溢出效应、边界选择与产业集聚测度

产业集聚程度的测度结果受所选产业加总水平与地理单元大小的影响,后者应以集聚过程中溢出效应的产业边界与空间边界为基础。当前产业集聚测度研究普遍存在产业加总水平与地理单元选择过大的问题,本文试图研究在这种情况下测度结果是否存在偏差,偏差的系统性、方向、大小和影响因素,以及有偏的测度结果对集聚经济研究的影响等。结果表明,超出空间边界或产业边界的测度结果将显著偏低,并导致对集聚经济的低估;按省测度的区位基尼系数比按市测度平均低了近40%,两位数产业的区位基尼系数通常低于其所含三位数产业的区位基尼系数的均值,甚至最低值。

Spillovers, Selection of Frontiers and the Measurement of Industrial Agglomeration

Measurement of industrial agglomeration relies on the choices of industry aggregation level and scale of geographical unit.The choices should be based on the frontiers of the existence of spillovers.Most measurement practices so far have the problems of choosing industry aggregation level and/or scale of geographical unit beyond the existence of spillovers.This paper is trying to prove the existence of deviate caused by such problems,and to find out the direction,degree and determinants of deviate,and to explore the influence of the biased results on the study of agglomeration economies.Result shows that measured degree of agglomeration will be downward biased when the chosen level of industry aggregation or scale of geographical unit is too large for the existence of spillovers,and the deviate will cause under-estimation of agglomeration economies.Empirical studies with the data of Chinese manufacturing industry verify the theoretical analysis,and find that:(i) the locational gini coefficients of 3 digit industries at province level are approximately 40% lower than those at city level,(ii) the locational gini coefficients of most 2 digit industries are lower than the means of the coefficients of the corresponding 3 digit industries,even lower than the minimum one.