Predictability and (co-)incidence of labor and health shocks

Setbacks such as dismissal or illness can turn the lives of people upside down. This study shows that such adverse events can be anticipated in advance and that their occurrence is strongly interrelated. These insights suggest that social security policy should consider the fact that vulnerable groups are likely to face multiple difficulties at the same time. Using machine learning techniques and anonymous data on millions of Dutch people, this study maps out the entire probability distribution of a wide range of labor market and health shocks. The degree of inequality in risk exposure across the population is striking. Most people have a low probability of becoming seriously ill or dependent on social benefits, while one percent of people bears up to thirty times more risk compared to the population average. People with a flexible employment contract, low income, little wealth and migration background are overrepresented within this high-risk group.