Using Expectations Data to Infer Managerial Preferences

Abstract

Forward-looking agents in econometric models of choice are generally assumed to have rational expectations so that the researcher can focus on the determinants of revealed preference. However, studies in psychology and economics find evidence that individuals may have systematic biases. We develop a structural econometric framework to recover behavioral parameters incorporating subjective expectations data that allows agents to have biased beliefs. We apply our methodology to analyze the advertising decisions of the marketing manager at a large non-profit performing arts center. A novel feature of this data is that the manager reports her expectations regarding advertising expenditure and ticket sales as as part of the annual budgeting process for each of the 146 shows we study. We jointly estimate the manager's utility parameters and the actual and expected demand functions. We find evidence that the manager is over-optimistic about the appeal of certain product characteristics and advertising effectiveness. Estimates of the manager's utility parameters are quite sensitive to the specification of expectations: a key behavioral parameter changes sign and is implausible when rational expectations is imposed. Our estimates also indicate that the manager manipulates advertising so that ex post ticket sales more closely correspond to her ex ante reported forecasts.

Keywords: Subjective expectations, biased beliefs, structural choice and demand models, advertising, non-profit, avant-garde art

JEL Classification: D8, L1, L3, L8, M37