CEO overseas experience, dynamic capabilities and corporate digital transformation: An imprinting theory perspective

The paper addresses why and how CEO overseas experience (CoE) affects corporate digital transformation (CDT) in Chinese listed manufacturing firms. Motivated by the high failure rate of digital transformation and executives’ strategic ambiguity, the study leverages imprinting theory to argue that CEOs’ formative overseas experiences imprint enduring cognitive and capability traits that shape strategic choices like CDT. The authors pose core questions: (1) What influence does CoE have on CDT? (2) Which contextual factors amplify or constrain this effect? They posit that CoE positively influences CDT (H1), that market competition intensity (MCI) and organizational slack resources (SR) strengthen this relationship (H2–H3), and that firms’ dynamic capabilities mediate the CoE–CDT link (H4). The study contributes by shifting attention from outcomes to antecedents of CDT, articulating imprinting mechanisms (cognitive and ability), and situating effects within China’s institutional context.

Prior work identifies external (e.g., digital technology availability, public R&D/technology spending, digital finance, regional/gender differences) and internal drivers (e.g., business model innovation, knowledge management, dynamic/organizational capabilities, leadership traits including overconfidence and myopia, chief digital officers) of digital transformation. Evidence shows public technology expenditure and digital finance spur CDT, with heterogeneity across ownership types and regions. Internally, leadership and managerial cognition shape CDT adoption and execution. However, top executives ultimately direct strategic choices, suggesting CEOs are pivotal. Imprinting theory posits that experiences during sensitive periods leave lasting cognitive and capability marks that guide later decisions. The authors argue CoE imprints broader cognition (global perspective, long-term orientation) and superior abilities (opportunity recognition, resource reconfiguration), which should favor CDT. They hypothesize positive main effects, moderation by competitive and resource contexts, and mediation via dynamic capabilities.

Design: Quantitative panel analysis of Chinese A-share listed manufacturing firms, 2007–2021. Data source: CSMAR; annual reports scraped via Python for dependent variable construction; historical missionary university data for instrumental variable. Sample: Listed manufacturing firms with exclusions (unclear firm nature; special treatment S, ST, ST, SST, PT; severe missing data). Panel is unbalanced; continuous variables winsorized at 1% and 99%. Measures: - CDT (dependent): Frequency of digital-transformation-related keywords in annual reports, covering two dimensions—underlying technologies (AI, blockchain, cloud, big data) and practical applications; summed and log-transformed to address right skew. - CoE (independent): Dummy =1 if CEO has any overseas experience (study or work). Further splits: CoESTD (overseas study), CoEWORK (overseas work). - Moderators: Market Competition Intensity (MCI) via Herfindahl-Hirschman Index; Slack Resources (SR) via equity–debt ratio (robustness: current ratio). - Mediator: Dynamic capabilities (DC) capturing opportunity sensing, resource integration, and organizational growth, operationalized with secondary financial/operational indicators; robustness: standardized mean of R&D intensity, technician ratio, ROA. - Controls: Firm size (ln assets), leverage (debt/asset), ROA, growth (sales growth), state ownership dummy, audit opinion, CEO duality, CEO gender, CEO age (ln), CEO education (coded 1–5), as well as year and industry fixed effects. Models: Fixed-effects panel regressions chosen via Hausman test; Driscoll–Kraay standard errors to address heteroskedasticity, autocorrelation, and cross-sectional dependence. Specifications: (1) CDT on controls; (2) add CoE; (3) add CoE×MCI and MCI; (4) add CoE×SR and SR; (5) DC on CoE and controls; (6) CDT on CoE, DC, and controls for mediation. Diagnostics: VIF ≈1.09 (no multicollinearity); Kao panel cointegration test indicates cointegration among variables. Endogeneity: Instrumental variable approach using the number/log of Christian missionary-founded universities per province (as of 1920) as instruments for CoE; 2SLS regressions confirm main results. Robustness: Alternative measures (CDT dimensions; SR via current ratio; DC alternative index), sub-sample analyses (high/low MCI; firm growth), time-window leads/lags, exclusion of 2008–2009 crisis years, varying sample periods, alternative model forms; results remain consistent.

• Main effect: CoE positively predicts CDT. Fixed-effects with D–K SEs show CoE coefficient β=0.0868 (p<0.05). - Moderation by competition: CoE×MCI is positive and significant (β=0.2242, p<0.05), indicating stronger CoE–CDT effects in more competitive industries. - Moderation by slack: CoE×SR is positive and significant (β=0.0032, p<0.05), implying greater slack amplifies CoE’s impact on CDT. - Mediation via dynamic capabilities: CoE → DC is positive (β=0.2854, p<0.01). DC → CDT is positive (β≈0.0250, p<0.10). In a joint model, DC remains significant (β≈0.0258, p<0.10) while CoE stays significant (β≈0.080, p<0.10), indicating partial mediation. - Further analyses: • Type of overseas experience: CoEWORK has a stronger effect (β=0.1175, p<0.01) than CoESTD (β=0.1104, p<0.01). • Firm growth: Effect is stronger in low-growth firms (β=0.1170, p<0.01) and insignificant in high-growth firms (β=0.0422, p>0.10). • CEO duality: Stronger when the CEO is also board chair (β=0.1709, p<0.01) vs non-duality (β=0.1118, p<0.01). • Academic background: Stronger for CEOs with academic backgrounds (β=0.1671, p<0.01) vs without (β=0.0873, p<0.05). - Robustness and IV tests uphold the main conclusions; most alternative specifications corroborate positive main and moderating effects (some SR interaction estimates vary under IV).

Findings support imprinting theory: CEOs with overseas experience exhibit cognitive imprints (broader perspectives, long-term orientation toward digital value) and ability imprints (opportunity recognition, resource orchestration) that translate into higher CDT intensity. Competitive markets heighten the salience of these imprints, making digital differentiation more valuable; organizational slack provides the resource cushion to act on CEOs’ imprinted preferences, enabling riskier, longer-horizon digital investments. Dynamic capabilities function as a capability base that connects CEO imprints to transformation outcomes: CoE enhances firms’ sensing, integrating, and reconfiguring abilities, which in turn facilitate CDT. Heterogeneity analyses show context and CEO attributes shape effect strength: practical overseas work experience, CEO duality (greater discretion), academic backgrounds (analytic and innovative proclivities), and lower firm growth (greater need or slackened short-term pressures) all magnify CoE’s impact. These insights underscore the strategic importance of executive life experiences for firm-level digital transformation in emerging-market contexts.

The study demonstrates that CEO overseas experience significantly accelerates corporate digital transformation in Chinese manufacturing firms, with the effect strengthened by higher market competition and greater organizational slack, and partially mediated by firms’ dynamic capabilities. Overseas work experience exerts a stronger influence than study experience; effects are more pronounced in low-growth firms, under CEO duality, and when CEOs have academic backgrounds. Contributions include: (1) identifying CEO overseas experience as a key antecedent of CDT from an imprinting perspective; (2) explicating cognitive and ability imprint mechanisms; (3) revealing how external (competition) and internal (slack) contexts and dynamic capabilities shape the CoE–CDT pathway. Future research could test generalizability across industries and countries, explore additional CEO traits (e.g., risk preferences, network centrality), incorporate richer capability measures, and examine regulatory and environmental contingencies.

• External validity: Sample restricted to Chinese listed manufacturing firms; results may not generalize to other industries or unlisted firms. - Time horizon: Fifteen-year window may be short for fully observing long-run imprinting and transformation dynamics. - Omitted variables: Despite extensive controls, unobserved factors may remain; endogeneity addressed via IV but not exhaustively. - Scope of mechanisms: Focuses on competition, slack, and dynamic capabilities; other CEO traits and environmental mechanisms (e.g., regulation) are not directly modeled.