This paper documents and attempts to explain the observed disparities
between unemployment rates computed from contemporaneous and retrospective CPS
data. The maintained hypothesis is that the discrepancies are consistent with
different definitions of unemployment between the two measures. The longitudinal
nature of the CPS, which allows a respondent's answers to be matched between one
year and the next, is exploited to examine two commonly expressed shortcomings
in the contemporaneous definition. I find that relative to the retrospective
measure, more workers with weak labor force attachment are considered unemployed
in the contemporaneous rate. In addition, discouraged workers, who are
classified as out of the labor force according to the contemporaneous definition,
may be counted as unemployed in the retrospective.

The US unemployment insurance system is financed by an experience-rated
payroll tax. The system imposes a layoff or firing cost on employers, leading
them to fire fewer workers in a downturn and hire fewer workers in an upturn.
Increases in the degree of experience rating are therefore predicted to reduce
the temporary layoff unemployment rate in a recession, and to lower employment
at the peak of the business cycle.
We combine Current Population Survey data for l979-1987 with a newly
assembled database of experience rating factors for individual states and
industries to measure the effects of imperfect experience rating on temporary
layoff unemployment rates over the cycle. We find a strong negative
correlation between the degree of experience rating and the rate of layoff
unemployment in recessionary years, but smaller and unsystematic correlations
in expansionary years. We also find that cyclical employment fluctuations are
dampened in states and industries with a greater degree of experience rating.

This paper assesses the ability of a simple search—theoretic
model to explain the results of two controlled reemployment bonus
experiments. The availability of two independent experiments
with substantially different treatments allows for a rigorous
test of the model. Parameters of the model are estimated by
minimizing the distance between the observed and predicted
aggregate response in each experiment, then cross-validated using
the observed and predicted treatment response from the other
experiment. The model is unable to predict an effect as large as
that observed in one of the experiments. In addition, the model
cannot explain the degree of individual—specific wage variability
found in the data. The relative success of models with and
without variable search intensity is also considered, but the
statistical procedures cannot distinguish between them.

In this paper, I consider the effect of changing the level of unemployment
insurance (UI) benefits on workers who do not receive UI. I hypothesize that
a spillover effect between insured and uninsured workers exists so that an
increase in the UI benefits, which leads to longer durations of unemployment
for insured workers, will lead to a reduction in the duration of unemployment
for the uninsured. This prediction is tested using data from several March
Current Population Surveys and the National Longitudinal Survey of Youth. In
both samples I find that an increase in UI benefits leads to a reduction in
the duration of unemployment for uninsured workers. Furthermore, using
several years of state level data, I show that the estimated effect on
unemployment for the entire labor force is roughly zero when I allow for the
spillover effect.