Consumption, Computation Mistakes, and Fiscal Policy
通过实验发现，人们在生命周期消费决策中会系统性地犯错误，例如过度折现未来收入，这对评估财政政策（如赤字政策）有重要启示。
An understanding of the correct model of intertemporal consumption choice is crucial to evaluating the effects of fiscal policies. The debates over whether deficit policy matters (Martin Feldstein, 1974; Robert Barro, 1974) and, if so, how to measure such policy (Robert Eisner and Paul Pieper, 1985; Kotlikoff, 1986) are fundamentally debates about the correct model of consumption. Unfortunately, distinguishing empirically between different consumption theories is a subtle business that has produced no strong conclusions. One problem confronting many tests of alternative consumption theories is that they require joint and quite specific assumptions about preferences, economic resources, and the consumer's information set that may not be justified. In such cases, what is described as a rejection of a particular model may simply be a rejection of restrictive assumptions placed on the model. A second problem that is also routinely swept under the rug involves the implicit assumption that consumers optimize perfectly given their preferences and resources, and that they correctly value their resources. In order to explore these more fundamental questions we, have conducted an experiment to determine whether individuals, when placed in a controlled life cycle setting, make consistent and coherent consumption choices, and whether they correctly value their future resources. The experiment provides negative answers to both of these questions; in the experiment subjects made significant and systematic errors in their consumption choice, reflecting, in part, an overdiscounting of future income. This paper reviews several of the findings from our 1987 working paper, and then discusses their implications for fiscal policy. We give a brief description in Section I of the experiment. Section II describes inconsistencies and errors in consumption choice and traces them to the overdiscounting of future labor income. Section III presents some regression results also pointing to an undervaluation of future resources. Section IV discusses the implications of these results for viewing fiscal policy and suggests the need for additional experiments as well as consumption models that acknowledge, rather than avoid computation problems.