The blocking and transmission effects of lower level system in patent transaction - evidence from Chinese colleges and universities

The recognition that law is a systematic existence is widely supported by the existing literature. Within the legal system, there are divisions between legal departments, between upper and lower laws, and between new and old laws. Harmony between these divisions is the pursuit of good law and is in itself good law. In a country's legal system, the status of the state organ that formulates legal norms is directly proportional to the hierarchy of legal authority. In China, for instance, laws enacted by the National People's Congress are positioned at a higher rank compared to administrative regulations formulated by the State Council, which in turn are superior to regulations enacted by local governments. However, inconsistencies between different legal provisions on the same matter are common due to different legislative subjects, themes, timing, and amendment frequency. The Legislation Law stipulates levels of effectiveness and applicable rules such as new laws superseding old laws and special laws superseding general laws, which helps address conflicts to some extent. As economic and social system reform deepens, legal revision has become a means for law to adapt to and guide social behavior. Yet revisions often occur in isolation rather than systematically, leading to situations where an upper law changes but lower-level regulations remain unrevised. While jurisprudentially the upper law should apply, the public and implementing organizations may struggle to correctly interpret and act under conflicting rules, potentially adhering to older lower-level rules out of inertia, thereby affecting implementation outcomes. In China’s science and technology policy domain, evaluation methods are mature and enable empirical analysis of legal conflicts. The 2015 Law on Promoting the Transformation of Scientific and Technological Achievements reformed university patent transaction processes, and in April 2019 the Ministry of Finance revised the Interim Measures for the Management of State-owned Assets of Public Institutions. The short-term conflict between upper and lower laws created by these staggered reforms provides a quasi-natural experiment to study how legal system conflicts affect patent transactions. Using mixed cross-section data from three periods around the reforms and a DID framework, this paper empirically assesses patent transaction efficiency to reveal the effects and mechanisms of legal system conflicts. The contributions are: moving conflict-of-laws research from conceptual to empirical analysis; demonstrating hindrance and conduction effects from lower rules to higher laws and the need for integrated multi-level legal design; and applying a two-stage DID within one study to show real-world institutional change impacts. Practically, systematic legal amendments and improved legislative techniques can reduce system frictions and enhance socioeconomic benefits, informing policy in reforming economies.

Research on legal conflicts has traditionally centered on conflicts of law in private international law, aiming to determine applicable national laws in cross-border civil relationships. However, numerous conflicts also arise within a single domestic legal system. In China’s rapid legislative environment, inconsistencies and contradictions among laws are prevalent and expanding. Domestic legal conflicts include formal (resolvable via interpretation) and true conflicts (requiring institutional resolution). Conflicts between new general provisions and old special provisions often persist even under the Legislation Law. Legal conflicts impede effective implementation by complicating judicial decisions, increasing compliance costs, and discounting the operation of law, particularly in interactive administrative processes where differing understandings between administrators and counterparts magnify impacts. In the university patent context, the 2015 Law on Promoting the Transformation of Scientific and Technological Achievements allowed flexible transaction modes (transfer, licensing, investment by agreement pricing with internal disclosure), while the still-effective 2006 Interim Measures for the Management of State-owned Assets of Public Institutions required evaluation and approval for disposal of state-owned patent assets. This inconsistency (2015–2019) likely increased compliance costs and uncertainty, hindering the upper law’s goals. In April 2019, the revised Interim Measures aligned lower-level rules with the 2015 law, granting universities autonomy on whether to evaluate assets and removing approval/filing requirements, potentially restoring efficiency. Prior studies note tensions between state-owned asset management and patent transaction reforms and propose solutions such as inventor ownership. Based on this, the paper posits: H1: After the 2015 law’s implementation, university patent transaction efficiency decreased. H2: After the 2019 revised measures, university patent transaction efficiency recovered and improved.

Design: The study exploits staggered reforms as a quasi-natural experiment and applies a two-stage difference-in-differences (DID) framework. University patents (treated group) are compared to enterprise patents (control group), as the 2015 law and 2019 measures targeted universities and public research entities, not enterprises. Two-stage DID: Stage 1 estimates the effect of the 2015 Law on Promoting the Transformation of Scientific and Technological Achievements on transaction efficiency. Stage 2 estimates the effect of the April 2019 revision of the Interim Measures for the Management of State-owned Assets of Public Institutions, which aligned lower-level rules with the 2015 law. In each stage, the DID model includes: outcome (transaction efficiency), group dummy (Treat: university=1, enterprise=0), time dummy (Post: after policy=1, before=0), their interaction (DID=Treat×Post), and controls. Outcome variable: Patent transaction efficiency is measured as the time interval (in months) between patent publication (disclosure) and the first recorded transaction (assignment or license). A larger interval indicates lower efficiency. The focus is on invention patents (dominant in universities), which are publicly disclosed prior to authorization, enabling potential transactions. Explanatory variables: Treat (university=1, enterprise=0); Post (for Stage 1: after Oct 2015=1; for Stage 2: after April 2019=1); DID interaction is the treatment effect estimator. Controls: Transaction type (exclusivity level: transference/exclusive license=3; sole=2; general=1). Patent characteristics capturing basic information, technological breadth, and value: number of claims; number of pages in the application; number of applicants; number of IPC classes; and size of patent family (homologous number). Data: Patent transaction data were extracted from the Incopat database for 2013–2021 using keywords indicating license and assignment years. Samples included only domestic university and enterprise patents; natural persons, foreigners, and jointly-owned university–enterprise patents were excluded. Observations with missing/extreme values on key variables were removed. Total effective samples: 10,487, including 3,674 samples between the two reforms used in staged analyses. For the DID regressions: Stage 1 N=5,902; Stage 2 N=8,259. Descriptively, on average, it takes about four years from disclosure to transaction record for invention patents across groups. After 2015, average transaction time increased, motivating DID analysis to isolate policy effects. Estimation and validation: Baseline DID regressions are estimated with and without controls. Robustness checks include: (1) Parallel trend tests using interaction terms across leads/lags around policy dates. Stage 1 satisfies parallel trends pre-policy; Stage 2 required reverse analysis and excluding 2021 to satisfy parallel trends, with results remaining significant. (2) Counterfactual timing tests shifting policy dates ±1 year; insignificant DID terms support that observed effects are not driven by other shocks. The design avoids staggered dynamic treatment complications by using untreated groups and extended pre-trend checks, in line with recent DID best practices.

- Stage 1 (post-2015 upper law, lower rules unrevised): DID interaction terms are positive and significant, indicating increased time from publication to transaction for university patents relative to enterprises (lower efficiency). Without controls: DID ≈ +3.129 months (p<0.05); with controls: DID ≈ +3.348 months (p<0.05). Post and Treat main effects are also significant in several specifications. - Stage 2 (post-2019 alignment of lower rules with upper law): DID interaction terms are negative and significant, indicating reduced time to transaction (higher efficiency) for university patents relative to enterprises. Without controls: DID ≈ −4.172 months (p<0.01); with controls: DID ≈ −5.677 months (p<0.001). The absolute magnitude in Stage 2 exceeds Stage 1, suggesting efficiency not only recovered but improved relative to pre-reform. - Descriptive context: Mean transaction interval Y rose from roughly 42.3 months (Stage 1 sample) to ~48.9 months (Stage 2 sample), consistent with initial slowdown post-2015 before improvements post-2019 in the treated group as captured by DID. - Parallel trends: Stage 1 pre-policy interaction terms are near zero and insignificant; Stage 2 required reverse analysis (excluding 2021) to meet parallel trends; core results remain significant. - Counterfactual timing: Shifting policy intervention by ±1 year yields insignificant DID terms across models, supporting that observed changes stem from the studied reforms rather than confounders. Overall, inconsistency between upper and lower laws created a blocking effect that reduced university patent transaction efficiency after 2015; subsequent lower-law alignment in 2019 generated a transmission effect that enhanced efficiency beyond pre-reform levels.

The two-stage DID analysis shows that legal system inconsistency—upper law reform without timely lower-level revision—produces a blocking effect that impedes the implementation of the higher law. Universities, accustomed to established evaluation–approval practices under the older asset management rules, faced uncertainty and higher compliance costs after the 2015 reform, leading to slower patent transactions. Once the Ministry of Finance revised the lower-level measures in 2019 to align with the 2015 law, a transmission effect emerged: clearer, coherent rules facilitated faster transactions, improving efficiency beyond pre-reform benchmarks. These findings empirically validate that multi-tier legal coherence is critical for realizing legislative intents. Policy implications include: (1) Legal systems are symbiotic across levels; conflicts cannot be fully cured by abstract applicability rules alone because ordinary actors lack the capacity to resolve contradictions in practice. (2) Legislative revisions should consider system-wide coherence; lower-level laws and regulations must be updated promptly when upper laws change. (3) Establish connective legislative mechanisms (e.g., synchronized agendas, joint motions, and explicit cross-references) to reduce temporal gaps between upper and lower revisions. (4) Strengthen rule-of-law capacity in grassroots units, including legal consultancy systems, to guide behavior under temporary conflicts and reduce compliance costs. The study also acknowledges evolving DID methodologies and adopts design choices to enhance internal validity.

This study extends conflict-of-laws research by applying a two-stage DID framework to quantify how inconsistencies between upper and lower laws affect patent transaction efficiency in Chinese universities. It shows that unrevised lower-level rules can block implementation of higher laws, reducing efficiency, whereas timely lower-level alignment transmits and amplifies the higher law’s intended effects, improving efficiency. Methodologically, it demonstrates how policy evaluation tools (DID) can illuminate mechanisms of legal system performance. Practically, the findings argue for systematic, coordinated legal revisions and improved legislative techniques to minimize institutional frictions. Future research should broaden to other types of legal conflicts (same-rank conflicts, intra-law inconsistencies) and leverage big data to deepen empirical legal studies.

- Outcome coverage: The transaction efficiency metric is based on patents that do transact; it omits patents that never transact, limiting inference on the full distribution of commercialization outcomes. - Scope of legal conflicts: The analysis focuses on conflicts between upper and lower legal systems; it does not examine same-rank or intra-law conflicts. - DID assumptions and dynamics: Although parallel trends were tested (including reverse analysis for Stage 2) and counterfactual timings checked, residual concerns typical of DID (e.g., heterogeneous treatment effects, timing issues) may remain. The study mitigates these by using untreated controls, extended pre-trends where possible, and avoiding staggered dynamic treatments.