Detailed Summary: "Development: Which Way Now?

" by Amartya Sen

(1983)

What is this paper about? (Simple Overview)

This paper asks a big question: Is development economics — the study of
how poor countries can grow and improve — a failed subject? Some
economists at the time were saying yes. Sen disagrees, but he also says the field
has serious problems. He argues that development economics got some things
right about how to grow economically, but completely missed the bigger point:
economic growth is not the same as development. Real development
means improving what people can actually do and be in their lives. The paper
then introduces two key ideas — entitlements and capabilities — as better
ways to think about development.

Section I: The Promise and the Default — Is Development Economics

Dead?
The Starting Point: Hirschman's Obituary
Sen opens by discussing Albert Hirschman's essay called "The Rise and Decline
of Development Economics." Hirschman argued that development economics
had lost its excitement and credibility. It had originally promised to solve the
problem of backwardness almost on its own, but had failed to deliver.
Sen acknowledges there is some truth in this, but pushes back strongly. He
argues the obituary is premature.
What Was Development Economics Originally About?
Sen explains that development economics, born roughly in the 1940s and 1950s,
was built around four main policy ideas:
1. Industrialisation — poor countries needed to build industry to grow.
2. Rapid capital accumulation — countries needed to save and invest a lot
of money.
3. Mobilisation of underemployed labour — many people in poor
countries were not fully productively employed ("disguised
unemployment"), and putting them to work was key.
4. Planning and an active state — the government needed to guide the
economy, since markets alone would not solve the problem of
underdevelopment.
These ideas challenged the standard neoclassical economics of the time, which
trusted markets to solve everything. Development economists said: markets
alone don't work well enough in poor countries.
The Neoclassical Comeback
By the 1980s, neoclassical economists were hitting back. They pointed to the
spectacular economic success of South Korea, Taiwan, Hong Kong, and Singapore
— all of which were seen as examples of market-based, trade-oriented growth —
as proof that free markets worked best. They also pointed to failures in many
planned economies as proof that government intervention was harmful.
Critics also argued that:
 Enterprise (entrepreneurship), not capital, was the real
bottleneck — so emphasising capital accumulation was wrong.
 Industrialisation had often led to inequality and exploitation, especially
in Latin America, where it benefited foreign companies more than local
people.
Sen says these criticisms have some merit but go too far. Moving from
"industrialisation had bad effects in some places" to "industrialisation is always
wrong" is a big and unjustified leap.

Section II: Traditional Themes in the Light of Recent Experience — Did

the Old Ideas Actually Work?
Testing the Four Main Themes
Sen does something very concrete here: he actually looks at data from 32
countries (14 low-income and 18 middle-income) covering the period 1960–
1980, drawn from the World Development Report 1982. He tests whether the
traditional themes — capital accumulation, industrialisation, labour mobilisation,
and state planning — actually correlate with growth success or failure.
Capital Accumulation
The results are striking. Among low-income countries, the top three growers
(China at 3.7%, Pakistan at 2.8%, Sri Lanka at 2.4%) also tended to have the
highest investment rates. Sri Lanka invested 36% of GDP, China 31%. Among
middle-income countries, the top three growers (Romania at 8.6%, South
Korea at 7.0%, Yugoslavia at 5.4%) were also the top three investors
(Yugoslavia 35%, Romania 34%, South Korea 31%). Meanwhile, the worst
performers (Uganda, Sudan, Mozambique in low-income; Ghana, Peru, Chile in
middle-income) also tended to have the lowest capital accumulation. Ghana,
the worst middle-income performer, invested only 5% of its GDP. This is almost
exactly what the old Harrod-Domar model predicted.
Conclusion: Capital accumulation still matters. The traditional view was
not wrong.
Industrialisation
The correlation here is slightly less clean but still holds. The top growers in both
groups also tended to have higher industrial shares of GDP. Uganda (lowest
grower in low-income) also had the lowest industrial share; Ghana (lowest in
middle-income) also had the lowest industrial share in that group. At the bottom
end, the picture is messier — Peru, for example, had a high industrial share but
low growth — but overall, industrialisation still seems relevant, especially for the
poorest pre-industrial countries.
Labour Mobilisation
The thesis about disguised unemployment has been attacked (notably by
Theodore Schultz), but Sen argues the attackers haven't fully won. The key point
isn't whether the marginal product of rural labour is exactly zero — it's whether
mobilising labour into productive use matters. And indeed, the high
growers — China and South Korea especially — had outstanding records of
labour-absorbing growth, using their large populations productively.
State Planning
This is harder to measure quantitatively, but qualitatively the picture is clear.
Among the top growers, China has an obviously activist state. Sri Lanka had
active government intervention in health, education and food. Romania and
Yugoslavia had central planning. And South Korea — often cited by free-market
advocates as a triumph of the market — was actually deeply state-directed. The
government controlled financial institutions including nationalised banks, forced
export-oriented growth after years of protectionist import substitution, and
restricted many imports. Sen famously quips: "If this is a free market, then
Walras's auctioneer can surely be seen as going around with a government white
paper in one hand and a whip in the other."
Sen's Balanced Conclusion from Section II
The data does not debunk traditional development economics. Its themes —
capital accumulation, industrialisation, labour use, active state — remain
relevant for economic growth. However, Sen adds an important qualifier: these
strategies were designed for the conditions of the 1940s and 1950s, and need to
be adapted to current circumstances. But the time to "bury" traditional
development economics has not arrived.

Section III: Fast Growth and Slow Social Change — The Real Problem
Growth Is Not Development
Here Sen makes his most important critical point about traditional development
economics. Its real failure was not that it misidentified the drivers of economic
growth. Its failure was that it treated growth as the goal itself, rather than as
a means to some deeper ends. Growth matters only because of what it does for
people's lives — their health, their education, their freedom, their dignity.
The Key Example: Life Expectancy vs. GNP
Sen presents a powerful and simple table comparing five countries around 1980:

Life Expectancy GNP per head

Country
(1980) (1980, US$)

Brazil 63 years $2,050

China 64 years $290

 Life Expectancy GNP per head
Country
(1980) (1980, US$)

Mexico 65 years $2,090

South
65 years $1,520
Korea

Sri Lanka 66 years $270

The numbers are startling. China and Sri Lanka, with per capita incomes less
than one-seventh of Brazil or Mexico, had similar or better life expectancy
than those much richer countries. South Korea, despite being more than five
times richer than China and Sri Lanka in terms of GNP per capita, had not
surpassed them in life expectancy. If a poor country wants its people to live long,
healthy lives, trying to achieve that through growth first is far too slow and
indirect.
The Sri Lanka Calculation
Sen cites his own earlier research (1981) showing that if Sri Lanka had tried to
reach its life expectancy level through growth alone (the way typical developing
countries approach it), rather than through direct public action in health, food,
and education, it would have taken between 58 and 152 years to get where it
already was. Direct public action — providing healthcare, education, subsidised
food — got Sri Lanka there far more efficiently. As Sen drily notes: "money
answereth all things, but the answer certainly comes slowly."

Section IV: Entitlements and Capabilities — The New Framework

The Biggest Failure of Traditional Development Economics
Sen argues that the deepest problem with traditional development economics is
its focus on national product, aggregate income, and total supply of
goods, rather than on what people actually can do. Development, properly
understood, is about expanding the capabilities of people — whether they
can live long, avoid disease, be well nourished, read and write, take part in
community life, and so on. In Marx's phrase (which Sen quotes approvingly), it is
about replacing the domination of circumstances and chance over individuals
with the domination of individuals over their circumstances.
What is "Entitlement"?
Sen introduces his concept of entitlement clearly and simply. Entitlement refers
to the total set of commodity bundles (goods and services) that a person can
actually obtain, given all their rights and opportunities in society. In a market
economy this is determined by:
 What you own (your endowment — land, labour, skills, assets).
 What you can exchange those things for (your exchange entitlement —
wages, prices, market access).
Together, these determine what you can actually command. A person earning
$200 can command any combination of goods costing no more than $200. But
entitlements go beyond market exchange — in socialist economies, what the
state provides matters; in any economy, social security systems extend
entitlements beyond what markets alone provide.
Entitlements and Capabilities
From entitlements, people derive capabilities — the actual ability to do and be
things (be well nourished, be healthy, be educated). Economic development can
therefore be understood as a process of expanding people's capabilities,
which in turn depends on expanding their entitlements. Entitlements are
important, but they are derivative — they matter because of what they enable
people to do and be.
Why Famines Occur: An Entitlement Analysis
Sen applies this framework to the problem of famine, one of the most
devastating failures of development. The traditional approach explains famine
through food supply vs. population — famine happens when there isn't
enough food. Sen argues this is fundamentally wrong.
His key example: The Bangladesh Famine of 1974. A massive number of
people died in a year when food availability per head was actually at a peak —
higher than any other year between 1971 and 1975. What happened? Floods
disrupted agricultural work — rural labourers lost their jobs in rice planting and
transplanting. Without income, they couldn't buy food, even though food was
available in the market. Inflation also destroyed the purchasing power of those
who did have some income. People starved not because food wasn't there, but
because their entitlements collapsed.
This is a pattern Sen had documented across multiple historical famines — the
Great Bengal Famine of 1943, the Ethiopian famine of 1973–74, and the Sahelian
famines — all cases where famine struck without major reductions in overall food
supply. The problem was always entitlement failure, not supply failure.
Why Income Is Not Enough
Sen also critiques the common practice of measuring development through
income (even income distribution). Income is a crude proxy for entitlements. It
measures purchasing power in a single number, but:
 It cannot capture entitlements to things that cannot be purchased on the
market (e.g., living in a malaria-free area, access to public education or
hospitals that don't exist nearby).
 As prices of key goods rise, income's purchasing power over those goods
falls.
 For health, education, social equality, self-respect, and freedom from
social harassment, income is — in Sen's words — "miles off the target."

Section V: Political Complexities — Entitlements and Democracy

Entitlements Are Not Just Economic
The analysis of entitlements must go beyond purely economic factors. Whether
starving people can access food relief, for example, is not just a matter of
income — it is a matter of political power and pressure. Food, in a very real
sense, is sometimes "purchased" not with money but with political voice.
The Ireland and Bengal Examples
The Irish during the famine of the 1840s starved partly because they lacked
political power over the British government. The Bengalis in 1943 similarly
lacked effective political leverage to force a swift response. The Ethiopians in the
Wollo famine of 1973 faced the same political powerlessness. In each case,
timely public action could have prevented or greatly reduced mass death, but
the political pressure to force that action was absent.
India vs. China: A Fascinating Contrast
Sen presents one of his most famous and striking comparisons: India has
prevented famines since independence, but China eliminated chronic
malnutrition better than India. Yet China suffered a catastrophic famine
in 1959–61 that India has never experienced.
India's strength: A free press, active opposition parties, and democratic
accountability make it almost impossible for a famine to occur. No government
wishing to stay in power can ignore dramatic reports of starvation. Rapid relief
measures — food redistribution, imports, food-for-work programmes — get
mobilised quickly. Several potential famines (Bihar 1967–68, Maharashtra 1971–
73, West Bengal 1978) were averted this way.
India's weakness: The chronic, quiet, everyday malnutrition affecting roughly a
third of the rural population generates no newspaper headlines, no political
crisis, no parliamentary uproar. It is not dramatic enough to be "newsworthy."
The system absorbs it without political response. India's death rate in 1980 was
14 per thousand, compared to China's 8 per thousand — a reflection of this
chronic deprivation.
China's strength: The political commitment of the socialist system ensures
general concern with distributing food and livelihood more equally. China had
succeeded in virtually eliminating regular malnutrition. In a normal year, the
Chinese poor were much better fed than the Indian poor.
China's weakness: When the government pursued disastrously wrong policies
during the Great Leap Forward period (1958–61), there were no opposition
parties, no free newspapers, and no public pressure to force a correction. The
result was catastrophic. Sen presents mortality data released by Chinese
sources: death rates rose from 10.8 per thousand in 1957 to an average of 16.58
per thousand during 1958–61. This implies 14–16 million extra deaths during
the famine years — far more than the approximately 3 million extra deaths in the
Great Bengal Famine of 1943. One Chinese economist, Sun Yefang, reported the
death rate rose as high as 25.4 per thousand in 1960 alone, implying around 9
million extra deaths in that one year, with the four-year total around 15 million.
Sen notes that Chairman Mao himself, in a speech to 7,000 cadres in 1962 — just
after the famine — acknowledged that without democracy and communication
from the masses, good policymaking is impossible.
The paradox: India's democratic system prevented acute famines but couldn't
deal with endemic malnutrition. China's socialist system eliminated chronic
hunger but couldn't prevent a catastrophic famine when policies went wrong.
The challenging question Sen leaves open is: can the advantages of both
systems be combined?

Section VI: Concluding Remarks

Sen provides a clean five-point summary:
1. Traditional development economics identified the drivers of
economic growth reasonably well — capital accumulation,
industrialisation, labour mobilisation, and active state planning all retain
analytical relevance.
2. It failed to characterise what development actually is —
development is the expansion of people's capabilities, not just growth of
GDP. Growth is only a means, and often not an efficient one.
3. Focusing on entitlements — what commodity bundles people can
actually command — is a much better framework for understanding
development than aggregate national income or even income distribution.
4. Famines are best understood as entitlement failures, not food
supply failures. A famine can strike even when overall food supply is
adequate, if specific groups lose their ability to command food.
5. Political factors are central — entitlements are shaped not just by
markets and incomes but by political systems, press freedom, opposition
parties, and social pressure. The contrast between India and China
illustrates how profoundly political arrangements affect people's ability to
survive and flourish.

Key Themes Summarised Simply

Question Sen's Answer

Is development economics a
No — its obituary is premature
failed field?

Did capital accumulation and

Yes — the data broadly supports the traditional
industrialisation matter for
view
growth?

Is growth the same as No — growth is only a means; development is

development? about expanding people's capabilities

What is "entitlement"? The total set of goods and services a person can
Question Sen's Answer

actually command, through ownership and

exchange

Entitlement failures — specific groups lose

Why do famines happen?
access to food — not necessarily food shortages

Is income a good measure of No — it misses health, education, dignity, and

development? many other dimensions of real well-being

Absolutely — press freedom, democracy, and

Do politics matter for
political accountability shape whether people
development?
survive and flourish

Why This Paper Matters

Written in 1983, this paper is an early and powerful statement of what would
later become Sen's full-fledged Human Development approach — the
foundation for the UNDP's Human Development Index (HDI) and for much of
subsequent development policy thinking worldwide. It challenged economists to
stop treating growth as a goal in itself and start asking what people can actually
do and be. It also planted the seeds of Sen's later work connecting democracy,
free press, and the prevention of famines — an idea that became globally
influential.

Detailed Summary: "The Capability Approach: A Theoretical Survey" by

Ingrid Robeyns (2005)
Bibliographic Details
 Author: Ingrid Robeyns, Research Fellow, Department of Political Science
and Amsterdam School of Social Sciences Research, University of
Amsterdam
 Published in: Journal of Human Development, Vol. 6, No. 1, March 2005,
pp. 93–117
 DOI: 10.1080/146498805200034266

What is this paper about? (Simple Overview)

This paper is a comprehensive beginner-friendly guide to the capability
approach. Because the capability approach is used across many different
subjects — economics, philosophy, political science, development studies, social
policy — its literature is scattered across many journals. Robeyns wrote this
article to bring all the core ideas together in one accessible place. She explains
what the capability approach is, what its key concepts mean, how Sen's and
Nussbaum's versions differ, and what the main debates and criticisms around it
are.

Introduction: Why This Survey Was Needed

Robeyns begins by noting there has been an explosion of interest in the
capability approach among researchers and policy-makers. But because the
approach is so interdisciplinary — crossing economics, philosophy, political
theory, development studies — its literature is scattered and hard to navigate.
This paper fills that gap. Importantly, the paper focuses only on theory and
concepts — it does not deal with how to measure capabilities or apply them
empirically, as those topics are covered elsewhere.

Part 1: What is the Capability Approach?

Basic Definition
The capability approach is a broad normative framework — meaning it is a
way of thinking about what is right and valuable — used to:
 Evaluate individual well-being
 Assess social arrangements and institutions
 Design policies
 Propose social changes
It is used in development studies, welfare economics, social policy, and political
philosophy. It can evaluate inequality, poverty, individual well-being, and group
well-being. It has also provided the theoretical foundation for the Human
Development paradigm used by the UNDP.
An Important Clarification
Robeyns is careful to point out that the capability approach is not a theory that
explains poverty or inequality. It is a framework and tool for thinking about
and evaluating these phenomena. If you want to explain why poverty exists, you
need to add other explanatory theories on top of the framework.
The Core Idea
The heart of the capability approach is its focus on what people are effectively
able to do and to be. This is very different from:
 Approaches that focus on happiness or desire-fulfilment
(utilitarian/welfarist approaches)
 Approaches that focus on income, expenditure, or consumption
Sen argues that policies should focus on what people can do and be, the quality
of their lives, and removing obstacles to their freedom to live the kind of life they
have reason to value. The approach has intellectual roots in Aristotle, Adam
Smith, and Karl Marx, though its current form was pioneered by Sen from 1980
onwards.
Means vs. Ends — A Key Distinction
One of the most important ideas in the capability approach is the distinction
between means and ends. Only the ends — people's capabilities and well-being
— have intrinsic (in-themselves) importance. Means (like income, resources,
goods) are only valuable because of what they help people achieve. However, in
practice the distinction can blur: being healthy, for example, is both an end
(valuable in itself) and a means (it helps you work and do other things).
Functionings
Sen uses the word "functionings" to describe the things people do and are —
like working, resting, being literate, being healthy, being part of a community,
being respected. These are the actual activities and states of being that make a
life valuable. Functionings include both simple things (being well-nourished) and
complex things (participating in public life).
Capabilities vs. Achieved Functionings
This is one of the most important distinctions in the whole approach:
 Achieved functionings = what a person actually does or is.
 Capabilities = the real, effective opportunities a person has to achieve
various functionings — whether or not they choose to use them.
Think of it this way: a person might have the capability (the real opportunity) to
eat a nutritious diet, but choose to fast for religious reasons. Their achieved
functioning (actually eating) differs from their capability (being able to eat). The
capability approach says that what ultimately matters is that people have
the real opportunities — the capabilities — to lead the lives they value. Once
they have those genuine options, they are free to choose among them.
An example Robeyns uses: every person should have the opportunity to be part
of a community and to practice a religion. But if someone prefers to be a hermit
or an atheist, they should also have that option. The point is having the genuine
freedom, not being forced into any particular outcome.
Liberalism
Robeyns notes that the capability approach fits within the liberal tradition of
political philosophy — a tradition that values individual freedom. She is careful to
clarify that "liberal" here means the philosophical tradition valuing freedom, not
the everyday political left-or-right use of the word, and definitely not neoliberal
economic policies promoting free markets.

Part 2: The Capability Approach as an Alternative Framework

What it Rejects
The capability approach is in part defined by what it opposes. It rejects several
other frameworks as normatively inadequate:
Critique of Utilitarianism/Welfarism
Sen rejects welfarist theories — those that rely exclusively on utility (happiness,
desire-fulfilment, mental satisfaction). His reasons:
 Utilitarianism only uses "utility information" in moral judgements,
excluding important non-utility information. For example, the principle that
men and women should be paid equally for the same work has no intrinsic
value for a utilitarian — women should be paid less as long as they are
satisfied with lower wages. But most people find this counter-intuitive and
morally wrong.
 Sen does NOT say happiness is unimportant. He says it is wrong to rely
exclusively on mental states for moral evaluation.
Critique of Income-Based Approaches
Economists often use income as a proxy for well-being in applied work. But
income is only a rough proxy. As the capability approach stresses, people need
different levels of resources to reach similar levels of well-being (due to physical
disability, social discrimination, environmental factors, etc.). Moreover, income
cannot capture access to things that cannot be purchased — like living in a
malaria-free area, having a meaningful community, or having genuine political
participation.
Critique of Rawls's Primary Goods
Sen acknowledges a great debt to philosopher John Rawls, but criticises Rawls's
use of primary goods (basic rights, opportunities, income, wealth, self-respect)
as the metric for interpersonal comparisons. The problem: primary goods are
means, not intrinsic ends. If all people were the same, an equal share of primary
goods would give them equal opportunities. But people are different — a
disabled person needs more resources to reach the same capability level as a
non-disabled person. Focusing on primary goods misses this diversity.
Martha Nussbaum has extended this critique further, also examining the deeper
problem that Rawls's theory belongs to the social contract tradition (which
assumes roughly equal, independent, fully capable people), while the capability
approach does not.
Critique of Dworkin's Equality of Resources
Sen has also criticised Ronald Dworkin's resource-based theory of equality,
generating a highly abstract philosophical debate about the precise differences
between the two approaches.
What the Capability Approach Is NOT
Robeyns clarifies two common misunderstandings:
1. Some economists think the capability approach should provide a neat
formula or algorithm for welfare comparisons. It doesn't, and that is not
its purpose. Criticising it for not providing one misunderstands what it is
for.
 2. Some political philosophers think it is only a foundation for a theory of
equality or justice. It is broader than that.

Part 3: Means vs. Functionings — The Conversion Factor Framework

Goods and Services
Goods and services are not just money-equivalents. A bicycle is interesting not
because of its physical materials, but because of what it enables — the
functioning of mobility, getting from one place to another faster than walking.
Three Types of Conversion Factors
Between a good and the functioning it enables stand conversion factors —
things that affect how well a person can turn a good into a capability. There are
three types:
1. Personal conversion factors — metabolism, physical condition, sex,
skills, intelligence. Example: a person who has never learned to cycle
cannot use a bicycle for mobility. A disabled person may not be able to use
it at all.
2. Social conversion factors — public policies, social norms, discriminatory
practices, gender roles, power relations. Example: if a government or
cultural norm says women cannot cycle without a male family member
present, then women cannot use the bicycle for mobility even if they
physically own one.
3. Environmental conversion factors — climate, geography. Example: if
there are no paved roads, the bicycle cannot enable mobility effectively.
This framework shows why knowing what goods a person owns is not
enough to know what capabilities they have. You need to understand the whole
personal and social context in which they live. This is how the capability
approach accounts for human diversity: through diverse functionings AND
through diverse conversion factors.
Beyond Goods: Social Institutions as Means
Goods and services are not the only means to capabilities. Social institutions,
political arrangements, cultural norms, public goods — all of these shape
people's capability sets. The capability approach therefore insists on examining
the broader social and institutional context, not just material resources.
Robeyns cites Drèze and Sen's own statement that working within the capability
approach does not exclude analysis of resources or other means — those means
are just evaluated in terms of their actual effectiveness in enhancing people's
freedom and capabilities, not treated as valuable in themselves.

Part 4: Achieved Functionings vs. Capabilities — Further Clarification

Terminological Note 1: The Word "Capability"
In Sen's earliest work, "capability" referred to a person's entire opportunity set —
equivalent to the capability set as a whole. Many other scholars (including
Nussbaum) use "capabilities" in the plural to refer to the individual elements of
that set. Sen's later writings use both meanings interchangeably. Robeyns says
the plural usage is more straightforward for most readers.
Terminological Note 2: "Basic Capabilities"
"Basic capabilities" in Sen's framework are a subset of all capabilities —
specifically, the freedoms needed for basic survival and to avoid extreme
poverty. They are most relevant for poverty analysis in developing countries. In
richer countries, capability analysis typically focuses on capabilities beyond basic
survival.
When to Focus on Functionings Instead of Capabilities
In most cases, the capability approach says capabilities (the opportunities) are
what matter — because people should be free to choose. But there are situations
where looking directly at achieved functionings makes more sense:
 Domestic violence: We can safely assume no one wants to be beaten up
at home. If someone is being abused, their achieved functioning (bodily
safety) is harmed, which directly tells us their capability of bodily safety
was absent.
 Children and the mentally disabled: People who cannot make complex
choices should often be evaluated in terms of achieved functionings.
 Famine situations: In contexts of extreme material deprivation, the gap
between capability and functioning may be negligible — focusing directly
on whether people are well-nourished is more useful.
Capabilities and Choice in Real Life
Two people with identical capability sets will make different choices and end up
with different achieved functionings — because they have different ideas of the
good life. The capability approach, as a liberal framework, respects these
differences. However, Robeyns makes a subtle point: in real life, our choices are
deeply shaped by our family, religion, community, and culture. A child raised in a
Jewish household is very unlikely to end up Muslim. Is this really a free choice, or
a constrained one? The capability approach must grapple with this — it needs to
ask whether people genuinely have access to all the capabilities in their set, and
whether they face punishment for making certain choices.

Part 5: Well-Being vs. Agency — Sen's Distinction

Standard of Living, Well-Being, and Agency
Sen makes important distinctions between three related but different concepts:
 Standard of living = personal well-being related to one's own life.
 Well-being = standard of living plus the benefits of helping others
(sympathy — feeling better because you helped someone).
  Agency = well-being plus commitments — actions taken for reasons
beyond one's own benefit, even if they don't improve one's own well-
being.
Each of these can further be divided into achievements (what was actually
achieved) vs. freedoms (what one was free to achieve, whether or not one used
that freedom).
A Concrete Example — Anna and Becca
Robeyns uses a vivid example to illustrate these distinctions:
Two sisters, Anna and Becca, both oppose the power of global corporations. Anna
uses her agency freedom to travel to Genova to protest at the G8 summit.
Italian police beat her up in prison and offer her a deal: sign a false document
(getting a criminal record) or stay in jail longer. Anna must choose between
agency freedom (refusing to sign, maintaining her integrity) and achieved
well-being (signing, getting out of jail). Becca had the same agency freedom to
protest but chose not to use it — she is concerned but doesn't want to sacrifice
her well-being.
This example shows why these distinctions matter: in any evaluation, you need
to ask what the relevant dimension of advantage is — standard of living,
achieved well-being, agency achievement, well-being freedom, or agency
freedom. The capability approach says whatever dimension you focus on, the
information base should be in terms of functionings and/or capabilities.

Part 6: Key Differences Between Sen and Nussbaum

Robeyns identifies five major differences between Sen's and Nussbaum's
versions of the capability approach:
Difference 1: Different Goals
 Nussbaum's goal: Develop a partial theory of justice — specifically,
political principles that every government should guarantee to its citizens
through its constitution. She comes from moral-legal-political philosophy.
Her approach is universalistic.
 Sen's goal: More varied — he was simultaneously working on the
"equality of what?" question in liberal philosophy, on practical poverty
analysis in developing countries, and on social choice theory. His work is
closer to economic reasoning, formal/mathematical approaches, and
quantitative empirical applications.
As a result, Nussbaum's work engages more with narratives, literature, and
human stories. Sen's work is more aligned with formal, axiomatic modelling.
Difference 2: Conception of Capability
 Sen: Capability is primarily a real or effective opportunity (as in social
choice theory).
  Nussbaum: Capability also pays more attention to people's skills and
personality traits. She also introduces three sub-categories:
o Basic capabilities = innate abilities (note: different from Sen's use
of "basic capabilities").
o Internal capabilities = developed states of a person that allow
exercising a capability if external conditions permit.
o Combined capabilities = internal capabilities together with the
external provisions that actually enable exercise. Both Sen and
Nussbaum agree that politics should focus on combined capabilities.
Difference 3: The List of Capabilities
 Nussbaum: Proposes a specific list of ten Central Human Capabilities
(life; bodily health; bodily integrity; senses, imagination and thought;
emotions; practical reason; affiliation; other species; play; control over
one's environment — as detailed in the Nussbaum paper summarised
above). She argues Sen must also endorse such a list for his approach to
have any bite regarding justice. She argues without a list, any capability
could be argued as valuable — even the capability to abuse power.
 Sen: Consistently refuses to endorse one fixed, universal list. He argues
that selecting capabilities is the task of democratic processes, not
theorists. Different purposes need different lists; different social and
cultural contexts influence which capabilities matter; and public discussion
can deepen understanding of what capabilities are important.
Robeyns notes that both approaches face democratic problems: Sen's approach
leaves unclear how fair democratic processes for selecting capabilities will be
guaranteed. Nussbaum's approach has capabilities at such a general level that
undemocratic local decision-making can still produce problematic outcomes.
Difference 4: Role of Government
 Nussbaum: Frames her work as providing citizens with justifications for
what they can demand from their government. She believes in a
benevolent government implementing the capability list through
constitutional guarantees.
 Sen: Broader scope. Inequalities in capabilities can be discussed without
necessarily knowing how to fix them or assuming government is
responsible for all redistribution. Sen's approach allows for social change
through many routes, not just government action. Nussbaum has been
criticised for her assumption of benevolent government, particularly by
post-structuralist and post-colonial scholars.
Difference 5: Agency vs. Well-Being
 Sen: Maintains a clear distinction between well-being and agency (as
described above).
 Nussbaum: Does not endorse this distinction. She argues all important
distinctions can be captured within the capability/functioning distinction
alone. In her approach, practical reason plays an architectonic role — it
 goes beyond direct well-being contribution and is probably the main site of
agency. But how agency concepts differ between the two remains to be
fully explored.

Part 7: The Question of the List — The Central Debate

Why This is Controversial
The question of which capabilities matter, who decides, and how is the
most hotly debated issue in the capability literature. Nussbaum argues Sen's
approach has no normative bite without a list. Some economists also argue the
approach cannot be made operational without knowing which capabilities to
select.
Sen's Response
Sen argues the problem is not with listing capabilities, but with endorsing one
predetermined universal list. Selection should happen through democratic
processes. Lists vary by purpose (poverty assessment needs a different list than
human rights analysis), by social/cultural/geographical context, and are always
subject to refinement through public reasoning and debate. (This is exactly what
Sen argues in the 2004 paper summarised above.)
Nussbaum's Defence
Nussbaum emphasises that her list is highly general and abstract — it needs to
be made more specific by local people and democratic processes.
Shared Problems
Robeyns points out that both approaches face challenges. Sen's approach needs
much more work on the specific principles and procedures by which fair
democratic selection of capabilities happens. Nussbaum's approach, despite its
claim to generality, can still be applied undemocratically at local level. Some
scholars (like Qizilbash) suggest the list debate gets too much attention — many
existing lists are actually quite compatible with each other. More importantly, not
enough work has been done on the democratic institutions the capability
approach would require in practice.

Part 8: Individuals, Groups, and Social Structures — Is the Approach Too

Individualistic?
Robeyns addresses three common criticisms:
Claim 1: The Capability Approach is Too Individualistic
To evaluate this claim, Robeyns distinguishes three types of individualism:
 Ethical individualism: Only individuals are the units of moral concern.
When evaluating states of social affairs, only effects on individuals matter.
  Ontological individualism: Only individuals and their properties exist;
society is nothing more than the sum of individuals; social entities can be
reduced to individuals.
 Explanatory (methodological) individualism: All social phenomena
can in principle be explained in terms of individuals and their properties.
Robeyns's verdict: The capability approach embraces ethical individualism
(individuals are the units of moral concern), but does NOT rely on ontological
individualism. It fully recognises that people are social beings embedded in social
relationships and structures. This is shown in two ways:
1. The conversion factor framework explicitly recognises social and
environmental factors shaping what capabilities people can achieve from
their goods.
2. The distinction between functionings and capabilities incorporates the
influence of social structures and constraints on people's choices.
Sen himself has studied profoundly collective processes — such as households as
sites of "cooperative conflict" (where household members cooperate but also
have conflicting interests, affecting women's capabilities especially). The
approach is people-centred, but "social" in a rich sense — individual
opportunities depend greatly on relations with others and on what the state and
institutions do.
Conclusion: Claim 1 is wrong. Once ethical individualism is distinguished from
ontological individualism, most critics accept that ethical individualism is
worthwhile.
Claim 2: The Capability Approach Does Not Pay Sufficient Attention to
Groups
Strong version (the approach cannot pay sufficient attention to groups) is
obviously false — much capability research compares capabilities of women vs.
men, studies women's collectives in India, includes capabilities related to
community membership (affiliation in Nussbaum, social relationships in
Robeyns's own work, relationships and participation in Alkire), and the UNDP has
produced Human Development Reports on gender and culture.
Weak version (the current literature doesn't pay enough attention to groups)
has some merit. While Sen believes strongly in individual rationality and
resistance to social pressure, other capability scholars (Alkire, Nussbaum,
Iversen, Robeyns) pay much more attention to how social norms and group
processes shape individual choices and well-being. More dialogue with sociology,
anthropology, history, and gender/cultural studies is needed — but there is no
inherent reason the approach cannot fully incorporate group dynamics.
Claim 3: The Capability Approach Does Not Pay Sufficient Attention to
Social Structures
Similar analysis: social structures and institutions can and do have important
effects on capability sets (shown in Figure 1). In practice, policies can only
change the means of capabilities — the social determinants. So knowing the
social determinants (including structures and institutions) is crucial for any
practical application. The capability approach includes social structures in its
conceptual framework.
However, Robeyns notes an important limitation Sen himself acknowledges: the
capability approach only captures the opportunity aspect of freedom and
justice, not the procedural aspect. Institutions must be not only good in their
outcomes, but also procedurally just — e.g., global trade agreements should not
primarily benefit the most powerful nations, and discrimination in hiring is wrong
regardless of outcomes. The capability approach is not equipped to account for
these procedural dimensions of justice, which is a genuine limitation.

Conclusion
Robeyns concludes by summarising the key takeaways:
 The capability approach literature is scattered and interdisciplinary, which
is why a survey was needed.
 Its main characteristics are interdisciplinarity and a focus on the plural,
multidimensional aspects of well-being.
 It highlights the difference between means and ends, and between
substantive freedoms (capabilities) and outcomes (achieved
functionings).
 It is not a panacea — it cannot solve everything — but it provides an
important framework for analysing development, poverty, justice, and
social policy.

Key Concepts at a Glance — Simple Summary Table

Concept Simple Explanation

A framework for evaluating well-being that focuses on

Capability approach
what people can actually do and be

The actual things people do or are (working, being

Functionings
healthy, being part of a community)

The real opportunities people have to achieve

Capabilities
functionings — whether or not they use them

Achieved functionings What someone actually does or achieves

The full range of options (potential functionings)

Capability set
available to a person

The minimum set of capabilities needed to avoid

Basic capabilities
poverty and survive

Personal conversion Individual characteristics (health, skills) that affect how

factors well goods convert to capabilities
Concept Simple Explanation

Social conversion Social norms, policies, discrimination that shape what

factors capabilities are available

Environmental
Climate, geography that affect capability conversion
conversion factors

Acting for goals beyond one's own well-being — for

Agency
commitments and principles

Personal flourishing, including outcomes from

Well-being
sympathetic actions for others

Only individuals are units of moral concern (the

Ethical individualism
approach endorses this)

Ontological Society is just the sum of individuals (the approach

individualism does NOT rely on this)

Key Debates Summarised

Robeyns's
Debate Sen's Position Nussbaum's Position
Assessment

Both face
No — democratic
Should there be a Yes — a general, democratic
process should
fixed list of revisable list is needed challenges; more
select, purpose-
capabilities? for justice procedural work
relative
needed

Embraces ethical
Is the approach
but not
too Same Claim 1 is wrong
ontological
individualistic?
individualism

Weak claim has

Does it pay Confident in More attentive to merit; more
enough attention individual group/social norm dialogue with
to groups? rationality influences other disciplines
needed

Does it pay Yes, as means to

Procedural justice
enough attention capabilities; but
Similar is a genuine
to social procedural justice
limitation
structures? is a gap

Not needed — all

Well-being vs. Important captured in Further exploration
agency distinction capability/functioning needed
distinction
Summary: "Bargaining and Gender Relations: Within and Beyond the
Household" by Bina Agarwal (1997)

What is this paper about?

This paper is about how women and men negotiate power — inside the home
and outside it. Agarwal argues that existing economic models of the household
are too simple and miss many important things about how gender inequality
actually works. She wants to build a richer, more realistic picture of how
bargaining between men and women happens, what shapes it, and how it can
change.

Part 1: The Problem with Old Models of the Household

Economists used to treat the household as a single unit — like one person
making decisions. This is called the "unitary model" (associated with economist
Gary Becker). It assumed:
 All household members share the same preferences
 All income is pooled together
 An "altruistic" male head of household distributes resources fairly for
everyone's benefit
Agarwal says this model is wrong and dangerous because:
 Households are actually made up of people with different and often
conflicting interests
 Resources are not shared equally — especially between men and women
 Evidence from many countries shows persistent gender inequality inside
families

Part 2: Alternative Models — The Bargaining Approach

Economists began proposing bargaining models as alternatives. These
recognize that household members both cooperate (because cooperation
benefits everyone) and conflict (because within that cooperation, some
outcomes are better for one person and worse for another).
The key concept is the "fall-back position" (also called the "threat point") —
how well a person would survive if cooperation broke down. The better your
outside options, the more bargaining power you have.
There are different types of bargaining models:
 Cooperative models — members bargain but assume they'll reach an
efficient agreement
 Noncooperative models — members may not cooperate fully; no
assumption of efficiency or income pooling
 Separate spheres models (Lundberg & Pollak) — people have their own
domains defined by social norms; they may cooperate on some things (like
meals, childcare) but default to separate spheres rather than threatening
divorce for every small dispute
 Collective models — only assume the outcome is efficient; try to
uncover decision-making rules through data
Agarwal says all these models are useful but too narrow. She wants to go
beyond them.

Part 3: What Determines Bargaining Power? (The Neglected Factors)

A. Eight Key Factors Affecting a Woman's Bargaining Power (for basic
subsistence)
Agarwal identifies eight factors that determine how well a rural woman can
survive outside the family — and therefore how much power she has inside it:
1. Ownership of assets, especially land
2. Access to employment and income
3. Access to communal resources (village commons, forests)
4. Access to social support systems (kin, caste networks, patronage)
5. Support from NGOs
6. Support from the State
7. Social perceptions about her contributions and needs
8. Social norms
The first six affect her material survival options. The last two affect bargaining
more directly — through how people value her contributions and what rules
govern distribution.
B. Why Land is the Most Important Factor
Agarwal argues that land ownership holds a specially privileged position
among all these factors, for six reasons:
1. Village commons and forests are shrinking, so that source of women's
independent subsistence is disappearing
2. Traditional kin and caste support networks are eroding, especially among
poorer communities
3. Access to non-farm income is itself often linked to having land
4. Elderly women who own land can bargain for better care from their
families
5. During famines and droughts, women's assets (jewelry, utensils) are
sold first while men retain land — leaving women with far weaker fall-back
positions, sometimes leading to abandonment
6. Women in communities with matrilineal inheritance (land passes
through women) historically had more freedom, better treatment, and
greater autonomy — showing the link between land rights and overall life
conditions
Women who have received land through government programs or NGOs report
greater self-confidence, better treatment from husbands, and improved
economic security.

C. Bargaining Over Land Itself

Getting a share of family land is a second, deeper level of bargaining. A woman's
ability to claim land depends on:
 Existing inheritance laws
 Whether her claim is seen as socially legitimate (not just legally valid)
 Her education and legal literacy
 Her access to government officials
 Her economic ability to survive outside the support of those contesting her
claim (e.g., brothers)
 Her access to legal machinery (courts, lawyers)
This shows that intra-household bargaining is connected to extra-household
bargaining — with the community and the State.

D. Social Perceptions and Bargaining Power

There is often a gap between what a person actually contributes or actually
needs, and what others perceive those contributions and needs to be. This
matters because:
  Women's work is often labeled "unskilled" simply because it is done by
women
 Unwaged, home-based work is seen as less valuable than paid work
 Women's needs are often assumed to be subordinate to "family needs"
 In the labor market, women are often assumed to be less productive —
even when studies show this is wrong (e.g., in rural South Asia)
These incorrect perceptions reduce women's bargaining power both
inside and outside the home. Even if a woman strengthens her fall-back position
(say, by gathering from village commons), her contribution may be undervalued
because it is not monetized.

E. Not All Factors Are Equal — Prioritization Matters

Agarwal stresses that different factors carry different weight in different
contexts. For policy, it is important to identify which factors matter most. In
agrarian economies, land is the most critical factor. In urban or industrial
contexts, other forms of property matter more.

Part 4: The Role of Social Norms

Social norms affect bargaining in four ways:
i. Norms Set Limits on What Can Be Bargained About
Some things are treated as beyond question — what Bourdieu calls "doxa" —
the natural, self-evident order that is never discussed or challenged. For
example:
 That women are responsible for housework and childcare
 That men eat first or get better food
 That women shouldn't work outside the home (purdah norms)
These define what is even allowed to be negotiated. Women must first fight to
bring an issue into the "arena of contestation" before they can bargain over it.
ii. Norms Determine Bargaining Power
Norms like seclusion (purdah) directly weaken women's bargaining power by:
 Preventing them from working outside the home
 Restricting their access to markets, legal institutions, and public spaces
 Defining childcare as solely women's responsibility
 Institutionalizing lower wages for women
 Pre-defining men as "household heads" in land reform programs
iii. Norms Shape How Bargaining Happens
Gender norms dictate how people are expected to behave. Assertive behavior is
more accepted in men and older women than in young women or daughters-in-
law. This means women are often pushed to use implicit or covert bargaining
strategies:
 Persistent complaining
 Pleading illness
 Threatening to return to their natal home
 Withholding sex
 Playing family members against each other
 Withdrawing into silence
These covert strategies are generally less effective than open bargaining. The
story of Nhanu, the Tamang woman from Nepal, illustrates this — she sat silently
as her bronze bowl was given away, and only got it back because a brother
spoke up for her.
iv. Norms Can Be Bargained Over (They Are Endogenous)
Most economists treat norms as fixed and external. Agarwal argues they can
change — but it requires:
1. Economic pressure — poverty forces women to work outside the home,
challenging seclusion norms
2. Group strength — women acting collectively are far more able to
challenge norms than individual women (who can easily be penalized,
ostracized, or shamed)
3. Interaction between arenas — changing a norm requires contestation
not just in the household but in the community and with the State
The BRAC example from Bangladesh shows this vividly. BRAC women
organized in groups challenged purdah norms, noting that rich women's
violations of purdah were ignored while poor women's were condemned.
Economic necessity + group solidarity helped them gradually change behavior
both inside their homes (husbands became less abusive, allowed more freedom)
and in the community.

What Affects the Ability to Challenge Norms?

Three factors:
1. Economic situation — desperate poverty creates motivation to
challenge restricting norms
2. Control over wealth and property — those who own property also tend
to control institutions (media, religion, education) that shape norms;
village leaders define purdah in ways that suit powerful men
 3. Group strength and collective action — organizations like Grameen
Bank, BRAC, and SEWA show that women in groups can contest norms
more effectively

Part 5: Self-Interest, Altruism, and Women's Motivations

Do Women Have False Perceptions of Their Own Interest?
Amartya Sen argued that women in "traditional societies" may not perceive their
own self-interest clearly — that they may be complicit in their own oppression.
Agarwal challenges this.
The empirical evidence shows the opposite: women do perceive inequality and
do resist it — but often covertly, because overt resistance is risky:
 South Asian women secretly store rice, raise goats secretly, hide earnings
in bamboo poles or roofing
 Sri Lankan women hide money in different places to give husbands only
some of it
 Women hold clandestine picnics to get food denied to them in joint family
systems
 Women feign spirit possession to extract food otherwise denied them
 Bangladeshi women serve weak tea to their husbands' friends to
discourage their return
These examples show that compliance is not the same as acceptance.
Women know the system is unfair; they simply have limited ability to challenge it
openly.
Agarwal therefore shifts emphasis: rather than helping women realize they
deserve better, policy should help women believe they can do better and
actually strengthen their bargaining position.
Are Women More Altruistic Than Men?
Women in many societies do appear to act more altruistically — spending income
on family, forfeiting land claims to brothers, eating last. But Agarwal argues this
behavior is better explained by self-interest under constraint rather than
genuine altruism:
1. Regional variation — women in matrilineal communities (like the Khasis
of northeast India) don't wait for husbands to eat; this variation
undermines the idea that altruism is simply socialized into all women
2. Dependence — women with limited outside options may rationally
maximize "family" welfare because their own survival is tied to the
family's wellbeing
3. Son preference is self-interested — investing more in sons is rational
given that sons provide economic security and social mediation for women
in old age or widowhood
 4. Gift-giving to kin = building social capital — women who give gifts
and suppress their own claims are investing in future security
Even seemingly altruistic behavior (like giving up land claims to brothers) can
shift across generations: daughters today are more likely to claim land rights
than their mothers were, because kin support networks are weaker now. Both
generations are acting in self-interest, just under different conditions.
Agarwal's conclusion: both self-interest and altruism motivate behavior in
both women and men, but we cannot predict a priori which will dominate. This
makes bargaining outcomes hard to predict without context.
Intra-Household Coalitions
Bargaining doesn't always happen between individuals — there can be
coalitions within households:
 Mothers and children against husbands
 Co-wives cooperating
 Women acting on behalf of their children, not just themselves
A woman may be altruistic toward her children while bargaining hard against her
husband on their behalf. Women and children often form a distinct unit with
interests separate from — sometimes opposed to — the husband's. This is
documented in South Asia and Africa.

Part 6: Beyond the Household — Market, Community, and State

Agarwal argues that we cannot understand household bargaining in isolation.
Women's power inside the home is shaped by what happens outside it.
A. The Market
Women's bargaining power in labor markets is constrained by:
 Employers' false assumptions about women's abilities and commitment
 Women's domestic responsibilities limiting job options
 Cultural norms restricting women's access to certain jobs or public spaces
 Barriers to joining trade unions
 Women's own (or their household's) property status affecting their
"reserve price" of labor
Gender ideology shapes market bargaining just as much as it shapes household
bargaining.
B. The Community
A community (defined by location, caste, religion, ethnicity, etc.) involves
cooperative conflict similar to the household. An individual cooperates with
the community because membership brings economic, social, and political
benefits.
Women's bargaining power within the community is weaker than men's because:
 Women are excluded from or underrepresented in public decision-making
bodies
 Their weak household bargaining position weakens their community
position too
 In patrilocal communities (where wives move to the husband's village),
women lack the kin support networks that men have
Women's fall-back position relative to the community depends on:
 Direct property rights
 Skills, education, and economic opportunities independent of the
community
 Support from family/household
 External support from NGOs, women's groups, and the State
Acting as a group dramatically increases women's bargaining power with the
community — a single woman who violates a norm can be ostracized; a group of
women is much harder to punish.
In multi-caste, class-heterogeneous villages, caste and class also affect
bargaining power — poor or low-caste women's groups have less leverage than
high-caste or wealthy women's groups.
C. The State
The State is both a potential ally and an opponent of women's interests. It can:
 Enact gender-progressive laws
 Increase women's access to resources, employment, and education
 Protect against gender violence
 Shape media and educational discourse
But the same State can also reinforce gender-retrogressive practices through its
laws, bureaucracy, and coercive apparatus.
What motivates the State to cooperate with women's groups?
 Political pressure and voting power
 International public opinion and aid conditionality
 Recognition that NGOs and women's groups help deliver development
programs more effectively
However, the State is unlikely to support major redistribution of economic
resources (like land) that threatens its main political constituencies.
Importantly, the State is not a monolith — there are gender-progressive
individuals within gender-retrogressive State structures; women's bureaus exist
within overall patriarchal systems; local bureaucrats may resist or support
gender-progressive policies.
Women's bargaining power with the State increases with:
 Group size and organization
 Media and opposition party support
 Allies within the State apparatus
 The degree of democracy and judicial independence in the country
D. The Four Arenas Are Interactive
The household, market, community, and State all influence each other
simultaneously. They can:
 Reinforce each other (e.g., State + community + family all impose
seclusion norms)
 Contradict each other (e.g., State passes progressive laws but
community resists implementation)
 Create spaces for resistance (e.g., a gender-progressive State allows
women to contest community practices)
Gender-progressive coalitions and collective action are important in all four
arenas. As one BRAC woman put it: "A house cannot stand on one post. Put a
post in each corner and it is strong!"

Conclusion: Four Key Neglected Dimensions

Agarwal summarizes what existing household models fail to capture:
1. Complex, qualitative determinants of bargaining power —
especially the special importance of land in agrarian societies, and the fact
that these determinants are themselves contestable and changeable
2. Social norms — they are not fixed; they limit what can be bargained
over, shape how bargaining happens, and can themselves be bargained
over (endogenous in the long run)
3. Co-existence of self-interest and altruism — both motivate behavior
in both women and men; we cannot assume either pure self-interest or
pure altruism
4. Embeddedness of households in wider institutions — bargaining
happens simultaneously in the household, market, community, and State;
groups and coalitions are crucial determinants of bargaining power in all
arenas
Some of these can be incorporated into formal models and tested empirically
(e.g., the importance of land, the role of coalitions). Others — like social
perceptions, the complexity of norms, or the mix of altruism and self-interest —
are harder to formalize but must still be recognized.
The paper closes with a challenge: the analysis of collective bargaining over
social norms — not just over economic resources — opens a whole new field of
theory and policy that remains to be developed.
Summary of "Cleaning House" by Nancy Folbre (1986)
What is this paper about?
This paper argues that economists have been getting the household wrong. Most
economists either ignore what happens inside families, or treat the whole
household as if it were one single person with one set of goals. Folbre says both
approaches are mistaken. She reviews a large body of theoretical and empirical
literature to make three core arguments:
1. Household activities (cooking, childcare, subsistence farming) can and
should be analyzed using economic logic.
2. Men, women, and children have meaningfully different economic positions
inside the household — it cannot be treated as a unified, harmonious unit.
3. To fully understand households, you must also look at the bigger picture:
patriarchy as a social system, class structure, and the global capitalist
economy.

Section 1: Introduction
Folbre opens by noting that a huge share of economic activity in developing
countries takes place inside households — yet household economics gets pushed
to the margins of development economics. She identifies two big problems with
how economists usually handle this:
 The household seems "outside" the economy. Many economists
assume that self-interest and rational calculation only operate in markets,
not in families. So they either ignore households or treat them as driven
purely by love and generosity.
 The household is treated as one unit. Economists who do study
households often talk about "the household's decisions" or "the
household's interests," as if every member shared the same goals. This
erases the reality of conflict and inequality between husbands, wives, and
children.

Section 2: Economic Approaches to the Household as a Whole

2.1 Theoretical Issues
Moral Economy vs. Political Economy
A major obstacle to studying households economically is the widespread belief
that families operate on "moral" rather than "rational" principles. Anthropologists
of the "substantivist" school (like Sahlins) argued that people outside capitalist
markets don't behave like economic actors — they share and reciprocate out of
kinship duty, not self-interest. Applied to households, this becomes the simple
assumption that family members selflessly care for one another.
Both Neoclassical and Marxist economists, despite their differences, tend to
accept this assumption implicitly. Neoclassicists focus their rational-actor models
on markets; Marxists focus theirs on the sphere of commodity production. In both
traditions, the household gets assigned to a kind of moral bubble where
economic logic supposedly doesn't apply.
Folbre challenges this boundary. She argues that the line between self-interest
and altruism does not neatly coincide with the front door of the house.
The Neoclassical Approach (Becker)
Gary Becker and his followers tried to bring the household inside economics.
Their approach treats the household like a small firm: it maximizes a "joint utility
function," produces goods (including children), and allocates time and resources
efficiently in response to prices and incomes.
Key features of the Beckerian approach:
 Utility functions are assumed to be exogenously given (tastes don't
change with circumstances), which makes it possible to predict how
behavior responds to price and income changes.
 Children are treated as a kind of good — either "producer goods" (they
work and support parents) or "consumer durables" (they provide
emotional satisfaction). There is a tradeoff between having more children
and investing more in each child's "quality."
 The household's joint utility function sidesteps the problem of conflicting
individual preferences by simply assuming everyone's interests are
aggregated together.
The problem Folbre identifies: this approach assumes all household members
somehow share one set of preferences, which is just as arbitrary as assuming a
dictator runs the household. It also struggles with interdependent utilities — if
family members genuinely care about each other, standard Neoclassical tools
break down because you can't cleanly separate or aggregate their preferences.
Many mainstream economists avoid household economics precisely because of
this problem. They're happy to apply "rational expectations" to markets but resist
applying it to their own family lives.
The Marxist Approach
Marxist economists are skeptical of Neoclassical methodology, but they have
their own blind spot: they tend to restrict economic analysis to the sphere of
commodity production. In Marxist theory, capitalists pursue profit; workers
struggle over the products of their labor. But what about what happens inside
households?
Marx himself never seriously considered whether domestic labor (cooking,
childcare, etc.) affected the value of labor power. The "domestic labor debates"
of the 1970s tried to fill this gap. Some Marxists (like Dalla Costa and James)
argued that women's domestic labor creates value by lowering the cost of
reproducing the workforce for capitalists. Others (like Himmelweit and Mohun)
rejected this, arguing that without market competition to weed out inefficiency,
you can't apply the concept of "socially necessary labor time" to housework.
A broader Marxist argument — "functional dualism" — held that rural subsistence
production (much of it done by women) is useful to capitalism because it keeps
wages low, since workers can partly survive without full wages. But this
framework, like the Neoclassical one, treats the household as a unit and ignores
who inside the household actually benefits or suffers.
Both traditions, Folbre concludes, allow economic self-interest to operate upon
households (from outside) but not within them.

2.2 Empirical Studies: The Household as an Economic Unit

Despite theoretical reluctance, empirical evidence shows that households do
behave in economically rational ways.
Non-market production is systematically undercounted. Census methods
and national income accounts exclude unpaid household work, which
dramatically understates women's economic contributions. A study of India's
National Sample Survey showed that if activities like collecting firewood, tending
kitchen gardens, and caring for livestock were counted as economic, the female
labor force participation rate would jump from 30.5% to 52.3% — much closer to
the male rate of 63.7%.
Household production of goods vs. services. The distinction between
"economic activity" and "domestic work" reflects an implicit bias: goods
production tends to get counted, service provision does not. But why should
cooking be less economic than growing food? Why is childcare less productive
than caring for livestock? These activities become labeled "economic" only when
they enter the market — a circular and arbitrary definition.
Peasants respond to economic incentives. Empirical studies across the
developing world show that peasant households are responsive to wages, prices,
and income opportunities. They're not passive or tradition-bound; they calculate
costs and benefits. Studies of agricultural innovation, labor migration, and supply
responses to prices all support this.
The household-market boundary is blurring. Farm households in developing
countries increasingly have members doing both household production and wage
labor. An all-India survey showed that nearly all cultivating households both
bought and sold labor services. The sharp line between "household sector" and
"market sector" doesn't hold up empirically.
Risk matters. Households don't just maximize income — they manage risk.
Studies of agricultural decision-making show peasants using sophisticated
heuristics to navigate uncertain environments. This is economic behavior,
broadly defined.
Migration. The standard Neoclassical model treats migration as an individual's
rational calculation of expected earnings. Marxists tend to see it as a forced
response to poverty. Both miss the household dimension: how does migration
affect and get affected by the distribution of labor and income within the family?
The empirical record suggests remittances are often much smaller and less
regular than expected, raising questions about intrahousehold dynamics.

2.3 Empirical Studies: The Household as a Demographic Unit

Fertility decisions are among the most studied household behaviors. The
dominant economic framework treats children as goods with a "price" (the time,
money, and opportunity cost of raising them). Key findings:
 As the cost of children rises, desired family size falls. This relationship
holds across many countries and methodologies.
 Female employment raises the opportunity cost of women's time, which
tends to lower fertility. Women's earnings, even when low, often exceed
the implicit value of their domestic labor.
 Women's education reduces fertility more reliably than men's education. It
also improves child nutrition and health outcomes, because education
makes women more effective producers of family wellbeing.
 A study by Rosenzweig and Schultz argued that Indian families allocate
more food to male than female children because male wages are higher —
making it "economically efficient" to invest more in boys. Regional female-
to-male infant mortality ratios were inversely linked to female labor force
participation rates. Folbre later challenges this interpretation as too
narrowly efficiency-focused, arguing bargaining power better explains the
pattern.

Section 3: Altruism Versus Self-Interest in the Household

3.1 Theoretical Issues
Over the preceding years, altruism within the family had become a major
theoretical debate. Becker's approach rests on the idea that a benevolent
household "head" (implicitly the father) maximizes a joint utility function that
incorporates everyone's wellbeing. His "Rotten Kid Theorem" argues that even
selfish children will act in the family's collective interest, because a truly
altruistic parent controls wealth transfers and children know that harming the
parent ultimately harms themselves.
Folbre identifies a critical flaw: the theorem doesn't solve the "rotten parent"
problem. What stops the wealth-holding parent from using that control for
personal gain rather than altruism? If bargaining power shapes the household's
objective function, the Neoclassical assumption that this function is fixed and
exogenous is violated. As economic development shifts the relative wealth and
income of men, women, and children, household decision-making will shift too —
making the objective function endogenous to development itself.
Nash Bargaining Models
To address this, economists like Manser and Brown (1980) and McElroy and
Horney (1978) applied Nash bargaining theory to households. In these models:
 Household members negotiate outcomes, dividing the gains from
cooperation.
 Each person's threat point is their utility if the household dissolves —
i.e., what they could get on their own outside the marriage.
 The better someone's outside options (wages, property rights, legal
protections), the stronger their bargaining position inside the household.
Similar models were applied to intergenerational conflict: fathers use wealth
control to extract labor from children; children's outside opportunities affect how
much they comply.
Folbre notes that while these models are theoretically valuable, they are
empirically difficult to test, and they raise but don't fully answer the deeper
question: what determines bargaining power in the first place?
Sociological and Marxist-Feminist Approaches
Sociologists (like Blood and Wolfe) have long studied power dynamics within
marriage. Boserup (1970) argued that the more economically valued women's
work, the more power women have. Later studies found the relationship more
complicated.
Socialist feminists go further. Folbre herself argues that domestic labor can be
analyzed in terms of exploitation: if you measure the hours worked against the
share of resources consumed, you can calculate a "rate of exploitation" for each
family member. Women and children may be doing more labor than the value of
goods they receive — subsidizing wage-earning men, who in turn subsidize
capitalists.
This framework also challenges the claim that domestic labor benefits only
capitalists. It may also benefit husbands directly. And when children are raised
but grow up to contribute to Social Security systems that pay out based on wage
earnings rather than parenting work, mothers are "publicly exploited" — their
reproductive labor benefits society but goes uncompensated.

3.2 Empirical Studies

Inequalities Within the Family
Time budget studies and surveys from across the developing world document
systematic inequalities inside households:
 In Ghana, women's incomes averaged only one-quarter of men's.
 In Western Kenya, men benefited from women's and children's labor in
both subsistence and cash crop production but shared only a small fraction
of cash income.
 In Africa more broadly, women consistently work longer hours than men in
their households.
  Similar patterns were found in Bangladesh, Nepal, and the Philippines.
 Girls and women in many developing regions receive a smaller share of
their recommended daily calories and protein than men and boys.
 Studies from India showed dramatic differences in male and female
mortality rates, significantly linked to unequal food distribution.
 Increases in women's income had a far larger positive impact on children's
nutrition than equivalent increases in men's income — suggesting that
who controls income matters enormously.
Folbre cautions that while these inequalities are real, they don't automatically
prove that bargaining power explains them. They could theoretically reflect
"voluntary choices" or cultural norms. The empirical evidence specifically linking
bargaining power to outcomes is still thin at this point — though studies from
Nepal (Acharya and Bennett) and the Philippines (Folbre's own research) offer
supportive examples.
Impact of Inequalities
Inequalities within the household have spillover effects on major development
outcomes:
Labor supply: Women's low wage acceptance could reflect their weak
bargaining position at home — a household that places lower value on women's
leisure pushes women to accept lower wages. Men may even benefit from wage
discrimination against women in the labor market because it keeps women
economically dependent at home.
Migration: Standard migration analysis ignores how migration affects and is
affected by household distribution. A migrant man gains market income but loses
unpaid household services from wife and children — services that are cheap only
because those family members have little bargaining power. Urban migration
may actually improve women's and children's autonomy, which could be "costly"
for the male head. This may explain why some men resist migration even when
urban wages are substantially higher.
Remittances: The assumption that migrants will send substantial remittances
home is empirically weak. Studies show remittances are often small and irregular
unless migrants expect to return. This is consistent with intergenerational
bargaining dynamics: as development weakens the economic leverage of elder
generations, income flows from children to parents diminish.
Fertility: Women bear a disproportionate share of the costs of childrearing,
especially as development progresses and men shift more of the economic
burden to women. Female-headed households — women raising children without
reliable male financial support — are among the most economically
disadvantaged groups in many countries. This observation alone seriously
undermines the idea that the "stereotypical household" is a valid unit of analysis.

Section 4: Social Institutions and the State

4.1 Theoretical Issues
Patriarchy as a System
Marxist-feminist scholars argue that gender inequality isn't just a household-level
phenomenon — it is a system, called patriarchy, that operates through social
institutions, laws, culture, and the state. Just as capitalism is organized around
control over the means of production, patriarchy is organized around male
control over land, property, and women's and children's labor.
Patriarchy extends beyond property ownership. Even in propertyless households,
a sexual wage differential gives men better access to subsistence, reinforcing
their power over women regardless of specific household wealth.
Patriarchal ideology reinforces this: gender inequality is justified by religion and
biology, children are socialized into gender roles early, and this makes it difficult
for women to take advantage of even small gains in economic autonomy.
Patriarchy and capitalism are not simply opposed. They often complement each
other. Both systems benefit from keeping female labor cheap. However, they can
come into conflict over control of young adults' labor — wage work tends to
weaken the authority of elder generations over young people, which capitalism
encourages but traditional patriarchy resists.
Beneria and Sen argue that capitalist development has contradictory effects
on gender inequality: it mitigates some forms while intensifying others. Stacey
shows that state socialist regimes like China also maintained patriarchal family
structures when it suited them.
Caldwell argues that cultural modernization (changing values, media, education)
has become somewhat autonomous from economic development, meaning that
shifts in intergenerational authority can happen even without major economic
transformation.
Patriarchy as a Mode of Production
Some scholars go further, treating patriarchy as a distinct mode of production
governing human reproduction. Meillasoux and Rey analyze "domestic" or
"lineage" modes of production that lower the cost of reproducing labor power for
capitalism, and through which elder men control younger men and women.
However, both scholars stop short of calling intergenerational relations
"exploitation" and inadequately address gender.
Caldwell's "family mode of production" concept explains how the weakening of
traditional family structures tends to reduce desired fertility: once the younger
generation gains economic independence and women gain more control over
reproductive decisions, the traditional economic rationale for having many
children disappears.
Folbre criticizes Caldwell for being overly optimistic about capitalism "liberating"
women. The persistence of a large gender wage gap and growing female poverty
in advanced capitalist economies like the U.S. undercuts this. She also notes that
the low valuation of women's reproductive labor — across both high- and low-
fertility economies — reflects a complementarity between patriarchy and
capitalism that goes beyond wage suppression: employers and taxpayers
without children benefit when public support for parents and children is minimal,
and the standard economic assumption that childcare is not "real work" serves to
justify this.

4.2 Empirical Studies

Institutional Studies
Structural gender inequalities in access to wealth are documented across
regions:
 Women's land rights are severely restricted almost everywhere. In most of
Africa, land is controlled by men even where women do most of the
farming. Under Islamic law, women inherit less than male relatives. In
Nepal, daughters cannot inherit leaseholds that sons can.
 In at least two countries (Senegal and the Philippines), husbands had a
legal right to prevent wives from working outside the home. In
Zimbabwe, married men legally controlled their wives' earnings.
 "Protective" labor legislation that appears benign often reinforces
occupational segregation and keeps women in lower-paid jobs.
 Legal and religious restrictions on contraception and abortion limit
women's control over their own fertility in many countries.
 Marital rape is rarely a legal offense anywhere.
 Punitive laws against non-procreative sexual practices disproportionately
control women's sexuality.
International organizations (UN, World Bank, ILO) collect less data on women's
legal and reproductive rights than on their economic status — a gap that makes
it harder to document or challenge patriarchal systems.
Policy Implications
Development policy has systematically neglected women:
 Agricultural extension programs have largely ignored women farmers,
even though women do a large share of agricultural work.
 The introduction of lucrative cash crops tends to push women out of cash
crop production.
 U.S. Peace Corps programs significantly underrepresented women as
beneficiaries, particularly in programs involving skills and resource
transfers.
 The "basic needs" development approach was partially undermined by its
failure to account for women's household labor, which is the primary
source of basic needs satisfaction for families.
 Female-headed households, a growing share of households in developing
countries, have not been targeted for specific aid despite being among the
most economically vulnerable groups.
New research was beginning to highlight:
 The underestimated benefits of prenatal care and comprehensive family
planning programs (far more effective than contraceptives alone).
 The possibility that fertility rates may actually be rising in some countries
(like Kenya) because women's economic opportunities remain too
restricted to raise the opportunity cost of childrearing.
 The political dimensions of patriarchal systems — Islamic fundamentalism
in Khomeini's Iran was partly driven by a desire to reassert control over
women and children; agrarian reforms in Ethiopia and Nicaragua were
beginning to extend formal rights to single women.

Conclusion
All the literature reviewed leads Folbre to a clear conclusion: the traditional
distinction between "moral economy" (the household) and "political economy"
(the market) does not hold up. Economic self-interest penetrates family life. This
does not mean households can be reduced to pure economic calculation — but it
does mean that the boundary between self-interest and altruism does not sit
neatly at the front door.
The paper calls for:
 Treating household production as economic activity and measuring it
properly.
 Abandoning the "black box" view of the household in favor of analysis that
looks at conflict, inequality, and bargaining within families.
 Situating microeconomic analysis of households within a broader
structural analysis of patriarchy, class, and capitalism — because
bargaining power inside households is shaped by laws, labor markets,
property rights, and social norms that operate far beyond any individual
family.
Summary of "The U-Shaped Female Labor Force Function in Economic
Development and Economic History" by Claudia Goldin (1994)
What is this paper about?
Goldin argues that women's participation in the paid labor force does not simply
rise or fall as countries get richer — it traces a U-shape. Women work a lot at
very low levels of development, then work less as incomes first rise, and then
work more again as development continues further. The paper explains why this
pattern exists, what drives the turn back upward, and uses both data from over
100 countries and U.S. history to support the argument.

The Core Idea: The U-Shape

When countries are very poor, women work heavily — on family farms, in home
workshops, and as own-account traders. Much of this work is unpaid but it is
economically real and gets counted in labor force statistics (to varying degrees).
As incomes rise with early economic development, women's labor force
participation tends to fall. Production shifts from households and family farms to
factories and firms. Families get richer, and the income effect kicks in — when a
husband earns more, the family can "afford" to have the wife stop working
outside the home. Additionally, factory and manual labor in agriculture, industry,
and construction carry a strong social stigma for married women in many
societies, so even if wages rise enough to attract women to factory work in
theory, many families refuse to accept it.
Eventually, at higher levels of development, women's participation rises again.
This happens when women gain access to education — particularly at the
secondary school level — and when this education opens up white-collar office
and clerical jobs for them. There is little or no social stigma against married
women working in offices, so the barrier that kept them out of factory work does
not apply. The substitution effect (working pays more, so work more) comes to
dominate the income effect (we're rich enough, so don't need to work), and
participation climbs.
This U-shape can be seen both across countries at a single point in time (richer
countries tend to be higher on the right side of the U) and within individual
countries over their economic histories.

Section 1: The U-Shape in Cross-Section and Theory

The Data
Goldin plots women's labor force participation rates (for women aged 45–59,
around 1980) against per capita GDP and against average years of male
schooling for over 100 countries. Both graphs show a clear U-shape. The pattern
is driven mostly by regional differences rather than variation within regions:
 Sub-Saharan Africa (East, Middle, West) sits at the high-participation left
end of the U — very poor, but women work a lot on farms and in
household production.
 North Africa and South Asia sit lower down the falling portion, as do parts
of Southeast Asia.
 Latin America sits near or at the bottom of the U.
 Southern and Eastern Europe begin the rising portion, with Western
Europe, Northern Europe, North America, and parts of the Pacific at the top
right.
An important technical note: the U-shape in the aggregate is driven by changes
in the sectoral composition of employment. Unpaid family work and self-
employment decrease with development, while wage employment stays flat until
quite late in development when it rises sharply. Within any single sector, the U
cannot be observed.
Why Does Participation First Fall?
Several forces push women out of the labor force during early development:
The income effect. As family income rises (initially because other family
members, usually men, get factory wages), women's labor supply falls. Richer
families don't need women's earnings as badly.
Structural change in production. As production moves from the family farm
and home workshop to the factory, the type of work available to women
changes. Home-based production disappears, but factory work is stigmatized for
married women.
Technology changes in agriculture. New agricultural technologies (like those
associated with the Green Revolution) sometimes reduce demand for female-
specific tasks (like weeding) while increasing demand for male-operated
machinery, effectively displacing women from farm work.
Education lags for women. Early in development, education increases mainly
for men. Women's relative productivity in the market therefore doesn't rise as
fast as family income does, reinforcing the downward pressure on their
participation.
The Stigma Model
Goldin's most important theoretical contribution is a formal model showing that a
social stigma against married women doing manual wage labor is central to the
U-shape. The model works as follows:
There is one good (G) that can be produced either by the family at home or by a
firm (a factory). Women have three uses of their time: home production of G, firm
production of G, and childcare. A woman's utility depends on goods and childcare
— and is reduced by a stigma cost (S) if she works in manual/blue-collar wage
employment.
The stigma is "all or none" — it applies the moment the wife takes a factory job,
regardless of how many hours she works. This means whether the stigma is
binding or not depends solely on whether the utility gain from taking the factory
job exceeds the utility cost of the stigma (S).
Three periods of development are modeled:
Period 1: No firms. Women work at home production. This is the high-
participation left end of the U.
Period 2: A firm appears offering factory wages. Family income rises because
men work for the firm. If there were no stigma, a woman might choose to work
for the firm (the substitution effect) while also producing less at home because
the family is richer (the income effect). The net result for labor force participation
is ambiguous. But with the stigma, many families will choose to stay at home
rather than face social shame, even as women's wages rise. This produces the
falling portion of the U.
The probability that the stigma binds increases with family income — richer
families have more to lose from the perceived shame, and lower-paid factory jobs
offer less utility gain. So as development raises incomes, more and more families
end up in the stigma equilibrium, pulling women out of market work.
Period 3: Women's wages rise because their education increases and they can
now access white-collar jobs in the firm. The stigma does not apply to office
work. The substitution effect now operates fully, and women's participation rises
— tracing out the rising portion of the U.
An important implication: because the stigma suppresses women's labor supply
in period 2, an economist measuring the wage effect on labor supply will
underestimate it during the falling portion and then see an apparently very large
wage effect when women finally enter white-collar work. The measured wage
effect is biased by the fettered equilibrium.
Why No Stigma for White-Collar Work?
Goldin offers an explanation: when a married woman takes a factory job, it
clearly signals that her husband is a poor provider — he is "lazy or negligent."
The stigma enforces the social norm that men must support their families. But
when a woman takes an office job, the signal is ambiguous. She could be an
educated woman who happens to want to work, married to a successful man —
there is no clear implication of male failure. The norm therefore cannot take hold
for white-collar work.
This means white-collar work "unlocks" the substitution effect while bypassing
the stigma constraint. And importantly, access to white-collar work requires
secondary education.
Evidence on the Rising Portion
Goldin presents several regression results to support the mechanism.
First, she shows that the ratio of female to male secondary school enrollment
converges toward equality (and often reverses, with girls exceeding boys) as
male adult education increases. Richer countries — where men are more
educated — tend also to have greater gender equality in secondary schooling.
Second, white-collar employment as a share of all employment rises strongly
with GDP per capita. As countries develop, the service and white-collar sectors
expand.
Third — and most telling — the ratio of female to male clerical workers rises
strongly with female secondary education. Countries where women are much
less educated than men have very few women in clerical work. As the education
gap narrows, women's representation in clerical work rises rapidly. A regression
confirms that the interaction between women's clerical employment and their
education levels is the key driver of older married women's labor force
participation: neither female education alone nor clerical employment alone is
sufficient — both are needed together.

Section 2: The United States as a Historical Case

The U in U.S. History
The U-shape is not just a cross-sectional pattern — it has been traced in the
historical record of developed countries. In the U.S., however, this was hard to
see for a long time because early censuses used the "gainful employment"
concept (having an occupation), and most women did not report occupations
even when they were genuinely productive.
In 1890, fewer than 3% of all married white women reported an occupation to
the census. But archival research shows that many more were working — on
family farms, in boardinghouses, in home manufacturing. An adjusted estimate
puts the true participation rate for married white women at close to 15% in 1890,
possibly higher earlier in the 19th century when women actively participated in
their husbands' home-based trades and farm production. By 1940, the
participation rate was 12.5%.
This suggests that participation had indeed been falling through the late 19th
and early 20th centuries before its long rise. Goldin believes the bottom of the U
was reached sometime in the 1920s.
The data from Table 1 of the paper illustrate the rise clearly: from 12.5% in 1940,
married white women's participation climbed to 20.7% by 1950, 29.8% by 1960,
38.5% by 1970, and 49.3% by 1980.
The Role of Secondary Schooling
The turning point in the U — the beginning of the sustained rise — coincides with
a dramatic expansion of secondary schooling in the United States. In the non-
South, high school graduation rates rose from under 10% of 17-year-olds in 1910
to about 50% by 1937. Nearly all of this growth happened in just seventeen
years, between 1920 and 1937.
Crucially, girls were more likely to graduate high school than boys throughout
this period — 1.5 times more likely in the 1910s and 1.25 times more likely in the
1920s. This gender difference in educational attainment is historically unusual
and helps explain why the U.S. saw such an early and sharp rise in women's
labor force participation compared to other countries.
Why did girls go to school more than boys? Because secondary school had a
higher private return for girls. The alternative for a girl who didn't complete
school was factory or domestic service work, which was poorly paid and
stigmatized. Secondary school opened the door to clerical jobs — typist,
stenographer, secretary — which were cleaner, more respectable, shorter in
hours, and paid about 30% more than manufacturing by the 1920s. For boys, the
alternatives included skilled trades that didn't require a diploma, so the relative
return to schooling was lower for them.
Office work became "the ambition of the school girl," as vocational counselors of
the era observed. A few years of high school transformed a young woman's
occupational trajectory from factory drudgery to office respectability.
The Feminization of Clerical Work
The rise of secondary schooling drove a rapid feminization of the clerical sector.
In 1890, only about 15% of all office workers were female. By 1910, 36% were
female. By 1920, 48% were female — office work had become a predominantly
female domain within just thirty years. By 1950, 62% of all clerical workers were
women.
But these women were initially unmarried. In 1890, only 4% of female clerical
workers were married. Many young women went to school, entered office work,
married, and then left the labor force — as social norms and, increasingly,
explicit employer policies (marriage bars) required.
Marriage bars — official firm policies of not hiring married women and firing
women who married — were already present in some firms in the 1920s and
became widespread during the Great Depression, when employers argued that
married women were "taking jobs" from men who needed to support families.
These bars delayed the full impact of the clerical revolution on married women's
labor force participation.
When marriage bars dissolved after World War II — partly due to labor shortages,
changing attitudes, and eventually legal pressure — the women whose cohorts
had entered clerical work in the 1920s and 1930s were now in their 40s and
could return to the labor force. The explosion in married women's labor force
participation in the 1940s and 1950s was primarily among women over 40 (Table
1 shows this clearly — the 45–54 age group went from 10.1% in 1940 to 22.2% in
1950 to 38.6% in 1960).
Goldin also presents evidence from a 1939 dataset showing that women who had
started their careers in manufacturing were 10 percentage points less likely to be
in the labor force later in life than women who had started in clerical, sales, or
professional work. This supports the argument that early white-collar work
created a lasting attachment to the labor force.
The Peak Substitution Effect
Mincer's famous 1962 study of female labor supply found a strong positive
substitution effect and concluded it dominated the income effect, which
explained why women entered the labor force despite rising family incomes.
Goldin argues that Mincer may have been measuring at the historic peak of the
substitution effect in the U.S. — the 1950s and 1960s — when the difference
between the wage elasticity and the income elasticity was at its highest. Cross-
sectional studies from 1900 to the 1960s, taken together, show the substitution
effect rising over time and the income effect declining — consistent with the U-
shaped story.

Section 3: Global Comparisons and Predictions

Latin America
Latin America is unique in the developing world: girls are educated at secondary
school to the same or greater degree than boys, mirroring what the United States
looked like in the early 20th century. And indeed, Latin American countries in the
early 1990s were sitting at or near the bottom of the U-shaped function —
female labor force participation rates were in the 20–40% range. The percentage
of female workers in clerical employment was in the 10–20% range and the
percentage of all clerical workers who were women was 30–40%, similar to the
U.S. in the 1920s.
Goldin predicts that Latin America was ripe for a transition to the rising portion of
the U and that married women's labor force participation would rise substantially
in the coming decade, paralleling what happened in the U.S. after World War II.
Asia
The Asian newly industrialized countries presented a mixed picture. Singapore
and Hong Kong had seen rapid educational advances and women were quickly
moving into white-collar occupations — the female share of clerical workers was
51% and 63%, respectively. Goldin predicted their female labor force
participation would rise strongly.
Korea and the Philippines, by contrast, had high education levels but women
were poorly represented in clerical work (only 33% and 42% of clerical workers
were women, respectively), suggesting slower future growth in female
participation despite educational gains. The key distinguishing variable was not
education alone, but the combination of education and access to white-collar
employment.

Conclusion
Goldin summarizes the paper's argument: women's labor force participation is U-
shaped over the course of economic development. The downward slope reflects
a combination of a strong initial income effect (as family incomes rise) and a
weak substitution effect, compounded by the social stigma against married
women doing manual wage labor and the structural shift of production from the
household to the factory.
The upward slope begins when women's education — particularly at the
secondary level — improves relative to men's, when the white-collar sector
expands to absorb educated women, and when white-collar work carries no
stigma. Access to education raises women's wages in non-stigmatized
occupations, strengthening the substitution effect. At the same time, the income
effect weakens as fewer families face the stigma constraint. The substitution
effect begins to dominate, and the upward portion of the U is traced out.
The deeper implication is that economic development does eventually promote
gender equality, but not smoothly or immediately. The falling portion of the U
can represent real deterioration in women's economic position and autonomy.
The key to the turn is education — particularly secondary education for girls —
combined with a structural transformation of the economy toward white-collar
service work. These are the conditions that allow women to escape the stigma
trap and fully enter the paid labor force.

Summary: "The Roots of Gender Inequality in Developing Countries" by

Seema Jayachandran (2015)
What is this paper about?
This paper asks a fundamental question: why is gender inequality so much
worse in poor countries? Is it simply because these countries are poor — and
inequality will fade as they grow richer? Or do the societies that happen to be
poor today also have specific cultural beliefs and practices that make gender
inequality worse, independent of poverty itself?
Jayachandran argues the answer is both — economic underdevelopment causes
much of the gender gap, but culture plays an independent role too, especially in
explaining things like strong son preference in India and China, and low female
employment in India, the Middle East, and North Africa.

Part 1: The Basic Facts — How Bad is Gender Inequality in Poor

Countries?
Jayachandran begins by documenting several areas where gender gaps are
larger in poorer countries.
Education
The ratio of male to female college enrollment falls sharply as GDP rises — in
fact, in rich countries, women now outnumber men in college. The relationship is
strong (R² = 0.44). The same pattern holds for primary and secondary school
enrollment.
Health
Women naturally live longer than men, but this female advantage in life
expectancy is smaller in poor countries. Even for the same diseases, women
have higher relative mortality in poor countries. Sub-Saharan Africa is a notable
outlier because HIV/AIDS has disproportionately cut female life expectancy there.
Employment
Here the pattern is surprising: female labor force participation (FLFP) does
not rise with income at the country level — the correlation is essentially zero.
However, attitudes about women working are more progressive in richer
countries. India stands out — men there are three times as likely as women to be
working. The Middle East and North Africa also have abnormally low FLFP.
Gender-Based Violence
Tolerance for domestic violence (measured by whether women say wife-beating
is sometimes justified) is much higher in poor countries — ranging from under
1% to over 85% across countries, but systematically higher where GDP is lower.
Decision-Making Power in Marriage
In poor countries, women are less likely to have a say in major household
spending decisions. This pattern holds both across countries and within countries
— wealthier women within the same country have more decision-making power.
Freedom of Choice and Life Satisfaction
Women in poor countries report having significantly less control over their lives
compared to men. This is especially pronounced in India, the Middle East, and
North Africa — the same places with very low female employment. Women's life
satisfaction relative to men's also tends to be lower in poorer countries.

Part 2: How Does Economic Development Reduce Gender Inequality?

Jayachandran identifies three main mechanisms through which the process of
development itself — independent of culture — should narrow gender gaps.
A. Shift Away from Brawn-Based Work
As economies grow, they shift from agriculture and manufacturing (which require
physical strength) toward services (which require mental and social skills). Men
have a physical comparative advantage in brawn-based tasks, so women's
relative productivity is lower in agricultural economies. As services grow,
women's relative earning power increases.
Key evidence:
 A study of China's agricultural reforms (Qian 2008) found that in regions
where tea-growing became more profitable (women have a comparative
advantage in picking tea leaves), there were fewer "missing girls" —
families had less reason to abort female fetuses or neglect infant
daughters, because women's income rose and they gained more
bargaining power in the household.
 Research in India (Carranza 2014) showed that in areas with soil suited for
deep tillage (which favors male labor), women's labor force participation
was lower and the sex ratio was more male-skewed.
 Historical research (Alesina et al. 2013) showed that regions that used the
plow (requiring male strength) in preindustrial times have less female
employment and more traditional gender attitudes today — meaning old
economic structures left lasting cultural imprints.
The paper also notes that FLFP follows a U-shaped pattern with development
(Goldin 1995): at very low development levels, women work on family farms; as
factory work grows, women withdraw (because factory work is seen as beneath
respectable women, and higher male wages mean families can afford not to send
women to work); and at even higher development levels, women re-enter the
workforce as service sector and white-collar jobs grow. Evidence from India
shows that the arrival of call center (BPO) jobs — which were seen as respectable
for women — led many women who would not otherwise have worked to take
jobs, raised girls' aspirations, delayed marriage, and increased school enrollment
among younger girls.
B. Technological Progress in Home Production
Women do the majority of housework in most societies. When development
brings better infrastructure — running water, electricity, household appliances —
housework takes less time, freeing women to work outside the home. This is a
major driver of increased FLFP.
Evidence:
 In the US, time women spent on home production fell from 47 hours per
week in 1900 to 29 hours in 2005.
 Electrification in post-apartheid South Africa (Dinkelman 2011) increased
FLFP, largely because women could switch from wood-burning to electric
stoves.
 Access to household appliances in the US between 1960–1970 was
associated with more women working.
 A program providing village water taps in Kyrgyzstan saved households 3
hours per day on water collection.
C. Lower and Safer Fertility
Richer countries have fewer children and safer pregnancies. Both help women.
Evidence:
 When Colombia rolled out a large family planning program in the 1960s–
70s, women who gained access to contraception delayed having children,
got more education, and worked more. Similarly, access to the birth
control pill in the US opened up long-duration careers in medicine and law
to women.
 In Sri Lanka in the 1940s–50s, a rapid decline in maternal mortality (due to
medical progress) led to meaningful increases in female life expectancy.
Because girls could now expect to live longer, parents had more reason to
invest in their education — and girls' schooling did rise, explaining about
one-third of the closing of the education gender gap.
 Medical advances also reduced post-pregnancy complications, allowing
more married women to return to work.

Part 3: Cultural Factors That Cause Gender Inequality Beyond Poverty

Even after accounting for development-related mechanisms, certain societies
have cultural practices that worsen gender inequality. Many of today's poor
countries share these norms, and poverty often makes them worse — but the
norms are not simply a product of poverty. Jayachandran identifies six key
cultural factors.
A. Patrilocality
In patrilocal societies, a married woman leaves her birth family and moves in
with her husband's family. This means:
 Parents invest in daughters less because they "lose" their daughter to
another family at marriage — captured in the Indian saying that "raising a
daughter is like watering your neighbors' garden" (and the Chinese
equivalent: "plowing someone else's field").
  Coresidence with adult sons is much more common in Asia, the Middle
East, and North Africa than elsewhere, which is strongly correlated with a
more male-skewed sex ratio.
 Within India, the north has a stronger patrilocal system than the south,
explaining why gender inequality is worse in the north.
Evidence: In one Indian study, 405 parents were told their child needed heart
surgery — 70% of boys received the surgery, but only 44% of girls did.
Poverty amplifies this: if resources are scarce, families prioritize investing in the
child (the son) who will remain with them.
B. Sons as Old-Age Support
Closely linked to patrilocality, sons traditionally support elderly parents in China
and India. This creates a strong economic reason to want sons. Evidence from
China (Ebenstein & Leung 2010) shows that when the government introduced a
rural pension program, households without sons were most likely to enroll — and
areas with pension access had a less male-skewed sex ratio. This suggests the
pension substituted for sons as old-age support, reducing son preference.
Importantly, the cultural norm didn't change — but its economic consequences
did.
C. The Dowry System
Dowry is a payment that the bride's parents make to the couple at marriage.
Unlike the historical European dowry (which the bride herself controlled as an
inheritance), today's dowry in South Asia is effectively a "price of a groom" — the
groom's family controls it.
This makes daughters a financial liability to parents:
 Parents cite dowry as a key reason for preferring sons.
 A billboard in Amritsar, when ultrasound was newly available, urged
parents to "invest Rs. 500 now, save Rs. 50,000 later" — meaning spend
on a sex-determination test and abort a girl to avoid paying dowry later.
 Dowry has actually been rising in real value in India and spreading in
Bangladesh, against the international trend.
 In extreme cases, dowry disputes lead to violence against new brides,
including so-called "dowry deaths."
D. Patrilineality
In patrilineal systems, property and family names pass through male
descendants. This disadvantages women in several ways:
 Indian widows traditionally do not inherit their husbands' ancestral land,
making them dependent on their sons to maintain their standard of living
— another reason women themselves often want sons.
 When four Indian states amended the Hindu Succession Act in the 1980s–
90s to give daughters equal inheritance rights, girls' schooling rose,
women married later (reflecting more financial independence), and their
 mothers were more empowered in the household. However, the reforms
also led to a rise in suicides, possibly because increased female bargaining
power triggered marital conflict.
E. Religious Rituals Requiring Sons
In both Confucianism (China, Vietnam) and Hinduism (India), sons play
irreplaceable roles in key religious rituals. In Hinduism, a son must light the
funeral pyre and secure the deceased's salvation. This creates a particularly
strong desire for at least one son.
Evidence: Jayachandran (2014) finds that parents in India strongly want to have
at least one son, but once they have one son, they prefer a roughly balanced
gender ratio among their children. The religious logic creates a floor desire for
sons rather than an unlimited preference for boys.
Upper-caste Hindus follow these norms more strictly, and accordingly show a
more skewed sex ratio than lower castes.
F. Concern for Female "Purity"
Many developing countries — especially in South Asia, the Middle East, and North
Africa — restrict women's physical mobility in the name of protecting their safety
and "purity." These restrictions directly reduce women's ability to work and go to
school.
Evidence:
 Distance to school matters far more for girls than boys. Building a village
school in Afghanistan essentially closed the gender enrollment gap.
 Giving bicycles to girls in India (Muralidharan & Prakash 2013) had a large
effect on girls' school attendance.
 Parents want single-sex schools and sex-segregated latrines — providing
these increases girls' enrollment.
 Girls benefit disproportionately from having female teachers, but the
supply of female teachers falls at higher grades, which can explain 10–
20% of the decline in girls' test scores as they advance in school.
 Pressure to marry daughters off early (to preserve chastity) keeps girls out
of school — in Bangladesh, delaying a girl's marriage by one year adds
0.22 years of schooling on average.
 Among Hindus, upper-caste women face stricter seclusion because caste
purity norms apply more strictly to them; lower-caste women typically
have more professional freedom.
 Muslim women similarly show low FLFP due to purdah norms, though
Muslim families in India show less son preference (smaller dowry norms,
inheritance rules that don't entirely exclude daughters) and a less skewed
sex ratio than Hindus.
G. Cultural Norms Persist Even When Economics Change
One powerful piece of evidence that culture matters independently of
economics: gender norms persist even when people immigrate to rich countries.
East Asian and South Asian immigrants in the US still show sex-selective
behavior (a male-skewed ratio of last births), even though many cultural
practices like dowry and nonemployment of women are abandoned after
immigration. The continued son preference suggests it is embedded in deep
preferences, not just current economic incentives.
Additional evidence: a woman's fertility and labor force participation are
predicted by the average rates in her country of origin, even after she moves to
the US. And if a mother works, her son's wife is more likely to work — showing
norms pass from generation to generation.

Part 4: The Special Problem of the Sex Ratio at Birth

This section gets special attention because the sex ratio at birth is a form of
gender inequality that is getting worse with development rather than better —
unlike education gaps, employment attitudes, or decision-making power.
The Scale of the Problem
The natural sex ratio at birth is 103–106 boys per 100 girls. But in China, it
reached 116:100 by 2012, and in India 111:100. These are enormous and
growing imbalances.
Two Distinct Dimensions of Son Preference
Jayachandran distinguishes between:
1. Wanting more sons (preferring to have sons rather than daughters)
2. Investing more in sons (giving sons more food, healthcare, education)
These often go together but are not the same. Notably, India and China are
extreme outliers in the sex ratio (wanting sons) but not particularly exceptional
in education gender gaps (investing in sons) compared to countries at similar
income levels. The sex ratio imbalance has much stronger cultural/geographic
patterning (concentrated in Asia) while education gender gaps have less
geographic clustering.
Also, today's rich countries historically showed other forms of gender bias (lower
female education, lower FLFP) but did not show strong son preference in the
same way. Historical US data shows no male-skewed "sex ratio of last births" (a
measure of son-biased fertility stopping behavior) — demonstrating that today's
developing countries' strong son desire is not simply a function of being at an
early stage of development.
Why is the Sex Ratio Getting Worse Despite Development?
Son preference (the desire for sons) actually falls with development. But the
actual sex ratio at birth is worsening with development. Why? Two reasons:
1. Ultrasound technology: Sex-selective abortion has become possible and
increasingly accessible. About half the increase in China's sex imbalance is
estimated to be explained by ultrasound access. This is a case where technology
interacts with existing cultural son preference to produce a worse outcome.
2. Declining desired fertility: When people want fewer children (3–4 vs. 1–2),
they face a higher risk of having no sons. A couple wanting 4 children is unlikely
to end up with zero sons by chance; a couple wanting 1–2 children faces that risk
more often. As female wages rise and opportunity costs of children increase,
desired family size falls, making sex-selective abortions more likely even without
any increase in son preference itself.
The conceptual upshot: the sex ratio at birth measures not just son preference,
but son preference combined with desired family size. Development reduces
both desired family size and son preference — but the family-size effect
dominates in the short run, worsening the sex ratio even as attitudes improve.

Part 5: Policy Approaches

Because cultural norms may persist even as economies grow, waiting for
development alone to close gender gaps is insufficient. Jayachandran outlines
several policy approaches.
Legal Rights for Women
Reserving political seats for women in India led female leaders to implement
policies better reflecting women's preferences, began changing attitudes toward
women as leaders, and raised girls' aspirations and long-term investments.
However, enforcement of legal reforms is often weak — land inheritance reforms,
bans on prenatal sex determination, bans on dowry, and bans on child marriage
are all frequently unenforced.
Financial Incentives for Having or Investing in Girls
Many Indian states offer financial incentives for having daughters. Conditional
cash transfer programs like Mexico's Progresa/Oportunidades give larger
subsidies for educating girls than boys, responding to girls' higher dropout rates.
Shifting Financial Resources to Mothers
Giving mothers more control over income improves girls' outcomes — women
tend to have less pro-boy bias than men in how they spend money. However, an
important caveat: the gender gap in attitudes is often surprisingly small. In India,
37% of women and 33% of men tolerated wife-beating in DHS surveys; 20% of
women and 19% of men wanted more sons than daughters. Women often hold
these attitudes because of practical concerns — they gain household status from
having a son — and because they lack role models showing that gender equality
is possible.
Changing Attitudes Directly
Creating female role models can shift women's (and men's) beliefs about what is
possible. Television in Brazil and India appears to have changed attitudes on
family size and gender relations, even without that being its explicit goal.
Mothers' gender attitudes are more influential than fathers' in shaping children's
beliefs, but fathers often have more decision-making power over girls' lives.
Conclusion
Jayachandran's main conclusions:
1. Gender inequality is systematically worse in poor countries across
education, health, employment, decision-making, and freedom of choice.
2. Economic development reduces gender gaps through three main
channels: the shift to service-sector work, time-saving technology in the
home, and lower, safer fertility — all of which increase female labor force
participation, which in turn leads to more investment in girls and greater
female autonomy.
3. But culture independently causes gender inequality, especially through
patrilocality, son-based old-age support, the dowry system, patrilineality,
religious rituals requiring sons, and purity norms. These explain why
India's FLFP is so low and why India and China have such extreme sex
ratios.
4. The skewed sex ratio at birth is a unique and worsening problem — driven
by the combination of cultural son preference, new technology
(ultrasound), and declining desired fertility. It is getting worse even as
economies grow and son preference falls.
5. Policy can and should accelerate the narrowing of gender gaps, rather
than waiting for growth alone — because some cultural norms are deeply
embedded and slow to change even when economic circumstances
improve.

Summary: "Economics and Identity" by Akerlof and Kranton

(2000)
What is the Paper About?
This paper argues that who you think you are — your identity — affects
your economic decisions. Standard economics assumes people make
choices based on money, prices, and preferences. Akerlof and Kranton say
that's incomplete. People also make choices to protect or affirm their
sense of self. They build a formal economic model around this idea and
apply it to three real-world areas: gender in the workplace, poverty, and
housework.

The Core Idea: Identity in a Utility Function

The authors propose that a person's utility (happiness/satisfaction)
depends not just on what they consume or earn, but also on their identity
— how well their actions match the expected behavior for their social
group.
 They formalize this as:
U = U(own actions, others' actions, identity)
And identity itself depends on:
 Which social category you belong to (e.g., man/woman, Black/white,
rich/poor)
 Your personal characteristics and how well they match the "ideal" of
your category
 How well your actions (and others' actions around you) match the
prescriptions (expected behaviors) for your category
If you follow your group's prescriptions, your identity is affirmed — you
feel good. If you violate them, you feel anxiety and discomfort. If someone
else in your group violates them, you may also feel disturbed — and even
retaliate.

Four Key Insights About Identity

1. Identity changes what YOU get from your own actions. A woman
working in a "man's job" may feel internal discomfort, not because the job
is bad, but because it conflicts with her gender identity. This affects her
willingness to do that job.
2. Identity changes what you get from OTHERS' actions. A man may
feel threatened when a woman does "his" job — his masculinity feels
diminished. This can cause him to harass or undermine her, even at a cost
to himself.
3. Third parties can change identity and its prescriptions.
Advertisers, political leaders, schools, and social movements can reshape
what a social category means and what behavior is expected of it. This
changes how people behave economically.
4. People can (sometimes) choose their identity. Some people have
flexibility — a woman can choose to be a "career woman" or a
"housewife." But this choice is often constrained. A Black person in a racist
society, for example, may never be fully accepted as part of the dominant
group no matter what they do.

Evidence and Examples Used

The authors offer a rich set of real-world examples:
 Body modification (tattoos, piercings, circumcision): People alter their
bodies to signal group membership, even at personal cost.
 Female lawyers: Studies show female lawyers feel deep conflict — being
a good lawyer meant "acting like a man," which conflicted with their
female identity.
 Alumni giving: People donate to their own colleges, not the most efficient
charity — driven by identity and belonging.
 Mountaineering: People endure extreme danger and discomfort to
enhance their sense of self.
 Workplace harassment: Surveys show people widely believe men
harass women coworkers because the women's presence makes them feel
"less masculine."
 Insult experiments (Nisbett & Cohen): Men from the U.S. South
responded much more aggressively to insults than Northerners — showing
how regional identity ("culture of honor") shapes reactions.
 "Acting white": Black students who study hard or speak Standard English
are sometimes taunted by peers — because it signals abandonment of
group identity.

The Prototype Model (Green vs. Red)

To formalize their theory, the authors build a simple game-theory model
with two groups: Greens and Reds.
 Everyone starts as Green. Greens are supposed to do Activity One.
 If someone does Activity Two (the "wrong" thing for a Green), they lose
identity (suffer a utility penalty I_s).
 If Person One sees Person Two do Activity Two, Person One also suffers
(loss I_o) because it threatens their own identity.
 Person One can respond (punish Person Two) at a cost c, restoring their
own identity.
This simple setup generates several interesting outcomes:
 Sometimes Person One deters Person Two from misbehaving.
 Sometimes deterrence fails but Person One responds anyway.
 Sometimes Person One doesn't bother responding.
 Sometimes Person Two doesn't deviate at all, because their own
internalized identity is strong enough.
Policy insight from this model: A tax on responses (e.g., anti-
harassment laws) can change equilibria — making it harder for Person One
to punish Person Two. This may benefit Person Two but hurt Person One.
There's an inherent conflict between protecting nonconforming individuals
and protecting those who feel threatened by nonconformity.

Application 1: Gender in the Workplace

The authors use their model to explain occupational segregation —
why men and women tend to cluster in different jobs.
 As recently as 1970, two-thirds of U.S. workers would have had to switch
jobs for full gender parity in occupations.
 Traditional economic models explain this through women's "lower
attachment" to the labor force (they expect to quit and become mothers).
But the evidence for this is weak.
The identity model says: Jobs have gender auras. Some jobs are seen
as "men's jobs," others as "women's jobs." When a woman enters a man's
job:
 She suffers an identity loss (feels unfeminine).
 Male coworkers suffer an identity loss (feel less masculine).
 Male coworkers may then harass her, reducing productivity.
Firms avoid this by maintaining separate job categories — keeping "men's
jobs" and "women's jobs" distinct. Changing these associations requires
society-wide shifts, not just individual firm decisions.
The Women's Movement is explained as a change in prescriptions — it
reduced the identity loss women felt from entering male jobs and reduced
the identity threat men felt from women entering. This directly explains
the large occupational shifts from 1970–1990.
Sex discrimination law is also reinterpreted. Anti-harassment laws (like
Title VII) are about protecting people who violate gender-job associations
— exactly what the model predicts will generate hostility.

Application 2: Poverty and Social Exclusion

Here the Greens represent the dominant culture and Reds represent
marginalized/excluded groups (based on race, class, ethnicity).
The key insight: If a person from a marginalized group tries to identify with
the dominant Green culture, they face rejection (they can never fully be
"Green"). This rejection has a cost, called r in the model.
If r is high enough, some people will rationally choose a Red
(oppositional) identity — rejecting mainstream values like education,
employment, and law-abidingness. This isn't irrationality; it's a rational
response to exclusion.
This explains behaviors like:
 Joining gangs
 Dropping out of school
 Drug use and crime
These look self-destructive from the outside, but they make sense as
identity affirmation within the Red group's prescriptions.
Three equilibria emerge:
1. All-Green: Everyone integrates and works — possible only if social
exclusion is low.
2. Mixed: Some people integrate (Green), others reject mainstream culture
(Red).
3. All-Red: Everyone adopts an oppositional identity, though some may still
work.
Policy implications:
 Job Corps works (residential training programs) partly because removing
people from their neighborhoods cuts off Red identity peer pressure and
reduces the cost of being Green.
 JOBSTART failed — it kept all the features of Job Corps except housing,
and produced almost no improvement. This supports the identity-based
explanation.
 School reforms like Central Park East Secondary School (East Harlem)
may succeed by separating Green and Red students, or by reducing the
identity cost of "acting mainstream" for Red-identified students.
 Affirmative action debate: Rhetoric matters. Portraying Black
Americans as victims may increase r and push more people toward Red
identities. But eliminating affirmative action sends a signal of exclusion
that also increases r. The model shows these symbollic effects can be
large, far exceeding the direct effects of the policy itself.

Application 3: Household Division of Labor

The authors present data from the Panel Study of Income Dynamics
(PSID) covering 1983–1992, showing something standard economic
models cannot explain:
 When a husband does all the paid work outside, he does about 10% of
housework.
 Even as his share of outside work falls and his wife works more, his
housework share rises only to about 37% — never reaching 50%.
 This asymmetry holds even when the wife earns more than her husband.
Standard models (based on comparative advantage) predict a symmetric
relationship — whoever works more outside should do less inside,
regardless of gender. The data shows this only holds on one side.
The identity model explains it: Prescriptions say men shouldn't do
"women's work" at home, and men should earn more than their
wives. When a wife earns more, the husband loses identity — and the
equilibrium is restored not by him doing more at home, but by her doing
more (even though she's already working outside).
Hochschild's ethnographic study The Second Shift provides vivid
illustrations — including one couple where the husband "took care of the
 lower half of the house" (basement and tools) while his full-time-employed
wife handled everything else plus the child.

Conclusion
The paper argues that incorporating identity into economics:
1. Explains behaviors that look irrational but are internally consistent (self-
destructive behavior, staying in bad jobs, not voting with one's economic
interest).
2. Creates a new type of externality — other people's choices threaten your
identity.
3. Reveals new levers for policy — changing social categories and
prescriptions can change economic behavior, sometimes more powerfully
than changing prices or wages.
4. Makes identity choice itself an important economic decision — and limits
on that choice a key driver of inequality.
The authors suggest future work could apply this framework to political
economy, organizational behavior, language, violence, education,
consumption, retirement, and labor relations. They also call for
comparative studies across countries and time periods to understand how
identity-based preferences form and change.

Why This Paper Matters

Before this paper, mainstream economics largely treated preferences as
fixed and ignored the social context of decision-making. Akerlof and
Kranton showed that who people think they are — and who society
says they are — is itself an economic variable, one that shapes labor
markets, poverty traps, household dynamics, and discrimination in ways
that previous models simply missed.

Economics and Identity — George Akerlof & Rachel Kranton (2000)

Notes
The paper argues that traditional economics is incomplete because it
ignores identity, meaning a person’s sense of self and social belonging.
People do not act only for income, prices, or material rewards; they also
care about dignity, status, gender roles, group membership, and social
approval. Identity therefore becomes an important determinant of
economic behavior. The authors combine economics with psychology and
sociology to explain outcomes like discrimination, occupational
segregation, poverty, political conflict, and household behavior.
The central idea is that society divides people into social categories such
as man/woman, black/white, rich/poor, citizen/immigrant,
professional/nonprofessional. Each category carries social expectations
and behavioral prescriptions. People gain utility when their behavior
matches these expectations and lose utility when they violate them. For
example, masculinity may be associated with aggression and physical
strength, while femininity may be associated with care and gentleness. If a
man behaves in ways society considers feminine, both he and others may
experience discomfort or anxiety. Thus identity changes the payoffs from
economic actions.
The authors explain four major reasons why identity matters in economics.
First, identity explains behavior that seems irrational or self-destructive.
People may undertake painful or economically harmful actions because
these actions strengthen identity or social belonging. Examples include
gang behavior, nationalism, extreme beauty practices, tattooing,
circumcision, steroid abuse, self-starvation, and plastic surgery. These
actions often cannot be understood through income maximization alone.
Second, identity creates externalities because one person’s behavior
affects another person’s self-image. For example, when women enter
occupations traditionally viewed as male jobs, some men feel their
masculinity threatened and may respond with harassment or exclusion.
Thus identity creates emotional and social spillovers beyond ordinary
market externalities.
Third, identity can be manipulated and socially constructed. Advertising,
education, politics, and religion attempt to shape people’s identities and
social norms. Marlboro advertisements, for example, associated smoking
with rugged masculinity, while Virginia Slims advertisements associated
cigarettes with feminine independence. Political movements similarly
reshape identity through nationalist or revolutionary symbols. Gandhi’s
Salt March helped create Indian national identity, while the French
Revolution transformed subjects into citizens.
Fourth, the choice of identity itself becomes an important economic
decision. People often choose what kind of person they want to become
and which social groups they belong to. Parents choose schools partly to
shape children’s identities. Students choose fraternities, sororities, or
theme-oriented dormitories because these shape self-image and behavior.
Immigrants choosing citizenship also experience identity change involving
anxiety, guilt, and belonging. However, identity choice is not fully free
because race, appearance, language, and accent often constrain
movement across social categories.
The paper incorporates identity directly into the utility function. Traditional
economics assumes utility depends only on actions or consumption.
Akerlof and Kranton instead argue that utility also depends on identity.
Identity itself depends on social category, personal characteristics, and
conformity with social prescriptions. People gain identity utility when their
characteristics and actions match social expectations and lose identity
utility when they deviate.
The paper draws heavily from social psychology. Psychologists argue that
individuals strongly identify with groups and derive self-esteem from group
membership. The Robbers Cave experiment showed that even arbitrary
divisions among boys at a summer camp created strong group identities,
hostility, and conflict. Minimal group experiments similarly showed that
even meaningless group assignments lead people to favor their own group
and discriminate against outsiders. These experiments support the idea
that identity strongly affects economic and social behavior.
The authors then discuss identity-related behavior in real life. Body
modification practices such as piercing, tattooing, circumcision, and
cosmetic surgery are viewed as attempts to match social ideals and affirm
belonging. Identity also explains occupational conflict. Occupations often
carry gender identities: lawyer and Marine are viewed as masculine, while
nurse is viewed as feminine. Women entering masculine professions often
report pressure to suppress femininity and behave aggressively to fit
occupational expectations. A study of female lawyers showed they felt
forced to “act like men” in order to succeed professionally.
Identity also explains behaviors like alumni donations and mountaineering.
Alumni donate not only because of economic efficiency but because
universities form part of their identity. Mountaineers undertake dangerous
climbs because extreme hardship strengthens self-image and personal
identity.
The paper gives important examples of identity externalities. Women
entering male jobs may provoke harassment because male workers feel
less masculine. In studies of power plants and construction work, male
workers refused to help female coworkers, mocked them, or harassed
them physically. Surveys showed many respondents believed harassment
occurred because men feared loss of masculinity, feared teasing from
other men, or believed women did not belong in male occupations.
The authors also discuss the “culture of honor” in the American South.
Experiments showed Southern men reacted more aggressively to insults
than Northern men because insults threatened masculine identity. Insulted
Southerners displayed more aggression, stress hormones, and concern
about how others viewed their masculinity. Thus violence and aggression
can partly arise from identity preservation rather than material incentives.
Identity also explains social ostracism. People who violate group norms are
mocked or excluded because their actions threaten others’ identities.
Students trying to escape poverty or move into elite education are often
ridiculed by peers for “acting differently.” Betty Friedan faced criticism
from suburban housewives because her feminism challenged traditional
female identity.
The paper next develops a prototype model using two social categories:
Green and Red. Each category has prescribed activities. Greens are
expected to choose Activity One while Reds are expected to choose
Activity Two. If someone violates these prescriptions, they suffer identity
loss. Others in the same group may also suffer because the violation
threatens shared identity. They may respond aggressively or punish the
violator to restore their own sense of identity. The model therefore
connects economics with psychological theories of anxiety, identification,
and ego defense.
The model generates several equilibrium outcomes. Individuals may
conform to social prescriptions because punishment or identity loss is too
costly. Others may violate prescriptions if material rewards exceed identity
costs. The model also shows that government policy can alter behavior
either by increasing the cost of punishment or by changing social norms
themselves. Public campaigns can therefore reshape identity and
economic outcomes.
The paper applies identity theory to gender and labor markets.
Occupational segregation is explained not simply by skill differences or
human capital but by social ideas about masculine and feminine work.
Men may avoid “female” jobs because such jobs reduce masculine
identity, while women may avoid “male” jobs because these reduce
feminine identity. Male workers may also harass women to defend their
own gender identity.
The authors explain that firms may create separate “male” and “female”
occupations to avoid productivity losses caused by identity conflict.
Historical examples include teaching. Elementary school teaching became
feminized because women were viewed as naturally caring and morally
pure, while higher administrative roles remained masculine. Gender-job
associations persist because changing social norms is costly and individual
firms may not benefit enough from changing them. Therefore broader
social change and legal intervention are often necessary.