尽管多因素定价模型为学术界广泛接受,但其因子有效性源于公司基本面风险差异还是投资者预期差,依然存在分歧.选取账面—市值比B/M测度价值溢价,将其分解为账面—价值比(B/V)和价值—市值比(V/M),研究已有记载的7种风险假说是否能够充分呈现价值溢价的因果逻辑.结果发现,B/M对预期收益率的解释力与V/M相当,而B/V的差异则不能解释预期收益率差异.进一步地,财务困境风险假说、个股现金流风险假说,既不能一致地解释B/M之间的差异,亦不能解释V/M之间的差异;股权隐含违约风险假说、系统性风险敏感性假说、市场现金流风险假说、营运杠杆风险假说、股权久期风险假说这五类理论,虽可解释B/V的差异,但无法解释V/M的差异.基于线性外推偏差假说和套利风险理论,讨论了V/M在市场上产生和持续存在的原因.研究支持行为金融学中价值溢价成因的预期差假说,对以Fama and French(1993)为代表的多因素模型在中国A股市场上是否可由风险理论所解释提出质疑,由此基于"价值溢价消失之谜"这一前沿学术问题提供了重要参考.
Origins of Value Premium from a Book-to-Market Decomposition Perspective:Risk or Expectational Differences
Although multifactor pricing models are widely accepted in academia,debates per-sist regarding whether their effectiveness stems from fundamental risk differences among compa-nies or from divergences in investor expectations.This study employed the book-to-market ratio(B/M)to measure value premium and decomposed it into the book-to-value ratio(B/V)and the value-to-market ratio(V/M)to examine whether seven documented risk hypotheses ade-quately explain the causal logic behind value premiums.The findings reveal that B/M's explana-tory power for expected returns is comparable to that of V/M,whereas differences in B/V do not account for variations in expected returns.Further analysis shows that hypotheses such as finan-cial distress risk and individual stock cash flow risk neither consistently explain the differences in B/M nor in V/M.Theories such as equity implied default risk,systemic risk sensitivity,market cash flow risk,operational leverage risk,and equity duration risk fail to account for differences in V/M,although they explain variations in B/V.The study also discussed the generation and per-sistence of V/M in the market based on linear extrapolation bias hypotheses and arbitrage risk theories.This study supports the behavioral finance perspective that differences in expectations drive value premiums and questions the applicability of risk-based explanations,such as those presented by Fama and French(1993),in China's A-share market.Hence,this study contrib-utes significantly to the discourse on the enigma surrounding the disappearance of value premi-ums.
NETL
NSTL
万方数据
