Herding and Male Dominance* Anke Becker November 11, 2017 (preliminary version)

Abstract Unlike other forms of pre-industrial subsistence, herding is a predominantly male task. Herding capital is mobile and hence needs to be protected, and there is a pronounced segregation between the sexes. This paper tests the hypothesis that, as a result of these characteristics, herding as a form of subsistence was conducive to male dominance over women in pre-industrial societies. First, it documents that in pre-industrial herding societies women indeed had less economic independence and a more subordinate position in their marriage. Moreover, kinship organization and cultural beliefs tended to favor men more with increased dependence on herding. Second, the paper provides empirical evidence for a persistence of this pattern until today. Using contemporary individual-level data, I show that descent from a herding society is associated with a preference for sons and less bargaining power of women within marriage. Moreover, women whose ancestral ethnic group depended more on herding animals are more likely to be have undergone the most invasive form of female genital cutting. Using data on ecological conditions measuring suitability for herding, the paper provides empirical support for a causal link between herding and male dominance.

JEL classification: I15, N30, Z13 Keywords: Gender inequality; cultural persistence; intra-household bargaining; female genital cutting;

* Harvard

University, Department of Human Evolutionary Biology; [email protected].

1

Introduction

The environmental conditions in which humans have lived historically have not only largely determined their way of living and their biology, but have also ultimately shaped their cultures (Boyd and Richerson, 1988, 2005; Harris, 1977; Henrich, 2015). This study examines the deep determinants of one particular dimension of culture that affects many aspects of our social lives and that varies considerably across cultures: the status of women relative to men. Thinking about the origins of an unequal division of power between men and women in society, anthropologists have long put forward a functional relationship between female status and subsistence style. Some have noted an association between matrilineal kinship structure and horticulture (Aberle, 1973) or between female subordination and having domesticated animals (Sanday, 1981). Others have pointed out the importance of female production to female empowerment, e.g. Sanday (1973) and probably most famously Boserup (1970). Ethnographies on pastoral people often mention the notoriety with which pastoralism is associated with low female status. This paper tests the hypothesis that herding, but not animal husbandry per se, was conducive to male dominance over women in pre-industrial societies. Using ethnic-group level data from two anthropological databases, I construct a measure for herding dependence and show that societies with stronger dependence on herding were indeed characterized by a male favoring kinship structure and a cultural belief in male superiority. Moreover, with increasing dependence on herding, women were economically less independent, had less control over household resources, and showed more deference to their husbands. Second, turning to contemporary measures of female subordination, I use individual-level data to explore the extent to which these measures are associated with historical dependence on herding. Based on information on individual’s ethnic affiliation I can assign historical herding dependence of ethnic groups to their descendants. The pre-industrial cultural pattern seems to persist until today. Women, whose ancestors depended more on herding have lower bargaining power in their marriage and a stronger preference for sons today. Moreover, they are more likely to have undergone infibulation, the most invasive form of female genital cutting. Conditioning on an extensive set of historical ethnicity-level and contemporary individual-level controls leaves the results unchanged. Motivated by a possibility that societies who already hold male-supremacist beliefs could be more likely to take up herding, the paper then tests whether it is indeed ecological conditions that determine historical herding intensity. Using data on land suitability from Beck and Sieber (2010) it documents that there is a strong and positive relationship between land suitability for herding and actual dependence on herding, providing empirical support for a causal interpretation of the association between female status and historical herding. In addition, using land suitability as an instrument for herding intensity generates estimates that are largely consistent with the OLS estimates, thereby providing further corroborating evidence for a causal relationship.

1

This paper adds to the literature on historical or cultural origins of contemporary variation in economically relevant outcomes today (Ashraf et al., 2016; Becker et al., 2017; Chen, 2013; Dell, 2010; Enke, 2017; Falk et al., 2017; Schulz, 2017; Teso, 2016). It is most closely related to studies examining more ultimate explanations, such as Alesina et al. (2013) who find that contemporaneous female labor force participation and corresponding cultural beliefs have their deep roots in pre-industrial plow use, or Galor and Özak (2016) who identify contemporary long-term orientation to be rooted in pre-industrial agro-climatic characteristics. In a similar spirit, this paper shows that it is the ecological conditions that determine whether a society takes up herding and hence ultimately generates a male-favoring culture. It also speaks to research on the persistence of cultural traits. Using data on second-generation American women, Fernandez and Fogli (2009) show that their beliefs and choices regarding labor market participation and fertility can be explained by their ancestral’s beliefs and choices. Nunn and Wantchekon (2011) examine contemporary levels of distrust in Africa and show that they can be traced back to the slave trade that generated such a culture. Similarly, Voigtländer and Voth (2012) use regional variation in anti-semitic violence in Nazi Germany and find robust associations with measures of anti-semitism during the Black Death period in the 14th century. Grosjean (2014) provides evidence for a persistence of a "culture of honor" (Cohen and Nisbett, 1994) introduced into the US South by immigrating Scot-Irish herders. By using similar measures of female autonomy in pre-industrial times and today and showing very similar associations of both classes of measures to historical herding intensity, this paper provides evidence that female subordination induced by a subsistence form that favors men persists until today. The remainder of the paper is structured as follows. Section 2 presents the historical and contemporary data sources. Section 3 describes pre-industrial herding both qualitatively and quantitatively, derives hypotheses, and examines the degree to which herding depends on ecological conditions as the identification strategy. In sections 4 and 5, I present the historical ethnicity-level and the contemporary individual-level results, respectively, as well as several robustness checks. Section 6 concludes.

2 2.1

Data Historical Data

All historical measures employed in this paper are taken from the Ethnographic Atlas or the Standard Cross-Cultural Sample. The Ethnographic Atlas (Murdock, 1967) is probably the most commonly used ethnographic database in cross-cultural research. It is an ethnicity-level database compiled by George Peter Murdock, drawing on work by many ethnographers, including himself. It contains information on historical subsistence, kinship organization, religious beliefs, settlement patterns, political organization, and institutional complexity for approximately 1,300 ethnic groups worldwide. The Ethnographic Atlas is regarded to reflect ancestral lifestyle as each society was portrayed prior to contact with industrialization, even when the exact timing

2

of observation differs across ethnic groups. For the empirical analysis, I will make use of two variables in the Ethnographic Atlas: the degree to which a society depended on animal husbandry and which animal was the predominant type in that society. Combining the two, I construct a measure of a society’s dependence on herding which will be the main explanatory variable in this paper. For a subset of 186 societies in the Ethnographic Atlas - the so-called Standard Cross-Cultural Sample - even more detailed information is available, in particular about gender related aspects of societies such as cultural beliefs or the nature of husband-wife interactions (Murdock and White, 1969). These will be my main measures of female status in pre-industrial societies.

2.2

Contemporary Data

For contemporary measures of female status, I will use the IPUMS sample of the Demographic and Health Surveys (DHS).¹ The DHS has detailed data on the health and well-being of women in developing countries and is therefore ideal to study contemporary female empowerment at the individual level. Until today, seven waves have been conducted, starting in 1984. The IPUMS sample consists of 22 African countries and India. For 13 African countries the data includes meaningful information on individuals ethnicity which can be matched to historical ethnicitylevel data from the Ethnographic Atlas or the Standard Cross-Cultural Sample and therefore to individuals’ ancestors’ dependence on herding. A detailed description of the matching of contemporary ethnic groups to the historical record is relegated to section D in the appendix.

3 3.1

Herding Herding in Pre-Industrial Societies

Herding is a form of subsistence that can be found in almost all regions of the world. Diverse ecologies such as the most northern regions of Scandinavia and Russia, the steppes of Eurasia, the deserts of the Arabian Peninsula and Northern Africa, or the Andes in South America were (and still are) homes to herding people. Herd animals cover a broad spectrum of species, from smaller ones like goats and sheep to larger ones like horses, reindeer, donkeys, camels and camelids, not to mention the large variety of cattle.² While herding societies in general are ¹https://www.idhsdata.org/idhs/ ²According to Barfield (1993), the Old World can be divided into five herding zones: (1) the cattle raising zone south of the Sahara, in the Sahel across the African continent, and in and around the Great Rift Valley in East Africa; (2) the camel herding zone in the Saharan and Arabian Desert; (3) a sheep and goat herding zone along the Mediterranean littoral through the Anatolian and Iranian Plateaus into mountaineous central Asia; (4) a horse herding zone in the Eurasian steppe running from the Black Sea to Mongolia; (5) the Tibetan Plateau with herding of yaks, sheep, goats, and horses at high altitude. For the New World, Kardulias (2015) identifies three to four zones. Here, the Andes are the only place where people tended herds in prehistoric times (camelids like llamas and alpacas). Horses were brought to Mexico and the Southwest of North Americas by the Spanish, from where they eventually spread to the Great Plains. Finally, next to horses, the Spanish also brought sheep to the North American Southwest.

3

almost never fully sedentary, some practice a more localized transhumance lifestyle, others are semi-nomadic, while again others are permanently mobile people (Hall, 2015). Typically, herding societies are not entirely dependent on animal resources, but most of them additionally also subside on horticulture or some other form of agriculture. For 1,202 societies, the Ethnographic Atlas contains information on the degree to which a society depends on animal husbandry, measured on a 10-point scale. In order to create a measure for how much a society depended on herding, I make use of a second variable: the predominant animal in a society. I generate an indicator that takes value 1 if the predominant type of animal is a herd animal (e.g. sheep, goats, donkeys, horses, reindeer, cattle, or camels/camelids), and 0 if it is an animal that is not herded (e.g., pigs, bees, poultry, or guinea pigs). Multiplying this indicator with dependence on animal husbandry gives my measure of historical dependence on herding which I will use throughout the paper. Figure 1 illustrates the variation in dependence on herding of 1,202 societies in the Ethnographic Atlas. About one third of societies do not herd at all, and about 5% do so to only very small extent. Most societies range between 10% and 50% in their dependence on herding, and there are only some societies that almost exclusively depend on herding.

Figure 1: Distribution of herding dependence across 1,202 societies in the Ethnographic Atlas.

Despite the variety of environments in which herding is practiced, the variation in animal species that are herded, and the differences in settlement patterns across herding societies, three aspects are characteristic of all forms of herding. (i) Men do almost all and often all the work related to subsistence, i.e. to the animals, resulting from the incompatibility of many of the tasks with child rearing, in particular with being pregnant or nursing an infant. To illustrate, many tasks either involve a certain risk of injury (e.g. protecting the herd from predators or thieves at night, castrating bulls, shepherding large animals like long-horn cattle) or long and potentially strenuous walks, e.g. finding new grazing grounds or watering places. Table 1 illustrates how pronounced the gender difference in terms of contribution to subsistence is in herding for societies in the Ethnographic Atlas. In only

4

Table 1: Sex Differences in Herding, Animal Husbandry, and Agriculture (Ethnographic Atlas)

Predominantly Male

Predominantly Female

Participation Equal but Differentiated

Participation Equal, No Differentiation

Herding

67.80%

6.12%

16.55%

6.58%

Animal Husbandry With Non-Herd Animals

17.91%

56.72%

1.49%

14.93%

Agriculture

33.02%

35.04%

12.35%

19.33%

Subsistence

6% of societies is herding a predominantly female activity.³ Note that, even when participation is equal, it is differentiated by gender more than twice as often than it is undifferentiated. Hence, women might be in a weaker position for two reasons. One, they simply contribute less to subsistence and hence have less say over produced output or surplus. Second, since most of the time their work is not related to the animals they potentially lack cultural knowledge that is of strategic importance to the survival of the group. Again, this potentially weakens their bargaining position relative to men. Sections 4 and 5 will provide evidence for the hypothesis that women are economically and politically less independent, have a subordinate position in marriage relative to their husbands, and that these aspects are then reflected in cultural beliefs in male superiority. (ii) In herding, capital is mobile and hence needs to be protected, a typically male task. It is also fungible and trading some animals in exchange for other resources, e.g. during dry season, is common in most herding societies. Hence, male ownership of the capital, which is common in most pre-industrial societies regardless of subsistence form, is of particular strategic relevance. Women depend on the goodwill of their male family or clan members to get access to strategically important resources and are therefore in a relatively weaker bargaining position. However, while capital is mobile and fungible, herd size has to be managed carefully. Not only are there economies of scale that make larger herds desirable (see for example Borgerhoff Mulder et al. (2010) and references therein) but it is also important to maintain a certain minimal herd size especially in larger species with longer birth intervals such as camels (Chatty, 1973), or as an insurance against adverse shocks such as droughts (Fratkin and Roth, 1996). Since men own and inherit the capital, patrilocal residence is advantageous over alternative forms of postmarital residence as this facilitates keeping the herd of the extended family together.⁴ As the following section will show, this seems to be borne out by the data. Another immediate hypothesis following from this is that sons should be preferred over daughters. For one, can they better help with important tasks of pastoral people such as protecting the herd. Moreover, for women, a preference for sons might also simply result from the fact that sons, who live with their parents after marriage, might serve as a guarantee for access to resources in widowhood when their husbands’ wealth has been transferred to their sons. ³Note that this contrasts strongly with animal husbandry without herd animals and agriculture. The former is a predominantly female form of subsistence in 57% of cases, and the latter in 35%. ⁴In patrilocal societies, husband and wife live with the husband’s family after marriage.

5

(iii) The third defining aspect of herding is a strong segregation between the sexes. Men not only spend the day out with the animals. They often have to stay with the herd at night to protect them from predators or from thieves. Sometimes, men will have to leave camp for a couple of days or even weeks to take the animals to new feeding grounds or water holes. As opportunities for mate guarding decrease, paternity uncertainty increases. Higher paternity uncertainty alone does not imply lower female status per se.⁵ However, in a situation in which women strongly depend on men but not vice versa, paternity uncertainty might cause women to be willing to take up costly signals of fidelity so as to ensure paternal investment into their offspring (and investment into themselves as well).⁶ Several cultural practices that can be interpreted to send such costly signals have been hypothesized to be associated to situations of increased paternity uncertainty: the complete veiling of women, Chinese footbinding, and infibulation, the most invasive form of female genital cutting (Hicks, 1996; Mackie, 2000; Shell-Duncan and Hernlund, 2000). Section 5 will present empirical results that are in line with this hypothesis. In sum, the pastoral lifestyle seems to impose features that put women at a disadvantage within the family and within society overall. The next section will shed light on the question to what extent herding is determined by ecological factors.

3.2

Ecological Determinants of Herding

While herding occurs in a broad variety of regions around the world, certain ecological conditions are highly favorable for herding whereas others make herding impossible to practice. A good example of the latter is regions in Africa where the TseTse fly is endemic as it transmits trypanosome disease that is lethal to livestock such as cattle (Alsan, 2015; Diamond, 1997). In a recent study, Beck and Sieber (2010) explore the extent to which climate and soil conditions determine the spatial distribution of basic landuse types (hunting-gathering, agriculture, sedentary animal husbandry, and nomadic pastoralism). The environmental data they employ includes detailed information about temperature and precipitation between 1961 and 1991 (Hijmans et al., 2005), altitude, and soil classification data from the United Nations Food and Agriculture Organization. Using these variables in maximum entropy modelling (Phillips et al., 2006; Phillips and Dudik, 2008), they estimate the probability with which each type of landuse occurs on 5x5 km grid cells for the Old World and Australia. Figure 2 shows the resulting heat map for herding. Using Beck and Sieber’s data, I can assign a measure of herding suitability to 750 societies in the Ethnographic Atlas to evaluate the degree to which actual dependence on herding is determined by ecological conditions. Even though this analysis likely underestimates the size of the true association between actual historical herding intensity and herding suitability as the suitability measure is based on contemporary conditions, it will potentially allow the identification ⁵In fact, high paternity uncertainty has been hypothesized to be a potential factor in explaining the evolutionarily puzzling occurrence of matriliny (Alexander, 1974; Hartung, 1985; Holden et al., 2003). ⁶Such practices can also be understood as credibility enhancing displays that evolve culturally to provide a credible measure of commitment to a value, belief, or preference (Henrich, 2009).

6

Figure 2: The map is taken from Beck and Sieber (2010) and shows the modeled land suitability for pastoralism. Red indicates better suitability (higher probability of occurrence), and blue indicates less suitable areas (lower probability of occurrence).

of a causal effect of herding on female status. As Table 2 shows, actual dependence on herding is strongly predicted by how suitable the land is for herding. More specifically, as column 1 shows, a substantial share (16%) of the variation in historical dependence on herding is explained by the suitability of the environment for herding alone. The R2 increases to 35% when conditioning on the continent and adding controls for the geographical location of the ethnic groups. The association between herding and herding suitability remains strong and statistically highly significant when adding historical controls or fixed effects for 24 world regions although, unsurprisingly, the coefficient becomes a bit smaller.

7

Table 2: Herding and Herding Suitability

(1)

(2)

∗∗∗

(5)

0.31

0.22

0.17∗∗∗

(0.04)

(0.04)

(0.04)

(0.03)

(0.03)

Continent FE

No

Yes

Yes

Yes

No

Region FE

No

No

No

No

Yes

Geographic Controls

No

No

Yes

Yes

Yes

Historical Controls

No

No

No

Yes

Yes

R2

∗∗∗

(4)

0.38

Observations

∗∗∗

(3)

0.42

Herding Suitability

∗∗∗

750

750

750

723

723

0.164

0.221

0.356

0.574

0.610

Notes. OLS estimates, robust standard errors. Geographic controls include distance to equator, longitude, and distance to nearest waterway (river or coast). Historical controls include dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation. Region FE are dummys for 24 world regions.∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Hence, even though other, more endogenous factors such as cultural beliefs could in principle also contribute to whether and how intensively an ethnic group practiced herding, the evidence presented here strongly suggests that exogenous factors such as the ecology at least play an important role. Assuming that land suitability for herding affects male dominance over women only through the channel of practicing herding it can be employed as an intrument for actual historical dependence on herding to lend empirical support to the causal claim made for herding on female status. I will conduct instrumental variable regression (IV) whenever possible and present the results in the appendix. In the main paper itself, I will stick to presenting reduced form regressions in order to avoid going back and forth between OLS and IV estimates. All IV estimates can be found in section ?? in the appendix and they will be discussed in the concluding section.

4

Male Dominance in Pre-Industrial Societies

The Standard Cross-Cultural Sample and the Ethnographic Atlas contain various society-level measures of female status. In this section, we will use these databases to test whether we find empirical support for an association between herding dependence and a subordinate position of women relative to men in pre-industrial societies. First, we will see whether we find empirical support for the claim that women in societies with a stronger dependence on herding are economically less independent, as suggested in the previous section. We make use of a variable (female power score) that was introduced into the Standard Cross-Cultural Sample by the anthropologist Peggy Sanday. It is a Guttman-scale that measures whether a society has flexible marriage mores, whether there is female non-domestic production, whether there is demand for female produce, whether women have economic con-

8

trol, whether women participate in the political sphere, and whether there are female solidarity groups. Thus, higher scores on this scale from 1 to 7 represent more female economic independence and political influence.⁷ Regressing the score on herding yields a clear result: stronger dependence on herding in a society implies lower values on the female power score and hence less economic independence of women. The effect becomes even stronger when conditioning on a set of society-level controls and including continent fixed effects, as columns 2 and 3 in Table3 show. Zooming in on the relationship between husband and wife we now test whether we find that women experience less economic independence also in the domestic sphere in societies with higher dependence on herding. Men control fruits of labor is the sum of two binary indicators, each of which takes the value 1 if the husband has the total or predominant control over the use of fruits of (i) his own labor and (ii) his wife’s labor, respectively, and 0 otherwise. They were introduced into the Standard Cross-Cultural Sample for 91 societies by Whyte (1978) as part of his cross-cultural work on the relative status of women in society. Regressing Men control fruits of labor on herding yields similar, albeit slightly weaker, results. The coefficient is positive and remains statistically significant at the 10-percent level when adding the usual set of controls (see columns 4 to 6 in Table 3). Next, we explore whether the way husband and wife interact reflects female subordination in herding societies. Again, Whyte (1978) gathered information ideal to study this question for 84 societies. Institutionalized deference from wife to husband is a Guttman-scale from 1 to 6 and measures whether the husband dominates domestic decision making, whether the wife is excluded from many social gatherings, whether the wife rarely disputes her husband, whether the husband has seating priority, and whether the wife kneels and bows when she greets her husband.⁸ The strength of institutionalized deference within marriage from wife to husband is positively associated with the degree to which a society depends on herding (columns 7 to 9 in Table 3). When including all controls the coefficient remains quite sizable and positive, but the p-value fails to reach statistical significance (p = 0.14). Note, however, that measures of within-marriage status or behavior will arguably be very noisy as ethnographers, who in the vast majority of cases were men, typically had only limited access - if any - to the female sphere in pre-industrial societies. This arguably holds true even more in societies characterized by a pronounced segregation of the sexes. Now, we turn to the question whether the beliefs a society holds are in line with the apparant subordinate position of women in the economic, political, and domestic sphere. Belief in male superiority (Whyte, 1978) is a binary variable that indicates whether the ethnographer regarded such a belief to be prevalent in a society. Confirming the evidence found so far, a stronger dependence of a society on herding animals comes with a general belief in male superiority. This association is sizable and robust to the inclusion of controls (columns 10 to 12 in Table 3). ⁷Figure 3 in section A in the appendix shows the distribution of values within the sample of 133 societies for which this information is available. ⁸Figure 4 in section A in the appendix depicts the distribution of values of this measure.

9

10 133 0.163

No

Yes

129 0.177

Yes

Yes

-0.32∗∗∗ (0.11)

(3)

91 0.014

No

No

0.13 (0.11)

(4)

91 0.057

No

Yes

0.23∗ (0.13)

(5)

89 0.138

Yes

Yes

0.26∗ (0.14)

(6)

84 0.088

No

No

0.30∗∗∗ (0.10)

(7)

84 0.097

No

Yes

0.31∗∗ (0.13)

(8)

82 0.150

Yes

Yes

0.22 (0.15)

(9)

Wife to Husband [Std.]

Institutionalized Deference:

93 0.125

No

No

0.17∗∗∗ (0.06)

(10)

93 0.172

No

Yes

0.15∗∗ (0.06)

(11)

91 0.235

Yes

Yes

0.15∗∗ (0.06)

(12)

Superiority [0/1]

Belief in Male

Notes. OLS estimates, robust standard errors. Historical controls include dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

133 0.041

No

Historical Controls

Observations R2

No

-0.29∗∗∗ (0.10)

-0.21∗∗ (0.10)

Continent FE

Dependence on Herding [Std.]

(2)

of Labor [Std.]

[Std.]

(1)

Men Control Fruits

Female Power Score

Dependent Variable:

Table 3: Male Dominance in Pre-Industrial Societies (Standard Cross-Cultural Sample)

Finally, we will examine the relationship between herding and kinship organization as a fundamental aspect of social structure, determining inheritance systems, how individuals define their group of relatives, and where a couple lives after they get married. Anthropologists have long noted that ecological conditions, such as the presence of domesticated animals, seem to vary systematically with kinship structure.⁹ A key distinction they make is between patrilineal and matrilineal societies. In patrilineal societies, descent is traced through the male line, i.e. an individual’s group of relatives will be her parents, her siblings, her father’s parents, her father’s siblings and their children, etc. Property will typically be inherited by the father’s sons, his siblings, or their sons. Post-marital residence tends to be with the husband’s parents (or his aunt or uncle). In analogy, in matrilineal societes descent is traced through the female line, inheritance goes (to men) through the female line, and post-marital residence tends to be with the bride’s parents. While matrilineality does not equate with matriarchy, it is conceivable that matrilineal kinship organization tends to be more favorable for women than patriliny. Here, I use data on kinship organization from 1,265 societies in the Ethnographic Atlas to look at three fundamental aspects of kinship structure: descent, inheritance, and post-marital residence. For descent and inheritance, I create an indicator that takes value 1 if they are patrilineal. For residence, the indicator is 1 if the couples reside with the groom’s parents after marriage. Almost half of the societies (46.79%) have patrilineal descent. A little more than half of the societies have patrilineal inheritance (51.59%) and patrilocal residence (51.89%). Quite naturally, societies that trace their descent through the male line also tend to have patrilocal residence (ρ = 0.81, p < 0.01) and patrilineal inheritance systems (ρ = 0.54, p < 0.01). Regressing each indicator on the dependence on herding reveals a very similar and stable positive association for each aspect of kinship structure in line with the picture that emerges from the anthropological literature: the more a society depends on herding, the more likely it is that its descent and inheritance are patrilineal and that post-marital residence is patrilocal. Table 4 illustrates the results.¹⁰ The associations are robust when conditioning on dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation, and including continent fixed effects.

⁹Using a sample of 564 societies, Aberle (1973) examines ecological correlates of matriliny and observes that matrilineal and bilineal descent systems are particularly infrequent among pastoral groups. Alluding to this observation, he concludes that "the cow is the enemy of matriliny, and the friend of patriliny". In more recent work, Holden and Mace (2003) test for coevolution between cattle and descent of Bantu- and Bantoid speaking groups in southern Africa and conclude that the acquisition of cattle led groups that formerly traced their descent through the matriline to switch to patrilineal or mixed descent. ¹⁰See Table 8 in section B.1 in the appendix for Probit estimates.

11

Table 4: Kinship Organization in Pre-Industrial Societies

Dependent variable: Patrilocal Residence [0/1] (1) Dependence on Herding [Std.]

∗∗∗

(2)

Patrilineal Inheritance [0/1] (3)

∗∗∗

0.082

∗∗∗

(4) 0.042

Patrilineal Descent [0/1] (5)

∗

∗∗∗

0.18

(6) 0.081∗∗∗

0.21

0.12

(0.01)

(0.02)

(0.02)

(0.02)

(0.01)

(0.02)

Continent FE

No

Yes

No

Yes

No

Yes

Historical Controls

No

Yes

No

Yes

No

Yes

Observations

1186

1146

884

863

1185

1149

R2

0.176

0.403

0.028

0.132

0.125

0.295

Notes. OLS estimates, robust standard errors. Historical controls include dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

In sum, this section provides strong evidence that herding animals implies a dominant position of men relative to women in a society. Stronger dependence on herding comes with less economic independence of women, subordinate positions of wives in marriage, and male supremacist beliefs in a society. In the next section, we will explore whether we find evidence for cultural persistence of these outcomes until today.

5

Male Dominance in Contemporary Societies

5.1

Empirical Strategy

In what follows, we will employ contemporary individual-level data to study whether a culture of male dominance induced by pre-industrial herding persists until today. For some countries and waves, the DHS information on individual’s ethnic affiliation. Based on this information, i.e. by matching contemporary ethnic groups to the historical record, individuals can be assigned the degree to which their ancestors depended on herding if their ethnic group is included in the Ethnographic Atlas, resulting in a sample of 312,357 women from 13 countries in Africa¹¹. Using this sample, we will test the association between historical herding intensity and current outcomes by regressing current individual-level outcomes on the historical dependence on herding of the individual’s ethnic group. In particular, the regression specification will be yi, j = α + β × dep_herding j + I i + H j + εi with yi, j denoting some outcome reported by individual i from ethnic group j, dep_herding j the ethnic group’s historical dependence on herding, I i a vector of individual-level controls, H j a vector of historical controls at the ethnic group level, and εi the error term. In particular, all regressions will include country-fixed effects so that only individuals facing similar institutional ¹¹Those countries are Benin, Burkina Faso, Cote d’Ivoire, Ghana, Guinea, Kenya, Malawi, Mali, Mozambique, Niger, Nigeria, Uganda, Zambia.

12

settings will be compared.

5.2

Main Results

Son Preference. Multiple aspects of herding suggest that sons might be preferred by parents in herding societies over their female siblings. Not only can sons better take up herding-related tasks than their sisters, but, as confirmed in section 4, herding is associated with patrilocal residence after marriage. Hence, while daughters move away from their parents when they are married, a son will stay with his parents which is of particular relevance for women in the light of potential widowhood. The DHS asks participants to state (i) the ideal number of sons and (ii) the ideal number of daughters. Using these two variables, I create a measure for son preference for each participant by subtracting the stated ideal number of daughters from the stated ideal number of sons and dividing by the sum of the two. If a participant wants all her children to be girls, the index takes value −1. If she wants all of them to be boys, the index takes value 1. A 0 means that she wants as many sons as daughters.¹². The sample includes 195,731 women from 13 African countries, covering 104 ethnicities. While the majority (n=160,808) want as many daughters as sons, the sample shows variation and there are more participants with a preference for sons than with a preference for daughters. On average, the index is slightly larger than 0 (0.032 ±0.243). Comparing only individuals who live in the same country, column 1 of Table 5 shows that women with stronger ancestral dependence on herding have a more pronounced preference for sons today. This effect becomes larger in terms of effect size and stronger in terms of statistical significance when adding individual level and historical society level controls (columns 2 and 3, respectively).

Table 5: Contemporary Son Preference and Historical Dependence on Herding

(1) Historical Dependence on Herding [Std.]

(2) ∗

(3) ∗∗

0.058∗∗∗

0.034

0.036

(0.02)

(0.02)

(0.02)

Individual Controls

No

Yes

Yes

Historical Controls

No

No

Yes

Country FE

Yes

Yes

Yes

195731

195724

195336

0.033

0.034

0.035

Observations R2

Notes. OLS estimates, standard errors clustered at the ethnicity level. The dependent variable (son preference index) is standardized. Historical controls: dep. on agriculture, jurisdictional hierarchy, settlement patterns, year of observation. Individual controls: age, age squared, urban residence, educ. attainment, marital status FE, religion FE, survey wave FE. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

¹²Table 5 in section A in the appendix shows the overall distribution of this index

13

Female subordination within marriage. Section 4 provided evidence for female subordinance within marriage in pre-industrial societies. In analogy, we will now explore whether this pattern persists until today. We will examine two very different domains of husband-wife interaction for which the DHS provides information: (i) decision-making over household purchases and (ii) the ability of the wife to refuse sex when her husband does but she does not want it. Both are direct measures in two distinct categories of whether a wife’s preferences and needs are secondary relative to those of her husband in their relationship. To assess household decision making over purchases, the participants (all women) were asked who has the final say when it comes to (i) big household purchases and (ii) daily household purchases. For both items, I created an indicator that takes value 1 if the wife has the final say, and 0 otherwise, e.g. if the husband or some other member of the family has the final say. The two indicators were then simply added to create a measure for how much control the wife can exert over the household income. If the participant does not have a final say in either category, the index takes value 0. It takes value 2, if she has final say in both categories, and 1 if she has a final say in one of the two categories. Most women (72.70%) do not have a final say in either category. 16.65% have a final say either when it comes to big or to daily household purchases and 10.65% have the final say in both categories. As Table 6 shows, ancestral dependence on herding is negatively associated with contemporary control over which goods an individual’s household purchases, and this association is robust to adding a large set of individual level and historical ethnicity level controls. Next, we test whether we find evidence for male dominance in the sexual domain. The DHS asks female participants Can you say no to your husband/partner if you do not want sexual intercourse? From answers to this question I created an indicator taking the value 1 if the woman answered "yes" and 0 if she answered "no". 45 % of the women state that they can not refuse sex even if they do not want to have sex. Again, answers to this question are related to ancestral herding intensity. Female descendants of ethnic groups with stronger dependence on herding are less able today to say no to their partner when he wants sexual intercourse and they do not and the effect is robust to the inclusion of a broad set of controls as columns 4 to 6 in Table 6 show.¹³

¹³See Table 9 in section B.1 in the appendix for Probit estimates.

14

Table 6: DHS: Female Subordination Within Marriage

Dependent variable: Wife has final say in HH purchases [Std.] (1) Historical Dependence on Herding [Std.]

(2) ∗∗∗

-0.067

-0.045

(3) ∗∗∗

-0.063

Wife can refuse sex [0/1] (4)

∗∗

-0.047

(5) ∗∗∗

(6) ∗∗

-0.023

-0.050∗∗∗

(0.01)

(0.01)

(0.03)

(0.01)

(0.01)

(0.02)

Individual Controls

No

Yes

Yes

No

Yes

Yes

Historical Controls

No

No

Yes

No

No

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

Yes

122685

122685

122373

95979

95976

95676

0.043

0.164

0.166

0.174

0.204

0.206

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Infibulation. As discussed in section 3, herding is characterized by a pronounced segregation between men and women. This segregation does not merely refer to a strict division of labor, but quite literally to a physical or geographical separation of the sexes. Men take the animals out on marches, which can last several hours to several days and sometimes even weeks, and often spend the nights outside of the camp with the animals in order to protect them from predators or thieves. All other things being equal, a more pronounced segregation of men and women decreases mate guarding opportunities for men and potentially increases uncertainty about biological relatedness between men and their wives’ offspring. Several cultural practices have been hypothesized to be a result of situations with high paternity uncertainty, e.g. the complete veiling of women or Chinese footbinding. Indeed, given the strong economic dependence of women on men in herding societies, it is conceivable that sending costly signals of willingness to be faithful to their husbands could be beneficial to them. Henrich (2009) shows how such costly signals ("credibility enhancing displays") might evolve culturally to transmit credible statements of commitments to values or beliefs shared at the group level. In what follows, I will empirically test whether one particular such practice – infibulation – is indeed associated with descent from a society with potentially higher levels of paternity uncertainty. Infibulation – sometimes also referred to as "pharaonic circumcision"¹⁴ – refers to the most invasive form of female genital cutting. In most cases, the procedure is performed by senior women on young girls between the ages of 4 and 10, in all cases well before puberty, often without anesthesia, using sharp instruments such as razors or knifes. Typically, it involves the complete removal of the clitoris, the labia minora and most or all of the labia majora. The raw edges of the vulva are then sutured together so that the opposite sides heal together to form a physical barrier over the vaginal opening. A small hole is left to allow for the passage of urin or menstrual blood. Before giving birth, infibulated women have to be "de-infibulated", ¹⁴Infibulation is often thought to originate in ancient Egypt but there is no direct evidence that it was practice there (Hicks, 1996; Shell-Duncan and Hernlund, 2000).

15

i.e. the physical barrier has to be opened surgically as it represents a clear obstruction to childbirth.¹⁵ Afterwards, they often undergo "re-infibulation". See Hicks (1996), Mackie (1996) or Shell-Duncan and Hernlund (2000) and references therein for more detailed descriptions and variations of the procedure. For some countries and waves, the DHS survey included questions about whether the respondents were infibulated. In some cases, respondents were asked to state which type of female circumcision – if any – they had undergone (with infibulation being one of the potential categories). In other cases, women were asked whether they had been circumcised, and if so, whether they had been infibulated.¹⁶ The resulting sample consists of 67,559 women from 51 ethnicities across 8 countries (Benin, Burkina Faso, Cote d’Ivoire, Guinea, Kenya, Mali, Niger, and Nigeria), of which 3,788 (5.61 %) stated that they had been infibulated. Columns 1 to 3 in Table 7 present the results from regressing a dummy indicating whether a women has been infibulated on the historical dependence on herding of her ethnic group. In all three specifications, the effect is very similar and quite sizable: the more a woman’s ancestors practiced herding, the more likely it is that she underwent infibulation today. See Table 10 in section B.1 in the appendix for results from Probit regressions. Next, I explore whether in addition to the incidence of infibulation I can also find evidence that the severity of the practice is associated with ancestral dependence on herding. To this end, I make use of data on the incidence of de-infibulation. The occurrence of de-infibulation is a good proxy for the invasiveness of the infibulation. The less invasive the initial procedure, i.e. the larger the vaginal opening that is left, the lower the likelihood that a woman has to be de-infibulated either to facilitate sexual intercourse after marriage or even before marriage to allow for the passage of menstrual blood during her first menstruation. In 1998 and 1999 the DHS asked participants "With your first period or when you got married, did someone have to make an incision to open the vaginal area?". Of the total sample of 8,579 women from 24 ethnicities in 5 countries¹⁷, 93 reported to have undergone de-infibulation (21 cases in Burkina Faso, 67 in Guinea, and 5 in Niger). Clearly, given the small number of countries and ethnicities, and the fact that we observe variation only in three of the five countries, this analysis is only tentative. Nevertheless, I find evidence for an association between ancestral dependence on herding and the severity with which infibulation is practiced today. As columns 4 to 6 in Table 7 show, the coefficient of dependence on herding is significantly positive and robust to adding individuallevel and ethnicity-level covariates as controls. Probit estimates can be found in Table 10 in ¹⁵Among the most common side effects associated with infibulation are obstructed or prolonged labor, which can cause fistulae (openings between the vagina and either the rectum or the bladder or both, so that feces or urin pass through the vagina without the woman having full control over it), see for example p. 14 in Shell-Duncan and Hernlund (2000). Indeed, I find a positive albeit small correlation between reports of having experienced fistula and being infibulated (ρ = 0.02, p < 0.05). ¹⁶Furthermore, I classified a woman as being infibulated if she stated that she had needed to undergo de-infibulation, even when they classified the type of circumcision that they had undergone as something other than infibulation. The results do not change if I do not count these cases as infibulation as Table 11 in the appendix illustrates. ¹⁷Guinea, Cote d’Ivoire, Niger, Egypt, and Burkina Faso.

16

section B.1 in the appendix. Table 7: DHS: Infibulation and Deinfibulation

Dependent variable: Underwent Infibulation [0/1] (1) Historical Dependence on Herding [Std.]

(2) ∗∗

0.080

0.072

∗∗∗

(3) 0.075

∗∗∗

Had to be Deinfibulated [0/1] (4) 0.011

(5) ∗∗

0.010

(6) ∗∗

0.017∗∗∗

(0.03)

(0.03)

(0.03)

(0.00)

(0.00)

(0.01)

Individual Controls

No

Yes

Yes

No

Yes

Yes

Historical Controls

No

No

Yes

No

No

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

Yes

Observations

67559

67555

67554

8579

8579

8579

R2

0.111

0.116

0.117

0.005

0.007

0.008

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

5.3

Robustness

Within-District Analysis So far, all analyses at the individual level employed country fixed effects in order to provide evidence that herding rather than insitutional factors varying at the country level drive the effects. However, it is conceivable that environments, e.g. institutions or climate, also vary at the subnational level, in particular in large countries such as Nigeria. Hence, we should include district-level fixed effects in order to hold constant people’s contemporary environment. On the other hand, note that it is not clear a priori whether such an analysis is appropriate when we examine cultural variation. After all, culture might not vary too much at the district-level such that conditioning on district-level fixed effects simply soaks up all the variation. Nevertheless, whenever possible, I replicate the above analyses using district-level fixed effects. Table 12 in section B.3 presents the results. In most cases, the effects found above even hold true when only comparing individuals who live in the same district.

Excluding Extreme Values of Herding Dependence A natural concern with the above results is that they are driven by observations at either end of the distribution of historical herding intensity. On the one hand, it could be that going from no herding to some herding is all it takes to produce a culture in which males are dominant. On the other hand, it could equally likely be that the association between herding and male dominance is mainly driven by those societies that entirely depend on herding, i.e. societies in which other forms of subsistence play only a marginal role or are completely absent. I therefore replicate the analyses excluding societies who exhibit a dependence on herding of (i) 5% or less (lowest category) and (ii) 85% or more (highest category), and both at the same time. Excluding those with very little or no dependence at all leaves the results unchanged, as Table 14 in the appendix shows. The same holds true when excluding societies who very strongly (85% or more) depend on herding (Table 15 in

17

the appendix). Even though the association between herding and infibulation becomes weaker the coefficient remains statistically significant at the 10%-level. The same picture emerges when excluding societies with either the lowest or the highest dependence on herding (Table 16).

Contemporary Presence of Herding Animals More or less implicit in this paper is the idea that the contemporary results are driven by the transmission of cultural beliefs and values induced by the historical form of subsistence, rather than by the transmission of the cultural practice of herding itself. For some waves and countries, the DHS included questions about whether the household owned various types of animals. This allows me to replicate the analysis after excluding people who reported to own an animal which would – at least historically – classify as a herding animal. In particular, in this part of the analysis, I exclude participants who reported to own cattle, horses, donkeys, camels, sheep, or goats. Depending on the outcome, this leads to a quite substantial reduction of the sample by 20-30 %. Only for de-infibulation, the number of observations remains unchanged. In all other cases and despite the decrease in sample size, all effects remain robust and largely unchanged both in terms of effect size and statistical significance. Table 13 in section B.3 illustrates the results.

Placebo Analysis: Animal Husbandry with Non-Herd Animals A further potential concern is that the effects are not driven by herding but rather by owning and living with animals per se. Put differently, animal husbandry as such but not herding in particular could be prone to fostering societal patterns that are unfavorable for women, regardless of what kind of animals the society tends. To test this idea, we can run the same type of analysis for a "placebo group", i.e. societies that practice animal husbandry with non-herd animals, such as pigs or poultry. In analogy to the herding measure, I combine two variables (dependence on animal husbandry and information on the predominant type of animal in a society) to construct a measure for historical dependence on animal husbandry with non-herd animals. This allows me to replicate the above analysis only for the two measures of bargaining power within marriage and son preference.¹⁸ If the association between herding and male dominance is not driven by having animals per se but by other aspects of herding, we should expect no effect of animal husbandry with non-herd animals on female status. If anything, we should expect to find opposite effects as this type of animal husbandry is a predominantly female form of subsistence (see Table 1 in section 3.) Indeed, as Table 17 in section B.3 shows, regressing measures for contemporary female bargaining power within marriage and son preference on historical dependence on animal husbandry with non-herd animals yields coefficients with opposite signs to those found above. With increasing historical dependence on this form of subsistence, women today are more likely to have a say about how household resources are spent and to be able to refuse sex, and they have a weaker preference for sons. Adding historical dependence on herding, the association with ability to refuse sex remains positive and significant, and in the other cases the coefficient ¹⁸In the (de-)infibulation samples there is no variation in animal husbandry with non-herd animals.

18

decreases to close to zero and loses statistical significance. A very similar picture emerges when running the same type of analysis on the historical ethnic-group level data (see Table 18 in section B.3 in the appendix). The coefficient on animal husbandry with non-herd animals is typically close to zero, and tends to have the opposite sign than the one estimated for herding intensity. In sum, these effects support the narrative for the association between herding and female status provided in this paper.

6

Conclusion

The anthropological literature has long noted a relationship between subsistence styles and female status. This paper tests and confirms the hypothesis that herding, but not animal husbandry per se, was associated with a subordinate position of women in society. It shows empirically that pre-industrial societies with higher dependence on herding were indeed more likely to have a patrilineal kinship structure and had lower female economic independence, less female control over household resources, and a more institutionalized deference of wives to their husbands. It furthermore provides strong evidence for a persistence of this pattern until today. Contemporary women whose ancestral ethnic group depended more on herding have less influence on how household resources are spent, are less able to refuse sex with their partners, and have a stronger preference for sons. Moreover, they are more likely to have undergone infibulation, a cultural practice potentially originating as a result of a pronounced segregation between men and women in herding societies. Replicating the analysis for dependence on animal husbandry with non-herd animals I find no, and in some cases opposite, associations with measures for female status. This supports the interpretation that characteristics of herding other than having animals per se are the mechanism behind the effect. Excluding individuals who own herd animals from the analysis leaves the results virtually unchanged. The contemporary effects are therefore not driven by a transmission of the cultural practice of herding across generations. Even when the form of subsistence has long changed, cultural values originating in pre-industrial lifestyles and ultimately in historical ecological conditions seem to persist until today. In order to identify causality, I test whether historical herding intensity is predicted by the suitability of the environment and find a strong positive effect. Due to data limitations, IV regressions are possible only for a subset of the analyses. However, the resulting IV coefficients are largely in line with the OLS results, supporting a causal link between herding and female status. Moreover, note that during human evolution a transition between different subsistence styles, e.g. transiting from hunting and gathering to agriculture or from horticulture to herding, did not happen in one discrete step that could potentially be explained by a society’s desire to choose a more or less gender-balanced form of subsistence. Instead, such transitions were long processes happening over many generations (Diamond, 1997). In that light, it seems implausi-

19

ble that reverse causality (fully) explains the results.

20

References Aberle, David F., “Matrilineal Descent in Cross-Cultural Perspective,” in David M. Schneider and Kathleen Gough, eds., Matrilineal Kinship, Berkeley, Los Angeles, London: University of California Press, 1973, pp. 655–730. Alesina, Alberto, Paola Giuliano, and Nathan Nunn, “On The Origins of Gender Roles: Women And The Plough,” Quarterly Journal of Economics, 2013, 128, 469–530. Alexander, Richard D., “The Evolution of Social Behavior,” Annual Review of Ecology and Systematics, 1974, 5, 325–383. Alsan, Marcella, “The Effect of the TseTse Fly on African Development,” American Economic Review, 2015, 105, 382–410. Ashraf, Nava, Natalie Bau, Nathan Nunn, and Alessandra Voena, “Bride Price and Female Education,” Working paper, 2016. Barfield, Thomas J., The Nomadic Alternative, Englewood Cliffs, NJ: Prentice-Hall, 1993. Beck, Jan and Andrea Sieber, “Is the Spatial Distribution of Mankind’s Most Basic Economic Traits Determined by Climate and Soil Alone?,” PLoS ONE, 2010, 5, e10416. Becker, Anke, Benjamin Enke, and Armin Falk, “Ancient Origins of the Global Variation in Economic Preferences,” Working Paper, 2017. Borgerhoff Mulder, Monique, Ila Fazzio, William Irons, Richard L. McElreath, Samuel Bowles, Adrian Bell, Tom Hertz, and Leela Hazzah, “Pastoralism and Wealth Inequality: Revisiting an Old Question,” Current Anthropology, 2010, 51, 35–48. Boserup, Ester, Woman’s Role in Economic Development, London, Allen & Unwin Ltd., 1970. Boyd, Robert and Peter J. Richerson, Culture and the Evolutionary Process, University of Chicago Press, 1988. and

, The Origin and Evolution of Cultures, New York: Oxford University Press, 2005.

Chatty, Dawn, “Structuring Forces of Pastoral Nomadism in S.W. Asia (with Emphasis on Camel Pastoral Nomadism),” Development and Change, 1973, 4, 51–72. Chen, M. Keith, “The Effect of Language on Economic Behavior: Evidence from Savings Rates, Health Behaviors, and Retirement Assets,” American Economic Review, 2013, 103, 690–731. Cohen, Dov and Richard E. Nisbett, “Self-Protection and the Culture of Honor: Explaining Southern Violence,” Personality and Social Psychology Bulletin, 1994, 20 (5), 551–567.

21

Dell, Melissa, “The Persistent Effects of Peru’s Mining Mita,” Econometrica, 2010, 78, 1863– 1903. Diamond, Jared M., Guns, Germs, and Steel: The Fates of Human Societies, New York: W. W. Norton & Company, Inc., 1997. Enke, Benjamin, “Kinship Systems, Cooperation and the Evolution of Culture,” Working Paper, 2017. Falk, Armin, Anke Becker, Thomas Dohmen, David Huffman, and Uwe Sunde, “Global Evidence on Economic Preferences,” Working Paper, 2017. Fernandez, Raquel and Alessandra Fogli, “Culture: An Empirical Investigation of Beliefs, Work, and Fertility,” American Economic Journal: Macroeconomics, 2009, 1, 146–177. Fratkin, Elliot and Eric Abella Roth, “Who Survives Drought? Measuring Winners and Losers among the Ariaal Rendille Pastoralists of Kenya,” in Daniel G. Bates and Sarah H. Lees, eds., Case Studies in Human Ecology, Boston, MA: Springer, 1996, pp. 159–173. Galor, Oded and Ömer Özak, “The Agricultural Origins of Time Preference,” American Economic Review, 2016, 106, 3064–3103. Grosjean, Pauline, “A History of Violence: The Culture of Honor and Homicide in the US South,” Journal of the European Economic Association, 2014, 12, 1285–1316. Hall, Thomas D., “The Ecology of Herding: Conclusions, Questions, Speculations,” in P. Nick Kardulias, ed., The Ecology of Pastoralism, Boulder: University Press of Colorado, 2015, pp. 267–280. Harris, Marvin, Cannibals and Kings: The Origins of Culture, New York: Random House, 1977. Hartung, John, “Matrilineal Inheritance: New Theory and Analysis,” Behavioral and Brain Sciences, 1985, 8, 661–670. Henrich, Joseph, “The Evolution of Costly Displays, Cooperation and Religion: Credibility Enhancing Displays and Their Implications for Cultural Evolution,” Evolution and Human Behavior, 2009, 30, 244–260. , The Secret of Our Success: How Culture is Driving Human Evolution, Domesticating Our Species, and Making Us Smarter, Princeton University Press, 2015. Hicks, Esther K., Infibulation: Female Mutilation in Islamic Northeastern Africa, New Brunswick, N.J.: Transaction Publishers, 1996. Hijmans, Robert J., Susan E. Cameron, Juan L. Parra, Peter G. Jones, and Andy Jarvis, “Very High Resolution Interpolated Climate Surfaces for Global Land Areas,” International Journal of Climatology, 2005, 25, 1965–1978.

22

Holden, Clare Janaki and Ruth Mace, “Spread of Cattle Led to the Loss of Matrilineal Descent in Africa: A Coevolutionary Analysis,” Proceedings Of The Royal Society London Series B - Biological Sciences, 2003, 270, 2425–2433. , Rebecca Sear, and Ruth Mace, “Matriliny as Daughter-Biased Investment,” Evolution and Human Behavior, 2003, 24, 99–112. Kardulias, P. Nick, “Pastoralism as an Adaptive Strategy,” in P. Nick Kardulias, ed., The Ecology of Pastoralism, Boulder: University Press of Colorado, 2015, pp. 1–15. Mackie, Gerry, “Ending Footbinding and Infibulation: A Convention Account,” American Sociological Review, 1996, 61, 999–1009. , “Female Genital Cutting: The Beginning of the End,” in Bettina Shell-Duncan and Ylva Hernlund, eds., Female "Circumcision" in Africa: Culture, Controversy, and Change, Boulder, Colorado: Lynne Riener Publishers, Inc., 2000, pp. 253–282. Murdock, George P., Ethnographic Atlas, University of Pittsburgh Press, 1967. and Douglas R. White, “The Standard Cross-Cultural Sample,” Ethnology, 1969, 8, 329– 369. Nunn, Nathan and Leonard Wantchekon, “The Slave Trade and the Origins of Mistrust in Africa,” American Economic Review, 2011, 101, 3221–3253. Phillips, Steven J. and Miroslav Dudik, “Modeling of Species Distributions with Maxent: New Extensions and a Comprehensive Evaluation,” Ecography, 2008, 31, 161–175. , Robert P. Anderson, and Robert E. Schapire, “Maximum Entropy Modeling of Species Geographic Distribution,” Ecological Modelling, 2006, 190, 231–259. Sanday, Peggy R., “Toward a Theory of the Status of Women,” American Anthropologist, 1973, 75, 1682–1700. , Female Power and Male Dominance: On the Origins of Sexual Inequality, Cambridge University Press, 1981. Schulz, Jonathan F., “The Churches’ Bans on Consanguineous Marriages, Kin-Networks and Democracy,” Working Paper, 2017. Shell-Duncan, Bettina and Ylva Hernlund, “Female "Circumcision" in Africa: Dimension of the Practice and Debate,” in Bettina Shell-Duncan and Ylva Hernlund, eds., Female "Circumcision" in Africa: Culture, Controversy, and Change, Boulder, Colorado: Lynne Riener Publishers, Inc., 2000, pp. 1–40.

23

Teso, Edoardo, “The Long-Term Effect of Demographic Shocks on the Evolution of Gender Roles: Evidence from the Transatlantic Slave Trade,” Working Paper, 2016. Voigtländer, Nico and Hans-Joachim Voth, “Persecution Perpetuated: The Medieval Origins of Anti-Semitic Violence in Nazi Germany,” Quarterly Journal of Economics, 2012, 127, 1339. Whyte, Martin King, “Cross-Cultural Codes Dealing with the Relative Status of Women,” Ethnology, 1978, 17, 211–237.

24

A

Additional Graphs and Figures

Figure 3: Distribution of the Female Power Score (N=133)

Figure 4: Institutionalized Deference from Wife to Husband (N=84)

25

Figure 5: Distribution of Son Preference Index

26

B B.1

Additional Analyses Probit Regressions Table 8: Kinship Organization in Pre-Industrial Societies

Dependent variable: Patrilocal Residence [0/1] (1) Dependence on Herding [Std.]

∗∗∗

(2) ∗∗∗

Patrilineal Inheritance [0/1] (3)

(4)

∗∗∗

Patrilineal Descent [0/1] (5)

∗

∗∗∗

(6)

0.71

0.45

0.21

0.12

0.51

0.25∗∗∗

(0.07)

(0.07)

(0.05)

(0.06)

(0.05)

(0.06)

Continent FE

No

Yes

No

Yes

No

Yes

Historical Controls

No

Yes

No

Yes

No

Yes

Observations

1186

1146

884

863

1185

1149

Pseudo R2

0.151

0.329

0.021

0.099

0.098

0.233

Notes. Probit estimates, robust standard errors. Historical controls include dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Table 9: DHS: Wife Can Refuse Sex [0/1] (Probit Estimates)

(1) Historical Dependence on Herding [Std.]

∗∗∗

-0.13

(2) -0.061

(3) ∗∗

-0.14∗∗∗

(0.04)

(0.03)

(0.04)

Individual Controls

No

Yes

Yes

Historical Controls

No

No

Yes

Country FE

Yes

Yes

Yes

Observations

95979

95976

95676

Pseudo R2

0.133

0.159

0.161

Notes. Probit estimates, standard errors are clustered at the ethnicity level. Historical controls: dep. on agriculture, jurisdictional hierarchy, settlement patterns, year of observation. Individual controls: age, age squared, a dummy for urban residence, educ. attainment, marital status FE, religion FE, survey wave FE. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

27

Table 10: DHS: Infibulation and Deinfibulation (Probit Estimates)

Dependent variable: Underwent Infibulation [0/1] (1)

(2)

(3)

∗∗∗

(4)

∗∗∗

(5)

∗∗

∗∗

(6)

0.33

0.26

0.37

0.31

0.28

0.53∗∗∗

(0.09)

(0.07)

(0.09)

(0.14)

(0.14)

(0.09)

Individual Controls

No

Yes

Yes

No

Yes

Yes

Historical Controls

No

No

Yes

No

No

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

Yes

Observations

67512

67483

67482

8533

8357

8357

Pseudo R2

0.181

0.188

0.189

0.039

0.058

0.069

Historical Dependence on Herding [Std.]

∗∗∗

Had to be Deinfibulated [0/1]

Notes. Probit estimates, standard errors are clustered at the ethnicity level. Historical controls: dep. on agriculture, jurisdictional hierarchy, settlement patterns, year of observation. Individual controls: age, age squared, a dummy for urban residence, educ. attainment, marital status FE, religion FE, survey wave FE. ∗∗

B.2

p < 0.05,

∗∗∗

p < 0.01.

Alternative Infibulation Measure Table 11: DHS: Infibulation (Alternative Measure)

(1) Historical Dependence on Herding [Std.]

(2) ∗∗

(3) ∗∗

0.073∗∗∗

0.078

0.071

(0.03)

(0.03)

(0.03)

Individual Controls

No

Yes

Yes

Historical Controls

No

No

Yes

Country FE

Yes

Yes

Yes

Observations

67559

67555

67554

R2

0.115

0.121

0.121

Notes. Probit estimates, standard errors are clustered at the ethnicity level. Historical controls: dep. on agriculture, jurisdictional hierarchy, settlement patterns, year of observation. Individual controls: age, age squared, a dummy for urban residence, educ. attainment, marital status FE, religion FE, survey wave FE. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

28

∗

p < 0.10,

B.3 B.3.1

Robustness Within-District Analysis

Table 12: Within-District Analysis

Dependent variable:

Historical Dependence on Herding [Std.]

Wife Final Say: Purchases [Std.]

Wife Can Refuse Sex [0/1]

Son Preference [Std.]

Infibulation [0/1]

(1)

(2)

(3)

(4)

∗∗

-0.029

-0.028

0.040

∗∗∗

0.025∗

(0.02)

(0.01)

(0.01)

(0.01)

Individual Controls

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

District FE

Yes

Yes

Yes

Yes

122373

95676

191140

63411

0.101

0.232

0.040

0.156

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

B.3.2

Excluding Individuals Who Own Herd Animals

Table 13: Restricting the Sample to People Who Do Not Own Herding Animals Today

Dependent variable:

Historical Dependence on Herding [Std.]

Wife Final Say: Purchases

Wife Can Refuse Sex

Son Preference

Infibulation

De-Infibulation

[Std.]

[0/1]

[Std.]

[0/1]

[0/1]

(1)

(2)

(3)

(4)

(5)

-0.073

∗∗∗

-0.052

∗∗∗

0.057

∗∗∗

∗∗

0.050

0.017∗∗∗

(0.02)

(0.02)

(0.02)

(0.02)

(0.01)

Individual Controls

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

Observations

99283

67093

158693

47607

8579

R2

0.182

0.186

0.034

0.074

0.008

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

29

B.3.3

Excluding Extreme Values of Herding Intensity

Table 14: Excluding Societies with < 5% Dependence on Herding

Dependent variable:

Historical Dependence on Herding [Std.]

Wife Final Say: Purchases

Wife Can Refuse Sex

Son Preference

Infibulation

De-Infibulation

[Std.]

[0/1]

[Std.]

[0/1]

[0/1]

(1)

(2)

(3)

(4)

(5)

∗

-0.052

-0.044

∗∗

0.059

∗∗∗

0.075

∗∗∗

0.017∗∗∗

(0.03)

(0.02)

(0.02)

(0.03)

(0.01)

Individual Controls

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

115783

92330

183681

67554

8579

0.160

0.208

0.033

0.117

0.008

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Table 15: Excluding Societies with > 85% Dependence on Herding

Dependent variable:

Historical Dependence on Herding [Std.]

Wife Final Say: Purchases

Wife Can Refuse Sex

Son Preference

Infibulation

De-Infibulation

[Std.]

[0/1]

[Std.]

[0/1]

[0/1]

(1)

(2)

(3)

(4)

(5)

-0.093

∗∗∗

-0.043

∗∗

0.060

∗∗∗

0.025

∗

0.017∗∗∗

(0.03)

(0.02)

(0.01)

(0.01)

(0.01)

Individual Controls

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

118845

93553

193031

66780

8579

0.164

0.208

0.035

0.055

0.008

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

30

Table 16: Excluding Societies with < 5% or > 85% Dependence on Herding

Dependent variable: Wife Final Say: Purchases

Wife Can Refuse Sex

Son Preference

Infibulation

De-Infibulation

[Std.]

[0/1]

[Std.]

[0/1]

[0/1]

(1)

(2)

(3)

(4)

(5)

Historical Dependence on Herding [Std.]

-0.084

∗∗

-0.035

∗

0.062

∗∗∗

0.025

∗

0.017∗∗∗

(0.03)

(0.02)

(0.01)

(0.01)

(0.01)

Individual Controls

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

112255

90207

181376

66780

8579

0.157

0.210

0.033

0.055

0.008

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

B.3.4

Placebo Analysis: Animal Husbandry with Non-Herd Animals

Table 17: Placebo Analysis: Animal Husbandry with Non-Herd Animals and Male Dominance in the DHS

Dependent variable: Wife Final Say: Purchases [Std.] (1) Historical Dependence on Animal Husbandry [Std.]

(2)

∗∗∗

0.023

0.013

(0.01)

(0.01)

Historical Dependence on Herding [Std.]

Wife Can Refuse Sex [0/1] (3) 0.022

∗∗∗

(0.00)

(4) 0.013

∗∗∗

(0.00)

Son Preference [Std.] (5) -0.0083

(6) ∗∗∗

(0.00)

0.0022 (0.00)

-0.054∗

-0.043∗∗∗

0.059∗∗∗

(0.03)

(0.02)

(0.02)

Individual Controls

Yes

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Yes

Country FE

Yes

Yes

Yes

Yes

Yes

Yes

122373

122373

95676

95676

195336

195336

0.165

0.166

0.205

0.206

0.034

0.035

Observations R2

Notes. OLS estimates, standard errors are clustered at the ethnicity level. Dependence on animal husbandry encompasses animal husbandry with non-herd animals such as pigs or poultry. Historical controls include dependence on agriculture, jurisdictional hierarchy, settlement patterns, and year of observation. Individual controls include age, age squared, a dummy for urban residence, educational attainment, marital status fixed effects, religion fixed effects, and survey wave fixed effects. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

31

Table 18: Placebo Analysis: Animal Husbandry with Non-Herd Animals and Male Dominance in the Standard Cross-Cultural Sample

Dependent Variable: Female Power Score

Men Control Fruits

Institutionalized Deference:

Belief in Male

[Std.]

of Labor [Std.]

Wife to Husband [Std.]

Superiority [0/1]

(3)

(5)

(6)

(7)

(8)

(1) Dep. on Animal Husbandry [Std.]

(2)

(4)

0.080

0.015

0.13

0.18

-0.12

-0.065

-0.073∗

-0.041

(0.10)

(0.10)

(0.15)

(0.14)

(0.11)

(0.12)

(0.04)

(0.04)

Dependence on Herding [Std.]

-0.32∗∗∗

0.30∗∗

0.20

0.14∗∗

(0.11)

(0.15)

(0.16)

(0.06)

Continent FE

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Historical Controls

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Observations R2

129

129

89

89

82

82

91

91

0.119

0.177

0.117

0.157

0.130

0.153

0.193

0.240

Notes. OLS estimates, robust standard errors. Controls include dependence on agriculture, traditional plow use, jurisdictional hierarchy, settlement patterns, and year of observation. ∗ p < 0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

32

C C.1

Data Description Historical Variables

Dependence on Herding [Source: Ethnographic Atlas] Dependence on Animal Husbandry [Source: Ethnographic Atlas] Herding Animals [Source: Ethnographic Atlas] Dependence on Agriculture [Source: Ethnographic Atlas] Traditional Plow Use [Source: Ethnographic Atlas] Jurisdictional Hierarchy [Source: Ethnographic Atlas] Settlement Patterns [Source: Ethnographic Atlas] Year of Observation [Source: Ethnographic Atlas] Patrilineal Descent [Source: Ethnographic Atlas] Patrilineal Inheritance [Source: Ethnographic Atlas] Patrilocal Residence [Source: Ethnographic Atlas] Female Power Score [Source: Standard Cross-Cultural Sample] Men Control Fruits of Labor [Source: Standard Cross-Cultural Sample] Institutionalized Deference: Wife to Husband [Source: Standard Cross-Cultural Sample]

Belief in Male Superiority [Source: Standard Cross-Cultural Sample]

C.2

Contemporary Variables

Son Preference [Source: DHS] Wife Has Final Say in HH Purchases [Source: DHS]

33

Wife Can Refuse Sex [Source: DHS] Infibulation [Source: DHS] De-Infibulation [Source: DHS]

D

Matching Contemporary Ethnicities to the Historical Record

34