Summary:
Innovation activities are characterized by high uncertainty, long development cycles, and strong specialization, which create substantial challenges for risk allocation and make it difficult for market mechanisms alone to effectively diversify and share innovation-related risks. As the primary spatial carriers of innovation resources and activities, cities play a crucial role in determining the quality and sustainability of regional economic development. Consequently, how to optimize urban innovation ecosystems and stimulate the vitality of innovation actors through institutionalized risk-governance instruments has become a critical issue in both theoretical and practical domains. As an institutional arrangement designed to support technological innovation, the Science and Technology Insurance (STI) Policy aims to reduce uncertainties associated with research and development, technology commercialization, and industrial application by providing insurance protection. The value signals embedded in such policies reflect local governments’ emphasis on innovation risk protection and their intentions regarding resource allocation. However, existing studies have rarely examined, from the perspective of central–local policy coordination, the specific mechanisms through which science and technology insurance signals influence urban innovation via local transformation effects. Taking STI policy signals as the research object and urban innovation performance as the analytical focus, this study systematically investigates how central policy signals drive urban innovation through local transformation effects. To address this question, a two-stage analytical framework of “macro-level value signals–local transformation effects–urban innovation-driven development” is constructed. The first stage examines the transmission of central policy signals to local policy signals, whereby macro-level value signals are transformed into local transformation effects. This process is verified through a comparison of the temporal distribution patterns of central and municipal policy signals. The second stage explores the impact of local policy signals on urban innovation by empirically assessing their mechanisms and effects using city-level panel data. The first stage provides the institutional cognition and signal-transmission foundation for the latter, while the second quantitatively evaluates the actual innovation outcomes. Together, these two stages form a coherent and progressive chain of empirical evidence. Using panel data from 284 prefecture-level cities in China from 2007 to 2022, this study employs the BERTopic model to conduct topic clustering and evolutionary analysis of 159 documents on science and technology insurance issued by the central government. The TF-IDF method is further used to measure the textual similarity between local and central policies and construct an indicator of urban policy signal intensity. Subsequently, two-way fixed-effects models, mediation-effect models, and spatial Durbin models are applied to empirically examine the impact of policy signals on urban innovation. The results reveal five major findings. First, the central science and technology insurance system exhibits clear evolutionary phases and structural characteristics. Its policy orientation has gradually evolved from a routine branch of property insurance into a key instrument serving the national innovation strategy. Through policy-text responses, local governments generate local transformation effects, which constitute a crucial link in central–local policy coordination. Second, urban policy signals significantly promote urban innovation, and this result remains robust after a series of robustness tests, confirming the ex ante guiding value of policy signals. Third, mechanism analyses show that policy signals stimulate innovation through three channels: promoting the development of the science and technology insurance market, increasing local government expenditure on science and technology, and encouraging regional R&D investment. Fourth, the policy effects exhibit significant structural heterogeneity. The innovation promoting effects are stronger in central regions and northern cities, more pronounced in cities with medium levels of innovation, whereas the effect is not significant in fifth-tier cities. Fifth, policy signals display spatial autocorrelation but do not generate significant spatial spillover effects. This study contributes to the literature in three respects. First, it broadens the scope of existing research by moving beyond the traditional paradigm that treats policies as homogeneous interventions and instead conceptualizes science and technology insurance as a dynamic signal-transmission system encompassing macro-level intentions, local transformation, and stakeholder responses. Second, it enriches both theoretical understanding and policy implications by empirically identifying the multiple pathways through which policy signals drive urban innovation and by identifying the spatial boundaries of their effects, thereby providing valuable references for optimizing policy design and enhancing its effectiveness in supporting regional innovation systems. Third, through the analysis of STI policy texts and the construction of indicators, this study performs clustering and dynamic thematic analysis on central STI policies, quantifies policy-text similarity, and constructs a continuous policy signal intensity indicator. This achieves a complete application of policy text mining from macro-structural analysis to quantitative measurement, offering a novel measurement approach for future research.
Białkowski, J., H. D. Dang and X. Wei, 2022, “High Policy Uncertainty and Low Implied Market Volatility: An Academic Puzzle?”, Journal of Financial Economics, 143(3) , pp.1185~1208.
[19]
Bianchi, M. and M. Bobba, 2013, “Liquidity, Risk, and Occupational Choices”, Review of Economic Studies, 80(2), pp.491~511.
[20]
García-Peñalosa, C.and J. F. Wen, 2008, “Redistribution and Entrepreneurship with Schumpeterian Growth”, Journal of Economic Growth, 13(1) , pp.57~80.
[21]
Greer, R. A., T. T. Moldogaziev, R. P. Scott, et al., 2023,“Signaling Resilience: A Computational Assessment of Narratives in Local Government Budgets”, Journal of Public Administration Research and Theory, 33(4) , pp.688~700.
[22]
Hepfer, B. F., H. W. Judd and S. C. Rice, 2025, “Signaling Innovation: The Nontax Benefits of Claiming R&D Tax Credits”, Journal of Accounting and Economics, 79(1) , 101718.
[23]
Jones, B. D.and F. R. Baumgartner, 2005, The Politics of Attention: How Government Prioritizes Problems, Chicago:University of Chicago Press.
[24]
Newman, A. F., 2007, “Risk-Bearing and Entrepreneurship”, Journal of Economic Theory, 137(1) , pp.11~26.
[25]
Spence, M., 1973, “Job Market Signaling”,The Quarterly Journal of Economics, 87(3), pp.355~374.
