Economic activity increased in the time dimension and focus on the spatial dimension of the two most prominent feature of economic performance in the development process,both are closely related. However, in a long time, growth and agglomeration of economic activity was divided, the lack of space factor in the study of economic growth, the study of economic concentration was difficult to find traces of growth. Until the emergence of new economic geography, the two theories began to be incorporated into the same system. In reality, the process of economic development, economic active region are often areas economic activities assemble, this remarkable economic phenomenon caused scholars to its extensive experience in research. However, existing empirical studies did not reach a consensus on the relationship between economic growth and spatial concentration of economic activity. Some studies suggest that the economic agglomeration between economic growth have significant positive correlation, and some studies have shown that economic concentration does not significantly promote economic growth and or even negative. In addition, the study by Williamson pointed out that when the low level of economic development, agglomeration can promote economic growth, when high levels of economic development, the concentration will inhibit economic growth, both of which exhibit nonlinear relationship, and this is the famous Williamson hypothesis.Given the current economic developmen of China, East Midwest is extremely unbalanced both in in the spatial dimensions and the time dimension of economic activity. How to promote the Midwest, especially the western region’s economic development, become an urgent task. And the western region of Yunnan on the one hand, the level of economic agglomeration and the level of economic development are relatively low, economic development faces enormous challenges. on the other hand, because of its unique natural conditions, location advantages, as well as the national strategy of "One Belt And One Road", yunnan’s development is full of opportunities. Due to different empirical conclusions, the original research can make a limited reference in yunnan, So this article attempts to verify relatively perfect Williamson hypothesis.This article selects 129 counties (area, city) of yunnan province in 2004-2014 panel data as samples, With employment density as a measure of economic agglomeration,using system GMM estimation method to solve the possible existence of endogenous problems and existing problems such as the explained variable lag variables. The results confirmed that within the scope of yunnan,there is nonlinear relationship between agglomeration and growth:The lower level of agglomeration, agglomeration can promote economic growth significantly; With the constant improvement of the degree of concentration, when the tipping point, more agglomeration will inhibit the function of growth, namely the formation of inverted "U" curve; Assume that GDP per capita is 27300 yuan in yunnan province (that is, the level of GDP per capita in yunnan province in 2014), the critical value of employment density size is 272 people/km2; From yunnan province as a whole, however, employment density per capita of 2014 is only 11 people/km2, only wuhua and panlong exceed the critical value. Therefore, the economic activities of agglomeration in yunnan province is far from the levels of suppressing economic growth; Meanwhile, higher primacy ratio in Yunnan province, highly unbalanced economic development uneven, leading to the radiation effect of central city to surrounding areas is weak, therefore, adjusting measures to local conditions to vigorously promote other urban agglomerations is conducive to further promote economic growth in yunnan province. |