The collusion behavior of venture capitalists and entrepreneurs based on "guanxi": evidence from China

Venture capital fosters entrepreneurship and innovation but may involve collusive behavior between venture capitalists (VCs) and entrepreneurs that harms minority shareholders and raises firm risk. Collusion in VC refers to unethical actions by VCs and entrepreneurs to advance self-interest, relaxing monitoring and potentially misleading investors. In China, where relational ties (guanxi) are salient and the secondary market has speculative features, guanxi may amplify such collusion, particularly at the exit stage. Research gaps identified include: limited focus on the exit stage, scant attention to VC–entrepreneur collusion versus other forms (e.g., shareholders, management), and incomplete understanding of how guanxi-based collusion forms and is governed. This study integrates social identity theory, social exchange theory, and unethical pro-relational behavior to ask: (1) Is guanxi conducive to inducing collusion between VCs and entrepreneurs during the exit stage? (2) What can restrain this collusive behavior? Using GEM-listed Chinese firms (2009–2021), the study examines whether guanxi increases entrepreneurial self-interest, VC returns, harms minority shareholders, and raises operational risk, and whether equity incentives can mitigate these effects. Contributions include focusing on the interactive VC–entrepreneur relationship, highlighting negative guanxi effects at exit, and introducing unethical pro-relational behavior to explain stakeholder harm.

Collaboration between VCs and entrepreneurs has been studied via principal–agent perspectives focusing on contracts and incentives; however, these formal mechanisms may be insufficient given information asymmetries and opportunism. Relationship governance, trust, social exchange, and compatibility influence cooperation and post-investment effectiveness. Guanxi, as strong ties formed through shared attributes (alumni, hometown, colleagues, political, association), improves information access and coordination but can weaken supervision and encourage irrational decisions or loose contracts, potentially harming performance. Evidence shows board–CEO guanxi can reduce monitoring, increase risk, and even facilitate misconduct; sharing insider information compromises market fairness. In VC, guanxi aids screening and syndication but may lower success rates due to friendly bias. Collusion is defined as illegitimate cooperation to maximize private gains at third parties’ expense; drivers include self-interest, low collusion costs, and weak governance. Prior prevention centers on external regulation and corporate governance. The study develops hypotheses: H1 guanxi increases entrepreneurial self-interest; H2 guanxi increases VC returns; H3 guanxi harms minority shareholders’ interests; H4 guanxi increases corporate operational risks; H5 equity incentives weaken these guanxi effects.

Design and sample: Empirical analysis on Chinese Growth Enterprise Market (GEM) firms with VC backing, covering listing year and two subsequent years (data through 12/31/2021). Exclusions: financial firms, ST/*ST, incomplete data. Final panel sample: 462 firms, 1,600 firm-year observations. Cross-sectional sample for VC exit returns: 715 observations. Data sources: CSMAR and CNRDS for financial/governance; CVSource and CNRDS for VC events; manual collection (annual reports, official sites, news) for personal backgrounds.

Measures:

• Collusion behavior (dependent dimensions):
  • Entrepreneurial self-interest behavior: Executive perks consumption (Cpc) = ln(sum of managerial expenses: entertainment, travel, office, socialization). Executive overcompensation (Overpay) = residual from compensation model with predictors: firm size, ROA, intangible asset ratio, state ownership, entrepreneur age, CEO-chair duality, and industry/year effects; entrepreneur pay proxied by ln(cash compensation).
  • VC returns (VCself): Average annual book value multiple at exit, BVM = (cumulative book value exit returns − cumulative investment amount)/cumulative investment amount.
  • Minority shareholders’ interests (Msexpr): Binary =1 if firm adopts both online voting and cumulative voting systems; 0 otherwise.
  • Corporate operational risk (Risk): Coefficient of variation of industry-adjusted ROA over rolling 3-year window (t to t+2), where ROA = EBIT/Assets; industry-year mean subtracted; standard deviation ×100.
• Independent variable: Guanxi = count of relationship types between VC leaders and entrepreneurs (0–5): alumni, colleagues, hometown, political, association.
• Moderator: Equity incentives (Hold) = chairman’s shareholding proportion.
• Controls: Entrepreneur characteristics (age, gender, education, experience), firm characteristics (size, profitability/ROE, leverage, intangible asset ratio, EPS), VC characteristics (gender, education, experience), VC institution characteristics (reputation/tenure, investment experience, ownership type). Year and industry fixed effects included. Logged transformations applied where indicated.

Models:

• Fixed-effects panel regressions for continuous outcomes Overpay, Cpc, Risk.
• OLS cross-sectional regression for VCself.
• Panel xtlogit for binary Msexpr.
• Moderation tests: hierarchical regressions with centered variables; inclusion of interaction Hold×Guanxi; significance assessed via change in R² or interaction coefficient.

Endogeneity and robustness:

• Propensity score matching (1:1 nearest neighbor) on firm/entrepreneur/VC covariates and industry to balance guanxi vs non-guanxi groups; re-estimate effects.
• System GMM for dynamic panels (Overpay, Cpc, Msexpr, Risk) including lagged dependent variables; Sargan and AR(2) diagnostics; Wald tests.
• Alternative guanxi proxy: binary indicator of guanxi presence to test robustness.

• Main effects of guanxi on collusion dimensions (Table 4):
  • Entrepreneurial self-interest: Overpay β=0.066 (p<0.01); Cpc β=0.092 (p<0.01). Supports H1.
  • VC returns: VCself β=1.782 (p<0.01). Supports H2.
  • Minority shareholders’ interests: Msexpr β=−0.127 (p<0.01), indicating impaired protection. Supports H3.
  • Corporate operational risk: Risk β=0.002 (p<0.01). Supports H4.
• Moderation by equity incentives (Table 5):
  • Overpay interaction Hold×Guanxi η=−0.112 (p<0.10); R²↑ 0.069→0.076.
  • Cpc interaction η=−0.580 (p<0.01); R²↑ 0.301→0.320.
  • VCself interaction η=−4.596 (p<0.10); R²↑ 0.159→0.163.
  • Msexpr interaction η=+3.965 (p<0.01); Chi²↑ 32.80→70.32 (weaker negative effect on minority protection).
  • Risk interaction η=−0.015 (p<0.01); R²↑ 0.194→0.203. These confirm H5 and sub-hypotheses H5-1 through H5-4: equity incentives weaken guanxi’s collusive effects.
• Robustness:
  • PSM-adjusted estimates (Table 6): Guanxi remains significant across outcomes (e.g., Overpay β=0.073 p<0.01; Cpc β=0.144 p<0.01; VCself β=1.823 p<0.05; Msexpr β=−0.864 p<0.01; Risk β=0.003 p<0.01).
  • System GMM (Table 7): Lagged dependent variables significant; Sargan p-values indicate valid instruments; no AR(2). Guanxi remains significant: Overpay β=0.048 (p<0.01); Cpc β=0.073 (p<0.01); Msexpr β=−0.106 (p<0.01); Risk β=0.002 (p<0.05).
  • Binary guanxi robustness (Table 8): Results consistent (e.g., Overpay β=0.217 p<0.01; Msexpr β=−1.568 p<0.01; Risk β=0.006 p<0.01; VCself positive and significant in alternative specs).

The findings directly address the research questions. Guanxi between VCs and entrepreneurs facilitates collusion at the exit stage by building trust and implicit reciprocal contracts, which relax monitoring and align incentives for short-term gains. This manifests in higher entrepreneurial self-interest (excess pay and perks), higher VC exit returns, reduced protections for minority shareholders (lower adoption of online and cumulative voting), and elevated operational risk. Introducing equity incentives aligns managerial interests with long-term firm value, mobilizes broader executive oversight, and improves transparency, thereby weakening guanxi’s collusive effects across all dimensions. The results underscore the ethical risks inherent in strong-tie relationships within venture transactions: guanxi can shift moral boundaries toward protecting insiders at outsiders’ expense (unethical pro-relational behavior), with adverse implications for market integrity and firm sustainability. These contributions nuance the predominant emphasis on guanxi’s benefits at the investment stage by highlighting negative consequences at exit and offering governance-based mitigation through equity incentives.

This study demonstrates that guanxi between venture capitalists and entrepreneurs fosters collusive behavior at the exit stage, increasing entrepreneurial self-interest, boosting VC returns, harming minority shareholders, and raising corporate operational risk. Equity incentives significantly mitigate these effects. The work extends theory by: (1) foregrounding interactive VC–entrepreneur dynamics and the negative role of guanxi during exit; (2) introducing unethical pro-relational behavior to explain stakeholder harm; and (3) evidencing equity incentives as an internal governance mechanism to curb collusion. Policy and practice should strengthen detection and deterrence of VC–entrepreneur collusion, and firms should adopt robust governance—especially equity incentives, clearer VC role definitions, and stricter controls on executive compensation and expenditures. Future research should broaden samples beyond listed GEM firms, refine and expand guanxi constructs (including digital ties), examine external governance (media, market, legal environments), and analyze ecosystem-wide interactions among venture participants using larger, cross-country datasets and network methods.

• Sample scope: limited to GEM-listed firms with VC backing; excludes non-listed startups and failures, constraining generalizability.
• Measurement of guanxi: restricted to five relationship types (alumni, colleague, hometown, political, association), potentially omitting digitally mediated or other relational forms.
• External governance not modeled: roles of media oversight, market governance, and legal/regulatory environments left for future work.
• Bilateral focus: emphasizes VC–entrepreneur dyad; broader ecosystem dynamics and multi-actor interactions warrant study; cross-country validation and big-data/network analytics recommended.