The Roots of Health Inequality and the Value of Intra-Family Expertise
Discussant: Marion Aouad
We use the presence of a health professional within the family as a broad measure of exposure to information about health. Intuitively, such exposure may capture intra-family transfers of knowledge (pure “information”), as well as persistent nudging or reminders about health investments (raising the “salience” of beneficial behaviors).
Our empirical setting is Sweden, in which the universal health insurance system allows us to shut down inequality in formal access to health care, and the administrative population-wide tax data linked to birth and medical records provide a rich set of information.
For identification, we exploit “admissions lotteries” into medical schools and variation in the timing of a medical degree. In Sweden, a centralized admission process usually generates sharp GPA thresholds for admission to any university program. Due to substantial grade inflation, however, the cutoff for all medical schools hit the top GPA starting in 2002, and admission was randomized within the group of applicants with the highest possible GPA. Our identification strategy leverages this randomization, by comparing family members of applicants to medical school with a top GPA who were admitted and not admitted to medical school. We also complement the admission lottery analysis with event studies that compare individuals' health before and after their child receives either a medical degree or a law degree.
We find that “getting” a doctor in the family leads to a substantial long-run improvement in health and mortality of older relatives. The parents of a medical doctor are 2.5 percentage points less likely to have died by age 80 twenty years after the child matriculates in medical school, which corresponds to a 16% decline in mortality. The parents of doctors are also significantly less likely to be diagnosed with heart attacks, heart failure, and diabetes. Similar improvements are documented in health and health capital investments in adolescence, early childhood, and in-utero. Further, we show that information affects individuals throughout the income distribution and that asymmetric exposure to information about health accounts for as much as 20% of the health-SES gradient.