Identification in multiple discrete-continuous extreme value models: An application to donation behaviours
(A2023-114516)

AUTHORS

Keyvan Dehmamy, Goethe University Frankfurt

ABSTRACT

Many studies use price variation as a sufficient input to identify a demand model motivating multiple discrete-continuous data. But identification is only partial: Models calibrated in this way cannot be used to make inferences about the focal customer after their disposable budget changes. A model in which all the parameters are estimated is no longer calibrated for one budget but can be extrapolated to an alternative budget set. Estimating the full model is possible only if the consumer's budget varies over repeated tasks. To the author, this study is the first to use this identification strategy. An intrinsic donation feeling is always given in a donation scenario, which evokes many different behaviours than conventional consumption. An empirical illustration confirms that intrinsic donations can be measured when the budget available for donation varies. This value differs by individual and depends on the type of organization, which determines the contributions to society that the consumer believes he or she is making. Furthermore, assuming the same intrinsic value for all individuals and the organizations they donate to is inaccurate; assuming a fixed budget for different donation sets also results in a worse model fit.