Essays on work and retirement

This dissertation is a collection of three chapters on employment and retirement. The first two chapters are joint work with Marike Knoef and Daniel van Vuuren and focus on retirement decisions. The last chapter focuses on the relationship between employment and health. All three chapters make use of Dutch labor market reforms and empirical econometric methods to gain insights into the aforementioned relationships. In this section, I will explain the policy reforms, the empirical methods, and the main findings of the papers in successive paragraphs. The first two chapters focus on the labor force participation of older workers. Most OECD countries saw an increase in life expectancy and a decrease in the fertility rate during the last decades. This puts pressure on the sustainability of (pay-as-you-go) pension systems. To keep the pension system sustainable, most countries decided to implement policies that increase the labor force participation of older workers. The first two chapters investigate two policies that aim at keeping workers longer in the labor force. The first chapter analyzes the effect of an increase in the first pillar statutory retirement age1 on the labor force participation for couples. More precisely, we investigate how the younger spouse’s labor force participation changes if the older partner reaches the pension eligibility age. Due to a stepwise increase in the first pillar pension eligibility age in the Netherlands, we are also able to analyze the effect for different retirement ages of the older spouse. Our results show that the net labor force participation of the younger partner decreases if the older spouse reaches the pension eligibility age. Moreover, the results indicate that the magnitude of the effect is similar for cohorts regardless of the oldest spouse’s pension eligibility age. In addition, we find that younger partners more often stop working in high wage income households when the oldest spouse reaches the pension eligibility age. Lastly, we find some evidence that the initial retirement age of 65 still plays a role in the retirement decision (e.g., a social norm effect). The second chapter analyzes a different policy that aims at keeping older workers longer in the labor force. More precisely, the second chapter investigates the employment effect of Incentivized Gradual Retirement Plans (GRPs)2. The goal of GRPS is twofold. First, GRPs offer older workers a reduction in work time with little reduction in salary, and little to no reduction in pension build-up. In this way, older workers should be able to reach the pension eligibility age healthily and prevent them from dropping out of the labor force. Second, as older workers reduce their number of hours worked, employers need to post vacancies to hire new (young) workers. Therefore, this chapter analyzes the employment effects of GRPs. To do so, this chapter compares employment outcomes between Dutch municipalities that did and did not implement a GRP. We show that there is a positive employment effect for older workers (in terms of hours worked and the extensive margin). This positive effect is mainly concentrated in the middle- and high-wage income groups. For younger workers, however, there is no significant effect. We do not find any effect in terms of the hiring of young employees. In addition, we do not find any evidence that younger employees climb the career ladder faster due to the implementation of GRPs. The third chapter analyzes the effect of employment on health. Unemployed workers tend to have worse health than employed workers for a variety of reasons (selection effects, social norms). This makes it impossible to compare the health of employed and unemployed workers with each other. To investigate the relationship between work and health, I make use of an unemployment insurance reform. Due to this reform, some groups faced a decrease in the maximum unemployment insurance duration while for others it did not change. Using an instrumental variable design, I determine how an increase in job finding affects health. I find a reduction in medication use for pain and inflammation, hyperlipidemia (high cholesterol), and mental diseases (e.g., anxiety, depression, and tension). Therefore, job finding seems to have a positive effect on health. From a policy perspective, two broader conclusions can be drawn from this thesis. First, financial incentives have different effects on the labor force participation decision for different wage income groups. The results in chapter 1 show that high-wage income households have more flexibility in determining their retirement decision when compared to low-wage income households. In addition, chapter 2 shows that mainly middle- and high-wage income workers make use of GRPs. The question arises whether the design of GRPs is effective, since the financial means mainly end up with groups that probably do not need it to remain active in the labor force. Second, this thesis also shows that non-financial incentives are important. Chapter 1 shows that the initial statutory retirement age of 65 is important in the retirement decision, indicating that social norms play a role. In addition, the findings of Chapter 3 imply that the optimal unemployment insurance design is different once healthcare costs are considered. All in all, this thesis shows that both financial and non-financial incentives should be considered to properly understand the effects of policies.
-

Dit proefschrift is een verzameling van drie hoofdstukken over werk en pensioen. De eerste twee hoofdstukken gaan over pensioenbeslissingen. In het laatste hoofdstuk staat de relatie tussen werk en gezondheid centraal. Alle drie de hoofdstukken maken gebruik van Nederlandse arbeidsmarkthervormingen en empirische econometrische methoden om inzicht te krijgen in de bovengenoemde relaties.
Vanuit een beleidsperspectief kunnen twee bredere conclusies worden getrokken uit dit proefschrift. Ten eerste laat dit proefschrift zien dat de rol van financiële prikkels in het uittredingsgedrag in belangrijke mate kan verschillen tussen verschillende inkomensgroepen. Ten tweede toont dit proefschrift aan dat niet-financiële determinanten van arbeidsmarktgedrag ook belangrijk zijn. Al met al laat dit proefschrift zien dat zowel financiële als niet-financiële prikkels moeten worden meegenomen om de effecten van beleid goed te doorgronden.