The effect of parent job loss on health insurance coverage of children Susan Busch Partha Deb William Gallo Yale University Background • The U.S. economy lost 2.6 million jobs in 2008, with at least another 1.5 million in 2009 • Most displaced workers spend some time unemployed – reduced income – severance of non-pecuniary benefits (e.g., health insurance) • Adults often lose health insurance coverage as a result of job loss (Gruber, 1997 etc.) Background: children’s coverage • Little information on the effects of parent job loss on children’s coverage • Children may: – Lose employer sponsored insurance (ESI) – Lose coverage through private individual market (if premiums become unaffordable) – Gain ESI coverage through spouse – Enroll in public insurance programs, which are now available at higher income levels • Job loss is a trigger event for changes in children’s health insurance coverage (Czajka and Olsen, 2000) – Does not distinguish voluntary from involuntary job changes Objective • To examine changes in the insurance status of children whose parents experience involuntary job loss • We examine both: – For those children insured at time of job loss, may lose coverage – For those children uninsured at time of job loss, may fail to gain coverage Data: Medical Expenditure Panel Survey • • • • • Nationally representative household survey 2.5 year panel survey (5 waves) Approximately 18,000 individuals each cohort We consider two panels (2004-2006) Includes monthly information on insurance coverage • The subsample we consider are children less than 19 years of age with at least one employed parent at baseline Study Design • Exploit the longitudinal nature of the MEPS, and examine changes in insurance status for children post parent job loss • Our outcomes of interest are coverage 6 and 12 months post job loss • Transitions in and out of coverage may occur in the absence of job loss • Children with parents employed at baseline with no job loss are used as a comparison group – We randomly assign a ‘job loss’ date to children in comparison group • Adjust for weighting and complex sampling structure of MEPS • Use logistic regression to predict insurance coverage Measures • Involuntary Job loss. Self-report of loss of main job due to plant closing, business sale or layoff (measured at each wave) – Parent must be employed at baseline (wave 1) – We consider job losses between waves 1 and 3 to ensure sufficient follow-up – Precise date of job loss known – In some analyses we consider only individuals who lost their job due to plant closing or business sale (reduces concerns about selection) • Insurance status. Private, public and uninsured • Control variables. Job tenure (# years), age, self reported health status; cohort/panel; single parent; five income categories Any insurance coverage (adjusted), all children (N=7241) Baseline Month Month 6 12 (%) (%) (%) Children experiencing parent job loss Children not experiencing parent job loss 90 81 84 89 89*** 88* Any Insurance coverage (adjusted), children insured at baseline Baseline (%) Month 6 (%) Month 12 (%) Job loss 100 87.9 88.8 No job loss 100 97.0*** 94.9*** Job loss 100 88.7 86.7 No job loss 100 98.1*** 96.8*** Job loss 100 87.3 90.8 No job loss 100 94.4* 90.7 Children Insured at Baseline Privately insured at baseline Publicly insured at baseline Any Insurance, Children Uninsured at Baseline (%) Month 6 (%) Month 12 (%) Job loss 0 29.6 39.7 No job loss 0 24.0 35.3 Baseline Children Uninsured at Baseline Results • Among children who were insured in the month of job loss, significantly fewer are insured six months later (88 versus 97 %) • We find no significant change in public coverage rates in either group at 12 months; entire decline is for private insurance • Among children who were not insured in the month of job loss, children in job loss group may be more likely to gain coverage • Robust to limiting to just business sold or plant closings Limitations • We have a relatively small sample of job losers, so it is difficult to do sub-analyses • Would like to follow children for longer than 12 months • Differences in labor market may affect estimate (especially to the extent it takes longer to regain employment) • Do not know the effects this uninsurance has on children or their families Discussion • COBRA health benefit provisions and the nongroup market are unaffordable to many families. • Administrative mechanisms related to unemployment insurance may be used to extend coverage to job losers – The efficiency of extending coverage through an existing infrastructure may make the idea more attractive to policymakers • Because the risk of job loss is greatest during economic downturns there is presently a window of opportunity to increase public support for legislation that includes this population as part of a broader reform proposal Conclusion/Future Work • By adding additional MEPS cohorts, we hope to examine the specific types of coverage lost/gained • Additional analyses will examine whether these effects result in declines in health status and the receipt of necessary care Rates of private and public coverage (adjusted) Baseline (%) Month 6 (%) Month 12 (%) 61 58 45 58*** 47 57*** 25 26 23 25 25 25 Privately insured Job loss No job loss Publicly insured Job loss No job loss