Centres of Gravity
: Gravity Models, Market Access, and the Drivers of Trade

  • Conrad Copeland

Student thesis: Doctoral ThesisDoctor of Philosophy (PhD)

Abstract

In this dissertation, I analyse the drivers of trade, both modern and historical, and how these forces influence trade flows and growth outcomes. In the first chapter, I investigate the effects of ethnic linkages on trade between countries within Africa. I construct a newly digitised and geo-referenced dataset of historical trade routes for pre-colonial Africa. The empirical strategy exploits the role of historical trade routes in shaping the distribution of ethnic groups across countries to estimate the impact of shared ethnicity on modern trade flows. The impact of ethnic links is quantitatively similar to that of formal trade agreements, with trade more than doubling between countries that share a common ethnic group. The results are driven by a few ethnic groups creating powerful links between countries despite often being minority groups. In the second chapter, I estimate a modified gravity model that incorporates the impact of conflict in adjacent countries into transport costs. I use geographic variation in the location of countries and conflicts to identify an adjacent conflict effect that negatively impacts trade. Countries with already precarious access to external markets (such as landlocked countries) are more adversely affected by adjacent conflict, experiencing up to twice as severe a drop in trade as other countries. In the third chapter, I exploit an ancient textual source, the Periplus of the Erythraean Sea, to measure and model ancient trade in the Indian Ocean. I construct sailing times between ancient cities and estimate a gravity model: distance matters for trade in the region, but there are significant non-linear effects due to resource endowments. Trade appears to encourage economic growth in the long run, particularly for those cities that export a greater variety of products. Additionally, some cities appear to be constrained by an ancient version of the ‘resource curse’.
Date of Award23 Mar 2021
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorYanos Zylberberg (Supervisor) & Helen D Simpson (Supervisor)

Cite this

'