Updated: Nov 7, 2022
I want to make a confession. In the past I got things wrong, seriously wrong. Allow me to share why I was so mistaken and how I came to revise my priors. Rather listen than read? Here's the podcast.
When I lacked globally comparative knowledge, I was blind to country-specific characteristics
"Women can do what men can do" can be heard across the Zambian Copperbelt (banamayo kuti babomba incito sha baume). Eager understand the causes of growing support for gender equality, I undertook 18 months of ethnographic research. I became fluent in Bemba, interviewed over two hundred people, lived in slums, swamps and middle-class homes, observed classrooms, markets and political meetings.
Worsening economic security - the collapsing price of copper, state cut-backs, male job losses, user fees, the scourge of HIV/AIDs - had propelled women into the labour market. As a multitude of women demonstrated equal competence in socially valued male-dominated domains, people slowly came to question their gender stereotypes. Market traders share stories of female mechanics, mine truck drivers, and widows, single-handedly providing for their children. Seeing this public support, others become emboldened to demand equality: "for the elections, we want women!".
My explanation was only half right.
Without globally comparative knowledge, I was blind to country-specific characteristics. Zambians' preference for female seclusion was very weak. Female labour supply thus rose quickly in response to shifting economic incentives. South Asia, the Middle East and North Africa by contrast, are caught in the Patrilineal Trap: female employment remains low because potential earnings are too low to compensate for the loss of male honour.
Culture - I quickly realised - mediates the rate at which women seize opportunities created by development and democratisation. So I sought to explain global cultural variation. Here again, I was initially wrong.
I overestimated the importance of labour demand in agriculture
Many cross-sectional studies find a strong correlation between crop systems and gender relations. Where women’s contribution to farm production was relatively significant (such as in labour-intensive foraging, horticulture, and wet paddy fields), they tended to have greater authority and independence. Girls were seen favourably and pre-marital sex was permitted. Whereas in societies where men's labour was more crucial (ploughing and herding cattle), there was more reverence for men.
Or so the standard argument goes!
It presumes that farm-work raises women's status as well as social acceptance of their labour market participation. By studying the global history of gender, I realised this is incorrect.
Farm-work does not guarantee women's esteem, autonomy or protection from violence. Even if women work long days harvesting crops, pounding grain and fetching firewood, their labour may be unrecognised and unappreciated. Ethnographies, focus groups, and surveys all tell us that rural women's contributions are scarcely considered 'work', even by women themselves. As a 19th century Haryana saying goes, "jeore se nara ghisna hai": women as cattle bound, working and enduring all. It is only through paid work in the public sphere that women can mingle with diverse others, discover more egalitarian ideas, forge more diverse friendships, and become emboldened to resist unfairness. But families that exploit women's labour on farms may nevertheless restrict it in cities. Rajasthan scores high for rural but not urban female employment. Why? Because male honour is contingent on female purity, which can only be maintained through careful surveillance. That's simply not possible in Udaipur's bustling bazaars.
Where seclusion is idealised, urban but not rural female employment is repressed.
In Turkey, female labour force participation is 14% lower among those that idealise seclusion, but this relationship only holds in cities not villages. Even if peasants are extremely patriarchal they still harness female labour, since male honour is not at risk. Yet even as labour demand has dwindled in agriculture, less educated women have shunned urban employment. Greek women by contrast are much more responsive to economic opportunities in town. They now comprise 28% of legislators, senior officials and managers - double that of Turkey.
So why did I get it so wrong?
Convention and correlations misled me. Labour demand is the standard narrative and I didn't know any better. Alesina, Giuliano and Nunn famously proved that Boserup was right: plough-cultivation is associated with men's political and economic dominance. Since men are physically stronger they ploughed, while women processed cereals back home. Gender divisions of labour became naturalised.
Highly plausible! By studying ten thousand years of cultural evolution, I realised that's not the whole story. Societies with the exact same geography, technology and socio-economic complexity can differ enormously on gender. Let me give four examples of contingency:
"Men plough, women weave" has been the Chinese convention for millennia, but this was just a brute fact not necessarily a lofty ideal. Women only became confined to the inner quarters under the sway of Confucianism.
The Minoan civilisation was technologically complex and socially stratified. Yet paintings show women occupying prominent social positions in outdoor assemblies, fraternising freely with men. In this exact same location, a little while later, the Ancient Greeks institutionalised one of the most patriarchal societies the world has ever known. Women’s names were not even uttered in public.
Before Islam, gender relations were diverse across MENA. In Ancient Egypt, women had equal rights under the law. Queen Zenobia (of what is now Syria) led 70,000 men into battle against the Roman Empire. The Amazigh (Berber) revered women leaders: goddess Tanit, warrior queen Tin Hinan, and military commander Dihya. Seclusion only became normative after the Arab-Islamic conquests and subsequent influence of Ghazali.
North India's low female employment is sometimes attributed to its loamy soils which enabled the plough. But gender relations were once significantly different – as indicated by 2nd century CE statues of naked goddesses in Uttar Pradesh.
Persistence studies cannot detect this cultural evolution or discern causal mechanisms because they typically rely on data gathered at one point in time. Cultural economists usually rely on the Ethnographic Atlas - which compiles 19th and early 20th century ethnographies of over a thousand groups worldwide. This is just a snapshot.
By studying ten thousand years of patriarchy, I realised that Alesina and colleagues were right, the plough is hugely important. However, the causal mechanism is not labour demand. The plough created inherited wealth amid insecurity, which in some places led to a contingent process of cultural evolution.
The Yamnaya, Semitic and Akkadian cultures were especially patriarchal. Through conquest, patriarchy was institutionalised in Europe, South Asia, and the Horn of Africa.
Inherited wealth was another major driver. Cattle, the plow and irrigation raised crop yields, making land itself a valuable asset. Cereals could be traded and stored. The plough raised crop yields and made land a valuable asset.
Wealth turned patrilineal inheritance into a key element of social organisation. The more wealth a son inherited, the greater his reproductive success (by attracting wives, concubines and rearing offspring). But this was threatened by raiders.
Lineage cooperation, male honour and inter-marriage alliances were maintained by controlling female sexuality.
When societies grew they were threatened by in-fighting. Men squabbled over women, wealth and property. Sexual jealousy may have been mitigated by religions that idealised sexual segregation, chastity, fidelity, and veiling. Compliance was promoted by praising female virtue, social policing, state laws, and moralising supernatural punishment. Male rulers and theologians blamed floods, droughts and earthquakes on disobedient women. Amid fears of eternal damnation there emerged cults of chastity. Islam became particularly patriarchal.
So the plough was important, but not because it suppressed demand for female labour in the fields. Rather, the emergence of inherited wealth amid insecurity encouraged patriliny and as these societies flourished they built religions which sanctified female seclusion.
But technology is not destiny. Societies with the exact same technology show tremendous diversity in gender relations. Contrast the Etruscans and Romans, the Minoans and Mycenneans, the Sumerians and Akkadians, as well as Northern Nigeria after the Fulla jihad. Each of the latter cultures were more patriarchal; they were institutionalised through conquest.
Cultural evolution was also mediated by geography. Oceans, mountains and parasites constrained the spread of draft animals, pastoralism and Islam. Without these major forces of patriarchy, the Americas, Southeast Asia, Gulf of Guinea and Southern Africa did not develop patrilineal inheritance of wealth. With little concern for paternity, women could move freely, exercise authority, build independent networks of solidarity, and propagate folklore that glorified their powers.
On a trip to Morocco, I realised another reason why we might have been misled by the Ethnographic Atlas. Hiking through the Atlas mountains, I was surprised to see village women tending goats and taking cattle to water. Higher up, I spied a nomad summer camp, where entire families migrate for several months a year. This rocked my priors, for the Ethnographic Atlas codes their pastoralism as "male alone".
I wondered, had a patriarchal society exaggerated men's importance? If so, the close correlations observed between low female contributions to ancestral farming and patriarchal dominance could simply reflect male bias in coding. The direction of causation remains murky.
I underestimated the role of religion
Why do North and South India differ so much on gender? Why are South and North-east Indian women much more likely to survive infancy, be educated, marry later, choose their own husbands, interact more closely with their husbands, bear fewer children, own more assets, exercise more control over their dowry, socialise with friends, move more freely in their communities and work alongside men?
Why is female employment so low in the Middle East and North Africa (MENA)?
Three years ago, I dismissed the possibility it might be due to Islam. Instead I praised Michael Ross's theory that oil booms lowered the cost of agricultural and manufacturing imports, which made local industries uncompetitive and thereby lowered labour demand for women. I was wrong.
Oil did not cause patriarchy. Female seclusion was integral to male honour long before oil booms. Even if there is a giant oil discovery in a highly Muslim country, female labor force participation only falls by two percentage points over 10 years. Plus, the countries with the globally highest increase in female labour market participation are oil rich Qatar and Kuwait.
By tracing cultural evolution over several millennia, I realised that female seclusion only became normative after the Arab-Islamic conquests and Shariah law. In the 7th century, Arabs conquered vast swathes of territory across MENA. Conquered people gained rights and tax exemptions if they converted to Islam, recited the Quran, gained an Arab patron, and adopted tribal lineages. Patrilineal kinship was simultaneously reinforced (by Shariah law’s recognition of male agnates in inheritance and patrilineal ownership of children) and also threatened (by Muslim women’s inheritance rights). Cousin marriage provided a solution: consolidating family wealth, strength, and trust. It remains especially high in Muslim countries formerly under the Umayyad Caliphate. As Egyptians shifted from bilateral to patrilineal tribes, they restricted women’s rights and freedoms.
Iraq became the seat of the Sunni Muslim empire: Persian theologians managed state institutions of learning, and played a crucial role in developing Islamic ethics. They constructed men as intellectually superior, uniquely capable of reason, and thus rightful patriarchs. Men could only achieve piety by policing women. Clerics repeatedly prescribed gender segregation: barring women from communal prayers in the mosque. Open dissent was increasingly inhibited by close-knit tribes, fear of eternal damnation, and religious authoritarianism. In the 13th century, Mamluk sultan Barsbay and clerics claimed that Egypt’s plagues were Allah’s punishment for women’s unIslamic practices, they were ordered to stay at home. Looking back, I realise I under-estimated the impact of Islam.
During Islamic rule, North Indian society became more gender segregated. Government administration was based on patriarchal texts by Persian ethicists. Muslim rulers practised purdah and upwardly mobile families followed suit to gain prestige. India's caste-based society was already concerned by purity, but now even more women withdrew from public life. Once seclusion became normative across north India, men preserved their honour by guarding female kin, for rumours of misconduct would soil the family name.
Religious diversity may help explain why purdah persisted in North India long after the invasions. Hindus sought to protect ‘their’ women from outsiders. In the early 1900s in Uttar Pradesh, Hindu publicists broadcast unsubstantiated allegations of Muslim men's rape, abductions, and conversions. Today, 67% of Hindu and 80% of Muslim Indians believe 'it is very important to stop women in their community from marrying outside their religion'. But in South India (where Arabs came as traders rather than rulers), this resistance is much lower (only 35%).
I was also wrong about Christianity.
Medieval Europe was patriarchal but possessed a latent advantage, which would prove important centuries later under job-creating economic growth. Families were nuclear, without cousin marriage. How did these emerge? My prior was that deep wage labour markets and urbanisation accelerated exogamy. Young men and women worked in service outside their natal villages and then eloped. An alternative hypothesis - espoused by Goody, Henrich and Schulz - credited Christianity. I was extremely sceptical.
Reading more widely, scrutinising the evidence, I realised I was mistaken. So here are are the facts to the best of my understanding:
Patrilineal clans emerged in Europe as a result of colonisation by horse-riding steppe peoples.
From 300-1300 CE, the Roman Catholic Church and Carolingian Empire tried to stamp out cousin marriage and polygamy. Noble families leveraged incest prohibitions to prevent their rivals from consolidating wealth.
English families were nuclear before the Black Death. Peasants disregarded lineage and rarely exchanged work with extended kin.
There was broad compliance with myriad Church strictures, which cannot be explained by anything but religion. In the 14th Century, English marriages seldom occurred during Lent nor if men had prior relations with her kinswoman (as was proscribed by the Church).
Young men and women often worked in service until they saved enough to establish their own nuclear households.
The age of marriage was thus unusually high in North-Western Europe (mid-twenties), especially when wages were low.
The powerful, draconian and interfering Church had enforced nuclear families. Europe's cultural evolution has been indeed influenced by Christianity.
The Women, Peace and Security Index (incorporating employment, representation, violence, son bias, laws, and discrimination) ranks the most gender equal and most patriarchal countries. There is a trend:
Why was I repeatedly wrong about religion?
Progressive group think is likely part of the explanation. When I studied International Development at the London School of Economics, no one even mentioned religion. Nor did my friends, family, or academic peers. The only time Islam really came up was in relation to discrimination post-9/11. We all loathed the lies and hate propagated by right-wing media. No one I knew and respected emphasised the enormous cultural influence of religion. I was in a secular bubble. Ironically, my ignorance may have been exacerbated by my atheism, empiricism and nihilism. I did not fully appreciate the cognitive impact of faith, religious righteousness, and fear of eternal damnation. I was strongly biased against the proposition that people stopped marrying their cousins because they were instructed by the Church.
Comparative historical analysis has helped me understand each country better. I learnt so much more about the drivers of gender equality in Zambia by asking why it differed from other countries. Fukuyama put it best, “he who only knows one country understands none”. Culture, I quickly realised, mediates the rate at which women seize opportunities created by development and democratisation. But my initial explanations of global cultural variation were doubly wrong. I over-estimated the importance of labour demand in agriculture and under-estimated religion.
My ignorance was largely due to the novelty of my research project. I freely admit that before starting I did not know everything about ten thousand years of patriarchy.
Bias and group-think also played a part. Now I strive to destroy these blinkers and blind-spots, by reading widely and discussing that evidence with brilliant minds. Advancing knowledge is a collective endeavour and I'm grateful to those who help me get it right, eventually!