# Distribution Regression in Duration Analysis: an Application to Unemployment Spells

## TL;DR

This paper develops new inference procedures for distribution regression models in duration analysis with censored data, allowing for varying effects of explanatory variables over time, and applies it to unemployment duration studies.

## Contribution

It introduces a generalized approach to duration models that accommodates time-varying effects of covariates and provides inference methods for such models.

## Key findings

- Finite sample properties are validated through Monte Carlo simulations.
- Application to unemployment benefits shows significant effects on unemployment duration.

## Abstract

This article proposes inference procedures for distribution regression models in duration analysis using randomly right-censored data. This generalizes classical duration models by allowing situations where explanatory variables' marginal effects freely vary with duration time. The article discusses applications to testing uniform restrictions on the varying coefficients, inferences on average marginal effects, and others involving conditional distribution estimates. Finite sample properties of the proposed method are studied by means of Monte Carlo experiments. Finally, we apply our proposal to study the effects of unemployment benefits on unemployment duration.