The anatomy of the aggregate labor supply elasticity

We show that the aggregate Frisch elasticity of labor supply can greatly exceed the corresponding individual-level parameter, and we illustrate the “anatomy” of the former in terms of intensive and extensive margins. The methodology consists of using micro data from the PSID to construct a panel of individuals and an aggregate time series obtained by aggregating these individuals each year. These two data sets represent exactly the same sample at different levels of aggregation, and we use them to identify the parameters of two distinct MaCurdy-type micro and macro equations. We find a micro elasticity of about 0.1 and a much larger macro elasticity that ranges from 1.1 to 1.7. There is no conflict between the two estimates: the micro one reflects only the intensive margin while the macro one reflects, in addition, the much more volatile extensive margin. Furthermore, aggregation of only continuously employed individuals allows us to provide a reliable estimate of the intensive margin elasticity in the range 0.3–0.4. This implies an extensive margin elasticity in the range 0.8–1.4. These findings suggest that micro evidence is not a benchmark for assessing how large the Frisch elasticity of labor supply should be in a model of the aggregate economy.

► We construct a time series aggregating individual units in the PSID. ► We estimate the micro and macro Frisch elasticities of labor supply in a mutually consistent way. ► We find that the macro elasticity greatly exceeds the micro one, due to extensive margin responses. ► We document which socioeconomic groups contribute to an important extensive margin. ► We suggest that parameter estimates from micro data are not always appropriate for calibrating an aggregate model economy.