|
|
|
| How Does Homogeneous Competition in Industrial Policy Affect the Spatial Allocation of Fiscal Equity Capital? Evidence from LLM-Based Identification and Micro-Level Investment Events |
| LIN Zhifan, WANG Maosen, TANG Zishen
|
| Bay Area International Business School, Beijing Normal University; Lingnan College, Sun Yat-sen University;Business School, Beijing Normal University |
|
|
|
Abstract Government investment funds in China serve as both a key vehicle of the sci-tech finance system and an increasingly important instrument for local governments to implement industrial policies. Governments at various levels contribute fiscal capital as limited partners, channeling private investment into strategic emerging industries and critical sectors, with the stated objectives of fostering new quality productive forces, supporting technological innovation, and promoting local industrial upgrading. Unlike venture capital and private equity models dominated by private capital in Western countries, China's government investment funds are mostly funded by local fiscal budgets and subject to local performance assessment, making their investment decisions heavily reflective of local industrial policy orientations. By 2025, over 2,100 government investment funds had been established nationwide, with a target scale of approximately RMB 11 trillion, and sub-provincial governments had become the main sponsors. Over the past two decades, local industrial policies have exhibited pronounced homogenization: provinces and cities compete to prioritize the same sectors, including new energy and intelligent connected vehicles, advanced manufacturing, new materials, biomedicine, and big data. This gives rise to the central question of this study: how does horizontal intergovernmental competition, triggered by industrial policy homogenization, affect the spatial allocation of fiscal equity capital?
This study proposes that industrial policy competition among local governments is a key constraint on the cross-regional allocation of fiscal equity capital, and explains this through two mechanisms. The first is economic rationality: industrial policy homogenization reduces the differentiated institutional rents available from out-of-region investments, homogenizes the project pool, raises screening costs, and reinforces expectations of overcapacity, thereby lowering the net returns on cross-regional investments. The second is intergovernmental competition: because the funding and performance evaluation of government investment funds are embedded in local agendas, when cities within a province compete in similar sectors, local officials have incentives to retain fiscal equity capital locally, and canimpose constraints on fund managers through seats on investment committees, performance appraisals, and commitment to future follow-on funding.
Accurately measuring the degree of industrial policy homogenization is a core challenge. This study leverages a large language model to overcome the limitations of keyword-based searches and subsidy proxies: we apply DeepSeek-V3, an open-source model developed domestically, to process 1.77 million policy documents from the PKULAW database (2011–2022). The model identifies industrial policy documents based on four criteria, including government agency, specific instruments, target industries, and long-term objectives, and extracts key industries, policy goals, and instruments by “quoting original texts” to mitigate hallucination. We then construct city-year policy vectors and compute cosine similarity between each city and other cities within the same province. This measure is matched with 20,836 government fund investment events identified through multi-layer equity penetration and manual verification, classified into local investments (11,383) and cross-regional investments (9,453), forming a panel dataset covering 300 prefecture-level jurisdictions and 3,350 city-year observations. Fund investment data are obtained from Qingke PEdata, and city-level fundamentals and additional test data are drawn from various statistical yearbooks, CNRDS, CnOpenData, and other sources with cross-validation.
Baseline regressions show that the more similar a city's industrial policy is to those of other cities in the same province, the significantly less cross-regional government fund investment its local firms receive. This result survives a series of robustness checks. Mechanism tests indicate that the negative effect stems from both economic rationality and intergovernmental competition. The intergovernmental competition effect is evidenced by: (1) in the first year of a new Party Secretary's tenure, the marginal effect of policy similarity on cross-regional investment is about 2.2 times that in other periods; (2) the baseline effect is driven solely by policy similarity with intra-provincial cities, not by similarity with extra-provincial cities. Heterogeneity analyses show that the effect is stronger in cities with tighter fiscal constraints, concentrated in industries prioritized by central and provincial governments, and primarily driven by homogenization in supportive policy instruments. Analysis of economic consequences finds that industrial policy homogenization significantly weakens the positive effects of government investment funds on firm entry and intellectual property transactions.
This study offers three policy recommendations. First, for strategic emerging and future industries with strong economies of scale, the central government should take the lead in coordinating major investment decisions and give national-level funds a stronger role in crossregional allocation. Second, local industrial policies should move away from imitation toward a more differentiated approach grounded in comparative advantage. Evaluations of official performance should give due weight to fostering locally distinctive industries that fit regional conditions, and a mechanism for crossregional policy coordination and information sharing should be developed within the unified national market framework. Third, transfer payments and supporting fund mechanisms need to be improved to bolster policy implementation in fiscally weaker areas. At the same time, information disclosure and accountability should be enhanced to create a more favorable institutional environment for the crossregional flow of fiscal equity capital.
|
|
Received: 28 January 2026
Published: 19 October 2026
|
|
|
|
|
| Cite this article: |
|
LIN Zhifan,WANG Maosen,TANG Zishen. How Does Homogeneous Competition in Industrial Policy Affect the Spatial Allocation of Fiscal Equity Capital? Evidence from LLM-Based Identification and Micro-Level Investment Events[J]. Journal of Financial Research,
2026, 555(9): 20-38.
|
|
|
|
| URL: |
| http://www.jryj.org.cn/EN/ OR http://www.jryj.org.cn/EN/Y2026/V555/I9/20 |
|
|
|
