基于系统动力学模型的建筑工程从业人员信用行为研究

Research on Credit Behavior of Construction Engineering Practitioners Based on System Dynamics Model

摘要: 本研究聚焦我国建筑业工程质量安全事故频发的行业痛点，基于资源依赖理论揭示其核心诱因——建筑工程执业资格人员信用失范行为(典型表现为资格挂靠)，并阐释执业资格作为稀缺性无形资产，驱动从业人员与挂靠主体形成利益耦合关系的内在逻辑。为明晰信用行为决策的动态机制，研究引入系统动力学(SD)方法，构建涵盖内部驱动(个人认知、利益需求等)与外部驱动(市场秩序、监管力度等)的多维分析框架，提炼出7条核心因果反馈回路，其中“监管有效性提升→失信成本增加→风险偏好降低→信用水平提高”为关键正向调控回路。研究借助Vensim软件搭建系统流图并设定变量方程，以1天为仿真步长开展5000天时长的动态模拟。结果显示：1) 失信收益增幅与信用水平呈显著负向关联，且会加剧市场秩序紊乱；2) 制度完善速率呈现信用水平依存特性，信用水平提升后期制度完善增速趋于平缓；3) 主体风险偏好从初始中性状态逐步向厌恶型转变。政策仿真分析表明，信用系统主体回路存在阶段性演化特征——管理初期以导致信用水平下滑的正反馈为主导，后期逐步转化为推动信用回升的负反馈，且制度约束对信用行为的规制效能显著优于个人道德约束。结论证实，所构建的SD模型可有效模拟建筑工程执业资格人员信用行为系统的动态演化过程，验证了系统动力学方法在该研究领域的适用性；同时指出，鉴于信用行为的复杂系统性，研究对部分影响因素进行了合理简化，为后续深化研究预留拓展空间。

Abstract: This study addresses frequent engineering quality and safety accidents in China’s construction industry. Based on Resource Dependence Theory, it identifies the core cause as credit anomie (e.g., qualification affiliation) among qualified construction practitioners, and explains that professional qualifications—scarce intangible assets—drive an interest-based connection between practitioners and affiliation entities. To explore credit behavior decision dynamics, it adopts System Dynamics (SD), builds a framework with internal (e.g., personal cognition) and external (e.g., supervision) drivers, and extracts 7 causal feedback loops. A key loop is: “better supervision → higher dishonesty costs → lower risk preference → improved credit”. Using Vensim software, the study built a system flow diagram, set variables, and simulated 5000 days (1-day step). Results show: 1) Higher dishonest gains correlate negatively with credit levels and disrupt market order; 2) Institutional improvement slows as credit levels rise; 3) Risk preference shifts from neutral to averse. Policy analysis finds the credit system’s main loop evolves—early-stage positive feedback (lowering credit) gives way to later negative feedback (raising credit). Institutional constraints outperform moral ones. Conclusions confirm the SD model effectively simulates qualified practitioners’ credit behavior dynamics, validating SD’s applicability. They also note that due to credit behavior’s complexity, some factors were simplified, leaving room for further research.