Psychobusiness: A Statistical Framework for Collective Psychological States and Economic Dynamics

This manuscript is a theory-building working paper. It proposes definitions, formal representations, hypotheses, and falsifiable research questions. It does not report empirical validation, predictive performance, or causal estimates. Psychobusiness is proposed as a theoretical and interdisciplinary research framework for studying how collective psychological states may interact with economic dynamics. The framework begins with a distinction between individual-level unpredictability and possible population-level statistical regularities, while rejecting the assumption that aggregation automatically creates predictability. A central object, the collective psychological state, Ψₜ, is defined as a statistical state of an economically relevant population rather than as a single sentiment score. The framework therefore separates latent individual psychological states from their observable indicators and from the population-level distribution obtained after an explicit aggregation procedure. It further distinguishes sampling or exposure weights from social-influence weights, thereby preventing the measurement of the collective state from being conflated with a proposed mechanism of influence. The resulting collective state may be summarized by central tendency, dispersion, dependence, entropy, polarization, and influence concentration, while preserving the possibility that the full cross-sectional distribution contains information that a scalar mean does not. Social interaction is represented through a time-varying network, and economic and psychological variables are modeled as a coupled dynamic system with feedback. The central predictive hypothesis is that the collective psychological state may contain incremental information about a future economic outcome after conditioning on relevant economic, exogenous, structural, and network information: I(Yₜ₊ₕ; Ψₜ | Xₜ, Gₜ) > 0. A second hypothesis focuses specifically on distributional information beyond the collective mean. The paper emphasizes identification, measurement validity, endogeneity, network uncertainty, temporal instability, and the reflection problem in social interactions. Predictive usefulness is explicitly separated from causality. The framework is intentionally theoretical; empirical validation is reserved for subsequent work.