Comments by Khee Giap Tan, on Location Proximity and Productivity Spillovers: The Case of Korean Manufacturing Plants

Abstract

Khee Giap Tan: Backhoon Song's paper went beyond the traditional growth and productivity achieved through economic clusters but argued along the line of the Marshall thesis on agglomeration economies based on positive externalities in response to economic growth and the geographical concentration of firms where common resources in terms of labor productivity, technical know-how, infrastructure build-up, and management expertise and government facilitation were pooled, shared, and spilled over to generate spatial externalities. The author concludes that their empirical findings are consistent with the Marshall thesis where a firm located in a region with higher median total factor productivity (TFP) gains higher productivity from other firms in the same region due to easier access to superior external knowledge to generate new ideas where knowledge can be acquired not just from a firm's internal R&D investment but also from external sources from other firms through location proximity and hence higher future productivity.A few observations pertaining to model specification are in order. Firstly, it is not clear how external knowledge stock can be justly approximated by the “median TFP of a plant's location” and “the number of plants in that location” especially when the level of the median TFP in the area where the plant is located is obviously determined and dependent upon the number of plants in that location. In some sense the two proxies are somehow related so long as median TFP is adopted. Furthermore, it may be too simplistic to assume more plants in a location necessarily reflect more opportunities to produce new ideas or a greater variety of external knowledge as firms are attracted to a location for many other reasons such as appealing tax allowances, efficient infrastructure facilities, and competitive logistics services, which can all lead to improvement in productivity.Secondly, a closer approximation for external knowledge stock other than the two current proxies may be useful. Presumably, spillover that originates from knowledge stock may also be viewed more directly by the turnovers of managers across plants through poaching of staff or measuring violation of patent rights through pirate-copying, which are quite common within-industry. Finally, a more general comment is about the estimated reduced form equations adopted versus a more structural model of equations so as to get a clearer or a more convincing empirical case for spatial externalities.Notwithstanding these observations, in terms of the empirical results, interpretations on the positive coefficient for both the entrant dummy variable and the age variable are interesting—that is, that the new entrant is well prepared with the advanced equipment and production process at the stage of start-up and more likely to survive in the next period than old firms, and the new entrant effect does not last for a long period as aging itself has a positive influence on a firm's productivity evolution.In terms of spillover effect from knowledge stock, the author found that a firm surrounded by lower productivity neighbors has more chance to survive and the effect of a firm's internal knowledge on its future TFP would be stronger than the effect of its other external knowledge. More significantly, it was empirically determined that spillovers associated with the group of high-productivity neighbors are greater than the group of low-productivity neighbors, and firms benefited most from combining their internal knowledge with the external knowledge of neighbors that have higher TFP on average. These findings on a firm's knowledge spillover seemed intuitively acceptable.