【摘要】 创新韧性是企业生存与发展的根基，董事网络紧密联结下所产生的数字化同群效应为赋能中国制造业企业创新韧性的提升开辟了全新路径。使用2012—2021年沪深A股上市企业数据，采用固定效应实证模型，探讨董事网络下数字化转型同群效应对企业创新韧性的作用影响、异质性和调节效应。结果显示，董事网络联结下数字化转型的同群效应有利于提升企业创新韧性。分别从制度环境、董事特征和行业特征三方面展开异质性分析，发现对于东部地区和所处行业竞争程度高的企业、董事成员含有女性和持股比例越高的企业、生产效率越高和所处网络的中心度越高的企业，董事网络下数字化转型同群效应对创新韧性的提升效果更为明显。此外，通过调节效应检验发现，当企业面临外部高度不确定性的环境时，将正向调节同群效应发挥的赋能作用。更多还原

【Abstract】 Innovation resilience—the capacity of firms to sustain innovation activities and even achieve capability leaps in the face of external shocks—is critical for survival and growth in the current era of global supply chain restructuring and intensifying technological competition. Digital transformation is widely regarded as a key driver of innovation,yet existing research offers conflicting views: some argue it empowers innovation by improving information processing and internal control, while others warn of information overload, technological dependency, and innovation silos. This study introduces a relational perspective by focusing on digital peer effects within director networks. In China’s relational culture,director interlocks serve as trust-based channels for information sharing, resource mobilization, and strategic imitation. The study aims to answer three questions:(1) Do digital peer effects within director networks enhance firm innovation resilience?(2) How do these effects vary across institutional environments, board characteristics, and industry features?(3)Does external environmental uncertainty moderate this relationship?The study constructs a panel dataset of Chinese A-share listed manufacturing firms from 2012 to 2021, drawing data from CSMAR, CNRDS, and Wind. After screening, the final sample consists of 3,215 firms with 17,410 firm-year observations.Innovation resilience is measured using Martin’s(2019) sensitivity index method. The core variable is the change in invention patent applications and grants, compared with the expected change based on city-level patent trends. The measure captures whether a firm outperforms the expected innovation trajectory.The digital peer effect within director networks is constructed in three steps. First, a director-firm two-mode network is built and transformed into a one-mode firm-firm network where ties indicate shared directors. Second, firm-level digitalization is measured by text analysis of annual reports using a Word2Vec-augmented dictionary of 639 digital keywords; the natural logarithm of keyword frequency is used as the digitalization index. Third, for each focal firm, the peer digitalization level is calculated as the weighted average of the digitalization scores of all firms sharing at least one director, with weights equal to the number of shared directors.Control variables include firm age, size, profitability, CED duality, accounts receivable ratio, managerial ownership,cash ratio, total asset turnover, and board size. Environmental uncertainty is measured as industry-adjusted volatility of abnormal sales over a five year rolling window.The empirical analysis employs two-way fixed-effects panel regressions with firm and year fixed effects. Robustness checks include alternative measures of innovation resilience, exclusion of firms with strategic disclosure behavior, controlling for industry-time and province-time fixed effects, and instrumental variable estimation using the one-year lagged length of provincial optical fiber cables.The baseline results show that digital peer effects within director networks have a positive and significant impact on firm innovation resilience. Specifically, the coefficient for Dig peer is 0.685(p<0.05) when using invention patent grants as the resilience measure, and 2.363(p<0.01) when using invention patent applications. These findings support Hypothesis 1.Robustness checks confirm the stability of the results. When alternative measures of innovation resilience(total patents, province-benchmarked) are used, the coefficients remain positive and significant. Excluding firms with zero digitalization scores or retaining only those with high information disclosure quality does not alter the conclusions. Controlling for industry-time and province-time fixed effects leaves the estimates unchanged. Instrumental variable estimation confirms the causal direction: the first-stage correlation between optical fiber length and Dig peer is positive and significant, and the second-stage results show that instrumented Dig peer still significantly promotes innovation resilience.Heterogeneity analyses reveal that the effect is stronger in firms facing high market competition and in those located in Eastern China. Firms with female directors or higher director ownership benefit more from digital peer effects. Additionally, firms with higher productivity and those occupying more central positions in the director network exhibit larger effects.The moderation analysis shows that environmental uncertainty positively moderates the relationship: the interaction term between Dig peer and Uncertainty is positive and significant for the grant based resilience measure(0.208, p<0.05),supporting Hypothesis 2. This indicates that the positive influence of digital peer effects on innovation resilience is amplified when the external environment is more uncertain.This study provides robust evidence that digital peer effects within director networks significantly enhance firm innovation resilience. By leveraging director interlocks, firms can access timely and diverse information, reduce uncertainty, and collectively manage the risks associated with digital transformation. The effect is particularly pronounced in competitive markets, economically developed regions, firms with diverse and incentivized boards, and those with high absorptive capacity and network centrality. Environmental uncertainty strengthens this relationship, highlighting the strategic value of director networks during turbulent times.Theoretical contributions include bridging digital transformation and innovation resilience through a relational lens, extending peer effect theory from natural groupings to socially constructed networks, reconciling conflicting views on digitalization’s role, and advancing the measurement of innovation resilience.For managers, the findings suggest that building and nurturing director networks should be a strategic priority. Board diversity and incentive alignment can enhance the receptivity to external ideas. In uncertain environments, actively learning from peer digital strategies becomes especially critical. For policymakers, fostering director network formation, tailoring policies to regional characteristics, and maintaining a competitive market environment can help strengthen firm innovation resilience.Limitations include the assumption of linear influence and the focus on manufacturing firms. Future research could explore nonlinear effects, other network types(e.g., supply chain or alumni networks), and deeper causal mechanisms.This study is among the first to empirically link director-network-based digital peer effects with innovation resilience,offering a novel perspective on how social capital can be leveraged for strategic advantage in the digital era. The findings challenge the notion that digital transformation is purely a technological or internal capability issue, highlighting instead its socially embedded and collective nature.更多还原