The Historical Evolution and Financial Applications of Behavioral Catering Theory: From Dividend Premiums to Multi-Asset Corporate Decisions
Behavioral Catering Theory recasts corporate financial policy as a supply response to time-varying investor demand for particular security or firm characteristics. Its canonical formulation, developed by Baker and Wurgler (2004a, 2004b), explains dividend initiation and omission by the relative valuation premium assigned to dividend-paying firms. This survey reconstructs the theory’s evolution from that dividend-premium framework to a broader behavioral corporate-finance architecture encompassing payout substitution, nominal share-price management, capital structure, investment, corporate identity, and environmental, social, and governance (ESG) positioning. The review emphasizes the three conditions that make catering economically coherent: sentiment-driven relative mispricing, limits to arbitrage, and managerial objectives that place nontrivial weight on near-term market valuation. It then organizes two decades of evidence into four eras: genesis and early extensions (2004–2008), the risk/lifecycle counter-critique and international debate (2008–2015), multi-channel corporate applications (2015–2020), and a post-2020 “Behavioral 2.0” frontier shaped by digital investor attention, retail trading platforms, ESG preferences, and large technology firms simultaneously undertaking extraordinary capital expenditure and initiating cash dividends. The survey is deliberately discriminating about evidence. Some modern phenomena are direct tests of catering; others are best interpreted as new empirical environments in which the theory can be tested against lifecycle, agency, signaling, tax, and rational-risk explanations. Methodologically, the literature has moved from aggregate time-series OLS and Fama–MacBeth-style propensity models toward panels, dynamic specifications, quasi-experimental identification, and text- or search-based sentiment measures. We also show how time-varying parameter VARs, quantile connectedness, and frequency-domain decompositions can be imported into future catering research, while stressing that these methods are methodological opportunities rather than established canonical tests. The resulting synthesis identifies unresolved paradoxes concerning risk versus sentiment, dividends versus repurchases, governance as discipline versus transmission mechanism, and green preferences versus genuine long-horizon value creation. A research agenda is proposed for causal, cross-asset, high-frequency, and regime-dependent tests in the mid-to-late 2020s.