Have imports or exports driven Korean productivity growth (since 1980)?

So, will a U.S.-Korea trade agreement affect Korean productivity growth more through increased exports to the U.S., or through Korea's own increased openness to imports? 

Sangho Kim of Honam University, Hyunjoon Lim of the University of Rochester, and Donghyun Park of Nanyang Technological University look at The Effect of Imports and Exports on Total Factor Productivity in Korea (Research Institute of Economy, Trade, and Industry, April 2007), and find that:

1. imports have been more important for growth in total factor productivity than exports - at least since 1980 (their data set covers the years 1980-2003;
2. imports of consumer goods and capital goods have had a significant effect, imports of raw materials have not;
3. imports from G7 countries have had a significant effect, imports from other countries have not;
4. "These findings imply that the beneficial impact of imports stems not only from competitive pressures arising from the imports of consumer goods but also from technological transfers embodied in the imports of capital goods and imports from developed countries";
5. "However, our GDP growth regression results suggest that the beneficial effect of imports on productivity arises primarily from the competitive effects of imported consumer goods".

What does it mean for policy:

Many earlier studies on the trade-growth nexus imply that exports enhance productivity growth because firms exposed to international competition tend to absorb best-practice technology. This argument served as a major rationale for why developing countries such as Korea erected trade barriers to protect their infant industries until they become internationally competitive.  However, our empirical results suggest that lower trade barriers and higher imports would have been beneficial for Korea’s productivity growth during 1980-2003....

Our empirical findings have important implications for Korean policymakers. There is still a widespread philosophical tendency in Korea to view exports as beneficial and imports as harmful. This is perfectly understandable in light of the central role of export-oriented industrialization in the transformation of Korea from a typical poor developing country into one of the most dynamic economies in the world. The Asian crisis of 1997-1998 has further reinforced mercantilist tendencies by pointing to the potential benefits of running trade surpluses and thereby accumulating large international reserves. However, our findings clearly show that the notion of desirable exports and undesirable imports may be misguided and counterproductive. In the case of Korea, we find quite robust evidence in favor of the import-growth nexus, which is supported by the results of our causality tests, productivity determination regressions, and alternative specifications of growth. More specifically, we find that imports, but not exports, have a significant positive impact on TFP growth. It is worth remembering that Korea has already achieved a high per capita income level so that productivity growth will become more important relative to factor accumulation as a source of economic growth. The unmistakable implication for Korean policymakers is the need to open up more to foreign imports, which will help to bring about institutional and technological progress conducive to TFP growth. In light of our findings, policymakers in other East Asian economies, which share Korea’s experience of successful export-oriented industrialization, would also do well to consider the potential benefits of both imports and exports in formulating their trade and other policies.