Relatedness, Knowledge Diffusion, and the Evolution of Bilateral Trade
RRelatedness, Knowledge Di ff usion, and the Evolution of BilateralTrade Bogang Jun a , Aamena Alshamsi b , Jian Gao a,c , C´esar A. Hidalgo a,1 a Collective Learning Group, MIT Media Lab, Massachusetts Institute of Technology, Cambridge, MA 02139, USA b Masdar Institute of Science and Technology, Abu Dhabi, UAE c CompleX Lab, Web Sciences Center and Big Data Research Center, University of Electronic Science and Technology ofChina, Chengdu 611731, China.
Abstract
During the last decades two important contributions have reshaped our understanding of inter-national trade. First, countries trade more with those with whom they share history, language,and culture, suggesting that trade is limited by information frictions. Second, countries are morelikely to start exporting products that are similar to their current exports, suggesting that knowl-edge di ff usion among related industries is a key constrain shaping the diversification of exports.But does knowledge about how to export to a destination also di ff uses among related productsand geographic neighbors? Do countries need to learn how to trade each product to each destina-tion? Here, we use bilateral trade data from 2000 to 2015 to show that countries are more likelyto increase their exports of a product to a destination when: (i) they export related products to it,(ii) they export the same product to the neighbor of a destination, (iii) they have neighbors whoexport the same product to that destination. Then, we explore the magnitude of these e ff ects fornew, nascent, and experienced exporters, (exporters with and without comparative advantage ina product) and also for groups of products with di ff erent level of technological sophistication.We find that the e ff ects of product and geographic relatedness are stronger for new exporters,and also, that the e ff ect of product relatedness is stronger for more technologically sophisticatedproducts. These findings support the idea that international trade is shaped by information fric-tions that are reduced in the presence of related products and experienced geographic neighbors. Keywords:
International trade, Collective learning, Economic complexity, KnowledgeDi ff usion
1. Introduction
For more than a century, the literature on international trade explained global commerce as aconsequence of di ff erences in factor endowments Heckscher and Ohlin (1991), product quality,and product di ff erentiation (Krugman, 1979, 1991; Anderson, 1979; Helpman, 1987). Morerecent streams of literature, however, have shown that there is more to international trade thanendowments, costs, and distance, since countries need to learn how to produce and export each ∗ Email address : [email protected]
Preprint submitted to arXiv September 19, 2017 a r X i v : . [ q -f i n . E C ] S e p roduct (Hidalgo et al., 2007; Hidalgo, 2015), and also, need to overcome important informa-tion frictions to enter each export destination (Rauch, 1999, 2001; Rauch and Trindade, 2002;Petropoulou, 2008; Portes and Rey, 2005; Casella and Rauch, 2002; Anderson and Marcouiller,2002; Garmendia et al., 2012).During the last two decades scholars have documented that volumes of bilateral trade de-crease with the presence of borders (McCallum, 1995), and increase with migrants, shared lan-guage, and social networks (Rauch and Trindade, 2002; Rauch, 2001; Combes et al., 2005;Chaney, 2014; Bailey et al., 2017). In fact, using the random re-allocation of the Vietnameseboat people –a population of 1.4 million Vietnamese refugees reallocated in the U.S.–, Parsonsand V´ezina (2017) showed that states who received a 10% increase in their Vietnamese popula-tion experienced a growth in exports to Vietnam of between 4.5% and 14%.But the evidence in favor of knowledge di ff usion is not only expressed on aggregated tradeflows, since scholars have also shown the e ff ects of language, social networks, and informal insti-tutions to be larger for di ff erentiated products (Rauch, 1999, 2001; Rauch and Trindade, 2002).This suggests that factors limiting knowledge and information di ff usion (from social networks tolanguage) play a more important role in the di ff usion of the knowledge and information neededto exchange more sophisticated goods.A second stream of literature has focused on the supply side, in particular, on the process bywhich countries learn how to produce the products they export. This literature has shown that theability of countries and regions to enter new export markets is limited by knowledge di ff usion,since countries and regions are more likely to start exporting products when these are relatedto their current exports (Hidalgo et al., 2007; Hidalgo and Hausmann, 2009; Hausmann et al.,2014; Boschma et al., 2013), or when their geographic neighbors are already exporting them(Bahar et al., 2014). The importance of knowledge di ff usion in the diversification of economicactivities, however, is not limited only to the export of products. It has also been observed inthe development of regional industries (Ne ff ke et al., 2011; Gao et al., 2017), research activities(Guevara et al., 2016), and technologies (Kogler et al., 2013; Boschma et al., 2014), suggestingthat relatedness between economic activities facilitates knowledge di ff usion in general, not onlyin the context of international trade.Together, these findings have given rise to a more nuanced picture of international trade. Apicture where factor endowments and transportation costs do not determine trade fully, becauseinformation frictions and knowledge di ff usion determine the knowledge a country has, and hence,the products it can produce and the partners it can trade with.Here, we contribute to this literature by combining the stream of literature on knowledgeand information frictions and that on relatedness by exploring the path dependencies a ff ectingthe evolution of the network of export destinations for each product. We use more than 15years of bilateral trade data, disaggregated into more than 1,200 products, to construct a gravitymodel that validates previous findings and expands them. Looking at hundreds of thousandsof bilateral trade links reveals that: (i) countries are more likely to increase their exports of aproduct to a destination when they already export related products to it; (ii) countries are morelikely to increase their exports of a product to a destination when they already export to thatdestination’s neighbors, and (iii) countries are more likely to increase the exports of a product toa destination when their neighbors export that product to that destination. Moreover, we find thatsharing a colonial past, a language, a border, or bilingual speakers (when the two countries sharea language), is also associated with an increase in the volume of trade.Yet, only some of these findings are novel. The e ff ects of common language, shared border,and shared colonial past that we reproduce here have been documented in the past (McCallum,2995; Rauch, 1999). Also, we know that countries are more likely to start exporting to a desti-nation when they export to that destination’s neighbors (Chaney, 2014). What is novel, are (i)the e ff ect of relatedness on bilateral trade volumes, (ii) the e ff ect of having a neighbor export thesame product to the same destination, and (iii) the e ff ect of bilingual speakers. In particular, wefind that the e ff ects of relatedness–the finding that countries increase their exports of a product toa destination when they already export related products to it–are especially strong. In fact, a onestandard deviation increase in relatedness is associated with a 20% increase in exports in a twoyear period. This e ff ect is about 50% larger than the e ff ect of exporting that product to a neighborof the target destination (Chaney, 2014), and more than 170% larger than the e ff ect of having aneighbor export the same product to the same destination. The e ff ect of bilingual speakers, whilesignificant after considering all controls, are much smaller than that of sharing a language.We also study these e ff ects by separating exporters into new exporters, nascent, and expe-rienced exporters, by considering as new exporters those without comparative advantage in aproduct. We find the e ff ects of product relatedness, and especially those of geographic related-ness among exporters, to be stronger for new exporters than for experienced exporters. Also, wetest the hypothesis that knowledge di ff usion should a ff ect more strongly products that are knowl-edge intensive (Rauch, 1999, 2001; Rauch and Trindade, 2002) by dividing products into the fivetechnological categories suggested by Lall (Lall, 2000): primary, resource-based manufactures,low-tech, medium-tech, and high-tech products. We find that exporting related products has astronger e ff ect on the increase of exports for technological sophisticated products, suggestingthat knowledge di ff usion is more important for knowledge intense products. Surprisingly, wefind no e ff ect of technological sophistication on both of our measures of regional relatedness.That is, neither exporting to a neighbor, nor having a neighbor export the same product, appearsto have an e ff ect on future exports that either increases or decreases with technological sophisti-cation. Also, we find that sharing a language and a colonial past has a larger e ff ect on the increaseof exports for more technologically sophisticated products, providing further evidence that thee ff ects are due to information and knowledge frictions. Moreover, we find the negative e ff ectof distance–which are correlated with social network connections (Breschi and Lissoni, 2009;Singh, 2005)–to be larger for technologically sophisticated products. These findings supportthe idea that establishing trade relationships requires overcoming information and knowledgefrictions, and that product and geographic relatedness help reduce these frictions.
2. Data
We use bilateral trade data from 2000 to 2015 from MIT’s Observatory of Economic Com-plexity (Simoes and Hidalgo, 2011). The data is disaggregated into the Harmonized System (HSrev 1992, four-digit level) and consists of imports and exports between countries. Because bothexporter and importer report their trade information, we clean the data by comparing the data re-ported by exporters and importers following the work of Feenstra et al. (2005). Also, we excludecountries that have population less than 1.2 million or have a trade volume in 2008 that is below1 billion in US dollar. Also, we exclude data from Iraq, Chad and Macau.Macroeconomic data (GDP at market prices in current US dollar and population) comesfrom the World Bank’s World Development Indicators. Data on geographical and cultural dis-tance (shared language, physical distance between most populated cities, sharing a border, andshared colonial past) comes from GEODIST data from CEPII (Mayer and Zignago, 2011). Forlanguage proximity, we use one of the global language networks of Ronen et al. (2014): the3 roduct 1 Product 1
Product 1 Product 1
Same Product
Time = tTime = t + 2 (B) Importer Relatedness
Product 1 Product 1
Product 1 Product 1
Same Product
Time = tTime = t + 2 (C) Exporter Relatedness
Product 1 Product 2
Product 1 Product 2
Related Product
Time = tTime = t + 2 (A) Product Relatedness ! opd = X p pp p · x op d x od ⌦ ( d ) opd = X d /D dd /D d · x opd x op ⌦ ( o ) opd = X o /D oo /D o · x o pd x pd Figure 1: Relatedness among products, exporters, and importers. (A) Product Relatedness: the similarity between aproduct and the other products that a country already exports to a destination, (B) Importer Relatedness: the fraction ofthe geographic neighbors of a country that import a product from the same origin, and (C) Exporter Relatedness: thefraction of neighbors of a country that export a product to the same destination. one considering the number of books translated from one language to another as a proxy for thenumber of translators, or bilingual speakers, between two languages.
3. Results
Does relatedness among products or geographic neighbors help facilitate the knowledgeflows needed to increase bilateral trade flows?To explore this question we introduce three measures of relatedness. We use these to esti-mate: (i) the similarity between a product and the other products that a country already exportsto a destination (Product Relatedness), (ii) the fraction of the geographic neighbors of a countrythat import a product from the same origin (Importer Relatedness), and (iii) the fraction of neigh-bors of a country that export a product to the same destination (Exporter Relatedness). ProductRelatedness should help us capture information about knowledge flows between products (whichrange from knowledge flows among industries to knowledge flows among product lines within afirm). Figure 1(A) illustrates Product Relatedness in the context of Korea and Chile. In the ex-ample, Korea exports Products I and II to Chile (Shirts and Pants), and this may a ff ect the futureexports of Product III (Coats) to Chile, when Product III (Coats) is highly related to Products Iand II (Shirts and Pants). Our hypothesis is that knowledge flows should be larger among relatedproducts, and hence, exports should increase faster when a country exports related products to adestination.Importer Relatedness helps us capture knowledge flows on how to: (i) import a product fromthe same origin than a neighbor, or (ii) export to a neighbor of a current destination. In theexample of Figure 1(B), Korea exports Product I (Shirts) to Peru and Argentina and that maya ff ect the future volume of exports of Product I (Shirts) to Chile (who is a geographic neighborof Peru and Argentina). Here, knowledge on how to import from an origin should be flowing4mong neighboring importers, or knowledge on how to export to the neighbor’s of a country’sdestinations should be flowing within the exporter.Exporter Relatedness captures (i) knowledge flows among neighboring exporters on how toexport to a destination, or (ii) knowledge flows on how to import from a neighbor of a countryfrom where you currently import. In the example of Figure 1(C), Chile imports Product I (Shirts)from China and Japan, and that may a ff ect the future volume of exports of Product I (Shirts) fromKorea (which is a neighbor of the places from where Chile is currently importing Product I). Thiswould be a knowledge flow on how to export to a destination among neighboring exporters, or aknowledge flow within an importer, of how to import from a neighbor of a current origin.Mathematically, we can construct the three measures of relatedness using a similar formula.The formula is a weighted average of the number of neighbors, or related products, that alreadyhave an active trade relationships. In the case of similarity between products, weighs are theproximity between products p and p (cid:48) , φ pp (cid:48) . φ pp (cid:48) is the minimum of the conditional probabilitythat two products are co-exported by multiple countries (see Hidalgo et al. (2007) and AppendixA). φ pp (cid:48) = p and p (cid:48) are always co-exported and φ pp (cid:48) = / D dd (cid:48) and 1 / D oo (cid:48) ), where D dd (cid:48) is thedistance in kilometers between the most populated cities in countries d and d (cid:48) .Formally, let x opd be a matrix summarizing the trade flow in US dollars of product p fromexporter o to destination d . Then, Product Relatedness is given by: ω opd = (cid:88) p (cid:48) φ pp (cid:48) φ p · x op (cid:48) d x od (1)where x od is the volume of trade between countries o and d ( x od = (cid:80) p x opd ) and φ p = (cid:80) p (cid:48) φ pp (cid:48) .Similarly, Importer Relatedness is given by: Ω ( d ) opd = (cid:88) d (cid:48) / D dd (cid:48) / D d · x opd (cid:48) x op , (2)where x op is the volume of trade of product p from origin country o ( x op = (cid:80) d x opd ), D dd (cid:48) is thegeographic distance between destination country d and its neighbors d (cid:48) , and 1 / D d = (cid:80) d (cid:48) / D dd (cid:48) .Finally, Exporter Relatedness is given by: Ω ( o ) opd = (cid:88) o (cid:48) / D oo (cid:48) / D o · x o (cid:48) pd x pd , (3)where x pd is the volume of trade of product p to destination country d ( x pd = (cid:80) o x opd ), D oo (cid:48) isthe geographic distance between origin country o and its neighbors o (cid:48) , and 1 / D o = (cid:80) o (cid:48) / D oo (cid:48) .Next, we use these three measures of relatedness, together with data on common cultural andgeographic factors, to construct an extended gravity model to study the marginal contributionof product, importer, and exporter relatedness, and of shared languages, borders, and colonialpast, to the growth of future exports. Our model predicts bilateral trade in a product in two yearstime while controlling for: (i) initial trade in that product between the same trade partners, (ii)total exports of the product by the exporter, (iii) total imports of the product by the importer,5iv-vii) the GDP per capita and population of exporters and importers, and (viii) their geographicdistance. Formally, our model is given by Equation 4: x t + opd = β + β ω topd + β Ω ( d ) topd + β Ω ( o ) topd + β x topd + β x top + β x tpd + β D od + β gd p to + β gd p td + β Population to + β Population td + β Border od + β Colony od + β Language od + β Lang . Proximity od + ε topd (4)where the dependent variable, x t + opd , represents the volume of trade in (US dollar) of product p from exporter o to destination d in year t +
2. Our main variables of interest are our three mea-sures of relatedness: Product Relatedness ω topd , Importer Relatedness Ω ( d ) topd , Exporter Relatedness Ω ( o ) topd , and shared border ( Border od ), shared language ( Language od ), language proximity (num-ber of bilingual speakers Lang . Proximity od ), and shared colonial past ( Colony od ). Border od , Language od , Colony od are binary (dummy) variables (0 or 1). The other factors in the model gd p per capita, population, and distance ( D od ), are standard gravity controls (Tinbergen, 1962;P¨oyh¨onen, 1963). Finally, by incorporating the total volume of exports of a country ( x op ), thetotal imports of a destination ( x pd ), and the present day trade flow for each product between anorigin and a destination ( x opd ), we capture the e ff ects of our variable of interest in the change intrade experience in the subsequent two years. In Equation 4, we make all variables comparable(except binary variables) by standardizing them by subtracting their means and dividing them bytheir standard deviations. (Please see Appendix B and C for summary statistics and correlationamong variables)Table 1 shows our main results divided into three periods: 2000-2006 (pre-financial crisis),2007-2012 (crisis period), and 2012-2015 (recovery period). Since our result are qualitatively thesame for all of these periods we will describe them in unison. (See detailed results, correlationtable and summary statistics in Appendix B and C)First, we find that the three relatedness variables correlate positively with future bilateraltrade. This means that a country is likely to experience an increase in their exports of product p to a destination d when: (i) the country is exporting related products to that destination, (ii) itis exporting the same product to the neighbors of a destination (confirming Chaney (2014)), and(iii) it has neighbors that are already exporting the same product to that destination. This extendsBahar et al. (2014), who showed that having geographic neighbors increases the probability ofexporting a new product, since Bahar et al. (2014) did not look at individual export destinations(they aggregate across all destinations). Our findings, therefore, complement Bahar et al. (2014)by showing that having neighbors that export a product does not only increase the total volumeof exports, but the volume of exports to the same destinations that the neighbors were exportingto. When comparing the e ff ects of product and geographical relatedness (variables are standard-ized), we find that the role of Product Relatedness ( ω topd ) is on average the largest, while that ofExporter Relatedness ( Ω oopd ) is the smallest. In addition to these, we find strong and positive ef-fects for the role of shared borders, colonial past, shared language, and to a lesser extent, numberof translations (language proximity). Other standard gravity factors (distance, GDP per capita,and population) behave as expected. 6 able 1: Bilateral trade volume after two years for periods 2000-2006 (pre-financial crisis), 2007-2012 (crisis period)and 2012-2015 (recovery period) Dependent variable : log x t + opd ω topd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Ω ( d ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Ω ( o ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x topd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x top . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x tpd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Distance − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . gdp to . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . gdp td . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population o . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population d . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Border od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Colony od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Language od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Lang . Proximity od . ∗∗∗ . ∗∗∗ − . . . . . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Note: ∗ p < ∗∗ p < ∗∗∗ p < Together, the finding that relatedness among products, the presence of knowledge among ge-ographic neighbors, language, colonial history, shared borders, and language proximity, all havea positive and significant e ff ect in the increase of trade flows for particular products and coun-tries, are evidence in support of the notion that knowledge on how to trade a specific productbetween a specific pair of countries needs to flow for that trade to be materialized. If this hypoth-esis is correct, we should also be able to study the varying importance of knowledge flows fornew and experienced exporters (exporters with or without comparative advantage), and also, forproducts with di ff erent levels of technological sophistication.Next, we test the e ff ects of the exporters’ level of competitiveness in the di ff usion of theinformation needed to trade by dividing exporters of each product into new, nascent, and experi-enced. We do this by calculating the revealed comparative advantage (RCA) of each exporter ineach product. RCA is the ratio between the exports of a country in a product, and the exports that7 able 2: Bilateral trade volume after two years for new, nascent, and experienced exporters Dependent variable : log x t + opd New Exporter Nascent Exporter Experienced Exporter ω topd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Ω ( d ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Ω ( o ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x topd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x top . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . x tpd . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Distance − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . gdp to . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . gdp td . ∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population o . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population d . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Border od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Colony od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Language od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Lang . Proximity od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Note: ∗ p < ∗∗ p < ∗∗∗ p < are expected based on a country’s total export market and the size of the global market for thatproduct. We classify as new exporters all countries with an RCA below 0.2 in a product (coun-tries that export less than 20% of what they are expected to export by chance). We classify asnascent exporters, all countries with an RCA between 0.2 and 1. We classify as the experiencedexporters of a product, all countries that have revealed comparative advantage in it (RCA > ff ects of product and geographic relatedness, especially Exporter Relatedness, arestronger for new exporters, suggesting that knowledge and information frictions impose largerconstraints for countries that are not experienced in the export of a product. Second, the overallexplanatory power of the model is considerably larger for experienced exporters ( R ≈
53% vs R ≈
46% for nascent exporters and R ≈
28% for new exporters. These are large di ff erences,even considering that the sample sizes are not the same). This suggests that inexperienced ex-porters face more uncertainty (less predictable because of lower R-square), and hence, benefitmore from relatedness (higher relatedness coe ffi cients).Finally, we explore the interaction between our three measures of relatedness and the tech-nological sophistication of products using Lall (2000)’s five technological categories: primary,8 .100.150.20 Primary Resource-based Low-tech Medium-tech High-tech
Low technology High technology C o e f fi c i e n t s ⌦ ( d ) opd A ! opd R = 0 . ⇤ R = 0 . R = 0 . (B) Contiguous od Language od R = 0 . R = 0 . ⇤⇤⇤ Primary Resource-based Low-tech Medium-tech High-tech R = 0 . ⇤⇤ R = 0 . B R = 0 . R = 0 . ⇤⇤⇤ Language od Border od Colony od log Distance Low technology High technology ⌦ ( o ) opd -0.5 Figure 2: Coe ffi cients of variables by the technological sophistication of products; (A) Coe ffi cients of ω opd , Ω ( d ) opd ,and Ω ( o ) opd , (B) Coe ffi cients of Border od , Language od , and Colony od , and (C) Coe ffi cients of log lang . Proximity andlog
Distance . The fitted lines in red are statistically significant, while the lines in blue are not statistically significant. ∗ p < ∗∗ p < ∗∗∗ p < resource-based manufactures, low-tech, medium-tech, and high-tech products. Since Lall’s clas-sification is based on the 3-digit Standard International Trade Classification (SITC-3) rev 2, wematch products to our data using the conversion table provided by the UN Trade Statistics site . Following Lall (2000), we also exclude “special transactions” such as electric current, cinemafilm, printed matter, fold, coins, and pets.We present the coe ffi cients for all of the variables we had in the previous model in the Ap-pendix E (Table A15) and summarize the main results graphically, by plotting the coe ffi cientsand their errors as a function of the technological sophistication of products in Figure 2. Trendsthat increase significantly with technological sophistication ( p < .
1) are presented in red, non-significant trends are shown in blue.Figure 2 A shows that the e ff ect of Product Relatedness, but not that of Importer Relatednessor Exporter Relatedness, increases with technological sophistication. This suggests that productrelatedness captures channels of knowledge and information flow that are relevant for the exportof sophisticated products. Also, Figure 2 B shows that the e ff ect of sharing a language and acolonial past, but not those of sharing a border, are larger for more technologically sophisticatedproducts. Once again, this reiterates the idea that borders and geographic distance a ff ect knowl-edge flows by limiting social interactions (Singh, 2005; Breschi and Lissoni, 2009), so we do notsee much of a geographic e ff ect once we take cultural and linguistic similarity into account. Thenegative e ff ect of distance is slightly larger for technologically sophisticated products, but thee ff ect is not strong enough to be significant. Together, these findings support the idea that tradeis driven partly by the di ff usion of knowledge and information on how to export each product toeach destination.
4. Discussion
During the last decades two ideas have re-framed our understanding of international trade.The first idea is that information and knowledge frictions, not just di ff erences in transportation Available at https: // unstats.un.org / unsd / trade / classifications / correspondence-tables.asp ff erences in productivity, shape global trade. The second ideais that countries need to learn how to produce the products they export, and hence, evolve theirproductive structures in a path dependent manner that is constrained by knowledge flows. Here,we use bilateral trade data, together with various measures of economic size, culture, and geo-graphic proximity, to put these two ideas together. Our findings confirm many existing theoriesinvolving the role of language, and culture, but also, add to the body of knowledge by show-ing that relatedness among products and countries shape future trade volumes. In particular,we showed that relatedness among products, and among geographic neighbors, explains a sub-stantial fraction of future bilateral trade: trade volumes increase when countries export relatedproducts to a destination, but also, when they share neighbors who export to that destination, orwhen they are already exporting to a destinations neighbors. When comparing these three formsof relatedness, we found that relatedness among products is the strongest, suggesting that theremay be product or industry specific learning channels that play an important role in the di ff usionof the knowledge needed to establish or increase trade relationships. Moreover, we found the ef-fects of relatedness to be stronger for new exporters, and the e ff ects of product relatedness to bestronger for more technologically sophisticated products. These additional considerations sup-port the idea that the presence of related activities facilitates the knowledge flows that countriesneed to learn how to produce and export products to specific destinations.Yet, our results leave unanswered many questions about the mechanisms underlying theseknowledge and information flows. The two channels we observed among geographic neigh-bors (Importer and Exporter Relatedness) could further be disaggregated into four channels: theknowledge flows among neighboring importers or exporters, or the knowledge flows within anexporter or within an importer. From there, we may be able to start learning about the specificmechanisms that underlie each of these knowledge flows. Also, the interpretation of relatednesshas a similar problem. On the one hand, one cannot know if the flow of knowledge is amongproduct lines, within the same firm or industry, among industries, or the result of spin-o ff s, for-eign direct investment, or migrations.Nevertheless, our results do provide some light in the long quest to understand how socialnetworks, culture, and knowledge flows, shape international trade. They tell us that productrelatedness plays an important role since the size of its e ff ect is larger than the one observedamong geographic neighbors. This suggests that looking at knowledge flows among productlines, and among industries, should be an avenue of inquiry for improving our understanding ofthe social and economic forces that shape global trade. Acknowledgement
We thank Mauricio (Pacha) Vargas and Alex Simoes for help with the data. We also thankCristian Jara Figueroa, Fl´avio Pinheiro, Tarik Roukny, and Dogyoon Song for helpful comments.This project is funded by the MIT Skoltech Program and by the Cooperative Agreement betweenthe Masdar Institute of Science and Technology and the MIT Media Lab Consortia.10 eferences
Anderson, J. E., 1979. A theoretical foundation for the gravity equation. American Economic Review 69 (1), 106–116.Anderson, J. E., Marcouiller, D., 2002. Insecurity and the pattern of trade: An empirical investigation. Review of Eco-nomics and Statistics 84 (2), 342–352.Bahar, D., Hausmann, R., Hidalgo, C. A., 2014. Neighbors and the evolution of the comparative advantage of nations:Evidence of international knowledge di ff usion? Journal of International Economics 92 (1), 111–123.Bailey, M., Cao, R. R., Kuchler, T., Stroebel, J., Wong, A., 2017. Measuring social connectedness. NBER Working PaperNo. 23608.Balassa, B., 1965. Trade liberalisation and “revealed” comparative advantage. Manchester School 33 (2), 99–123.Boschma, R., Balland, P.-A., Kogler, D. F., 2014. Relatedness and technological change in cities: The rise and fallof technological knowledge in US metropolitan areas from 1981 to 2010. Industrial and Corporate Change 24 (1),223–250.Boschma, R., Minondo, A., Navarro, M., 2013. The emergence of new industries at the regional level in Spain: Aproximity approach based on product relatedness. Economic Geography 89 (1), 29–51.Breschi, S., Lissoni, F., 2009. Mobility of skilled workers and co-invention networks: An anatomy of localized knowl-edge flows. Journal of Economic Geography 9 (4), 439–468.Casella, A., Rauch, J. E., 2002. Anonymous market and group ties in international trade. Journal of International Eco-nomics 58 (1), 19–47.Chaney, T., 2014. The network structure of international trade. American Economic Review 104 (11), 3600–3634.Combes, P.-P., Lafourcade, M., Mayer, T., 2005. The trade-creating e ff ects of business and social networks: Evidencefrom France. Journal of International Economics 66 (1), 1–29.Feenstra, R. C., Lipsey, R. E., Deng, H., Ma, A. C., Mo, H., 2005. World trade flows: 1962-2000. NBER Working PaperNo. 11040.Gao, J., Jun, B., Pentland, A., Zhou, T., Hidalgo, C. A., 2017. Collective learning in China’s regional economic develop-ment. arXiv:1703.01369.Garmendia, A., Llano, C., Minondo, A., Requena, F., 2012. Networks and the disappearance of the intranational homebias. Economics Letters 116 (2), 178–182.Guevara, M. R., Hartmann, D., Aristar´an, M., Mendoza, M., Hidalgo, C. A., 2016. The research space: Using careerpaths to predict the evolution of the research output of individuals, institutions, and nations. Scientometrics 109 (3),1695–1709.Hausmann, R., Hidalgo, C. A., Bustos, S., Coscia, M., Simoes, A., Yildirim, M. A., 2014. The atlas of economiccomplexity: Mapping paths to prosperity. MIT Press.Heckscher, E. F., Ohlin, B. G., 1991. Heckscher-Ohlin trade theory. MIT Press.Helpman, E., 1987. Imperfect competition and international trade: Evidence from fourteen industrial countries. Journalof the Japanese and International Economies 1 (1), 62–81.Hidalgo, C. A., 2015. Why information grows: The evolution of order, from atoms to economies. Basic Books.Hidalgo, C. A., Hausmann, R., 2009. The building blocks of economic complexity. Proceedings of the National Academyof Sciences of the United States of America 106 (26), 10570–10575. idalgo, C. A., Klinger, B., Barab´asi, A.-L., Hausmann, R., 2007. The product space conditions the development ofnations. Science 317 (5837), 482.Kogler, D. F., Rigby, D. L., Tucker, I., 2013. Mapping knowledge space and technological relatedness in US cities.European Planning Studies 21 (9), 1374–1391.Krugman, P., 1979. Increasing returns, monopolistic competition, and international trade. Journal of International Eco-nomics 9 (4), 469–479.Krugman, P., 1991. Increasing returns and economic geography. Journal of Political Economy 99 (3), 483–499.Lall, S., 2000. The technological structure and performance of developing country manufactured exports, 1985-98. Ox-ford Development Studies 28 (3), 337–369.Mayer, T., Zignago, S., 2011. Notes on CEPII’s distances measures: The geodist database. CEPII Working Paper No.2011-25.McCallum, J., 1995. National borders matter: Canada-US regional trade patterns. American Economic Review 85 (3),615–623.Ne ff ke, F., Henning, M., Boschma, R., 2011. How do regions diversify over time? Industry relatedness and the develop-ment of new growth paths in regions. Economic Geography 87 (3), 237–265.Parsons, C., V´ezina, P.-L., 2017. Migrant networks and trade: The vietnamese boat people as a natural experiment. IZADiscussion Paper No. 10112.Petropoulou, D., 2008. Information costs, networks and intermediation in international trade. CEP Discussion Paper No.848.Portes, R., Rey, H., 2005. The determinants of cross-border equity flows. Journal of International Economics 65 (2),269–296.P¨oyh¨onen, P., 1963. A tentative model for the volume of trade between countries. Weltwirtschaftliches Archiv, 93–100.Rauch, J. E., 1999. Networks versus markets in international trade. Journal of International Economics 48 (1), 7–35.Rauch, J. E., 2001. Business and social networks in international trade. Journal of Economic Literature 39 (4), 1177–1203.Rauch, J. E., Trindade, V., 2002. Ethnic Chinese networks in international trade. Review of Economics and Statistics84 (1), 116–130.Ronen, S., Gonc¸alves, B., Hu, K. Z., Vespignani, A., Pinker, S., Hidalgo, C. A., 2014. Links that speak: The globallanguage network and its association with global fame. Proceedings of the National Academy of Sciences of theUnited States of America 111 (52), E5616–E5622.Simoes, A. J. G., Hidalgo, C. A., 2011. The economic complexity observatory: An analytical tool for understandingthe dynamics of economic development. In: Proceedings of the 17th AAAI Conference on Scalable Integration ofAnalytics and Visualization. AAAIWS’11-17. AAAI Press, pp. 39–42.Singh, J., 2005. Collaborative networks as determinants of knowledge di ff usion patterns. Management Science 51 (5),756–770.Tinbergen, J., 1962. Shaping the World Economy. New York: Twentieth Century Fund. ppendix A. Building a product space for 2000-2015 To calculate the ω opd , we need firstly build a product space. We define the product space bylooking at all proximity measures between products (Hidalgo et al., 2007) after aggregating allthe data that covers from 2000 to 2015. To capture the significant trade flow, we calculate therevealed comparative advantage (RCA) following Balassa (1965): RCA o , i = x o , i (cid:80) i x o , i (cid:30) (cid:80) o x o , i (cid:80) o , i x o , i (A1)Based on the result of RCA, we measure the proximity between product by calculating φ i , j between product i and j (Hidalgo et al., 2007). φ i , j = min (cid:110) P ( RCA i | RCA j , P ( RCA j | RCA i )) (cid:111) (A2)Using this significant trade flow over 2000-2015, we can create 1242 × Wood ProductsMetalsStone and GlassAnimal and Vegetable Bi-productsVegetable ProductsPapaer GoodsAnimal ProductsMineral ProductsFoodstuffsFootwear and HeadwearWeaponsTextilesTransportationMachinesArts and AntiquesPrecious MetalsChemical ProductsInstrumentsPlastics and RubbersMiscellaneousAnimal Hides
Industry ID I n d u s t r y I D P r o x i m i t y . . . . . . Histogram of log10(proximityProduct$proximity) − − − N u m b e r o f l i n k s value threshold S h a r e o f l i n k s B C ¹ .5 10- ¹ ¹ .5 10- ¹ ¹ .5 10- ¹ A D
Figure A1: Product space over 2000 to 2015: (A) Network representation of product space, (B) Cumulative distributionof proximity values, (C) Density distribution of proximity values, and (D) the product space matrix sorted in increasingorder of the is numerical code. ppendix B. Regression Results Table A1: Bilateral trade volume after two years for periods 2000-2006
Dependent variable : log x t + opd (1) (2) (3) (4) (5) (6) (7) ω topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . Ω ( d ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . Ω ( o ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . x topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . . x top . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . . x tpd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . . Distance − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . . . . . gdp to . ∗∗∗ . ∗∗∗ (0 . . gdp td . ∗∗∗ . ∗∗∗ (0 . . Population o . ∗∗∗ . ∗∗∗ (0 . . Population d . ∗∗∗ . ∗∗∗ (0 . . Border od . ∗∗∗ (0 . Colony od . ∗∗∗ (0 . Language od . ∗∗∗ (0 . Lang . Proximity od . ∗∗∗ (0 . . ∗∗∗ . ∗∗∗ − . ∗∗∗ . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . . . . . Note: ∗ p < ∗∗ p < ∗∗∗ p < able A2: Bilateral trade volume after two years for periods 2000-2006 (pre-financial crisis), 2007-2012(crisis period)and 2012-2015 (recovery period) Dependent variable : log x t + opd (1) (2) (3) (4) (5) (6) ω topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . Ω ( d ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . Ω ( o ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . x topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . x top . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . x tpd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . Distance − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . . . . gdp to . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . gdp td . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population o . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Population d . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Border od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Colony od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Language od . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . Lang . Proximity od . ∗∗∗ . ∗∗∗ − . . . . − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . . . . Note: ∗ p < ∗∗ p < ∗∗∗ p < ppendix C. Summary statistics and correlation table Table A3: Summary statistics: 2000-2006
Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − able A4: Summary statistics: 2007-2012 Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − − able A5: Summary statistics: 2012-2015 Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − a b l e A : C o rr e l a ti on M a t r i x : - ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . - . - . - . - . - . - . - . - . - . - . Ω ( d ) opd . . . - . - . - . - . . - . - . . . . . Ω ( o ) opd . . . - . - . - . . . - . - . . . - . . l og x t opd . . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . . l og x t pd - . - . - . . . . - . . - . . - . - . - . . l og D i s t an ce - . - . - . - . . . - . - . . . - . - . - . - . l og gdp t o - . - . . . . - . - . . - . - . - . . - . . l og gdp t d - . . . . - . . - . . - . - . - . - . - . . l og P opu l a ti on o - . - . - . . . - . . - . - . - . - . . - . . l og P opu l a ti on d - . - . - . . - . . . - . - . - . - . . - . . B o r d e r od - . . . . - . - . - . - . - . - . - . . . . C o l on y od - . . . . - . - . - . . - . . . . . . L anguag e od - . . - . . - . - . - . - . - . - . - . . . - . l og L ang . P r o x i m it y od - . . . . . . - . . . . . . . - . T a b l e A : C o rr e l a ti on M a t r i x : - ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . - . - . - . - . - . - . - . - . - . - . Ω ( d ) opd . . . - . - . - . - . . - . - . . . . . Ω ( o ) opd . . . - . - . - . . . - . - . . . - . . l og x t opd . . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . . l og x t pd - . - . - . . . . - . . - . . - . - . - . . l og D i s t an ce - . - . - . - . . . - . - . . . - . - . - . - . l og gdp t o - . - . . . . - . - . . - . - . - . . - . . l og gdp t d - . . . . - . . - . . - . - . - . . - . . l og P opu l a ti on o - . - . - . . . - . . - . - . - . - . . - . . l og P opu l a ti on d - . - . - . . - . . . - . - . - . - . . - . . B o r d e r od - . . . . - . - . - . - . - . - . - . . . . C o l on y od - . . . . - . - . - . . . . . . . . L anguag e od - . . - . . - . - . - . - . - . - . - . . . - . l og L ang . P r o x i m it y od - . . . . . . - . . . . . . . - . a b l e A : C o rr e l a ti on M a t r i x : - ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . - . - . - . - . - . - . - . - . - . - . Ω ( d ) opd . . . - . - . - . - . . - . - . . . . . Ω ( o ) opd . . . - . - . - . . . - . - . . . - . . l og x t opd . . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . . l og x t pd - . - . - . . . . - . . - . . - . - . - . . l og D i s t an ce - . - . - . - . . . - . - . . . - . - . - . - . l og gdp t o - . - . . . . - . - . . - . - . - . . - . . l og gdp t d - . . . . - . . - . . - . - . - . . - . . l og P opu l a ti on o - . - . - . . . - . . - . - . - . - . . - . - . l og P opu l a ti on d - . - . - . . - . . . - . - . - . - . . . . B o r d e r od - . . . . - . - . - . - . - . - . - . . . . C o l on y od - . . . . - . - . - . . . . . . . . L anguag e od - . . - . . - . - . - . - . - . - . . . . - . l og L ang . P r o x i m it y od - . . . . . . - . . . - . . . . - . ppendix D. Relationship between bilateral trade volume after two years and the threelearning channels by products’ competitiveness Table A9: Summary statistics: New exporters
Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − able A10: Summary statistics: Nascent exporters Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − able A11: Summary statistics: Experienced exporters Statistic N Mean St. Dev. Min Max ω topd − Ω ( d ) opd − Ω ( o ) opd − x topd − x top − x tpd − Distance − gd p to − gd p td − Population o − Population d − Border od Colony od Language od Lang . Proximity od − a b l e A : C o rr e l a ti on M a t r i x : N e w e xpo r t e r s ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . - . - . . - . - . - . . - . . - . Ω ( d ) opd . . - . - . - . - . . - . - . - . . - . . . Ω ( o ) opd . . . - . - . - . . - . - . - . . . - . . l og x t opd . - . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . . l og x t pd - . - . - . . . . . . . . - . . - . . l og D i s t an ce - . - . - . - . . . - . . . . - . - . - . - . l og gdp t o . . . . . . - . . - . - . - . - . - . . l og gdp t d - . - . - . . - . . . . - . . - . . - . . l og P opu l a ti on o - . - . - . . . . . - . - . . - . . - . . l og P opu l a ti on d - . - . - . . - . . . - . . . - . . - . . B o r d e r od . . . . - . - . - . - . - . - . - . . . . C o l on y od - . - . . . - . . - . - . . . . . . . L anguag e od . . - . . - . - . - . - . - . - . - . . . - . l og L ang . P r o x i m it y od - . . . . . . - . . . . . . . - . T a b l e A : C o rr e l a ti on M a t r i x : N a s ce n t e xpo r t e r s ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . - . - . - . - . - . - . . - . . - . Ω ( d ) opd . . . - . - . - . - . . - . - . . . . . Ω ( o ) opd . . . - . - . - . . . - . - . . . - . . l og x t opd . . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . - . l og x t op - . - . - . . . . . . - . . - . . - . . l og D i s t an ce - . - . - . - . . . - . - . . . - . - . . - . l og gdp t o - . - . . . . . - . . - . - . - . - . - . . l og gdp t d - . . . . - . . - . . - . - . - . . - . . l og P opu l a ti on o - . - . - . . . - . . - . - . . - . . . - . l og P opu l a ti on d - . - . - . . - . . . - . - . . - . . - . . B o r d e r od . . . . - . - . - . - . - . - . - . . . . C o l on y od - . . . . - . . - . - . . . . . . . L anguag e od . . - . . - . - . . - . - . . - . . . - . l og L ang . P r o x i m it y od - . . . . - . . - . . . - . . . . - . a b l e A : C o rr e l a ti on M a t r i x : E xp e r i e n ce d e xpo r t e r s ω t opd Ω ( d ) opd Ω ( o ) opd l og x t opd l og x t op l og x t pd l og D i s t an ce l og gdp t o l og gdp t d l og P opu l a ti on o l og P opu l a ti on d B o r d e r od C o l on y od L anguag e od l og L ang . P r o x i m it y od ω t opd . . . . . - . - . - . - . - . - . - . - . - . Ω ( d ) opd . . . - . . - . . . - . - . . . - . . Ω ( o ) opd . . . - . . - . . . - . - . . - . - . . l og x t opd . . . . . - . . . . . . . . . l og x t op . - . - . . . . . - . . - . - . - . - . - . l og x t pd . . . . . . . . - . . - . - . - . . l og D i s t an ce - . - . - . - . . . - . - . . . - . - . . - . l og gdp t o - . . . . . . - . . - . - . - . . - . . l og gdp t d - . . . . - . . - . . - . - . . - . - . . l og P opu l a ti on o - . - . - . . . - . . - . - . - . - . . - . - . l og P opu l a ti on d - . - . - . . - . . . - . - . - . . - . . . B o r d e r od - . . . . - . - . - . - . . - . . . . . C o l on y od - . . - . . - . - . - . . - . . - . . . . L anguag e od - . - . - . . - . - . . - . - . - . . . . - . l og L ang . P r o x i m it y od - . . . . - . . - . . . - . . . . - . ppendix E. Bilateral trade volume after two years for primary products, resource-based,low-tech, medium-tech, and high-tech manufactures Following Lall (2000), we exclude ”special transactions” such as electric current, cinemafilm, printed matter, fold, coins, and pets.
Table A15: Lall’s classification
Dependent variable : log x t + opd Primary Product Resource-based Manufactures Low-tech Manufactures Medium-tech Manufactures High-tech Manufactures ω topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Ω ( d ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Ω ( o ) opd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . x topd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . x top . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . x tpd . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Distance − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ − . ∗∗∗ (0 . . . . . gdp to . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . gdp td . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Population o . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Population d . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Border od . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Colony od − . ∗∗∗ − . ∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Language od . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Lang . Proximity od . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ . ∗∗∗ (0 . . . . . Note: ∗ p < ∗∗ p < ∗∗∗ p <0.01