David Hugh-Jones

Under social-genetic assortative mating (SGAM), socio-economic status (SES) and genetically inherited traits are both assets in marriage markets, become associated in spouse pairs, and are passed together to future generations. This gives a new explanation for persistent intergenerational inequality and the “genes-SES gradient” – observed genetic differences between high- and low-SES people. We model SGAM and test for it in two large surveys from Great Britain and Norway. Spouses of earlier-born siblings have genetics predicting more education. This effect is mediated by individuals’ own education and income. Under SGAM, shocks to SES are reflected in the DNA of subsequent generations, and the distribution of genetic variants in society is endogenous to economic institutions.

Existing theories of the effects of the printing press treat it as speeding up the transmission of technical knowledge. This cannot explain why a large proportion of both manuscripts and early printed books was religious. We argue that books transmit prudential and moral rules as well as technical information. These culturally transmitted rules provide a foundation for economic rationality, and solve problems of trust in large markets. In Europe, cheaper book production stimulated not only scientific progress, but also new forms of religion, which used book reading to inculcate rules appropriate to the emerging modern economy. We model the effect of the printing press on economic growth. Initially religious works dominate, but as the stock of technical knowledge grows, the proportion of technical works increases.

We investigate natural selection on polygenic scores in the contemporary US, using the Health and Retirement Study. Across three generations, scores which correlate negatively (positively) with education are selected for (against). However, results only partially support the economic theory of fertility as an explanation for natural selection. The theory predicts that selection coefficients should be stronger among low-income, less educated, unmarried and younger parents, but these predictions are only half borne out: coefficients are larger only among low-income parents and unmarried parents. We also estimate effect sizes corrected for noise in the polygenic scores. Selection for some health traits is similar in magnitude to that for cognitive traits.