The Dynamics of the Gender Gap at Older Ages
Paper Session
Sunday, Jan. 9, 2022 3:45 PM - 5:45 PM (EST)
- Chair: Claudia Olivetti, Dartmouth College
Caregiving and Labor Force Participation: New Evidence from Administrative Data
Abstract
A large share of the growing demand for long term service and supports is met informally by family members. While family-provided care may be an affordable and even preferable alternative to formal care, its toll on the economic well-being of caregivers, many of whom are also formally employed, may be significant. Family caregiving obligations may also contribute to gender gaps in employment and retirement outcomes, as the majority of family caregivers are women in their 50s, and women take on multiple caregiving roles across their lives. In this paper, we combine cross-sectional survey data about current and past family caregiving roles with longitudinal administrative data on labor market outcomes to estimate the dynamic relationship between lifecycle caregiving and employment. We identify current family caregivers using the 1996-2008 panels of the Survey of Income and Program participation (SIPP) and use family information on children’s ages and fertility histories to infer past caregiving roles. We then link SIPP respondents to their Social Security Administration administrative records to create a panel which allows us to observe annual employment-related outcomes for caregivers well outside the range of the SIPP survey—up to 30 years before and 10 years after the start of a caregiving spell. Our approach allows us to consider the pre-caregiving effect of anticipated care needs (“up-front” costs) separately from the ex post effect of realized caregiving. Additionally, the richness of the long administrative panel combined with detailed survey data allows us to consider various caregiving roles over the life course in tandem and evaluate their cumulative effect on the life-cycle labor supply of caregivers. We estimate how these multiple caregiving roles contribute to gender gaps in employment and retirement outcomes at older ages.Gender Inequalities in the Dynamics of Health and Employment near Retirement
Abstract
Declining health is a major risk faced by older workers. It carries consequences for productivity and employment and is potentially a crucial driver of inequalities in the years around retirement. This paper investigates gender inequalities in health and employment among older workers, how they are dynamically interlinked, and the mechanisms underlying the heterogeneous effects. These are important issues to study, for several reasons. First, the employment gap between those who suffer periods of ill-health and those who do not may exacerbate existing inequalities, especially along dimensions of gender and education. Second, the link between health and employment shapes the optimal design of retirement and health policies for older workers.We start by providing new evidence on differential exposure to health risks within and across gender, as well as on the short- and long-term effects of health on employment. We do this for the US and England, using comparable data from the Health and Retirement Study and the English Longitudinal Study of Ageing. We develop an innovative and rich empirical model to study the dynamic interactions between health and employment. This model has two main components. The first is a health process, which we formalize as a latent factor with a flexible specification that separates temporary and persistent health shocks. We use objective and subjective health measures to deal with concerns about endogeneity that arise when using self-reported, subjective measures of health. The second component accounts for labour supply at the extensive margin, with transitions in and out of employment dependent on past working status, monetary work incentives, and current and past values of health. Importantly, we allow for the transitory and persistent components of the heath shock to affect employment differently in a flexible specification. This new split of health shocks by their persistency is crucial to identify differential vulnerabilities to health shocks and to study the long-lasting effects of health on employment (Au et al., 2005, Disney et al., 2006, Bound et al., 1999 and 2010, or French, 2005). We estimate the health and employment parameters separately by gender, splitting as well by education to highlight key dimensions of inequality in these effects.
We then investigate and quantify the importance of some key mechanisms underlying the heterogeneous vulnerabilities to health across gender and education groups. We focus on three main drivers of these differences. First, different groups are differently exposed to severe health shocks. Our flexible model of health will allow us to inspect these risks. Second, the types of jobs that workers do may vary systematically in terms of their health requirements. A similar health shock may have very different consequences for ones’ ability to continue working depending on the type of job one has. And third, the financial incentives to continue working change quickly around retirement and in relation to health status. These financial incentives will likely compound the impacts of health on employment.
To disentangle and quantify the importance of these mechanisms, we build on previous work (Blundell et al. 2020) and develop a structural dynamic model of health, employment and earnings for older workers. This model formalises the both short- and long-run employment responses to changes in health. It features our flexibly specified health process, a detailed characterization of the monetary incentives to work and how they change as workers age, and a measure of the health content of jobs. We estimate this model for the US and England to study the relative importance of these mechanisms in different institutional environments.
JEL Classifications
- J1 - Demographic Economics
- J2 - Demand and Supply of Labor