Differentials in time preferences and related constructs are frequently invoked as unobserved potential reasons for socioeconomic gradients in economic outcomes. However, there is much less work actually measuring this. A number of recent papers have shown that measures of time preferences are predictive of various different types of behaviours and outcomes (see e.g. some papers involving me and colleagues here and here). However, there are no papers I am aware of that show that such factors can explain social gradients. A recent paper below by Van der Pol in Health Economics finds that controlling for measures of risk and time preference explains only a small fraction of the education gradients in health. I think this is an important result, which bears out the importance of thinking further about what is actually transmitted from parents to children in terms of determining their economic and health outcomes.
Health, Education and Time Preferences
Education has been shown to be the most important correlate of health. However, the mechanism through which education influences health has been largely unexplained. Grossman argued that education improves health production efficiency. In contrast, Fuchs argued that the association between health and education is not primarily causal but reflects unobserved causes of both outcomes. Instead of education causing better health, some ‘third’ variables may be related to both education and health. The ‘third’ variable most frequently mentioned is time preference. The aim of this paper is to investigate the role of time preference in the relationship between education and health. The role of risk attitude is also investigated. The paper exploits a unique data set of households that incorporated stated preference questions eliciting individuals' time preferences. The results show that the effect of education reduces but does not disappear when controlling for individuals' time preferences. Copyright © 2010 John Wiley & Sons, Ltd.