@TechReport{iza:izadps:dp18718,
            	author={Castro, Rui and Kim, Jiyoung and Lange, Fabian and Larivière, Jérôme and Poschke, Markus},
            	title={Predicting Labor Force Types},
            	year={2026},
	month={Jun},
            	institution={Institute of Labor Economics (IZA)},
            	address={Bonn},
            	type={IZA Discussion Paper},
            	number={18718},
            	url={https://www.iza.org/index.php/publications/dp18718},
	abstract={A small group of people accounts for a large majority of flows between labor market states and of spells in unemployment and non-employment. In this paper, we ask whether it is possible to identify those who will be weakly attached to the labor market during their prime working-age years using information available earlier in life. Using the long panel of the NLSY79, we first identify individuals with weak labor market attachment between ages 30 and 50 based on their observed employment, unemployment, and non-employment histories. This reveals a group that is disproportionately female, less educated, and in poorer health. We then examine whether membership in this group can be predicted using information collected before age 30 that is not used in the classification itself. We find that information from ages 22 to 29 predicts later labor market attachment with high accuracy. Early labor market experiences and health are particularly important predictors. The ability to predict labor market attachment decades in advance suggests that persistent individual differences play an important role in shaping labor market outcomes throughout the life cycle.},
	keywords={labor market flows;labor market dynamics;labor market types;labor force attachment},
            }