The present study examines whether the Preston curve reflects a causal impact of income on longevity or, for example, factors correlated with both income and life expectancy. In order to understand the Preston curve better, we develop a model of optimal intertemporal consumption in which the representative consumer is subject to physiological aging. In modeling aging we draw on recent research in the fields of biology and medicine. The speed of the aging process, and thus the time of death, are endogenously determined by optimal health investments. We calibrate the model to US data and proceed to show that the model accounts for nearly 80% of the cross-country differences in life expectancy that the Preston curve captures.