17 November 2007 (Sat.), GRIPS Campus in Tokyo, 14:00-17:00

"The Impact of the Components of Trade Costs on Export Growth: An Empirical Measurement"

Mr.  Imran Ullah Khan

PhD candidate, GRIPS-FASID Joint Program


 To expand our research networking with other people from different countries, this time VDF-Tokyo invited Mr. Imran Ullah Khan, a PhD candidate of the GRIPS-FASID Joint Program, to share his empirical findings for the export growth in Pakistan during the past decade.


According to Mr. Imranfs overview, there had been a variety of studies on the impacts of trade liberalization on economic growth, poverty reduction, income inequality, and gender equality. He also emphasized that the links between trade liberalization and aforementioned indicators could be analyzed meaningfully only if the determinants of trade were precisely identified. In this research, he would like to explore the impacts of trade costs -- which were considered as a determinant of export growth, and were usually proxied by geographical distance -- on the export growth in Pakistan. Going further with definition and specification of trade costs in various studies, Mr. Imran showed that trade costs had been defined differently in literature, and they had various levels of impacts on export growth in different countries. Among the mostly cited studies, Mr. Imran focused more on such important analyses as McCallum (1995), Anderson and van Wincoop (2003), and Balisteri and Hillberry (2007).


Before presenting the main analysis for the case of Pakistan, the presenter made a brief summary on the growth trend of exports and imports of Pakistan in 1999-2004, which was the period of study. Though Pakistan had impressive export growth during the period, it was shown that the country had been still a relatively closed economy if we compared export volume with total population and gross national product.


Mr. Imran then provided the analytical framework for his study, and analyzed some constraints of trade costs, including natural constraints (or distance between trading points), behind the border constraints (e.g. transaction costs, port and custom procedures, and taxes and duties), explicit beyond border constraints (e.g. tariffs and exchange rates), and implicit beyond the border constraints (e.g. marketing and retailing). Each constraint had a lot of determinant factors. In addition to these constraints, Mr. Imran also discussed about the estimation issues by the gravity model of international trade under ordinary least squares (OLS) or generalized least squares (GLS), and showed that these methods could not be able to model all the determinant factors of trade, as they included multilateral resistance terms and produced biased results due to heteroskedasticity. To overcome the problem, Mr. Imran used the gravity model with composed error term, which was proposed by Kalirajan (2007). This method would decompose the trade costs between Pakistan and its trading partners into different costs due to aforementioned constraints.


The presentation was then continued with some initial findings, including the estimates for export losses in 1999 and 2004 in top 10 countries, in which the two biggest trading partners, i.e. China and India, showed the highest levels in both points of time. These losses were then decomposed by the changes due to the above-mentioned constraints.


[For further information of mathematical and graphical illustrations, as well as detailed findings, please refer to his presentation attached to this summary]


As usual, the Q&A session started by comments and questions from Prof. Kenichi Ohno (GRIPS & VDF).  He suggested Mr. Imran use similar products exported to different countries when decomposing trade costs, as they might provide more comparable data on such important costs as transportation and tariffs. Providing further insights for the suggestion, Prof. Ohno said that there would be significant differences between high income-elasticity products (such as plasma TV) and low income-elasticity products (such as T-shirt), as well as between intermediate and final products. As such, both behind and beyond border constraints of these products must be substantially different. In his response, Mr. Imran explained that majority of exports from Pakistan to other countries were cotton and garment products, and they were the main products he considered in the analysis. However, he acknowledged that he would also need to consider other products in the export categories, so that the estimates would be more precisely reflected by the changes in their determinant factors.


Going in detail with the model, Prof. Ohno said that there might be a lot of factors, beside trade costs, influencing error terms proposed in the model, and thus the estimates might ignore these omitted factors. Based on this comment, Mr. Giang Thanh Long (GRIPS & VDF-Tokyo) suggested Mr. Imran conduct a factorial model to capture the possible factors, such as certain trade policy changes during the study period, of the error terms. As supervisor of Mr. Imran, Prof. Kaliappa Kalirajan (GRIPS-FASID) agreed with these comments, and said that such model would be a necessary step to get further policy implications for the results.


Ms. Sachiko Kondo (JBIC) wondered whether there was a tendency for trade changes between 1999 and 2004 in Pakistan, and there were sudden changes in trade policies, such as tariffs, because she thought that these factors must also have significant impacts on export growth, particularly under strong globalization process. In his answer, Mr. Imran said that there was almost no evidence for such changes, and thus the trend was relatively stable, especially when considering trade volumes between Pakistan and other South Asian countries.


Going back with the graph showing exports and imports of Pakistan, Prof. Ohno suggested Mr. Imran look for the reasons why exports were growing faster than imports during the study period. As for other countries, some of possible reasons include the increasing inflows of foreign direct investment (FDI), or increasing fund from the official development assistance (ODA). Mr. Imran said that he would consider these factors in the factorial model of the error terms to provide more concrete results.


At the end of the presentation, Mr. Imran thanked all the participants of the workshop. Then, we had 1-hour informal meeting with him and other people to exchange various sources of information.


In the next workshop on Saturday, December 15, 2007, we will also welcome a graduated student of GRIPS-FASID Joint Program -- Ms. Pham Thu Hien (a Vietnamese) -- to present her findings on the relation between ODA in infrastructure and provincial FDI inflows in Vietnam.

Extended Abstract (PDF30KB), Slides (PDF78KB) 

 (By Giang Thanh Long)


WS Schedule | VDF-Tokyo HOME