Académique Documents
Professionnel Documents
Culture Documents
Lecture Notes
Todaro-Smith ch. 5 notes
- NOTE: 0% of people receive 0% of total income; 100% (all) people receive 100% (all) of
total income
- if all people had equal income, poorest 10% would receive 10% of total income, poorest 20%
will receive 20% of total income etc.
- thus, perfect equality will produce a LC which is a straight line, the 45-degree line.
- how about if one person had all the income in the country? Then all but the richest person
(99.99% on the horizontal axis will have 0 income, i.e. a flat line at 0, and the richest person
will have 100% - a vertical line). The LC is a right-angled curve.
- realistically were in between: the LC is an increasing line (Why?) and convex (i.e. increasing
at an increasing rate) (WHY?).
- a LC closer to the 45-degree line corresponds to more equal society, opposite is true for LC
further away from the 45-degree line (closer to the bottom and right sides of the box) figure
5.2.
Gini coefficient
* The LC is a very detailed measure of inequality but hard to use in practice, would be nice to
somehow aggregate the information carried by it into a single number we can compare across
countries.
* this is where the GINI COEFFICIENT measure of inequality comes in.
* GC is an aggregate measure of inequality that can be obtained as (see fig. 5.3) as the ratio of
the shaded area A (the area between the LC and the 45-degree line) and the total area BCD.
* NOTE: GC is by construction a number between 0 and 1.
* GC = 0 corresponds to perfect equality; GC = 1 corresponds to perfect inequality.
* Caution: GC is an aggregate measure two different LCs can have same GC (but then they
must cross) (Explain why). Alternatively, two societies can have different GCs but their LCs
cross how to rank?
* GC has some nice properties which is why it is most commonly used to measure income
inequality in development economics:
- anonymity: GC does not depend on who is who (just their income level)
- scale independence: if we measure income in different units, GC remains the same. If
the economy rich on average or poor on average does not matter too.
- population size independence - GC is units-free; if we clone each person into 2, each
with the same income as the original person, GC remains the same;
- transfer principle: holding all else constant, if we transfer some income from a richer
person to a poorer person, GC declines (the resulting income distribution is more equal).
Functional inequality
* sometimes important to know not only the size of income a person has, but its factor share
distribution (functional inequality). E.g. what fraction of income labor vs. capital receive.
* gives additional, more detailed information on inequality in a country. May be important for
policy: e.g. if a person receives a certain income as unemployment pay vs. working at
minimum wage, etc.
POVERTY measurement
* the inequality measures give idea about relative poverty (how much income people in a
country have relative to each other)
* here: absolute poverty the number of people unable to satisfy some basic needs
*basic idea: define a poverty line (PL) minimum amount of income (PPP adjusted) that
can be used to compare poverty internationally; typically $1 a day or $2 a day.
Headcount (the nr of people below the PL) and headcount index (ratio) the proportion of
people below PL from the whole population.
* headcount H (e.g. 5 mln people), headcount index is H / N (e.g. 10%)
* why poverty line? Simple. But there are issues: e.g. it matters if most people are right below
it, or far from it (policy bias can result from that, explain)
* to measure the amount of poverty more precisely (instead of the binary measure of being
below or above the PL) we use another measure: total poverty gap (TPG)
* TPG measures the total amount of income necessary to lift everyone below the PL to that line
(see fig. 5.6); can think of it as the amount of money per day needed to bring every poor person
to the PL.
* TPG =
(Y
i|Yi <Y p
* creates a measure both different from the standard PL and also from HDI;
*Some examples: Cote dIvoire 26 places worse if poverty measured by HPI than in income;
Morocco 37 places worse; i.e. for these countries poverty is bigger problem than income
measures indicate;
* in contrast, Nigeria, Nicaragua, Bolivia, Tanzania perform better on HPI.
POVERTY, INEQUALITY AND WELFARE
* most people agree that absolute poverty is bad and should be eradicated (e.g. in all religions,
all govt policies); Do you agree?
* but how about inequality? Should relative inequality be a concern? (above the poverty line)
1. economic efficiency: income inequality can lead to inefficiencies
Example 1: credit markets a poor person with a great business idea but no collateral the
idea will never be implemented a loss to society;
Example 2: education; if costly (and there always is an opportunity cost even if no fees) poor
parents may not educate their smart children who may otherwise become doctors, inventors,
etc. loss to society.
Example 3: the saving rate if many poor overall saving rate in the economy can be very low
(less domestic resources to invest in the economy)
Example 4: farming - large farms run by hired labor may be unproductive due to incentive
problems; smaller, family-run farms usually more productive. However, large farms can be
even more productive by using machines.
NOTE: for most of the above to be true we need some type of credit or other market
inefficiency; if all markets are perfect, inequality would not matter (give examples).
2. political and social stability
* higher levels of inequality may undermine political and social stability; inequality makes the
rich richer, raises their power and can yield to outcomes that further exacerbate inequality.
* high inequality can facilitate rent seeking (incl. excessive lobbying, political donations,
bribery, cronyism). Resources devoted to such activities are unproductive! Again economic
inefficiency.
* hard to make reforms and changes: the losers are typically the rich who have the power.
* on the other hand, lots of poor can also lead to populist (redistributionary) policies that are
also bad for the economy in the longer run (e.g. certain land reforms, nationalizations,
unaffordable social policies).
* with more inequality focus of politics is mostly on redistribution rather than increasing the
size of the pie. (more in ch. 11)
3. moral and fairness objections to inequality
Rawls veil of ignorance criterion: what level of inequality would you vote for before being
born. Most people vote for some intermediate amount.
- women and children experience harshest deprivation, more likely to be undernourished, less
likely to receive medical services, clean water, sanitation, etc.; less access to education, formal
sector employment, social security
- poorest segments in LDCs are in women-headed households.
- women paid less for same tasks but also effectively precluded (banned) from high-earning
occupations
- legislation and social customs often preclude women from owning property, businesses,
signing contracts;
* this inequality also exists within households
- strong intra-household bias against women in nutrition, medical care, education, inheritance
e.g. in India girls are 4 times more likely than boys to suffer from acute malnutrition; 40 less
likely to be taken to hospital when ill.
- these gender biases are possible reason for high sex ratios (#men to #women) in some
countries (the missing women) - HICs .95-.98 (many LICs 1.02-1.08) China = 1.06; India
= 1.06; Kuwait = 1.39; Canada = .98; USA = .97; France = .95; Japan = .96;
* why?
- much work performed by women is unremunerated / intangible (parenting, housework)
- often socially unacceptable for women to contribute to money earning leads to their low
bargaining power in the family
- programs: mostly oriented to men (agricultural extension services; job training)
* much more needs to be done focusing on women (education, microfinance)
3. ethnic minorities
* over-represented among the poor
* data difficult to obtain for political reasons, but see Table 5.7 for Latin America; e.g. in
Mexico over 80% of indigenous population is poor vs. 18% of non-indigenous
4. people in the poor countries
* seems obvious but important implication is that growth can help (see more below)
POLICIES (see slides too)
* do some policies maximize poverty reduction?
* Dollar and Kraay try different macro policy variables (rule of law, openness to trade,
inflation stabilization) find no or weak evidence in promoting poverty reduction
* human capital determinant of growth and good for reducing inequality
* seems increasing secondary enrollment matters (but quality of education is key!)
* Besley et al. increasing school enrollment for girls
* How to increase school enrollment?
- by increasing supply of schooling (but trade-off between quality and quantity of schools and
teachers)
- by increasing demand for education (conditional transfer programs in Mexico, Brazil, etc.
that encourage school enrollment)
* rural development?
- not all rural poor are *farmers*
- land reforms? Incentives vs. land parcelization (splitting land into small pieces)
- rural infrastructure (roads, electricity, water)
- transaction costs better connection to markets
What should work?
- productivity enhancing technologies for small farmers (a new Green revolution for Africa?)
- food crops research
- extensions systems
- risk reduction devices for small farmers
- increasing non-farm employment