An industry level analysis of causality between export and productivity: the case of Korea

Abstract

The export-led growth hypothesis is one of the most popular topics in growth economy. Many researchers have tried to test the hypothesis by using different variables and a range of test methods. The results were mixed. This study emphasizes the specific industries by using industry-level data to test Granger type causality. We used a five-variable Lag Augmented-VAR (LA-VAR) model to test the export-led growth hypothesis in the case of Korea. We found that all light manufacturing industries show bi-directional causality and the result is robust to lag structure and data characteristic. The Non-metallic Mineral Products (industry 36) and The Motor Vehicles and Transport Equipment (industry 385) supported the export-led growth hypothesis in the sense that there is unidirectional causality from export to productivity. In heavy industry, we found more industries that have clear causal direction from productivity to export, rather than from export to productivity. However, some heavy industries showed sensitivity in lag structure, meaning the causal direction could not be determined. No causality at all was found in the Medical, Precision and Optical Instruments industry. We suspect that there is "productivity paradox" in Electric and Electronic Equipment (industry 383) and Medical, Precision and Optical Instruments (industry 384). Industry-specific industrial policy turns out to be superior to across-the-board industrial policy because policy implications differ by industry. Promotion of export behavior is not always the path to economic growth in every industry. As can be learned from the analysis of light manufacturing industries, promotion of competition and better networking and the reduction of sunken cost for industry would be a good focus for future policy. Better understanding of the mechanism for transmission of externalities could lead to better policy analysis.