中文摘要：

威廉姆森假说认为,空间集聚在经济发展初期能显著促进效率提升,但达到某一门槛值后,空间集聚对经济增长的影响变小,甚至不利于经济增长,拥挤外部性更倾向于分散的地理空间结构。本文基于巴罗（Barro）增长模型建立了门槛回归模型,利用中国30个省域1978—2008年的数据对威廉姆森假说进行了实证检验,考察了空间集聚与中国区域经济增长之间的关系。研究结果表明,空间集聚对经济增长具有非线性效应,即没有达到门槛值以前,集聚对经济增长具有正效应,但超出门槛值后,集聚会降低经济增长率,即威廉姆森假说在中国显著存在。研究同时还表明,中国经济发展呈现出趋同和发散并存的态势。最后,本文据此提出了相应的政策建议。

英文摘要：

Williamson hypothesis suggests that efficiency can be significantly enhanced by spatial agglomeration at the early stage of development.But after reaching a threshold value,the effect of spatial agglomeration on growth is weaker,even adverse.Congestion externalities may favor a more scattered geo-spatial structure.Based on Barro＇s growth model,this paper used a threshold regression model and 30 provincial data in China from 1978 to 2008 to make an empirical test of Williamson hypothesis and explored the relationship between spatial concentration and economic growth.The results showed that spatial agglomeration had significant non-linear effect on China＇s economic growth.That is,while not reaching the threshold value,agglomeration had a positive effect on economic growth;once exceeding the threshold value,agglomeration would reduce the rate of economic growth.That is to say,there existed a significant Williamson hypothesis.Meanwhile,the trend of economic growth in China had shown the coexistence of convergence and divergence.Accordingly,this paper presented some corresponding policy recommendations in the final part.