This project deals with how the pension system design and the presence of long-term care risk affects individual homeownership decisions during retirement. It uses dynamic macroeconomic models to answer these questions and consists of two major parts: The first uses a stochastic lifecycle model to show how real estate can have an insurance function against long-term care risk and provides with its solution algorithm a detailed, economically intuitive model of the individual decision process. The second part is based on the steady state Overlapping Generations model with housing choice, extended to account for transitional effects resulting from changes in either the design of the pension system or coverage of long term care through the social security system. In their baseline version the models are parameterized to match German data, but can also be extented to other economies in order to identify the drivers of in homeownership-rate differences across countries.
The lifecycle model is closest in structure to the work of Cocco (2005) or Yao and Zhang (2004). Its approach differs from their work in two important ways: The first and most important difference lies in the way real estate is modeled. Both assume in contrast to my work that real estate is a risky asset and model housing services derived from owned and rented properties either to be equivalent (Yao and Zhang 2004) or even abstract completely from the possibility of rented shelter (Cocco 2005). In the baseline version rented properties are modeled to yield lower housing services compared to those derived from owned properties, which captures the fact that owner-occupied real estate usually exhibits a higher degree of customization than rental properties. The assumption of risk-free real estate is a direct result of my calibration of the model to Germany and does not need to hold true for other countries.
Second, long term care is included as an additional source of risk into the framework to show the insurance function real estate can serve. Long-term care is hereby modeled to induce cost in the form of a lump-sum payment and to be a permanent shock. Once an individual is hit by an adverse shock, she is subject to care for the rest of her life. The higher utility of owner-occupied housing services relative to rented housing services in combination with the assumption that real estate is a risk free asset makes it particularly attractive for individuals to own real estate up to high ages until they need long term care. Even though the literature has treated the link between medical expenses, housing and portfolio choice (Yogo 2016), the insurance component of housing has not yet been treated.
Fehr, H. and Hofmann, M. (2020): Tenure Choice, Portfolio Structure and long-term Care: Opimal Risk Management in Retirement, Journal of the Economics of Ageing, fothcoming