Abstract:     Inclusive institutional arrangement and selective direct support are two main ways for governments exerting their influence on enterprises′ innovation activities. The former refers to governments managing to create a suitable external environment for regional innovation activities through industrial, legal, talent, financial and other functional institutional arrangements. The latter is the government′s selective and differential funding support for specific industries or specific enterprises, which aims to enhance industrial security, and speed up technology catch-up or radical innovations in key fields.
    Based on the data of manufacturing companies listed in China′s mainland and text analysis of the work report of local government from 2008 to 2017, fixed effect models are employed to investigate the impact of different ways of government′s participation on the innovation output of enterprises. The firm-level data such as patent authorization and government subsidies required for empirical research comes from the CSMAR database. The data required for regional level variables such as economic development level and industrial structure characteristics extract from the China Economic Database (CEIC). This paper sorts out the government work reports for the corresponding year in the 225 cities where the sample companies are located, and text analysis based on 10 keywords related to institutional arrangement provide a reasonable measurement of institutional arrangement. 
    Inclusive institutional arrangement can promote high-quality innovation output of enterprises more precisely. "Spill-over effect" of the policy on low-quality innovation output is not significant. The possible reason is that China′s social innovation resources are still relatively insufficient in current period, and the optimization of institutional environment promotes the free flow of innovation elements and leads resources to flow more to high-quality innovation activities. On the other hand, inclusive institutional arrangement is less likely to induce companies to "whitewash" their R&D performance with low-quality innovation output.
     Selective direct support can promote high-quality innovation output as well, while this policy may induce enterprises′ opportunistic behavior of whitewashing their R&D performance with low-quality innovation output. Information asymmetry between government and support targets is higher in the innovation process and output quality, and opportunistic behavior is more likely to occur.
     Synergistic effects between two ways of government participation exist in high-quality innovation output only. Inclusive institutional arrangements can inhibit the opportunistic behavior of enterprises in the process of policy implementation, and improve the governance effects of government direct support. Selective direct support may guide the allocation of social resources to specific enterprises or industries, which can magnify and strengthen the promotion effect of institutional arrangements.
     Compared with selective direct support, effects of inclusive institutional arrangement are more durable. The innovation promotion effect of inclusive institutional arrangement has a sustainable impact lasting for at least three years. Although the effect of selective support has deferred effects also, its durability is lower. Based on the above analysis, inclusive institutional arrangement is a kind of long-term mechanism, which can promote high-quality innovation output of enterprises.
     Inclusive institutional arrangement can alleviate the negative effect of poor market environment to some extent, while selective direct support for innovation promotion can only be fully realized at a high level of marketization. The policy effect of institutional arrangements is inclusive, which have no significant difference among enterprises of different life cycle. The innovation promotion effect of selective support is more significant for enterprises in the growing stage, whose resources are relatively insufficient. 
     This paper mainly contributes to the area in the following aspects: First, it introduces the inclusive and functional institutional arrangements into the analytical framework of government intervention and social innovation activities, and expands the research horizon in related fields. Second, the systematic analysis of the two ways of government participation in innovation on the promotion of different quality R&D output of enterprises and the persistence of their policy effects has clarified the heterogeneous innovation governance effects of the two participation methods, providing an analytical framework for reference for subsequent research.
    In the backgrounds of global changes in the trend of technological innovation, the research findings mentioned above can provide references for China to improve its national scientific and technological innovation governance system and promote the modernization of government technological innovation governance capabilities. In this process, government should pay more attention to regional innovation environment shaping through functional and inclusive institutional arrangements while continuing to refine the existing selective and differentiated innovation policies. At the same time, the public innovation policy portfolios should be adjusted based on the regional market-oriented environment, the life cycle of enterprises and resource capabilities， to fully realize its synergistic governance effect.

Key words:  government participation, inclusive institutional arrangement, selective direct support, R&D performance, heterogeneity in the innovation governance effect