Studies of intergenerational mobility usually compare the status of a sample of adults (most often men) with that of their parents (generally fathers) to gauge the strength of the relationship between them. For sociologists the status in question is typically occupation or education, for economists, income or earnings. From these studies we know that intergenerational associations tend to be stronger in the United States than in social democratic countries like Sweden and Denmark, but not as strong as in developing countries like Mexico, Brazil, and China. In The Son Also Rises Gregory Clark addresses the estimation of the association in a novel way. He determines the social status of the holders of rare surnames at a point in the past and compares it with the status of those who bear the same names in later generations. Drawing on diverse sources of surnames he discerns a universal pattern: People with certain distinctive names occupy much the same position in the status distribution for many, perhaps hundreds, of years. The finding holds not only for countries often depicted as class-ridden, such as Great Britain and, latterly, the United States, but also for egalitarian Sweden, and it holds for western and non-western societies alike.

Clark documents a very slow decline in the over-representation of formerly elite names at the top of the status distribution, consistent with a parent–child correlation of social status of around 0.75, roughly double the usual estimates of the intergenerational correlations of incomes found in western countries. He reconciles the long-term persistence he documents with the apparently greater short-term mobility reported in most other research by arguing that measures of social position such as income, wealth, and education are manifestations of an underlying latent “social competence.” But because they are an imperfect proxy for social competence, estimates of intergenerational mobility based on them overstate the true underlying rate.

Clark points out that if intergenerational transmission is more or less constant over time and countries, attempts to explain change or variation in mobility are pointless, since what looks like differences between mobility regimes are, in fact, differences in how well the measured status in question reflects social competence. And the policy implications of this are stark: Contrary to what most of us think, differences in the organization of labor markets, in educational systems, and in welfare regimes make no impression on the intergenerational transmission of social competence. Clark's own policy prescription follows directly: What matters is not equality of opportunity but greater equality of condition, which can be secured by “reducing the rewards society generates for those of different abilities” (p. 274). This is strikingly at odds with the accepted wisdom of the past 30 years and, though it might appeal to those on the left, they are unlikely to find Clark's pathway to this conclusion congenial.

Clark's book is based on extensive and comprehensive empirical evidence. Successive chapters apply the analysis of surname data to Sweden, the United States, Great Britain, India, China, Taiwan, Chile, Japan, and Korea. This is impressive, even if the repetition of the same analytical approach tends to make for unexciting reading. Yet, there are reasons for skepticism about some of Clark's claims.

“(S)tatus inheritance,” Clark writes on page 126, “is indistinguishable in form from the inheritance of genetically controlled attributes. This is not to say that social status is determined genetically. But whatever drives it is, on the tests performed here, indistinguishable from genetic inheritance. Status may or may not be genetically inherited, but for all practical purposes, nature dominates nurture.” But genetic relatedness disappears quickly even among direct descendants. Someone in twenty-first century Britain named Pepys will share virtually no genes with a direct ancestor called Pepys who lived in the seventeenth century, unless the Pepyses have practiced a truly remarkable degree of genetic assortative mating over the centuries. The very long-term persistence of high social status probably has less to do with genes than to the advantages— material, cultural, and symbolic—that accrue to status and can be transmitted to subsequent generations.

In making the claim that the true intergenerational correlation of social competence is much higher than had been thought, Clark has to derive the correlation for the whole population from data showing rates of representation of surnames in particular occupations, such as medical doctors, over long periods of time. He does this in an ingenious way, but he relies on the assumption that the mobility of people with unusual names is representative of mobility as a whole. But people with unusual names are probably not representative. The link between parents and children's status is much stronger at the top, and probably also at the bottom, of the income distribution than it is on average. Clark's names come from these extremes and so his results for elite names, for example, should come as no surprise, but it is doubtful that they apply to whole populations rather than just to the elite.