Family Ties

In his new book, The WEIRDest People in the World (500 pages of text plus 180 pages of endnotes and references), evolutionary biologist Joseph Henrich says Westerners are psychologically different from other peoples of the world. He argues that the difference has its origins in teachings of the Catholic Church—in particular, the Church teaching in late antiquity and the Middle Ages that cousins shouldn’t marry each other. More, Henrich lays out the consequences of this specifically Western psychology and how it leads to advanced Education and Industrialization, to Riches, and to Democracy. (Get it? WEIRD.) These consequences arose because when cousins don’t marry each other, the intensity of kinship ties is lower, redounding to less mistrust of outsiders. Mistrust of outsiders operates like a kind of tax on commerce and on the economic development that is part and parcel of commerce. Where the tax is absent, commerce and development have more room to thrive.

The upshot, he claims, is that measures of economic development are high today in the places where proximity to the Catholic Church was high in the Middle Ages (and where, as a result, cousins didn’t marry cousins).

That’s a load to unpack, so let’s first put it in intellectual historical context.

In the High and Far-off times, O Best Beloved, as Kipling would have begun the story, we were all dirt-scratchingly poor. How, then, did it come to pass that some—but not all—of us are rich today? This central question of economic history remains open.

In the 4^th century B.C.E., Aristotle offered an explanation: It was variations in climate. In the 14^th century, Tunisian diplomat and scholar Ibn Khaldun weighed in; Montesquieu took a stab at it in the 18^th. Both of them cited climate variations, climate-driven culture, geography, and human physiology. In the 1830s, Tocqueville echoed the importance of culture while gently deriding the primacy of climate.

By the late 20^th century, however, culture and climate had been largely written out of economic development theory by the New Institutional Economics. Daron Acemoglu and James Robinson, two leading proponents of that school, summarize it in this way: “Although cultural and geographical factors may also matter for economic performance, differences in economic institutions are the major source of cross‐country differences in economic growth and prosperity.”

The term “economic institutions,” of course, is not well defined. The Nobelist Oliver Williamson described these institutions as “executive, legislative, judicial, and bureaucratic functions of government” whose salient features include the “enforcement of property rights and contract laws.” Williamson and another Nobelist, Douglass C. North, were perplexed, however by the role of non-institutional factors: specifically, the “pervasive influence” of cultural factors on “the long run character of economies.” Henrich’s analysis is intriguing precisely because he makes causal connections between those centuries-old cultural factors and the development of contemporary economic institutions.

Should we be convinced by his argument? We can’t examine all his references, but we can check a couple of them.

One is the connection between kinship and trust. Henrich cares about trust because trust is a component of the social capital on which societies are founded. Whom we trust, whom we don’t, and the way we treat those in one group or those in the other are the sum total of life experience begun in the arms of your mother—your first teacher, who was taught by her mother, and so forth. In this way, social conventions, a society’s defining features, are passed down from generation to generation. In general, we do not know why one society has one set of conventions and another a different set. But Henrich offers a testable hypothesis about the conventions of trust: It is that outsiders are distrusted more in societies where the bonds of kinship are strong than in those where such bonds are weak.

Henrich tests this hypothesis by looking for a correlation between kinship intensity and a measure of outsider versus insider trust. He estimates kinship intensity by averaging five different dimensions of social organization: preference for cousin marriage, polygamy, co-residence of extended families, rules of descent, and community organization. Then, he estimates outsider versus insider trust by measuring differences in responses to survey questions in roughly seventy-five different countries about how respondents trust insiders and outsiders.

Self-reported trust is unreliable, and differences between two unreliable data sets are doubly unreliable. Henrich makes no effort to convince us that his trust measure is worth taking seriously. Nevertheless, let’s ask: Does kinship intensity, as described by Henrich’s measures, correlate in the way he predicts with the differences between trust of outsiders and trust of insiders?

The first-level answer is yes: The kinship differences “explain” some 18 percent of the “trust” differences. Still, testing a hypothesis demands that we compare it to reasonable alternative hypotheses. Unfortunately, Henrich offers us no such alternatives. Can we fill the gap by finding and testing some reasonable alternatives to Henrich’s kinship hypothesis?

In an exceptionally long November 2019 research article in Science, Henrich and three co-authors developed the measures of kinship and trust that we discuss here. In December 2019 the article’s lead author, Jonathan Schulz, generously provided us with the raw data files. They show, among other things, that almost a third of the countries in the “trust” sample once had communist regimes and that almost a third were Muslim-majority nations.

Communism sowed distrust at every level of society, and Islam is not famous for its embrace of outsiders and their ideas. The notion that societies rich in communist legacy or Muslim tradition would show especially high mistrust of outsiders is, to say the least, not outlandish. So, how does a communist/Muslim hypothesis compare with a kinship hypothesis in explaining the difference between outside and insider trust? And what if we combine all three hypotheses?

It turns out that the communism/Muslim hypothesis explains 33 percent of the trust data. It’s better, in the sense of being more likely, than the kinship hypothesis by a factor of 100 to 1. As for combining all three hypotheses, that leads to statistical rubbish.

Thus, the kinship explanation for the difference between outsider and insider trust is, colloquially and technically, highly improbable. Maybe Henrich has inadvertently exposed something important about communism, Muslim populations, or both. Maybe other, unknown factors are at work. Maybe the data on outsider versus insider trust are themselves nothing but noise. We can’t know.

There is no doubt, though, that the failure of the kinship hypothesis undermines the larger thesis of Henrich’s book, which is the idea that lower kinship intensity born of Church doctrine drives the accumulation of social capital and, in turn, formation of the institutions responsible for the prosperity of WEIRD societies.

We can also examine another part of Henrich’s thesis: the direct connection between kinship and institutions, without stopping at the intermediate factor of outsider versus insider trust. A diverse set of measures of quality of life and governance institutions are published by the United Nations and World Bank. Thanks to high correlations, we may combine them all to form a unified measure, I. Then, we construct a large family of causal hypotheses about I and compare them.

Do the best—i.e., most likely—of these hypotheses involve kinship intensity? No. What if we compare the best hypothesis overall to the best hypothesis involving kinship? Here, the ratio of likelihoods is not 100 to 1, as it was when we compared the communist/Muslim hypothesis to the kinship hypothesis, but 100 to a vanishingly small number.

In other words, the thesis of Henrich’s book is false. Contemporary economic institutions are not the product of kinship intensity, and WEIRD outcomes are not an echo of medieval Church teachings.

In the 19^th century, when Durkheim, Engels, Marx, and others founded modern sociology with sweeping narrative theories of the clockworks driving us all, we did not ask—certainly for lack of data and method and possibly for lack of inclination as well—whether these theories were quantitatively sound. Fortunately, we are no longer so constrained. Henrich’s study of kinship is interesting for its attempt to connect the world of norms, particularly those driven by marriage patterns, to the world of economic institutions. Thomas Kuhn would have regarded this endeavor in economic history as an example of pre-science, a state of affairs in which competing schools of thought differ about basic assumptions, methods, and facts. Karl Popper, for whom falsifiability was the sine qua non, would have applauded it for its offering of hypotheses. But because the central hypothesis is demonstrably false, we must look elsewhere for answers to the central challenge of economic history—that is, the question of why the haves have, and the have-nots have not.

Daniel Seligson is a physicist who studies the role of nature, culture, and institutions in social and economic development.