Mon, Feb 13, 2017, noon:
Daniel Almirall, "Getting SMART about adaptive interventions"
Stephens, Melvin. 2007. "Are there treatment duration differences in the Seattle and Denver income maintenance experiments?" B.E. Journal of Economic Analysis and Policy, 7(1): 58.
This paper re-examines the labor supply responses in the Seattle and Denver Income Maintenance Experiments (SIME/DIME). Specifically, the original experimental results show a significantly larger labor supply response for men and women from dual-headed households in the five-year Negative Income Tax (NIT) treatment relative to those in the three-year NIT treatment. Although typically thought of only as an NIT experiment, the SIME/DIME also included a job training experiment that enrolled roughly 60 percent of households, including both NIT treatment and control households. The original empirical specification imposed strong assumptions on the treatment response to the job training experiment in order to increase the precision of the estimated parameters. Once these assumptions are relaxed, the labor supply differences between men in the three- and five-year NIT treatments fall by over 50 percent in magnitude and become statistically insignificant. The analogous differences for women are almost entirely explained by these specification changes. Whereas the original findings of the SIME/DIME were inconsistent with the standard life-cycle labor supply model, the results of the re-analysis are mostly consistent with the model.