The Varying State of Central Asian Intra- and Extra-Regional Trade

The purpose of this paper is to measure the extent to which Central Asian countries are currently trading more or less among themselves, as well as with other regions of the world, than would be expected on the basis of the standard empirical model of bilateral trade patterns, namely, a sophisticated gravity model. Such comparisons between actual and expected trade patterns are used to offer suggestions as to how certain kinds of trading patterns can be strengthened. Of critical importance to this task is to examine how the gaps between actual and expected trade within the Central Asian region and between it and other major regions vary according to the breadth in the definition of Central Asia and over time since the time that many countries of the region were part of the Soviet Union. We also attempt to identify some of the factors lying behind these variations over time. Understanding the drivers of these changing patterns could be used to identify priorities among factors that could overcome trade shortfalls, ranging from ports and customs improvements to strengthening trade agreements within the region and forging new ones with other regions and countries. Since recorded annual trade flows are subject to occasional measurement errors, to minimize the influence of measurement error our empirical work is based on a panel data of observed bilateral trade using three year averages for each of the five year intervals from 1990 to 2010. We apply a Poisson pseudo-maximum-likelihood estimation technique –which addresses potential bias issues in a typical bilateral trade dataset– over both narrow (traditional) and broad (and more sophisticated) gravity models that include differences in production structure as well as various historical and geographical factors, and their interactions. We find that overall, Central Asian countries are –relatively large– over-traders when it comes to intra-regional trade. Yet, the extent to which Central Asian countries trade more among each other than would be expected is substantially reduced when the region is defined more broadly to include Iran and Turkey. When it comes to the region’s trade with the rest of the world, the extra-regional trade, the picture is reversed and the region is shown to be a relatively large under-trader. Indeed, overall the extra-regional trade of Central Asia is only about 40 percent of what it would be expected to be by our gravity model, though by 2010 the extent of under-trading with the rest of the world fell to about 20 percent. While the numbers differ for imports and exports, we have found the largest extra-regional under-trading of Central Asia to take place with the United States and EU countries and to a less extent with East Asia. An important exception to this extra-regional under-trading by Central Asia is that it is importing from East Asia more than predicted by the gravity model. Our study of the interactions with various economic structure differences, cultural and geographical variables reveal the importance of these factors in explaining the patters explained above. Historical and cultural commonalities, most notably ties to Russia, seem to have played an important role in this region’s trade behavior and making it trade more than what the factors in a gravity model would predict. This group of countries seems to be less subject to the border effect as they once shared a common border and this is what distinguishes them from other developing countries the under- or over-trading patterns of which have been studied. The study concludes with policy suggestions as to means of strengthening trade within the broader definition of Central Asia and with respect to the rest of the world. See above See above