发表期刊：Management Science
发表时间：June 2025
作者及单位：Erica X. N. Li, Guoliang Ma(Paula and Gregory Chow Institute for Studies in Economics and Department of Finance in School of Economics, Xiamen University), Shujing Wang, Cindy Yu
摘要：This paper proposes a portfolio-independent method to estimate q-theory models, in which parameters are obtained using Bayesian Markov chain Monte Carlo (MCMC) to match firm-level stock returns. Our methodology addresses a previous critique on prior studies that model parameters are chosen to fit a specific set of anomalies and different values are needed to fit each anomaly. By targeting the entire sample of firm-level returns and allowing industry and time variations in parameter values, our estimations yield higher correlations between realized and fundamental portfolio returns compared with prior literature. Additionally, the estimated two-capital model generates large and significant size, momentum, profitability, investment, and intangibles premiums, but falls short in explaining the value and accruals anomalies. This limitation underscores the importance of portfolio-independent parameter estimation in evaluating a model’s capability to generate return anomalies.
关键词：q-theory; anomalies; Bayesian MCMC estimation; investment; profitability
