Boundedly Rational Meta-Learning in Sequential Consumer Choice

2026-05-15

AI Summary

This study addresses the challenge of accurately modeling how consumers transfer previously learned brand regularities across decision contexts—a process inadequately captured by existing models. To this end, the authors design a hierarchical experimental task and propose the Boundedly Rational Meta Dynamic Programming (BRMDP) model, which uniquely integrates bounded rationality with Bayesian meta-learning to represent consumers’ coarse encoding of prior uncertainty and their mechanisms for cross-contextual knowledge transfer. Combining behavioral experiments, Bayesian inference, and trial-level likelihood comparisons, the research demonstrates that BRMDP(1) best accounts for human behavior, revealing that while consumers do engage in cross-contextual knowledge transfer, they do not perform fully Bayesian integration. These findings offer novel theoretical insights into consumer learning and implications for marketing decision-making.

Abstract

Many consumer decisions are repeated choices under uncertainty. Standard models capture these decisions using Bayesian learning and dynamic programming: consumers update beliefs from feedback and use those beliefs to guide future choices. In many markets, however, learning does not restart when consumers enter a new context: prior experience with a brand, product, or provider can shape beliefs in later, related decisions. We study this cross-context knowledge transfer, or meta-learning, in sequential choice. We design a hierarchical laboratory task in which participants repeatedly choose among airlines across routes and observe noisy binary outcomes. Reduced-form evidence shows that participants improve not only within routes, but also across routes: they choose better airlines earlier in later routes and reduce pseudo-regret. To identify the mechanism behind this transfer, we compare human choices to a no-transfer benchmark and a fully integrated Bayesian meta-learning benchmark. In particular, we introduce a class of boundedly rational meta dynamic programming policies, BRMDP(D), that approximate full integration using a limited number of hyper-posterior draws, denoted by D. Trial-by-trial likelihood comparisons show that low-D boundedly rational meta-learning, especially BRMDP(1), fits participant behavior better than both no transfer and fully integrated Bayesian transfer. Consumers, therefore, transfer brand-level regularities across contexts, but through coarse representations of prior uncertainty. The findings imply that models of consumer learning should allow for approximate cross-context transfer, and that managerial counterfactuals based on either no-transfer or fully integrated learning can be misleading.

Problem

• meta-learning
• sequential choice
• cross-context transfer
• consumer learning
• bounded rationality

Innovation

• meta-learning
• bounded rationality
• dynamic programming
• cross-context transfer
• Bayesian learning