Degree Name

Doctor of Philosophy


School of Economics


The main objective of this thesis is to ascertain the determinants of trade among nations within a framework of recent and theoretically motivated trade models. It specifically focuses on trade costs facing exports from a country. Moreover, the study utilises recent cross-sectional data for the year 2007. This study integrates a somewhat scattered literature spanning more thanthree decades (from 1979 onwards) to derive the so called gravity model of trade. It then presents a comprehensive empirical analysis of internatonal trade using various formulations of the gravity equation.

Trade facilitation in recent years has become a subject of debate at various bilateral and multilateral forums. It is therefore not surprising that empirical research on the issue of trade facilitation has experienced resurgence. This study presents a comprehensive quantitative assessment of trade costs using recent trade models and their extensions. The study estimates the benchmark gravity specification that appeared in Anderson and Wincoop (2003) using country fixed effects andyear 2007 data, in a global setting of 170 countries. Noticeably, our results indicate that both bilateral distance and tariffs still have a significant negative influence on exports. Moreover, better trade-related infrastructure and border governance promotes trade.

This study also uses advanced econometric techniques by using Poisson and modified Poisson models such as negative binomial (NB), zero-inflated Poisson (ZIP), zeroinflated negative binomial (ZINB) models. These models were recently suggested for trade gravity estimation in Silva and Tenreyro (2006) and Burger et al (2009) and lend further credibility to our results. Assuming heterogeneity (in productivity) of exporting firms as suggested in Melitz (2003) and Helpman et al (2008), our results indicate that the extensive margin ( that is the number or proportion of exporting firms) is a significant determinant of aggregate exports for the year 2007.

The somewhat forced use of country fixed effects precludes including countryspecific variables into the analysis. This study utilises an extension of the gravity model presented in Baier and Bergstrand (2009) that uses a first order Taylor-series expansion to approximate for theoretically motivated multilateral resistances. This approximation approach notonly allows us to include policy relevant country-specific variables in analysis, but also facilitates counter-factual simulations. Our results indicate that both exporting and importing country’s quality of infrastructure and governance is a significant determinant of exports. Moreover, lack of access of a country to the sea has a negative influence on trade. Interestingly, our results indicate that when an exporting or importing country is landlocked, it has a dampening impact on exports and more so if it is the importing country that is landlocked. The so-called approximation approach also allows us to conduct a general equilibrium comparative static exercise. As an example of methodology this study provides counterfactual simulations involving South Asian Association for Regional Cooperation (SAARC) countries. Our results indicate that a reduction in tariffs and improvement in traderelated logistics can result in enormous trade benefits for the SAARC region.

Lastly, this study revisits the results presented in Helpman et al (2008) by providing a disaggregation analysis of the extensive margin of trade. This study disaggregates exports data to HS 2-digit level based on classification presented in Rauch (1999) for the year 2007. Utilising a two-stage estimation proposed in Helpman et al (2008) our results suggest that extensive margin is not a significant determinant of trade for homogeneous sectors characterised by high elasticity of substitution. In contrast, extensive margin is a significant determinant of trade for differentiated sectors characterised by low elasticity of substitution. These findings have important empirical and theoretical implications and support recent theoertical trade models such as Chaney’s (2008).

Overall the study provides fresh insights into the determinants of trade with a focus on trade costs. For the purpose of analysis, this study brings together and further extends the literature presented in Anderson and Wincoop (2003), Helpman et al (2008) and Baier and Bergstrand (2009). Our estimations, however, shouldonly be used to give a sense of the order of magnitude by which a policy can impact trade, Policy makers should, therefore, exercise judgement on the extent to which model results should derive policy making.



Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.