BONSANG Eric

We investigate the impact of cognitive load on occupational injuries using survey data. Cognitive load is defined in the literature as a tax on bandwidth which reduces the amount of cognitive resources available for engaging in logical reasoning. We proxy cognitive load with the number of nonprofessional tasks that individuals perform every day, conditional on the time they spend on them. The underlying assumption is that when individuals perform a large number of those tasks, this requires mental organization which keeps part of their working memory busy. We show that cognitive load increases the risk of occupational injury for both males and females. The effect is stronger for individualsin high-risk occupations and, among those, for low-educated workers.

The population of most countries is aging and the main caregivers are children. Policy makers are also promoting the provision of care by relatives to delay institutionalization of the elderly. We study the provision of care by children, how assistance is distributed among them, and its effectiveness in delaying institutionalization of the elderly. The main contributions can be summarized as follows. Retirement policies can have negative consequences for frail people with a high number of assistance needs who require daily attention. Children do not coordinate when deciding what to provide for their parent, and the allocation of the help they provide is then inefficient. The results suggest that children are strategic substitutes, but also that the strategic effect is greater for a sister than a brother. This can be explained theoretically by greater productivity in helping women than men. Policy makers should expect that stimulating informal care may accelerate the admission of elderly people with poor health.

Demographic aging raises questions about the sustainability of the pension system and support for the elderly. These issues have led to pension reforms and several developments in policies to support the elderly. This thesis aims to improve knowledge of the implications of these policies. The first part focuses on the impact of retirement on health. Chapter 1 summarizes the results of previous studies regarding the impact of retirement on health. Chapters 2 and 3 focus on the impact of delayed retirement on mortality and on health expenditures, respectively. They show that retiring later as a result of the 1993 reform in France has no impact on mortality and decreases health care expenditures for men. The second part of this thesis focuses on support policies for the elderly in France. Chapter 4 uses a microsimulation model to project the future needs for assistance to the elderly. The number of dependent elderly is estimated at 3.63 million people in 2060, with an expansion of morbidity. Chapter 5 looks at the impact of a change in formal assistance on the use of informal assistance. It shows that an increase in the amount of formal assistance decreases the probability of reporting receiving informal assistance.

Planning for old age needs involves ‘high stakes’ decisions such as the choice of a retirement plan, generic saving and investment decisions alongside the take up of insurance for long-term care . We argue that these individual decisions are formed upon limited information and, result from beliefs about future needs that are often not well understood. This paper provides an overview of a number of behavioral ‘regularities’ that illustrate areas where traditional economic approaches should include behavioral economic explanations to better understand and describe decisions for old age. These include the role of framing and reference points, the effects of misperception and behavioral learning, as well as the role of several biases such as optimism, present bias and overconfidence. We contend that choice architectures that incorporate such regularities can help making old age decisions.

We analyse the effects of retirement of one partner on time devoted to home production by both partners in a couple. Using longitudinal data on couples in Germany, we control for fixed household specific effects and allow for endogeneity of retirement using both a regression discontinuity approach and variation in the early and full retirement age across cohorts and with labour market history. Furthermore, we separately estimate the effect of retirement and the effect of lay-off due to an exogenous shock (i.e. plant closure) for each spouse. For both men and women, a transition from work to retirement significantly increases the amount of their own home production. Similar effects are found for other (layoff induced) exits from work. We find much smaller negative cross-effects of retirement or other exits from work on time spent on home production done by the male partner. the effect of the husband's retirement on the wife's home production is not statistically significant.

Policies to reduce health inequalities based on income redistribution or on reducing the costs of access to care for the poorest are common to many countries. These policies are based on the fact that there is a relationship between individual income and health status. This dissertation analyzes this relationship using the ≪ Survey of Health Ageing and Retirement in Europe ≫ which covers Europeans aged 50 years and older. We show that individual income is positively and concavely associated with health (absolute income hypothesis), but also that income inequalities within a country affect all individuals in that country (strong version of the income inequality hypothesis). The underlying mechanisms of this hypothesis show that to reduce health inequalities associated with income inequalities, governments must promote investments in human and social capital. Also, individuals are sensitive to the lifestyle followed by the majority of people. Subsequently, we implement a simultaneous analysis of health and income using a full information maximum likelihood estimator. The bidirectional causality of income and health is highlighted, as well as the presence of unobservable individual characteristics common to them. Finally, on a concrete example, that of anti-smoking policies, this thesis simultaneously analyzes tobacco consumption, individual income, and risk aversion. The results highlight the importance of individual preferences in the decision to smoke. Indeed, European smokers aged 50 and over are present-oriented, such that they do not consider the harmful effects of smoking on health, and are either risk averse due to anxiety, or risk-averse because they accept to alter their health.

Some studies have found that women outperform men in episodic memory after midlife. But is this finding universal, and what are the reasons? Gender differences in cognition are the result of biopsychosocial interactions throughout the life course. Social-cognitive theory of gender development posits that gender roles may play an important mediating role in these interactions. We analyzed country differences in the gender differential in cognition after midlife using data from individuals age 50 and above (N = 226,661) from 27 countries. As expected, older women performed relatively better in countries characterized by more equal gender-role attitudes. This result was robust to cohort differences as well as reverse causality. The effect was partially mediated by education and labor-force participation. Cognition in later life thus cannot be fully understood without reference to the opportunity structures that sociocultural environments do (or do not) provide. Global population aging raises the importance of understanding that gender roles affect old-age cognition and productivity.

This thesis is devoted to mixed health insurance systems where compulsory public coverage can be improved by a complementary health insurance. The questions addressed concern the inflationary effect of the complementary health insurance on the price of care and the impact of age-based pricing on solidarity between sick and healthy people and between income groups. The empirical analyses are based on French data. This original database includes the health care consumption of 99,878 MGEN members over the 2010-2012 period. Chapter 1 estimates the causal effect of better coverage on the consumption of excess fees and demonstrates the inflationary effect of complementary health insurance on the price of care. Chapter 2 considers the heterogeneity of the impact of better coverage on overcharges and its correlation with insurance demand. In fact, the inflationary effect of complementary insurance is accentuated by selection effects. Chapter 3 shows that age-based pricing maximizes transfers between the sick and the healthy at the expense of solidarity between high and low incomes.

In the context of a rapidly ageing population, this thesis explores the links between health and care methods for dependent elderly people and addresses the issue of financing dependence. Meeting the assistance needs of dependent elderly people is a central objective of public policy. Chapter 1 estimates the effect of informal (i.e. family) and formal (i.e. professional) home help on the mental health of dependent elderly people in France. The results show that informal help reduces the risk of depression and that formal help can improve general mental health. Recent studies recognize that helping a dependent relative has negative effects on the health of caregivers and emphasize the importance of supporting them. Chapter 2 focuses on the effect of social support on the health of informal caregivers. Finally, given the financial and fiscal pressure on public systems, Chapter 3 examines the extent to which Europeans would be able to finance their periods of dependency on the basis of their income and their financial and real estate assets. It also looks at the role of lifetime mortgages. The simulations show that only a small proportion of individuals would be able to finance all their LTC expenses. On the other hand, real estate assets could play an important role in financing LTC.