Research

The Inequality of Labor and Health Risks: A Probability-Impact Perspective

CPB Discussion Paper 461, coverage in ESB, Binnenlands Bestuur, Trouw, RTL, joint work with Brinn Hekkelman and Suzanne Vissers

We decompose the risk of labor and health shocks along an extensive margin (probability of incidence) and intensive margin (impact given incidence, measured as shock persistence). The two margins turn out to be positively correlated, revealing a group of individuals that is particularly vulnerable and poorly resilient to adverse events. Leveraging machine learning techniques and Dutch administrative data, we show that shock persistence is predictable at the individual level and that it is unevenly distributed. By integrating our analysis with individual-level shock probability estimates from previous research, we reveal a triple-layered accumulation of risk: (1) ex-ante probabilities of shock incidence are correlated across the domains of labor and health; (2) higher ex-ante shock exposure is associated with greater shock persistence; and (3) experiencing multiple shocks amplifies the likelihood of persistent impacts. These insights can help inform the design of pro-active and unified policies that support vulnerable groups.

Predictability and (Co-)Incidence of Labor and Health Shocks

CPB Discussion Paper 453, coverage in ESB, NRC, FD, Trouw, joint work with Emile Cammeraat, Brinn Hekkelman and Suzanne Vissers

This paper employs machine learning techniques to estimate the probability that individuals will be confronted with adverse labor and health events. We separate predictable components of shocks, interpreted as ex-ante risk types, from random components by contrasting ex-post shock incidence with ex-ante shock probabilities. Using rich administrative data on the entire Dutch population, we uncover three results. First, shock incidence is predictable especially in the labor domain and to a lesser degree also in the domain of mental and physical health. Second, risk exposure to shocks is very unevenly distributed, persistent over time and correlated with socioeconomic characteristics such as employment status, educational attainment, income and wealth. Third, bringing together risk estimates for twelve different shocks highlights that risk concurrence is sizable, monotone, non-linear and extended across domains. We discuss how our findings can help improve the targeting of prevention policies.

Wealth and consumption over the life cycle with pension plans and durable goods

Under revision [Working Paper]

Households allocate large portions of their income to illiquid durable goods and pension plans. Governments often subsidise or mandate participation in the latter to ensure adequate retirement saving. This comes at the expense of decreased resilience to adverse shocks and distorted accumulation of durables during working life. We construct a life-cycle model with pensions and durables to study the effects of different pension plan designs. Contributions are irreversible until retirement and non-collateralisable, while purchasing durables is subject to adjustment costs and borrowing constraints. Despite this illiquidity, households keep little precautionary savings because of utility from durables and longevity insurance from pension plans. We calibrate the model to the pension system of the Netherlands where households face a rigid contribution mandate. Our model replicates the illiquid financial portfolios of Dutch households and indicates that flexibilisation policies (such as voluntary participation, richer contribution mandates and early withdrawal) improve household liquidity and welfare.

Comparing imputation methods using expenditure surveys and unlinkable administrative data: an application to household consumption behaviour in the Netherlands 

Work in progress [Working Paper]

Imputation methods are often applied to study household consumption behaviour since administrative data based on tax records does not contain information on consumption, while expenditure surveys generally do not include sufficiently rich income and wealth information. Previous literature has imputed consumption with administrative data using the household budget constraint. We introduce an alternative approach where expenditure surveys are supplemented with information from unlinkable administrative data using overlapping variables. We compare the two imputation methods by exploiting a Dutch expenditure survey that is linkable to administrative data and apply both methods to study household consumption behaviour. We present evidence for heterogeneity in the marginal propensity to consume out of income and housing wealth due to the effects of low cash-on-hand, negative housing equity and mandatory pension fund participation. We conclude that more flexibility in the Dutch pension fund system would be beneficial to households facing adverse shocks and borrowing constraints.

Pension fund restoration policy in general equilibrium 

Published in Macroeconomic Dynamics (2020), Vol. 24 (7), pp. 1785-1814, with Ward Romp [Paper (Open Access)]

When the financial positions of pension funds worsen, regulations prescribe that pension funds reduce the gap between their assets (invested contributions) and their liabilities (accumulated pension promises). This paper quantifies the business cycle effects and distributional implications of various types of restoration policies. We extend a canonical New-Keynesian model with a tractable demographic structure and, as a novelty, a flexible pension fund framework. Fund participants accumulate inflation-indexed or non-indexed benefits and funding adequacy is restored by revaluing previously accumulated pension wealth (Defined Contribution) or changing the pension fund contribution rate on labour income (Defined Benefit). Economies with Defined Contribution pension funds respond similarly to adverse capital quality shocks as economies without pension funds. Defined Benefit pension funds, however, distort labour supply decisions and exacerbate economic fluctuations. While Defined Benefit pension funds achieve intergenerational risk-sharing, welfare analyses indicate that the negative effects of the induced distortions are sizeable.