NBER Publications by Guan Gong

This paper investigates to what extent and through which channels that health and educational attainment are interdependent. A dynamic model of schooling, work, health expenditure, and savings is developed. The structural framework explicitly models two existing hypotheses on the correlation between health and education. The estimation results strongly support the interdependence between health and education. In particular, the estimated model indicates that an individual's education, health expenditure, and previous health status all affect his health status. Moreover, the individual's health status affects his mortality rate, wage, home production, and academic success. On average, having been sick before age 21 decreases the individual's education by 1.4 years. Policy experiments indica...

This paper investigates whether subjective expectations about future mortality affect consumption and bequests motives. We estimate a dynamic life-cycle model based on subjective survival rates and wealth from the panel dataset Asset and Health Dynamics among Oldest Old. We find that bequest motives are small on average, which indicates that most bequests are involuntary or accidental. Moreover, parameter estimates using subjective mortality risk perform better in predicting out-of-sample wealth levels than estimates using life table mortality risks, suggesting that decisions about consumption and saving are influenced more strongly by individual-level beliefs about mortality risk than by group level mortality risk.

This paper investigates to what extent the differences in education between black and white men can be explained by the differences in their mortality risks. A dynamic optimal stopping-point life cycle model is examined, in which group-level mortality risk plays an important role in determining individual-level mortality risk, health expenditure,and the amount of schooling. The model is calibrated to quantify the effect of mortality risks on schooling by taking the black and white male population as the respective reference groups for black men and white men. We find that the impact of mortality risk on schooling explains more than two-thirds of the empirical education differences between black and white males. This conclusion is robust to a set of plausible parameter values.