1.57 Food 47.82 25.24 15.50 19.03 14.48 24.94 1.45 1.24 16.68 23.20 1.67 2.19 2.41 4.16 Raw 35.25 19.91 25.90 29.68 11.00 19.92 1.12 0.90 23.28 25.18 1.06 1.00 2.39 3.41 Min. 29.15 12.70 15.18 15.41 19.77 35.08 3.19 2.25 24.51 25.03 3.63 2.64 4.57 6.89 Iron Ore 10.55 0.99 6.17 3.26 34.41 65.04 0.29 0.06 31.87 23.09 12.57 2.26 4.13 5.31 Coal 20.11 5.69 13.52 12.21 6.49 8.20 0.22 0.08 51.42 61.50 0.02 0.01 8.22 12.30 Petrol 34.07 25.15 6.39 3.48 12.91 12.42 32.06 45.74 12.47 9.45 0.67 0.92 1.43 2.85 Gas 24.86 13.42 27.70 4.86 4.63 6.81 24.99 41.10 17.78 33.51 0.01 0.01 0.04 0.29 Manuf. 42.50 29.59 15.04 16.75 4.33 4.79 1.22 1.33 35.15 44.81 1.26 1.76 0.51 0.96

Table 5: Percentage Change of Export Values When Locational Advantages/Disadvantages Are Removed Rank Country Food Raw Minerals Coal Petrol Gas Total 1 Chile 14.61 5.31 30.53 0.01 0.13-0.21 50.38 2 New Zealand 33.48 10.24 0.64 1.40 0.73 0.00 46.50 3 South Africa 8.41 1.85 12.57 12.77 2.38 0.19 38.17 4 Brazil 12.61 6.97 16.31 0.02 2.24 0.01 38.15 5 Australia 9.35 2.36 6.19 14.48 3.31-0.71 34.99 6 Paraguay 6.93 19.61 0.55 0.00-0.25-0.05 26.79 7 Uruguay 16.68 7.07 0.38 0.00-0.75-0.03 23.35 8 Mauritius 22.35 0.32 0.23 0.00 0.03 0.01 22.93 9 Peru 4.24 1.36 21.03 0.01-3.73-0.11 22.80 10 Zambia 4.28 2.94 6.65 0.00 0.15 0.05 14.07 11 Argentina 12.43 7.12 2.23 0.00-6.10-3.71 11.97 12 Guyana -1.35 0.67 8.46 0.00 0.00 0.00 7.77 13 Gabon -0.01 0.23 0.57 0.00 5.33 0.00 6.13 14 Tanzania 2.59 0.14 0.77 0.00 0.30 0.00 3.79 15 Burundi 0.58-0.02 0.36 0.00-0.03 0.00 0.90 16 Kenya -0.34-0.51 0.00 0.00 1.23 0.01 0.40 17 Cameroon -1.23-1.01 0.00 0.00 0.83 0.00-1.42 18 Uganda -3.07-0.72 0.00 0.00 0.03 0.00-3.76 19 Maldives -3.67-0.28-0.31 0.00 0.00 0.00-4.26 20 Ghana -5.19-0.25 0.47 0.00-0.10-0.03-5.10 21 CAF -1.15-4.50-0.02-0.02 0.00 0.00-5.69 22 Cote d Ivoire -4.85-0.15 0.03 0.00-1.07-0.02-6.06 23 Benin -1.62-2.93 0.01 0.00-1.93-0.17-6.64 24 Mali -0.76-7.51-0.02 0.00-0.03 0.00-8.32 25 Yemen -0.65-0.06-0.07 0.00-7.53-0.06-8.36 26 Senegal -6.93-0.73 0.15 0.00-2.46-0.18-10.15 27 Jamaica -11.02 0.00 4.97 0.00-4.38-0.03-10.46 28 Sudan -1.38-1.31-0.07-0.01-8.80-0.06-11.61 29 Gambia -10.54-2.15 0.06 0.00-0.12 0.00-12.74 30 United States -4.21 0.01-0.54-0.02-6.84-2.85-14.44 31 Costa Rica -15.01 0.52 0.24 0.00-0.31-0.13-14.68 32 Bolivia 3.02 3.75 8.29 0.00-2.50-29.52-16.97 33 Kuwait -0.04-0.01-0.18 0.00-18.29-1.97-20.50 34 Niger -1.04-0.21-0.14 0.00-19.75-0.06-21.20 35 Saudi Arabia -0.09-0.02-0.16 0.00-19.14-2.02-21.45 36 Colombia -2.34 0.49 0.21 4.69-24.50-0.07-21.53 37 Ecuador -0.36 0.70 0.09 0.00-22.62-0.24-22.42 38 Honduras -23.17-0.43 0.79 0.00-0.16-0.34-23.31 39 Nicaragua -18.97-0.01 0.29 0.00-4.99 0.00-23.70 40 Qatar -0.01 0.00-0.08 0.00-13.16-13.57-26.82 41 India -6.18-2.69-10.48-0.01-7.88-0.02-27.25 42 Panama -9.72 0.09 0.70 0.07-17.58-1.35-27.79 43 El Salvador -23.01-0.07 0.34 0.00-4.91-0.95-28.60 44 Pakistan -13.33-6.07-2.21-0.02-7.34 0.00-28.98 45 Guatemala -21.35-0.59 0.21 0.00-8.14-0.13-30.00 46 Iran -0.69-0.18-0.53 0.00-28.08-0.88-30.36 47 Belize -24.54-0.14 0.04 0.00-8.31-0.03-32.99 48 Thailand -14.70-6.66-1.40 0.00-9.62-1.03-33.41 49 Barbados -2.25 0.06 0.16 0.00-31.71-0.09-33.82 50 Philippines -11.76-5.97-12.73 0.00-5.74-0.92-37.11

Rank Country Food Raw Minerals Coal Petrol Gas Total 51 Egypt -4.99-1.81-1.95-0.82-19.35-10.07-38.99 52 Indonesia -2.23-2.83-3.93-0.40-18.00-12.10-39.49 53 Iceland -37.09-1.45-0.80-0.01-0.15-0.01-39.51 54 Kyrgyz Rep. -11.94-9.36-18.62-0.01-0.39 0.00-40.32 55 Kazakhstan -1.57-0.58-4.79-1.43-30.64-1.44-40.47 56 Israel -20.06-7.03-6.90-0.01-10.06-0.01-44.07 57 Trinidad -0.25 0.01 0.15 0.00-26.21-17.89-44.20 58 Venezuela -0.13 0.01 0.59 0.23-44.40-0.92-44.62 59 Syria -3.80-1.83-0.69-0.01-38.84 0.00-45.17 60 Jordan -10.01-1.23-33.90 0.00-0.80-0.16-46.10 61 Morocco -26.23-1.95-12.08 0.00-5.76-0.24-46.26 62 Cyprus -29.34-2.69-10.47-0.30-6.93-0.46-50.18 63 Vietnam -10.27-1.50-0.80-1.22-36.58-0.01-50.38 64 China -18.76-4.13-4.28-8.91-15.17-0.08-51.33 65 Turkey -29.50-4.26-8.34-0.03-8.67-0.53-51.33 66 Russia -1.88-2.04-2.11-1.88-37.54-7.08-52.53 67 Malta -9.93-1.31-1.69-0.02-40.09-0.38-53.43 68 Portugal -24.10-11.80-7.29-0.10-9.81-0.52-53.62 69 Greece -28.53-9.09-4.51 0.00-10.76-1.64-54.54 70 Finland -7.43-22.56-3.72-0.19-20.60-0.10-54.61 71 Spain -37.19-5.95-3.29-0.42-5.78-2.39-55.03 72 Japan -4.80-10.78-17.33-1.93-19.89-0.58-55.31 73 Ukraine -15.22-6.49-15.03-3.76-12.85-2.34-55.69 74 Singapore -2.59-1.06-1.24-0.05-51.32-1.29-57.55 75 Tunisia -8.89-7.94-5.32-0.01-35.56 0.00-57.71 76 Romania -10.51-13.33-11.30-0.10-24.36-0.54-60.14 77 Albania -16.56-16.16-18.62-0.02-9.16 0.00-60.51 78 Bulgaria -23.50-8.42-8.59-0.16-21.70-0.09-62.46 79 Latvia -5.69-15.09-1.46-2.04-38.03-0.19-62.50 80 Estonia -8.57-9.64-4.41-2.28-38.18-0.02-63.09 81 Sweden -12.18-20.59-8.58-0.14-21.26-0.45-63.21 82 Macao -27.85-9.11-23.09 0.00-3.72 0.00-63.77 83 Lithuania -12.82-5.17-5.10-1.04-38.48-1.52-64.13 84 Italy -37.11-6.37-3.16-0.11-17.93-0.81-65.49 85 Malaysia -4.30-10.75-1.28 0.00-36.43-12.76-65.51 86 United Kingdom -16.96-2.25-3.55-0.17-40.24-3.49-66.66 87 Mexico -9.90-0.45-0.23 0.00-56.56-0.24-67.37 88 Norway -4.88-0.66-1.27-0.14-53.55-7.31-67.81 89 Canada -10.19-6.99-2.03-0.62-26.30-22.35-68.48 90 Ireland -56.54-2.55-7.25-0.66-2.49-0.15-69.63 91 France -48.12-5.80-4.43-0.12-10.74-1.53-70.73 92 Croatia -22.16-15.57-9.67-0.09-18.11-5.32-70.92 93 Slovenia -34.31-21.99-12.85-0.13-2.79-0.08-72.14 94 Poland -32.49-5.61-6.11-20.37-7.75-0.06-72.38 95 Denmark -40.90-6.30-2.02-0.05-21.98-1.34-72.59 96 Hungary -41.99-8.55-4.57-0.58-17.81-0.46-73.98 97 Algeria -0.08-0.02-0.31 0.00-46.48-27.33-74.22 98 Germany -38.64-9.15-6.68-0.62-14.55-4.62-74.25 99 Korea -8.80-6.34-2.07-0.02-57.93-1.05-76.20 100 Austria -37.57-18.70-9.26-0.07-9.30-1.56-76.46 101 Netherlands -38.09-10.54-4.13-0.37-22.83-2.06-78.02 102 Czech Rep. -25.42-15.76-9.73-15.60-12.22-0.43-79.16 103 Switzerland -39.48-8.95-15.75-2.13-12.89-0.96-80.17 104 Slovak Rep. -14.20-8.87-6.45-0.43-52.82-0.15-82.92

Table 6: PPML Results: Natural Gas (1) (2) (3) Nat. Gas LNG Gaseous ln gdp 0.184 1.331-3.303 (0.873) (0.235) (0.156) p ln gdp 0.390 1.774 0.126 (0.772) (0.224) (0.956) ln pop 0.836-0.052-3.576 (0.349) (0.949) (0.648) p ln pop 8.388** 11.381* 4.609 (0.044) (0.094) (0.463) ln distwces -3.713*** -2.672*** -5.674*** (0.000) (0.000) (0.000) ln land paire 0.110-0.332 0.324 (0.699) (0.183) (0.624) landlocked 1 11.216** 16.958*** 10.299 (0.013) (0.003) (0.226) landlocked 2 22.200** 39.743*** 19.520 (0.015) (0.000) (0.253) contig 0.780 0.709-0.708 (0.150) (0.605) (0.204) island 1 27.179** 38.292* 18.686 (0.023) (0.051) (0.297) island 2 54.999** 76.786** 42.732 (0.021) (0.050) (0.232) comlang ethno 0.145-0.312 0.545 (0.754) (0.514) (0.484) colony 0.558 1.471* -0.912 (0.241) (0.096) (0.211) hum k 0.291-0.318 0.029 (0.453) (0.327) (0.936) p hum k -0.057-0.654 0.091 (0.739) (0.144) (0.594) Constant -30.152-97.625** 79.478 (0.455) (0.018) (0.335) Observations 52,876 52,876 52,876 Notes: WITS/COMTRADE DATA using SITC-3 classification, for the period 1999-2003 and 2006. Country, partner country and year fixed effects are controlled for. Standard-errors are adjusted for clustering at the country-pair level. ***, p-value< 0.01; **, p-value< 0.05; *, p-value< 0.10. To allow the impact of distance to vary over time, interaction variables between distance and time are included.

Table 7: PPML Results: Natural Gas over Time (1) (3) (4) (5) All Nat. Gas LNG Gaseous ln gdp 1.203*** -0.100 0.764-2.145 (0.000) (0.928) (0.393) (0.283) p ln gdp 0.821*** 0.688-0.160-0.137 (0.000) (0.625) (0.928) (0.950) ln pop -0.260*** 0.711-0.263-3.133 (0.000) (0.420) (0.700) (0.673) p ln pop -0.016 9.210* 29.756*** 5.187 (0.845) (0.062) (0.003) (0.419) ln distwces -0.782*** -3.696*** -3.185*** -5.583*** (0.000) (0.000) (0.000) (0.000) ln land paire -0.084 0.109-0.329 0.329 (0.142) (0.699) (0.184) (0.620) landlocked 1 0.438 12.825** 34.200*** 9.856 (0.544) (0.038) (0.002) (0.231) landlocked 2 1.180 25.417** 74.251*** 18.631 (0.416) (0.041) (0.001) (0.260) contig 0.237*** 0.779 0.605-0.699 (0.001) (0.149) (0.633) (0.211) island 1 0.257** 30.586** 96.711*** 19.781 (0.032) (0.046) (0.004) (0.282) island 2 0.876*** 61.813** 193.615*** 44.877 (0.000) (0.044) (0.004) (0.220) comlang ethno 0.306*** 0.144-0.351 0.540 (0.000) (0.755) (0.461) (0.489) colony -0.105 0.555 1.399* -0.902 (0.304) (0.244) (0.093) (0.217) hum k 0.172*** 0.342 0.218 0.056 (0.000) (0.408) (0.610) (0.892) p hum k 0.119*** -0.072-0.046 0.037 (0.000) (0.688) (0.935) (0.856) dist y2000 0.018*** 0.030 0.137 0.043 (0.000) (0.586) (0.220) (0.751) dist y2001 0.008-0.143 0.129-0.115 (0.151) (0.131) (0.361) (0.563) dist y2002-0.008-0.095 0.201-0.029 (0.343) (0.283) (0.188) (0.901) dist y2003-0.002-0.033 0.288 0.114 (0.847) (0.783) (0.122) (0.688) dist y2006 0.006 0.064 0.849** -0.310 (0.671) (0.777) (0.011) (0.552) Constant -38.905*** -33.642-108.177*** 60.250 (0.000) (0.439) (0.001) (0.422) Observations 52,876 52,876 52,876 52,876 Notes: WITS/COMTRADE DATA using SITC-3 classification. Country, partner country and year fixed effects are controlled for. Standard-errors are adjusted for clustering at the country-pair level. ***, p-value< 0.01; **, p-value< 0.05; *, p-value< 0.10.