Chap4 Notes Econ-25
Chap4 Notes Econ-25
Measuring Inequality
A. Size Distributions
Personal distribution of income (size distribution of income) – the distribution of income
according to size class of persons—for example, the share of total income accruing to the
poorest specific percentage of a population—without regard to the sources of that income.
Common methods to arrange all individuals by ascending personal incomes and then divide the
total population into distinct groups or sizes are by dividing the population into:
Quintile (fifths) – a 20% proportion of any numerical quantity. A population divided into
quintiles would be divided into five groups of equal sizes.
Deciles (tenths) – a 10% portion of any numerical quantity; a population divided into
deciles would be divided into ten numerical groups.
Income inequality – the disproportionate distribution of total national income among households.
It can be derived from the ratio of incomes received by the top 20% and bottom 40% of the
population, also called as Kuznets ratio. Kuznets ration has been used as a measure of the
degree of inequality between high- and low-income groups in a country.
B. Lorenz Curve – a graph depicting the variance of the size distribution of income from perfect
equality. It was developed by Max O. Lorenz in 1905 for representing inequality of the wealth
distribution.
The more Lorenz line curves away from the diagonal (perfect equality), the greater degree of
inequality presented. The greater the degree of inequality, the greater the bend and the closer
to the bottom horizontal axis the Lorenz curve will be.
Total Poverty Gap (TGP) – the sum of the difference between the poverty line and actual
income levels of all people living below that line. It measures the total amount of income
necessary to raise everyone who is below the poverty line up to that line
𝐻
TPG = ∑(𝑌𝑝 − 𝑌𝑖 )
𝑖=1
On a per capita basis, the average poverty gap (APG) is found by dividing the TPG by the total population:
TPG
APG =
𝑁
Normalized poverty gap (NPG): NPG = APG/Yp; this measure lies between 0 and 1 and so can be useful
when we want a unitless measure of the gap for easier comparisons.
Average income shortfall (AIS), the total poverty gap divided by the headcount of the poor: AIS=TPG/H.
The AIS tells us the average amount by which the income of a poor person falls below the poverty line.
This measure can also be divided by the poverty line to yield a fractional measure, the normalized income
shortfall (NIS): NIS=AIS/Yp.
Poverty Gap Index (P1) – measures the extent to which individuals fall below the poverty line
(the poverty gaps) as a proportion of the poverty line. The sum of these poverty gaps gives the
minimum cost of eliminating poverty, if transfers were perfectly targeted. The measure does not
reflect changes in inequality among the poor.
Criteria for a desirable poverty measure that are widely accepted by development economists:
1. Anonymity principle
2. Population independence principle
3. Monotonicity principle – if you add income to someone below the poverty line, all other
incomes held constant, poverty can be no greater than it was
4. Distributional sensitivity principle – other things being equal, if you transfer income from a
poor person to a richer person, the resulting economy should be deemed strictly poorer.
The headcount ratio measure satisfies anonymity, population independence, and monotonicity,
but it fails on distributional sensitivity. The simple headcount fails even to satisfy the population
independence principle.
The Foster-Greer-Thorbecke Index (Pα class of poverty measures)– a class measures of the level
of absolute poverty. FGT class is based on the normalized gap of a poor person, which is the income
shortfall expressed as a share of the poverty line.
𝐻
1 𝑌𝑝 − 𝑌𝑖
𝑃𝛼 = ∑( )𝛼
𝑁 𝑌𝑝
𝑖=1
Squared Poverty Gap (poverty severity) Index (P2) – averages the squares of the poverty
gaps relative to the poverty line. It has become a standard of income poverty measure used by
the World Bank and other agencies, and it is used in empirical work on income poverty because
of its sensitivity to the depth and severity of poverty. Another reason to prefer P2 (or at least P1)
over P0 is that standard headcount measures also have the perverse property of creating
incentives for officials to focus efforts on the poor who are closest to the poverty line—because
that is the easiest and cheapest way for them to demonstrate progress.
Pearson-Equivalent Headcounts. Given a political need to feature “headline” headcount measures, a
partial improvement is to convert changes in the poverty gap into its headcount-equivalent (based on
the initial average income shortfall). If aid agencies featured a supplementary headcount-equivalent,
they could report in terms of numbers of people while accounting for changes in poverty depth.
Estimates using this approach show progress against poverty in many countries is significantly greater
than revealed using conventional headcount measures alone.
Multidimensional Poverty Measurement. Sabina Alkire and James Foster filled the gap on which
poverty cannot be adequately measured with income alone by extending the FGT index to multiple
dimensions.
A poor person is identified as poor through what is called the “dual cutoff method”: 1) the cutoff
levels within each of the dimensions (analogous to falling below a poverty line such as $1.25
per day if income poverty were being addressed); 2) the cutoff of the number of dimensions in
which a person must be deprived (below the line) to be deemed multidimensionally poor.
Multidimensional headcount ration HM – fraction of the population in multidimensional poverty
M0 (adjusted headcount ratio) –uses ordinal data and is similar conceptually to the poverty
gap P1 (which again can be expressed as the headcount ratio times the normalized income
shortfall). It may be represented by the product of the multidimensional headcount ratio times
the average fraction of dimensions in which the poor are deprived (or “average intensity of
poverty” A, that is, M0 = HM * A. The adjusted multidimensional headcount ratio satisfies the
desirable property (dimensional monotonicity) that if average fraction of deprivations increases,
so does M0.
Often, the poor lack access to credit, which constrains entrepreneurship, children’s education, and
fertility reduction
Social exclusion/injustice associated with poverty also leads to bad government policies that can
reduce growth
Multidimensional Poverty Index (MPI) – a poverty measure that identifies the poor using dual cutoffs
for levels and numbers of deprivations, and then multiplies the percentage of people living in poverty
times the percent of weighted indicators for which poor households are deprived on average. It takes
into account that there are negative interaction effects when people have multiple deprivations—worse poverty
than can be seen by simply adding up separate deprivations for the whole country, then taking averages, and only
then combining them.
First, cutoff levels within each dimension (analogous to falling below a poverty line for
example $1.25 per day for income poverty);
Second, cutoff in the number of dimensions in which a person must be deprived (below
a line) to be deemed multidimensionally poor.
MPI focuses on deprivations in health, education, and standard of living; and each
receives equal (i.e., one-third of the overall total) weight. MPI indicators are as follows:
1. Health – two indicators with equal weight – whether any child has died in the family,
and whether any adult or child in the family is malnourished – weighted equally (each
counts as one-sixth toward the maximum deprivation in the MPI)
2. Education – two indicators with equal weight – whether no household member
completed 5 years of schooling, and whether any school-aged child is out of school
for grades 1 to 8 (each counts one-sixth toward the MPI).
3. Standard of living, equal weight on 6 deprivations (each counts as 1/18 toward the
maximum): lack of electricity; insufficiently safe drinking water; inadequate sanitation;
inadequate flooring; unimproved cooking fuel; lack of more than of 5 assets—
telephone, radio, TV, bicycle, and motorbike.
Interaction of the deprivations:
o Building the index from household measures up to the aggregate measure (rather
than using already – aggregated statistics), MPI approach takes account of
multiplied or interactive harm (complementarity) done when multiple deprivations
are experienced by the same individual or family
o The MPI approach assumes an individual’s lack of capability in one area can only
to degree be made up by other capabilities – capabilities are treated as
substitutes up to a point but then as complements.
Computing the MPI
o The MPI for the country (or region or group) is then computed
o A convenient way to express the resulting value is H*A, i.e., the product of the
headcount ratio H (the percent of people living in multidimensional poverty), and
the average intensity of deprivation A (the percent of weighted indicators for
which poor households are deprived on average).
o The adjusted headcount ratio HA is readily calculated
o HA satisfies some desirable properties. Dimensionally monotonicity: If a person
already identified as poor becomes deprived in another indicator she is measured
as even poorer—not the case using a simple headcount ratio.
Multidimensional poverty tells a different story than income poverty
o The results showed that knowing income poverty is not enough if our concern is
with multidimensional poverty.
o Multidimensionally, Bangladesh is substantially less poor – but Pakistan
substantially poorer – than would be predicted by income poverty
o Ethiopia is far more multidimensionally poor, and Tanzania much less so, than
predicted by income poverty.
o Most Latin American countries e.g. Brazil rank worse on multidimensional poverty
than on income poverty; but Colombia’s income and MPI poverty ranks are about
same. (refer Table 5.5 Income Poverty Indicated in Selected Countries)
On the average, in Africa and Asia, about 80% of all target poverty groups are located in the
rural areas, as are about 50% in Latin America. (see Table 5.7 for the data of specific countries in
relation to poverty in rural vs urban)