Lecture 3

Poverty and inequality

The number of people in extreme poverty
Our world in Data
 Average Poverty Gap at $2.15 a day (2017 PPP) (%)

1990 2000 2015 2019

Regions
East Asia and Pacific 25.6 13.0 0.5 0.2 𝐴𝑃𝐺
=
Europe and Central
σ𝑦𝑖<𝑧(𝑧 − 𝑦𝑖 )൙
1.2 3.3 0.9 0.6
Asia 𝑁. 𝑧
Middle East and North 1.3 0.7 1.4 --
Africa

South Asia 14.3 - 3.3 1.6

Sub Saharan Africa 23.9 25.5 13.5 12.5

World 14.0 9.8 3.1 2.6

India - - 3.7 1.8
China 28.8 - 0.2 0
Measurement and comparison of income inequality
Two goals:
• Measure the degree/nature of income inequality within a country
• Compare inequality between countries.
A natural way to study/investigate income inequality in a country is to look at
the income distribution (e.g., frequency distribution).
It is customary among researchers (studying income inequality) to calculate the
size distribution (from the underlying frequency distribution of income) and
analyze this.
• Size distribution tells us the share of national income received by different
groups of the population ranked according to their income levels.
• Look at the following table.
Size distribution
Share of total income, source: World Bank
Quintiles Bangladesh Mexico

2010 2016 2010 2016

Poorest 20% 8.9 8.6 4.7 4.7

Second poorest 20% 12.5 12.4 9.0 8.8

Third 20% 16.0 16.1 13.2 13.1

Fourth 20% 21.2 21.4 20.3 20.0

Richest 20% 41.5 41.4 52.9 53.4

• Just by looking at the above Table, one would intuitively conclude that ‘income
inequality’ is larger in Mexico than in Bangladesh. (On what basis?)
• One may also conclude that income inequality has risen slightly in both Bangladesh and
Mexico between 2010 and 2016. (Why?)
Quantile measures
What are quantiles?

• Quantiles are a ‘general term’ for the following: quartiles (4-groups), quintiles (5-
groups), deciles (10-groups), and so forth.

Quantile ratio as a measure of inequality:

• A quantile ratio is the ratio of the average income of the highest quantile to the
average income of the lowest quantile. For example,

𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑖𝑛𝑐𝑜𝑚𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑖𝑐ℎ𝑒𝑠𝑡 20%

𝑄𝑢𝑖𝑛𝑡𝑖𝑙𝑒 𝑟𝑎𝑡𝑖𝑜 =
𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑖𝑛𝑐𝑜𝑚𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑝𝑜𝑜𝑟𝑒𝑠𝑡 20%

The quintile ratio for Brazil in 2018 was about 19 (!)

While quantile ratios are easy to compute, these do not use ‘all the available information
on income distribution’ – a shortcoming?

(In the quintile ratio, no information is used from the 2nd, 3rd and 4th income quintiles).
Measurer of income inequality
The Lorenz curve (Max O. Lorenz, 1905)

The preferred way for economists to present information on income

distribution is via a Lorenz curve.

The Lorenz curve shows the cumulative income share held by cumulative
(quantile) share of the population.
 Measurement of inequality: Individuals Income share (% of Quintile Quintile,
national income) (one fifths)
Distribution of individual income cumulative
1 1.8

100 2 3.2 5 5
Proportion of national income,

3 3.9
80
4 5.1 9 14
Line of
60 equality 5 5.8
Lorenz
cumulative (%)

kurve 6 7.2 13 27
40
(80,49)
(60,27) 7 9.0
(40,14)
20
(20,5) 8 13.0 22 49
45⁰
0 20 40 60 80 100 9 22.5
Proportion of population (%)

• We arrange population according to increasing income along x-axis. 10 28.5 51 100

• Along y-axis, we measure proportion of national income that goes to the

Total 100 100
corresponding proportion of population measured along x-axis.
Income distribution: Lorenz curve
Lorenz curve shows a quantitative relationship between percentages of
population and the percentages of national income they receive.

Lorenz criterion:
L(A) and L(B) are two Lorenz curves for two different
income distribution, A and B, (say, A and B are two
Proportion of national income (%)

100 countries).

If L(B) lies below (or to the right) of L(A) then the

poorest X% of population (0<X<100) in country B
owns less of its national income than in country A.
L(A)
This implies that B represents a more unequal
income distribution compared to A. (Lorenz-
L(B) criterion)

0 20 40 X=60 100 The diagonal line represents the Lorenz curve for an
Proportion of population income distribution where everyone has the same
(%) from poor to rich income (perfect equality) 10
Income inequality
Lorenz curve for complete inequality

national income (%) 100 Consider the following scenario.

The entire population has no

80
income (or income share) except
Proportion of

one person.
60 The entire national income
accrues to just one person.

40 Then the Lorenz curve will

coincide with the blue line – the
horizontal and the vertical
20
segment. Given by the blue line.

0 20 40 60 80 100
Proportion of population (%) from
poor to rich
Income inequality, cont’d
Lorenz curve for perfect equality

100%
national income (%)
The poorest X%, (0 < X ≤100)
has exactly X% of national
Proportion of

income.

Lorenz
curve

0 20 40 60 80 100
Proportion of population (%) from
poor to rich
Lorenz criterion (for comparing income distributions)

• Lorenz criterion: If the Lorenz curve for income

distribution B lies to the right of the Lorenz for income
distribution A, then distribution B is more unequal than
distribution A.

13
Income inequality
Gini coefficient: An aggregate measure of inequality
= (area between the Lorenz curve and the diagonal) / (area of the triangle ABC)

% share of income
C

100
Gini coefficient Lorenz curve
=

A B
0 100 % share of population

Gini coefficient varies between 0 (perfect equality) and 1 (perfect/complete inequality).

Why? Should explain this in some detail. 14
 Patterns of inequality
Gini-coefficient ×100 (2005) Gini-coefficient×100
(2010)
Sub-Saharan Africa
Ethiopia 33.6 (2011)
Guinea 33.7 (2012)
South Africa 63.1 (2009)

Latin America & Caribbean

Peru 51.1 48.1
Mexico 51.2 47.2
Brazil 57.4 54.7 (2009)

East and South Asia

Vietnam 36.8 (2004) 35.6 (2008)
Bangladesh 33.2 32.1
Malaysia 37.9 46.2 (2009)

Europe and Central Asia

Hungary 30.0 (2004) 31.2 (2007)
Slovenia 24.6 24.9
Turkey 42.6 40.0

High income
Japan 25.0 (1993?)
Germany 32.1 30.2 15
Economic determinants of inequality

We observe large differences in income inequalities between countries.

The main questions we address below are:

• What are the underlying causes that generate more inequality in some
countries and less in others?

We discuss two issues here:

1. Inequality in educational achievements (human capital) among citizens

leading to income inequality

2. Inequality in land ownership causing income inequality.

1. Education and income inequality
• High participation/free-access to education, (for example, free/state-funded
education system), will mean that the entire population, in general, will have high
skill/education levels (- human capital -). That is, free access will mean that there will
be low inequality in the level of human capital in the population. This will lead to low
earnings differential among citizens.

• If, on the contrary, only a few in the population are educated (- that is, many are
uneducated/illiterate -) then large earnings/income inequality will arise.

• Further, the poor with low levels of education are less like to participate in political
debate on inequality. This means less pressure will be put on the authorities for
inequality reduction – income redistribution, etc.

• That is, low levels of education among the poor help perpetuate education and
income inequality.
2. Land/asset ownership and income inequality
Another determinant of income inequality is the unequal distribution of land,
i.e., inequality in property ownership.
• Latin American countries historically have had highly unequal land distribution.
Asian countries on the other hand have more equitable distribution of land.
This appears to translate into relatively high Gini coefficient (high income
inequality) in South America relative to Asia.
• In poor countries, especially with a large rural/agrarian sector, land is the main
source of income.
• Figure 2.5 in your textbook shows a plot of “Gini coefficient for land
ownership” against “Gini coefficient for income”, for different countries. A
strong positive correlation was found between the two – i.e., high inequality in
land distribution is correlated with high income inequality.
Inequality, growth and development
Question: Does income inequality lead to low/high economic growth, or is
it the other way round? That is, it is the economic growth that causes
inequality.

This is a question about the direction of causality between income inequality

and economic growth.

Kuznets’ inverted-U hypothesis

▪ Economic growth causes income inequality to rise in the early stage of
development, that is, when the economy is still relatively poor.
▪ As growth continues, and the economy matures to a developed stage, the
inequality declines.
Kuznets inverted-U hypothesis
Figure 2.6 from text:

Ratio of income share of the richest 20%

to the share of the poorest 60%

GDP per capita

Cross-sectional country data

Kuznets hypothesis: What are the economic arguments/mechanisms?
Consider a [primitive/pre-industrial] poor economy dependent on agriculture. The
distribution of land is not ‘too’ unequal and income inequality is relatively low.

• Suppose this rural economy starts to industrialize. It is likely that there will only be a
few entrepreneurs/capitalists in the beginning of the industrialization process.

• This means that the factory owners/capitalists will earn large profits. Worker wages
will still remain low as the supply of labour is plentiful.

• As a result, inequality will arise over time.

• As the process of industrialization takes hold, new entrepreneurs enter the market.
Competition among entrepreneurs will lower profits.

• Worker wages, on the other hand, will rise as many entrepreneurs will be bidding
for a decreasing supply of available workers.

• Inequality will gradually decline.

Kuznets hypothesis
This hypothesis finds some support in cross-sectional data on income inequality
(Look at slide #15, Gini coefficient for countries).

• Slide 15 shows that the poor economies of Africa and Asia have low Gini
coefficients.

• The south American countries which belong to middle income group have the
highest Gini of all countries.

• This rich industrialized countries also have relatively low Gini coefficient.

However, this does not necessarily mean that all countries will go through these
phases of income inequality, ( low – high – low), in the process of development (over
time).

That is, the data does not necessarily show a causation, i.e., that growth is causing
changes in income inequality.
Kuznets hypothesis

• There is indeed little support for Kuznets hypothesis in longitudinal/time

series data!

• That is, if we follow the historical development process of countries over

time, we find little or no statistical evidence that income inequality for a
given country rises first before declining (through the process of
economic growth).
Income inequality and economic growth
Does higher income inequality cause lower growth?
Why inequality matters? Some observations:

• Large differences in income between people is a problem in itself. - A

moral/ethical issue.

• It may appear unjust and unfair that some live in absolute luxury while
others are struggling just to survive - ”Intrinsic problem”

• But, more importantly, we, as economists, want to know if income inequality

has consequences for saving, education, entrepreneurship, political stability,
criminal activities/corruption, that is, if and how income inequality affects
economic growth and development.
24
Income inequality and economic growth
Look at Figure 2.7 in your text.
• It plots “long-term economic growth” against Gini coefficient (for
income) for a large number of countries.
• The data show a negative correlation. That is, countries with high Gini
coefficient tend to have low growth.
Based on empirical evidence from past 30 years, economists largely
agree that high income inequality will most definitely have a negative
effect on economic growth.
What are the economic arguments/models/theories behind this?
Income inequality and economic growth
Why/how inequality may affect economic growth
There are a number of views.
1. Imagine a poor economy with high degree of income inequality: That is, the country has a
few very rich and many poor.
• Rich are the ‘landed gentry’ and are not interested in investment activities!! (Assumption)
• Further, most of the poor, especially in the agricultural sector, (although willing) cannot
undertake simple productivity increasing investment because they lack the means, i.e.,
savings.
• Assume that credit markets do not exist in this country or are poorly developed. The poor
will then have no access credit → no (or low) investment → no (or low) growth.
• Many of the poor’s investment projects are inexpensive but give high returns. A little less
inequality (i.e., a little bit more to the poor meaning poor having some ability to save)
and/or opportunities for borrowing (through credit markets) could make many of these
investments affordable/feasible.
• High inequality → low investment → low growth. 26
Income inequality and economic growth
2. “Political economy”

(a) In a highly unequal society, the rich can buy political influence.

▪ The rich can affect government budgetary allocation for education, health care, trade policies,
etc., for example, in favour of universities (rather than primary education), urban hospitals
rather than basic health care in rural areas. This can affect economic growth negatively.
▪ Note that the rich generally send their children to universities and use urban hospitals.
This leads to low levels of education and poor health for masses of poor, especially in
rural areas, and low economic growth.

(b) Another political economy view (read text, page 50)

▪ (i) Consider a poor country with high inequality and a social democratic government. The
constituents/voters will demand income redistribution through high taxes on entrepreneurial
profits and income. This may lower both investment and work incentive and leading to lower
growth. Go to next slide…..
Income inequality and economic growth
(bii) Another political economy view (continued from previous slide )

▪ On the other hand, the poor, under democracy, may vote for higher public
expenditure on schools, public health care, public transportation and other
infrastructure. These would have positive impact on growth.

(c) Yet another view: High inequality is good for growth (?!)

▪ High inequality means income/wealth is concentrated in the hands of a few.

Rich, with a lot of savings, can contribute to large/rapid capital accumulation
(investments) necessary for growth.

We have argued on the both sides on the effects of inequality on economic

growth. The net effect of all of these hypotheses – i.e., whether inequality affects
growth or not - remains an empirical question.