(i) Bilateral exports are measured in constant PPP; source: UNCTAD (Comtrade database)
(ii) GDPs are measured in constant PPP; source: Cepii (Chelem database)
(iii) T A R 𝑖 𝑗 is captured by the applied weighted average tariffs charged on the importing country 𝑗 ; source: World Bank
(World Development Indicators). As a sensitivity analysis, the Market Access Overall Trade Restrictiveness Indexes (MAOTRI)
captures the barriers faced by each exporter when selling in other countries; source: Kee et al. (2008), [2].
(iv) NTBs are proxied by ad-valorem equivalents (AVEs) calculated in Kee et al. [2]. The methodology is fully described in
the Appendix.
(v) L P I 𝑖 is a larger proxy than transport costs; this indicator is built from the information gathered in a worldwide survey of the
companies involved in logistics services. Seven areas are covered by this index, namely: efficiency of the clearance process by
customs and other border agencies, quality of transport and information technology infrastructure for logistics, ease of arranging
international shipments, efficiency of the local logistics industry, ability to trace and check international shipments, domestic
logistics costs, as well as timeliness of shipments in reaching destination. The LPI is a weighted average of these variables.
It ranges between 1 (worst) to 5 (best); source: World Bank [3]
(vi) L A N G 𝑖 𝑗 is a dummy variable which takes the value of 1 if a common language is spoken by at least 10% of the population in
each country pair (exporter and importer) and 0 otherwise; source: Cepii (Chelem database).
(vii) C O L 𝑖 𝑗 reflects colonial relationships over a long period of time with substantial participation in the colonized country’s
governance. This variable is equal to 1 in case of colonial links and 0 otherwise. It accounts for cultural and historical
relationships which are expected to increase trade flows between some EU countries and Mediterranean countries. Source: Cepii
(Chelem database).
(viii) R E G 𝑖 𝑗 (Regional integration) is proxied by a bilateral dummy; source: WTO (RTA database)
(ix) M I G 𝑖 𝑗 𝑡 is measured by the bilateral migrant stock in country i originating from country 𝑗 ; as a sensitivity analysis, it is
measured both in value and as a percentage of the exporting country’s population; source: World Bank (Bilateral Migration and
Remittances 2010)
(x) F D I 𝑖 𝑗 is captured by several proxies: inward and outward bilateral stock or flows (in value and as a percentage of GDP);
World Bank (World Development Indicators); UNCTAD (Handbook of Statistics and World Investment Report).
(xi) S E R V 𝑖 𝑗 is measured by trade in services (source: UNCTAD [4])
(xii) P O L 𝑖 𝑗 is captured by several alternative proxies: voice and accountability, which measures to what extent citizens are able to
participate in selecting their government, political stability and the absence of violence and terrorism, government effectiveness,
regulatory quality, rule of law as well as the control of corruption (for more details, refer to [5]).
Box 1: Data and sources.