> For the complete documentation index, see [llms.txt](https://laurence-wilse-samson.gitbook.io/textbooks/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://laurence-wilse-samson.gitbook.io/textbooks/world-economic-history/part-i-foundations-pre-history-and-antiquity/chapter03_neolithic_revolution.md).

# Chapter 3: The Neolithic Revolution

> **Week 3**: The agricultural transition and its profound effects on gender roles, genetic diversity, and state formation
>
> **Key Papers**: Alesina, Giuliano & Nunn (2013), Ashraf & Galor (2013), Mayshar, Moav & Pascali (2022)
>
> **Methods Focus**: Geographic Instruments, Persistence, Hump-Shaped Relationships

***

## 1. Opening Vignette

**The First Harvest**

It is 10,000 BCE in the Levant, specifically in the settlement of Jericho. For generations, the Natufian people here have lived in a way that no humans before them had: they have settled down. They gather wild emmer grain, hunt gazelle, grind the grain into flour, and bake bread. But they are still gatherers, dependent on the whims of nature. If the climate cools—as it recently has during the Younger Dryas—the wild stands of emmer and barley will thin, and the community will face starvation.

One year, a group of Natufians makes a decision that will change the trajectory of the human species. Instead of consuming all the seeds they have gathered, they save the largest, plumpest kernels. They clear a patch of land near the spring, break the soil with stone tools, and deliberately plant these seeds. They weed the plot, protect it from birds, and wait.

When the harvest comes, the yield is miraculous. The planted seeds produce more grain than the wild stands ever did. The community has a surplus. They can store food for the winter. They can feed more children. They can build permanent stone houses. But this abundance comes at a cost. The work is back-breaking. The women spend hours every day kneeling to grind the grain, deforming their toes and knees (pathologies visible in their skeletons today). The population grows, but individual health declines; people are shorter, have worse teeth, and die younger than their hunter-gatherer ancestors.

This is the **Neolithic Revolution**. It was not a single event but a process that unfolded independently in at least seven locations around the world: the Fertile Crescent, the Yangtze and Yellow River valleys, New Guinea, Mesoamerica, the Andes, the Amazon, and the Eastern United States.

For 99% of human history, we were foragers. In a blink of evolutionary time, we became farmers. This transition is arguably the most important event in economic history. It created the economic surplus that allowed for the emergence of cities, writing, states, and eventually, modern civilization. But it also sowed the seeds of inequality, patriarchy, and infectious disease.

This chapter traces the "deep roots" of comparative development, asking a radical question: Could the economic differences between nations today—the wealth of the United States versus the poverty of the Congo—originate in decisions made by our ancestors thousands of years ago? The technology of the plow may have created gender norms that persist to this day. The "Out of Africa" migration shaped genetic diversity and development in ways still visible in the data. And the very crops societies chose to grow—cereals versus tubers—determined whether they would be ruled by powerful states or live in stateless societies.

***

## 2. Historical Context

### 2.1 The Agricultural Transition

The Neolithic Revolution (literally "New Stone Age") refers to the transition from hunting and gathering to agriculture and settlement. It began around 10,000 BCE in the Fertile Crescent and spread to Europe, North Africa, and South Asia.

This transition involved a "package" of interconnected innovations. Humans domesticated plants—wheat, barley, peas, and lentils in the Near East—selecting for larger seeds, easier harvesting, and higher yields over countless generations. They domesticated animals as well: sheep, goats, cattle, and pigs provided meat, milk, leather, and eventually draft power. These changes demanded sedentarism, the shift from nomadic wandering to permanent villages where crops could be tended and livestock kept. Settled life in turn enabled pottery for storing grain and cooking, and polished stone tools—hoes, sickles, and grinding stones—for working the fields and processing harvests.

Agriculture was not necessarily "better" for the average individual. Hunter-gatherers had more varied diets, worked fewer hours, and suffered less from communicable diseases. The "adoption puzzle"—why humans switched to a lifestyle that reduced their standard of living—remains a major topic of debate. The leading theory holds that agriculture allowed for higher **population density**, giving farming communities a military advantage over foragers, despite lower individual well-being.

### 2.2 The Rise of Complexity

Agriculture created a **surplus**—a single farmer could produce more food than their family needed, freeing others to become potters, priests, soldiers, or kings.

This surplus drove the emergence of **social stratification**. Hunter-gatherer societies are typically egalitarian; agricultural societies are hierarchical. A central authority could tax, store, and redistribute the surplus, giving rise to the **state**—an organization with a monopoly on violence and the power to tax.

### 2.3 Divergence in Timing

For economic history, the critical fact is that this transition did not happen everywhere at the same time. The Fertile Crescent led the way around 10,000 BCE, with China following independently around 8,000 BCE. Mesoamerica developed agriculture by roughly 5,000 BCE, while the eastern United States saw indigenous cultivation emerge only around 2,500 BCE. Australia, isolated by ocean and lacking easily domesticable species, never developed agriculture indigenously at all.

Jared Diamond, in his famous book *Guns, Germs, and Steel*, argued that these differences in timing—driven by the availability of domesticable plants and animals—explain modern inequality. The societies that adopted agriculture first had a head start in developing "guns, germs, and steel" (military technology, immunity to crowd diseases, and metal tools), allowing them to colonize others.

***

## 3. Core Economic Questions

### 3.1 The Big Questions

**Question 1: Persistence** Do historical events from thousands of years ago still influence economic outcomes today? If so, through what mechanisms—geography, institutions, or culture?

**Question 2: Culture and Norms** Where do cultural norms come from? Specifically, why do some societies have strong patriarchal norms while others are more egalitarian? Can we trace this to the type of agriculture they practiced?

**Question 3: Genetic Diversity** Does the genetic composition of a population affect its economic performance? Is there a "goldilocks" level of diversity that is optimal for development?

### 3.2 Theoretical Framework

**The Persistence of Culture**: Economists usually take preferences (culture) as given. In this chapter, we treat culture as endogenous—it is determined by the economic environment of the past. Once formed, culture is "sticky"; parents transmit it to children, and it persists even when the economic environment changes.

**Malthusian Dynamics**: (Recall Chapter 1) For most of this period, the technological improvements of the Neolithic did *not* increase living standards. They increased population size. We are studying the *aggregate* effects (state formation, total GDP) rather than *per capita* improvements.

### 3.3 Challenges to Answering These Questions

1. **Data Scarcity**: No GDP or census data exists for 5,000 BCE. We must rely on archaeological proxies and ethnographic data.
2. **Identification**: Geography determines agriculture, but geography also determines trade, disease, and climate. How do we isolate the effect of *farming* from the effect of *place*?
3. **Reverse Causality**: Did the state arise because of agriculture, or did powerful leaders force people to farm?

> \[!NOTE] **Methodological Preview**: All three papers in this chapter address these challenges using **geographic instrumental variables** (geographic IVs)—features of the physical environment (soil type, climate, crop suitability) that predict historical choices but are themselves determined by geology and climate millions of years before humans arrived. This approach, introduced in Chapter 2, is the workhorse of modern cliometrics.

***

## 4. Data and Measurement

### 4.1 The Ethnographic Atlas

A primary data source for this chapter is the **Ethnographic Atlas**, compiled by anthropologist George Peter Murdock. It contains coded data on the cultural characteristics of 1,267 ethnic groups around the world, observed mostly in the 19th and early 20th centuries, before full industrialization.

**Key Variables**:

* **Subsistence**: % dependence on hunting, gathering, fishing, agriculture.
* **Agricultural Intensity**: Intensive (plow) vs. Extensive (hoe).
* **Family Structure**: Nuclear vs. Extended, Patrilocal vs. Matrilocal.
* **Political Integration**: Hierarchy levels beyond the local community.

**Connecting to Modern Data**: To use this for cross-country regressions, economists "match" these ethnic groups to modern country boundaries. If a country is composed of 40% Group A and 60% Group B, its "ancestral score" is a weighted average of the two.

### 4.2 Genetic and Biological Data

**Migratory Distance**: To measure genetic diversity without contamination from endogenous modern migration, scholars use a **proxy variable**—an observable quantity that stands in for an unobservable one. Here, the proxy is the walking distance from Addis Ababa (the cradle of humanity) to the indigenous location of a population, which predicts genetic diversity through the "Founder Effect" (see Section 5.2).

**Caloric Suitability**: Data from the FAO (Food and Agriculture Organization) allow us to calculate the *potential* yield of different crops (wheat vs. potatoes) for every grid cell on Earth, based on soil and climate. This provides a powerful **instrumental variable** (IV)—a variable correlated with the treatment of interest but not directly caused by the outcome. Here, crop suitability predicts what a society *should* have grown, based on geography alone, independent of their actual choices.

**BOX 3.1: DATA SPOTLIGHT - THE ETHNOGRAPHIC ATLAS**

> **Source**: George Peter Murdock (1967)
>
> **Coverage**: 1,267 societies observed 1800-1950. Global coverage but heavy on Africa and North America, lighter on Europe/Asia (where traditional cultures had already changed).
>
> **Key Variables**:
>
> * `v39`: Type of agriculture (Plow, Hoe, None)
> * `v33`: Jurisdictional hierarchy (0-4 levels)
> * `v43`: Descent (Patrilineal, Matrilineal)
>
> **Usage**: Scholars map these points to coordinates, then use "spatial matching" to link them to modern regions or countries.
>
> **Limitations**: It is a snapshot of "traditional" life, but observed often after colonial contact. It captures the "ethnographic present," not necessarily the deep past, though traits are assumed to be old.

***

## 5. Empirical Evidence

### 5.1 The Plow and Gender Roles: Alesina, Giuliano & Nunn (2013)

**Full Citation**: Alesina, Alberto, Paola Giuliano, and Nathan Nunn. 2013. "On the Origin of Gender Roles: Women and the Plough." *The Quarterly Journal of Economics* 128 (2): 469–530.

#### 5.1.1 Research Question

Why do we see such vast differences in female labor force participation (FLFP) and attitudes toward women's work across the world? In some countries (e.g., Rwanda, Madagascar), women work at high rates. In others (e.g., Egypt, India), they participate much less. The authors hypothesize that this divergence stems from the *type* of agriculture historically practiced.

#### 5.1.2 Historical Background: Boserup's Hypothesis

Ester Boserup (1970) distinguished between two types of farming:

1. **Shifting Cultivation (Hoe/Digging Stick)**: Used in Sub-Saharan Africa and parts of the Americas. Compatible with child-rearing. Women were the primary farmers.
2. **Plough Cultivation**: Used in Europe, North Africa, and Asia. The plow requires significant upper-body strength and is incompatible with carrying an infant. Men became the primary farmers; women were relegated to the home.

The hypothesis is that societies that adopted the plow developed norms that "women belong in the home." These norms persisted even after the plow became obsolete (e.g., in the industrial office economy).

#### 5.1.3 Data

The independent variable is ancestral plow use, constructed from the Ethnographic Atlas as the fraction of a country's ancestors who practiced plow agriculture. The outcome variables include current female labor force participation rates from the World Bank and gender attitudes from the World Values Survey (respondents' agreement with statements like "When jobs are scarce, men should have more right to a job").

#### 5.1.4 Empirical Strategy

**OLS**: Regress current FLFP on historical plow use, controlling for geography (tropical climate, soil type), religion, and development level.

$$
FLFP\_{c} = \alpha + \beta Plow\_{c} + \mathbf{X}'*{c}\gamma + \epsilon*{c}
$$

**Instrumental Variable (IV)**: The choice to use the plow is **endogenous**—meaning it may be correlated with the error term because unobserved factors (maybe sexist societies chose the plow?) drive both the treatment and the outcome. **Instrument**: *Plow Suitability*. Some crops (wheat, barley, rye) require the plow (deep roots). Others (sorghum, millet, roots) do not (shallow roots). The authors construct an index of "Plow Suitability" based on the FAO agro-ecological suitability of plow-positive vs. plow-negative crops. This relies only on climate and soil, which are exogenous.

#### 5.1.5 Results

The OLS regressions confirm that ancestral plow use is strongly associated with lower female labor force participation today. The instrumental variable estimates sharpen this finding into a causal claim: a one standard deviation increase in plow suitability reduces FLFP by significant margins. Most strikingly, the effect persists even among children of immigrants in the United States. Women descended from plow-using cultures work less in the US despite facing the same institutional environment as everyone else. The transmission mechanism is culture—internal beliefs passed across generations—not local institutions or economic constraints.

> \[!NOTE] **KEY FINDING: The Plow Shaped Gender Norms for 3,000 Years**
>
> Societies that adopted the plow have significantly lower female labor force participation today. The effect persists among children of immigrants in the U.S.—women whose ancestors used the plow work less even in America's institutional environment.
>
> **Implication**: The transmission mechanism is *culture*, not institutions. A technological choice millennia ago became a norm that outlived the technology itself.

<figure><img src="/files/4UPS5gwqs9TJe3YdgAsa" alt="Historical Plow Use and Female Labor Force Participation"><figcaption><p><em>Figure 3.1: Scatter plot showing the negative relationship between ancestral plow use and contemporary female labor force participation across countries. Societies that historically used the plow exhibit lower female workforce participation today, demonstrating the persistence of gender norms across millennia.</em></p></figcaption></figure>

#### 5.1.6 Interpretation

Technological choices made 3,000 years ago created a division of labor. This division solidified into a cultural norm ("men work, women stay home"). This norm persists today, affecting economic outcomes for billions of women, long after the technology that spawned it vanished.

**BOX 3.2: METHOD SPOTLIGHT - Geographic Instrumental Variables**

> **The Challenge**: How do we prove that historical plow use *caused* modern gender norms, rather than just being correlated with them?
>
> **The Solution**: Geographic IVs use features of the physical environment that:
>
> 1. **Predict** the historical practice (Relevance)
> 2. **Don't directly affect** modern outcomes except through that practice (Exclusion Restriction)
>
> **In This Paper**:
>
> * **Instrument (Z)**: Plow Suitability = FAO data on which crops grow well in a region's climate/soil
> * **Endogenous Variable (X)**: Ancestral Plow Use
> * **Outcome (Y)**: Modern Female Labor Force Participation
>
> **Why It Works**:
>
> * Relevance: Wheat-suitable climates adopted the plow (First Stage F-stat > 40)
> * Exclusion: Soil chemistry 3,000 years ago shouldn't affect modern office work, except through the cultural norms it created
>
> **Two-Stage Least Squares (2SLS)**: $$\text{Stage 1: } Plow\_c = \pi\_0 + \pi\_1 Suitability\_c + \mathbf{X}'\_c\pi\_2 + \nu\_c$$ $$\text{Stage 2: } FLFP\_c = \beta\_0 + \beta\_1 \widehat{Plow}\_c + \mathbf{X}'\_c\beta\_2 + \epsilon\_c$$
>
> **Interpretation**: $$\beta\_1$$ gives us the causal effect of plow adoption on modern gender norms, purged of reverse causality and omitted variables.
>
> **Threats to Validity**:
>
> * Exclusion violation: What if wheat-suitable climates also had other features (trade routes, urbanization potential) that independently affected gender norms?
> * Authors test: Control for agricultural productivity, climate variables, distance to trade routes. Effect persists.
>
> **Power of This Approach**: Geographic IVs are the workhorse of historical economics because geography is truly exogenous—determined millions of years ago, immune to human manipulation.

Figure 3.4 provides a visual representation of the instrumental variables logic described above. The diagram traces the causal chain from the instrument (plow suitability, determined by geography and soil) through the endogenous treatment (actual plow adoption) to the outcome of interest (modern gender norms and female labor force participation). Crucially, the instrument must affect the outcome *only* through the treatment--the exclusion restriction shown by the blocked direct path. This visual framework applies not just to the plow study but to the geographic IV strategy used throughout this chapter and across modern economic history.

<figure><img src="/files/a7DS33VnObqi2OUAPwye" alt="Instrumental Variables Diagram: Plow Suitability as Instrument"><figcaption><p><em>Figure 3.4: Instrumental variables identification strategy for the plow-gender hypothesis. Plow suitability (determined by geography and crop ecology) serves as the instrument (Z) for actual plow adoption (X), which in turn affects modern female labor force participation (Y). The exclusion restriction requires that plow suitability affects modern outcomes only through the channel of historical plow use and the cultural norms it generated--not through any direct path from ancient soil conditions to contemporary labor markets.</em></p></figcaption></figure>

#### 5.1.7 Robustness Checks and Alternative Explanations

Any finding this striking—that agricultural technology from 3,000 years ago affects gender norms today—invites skepticism. Alesina, Giuliano, and Nunn conducted extensive robustness checks to address the most serious concerns.

**Alternative Measures of Plow Use**

The authors test whether their results depend on how they measure ancestral plow use:

1. **Binary vs. Continuous**: Instead of the fraction of ancestors using the plow (0 to 1), they use a simple binary indicator (any plow use vs. none). Results hold.
2. **Ethnographic Atlas coding**: They verify that different ways of coding the original Ethnographic Atlas data (which has some ambiguous cases) don't drive the results. The effect remains consistent.
3. **Timing of adoption**: Some societies adopted the plow recently (19th century), others millennia ago. Do both matter equally? They find that *earlier* adoption has stronger effects, consistent with the idea that norms take time to solidify.

**Addressing Omitted Variables**

Perhaps plow-using societies differ in other ways that independently affect gender norms?

1. **Agricultural productivity**: Maybe plow societies are simply richer, and rich societies keep women at home? The authors control for agricultural suitability and GDP per capita. The plow effect persists.
2. **Religion**: Abrahamic religions (Christianity, Islam, Judaism) have patriarchal elements and historically used the plow. Is it religion or technology? The authors control for religion. The plow effect remains, and is found even within religious groups (e.g., comparing Christian societies with and without plow history).
3. **Political institutions**: Perhaps centralized states (which often used the plow) restricted women's rights? Controlling for historical state development, the effect persists.
4. **Climate and geography**: Maybe hot climates discourage outdoor work for women? The authors control for temperature, rainfall, latitude, and ruggedness. Results unchanged.

**Subsample Analyses**

Does the plow effect work everywhere, or only in certain contexts?

1. **Regional heterogeneity**: The effect is found in every major region (Europe, Asia, Africa, Americas), though strongest in Asia and Europe where plow use was most intensive.
2. **Developed vs. developing**: The effect holds in both rich and poor countries, suggesting it's not confounded with development level.
3. **Pre-industrial vs. industrial economies**: Remarkably, the effect persists even in fully industrialized economies where agriculture employs <5% of the workforce. This proves it's cultural transmission, not current agricultural practice.

**Immigrant Analysis: The Smoking Gun**

The most compelling robustness check uses second-generation immigrants in the United States—people born in the U.S. whose parents came from different countries.

* These women all face the *same* U.S. legal system, labor market, and institutions
* They differ only in their *ancestral* culture
* If plow use matters, it can only be through transmitted cultural beliefs, not contemporary institutions or economics

**Results**: Women whose ancestors used the plow work significantly less in the U.S., hold more traditional gender attitudes, and are more likely to say "men should have priority in jobs." This persists even in the second generation (grandchildren of immigrants), though the effect weakens over time.

> **Key Finding:** The immigrant analysis is the "smoking gun" for cultural transmission. Women who share the same U.S. institutions, labor markets, and legal system still differ in labor force participation based solely on their ancestors' agricultural technology. Culture, not institutions, carries the plow's legacy forward.

**Falsification Tests**

A good falsification test checks whether your instrument affects outcomes it *shouldn't* affect.

* **Male labor force participation**: If plow suitability directly affects modern labor markets (rather than working through gender norms), it should affect men too. It doesn't. Male LFP shows no relationship with ancestral plow use.
* **Attitudes on non-gender issues**: The authors check whether people from plow cultures differ on attitudes about democracy, trust in government, or environmental protection. They don't. The effect is specific to gender attitudes, not a general cultural conservatism.

These falsification tests strengthen the case that the mechanism is specifically about gender norms created by the sexual division of labor under plow agriculture.

**Measurement Error Concerns**

The Ethnographic Atlas was compiled from diverse sources, some more reliable than others. Could measurement error bias the results?

The authors show that:

1. Using only societies with "high quality" ethnographic data strengthens the results
2. Instrumental variables estimation (using plow suitability) corrects for measurement error in the endogenous variable (actual plow use)
3. Results are robust to excluding outliers or influential observations

Taken together, the evidence points to a robust, causal relationship rather than a statistical artifact.

***

#### 5.1.8 Policy Implications: Can We Overcome Deep Cultural Norms?

If gender norms are shaped by millennia-old agricultural practices, does this doom plow societies to persistent inequality? Or can policy break the path dependence?

**The Challenge of Norm Change**

Cultural norms are "sticky" because they're transmitted intergenerationally and reinforced by social pressure. A woman who wants to work may face:

* Family disapproval ("women belong at home")
* Employer discrimination (believing women are less committed workers)
* Lack of role models (few women in leadership positions)
* Structural barriers (lack of childcare, inflexible work hours)

These factors reinforce each other, creating a self-sustaining equilibrium. Breaking out requires coordinated changes across multiple dimensions.

**Evidence on Policy Interventions**

Despite this persistence, some policies have shown success in changing gender norms:

**1. Gender Quotas in Politics and Corporate Boards**

* **Norway's corporate board quota** (2003): Required 40% women on boards of public companies. Initially controversial, but research shows it:
  * Increased female representation in corporate leadership
  * Changed attitudes: younger women became more likely to aspire to leadership
  * Had spillover effects: more women in senior management below the board level
* **India's village council quotas**: Randomized quotas for female village leaders (Chattopadhyay & Duflo 2004) found:
  * Female leaders invested more in public goods valued by women (water, sanitation)
  * Exposure to female leaders reduced biases about women's effectiveness
  * Girls' aspirations and educational attainment increased in villages with female leaders

**2. Legal Reforms**

* **Inheritance laws**: Reforms giving daughters equal inheritance rights (India 2005, Kenya 2014) increased women's bargaining power within households and labor force participation
* **Divorce laws**: No-fault divorce and alimony protections increase women's outside options, encouraging labor force attachment

**3. Economic Interventions**

* **Childcare subsidies**: Quebec's universal childcare program (1997) significantly increased maternal employment
* **Conditional cash transfers**: Programs like Mexico's Progresa that condition payments on girls' school attendance have increased female education
* **Microfinance for women**: Evidence is mixed, but some programs show increased female autonomy and business ownership

**The Plow Context: Does History Matter for Policy?**

Alesina et al.'s findings suggest that the *same* policy may have different effects depending on cultural context:

* Gender quotas might face more resistance in plow societies (backlash, token compliance)
* But they might also be more *necessary* in these societies precisely because informal norms are so strong
* Complementary policies (childcare, flexible work) may be especially important in plow societies to overcome both normative and structural barriers

**A Case Study: South Korea vs. Sweden**

Consider two countries with different plow histories:

**South Korea** (intensive plow agriculture):

* Female LFP: 53% (2020)
* Persistent gender wage gap: 32% (highest in OECD)
* Strong cultural norms: "Women belong at home after marriage"
* Recent policies: Gender quotas, parental leave reforms
* Result: Slow improvement, but norms changing gradually among younger cohorts

**Sweden** (less intensive plow agriculture, hoe-based farming in medieval period):

* Female LFP: 80%
* Gender wage gap: 4%
* Egalitarian norms: Dual-earner model is default
* Policies: Generous parental leave, subsidized childcare, gender quotas

The comparison suggests historical norms set different baselines, but policy can shift equilibria—just more slowly in plow societies.

**Lessons for Development Policy**

1. **Patience is required**: Norm change takes generations. South Korea's female LFP has risen steadily since the 1970s, but still lags. Quick fixes don't work for deep cultural change.
2. **Comprehensive approaches work better**: Single interventions (quotas alone, childcare alone) are less effective than packages addressing multiple barriers simultaneously.
3. **Youth and education are key**: Norms change most easily in childhood. Education for girls, exposure to female role models, and gender-sensitized curricula can shift the next generation's beliefs.
4. **Economic incentives matter**: When women's work becomes economically essential (war, labor shortages, industrialization), norms adapt faster. Development policies should create economic opportunities for women.
5. **Backlash is real but temporary**: Strong interventions (quotas) often face initial resistance, but longitudinal evidence shows attitudes adjust within a decade.

**Can the Past Be Overcome?**

The plow paper demonstrates the power of historical persistence, but the policy evidence shows path dependence is not destiny. Societies can escape historical equilibria through deliberate intervention. However, the depth of historical roots matters for:

* How much effort is required
* How long change takes
* What resistance will be encountered

Economic history thus informs both the *diagnosis* (why inequality persists) and the *prognosis* (how hard it will be to change). Understanding that gender norms stem from agricultural technology 3,000 years ago doesn't excuse continued inequality—it explains why overcoming it requires sustained, comprehensive policy efforts rather than assuming markets or modernization will automatically erode traditional norms.

***

### 5.2 The "Out of Africa" Hypothesis: Ashraf & Galor (2013)

**Full Citation**: Ashraf, Quamrul, and Oded Galor. 2013. "The 'Out of Africa' Hypothesis, Human Genetic Diversity, and Comparative Economic Development." *American Economic Review* 103 (1): 1–46.

#### 5.2.1 Research Question

Does the genetic diversity of a population affect its economic development?

#### 5.2.2 The Hypothesis

The authors propose a "hump-shaped" (inverted-U) relationship between genetic diversity and economic development. Very low diversity harms development through lack of cross-fertilization of ideas, homogeneity, and groupthink. Very high diversity also harms development, but for opposite reasons: it undermines social cohesion, breeds mistrust and conflict, and makes coordination difficult. Intermediate diversity represents the "Goldilocks" zone—enough diversity to spark innovation, enough homogeneity to enable cooperation.

#### 5.2.3 Data and Identification

The authors measure genetic diversity using **heterozygosity**—the probability that two randomly chosen individuals from a population have different alleles (variant forms of a gene) at a given genetic locus. To address concerns about endogeneity in modern genetic data (shaped by recent migration and mixing), they instrument diversity using migratory distance from East Africa. This IV strategy follows the same logic introduced in Chapter 2 (Acemoglu, Johnson & Robinson): find an exogenous variable that predicts the endogenous treatment but does not directly affect the outcome.

The identification strategy exploits the "Founder Effect" from population genetics. As subgroups migrated away from the parent population, they carried only a subset of the original genetic diversity. Each subsequent migration wave further reduced diversity through random genetic drift. This creates a remarkably clean relationship: diversity declines almost linearly with walking distance from Addis Ababa, the cradle of humanity. Africa retains the highest diversity, Europe and Asia show intermediate levels, and the Americas and Oceania—settled last and via the longest migration routes—exhibit the lowest diversity.

#### 5.2.4 Results

The authors find a strong, robust hump-shaped relationship between predicted diversity and economic outcomes across multiple time periods. The relationship holds for population density in 1500 CE—the standard measure of development in the Malthusian epoch before modern growth—as well as for GDP per capita in 2000 CE. The persistence of this pattern across five centuries suggests the diversity effect is not a statistical artifact of modern data but reflects deep historical processes.

The "peak" of the inverted-U curve, representing optimal diversity for development, roughly corresponds to the diversity levels found in European and Asian populations. This finding is provocative: it suggests these regions may have enjoyed a biological advantage in the form of diversity levels that balanced innovation benefits against coordination costs.

> \[!WARNING] **KEY FINDING: The Diversity-Development Hump**
>
> Genetic diversity (instrumented by migratory distance from Africa) shows a **hump-shaped** relationship with development. Too little diversity → groupthink. Too much → coordination failures. The "Goldilocks zone" roughly corresponds to European/Asian diversity levels.
>
> **Caution**: This controversial finding proxies *cultural/trait* diversity, not racial differences. The authors emphasize the innovation-cohesion tradeoff, not biological determinism.

<figure><img src="/files/6FJ55zcgNXP61o5oDR3g" alt="The Hump-Shaped Relationship: Genetic Diversity and Development"><figcaption><p><em>Figure 3.2: The inverted-U relationship between genetic diversity (measured by expected heterozygosity) and economic development. Both very low diversity (e.g., indigenous populations in the Americas) and very high diversity (some African populations) are associated with lower development, while intermediate diversity shows optimal outcomes.</em></p></figcaption></figure>

#### 5.2.5 Synthesis and Critique

This paper is among the most controversial in modern economics. Linking biology to economic outcomes raises legitimate concerns about genetic determinism, and critics have challenged both the identification strategy and the interpretation. The authors emphasize that "diversity" here proxies for *cultural* and *trait* diversity, not racial superiority—the mechanism is the trade-off between the *benefits* of diversity (innovation, cross-fertilization of ideas) and the *costs* (coordination failures, social friction).

#### 5.2.6 Detailed Robustness Checks and Addressing Concerns

The Ashraf & Galor paper generated intense debate, partly because using genetic data in economics raises sensitive questions about determinism and racial differences. The authors conducted extensive robustness checks to address methodological and ethical concerns.

**Alternative Measures of Genetic Diversity**

The main measure—expected heterozygosity—comes from population genetics. But is the result robust to different measures?

1. **Ancestry-adjusted heterozygosity**: Using genetic markers that explicitly exclude recent admixture (mixing from colonization and migration post-1500), the hump-shaped relationship persists. This shows the effect stems from deep ancestral diversity, not recent population movements.
2. **Alternative genetic datasets**: The authors use data from multiple sources (HGDP-CEPH, Pemberton et al.). Results are consistent across datasets.
3. **Number of genetic markers**: With 50 markers vs. 500 markers, the diversity measures are highly correlated and produce similar results. This addresses concerns about measurement error in genetic data.

**Addressing Selection Bias in Migration**

A key concern: maybe the "smartest" or "most cooperative" groups migrated out of Africa, and that—not diversity per se—explains development?

The authors test this by:

1. **Within-continent variation**: Even within Africa, the hump-shape holds. East African populations (closer to the migration origin, higher diversity) perform differently than optimal-diversity populations within Africa.
2. **Path of migration**: They control for the specific migration route (northern route via Middle East vs. southern coastal route). The diversity effect persists, suggesting it's not about who migrated but how much diversity they retained.
3. **Time since migration**: Populations that left Africa earlier (Europe, Asia) vs. later (Americas, Oceania) differ in diversity but face similar developmental outcomes at similar diversity levels. This argues against time-dependent selection effects.

**Controlling for Confounding Factors**

Genetic diversity might correlate with other determinants of development:

1. **Agricultural timing**: Populations that left Africa later adopted agriculture later. The authors control for years since agricultural transition. The diversity effect remains.
2. **Disease environment**: High diversity populations (Africa) face higher disease burdens. Controlling for malaria ecology, pathogen diversity, and historical mortality, the hump-shape persists.
3. **Geographic factors**: Diversity correlates with distance from equator (latitude), access to coasts, and terrain ruggedness. Including comprehensive geographic controls doesn't eliminate the effect.
4. **Institutional quality**: Modern institutions might drive both diversity preservation and development. The authors show the relationship holds in 1500 CE (before modern institutions) and when controlling for current institutional quality.

**The Hump-Shape: Is It Real?**

A quadratic (U-shaped) relationship can be a statistical artifact if the data have outliers. The authors demonstrate the hump is robust by:

1. **Semi-parametric estimation**: Instead of assuming a quadratic form, they use flexible non-parametric regressions (local polynomial regressions). The hump-shape emerges from the data, not the functional form assumption.
2. **Piecewise linear specifications**: Breaking the diversity scale into three segments (low, medium, high), the effect is positive in the first segment, negative in the second—consistent with the hump.
3. **Excluding influential observations**: Removing Africa (high diversity, low development) or Western Europe (medium diversity, high development) weakens the relationship but doesn't eliminate it. Both tails of the distribution matter for the hump-shape.

**Temporal Persistence**

Does the diversity effect operate only in the distant past, or does it persist to the modern era?

The authors show the hump-shaped relationship for:

* Population density in 1500 CE (pre-colonial, Malthusian era)
* GDP per capita in 1500 CE (estimated from urbanization rates)
* GDP per capita in 2000 CE (World Bank data)

The effect is present in all periods, suggesting genetic diversity has been a persistent factor for millennia. This argues against the idea that modern institutions or technology have made diversity irrelevant.

***

#### 5.2.7 Contemporary Relevance and Policy Implications

While Ashraf & Galor study deep historical processes, their findings have sparked debates about modern policy, particularly around immigration and diversity.

**The Immigration Debate**

Does this research suggest countries should limit immigration to maintain "optimal" diversity?

**The authors' position**: No. The paper measures *ancestral* diversity (established over tens of thousands of years), not the effects of recent immigration. Contemporary migration happens on timescales far too short for the mechanisms in the paper (genetic drift, selection, cultural consolidation) to operate.

However, the paper does contribute to broader debates:

1. **Benefits of diversity**: The positive slope of the left side of the hump supports arguments that diversity fosters innovation. Homogeneous populations may lack the variety of perspectives needed for creative problem-solving.
2. **Costs of diversity**: The negative slope of the right side aligns with research showing excessive heterogeneity can reduce social cohesion, trust, and public goods provision (Alesina & La Ferrara 2005, Putnam 2007).
3. **Optimal diversity exists**: There may be a "Goldilocks" level of diversity where the benefits and costs balance optimally. But this is a long-run equilibrium, not a policy target for immigration control.

**Modern Diversity: Beyond Genetics**

In contemporary societies, *cultural* and *linguistic* diversity matter more than genetic diversity:

* Silicon Valley's success stems from immigration bringing diverse skills and perspectives, not genetic variation
* European Union integration creates economic benefits from cultural exchange, even among genetically similar populations
* Conflict often arises from cultural/religious differences (e.g., Yugoslavia), not genetic differences

The paper's mechanism—balancing innovation benefits against coordination costs—applies to any form of diversity, not just genetic. This makes it relevant for thinking about:

* Urban planning (diverse vs. segregated neighborhoods)
* Team composition in firms (homogeneous vs. diverse teams)
* Immigration policy (skill-based vs. origin-based selection)

**State Capacity and Diversity**

One mechanism the authors propose is that high diversity reduces state capacity (harder to coordinate taxation, enforce laws, provide public goods when the population is very heterogeneous). This has modern implications:

* Sub-Saharan Africa's state capacity challenges partly reflect high ethnic diversity (Michalopoulos & Papaioannou 2013)
* Nation-building efforts post-colonialism struggled partly due to arbitrary borders dividing or lumping ethnic groups
* Federal systems (U.S., India, Switzerland) manage diversity by allowing local autonomy

**Can Diversity Be Changed?**

Unlike the plow (a technology that can be abandoned), genetic diversity is essentially fixed for contemporary policy. The paper thus highlights a form of path dependence that's even more immutable than institutions or culture.

However:

1. **Mixing reduces diversity disparities**: Over centuries, global migration and intermarriage are homogenizing diversity levels. The extremes (very low diversity in isolated populations, very high in African populations) are gradually converging toward intermediate levels.
2. **Cultural integration can compensate**: Even genetically diverse populations can develop shared identities, languages, and norms that facilitate coordination. Nation-building, education, and inclusive institutions can mitigate the coordination costs of diversity.
3. **Technology changes the equation**: Digital communication, translation tools, and global markets may reduce the coordination costs of diversity, shifting the optimal level rightward.

#### 5.2.8 Methodological Critique: Is Migratory Distance a Valid Instrument?

While creative, the migratory distance instrument faces scrutiny. Does it truly satisfy the exclusion restriction?

**The Concern**: Distance from East Africa might affect modern development through channels *other than* genetic diversity:

1. **Agricultural timing**: Populations that migrated farther from Africa encountered agriculture later (Americas adopted agriculture 5,000 years after the Fertile Crescent). Late agriculture means late state formation, late urbanization, late technological development. This directly affects modern GDP.
2. **Disease environment**: Distance from Africa correlates with disease burden. Africa has the highest pathogen diversity (malaria, sleeping sickness, etc.). Disease directly reduces productivity and economic development.
3. **Biogeographic endowments**: Jared Diamond's *Guns, Germs, and Steel* argues that Eurasia's east-west orientation and abundance of domesticable species—not genetic diversity—explain its developmental advantage. Migration distance correlates with these geographic factors.
4. **Historical connections**: Populations closer to Africa had earlier contact with Eurasian trade networks (Silk Road, Indian Ocean), facilitating technology diffusion. Distance thus proxies for integration into the global economy.

**The Authors' Defense**:

Ashraf & Galor attempt to address these concerns by:

1. **Controlling for agricultural timing**: Including years since agricultural transition as a control variable. The diversity effect persists.
2. **Controlling for disease**: Including malaria ecology, pathogen richness, and climate controls. Results are robust.
3. **Controlling for geography**: Comprehensive controls for latitude, longitude, island status, landlocked status, terrain, and crop suitability. The hump-shaped relationship remains.
4. **Subsample analysis**: Looking only at the Old World (Africa-Europe-Asia), where populations had similar access to trade and technology, the relationship holds.

**Remaining Skepticism**:

Critics (see Easterly & Levine 2013) argue the IV may still be invalid. Some specific concerns:

* **Measurement error in dependent variable**: 1500 CE GDP is estimated with huge uncertainty. Genetic diversity might be capturing something else that correlates with better 1500 GDP estimates.
* **Omitted mediating variables**: Perhaps diversity → institutions → development. Controlling for modern institutions might be "over-controlling" (blocking the causal path), while not controlling for them leaves omitted variable bias.
* **Multiple testing**: With so many robustness checks and specifications, there's risk of selective reporting or data mining.

**The Broader Debate**:

This paper exemplifies a tension in empirical economics:

* **Supporters** praise it as creative use of exogenous variation from deep history to test a novel hypothesis
* **Critics** worry about genetic determinism, questionable IV validity, and policy misinterpretation

The consensus view: the paper demonstrates a **provocative correlation** that warrants further investigation. Whether it identifies a **causal mechanism** remains debated. The hump-shaped relationship is robust in the data, but whether genetic diversity *causes* development or merely correlates with unmeasured factors is unresolved.

> **Key Takeaway:** The Ashraf & Galor debate illustrates a broader lesson: even well-designed instruments face exclusion restriction challenges when the causal chain spans millennia. Readers should hold this finding with productive skepticism—appreciating its ingenuity while recognizing its limits.

**Lessons for Students**:

This debate teaches us:

1. **No instrument is perfect**: Even clever geographic/historical instruments have potential exclusion restriction violations
2. **Controversial findings invite scrutiny**: The more provocative the result, the higher the bar for evidence
3. **Robustness is necessary but not sufficient**: Passing many robustness checks increases confidence but doesn't prove causality
4. **Context and interpretation matter**: The same finding can be interpreted as genetic determinism or as evidence for optimal cultural diversity, depending on framing

***

### 5.3 Origins of the State: Mayshar, Moav & Pascali (2022)

**Full Citation**: Mayshar, Joram, Omer Moav, and Luigi Pascali. 2022. "The Origin of the State: Land Productivity or Appropriability?" *Journal of Political Economy* 130 (4): 1091–1144.

#### 5.3.1 The Puzzle

Why did states (taxation, bureaucracy) emerge in some places (Egypt, China, Peru) and not others (New Guinea, Congo)? **Standard Theory**: *Land Productivity*. Productive land supports density $$\rightarrow$$ State. **Authors' Theory**: *Appropriability*. States emerge only where the surplus is *easy to steal/tax*.

#### 5.3.2 The Mechanism

* **Cereals (Wheat, Rice, Corn)**: Harvested once a year. Stored in visible granaries. Dry and preservable. **High Appropriability**. (Easy to tax/confiscate).
* **Tubers (Potatoes, Cassava, Yams)**: Buried underground. Harvested year-round as needed. Rot quickly if dug up. **Low Appropriability**. (Hard to tax).

A bandit/king cannot tax a cassava farmer easily ("Come back next week, I haven't dug it up"). He can easily tax a wheat farmer ("The pile is right there").

#### 5.3.3 Evidence

Like Alesina et al. (Section 5.1), the authors use FAO suitability data as a geographic instrument.

* **Instrument**: Difference between *Cereal Suitability* and *Tuber Suitability*.
* **Outcome**: Political hierarchy (from the Ethnographic Atlas, described in Section 4.1) and modern state capacity.

#### 5.3.4 Results

Productivity (total calories available) does *not* explain state formation. **Appropriability** (cereal advantage) strongly predicts the emergence of hierarchy. Places that grew tubers remained stateless—not because they were backward, but because their surplus was "unlootable."

> \[!TIP] **KEY FINDING: States Emerged Where Surplus Was "Lootable"**
>
> Productivity alone doesn't explain state formation—*appropriability* does. Cereals (wheat, rice) are harvested once yearly and stored visibly → easy to tax → states emerged. Tubers (cassava, yams) stay underground and rot quickly → hard to tax → stateless societies.
>
> **Implication**: Ancient New Guineans weren't "backward"—their yams were just untaxable. The origin of the state is partly about what crops you grew.

<figure><img src="/files/anSsV2StrfJw4CuR0e7o" alt="Cereal Suitability and Political Hierarchy"><figcaption><p><em>Figure 3.3: Relationship between cereal crop suitability (relative to tubers) and the emergence of political hierarchy across societies. Regions where cereals were the dominant crop developed complex state structures because their annual harvests were easy to appropriate through taxation, while tuber-growing societies remained more egalitarian.</em></p></figcaption></figure>

#### 5.3.5 Robustness Checks and Alternative Explanations

The appropriability theory is elegant, but does it hold up to scrutiny? Mayshar, Moav, and Pascali conduct extensive robustness checks to distinguish their theory from alternatives.

**Separating Productivity from Appropriability**

The key challenge: cereals are often both productive *and* appropriable. How do we know appropriability—not just productivity—drives state formation?

The authors construct two separate measures:

1. **Total caloric suitability**: Maximum potential calories per hectare (productivity)
2. **Cereal advantage**: Difference between cereal suitability and tuber suitability (appropriability)

**Horse race regression**:

$$
Hierarchy\_i = \alpha + \beta\_1 Productivity\_i + \beta\_2 CerealAdvantage\_i + \mathbf{X}'\_i\gamma + \epsilon\_i
$$

**Results**:

* $$\beta\_1$$ (productivity): Small and statistically insignificant
* $$\beta\_2$$ (cereal advantage): Large, positive, and highly significant

**Interpretation**: Conditional on productivity, cereal-growing societies developed states while equally productive tuber-growing societies did not. This supports the appropriability mechanism over pure surplus theory.

**Geographic Variation**

Does the result hold across different regions, or is it driven by specific continents?

1. **Within Africa**: Among African societies, those in savanna regions suitable for sorghum (a cereal) developed hierarchies (Ghana, Mali, Songhai empires), while forest societies growing yams and cassava remained stateless (Igbo, Kikuyu). The pattern holds within a single continent.
2. **Within Americas**: Mesoamerican maize growers (Maya, Aztec) developed complex states, while Amazonian cassava growers largely remained in tribal organizations. Same continent, different crop types, different political outcomes.
3. **Old World vs. New World**: The effect operates in both hemispheres, suggesting it's not confounded with Eurasian advantages (domesticable animals, disease immunity, etc.).

**Addressing Reverse Causality**

Perhaps causality runs backward: powerful states *forced* farmers to grow cereals because they're easier to tax?

The authors address this by:

1. **Using pre-state suitability**: They measure crop suitability based on climate and soil before state formation. States couldn't have altered geography retroactively.
2. **Timing of adoption**: Archaeological evidence shows cereals were domesticated in regions where they grew wild naturally (Fertile Crescent for wheat, China for rice, Mesoamerica for maize). This suggests ecology drove crop choice, not state preference.
3. **Instrumental variable approach**: Using climate variables (rainfall timing, temperature ranges) that predict cereal vs. tuber suitability but couldn't be influenced by political decisions. Results hold.

**Alternative Mechanisms**

Could something other than appropriability explain the cereal-state correlation?

1. **Storage technology**: Maybe cereals require granaries, and granaries facilitate coordination, leading to states? The authors show that controlling for storage potential (dry climate), the cereal effect persists.
2. **Population density**: Cereals support higher population densities, and dense populations need governance? Controlling for population density (using Ashraf & Galor's data), appropriability still predicts statehood.
3. **Trade opportunities**: Cereals are easier to transport and trade, facilitating merchant classes and urban centers. Controlling for distance to navigable rivers and coasts, the effect remains.
4. **Warfare technology**: Cereal agriculture creates sedentary populations vulnerable to raiding, necessitating defensive institutions? The authors find the effect holds even for societies in isolated regions with limited warfare.

**Temporal Persistence**

Does appropriability matter only at the moment of state formation, or does it have lasting effects?

The authors show that:

* Ancestral cereal reliance predicts modern state capacity (tax collection efficiency, bureaucratic quality)
* Former tuber-growing regions have weaker governments today, even after controlling for GDP
* The effect persists across millennia, demonstrating deep institutional persistence

#### 5.3.6 Mechanisms: How Does Appropriability Lead to States?

The paper identifies appropriability as the key variable, but *how* exactly does a pile of wheat create a king? The authors propose and test several mechanisms.

**Mechanism 1: Elite Emergence Through Extraction**

**Theory**: When surplus is appropriable, a military elite can seize it and establish dominance.

In cereal societies:

* Harvest happens once per year in a concentrated period
* Grain is stored in visible granaries
* A small armed group can confiscate the granary contents, controlling the community's food supply

In tuber societies:

* Crops remain underground, harvested as needed year-round
* No visible stockpile to seize
* To extract surplus, elites would need to monitor every household constantly—prohibitively costly

**Evidence**:

* Societies with annual harvest crops (cereals, but also some nuts like acorns) developed stratification
* Year-round harvest crops (cassava, taro, breadfruit) correlated with egalitarian structures
* The timing of harvest matters as much as the crop type

**Mechanism 2: Taxation Technology**

**Theory**: States require a fiscal base. Cereals provide a "taxable technology" that makes bureaucratic extraction feasible.

Key features of cereal taxation:

1. **Observable**: Tax collectors can see the harvest in the field or count sacks in storage
2. **Predictable**: Annual cycles allow systematic tax assessment
3. **Storable**: Collected grain can fund armies, priests, and bureaucrats for months
4. **Divisible**: Grain can be redistributed to non-producing specialists (soldiers, artisans, priests)

Historical evidence:

* Early Mesopotamian cuneiform tablets (3000 BCE) are tax records for barley and wheat
* Egyptian hieroglyphs describe granary inventories and tax collection
* Chinese oracle bones track millet and rice tribute

In contrast, tuber taxation is logistically difficult:

* Roots rot quickly once harvested (cassava lasts days, not months)
* Underground storage makes inventory assessment impossible
* Year-round harvest makes systematic collection timing difficult

**Mechanism 3: Elite Coordination and Legitimacy**

**Theory**: States require not just extraction but also coordination among elites and legitimacy among subjects.

Cereals facilitate both:

1. **Elite coordination**: Large granaries become focal points for elite competition and cooperation. Control of the granary becomes the objective, creating incentives to formalize rules of succession and taxation.
2. **Legitimacy building**: Elites can claim roles as protectors of the granary, organizers of irrigation (for cereals), or mediators with gods for good harvests. Ritual and religious legitimation develop around harvest festivals.
3. **Specialization**: Cereal surplus funds full-time priests, scribes, and administrators. These specialists develop ideologies justifying the state (divine kingship, bureaucratic rationality).

Tuber societies lack these focal points:

* No granary to protect or distribute
* Less need for irrigation coordination
* Harder for elites to monopolize food security roles

**Mechanism 4: Military Organization**

**Theory**: States monopolize violence. Cereals enable standing armies.

* Cereals provide field rations for soldiers (dried grain, bread)
* Centralized granaries can provision armies during campaigns
* Agricultural surplus allows some men to train full-time for warfare rather than farm

Evidence:

* Early states (Egypt, Mesopotamia, China, Inca) all had professional militaries funded by cereal taxation
* Stateless societies typically relied on part-time warrior groups
* The shift from raiding bands to standing armies coincides with cereal agriculture and state formation

***

**Quantitative Test of Mechanisms**

The authors use **mediation analysis**—a technique that decomposes a total effect into pathways through intermediate variables—to test which mechanisms explain the cereal-state relationship:

1. **Does cereal advantage → population density → hierarchy?** Partial mediation (\~30% of effect)
2. **Does cereal advantage → surplus accumulation → hierarchy?** Strong mediation (\~50% of effect)
3. **Does cereal advantage → warfare → hierarchy?** Weak mediation (\~15% of effect)

**Interpretation**: Multiple mechanisms operate simultaneously. The primary channel is the ability to accumulate and control surplus (appropriability), which enables both elite extraction and state-building functions (military, redistribution, public goods).

#### 5.3.7 Modern Implications: State Capacity and Development Today

If crop types thousands of years ago shaped state formation, do these effects persist in the modern world? The answer appears to be yes, with important implications for development policy.

**State Capacity in Former Tuber-Growing Regions**

The authors document that regions historically dependent on tubers have weaker state capacity today, measured by:

* Tax revenue as % of GDP (lower in tuber regions)
* Bureaucratic quality indices (weaker)
* Legal system effectiveness (less developed)
* Property rights enforcement (more informal)

**Case Studies**:

**Sub-Saharan Africa (Mixed Cereal/Tuber)**:

* Coastal West Africa (yam belt): Historically stateless or small kingdoms. Modern states struggle with tax collection, rely on resource rents (oil), have informal economies.
* Sahel (sorghum/millet belt): Historical empires (Ghana, Mali, Songhai). Modern states have stronger (though still weak by global standards) administrative capacity.

**New Guinea (Tuber-Dominant)**:

* Highland societies grew taro and sweet potato for millennia
* Never developed large states pre-colonially
* Post-independence Papua New Guinea has among world's weakest state capacity: <15% tax/GDP ratio, limited government presence in rural areas, tribal governance persists

**Andean Contrast (Peru)**:

* **Highland Peru**: Potato-growing regions under Inca Empire developed state capacity through labor taxation (mit'a system) rather than crop confiscation
* **Coastal Peru**: Maize-growing regions had stronger state traditions
* Modern Peru: Highlands still have weaker tax compliance and more informal economies than coast

**Policy Implications**

Understanding the appropriability constraint helps explain persistent development challenges:

**1. Fiscal Capacity Building**

Traditional development advice: "Improve tax collection through better technology and enforcement."

Appropriability insight: In regions with historically informal economies (former tuber-growing areas), taxation faces cultural resistance because:

* Historical norms didn't include surplus extraction by authorities
* Trust in government is lower (no tradition of redistribution)
* Informal economy networks developed as substitutes for state services

**Policy adjustments**:

* Build fiscal capacity gradually, starting with taxes on visible, appropriable resources (property, imports, natural resources)
* Invest in state legitimacy (service delivery) before expanding taxation
* Consider alternatives to direct taxation (user fees, community contributions)

**2. Property Rights Formalization**

Development agencies often push for formal land titling in Africa and Latin America. But in historically stateless societies:

* Customary land tenure systems evolved without state enforcement
* Trust in community elders exceeds trust in government registries
* Formal titling can create conflict rather than security

**Appropriability perspective**: The "demand for state" in property rights enforcement varies with historical experience. Policies should:

* Work with existing informal systems rather than replacing them wholesale
* Build state capacity in dispute resolution gradually
* Recognize that different regions need different mixtures of formal/informal governance

**3. Decentralization vs. Central State Building**

Aid debates often frame this as "strengthen central government" vs. "empower local communities."

Appropriability insight: The optimal balance may depend on agricultural history:

* **Former cereal regions**: May benefit from rebuilding historical state capacity traditions
* **Former tuber regions**: May perform better with federal/decentralized governance that mirrors historical statelessness

**Example**: Ethiopia's ethnic federalism may be more appropriate for its diverse agricultural zones than a unitary state model imported from Europe.

**4. Understanding "Fragile States"**

Why do some states persistently fail to consolidate authority?

Standard explanation: Colonial boundaries, ethnic diversity, resource curse, conflict

Appropriability adds: Some regions never had strong centralized states because their agricultural base didn't require or enable them. "State failure" may partly reflect mismatch between imported Westphalian state models and local economic/agricultural realities.

**Implication**: State-building interventions should be calibrated to local capacity and history rather than assuming all countries should converge to the same governance model.

**Looking Forward: Does Appropriability Still Matter?**

In modern economies where agriculture is <10% of GDP, does crop type still matter?

The authors argue **yes**, through path dependence:

1. **Institutional persistence**: States built on cereal taxation developed bureaucracies, legal systems, and fiscal technologies that persist
2. **Cultural norms**: Attitudes toward government authority and taxation reflect millennia of experience
3. **Infrastructure**: Physical capital (roads, administrative centers) and human capital (trained bureaucrats) create increasing returns to state capacity

However, **opportunities for change exist**:

* Digital taxation (VAT, mobile money taxes) creates new appropriability for previously hard-to-tax sectors
* Globalization and aid provide alternative fiscal bases beyond domestic agriculture
* Democratic accountability can build state legitimacy even without historical state traditions

**Lessons for Students**:

The appropriability theory teaches us:

1. **Geography shapes institutions, which persist**: Soil type → crop choice → state formation → modern capacity
2. **Not all paths to development are the same**: Stateless societies weren't "backward"—they adapted to different agricultural constraints
3. **Policy must account for deep history**: Cookie-cutter state-building often fails because it ignores agricultural and institutional legacies
4. **Change is possible but path-dependent**: Modern technology and institutions can overcome historical constraints, but the starting point matters for the speed and form of development

**BOX 3.3: METHOD SPOTLIGHT - Testing Competing Theories**

> Mayshar, Moav & Pascali face a classic challenge: two theories predict the state.
>
> 1. **Productivity Theory**: Rich land → Surplus → State
> 2. **Appropriability Theory**: Taxable crops → State
>
> Both variables are highly correlated! (Cereals are often productive too)
>
> **Their Solution**: Horse Race Regression $$Hierarchy\_i = \alpha + \beta\_1 Productivity\_i + \beta\_2 Appropriability\_i + \mathbf{X}'\_i\gamma + \epsilon\_i$$
>
> * If $$\beta\_1 > 0$$ and $$\beta\_2 = 0$$: Productivity wins
> * If $$\beta\_1 = 0$$ and $$\beta\_2 > 0$$: Appropriability wins
> * If both > 0: Both matter
>
> **Result**: $$\beta\_2$$ is large and significant, $$\beta\_1$$ is near zero and insignificant
>
> **Lesson**: When two theories make similar predictions, you need variation that distinguishes them. Here: some places had high productivity tubers (productive but not appropriable), others had low productivity cereals (not productive but appropriable). This variation breaks the correlation and identifies the mechanism.

***

## 6. Synthesis and Debates

### 6.1 The Deep Roots of Divergence

These three papers build a unified argument: "Neolithic" choices—plow versus hoe, cereals versus tubers, migration paths out of Africa—cast a long shadow over modern development. Agricultural technology shaped culture: societies that adopted the plow developed patriarchal divisions of labor that persist millennia later. Crop type drove institutions: cereals enabled taxation and hierarchy, while tubers preserved egalitarianism. And migration history determined the distribution of human capital: the serial founder effect created predictable variation in genetic diversity that correlates with development outcomes today.

### 6.2 Determinism vs. Agency

Critics often charge this literature with **geographic determinism**: if development is determined by soil type or distance from Africa, where does policy fit?

The answer: these are "initial conditions," not destiny. They create friction or advantage. Policy *can* overcome them—promoting women's rights in plow societies, building state capacity in tuber regions—but it swims upstream against deep cultural currents. The distinction between "hard to change" and "impossible to change" matters enormously for development practice.

### 6.3 Methodological Lesson: The IV Revolution in History

All three papers use the same trick: **Geographic IVs**. They solve the endogeneity of history (people chose to farm) by using the exogeneity of geography (soil *could* support farming). This is the hallmark of modern cliometrics.

**BOX 3.4: SCHOLAR PROFILE - Nathan Nunn**

> **Nathan Nunn** (Ph.D. Toronto, 2005) is Professor of Economics at Harvard and one of the most influential practitioners of historical economics.
>
> **Research Agenda**: Nunn's work consistently asks: "How does history shape development?" He has shown that:
>
> * The slave trade (Chapter 2, Chapter 11) destroyed trust and institutions in Africa
> * The plow created gender inequality (this chapter)
> * Colonial forced labor systems persist for centuries (Chapter 11: Mita)
> * Potato adoption in the Old World drove population growth and urbanization
>
> **Methodological Signature**: Nunn is known for:
>
> 1. **Creative Instruments**: Using geographic features (distance to coast, climate suitability) or historical accidents (where Europeans could settle) to isolate causal effects
> 2. **Long-Run Persistence**: Demonstrating that shocks centuries ago still matter today
> 3. **Mechanism Testing**: Not just showing correlations, but testing *why* history persists (trust, norms, institutions)
>
> **Major Awards**: John Bates Clark Medal Finalist, Guggenheim Fellow, Canadian Economics Association Best Paper
>
> **Why He Matters**: Nunn has transformed economic history from a descriptive field into a causal science. His papers are models of clear research design and have spawned hundreds of imitators using the "deep roots" approach.

### 6.4 Integration with Broader Theories of Development

The three papers in this chapter—on the plow, genetic diversity, and state origins—are part of a larger "deep roots" literature. How do they fit with other major theories of economic development?

#### Diamond's Geographic Determinism

Jared Diamond's *Guns, Germs, and Steel* (1997) argues that geography—specifically, the availability of domesticable plants and animals—explains why Eurasia dominated the Americas, Africa, and Oceania.

**Points of agreement with this chapter**:

* Geography (soil, climate) shapes agricultural possibilities
* Agricultural choices (plow vs. hoe, cereals vs. tubers) have lasting institutional consequences
* The Neolithic Revolution's timing and form explain modern inequality

**Points of tension**:

* Diamond emphasizes biogeography (what *could* be domesticated), while Mayshar et al. emphasize crop *characteristics* (appropriability)
* Diamond focuses on continental-scale differences, while our papers find variation *within* continents (e.g., yam belt vs. sorghum belt in Africa)
* Diamond's mechanism is military conquest ("guns, germs, steel"), while our papers emphasize institutional development (state capacity, gender norms)

**Synthesis**: Geography provides the initial conditions (crop options), but institutions and culture mediate how those conditions translate into modern outcomes. Both matter.

#### Acemoglu & Robinson's Institutional Theory

*Why Nations Fail* (2012) argues that inclusive institutions (property rights, democratic participation) drive prosperity, while extractive institutions (elite exploitation) cause poverty.

**How our papers complement AJR**:

* **Origins of extractive institutions**: Mayshar et al. show *why* some societies developed extractive states (appropriable surplus). AJR show how these early states evolved into modern extractive regimes.
* **Persistence mechanisms**: Both literatures emphasize path dependence. AJR focus on elite resistance to change; our papers add cultural transmission (gender norms) and state capacity constraints.
* **Policy implications**: AJR advocate for institutional reform; our papers suggest the difficulty depends on how deep the roots go (harder to change norms from millennia of plow use than institutions from centuries of colonialism).

**Points of tension**:

* AJR downplay geography ("institutions trump geography"); our papers show geography → institutions → outcomes, making geography indirectly decisive
* AJR emphasize contingency and critical junctures (things could have been different); our papers emphasize deterministic geographic constraints

**Synthesis**: Institutions are the proximate cause of prosperity (AJR correct), but institutions themselves are shaped by deep historical factors like agricultural technology and biogeography (our papers correct). Both levels of causation matter.

#### Cultural Evolution and Gene-Culture Coevolution

Biologists and anthropologists study how cultural traits evolve through transmission, selection, and drift—analogous to biological evolution.

**Boyd & Richerson's framework** (dual inheritance theory) posits:

* Cultural traits are transmitted vertically (parent to child), horizontally (peers), and obliquely (teachers, media)
* Cultural selection operates: societies with adaptive traits outcompete those without
* Gene-culture coevolution: cultural practices (dairying) can select for genetic changes (lactose tolerance)

**How our papers fit**:

1. **Plow and gender norms**: Classic cultural persistence through vertical transmission. Parents teach children gender roles; these norms are "sticky" because they're reinforced across generations.
2. **Genetic diversity**: Ashraf & Galor bridge biology and culture. Genetic diversity creates trait diversity → affects innovation/coordination → shapes cultural/institutional outcomes.
3. **State formation**: Mayshar et al. could be interpreted through cultural group selection. Cereal-growing groups developed states, which gave them military advantages → states spread through conquest and imitation.

**Implications for mechanism**:

* Is plow persistence *genetic* or *cultural*? Evidence from immigrant studies shows it's cultural (second-generation immigrants from plow societies differ from neighbors with same genes but different culture).
* Gene-culture coevolution likely operates over very long timescales (lactose tolerance took 7,000 years). Institutional persistence can happen faster through pure cultural transmission.

#### Unified Growth Theory (Galor)

Oded Galor's *The Journey of Humanity* synthesizes the transition from Malthusian stagnation to modern growth.

**Key stages**:

1. **Malthusian era** (10,000 BCE - 1800 CE): Technology → population, not income
2. **Post-Malthusian transition** (1800-1870): Population boom + slow growth
3. **Modern growth** (1870+): Demographic transition + sustained per capita growth

**How Neolithic papers fit**:

* **Chapter 1** (Malthusian trap) sets the context: For 99% of agricultural history, surplus → more people, not richer people
* **This chapter**: Explains *variation within* the Malthusian era. All societies were poor per capita, but they differed in:
  * Population density (Ashraf & Galor's diversity)
  * Institutional complexity (Mayshar's states)
  * Gender roles (Alesina et al.'s plow)
* These differences proved decisive in the transition to modern growth (states could coordinate industrialization; gender norms affected female education and fertility)

**Synthesis**: The Neolithic Revolution created the conditions for the Malthusian era. Variation in Neolithic technologies (plow, crop type) created institutional diversity that persisted into the modern era and shaped how different societies escaped the Malthusian trap.

### 6.5 Open Questions and the Research Frontier

Despite the impressive evidence in these papers, many fundamental questions remain unanswered. This section outlines the frontier where you, as students, might contribute.

#### Question 1: Genetic vs. Cultural Transmission—Can We Distinguish Them?

**The puzzle**: When we see persistence across generations, is it transmitted through genes, culture, or both?

**What we know**:

* Immigrant studies (Alesina et al.) show culture persists even when people move to new environments → cultural transmission matters
* Genetic diversity (Ashraf & Galor) correlates with outcomes → genes might matter too
* But correlation ≠ causation: genetic diversity could proxy for cultural diversity rather than having direct biological effects

**Research frontier**:

* **Adoption studies**: Compare adopted children (different genes, same culture) to biological children (same genes, same culture). If adopted children from plow societies behave like their adoptive parents, it's culture; if like their biological ancestors, it's genes.
* **Twin studies**: Fraternal twins (50% shared genes) vs. identical twins (100% shared genes) raised apart. Do identical twins from plow societies have more similar gender attitudes than fraternal twins?
* **Epigenetics**: Can historical shocks (famine, trauma) create epigenetic changes that transmit across generations? This is biology, but not classic genetic selection.

**Why it matters**: If persistence is genetic, it's essentially permanent. If cultural, policy interventions can change it (with enough effort).

#### Question 2: What Triggers Norm Change?

**The puzzle**: If norms persist for millennia (plow → gender inequality), why do some societies change while others don't?

**What we know**:

* South Korea's female LFP rose from 35% (1970) to 53% (2020)—dramatic change despite deep plow history
* Saudi Arabia only recently (2018) allowed women to drive—slow change despite modernization
* Nordic countries shifted from patriarchal to egalitarian in a century

**Potential triggers**:

1. **Economic necessity**: Wars (WWII) drew women into factories, shifting norms permanently
2. **Education**: Female education creates aspirations and economic opportunities that challenge traditional roles
3. **Urbanization**: Cities are anonymous, reducing social pressure to conform to traditional norms
4. **Legal reform**: Quotas, anti-discrimination laws change behavior, which gradually shifts attitudes (contact hypothesis)
5. **Media and role models**: Seeing women in leadership normalizes it (Jensen & Oster 2009: cable TV in India reduced gender bias)

**Research frontier**:

* **Natural experiments in norm change**: Identify exogenous shocks (sudden policy changes, media access, migration) and track norm evolution
* **Tipping points**: Do norms change gradually (linear) or suddenly (nonlinear) once a threshold is crossed?
* **Backlash dynamics**: When do rapid changes trigger conservative backlash that reverses progress?

**Why it matters**: If we understand what shifts norms, we can design better interventions to promote gender equality, reduce ethnic conflict, and build trust.

#### Question 3: Why Do Some Institutions Persist While Others Change?

**The puzzle**: States formed on cereal taxation 5,000 years ago still have higher capacity today. But many other ancient institutions (slavery, theocracy, feudalism) have disappeared. What determines which persist?

**Hypotheses**:

1. **Increasing returns**: Some institutions get "stickier" over time because they create complementary investments (physical infrastructure, human capital, cultural identity). Once built, the cost of switching is prohibitive.
2. **Power and vested interests**: Elites benefit from extractive institutions and resist change. Inclusive institutions might be less sticky because more people benefit from reform.
3. **Compatibility with new technology**: Institutions that mesh with industrialization persist (bureaucratic states, property rights). Those incompatible disappear (guilds, serfdom).
4. **External competition**: Institutions that provide military/economic advantages spread through conquest or imitation. Inefficient institutions get selected out.

**Research frontier**:

* **Comparative case studies**: Why did European feudalism end but Indian caste persist? Both were hierarchical labor systems, but one disappeared and one endured.
* **Institutional complementarities**: How do bundles of institutions reinforce each other? (e.g., property rights + courts + banking)
* **Critical junctures**: When do shocks (Black Death, colonization, world wars) break path dependence vs. reinforce it? (Chapter 8 examines the Black Death as a critical juncture; Chapter 11 treats colonization.)

**Why it matters**: Development policy often tries to transplant institutions (democracy, markets) from rich to poor countries. Understanding persistence helps predict when transplants will "take" vs. be rejected.

#### Question 4: Inter-Group Competition and Institutional Selection

**The puzzle**: If cereal states had military advantages (professional armies, taxation), why didn't they conquer all tuber societies? Why does diversity persist?

**Possible explanations**:

1. **Geography protects**: Rainforest and highlands shield tuber societies from conquest (New Guinea, Amazon)
2. **Diminishing returns to conquest**: Tuber societies produce no surplus to extract, making them unprofitable to govern
3. **Adaptation**: Stateless societies develop guerrilla warfare, migration, and resistance tactics that neutralize state military advantages

**Research frontier**:

* **Conflict archaeology**: Studying ancient warfare patterns to see if cereal states systematically conquered tuber societies
* **State formation under threat**: Do tuber societies form states when faced with external threats (defensive state formation)?
* **Institutional diffusion**: Can states spread through imitation rather than conquest? (e.g., African kingdoms adopting European state models)

**Why it matters**: Modern state-building efforts (Afghanistan, Somalia) often fail. Understanding when states emerge organically vs. are imposed from outside informs intervention design.

#### Question 5: Measurement and Data Challenges

**The puzzle**: How accurate are our measures of ancient institutions, genetic diversity, and cultural traits?

**Concerns**:

* **Ethnographic Atlas**: Observed 1800-1950, not ancient times. How much change occurred between Neolithic and modern era?
* **Genetic diversity**: Based on small samples. What about unmeasured genetic variation?
* **Crop suitability**: FAO data assumes modern crop varieties and farming techniques. Did ancient farmers face the same constraints?

**Research frontier**:

* **Archaeological verification**: Can we validate Ethnographic Atlas codings with archaeological evidence (settlement patterns, grave goods, artwork)?
* **Ancient DNA**: aDNA from archaeological sites can directly measure historical genetic diversity without relying on modern proxies
* **Historical crop yields**: Experimental archaeology (growing ancient crop varieties with ancient tools) can validate suitability assumptions

**Why it matters**: If our measures are noisy, estimated effects might be biased. Better measurement improves both scientific understanding and policy relevance.

### 6.6 Policy Implications: Navigating Deep Historical Constraints

If development is shaped by decisions made 10,000 years ago, what can policy do? This section synthesizes lessons across all three papers.

#### Implication 1: Diagnosis Before Prescription

**Standard development approach**: Apply universal best practices (democracy, markets, education) everywhere.

**Deep roots insight**: Identical policies will have different effects depending on historical context.

**Example: Gender Equality Policies**

* **Plow societies** (Egypt, India, Pakistan): Quotas face strong cultural resistance. Complementary policies needed:
  * Childcare subsidies (address structural constraints)
  * Public awareness campaigns (challenge norms)
  * Economic incentives (make female employment lucrative)
  * Patience (change takes generations)
* **Hoe societies** (Rwanda, Ghana, Botswana): Quotas face less resistance. Focus on:
  * Education (maximize existing egalitarian potential)
  * Economic opportunities (formal sector jobs)
  * Legal protections (prevent backsliding)

**Lesson**: Tailor interventions to historical starting points. One size doesn't fit all.

#### Implication 2: Work With, Not Against, Historical Legacies

**Standard approach**: Replace "bad" institutions wholesale (e.g., impose Western-style states in post-colonial Africa).

**Deep roots insight**: Institutions evolved for reasons. Replacing them creates governance vacuums.

**Example: State-Building**

* **Former cereal regions** (Ethiopia, Nigeria): Historical state traditions exist. Rebuild and modernize existing capacity:
  * Strengthen tax administration (but make it fair and accountable)
  * Invest in bureaucratic training
  * Build on existing centralization
* **Former tuber regions** (Congo, PNG): Statelessness was adaptive. Don't force centralization:
  * Hybrid governance: Mix formal state and customary authority
  * Decentralization and local autonomy
  * Bottom-up development (community-driven projects)

**Example: Property Rights**

* **High historical state capacity**: Formal titling and land registries can work
* **Low historical state capacity**: Customary tenure with state recognition may be more effective than full formalization

**Lesson**: Institutions have historical logic. Understand it before redesigning.

#### Implication 3: Leverage Positive Historical Legacies

Not all historical legacies are negative. Some can be policy assets.

**Example: Diversity**

* **Optimal diversity societies** (Western Europe, parts of Asia): Leverage innovation potential:
  * Invest in R\&D
  * Support entrepreneurship
  * Build innovation clusters
* **Low diversity societies** (Iceland, some Polynesian islands): Recognize cohesion advantages:
  * High trust enables cooperation
  * Public goods provision easier
  * Focus on mobilizing collective action for development

**Example: Gender Egalitarianism**

* **Hoe societies** with historical female labor force participation can:
  * Build on existing norms to achieve gender parity faster
  * Position themselves as leaders in gender equality
  * Attract gender-conscious investors and talent

**Lesson**: Historical advantages are as real as historical constraints. Identify and build on them.

#### Implication 4: Invest in Norm Change, But Be Patient

Cultural norms are slow to change, but change is possible.

**Short-term** (5-10 years):

* Policy can change behavior through incentives and mandates
* Behavior change doesn't immediately shift attitudes (people comply grudgingly)

**Medium-term** (10-30 years):

* Exposure to new behaviors gradually shifts attitudes (contact hypothesis)
* Generational replacement: young people with new norms replace old

**Long-term** (30+ years):

* New norms become internalized
* Behavior shifts from externally motivated to intrinsically valued
* Path dependence on new equilibrium sets in

**Implication for policymakers**:

* **Commit for the long haul**: Gender quotas, state-building, and institutional reform take decades to show full effects
* **Don't judge too quickly**: Resistance and backlash in year 5 doesn't mean failure if year 20 shows progress
* **Invest in the next generation**: Education and socialization are the most effective long-run levers

#### Implication 5: Some Constraints Are Immutable, But Can Be Managed

Genetic diversity and geography are fixed. But their effects can be mitigated.

**High diversity societies**:

* Can't reduce diversity (nor should they)
* Can build institutions that facilitate coordination despite diversity:
  * Federal systems (local autonomy reduces coordination costs)
  * Proportional representation (inclusive political institutions)
  * Lingua francas and common education (build shared identity)

**Low diversity societies**:

* Can't increase historical diversity
* Can import diversity through immigration
* Can build networks for external knowledge exchange

**Tuber-growing regions**:

* Can't change ancestral crop patterns
* Can build state capacity through alternative fiscal bases (resource revenues, VAT, external aid)
* Can leverage digital technology to lower tax administration costs

**Lesson**: Path dependence doesn't mean determinism. Historical constraints can be worked around with creative policy.

#### Implication 6: Beware of Unintended Consequences

Policies that ignore deep historical roots can backfire.

**Example: Forced Centralization**

* Imposing unitary states on historically stateless societies (Somalia, Afghanistan) led to:
  * Resistance and civil war
  * State collapse
  * Reversion to clan/tribal governance
* Better: Recognize historical governance patterns and build hybrid systems

**Example: Rapid Gender Norm Change**

* Aggressive campaigns that attack traditional values without providing alternatives can trigger:
  * Backlash and conservative mobilization
  * Increased polarization
  * Reversal of gains
* Better: Frame changes as consistent with other values (e.g., economic prosperity, family wellbeing) rather than wholesale cultural assault

**Lesson**: History creates vested interests and identities. Reforms that threaten these face resistance. Smart policy minimizes disruption while achieving goals.

***

> \[!TIP] **Chapter Summary**
>
> * The **Neolithic Revolution** (beginning c. 10,000 BCE) was arguably the most important event in economic history: the transition to agriculture created surplus, enabled cities and states, but actually worsened individual health -- early farmers were shorter, sicker, and died younger than hunter-gatherers.
> * **Alesina, Giuliano & Nunn (2013)** showed that societies that historically adopted the plow have significantly lower female labor force participation today, with the effect persisting even among second-generation immigrants in the United States -- proving that culture, not institutions, transmits the plow's legacy across millennia.
> * **Ashraf & Galor (2013)** found a hump-shaped relationship between genetic diversity (instrumented by migratory distance from East Africa) and economic development: too little diversity limits innovation, while too much undermines social cohesion, with the "Goldilocks zone" roughly corresponding to European and Asian diversity levels.
> * **Mayshar, Moav & Pascali (2022)** overturned the standard theory of state formation by showing that **appropriability**, not productivity, explains where states emerged: cereals (visible, storable, taxable) gave rise to hierarchical states, while equally productive tuber-growing societies remained stateless because their crops were effectively untaxable.
> * Key methods introduced include **geographic instrumental variables** (using FAO crop suitability data as exogenous variation), **immigrant analysis** as a "smoking gun" for cultural transmission, and **falsification tests** that check whether instruments affect outcomes they should not (e.g., male labor force participation shows no plow effect).
> * The Ethnographic Atlas (1,267 societies coded by Murdock) serves as a primary data source, linked to modern countries through spatial matching -- enabling cross-country regressions that trace the effects of ancestral practices on contemporary outcomes.
> * The central takeaway is that decisions made thousands of years ago -- which crops to grow, which tools to use, which migration paths to follow -- created cultural norms, genetic compositions, and institutional structures that persist to the present day, demonstrating the power of "deep roots" in comparative development.

***

## 7. Looking Forward

The Neolithic Revolution set the stage, creating the agricultural societies that would dominate the next 10,000 years. The next chapter zooms in on one of the first complex economies to emerge from this revolution: **Mesopotamia**. These early states did not merely farm—they invented finance, trade contracts, and business law. The methods introduced here (geographic IVs, persistence analysis) will recur throughout the book, while Chapter 4 introduces new tools: structural estimation and time series analysis of ancient price data.

***

## Chapter Summary

**Key Takeaways**:

1. The Neolithic Revolution (10,000 BCE) was the shift from foraging to farming, creating the economic surplus.
2. **Persistence**: Choices made millennia ago (plow adoption) affect outcomes today (gender norms).
3. **Gender**: The plow made men the primary farmers, leading to patriarchal norms that persist.
4. **The State**: States arose not just from productivity, but from *appropriability* (cereals vs. roots).
5. **Diversity**: Genetic diversity, determined by migration distance from Africa, has a non-linear effect on development.

**Empirical Methods Learned**:

* **Geographic Instrumental Variables**: Using soil/climate to predict historical choices.
* **Spatial Matching**: Linking ethnographic data to modern country boundaries.
* **Hump-Shaped Relationships**: Testing for non-linear effects (quadratic terms).

***

## Discussion Questions

1. **The Neolithic Paradox**: If agriculture led to lower individual health, harder work, more disease, and shorter stature compared to hunter-gatherer life, why did humans adopt it? *Contemporary parallel*: Why do people work long hours in stressful jobs when research shows leisure and relationships better predict happiness? Does the Neolithic paradox teach us something about collective action problems and individual vs. group selection?
2. **Gender norms and path dependence**: Alesina et al. show that millennia-old plow use still affects female labor force participation today. Can you think of policies that might be specifically needed in plow societies to counteract this deep cultural bias? *Contemporary examples*: Compare gender quota policies in Norway (relatively successful) vs. India (mixed results). Both have quotas, but Norway has weak plow history while India has strong plow history. Does the historical difference matter for policy design? What about childcare subsidies, parental leave, or education campaigns—which might be more vs. less effective in plow societies?
3. **Appropriability vs. public goods**: How does the "Appropriability" theory of the state differ from the idea that the state exists to provide public goods (Olson's "stationary bandit" theory)? *Contemporary test case*: Why does Norway have very high state capacity despite being historically a low-cereal, fishing/pastoral economy? Does modern tax technology (digital records, financial monitoring) make appropriability less relevant? Or do oil revenues simply create a new form of appropriable surplus?
4. **Genetic determinism and ethics**: Critique the "Out of Africa" paper. What are the ethical and scientific risks of using genetic data in economics? *Contemporary debate*: Ashraf & Galor faced intense criticism, including calls to retract the paper. Some argued any use of genetic diversity in economics research is inappropriate regardless of findings. Others defend using biological data if handled carefully. Where do you stand? Can genetic data inform development policy without promoting genetic determinism? How should economists navigate this tension?
5. **Diversity trade-offs**: Ashraf & Galor find an inverted-U relationship between diversity and development—both very high and very low diversity are suboptimal. *Contemporary application*: How might this framework inform debates about immigration policy? Does the U.S. (already quite diverse) face different diversity trade-offs than Japan (very homogeneous)? What about the European Union—does diversity across countries create innovation benefits while diversity within countries creates coordination costs?
6. **State capacity and development**: Mayshar et al. show that former tuber-growing regions have persistently weaker states. Is weak state capacity always bad for development? *Contemporary examples*: Singapore and Switzerland both have small governments but high GDP. Somalia and Congo have weak states and low GDP. What's the difference? Does the TYPE of state capacity (bureaucratic efficiency vs. territorial control vs. fiscal extraction) matter more than the overall level?
7. **Policy transplantation failures**: The papers suggest that identical policies have different effects depending on historical context. *Contemporary case study*: Afghanistan received billions in state-building aid from 2001-2021, attempting to create a Western-style bureaucratic state. It collapsed in weeks when U.S. support ended. Would the papers in this chapter have predicted this failure? What would a history-informed state-building strategy have looked like instead?
8. **Gender norms vs. economic incentives**: Which is stronger—deep cultural norms (plow effect) or economic incentives? *Natural experiment*: When oil prices spiked in the 1970s, Middle Eastern countries became wealthy but gender norms barely changed (women's labor force participation remained low). When China industrialized, women's labor force participation increased dramatically despite some plow history. What explains the difference? Does the TYPE of growth (resource windfall vs. labor-intensive manufacturing) matter?
9. **Norm change mechanisms**: Section 6.5 discusses triggers for norm change (education, urbanization, media, quotas). *Contemporary evidence*: South Korea's fertility rate dropped from 6 children per woman (1960) to 0.72 (2023)—the world's lowest. This is a massive norm shift in two generations. What triggered it? Can the same mechanisms that changed fertility norms change gender role norms? Why might fertility be easier to shift than gender attitudes?
10. **Measuring historical persistence**: All three papers rely on instrumental variables from deep history. *Methodological question*: Is migratory distance from East Africa (Ashraf & Galor's instrument) more or less credible than settler mortality (AJR's instrument from Chapter 2)? What makes a good instrument for testing historical persistence? Can we ever definitively prove causation rather than just correlation when looking at millennia-old factors?
11. **Inequality origins**: The three papers explain inequality through very different mechanisms: gender (plow), diversity (migration), and state capacity (crops). *Synthesis question*: Are these competing explanations or complementary? Could a single country be disadvantaged by all three (e.g., high plow use + suboptimal diversity + tuber agriculture)? Or do they operate at different levels (individual vs. group vs. institutional)?
12. **Development pessimism vs. optimism**: Do these papers make you more pessimistic or optimistic about development? *Debate*: **Pessimistic view**: If 10,000-year-old decisions still constrain us, poor countries are doomed by geography and history. **Optimistic view**: Understanding deep constraints helps us design better policies; South Korea and Botswana escaped historical disadvantages, so others can too. Which view is more supported by the evidence? What would it take to change your mind?
13. **Climate change and the Neolithic**: The Younger Dryas cold period (10,800-9,500 BCE) may have triggered agriculture adoption in the Levant by reducing wild food availability. *Contemporary parallel*: Could climate change trigger another fundamental economic transformation? What would a "post-agricultural" economy look like? Are modern debates about degrowth or sustainability comparable to the Neolithic transition?
14. **Cultural evolution vs. institutional design**: The papers show cultural persistence (gender norms, trust, state legitimacy). *Policy question*: Should development policy focus on changing culture or designing institutions that work with existing culture? *Example*: Rwanda post-genocide promoted gender equality through quotas and achieved 61% female parliamentary representation despite some plow history. Did they change culture, or build institutions that bypassed culture, or both?

***

## 9. Data Exercises

### Exercise 1: Replicating the Plow Result (Intermediate)

**Data**: Replication dataset from Alesina, Giuliano & Nunn (2013) (synthetic data available at `data/chapter03/plow_gender.csv`).

**Tasks**:

1. Load the data.
2. Regress current Female Labor Force Participation (FLFP) on `ancestral_plow_use_fraction`.
3. Add controls for `agricultural_suitability` and `tropical_climate`.
4. Observe if the coefficient on `ancestral_plow_use_fraction` remains negative and significant. Discuss the interpretation.

***

### Exercise 2: The Diversity Curve (Advanced)

**Objective**: Visualize and test the hump-shaped relationship between genetic diversity and development.

**Data**: Synthetic data based on Ashraf & Galor (2013) (`data/chapter03/genetic_diversity.csv`).

**Tasks**:

1. Load the data.
2. Create a scatterplot of `log_population_density_1500` (y-axis) against `genetic_diversity` (x-axis).
3. Fit a linear regression line to this scatterplot. Is it a good fit?
4. Now, fit a quadratic regression line (`y ~ x + x^2`). Check if the coefficient on the quadratic term is negative, indicating an inverted-U shape.
5. Identify the approximate "peak" of the curve where diversity is optimal for population density.

***