作　　者： ;

机构地区： 中山大学国际金融学院

出　　处： 《经济研究》 2014年第12期18-32,共15页

摘　　要： 本文论证表明现有测算跨地区边界效应的文献关于市场一体化的前提假设可能不成立，且实证分析中还存在数据的加总偏误和样本选择偏误等问题。本文选取2011年5月至2014年5月长三角15个城市224个市场37种农产品的161个周度价格及成本数据，在一价定律的框架内重新考虑地区间市场分割情况，运用断点回归方法对省际边界和价格差异的关系进行实证研究。结果表明：（1）价差变动来源于成本差异和利润差异的变动，跨边界市场分割程度更大；（2）控制了当地市场供给和需求特征后，长三角地区省际边界效应显著存在，上海-江苏、上海-浙江和江苏-浙江的边界效应分别为42％、45％和32．4％；（3）由于跨边界农产品零售市场的分割程度比较大，本文估计的边界效应值只能反映真实边界效应分布的下界。经检验，本文断点回归设计有效，实证结果稳健。 The theory framework of ＂The Law of One Price＂, which is often used to estimate border effect, is established on the assumption of market integration, but the assumption is not necessarily held in fact. And empirical studies have problems of aggregation bias and sample selection bias. This paper selected weekly price and cost data of 37 kinds of agricultural products in 224 markets of 15 cities in the Yangtze River Delta Region from May 2011 to May 2014, combined market segmentation analysis into theoretical study, then used Regression Discontinuity to infer the causal relationship between border and price and re-examined provincial border effect of China. The empirical results show that： （ 1 ） Price difference fluctuation comes from both cost fluctuation and profit fluctuation. Border makes the deviation from Law of One Price more serious, and cross-border market segmentation is greater than that within border. （2） Border effect is still significant after controlling market demand and supply characteristics. Shanghai-Zhejiang border effect is 42% , Shanghai- Jiangsu is 45% , and Jiangsu-Zhejiang is 32. 4%. （3） Cross-border market segmentation is serious, so our border effect estimates only reflect the lower bound of border effect distribution. Our Regression Discontinuity Design is valid, and the empirical results are robust.

关 键 词： 微观商品价格差异 市场分割 边界效应 断点回归

分 类 号： [F114.41]

领　　域： []