Landlockedness and economic development : analyzing subnational panel data and exploring mechanisms

This paper revisits the hypothesis that landlocked regions are systematically poorer than regions with ocean access, using panel data for 1,527 subnational regions in 83 nations from 1950-2014. This data structure allows us to exploit within-country-time variation only (e.g., regional variation within France at one point in time), thereby controlling for a host of unobservables related to country-level particularities, such as a country's unique history, cultural attributes, or political institutions. Our results suggest lacking ocean access decreases regional GDP per capita by 10 - 13 percent. We then explore potential mechanisms and possible remedies. First, national political institutions appear to play a marginal role at best in the landlocked-income relationship. Second, the income gap between landlocked and non-landlocked regions within the same nation widens as i) GDP per capita rises, ii) international trade becomes more relevant for the nation, and iii) national production shifts to manufacturing. Finally, we find evidence consistent with the hypothesis that national infrastructure (i.e., transport-related infrastructure and rail lines) can alleviate the lagging behind of landlocked regions.