Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Dr. James MacGee

2nd Supervisor

Dr. Ananth Ramanarayanan

Joint Supervisor


My thesis consists of three chapters relating to topics in International Economics. In the first essay, I use bilateral trade data from Canada, Germany, Japan, Mexico, the U.S. and the U.K. to decompose the patterns of trade growth across various goods classifications during episodes of rapid growth in bilateral trade. I find that bilateral trade growth during these episodes is granular- less than 5\% of goods classifications account for over 65\% of overall bilateral trade growth. I quantitatively assess whether ``Melitz-style" trade models, with heterogeneous productivity firms, CES demand and fixed and variable costs of exporting, can match the observed granularity of bilateral trade growth. I find the standard model generates only 10\% of the observed granularity in the data, as measured by the share of total trade growth accounted for by various quantiles of goods classifications. However, by incorporating heterogeneous productivity changes and tariff reductions imputed from the U.S. production and export data, I find that the model generates roughly 70\% as much granularity of trade growth across goods as in the data.

When firms export their goods to foreign markets, they often choose between multiple distribution technologies in transporting their goods to their final destination. The second essay extends the standard trade model by incorporating a choice among two distribution technologies in the exporting process- one low-fixed, high-variable cost method, and one high-fixed, low-variable cost method- and assessing the implications for trade growth across goods. In this model, I find that heterogeneous productivity or tariff changes may lead firms to ``switch'' their optimal distribution method- from not-traded to traded, or from the low-fixed cost to the high-fixed cost technology. This results in disproportionately larger trade growth for these types of firms, since they benefit from a double reduction in the variable costs of exporting- the direct effect of the fall in trade costs, and the indirect effect of switching to a lower variable cost distribution method. Calibrating this model to bilateral trade flows, I find that model simulations with multiple distribution technologies generate up to 90-95\% of the granularity in trade growth observed in the data.

The third essay examines the role of variation in transportation options- what I denote the ``supply network''- on observed price differences between locations for a specific good, retail gasoline. I use a unique data set of weekly gasoline prices across 44 Canadian cities to analyze how the existence of variation in the available modes of transportation for gasoline between cities (via pipeline, marine tanker, rail or truck) accounts for observed price differences across locations. I find that the supply network is significant- cities connected by lower per-unit cost methods like pipelines or seaports exhibit smaller mean- and weekly-price differences than those connected only by road or rail, after controlling for variables such as distance, regional effects and market size.