A model of labor supply under uncertainty is developed, and comparative statics of current labor are carried out with respect to temporary and persistent wage change. This and a complementary analysis of measurement error suggest that individual wage growth leads to downward-biased estimates of intertemporal labor substitution. An alternative strategy, namely, the use of short-lived industry wage pulses in place of individual wage growth, is free of the above biases. Findings presented in the paper support this point of view. These results also suggest that intertemporal substitution has been undervalued as a source of cyclical changes in unemployment.

I analyze job search models with random layoffs in which employment opportunities are characterized by a wage and some measure of risk. Intuition suggests that a worker ought to demand a higher wage if he is to accept a job with a higher layoff rate; but this is not true in several models analyzed in the literature. I demonstrate here that assumptions about what happens immediately after a layoff and after a quit are critical in determining the relation between reservation wages and risk. Making these assumptions explicit clarifies the reasons why different models imply quite different predictions.