Does Social Status Run in Families?

To Have and Have Not: The Determinants of Social Status in England, 1600–2026, by Gregory Clark (Econ Journal Watch, 320 pp.)

Gregory Clark, the economic historian behind A Farewell to Alms and The Son Also Rises, has released the final installment of his trilogy on economic development and social mobility. Continuing the series’ Hemingway theme, this one is called To Have and Have Not: The Determinants of Social Status in England, 1600–2026.

Unlike its predecessors, To Have and Have Not did not come out through Princeton University Press—nor through any academic book publisher, for that matter. The peer reviewers just aren’t having it. Some won’t even read it.

Clark’s manuscript instead appears online in PDF form, free of charge but with the occasional typo still intact, via Econ Journal Watch, a publication focused on debates within economics. Alongside it are a handful of responses from critics and supporters.

The book advances a provocative argument, drawing together several of Clark’s recent academic papers—and several enormous datasets—to claim that social status not only persists strongly within families but does so specifically through genetic inheritance, to a far greater degree than virtually anyone has claimed. Indeed, despite protestations to the contrary, Clark may come closer than any social scientist in the past half-century to warranting the label of “genetic determinist.”

To Have and Have Not is far from thoroughly convincing, even to readers who are open to the idea that genes play a large role in determining social outcomes. Yet one cannot help but be awed by the scope of Clark’s undertaking—and his willingness to stake out a bold position, insisting that vast swathes of existing research and conventional wisdom are wrong.

To Have and Have Not is built on an incredible trove of historical information, including centuries’ worth of English genealogies and marriage records, as well as the population registers of modern Sweden and Denmark. The sheer scale of these data allows Clark to see how social status runs in families—across many links in the tree, both horizontal (from siblings to distant cousins) and vertical (from parents to kids to grandkids)—and to measure status in several ways, including occupational prestige, education, and housing quality.

Of course, these measures are only loose proxies for the broad concept of “social status,” and even very close relatives often differ on them. Those differences—and how they change when looking at different types of relatives—are the key to understanding Clark’s analysis.

To take a typical example, in Clark’s data on Britons born from 1910 to 1996, siblings’ home values correlated at about 0.33. In other words, if someone has an extremely expensive house, three standard deviations above the mean, our best guess would be for their sibling to have a merely nice house, one standard deviation above the mean—closer to the overall average than to the sibling.

Most alleged “genetic determinists” of years past wouldn’t be too bothered by such a finding. They might point out, for example, that identical twins have much more similar economic outcomes than fraternal twins or normal siblings, a strong indicator that genes matter. But they would also happily concede that even identical twins often turn out differently, owing to things like luck and free will and fate and peer influences.

Clark, instead, looks at what happens as you move further out on the family tree. When one moves from siblings all the way to cousins in that same housing-value analysis, the correlation drops only modestly, to 0.21. The general rule he discovers—across many different status measures, and across several centuries notwithstanding massive changes to society and technology—is that these correlations, despite starting out at a modest level, drop by only about one-fifth with each extra degree of relatedness. (Cousins are two steps more distant than siblings in this sense.)

As Clark explains, the weak initial correlations likely stem from the fact that status measures are messy and imprecise. How many years of education an individual has is a good example here: as Clark puts it, “a first-class degree in Physics from Cambridge University is counted as the same educational attainment as a third-class degree in Mass Communications from the University of Bedfordshire” because both take the same number of years to complete. When variables are measured poorly, or when they fail to capture important details of an underlying concept, the resulting correlations will appear artificially weak.

Thus, our focus shouldn’t linger on the modest correlations among immediate family members. Those aren’t a beautiful testament to the randomness of life and the possibility of change; they’re just misleading. Rather, the very slow decay of the correlations as we move through family trees is where the action is. It reveals a “latent” status that is extremely persistent. Clark finds measurable links in status even among fourth cousins, whose shared ancestors are five generations back.

The high persistence of social status through extended families is one of the most important findings of Clark’s career and a major contribution to the study of social mobility. But the next question is: Why is status so persistent?

Genes, says Clark. And that’s where everything gets even messier.

Genes might seem like an odd choice to explain Clark’s findings. That’s for a simple mathematical reason: with each step through the family tree, genetic relatedness drops by half, but his correlations fall only in the ballpark of 20 percent. Siblings share four times as much genetic material as cousins, for example, so why would the aforementioned housing-value correlation decline just a third between them, if genes are mostly the determinative factor?

Clark’s answer is assortative mating. If people marry others who are similar in their “latent” genetic social status—presumably by matching on a mix of education, intelligence, social graces, work ethic, family connections, and so on—each link in the family tree doesn’t simply add genetic randomness but rather springs from a careful romantic match of like with like. Using a formula that goes back to the pioneering British statistician and geneticist Ronald Fisher, Clark is even able to calculate how strong matching on this latent status must be: to explain the slow decay of social status across degrees of relatedness through genes alone, spouses must be correlated on their genetic status at 0.57. If someone has genetic social status one standard deviation above the mean, he’ll marry someone of status 0.57 standard deviations above it, on average.

As some of the responses to Clark note, this is toward the high end of what appears plausible based on studies aiming to quantify assortative mating on status-relevant genetics. Nor is it the only possible explanation that would fit the data. Instead of extreme assortative mating plus highly genetic transmission, we could be seeing any number of social phenomena at work. Genes are not the only thing that family members share to a declining extent as you move outward through degrees of relatedness.

Clark’s case against this possibility is circumstantial but wide-ranging, combining further analysis of relatives’ status correlations with case-by-case attempts to swat down other social factors that could be at work.

For example, if status is transmitted socially instead of genetically, we might expect siblings (who share a social upbringing) to be more similar than parent-child pairs, despite both relationships representing just one step on the family tree. We would also expect relatives’ influence to vary based on whether the individual actually met the relative (such as in cases involving the death of a parent or grandparent), or for mothers and fathers to transmit status to different degrees, maybe even varying over time as women’s roles changed. In general, such differences are not evident in Clark’s data.

Social transmission might also imply birth-order effects. Younger siblings have to share parents’ attention for a greater portion of their lives, and so should be less able to soak up their parents’ status. This is also not generally evident in Clark’s data.

Perhaps similarities in family outcomes are partly driven by geography, such as the south of England (including London) being richer than the north? To the contrary, Clark argues that the north-south gap itself stems from a self-sorting process, in which the high-status have headed to the richer region.

Compulsory education expanded during the time period Clark studies, but it didn’t seem to lead to expanded social mobility. The research literature finding large returns from education expansion seems plagued by publication bias (meaning studies with small or negative results tend not to be published), and the true benefit appears small.

Clark’s own takeaways from his analysis include that “England is, and has been, close to a meritocracy for many generations,” that “much of the 10 percent of national income typically spent on education by high income economies is largely a social waste,” that genetic transmission of status implies “an incentive to select high status immigrants, and reject low status ones,” and that the “relative fecundity of people at the top and the bottom of social abilities” will affect future social outcomes. As if his main results weren’t enough to alienate the lefty eggheads who oversee academic publishing!

To buy Clark’s argument in full is to accept a lot of highly debatable claims that conflict with other strands of research. Beyond the natural skepticism one should feel toward Clark’s concept of a latent genetic status whose power relies on high levels of assortative mating—“an unobserved trait, or set of traits, that has some quite extreme properties,” as one response put it—one might point to other studies about the effect of growing up in some places rather than others, or getting more schooling, or even being born as an older rather than younger sibling.

Nonetheless, To Have and Have Not convincingly shows that social status persists strongly within families, and it expands the discussion on genetic causes, which almost certainly do play a role in this phenomenon. As the dust settles on these debates in the coming decades, I doubt Clark’s interpretation will be left standing in all its particulars, but his work and ensuing arguments about it deserve to shape the conversation in the meantime.