Emotional intelligence, not technical literacy, best predicts worker productivity: well-being and engagement channel the gains, while AI intensity amplifies but cannot replace human-centered capabilities.

Main Finding

Emotional intelligence (EI) functions as a measurable form of human capital that strongly predicts labor productivity, employment quality and economic resilience. Using a machine‑learning approach, the authors find EI to be a dominant predictor that outperforms personality traits, AI/digital literacy and work‑environment factors. Psychological well‑being and work engagement substantially mediate the EI → productivity link. Digital/AI intensity and work environment amplify EI’s effects but do not substitute for EI. Ensemble ML models deliver better predictive performance than traditional econometric methods and provide explainable, nonlinear insights.

Key Points

• Conceptual contribution: Extends human capital theory by treating emotional intelligence as an economically relevant, intangible asset and situates it inside a behavioral‑economics framework (bounded rationality, adaptive performance).
• Mechanisms: EI influences productivity primarily through psychological well‑being (stress regulation, sustained mental health) and higher work engagement (vigor, dedication, absorption).
• Relative importance: EI explains more variation in productivity than personality traits, AI literacy, or simple measures of work environment—especially in knowledge‑intensive and service sectors.
• Context matters: Supportive organizational environments and higher digital/AI intensity strengthen the productive returns to EI but cannot fully replace socio‑emotional competencies.
• Methodology: The study emphasizes the value of machine learning—particularly ensemble models—to capture complex, nonlinear relationships and outperforms standard linear/econometric approaches.
• Policy orientation: Recommends integrating well‑being and EI development into education, workforce training, organizational practice and sustainable growth strategies (aligned with SDGs).

Data & Methods

• Analytical approach: Machine‑learning–based analysis using ensemble models (authors report that ensembles outperform traditional methods). The paper highlights the use of explainable ML techniques to interpret complex, nonlinear relationships among emotional, psychological and economic variables.
• Model structure: Independent variables include emotional intelligence and personality traits; mediators are psychological well‑being and work engagement; moderators include work environment and digital/AI intensity; outcomes are labor productivity, employment quality and resilience.
• Inference: ML models are used both for prediction and for extracting interpretable feature importance and mediation patterns (nonlinear effects emphasized).
• Methodological claim: Ensemble ML provides more robust, flexible modeling of behavioral determinants of productivity than linear econometric approaches.
• Limitations / missing details (in the provided text): the excerpt does not report data source(s), sample size, country/sector coverage, survey/instrument measures (how EI, well‑being, AI literacy, productivity were operationalized), specific ML algorithms, training/validation procedures, or performance metrics. These are important for assessing external validity and should be checked in the full paper.

Implications for AI Economics

• Complementarity with AI: Findings imply that socio‑emotional skills (EI) are complementary to AI and digital technologies—AI intensity increases the returns to EI rather than making EI obsolete. Models of labor market adjustment to automation should incorporate EI as a moderating factor.
• Human-AI collaboration: Policy and firm strategies for AI deployment should invest in EI development (training, job design, psychosocial supports) to maximize productivity gains from automation and reduce technostress.
• Measurement and modeling: AI economics research should routinely include measures of emotional intelligence, well‑being and engagement when estimating productivity returns to AI adoption; machine‑learning methods (with explainability tools) are useful to uncover nonlinearities and interaction effects.
• Welfare and growth models: Incorporate psychological well‑being as both an input and an outcome in models linking technology adoption to long‑run welfare and sustainable growth; EI-driven resilience can affect macroeconomic stability and human capital accumulation.
• Policy design: Workforce development programs should balance digital/technical upskilling with EI and mental‑health interventions. Evaluation of AI policy (retraining, reskilling) should measure psychological outcomes and engagement, not only technical skill gains.
• Research agenda: Empirical AI economics should test heterogeneity of EI effects across sectors, occupations, levels of automation, and cultural contexts; use longitudinal and causal designs to estimate returns to EI interventions in AI‑intensive environments.

If you want, I can: - Extract or approximate likely measurement instruments and ML methods the authors might have used (e.g., common EI scales, ensemble types, explainability tools), or - Draft specific empirical checks and robustness tests one should request or run to evaluate the paper’s claims (e.g., sample stratification by AI intensity, mediation tests, out‑of‑sample validation).