Three Essays on Retirement: The Patterns and Consequences of Job Insecurity and Employer Risk-Shifting

Abstract

This dissertation investigates how retirement behavior is affected by job insecurity and the shift of economic risk from employers to employees. Chapter I, “Household Responses to a Late-Career Job Loss,” demonstrates that spousal earnings affect an individual's decision to retire. I find that husbands with high-earning spouses are more likely to retire following an involuntary job loss. I augment earlier studies by considering how spousal earnings and household assets affect a worker's post-displacement labor supply. By developing a stylized two-period model, I illustrate how labor supply responds to spousal earnings and household assets in an uncertain environment. I test my theoretical model's predictions using data from the Household Retirement Survey (HRS) and a reduced-form empirical specification. Husbands with high-earning wives are more likely to exit the labor force following displacement than those married to low-earners, but this does not hold for older women. In both populations, a displaced worker with higher household assets is less likely to return to the labor force. For both sexes, job loss has a profound impact on retirement well-being. At a broader level, a reduction in the labor supply of older workers has negative fiscal consequences.

Chapter II, “Trends in Retiree Health Insurance: New Evidence from Household Surveys” deviates from past studies of individuals’ trends in access to employer-sponsored, early-retiree health insurance (ESERHI) that largely relied on employer-provided data. Because of this, direct estimates of levels and trends in access to ESERHI have been unavailable. Using data from seven waves of the HRS, I provide precise population estimates of levels and trends in the availability of ESERHI for individuals aged 55-64 between 2002 and 2014. I find that (1) declines in access measured at the individual level mirror offer rates at the employer level; (2) the diminishing probability of early-retiring individuals having health insurance coverage is less affected by their own employers withdrawing this coverage, and more affected by declines in the number of employers ever offering such coverage; and, (3) access to ESERHI declines more for some population groups, e.g., men, than others. I discuss the possible effects of declining access on individuals making the choice to opt for early retirement.

Chapter III, “Retirement Savings Responses to Liquidity Change and Consumption Needs” studies factors that lead households to contribute to their retirement savings or withdraw money from these plans (viz. defined contribution 401(k) and 403(b) plans, traditional and Roth individual retirement accounts (IRA), and annuities) prior to age 65. Households could be using these retirement accounts as another form of financial net worth accessible as a buffer stock to a range of cash flow and expenditure outlays prior to retirement. I find that the overall economic climate, household-level events, e.g. large unexpected out-of-pocket medical expenses, and household cash flow changes influence the decision to add or withdraw funds from defined contribution (DC) pensions. The ability to take pre-retirement withdrawals from these plans and reduce contributions raises the important question of how plan liquidity and discretionary participation affects retirement security.