Microfoundations and Macro Implications of Indivisible Labor

I show that the indivisible labor' models of Diamond and Mirrlees (1978, 1986), Hansen (1985), Rogerson (1988), Christiano and Eichenbaum (1992), and many others are, when aggregated across persons with the same marginal utility of income, equivalent to the divisible labor model of Lucas and Rapping (1969); any data on aggregate hours and earnings generated by the divisible (indivisible) model can be generated by some parameterization of the indivisible (divisible) model. The same is true when macro' data is obtained by aggregating over time and across people. This equivalence means that the indivisibility of labor per se does not have implications for macroeconomics. Nor does indivisibility have aggregate' normative implications. I then build a micro model of the bunching of work in continuous time as the consequence of fixed costs and fatigue effects.' Only in a special case does the micro model has as its reduced form the indivisible labor model. In other cases, the bunching of work in time may have unique macro implications. Indivisible and bunching models of labor are shown to have implications for public finance.