Robust-efficient fitting of mixed linear models: Methodology and theory

In many areas of application mixed linear models serve as a popular tool for analyzing highly complex data sets. For inference about fixed effects and variance components, likelihood-based methods such as (restricted) maximum likelihood estimators, (RE)ML, are commonly pursued. However, it is well-known that these fully efficient estimators are extremely sensitive to small deviations from hypothesized normality of random components as well as to other violations of distributional assumptions. In this article, we propose a new class of robust-efficient estimators for inference in mixed linear models. The new three-step estimation procedure provides truncated generalized least squares and variance components' estimators with hard-rejection weights adaptively computed from the data. More specifically, our data re-weighting mechanism first detects and removes within-subject outliers, then identifies and discards between-subject outliers, and finally it employs maximum likelihood procedures on the “clean” data. Theoretical efficiency and robustness properties of this approach are established.     

Long-Term Effects of a Plant Closure: a Multistate Duration Analysis of Event History Data in a Ten-Year Follow-Up Study

This paper discusses the possible effects on future prospects in the labour market of those who suffer a plant closure. A multistate duration model is estimated on monthly event data comprising 172 individuals over ten years, with the occurrence of the plant closure among the regressors. An individual can escape to one of several destinations when exiting one of three origin states (unemployed, sick leave, outside the labour force). Our prime concern has been transitions leading to a new job or to destination states outside the labour force (disability pension and at home, without any social security benefit). We find indications of severe negative long-term effects on employment from a plant closure: long durations in unemployment when searching for new jobs, higher incidence of leaving the labour force, longer periods of sick leave and higher incidence of disability pension. Thus, an adverse shock reducing employment may have negative long-term effects on the labour market through long spells of unemployment and a shrinking labour force. Our findings thus support the hypothesis of hysteresis dynamics in the labour market.