Background:
There is evidence of robust intergenerational correlates of body mass index (BMI) suggesting that health is persistent across generations. Since BMI is closely related to health behaviors that become healthier with higher socioeconomic status (SES), this raises the question of inequalities when poor health persists. We assume that social mobility (SoM) is associated with changes in health behaviors and hypothesize that the degree of transmission varies between mobile groups.

Methods:
We use longitudinal household-level data from the German Socio-Economic Panel. It provides information on self-reported height and weight, starting in 2002 (two-year cycle). We build a sample of individuals aged 15 to 35 years (12,108 person-years; 50.7% male) and link information on their parents. We calculate BMI as body weight (in kilograms) divided by height (in meters) squared. We measure SoM using offspring and parental education. Comparison leads to three groups: downward mobile (D), no mobility (N), and upward mobile (U). We estimate fixed effects panel models to obtain elasticities by SoM groups.

Results:
After controlling for sociodemographic and household characteristics, we obtain elasticities of 0.156 (D; standard error (SE): 0.043), 0.050 (N; SE: 0.022), and 0.047 (U; SE: 0.025) between mothers and offspring. A comparison with the father-offspring results shows similarities: Elasticities range from 0.137 (D; SE: 0.046), 0.022 (N; SE: 0.023), and 0.068 (U; SE: 0.036). The F-test shows significant group differences between the elasticities of downward and immobile individuals (p < 0.05).

Conclusion:
We obtain differences in BMI elasticities between SoM groups. While elasticities of immobile and upwardly mobile individuals are modest and not significant at the 5% level, a considerable degree of persistence can be found for downwardly mobile ones. For the downward movers, the change in BMI of the offspring is still influenced by the shared family environment, highlighting the need for behavioral prevention.