In recent years, preferential policies such as industrial support, financing support and tax relief become the main "policy means" of local governments' development of industrial agglomeration (Zhao and Wei, 2015). Many studies believe that industrial agglomeration is mainly featured by congestion effect rather than agglomeration effect (Zhou and Zhu, 2013; Shen, et al, 2014; Sun and Zhang, 2015). However, the key to the generation and development of industrial agglomeration and its promotion role in economic growth lie in its unmatched competitiveness compared with non-agglomeration enterprises, that is, innovative advantage (Wu and Wei, 2004). Here is a new question:due to the high congestion effect, whether the prevailing "policy rent" and the industrial agglomeration phenomenon depending on its growth are not conducive to the formation of innovative ecology, thus affecting the quality and level of industrial development and further inhibiting the innovation dynamics of industrial agglomeration.

In view of the prevailing congestion effect of industrial agglomeration in China's manufacturing industries and the dilemma of transformation and upgrading, this paper argues that, industrial agglomeration provides different rent sources for the development of enterprises in the context of government intervention and market induced, so as to provide different incentives for the choice of corporate competitive strategy, thereby leading to different preferences of enterprises for technology innovation modes. Specifically speaking, the market-led industrial agglomeration can stimulate enterprise innovation activity, and the industrial agglomeration formed by excessive government intervention or even government domination not only is difficult to cultivate the agglomeration rents of external economic form, but also leads to the lock-in of low-end innovation mode in the vicious circle of low-level competition, resulting in the prevailing "innovation inertia" within the concentration scope of enterprises.

Based on the CFPS2012 micro survey data and China's industrial enterprise database, we inspect the role of industrial agglomeration in technology innovation modes of manufacturing enterprises and microcosmic effects. After controlling enterprise characteristics, product characteristics, business environment and other impact variables, the empirical results show that industrial agglomeration significantly reduces the probability of choosing high-end innovation mode by enterprises. As for reasons, the behavioral inertia of government intervention creates a living environment that depends on low cost competition for the agglomeration of manufacturing enterprises. This "greenhouse effect" leads to the tendency towards low-end enterprise technology innovation mode, which is defined as a specific form of "innovation inertia" in this paper. Further examination of its transmission mechanism shows that the low cost competition strategy is indeed a possible way for industrial agglomeration to lead to the low-end innovation mode. We also find that active government intervention is very important to reverse this phenomenon. For example, effective control in favor of market competition and property rights protection are able to create a healthy environment for the development of industrial environment and are conducive to the promotion of enterprise willingness to innovate independently. It shows that the clarification of the boundaries between the governments and the market and the cultivation of market forces for the healthy development of industrial agglomeration are long-term tasks which cannot be neglected, whether for the sustainable industry development or to encourage enterprises in innovation vitality.

The marginal contribution and innovation points of this paper lie in that:based on the analysis of different competitive strategies, the distinction between technology innovation modes of micro-enterprises and a study of the structural imbalance of the impact of industrial agglomeration on technology innovation make up for the shortcomings of the existing research; using the database of enterprises as the source of the sample, the agglomeration degree of the manufacturing sub-sectors from a perspective of the urban spatial scale is calculated, and the industrial agglomeration data at city and industry levels are matched with the micro-sample data from the World Bank survey, to better examine the impact of meso-industry environment on micro-enterprise innovation behavior; the definition of "innovation inertia" and the analysis of the causes under the lock-in of low-end technology innovation mode provide a new perspective for the formulation of industrial upgrading policy in Chinese manufacturing.