Professor Shujing WANG from the Department of Economics and Finance at Tongji University’s School of Economics and Management (SEM) has published a collaborative research paper in Management Science, a world-leading management journal and one of the UTD 24 top academic journals. Titled Fundamental Anomalies and appearing in the journal’s February 2026 issue (Volume 72, No.2), the study tackles the long-standing “portfolio dependence” issue in traditional asset pricing models, providing a more rigorous framework to assess how firm fundamentals account for stock return anomalies.

The research team proposes a portfolio-independent parameter estimation framework based on Bayesian Markov Chain Monte Carlo (MCMC) methods, a departure from the conventional Generalized Method of Moments (GMM) that relies on test portfolio average returns. Instead, the study directly matches individual stock returns from 1967 to 2017, fundamentally eliminating parameter estimation’s reliance on specific portfolios.

Adopting the two-capital q-theory model and testing four parameter specifications, the empirical results show that the model with industry-specific and time-varying parameters delivers far stronger explanatory power: the average cross-sectional correlation coefficients between model-implied fundamental returns and actual returns hit 0.69 for decile portfolios and 0.43 for long-short portfolios, outperforming existing literature.

This estimated model generates significant size, momentum, profitability, investment and intangible asset premiums, covering most of the 12 widely studied asset pricing anomalies. However, it cannot produce notable value premiums nor explain the accrual anomaly even in its optimal form. The study also finds that investment adjustment cost parameters estimated via the portfolio-independent approach (0.14-0.73) are much lower than previous portfolio-based figures (2.84), aligning with direct estimation results in relevant research and validating Campbell’s (2017) critique that model explanatory power weakens when parameters are not optimized for specific anomalies.

The paper makes three pivotal contributions to asset pricing research: it develops a portfolio-independent parameter estimation method, offering a more impartial evaluation framework for investment-based asset pricing models; its empirical findings illustrate how parameter selection impacts assessments of model explanatory power, clarifying the applicability and limitations of q-theory models; and it identifies that parameter heterogeneity across industry and time dimensions is critical for models to match actual stock returns accurately.

Shujing WANG is a Tenured Professor and PhD Supervisor at Tongji SEM. She holds a PhD in Finance from the Hong Kong University of Science and Technology, a PhD in Physical Chemistry from the University of Rochester (USA), and a Bachelor’s Degree in Physical Chemistry from Peking University.

She is a recipient of numerous prestigious talent program honors, including the National Ten-Thousand Talents Program for Young Elite Talents, Shanghai Young Eastern Scholar and Shanghai Pujiang Talent Program. She has presided over Youth and General Programs of the National Natural Science Foundation of China, with her youth project rated Excellent in post-evaluation.

Her research focuses on big data analytics, asset pricing and FinTech, with publications in top international and Chinese academic journals – including five papers in Management Science – as well as Journal of Financial and Quantitative Analysis, Management World, China Industrial Economics and Journal of Financial Research, among others.