It is widely believed that health plays a major role in retirement decisions. The most important problem in including health in retirement models is the lack of availability of a good measure of health at the individual level in existing data sets. This problem is exacerbated when a model spanning multiple countries is desired, because self-reports on health may not be comparable across countries. Arguably, physical measures are less influenced by cultural and linguistic differences than self-reports on general health or even on health conditions. The authors develop a cross-country measurement model for health in which the relations between functional limitations, self-reports, and physical measures like grip strength are used to construct health indexes. Comparability across countries is achieved by using the physical measurements to define the measurement scales, and allowing other parameters to vary across countries to account for cultural and linguistic differences in response patterns. The usefulness of the health indexes is then investigated by including it in some simple retirement models.
Meijer, Erik, Arie Kapteyn, and Tatiana Andreyeva, Health Indexes and Retirement Modeling in International Comparisons. Santa Monica, CA: RAND Corporation, 2008. https://www.rand.org/pubs/working_papers/WR614.html.
Meijer, Erik, Arie Kapteyn, and Tatiana Andreyeva, Health Indexes and Retirement Modeling in International Comparisons, RAND Corporation, WR-614, 2008. As of January 11, 2023: https://www.rand.org/pubs/working_papers/WR614.html