We present new Monte Carlo evidence regarding the feasibility of separating causality from selection within non-experimental interval-censored duration data, by means of the nonparametric maximum likelihood estimator (NPMLE). Key findings are: i) the NPMLE is extremely reliable, and it...

We develop a new indicator of labour market tightness, based on the pure calendar time changes in individuals’ transition rates from unemployment to employment.Based on Norwegian register data from the 1989-2002 period, we show that this indicator,in contrast to the aggregate rate of...

We examine empirically the impacts of labor market policies - in terms of unemployment insurance (UI) and active labor market programs (ALMP) - on the duration and outcome of job search and on the quality of a subsequent job. We find that time invested in job search tends to pay off in the form...

The application of continuous time Weibull models on discrete unemployment duration data may produce bias in the estimated shape of the hazard rate. The bias can be substantial even for weekly duration data, and it is seriously aggravated if the Weibull model is erroneously mixed with a Gamma...

Building on register data describing monthly labour market status for the whole Norwegian population 1992-95, we estimate grouped competing risk hazard rate models for transitions between employment, unemployment and non-participation. The models impose no parametric restrictions on either...

We investigate how unemployment exit probabilities are affected by economic incentives, spell duration and macroeconomic conditions. Building on a database containing all registered unemployment spells in Norway in 1989-1998, we apply an econometric model in which exit probabilities vary freely...

We investigate how transitions from unemployment are affected by economic incentives and spell duration. Based on unique Norwegian register data that exhibit the rarity of random-assignment-like variation in economic incentives, the causal parameters are identified without reliance on...

When doing two-way fixed effects OLS estimations, both the variances and covariance of the fixed effects are biased. A formula for a bias correction is known, but in large datasets it involves inverses of impractically large matrices. We detail how to compute the bias correction in this case.

We present some theoretical results which simplifies the estimation of linear models with multiple high-dimensional fixed effects. In particular, we show how to sweep out multiple fixed effects from the normal equations, in analogy with the common within-groups estimator.

Large waves of refugees have arrived in Europe on a regular basis in recent decades. We know little about the impact of labour market policies intended to improve the labour market integration of refugees and their reunited family members. Using rich longitudinal data from Norway of the past 30...