Dynamic Models of the Labor Force Behavior of Married Women Which Can Be Estimated Using Limited Amounts of Information

Pure cross-sectional model cannot capture the observed continuity over time in the work behavior of wives. Proposed alternative models generally require panel data not available in most countries. We present simulation results for models incorporating more limited information about previous behavior which could be collected on a recall basis in cross-sectional surveys, and establish that these models yield good forecasts of employment, hours of work and earnings of wives over ten years. One of our models may be viewed as a difference transformation of a standard work behavior model incorporating an inequality decision rule and modified to include fixed effects