This paper studies approximation error in the quasilinear utility model. Error arises because individuals do not perfectly optimize, and instead satisfice. We investigate the consequences of individual satisficing for modeling aggregate demand, providing an approximate aggregation theorem. We present a simple method for statistical inference on the minimal level of satisficing needed to explain aggregate data. In an illustrative application to scanner panel data, we find that individual‐level data require a nontrivial degree of satisficing, but aggregate data admit a representative agent that maximizes a quasilinear utility function.
Allen et al. (Thu,) studied this question.