Working Paper: NBER ID: w4434
Authors: James M. Poterba; Lawrence H. Summers
Abstract: This paper develops an algorithm for analyzing discrete events, such as labor market transitions, when some of these transitions are spurious because of measurement errors. Our algorithm extends the standard multinomial logit model, although our basic approach could be used with other stochastic models as well. We apply this algorithm to study the effect of unemployment insurance (UI) on transitions from unemployment to employment and out of the labor force. Our results suggest that VI lengthens unemployment spells by reducing both transition rates, and show that correcting for measurement error strengthens the apparent effect of VI on spell durations.
Keywords: unemployment benefits; labor market transitions; measurement error; multinomial logit model
JEL Codes: J64; H53
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Unemployment Insurance (UI) (J65) | Duration of Unemployment Spells (C41) |
| Unemployment Insurance (UI) (J65) | Transition Rates from Unemployment to Employment (J69) |
| Transition Rates from Unemployment to Employment (J69) | Duration of Unemployment Spells (C41) |
| Receipt of Unemployment Benefits (J65) | Reemployment Probabilities (J68) |
| Unemployment Insurance (UI) (J65) | Labor Force Exit Rates (J63) |