Inequality, Poverty and Welfare

1
Inequality, Poverty and Welfare
2
Various issues

• Household or macro data
• Between nations and within nations
• Between groups (religion/ethnicity)
• Income or consumption
• Crossectional and time series data
• Historical data

3
Inequality scenario in the world

• Sri Lanka Equal distribution of income
• South Korea and Taiwan Equal distribution
  combined with rapid growth
• Latin America High inequality
• Ethiopia, Kenya, Zambia, Philippines, Colombia,
  Peru, Thailand, Brazil, Malaysia, Venezuela and
  Mexico Unusually high inequality
• Bulgaria, Poland and Hungary (transition
  economies)
• Australia, Singapore and USA (developed countries)

4
The inverted-U hypothesis

• According to Kuznets, a country in its initial
  stage of development exhibits low per capita
  income level and relatively low inequality level.
  
• As the country develops and per capita income
  increases, inequality tends to increase as well.
• At a more advanced stage of the development
  process, however, the per capita income-
  inequality relationship turns from positive to
  negative.
• Using data from developed and developing
  countries in the early 1960s, he plotted it in a
  graph where the vertical axis represents
  inequality (given by the ratio of income of the
  richest 20 of the population to the poorest 60
  of the population) and the horizontal axis, per
  capita income levels.
• The plotted data had the shape of an inverted-U
  curve.

5
Problems with inverted-U hypothesis (1)

• Pooling countries in a cross-sectional study may
  cause an artificial Latin effect.
• All the middle-income countries tend to be Latin
  American, so it may be something particular about
  Latin American regimes or culture that causes
  higher inequalities in this middle-income group,
  rather than their stage of development.
• Thus it is necessary to run tests using
  country-specific dummies.

6
Problems with inverted-U hypothesis (2)

• Doing this, Deininger and Squire (1998) found
  coefficients that do not fit Kuznets inverted-U
  hypothesis.
• The findings of Fields and Jakubson (1994), who
  suggest that if there is any relationship it is
  that inequality tends to fall with development,
  also refute the inverted-U law, as do Dollar and
  Kraay(2000) and Ravallion(2001).
• The growth in South East Asia did not fit the
  inverted-U hypothesis.
• government policies
• For example, the focus on education allowed the
  skills base of the poor to quickly catch up with
  what was demanded by the growing sectors.

7
Reasons for inequality

• Ravallion and Datts study (1999) of 15 of
  Indias states looks more closely at the
  questionwhy inequalities in some countries are
  rising?
• They find that the growth of the non-farm sector
  had a better impact on the poor in states with
  initially higher farm productivity, higher rural
  living standards relative to urban areas, and
  higher literacy, as these factors all make it
  easier for those living in that area to reap
  benefits from the growing non-farm sector.
• Ravallion and Datt also noted that rural economic
  growth cuts poverty more than urban growth.
• In Brazil equalisation of the distribution of
  education amongst income groups between 1976-96
  resulted in a fall of income inequality.

8
Range and Range Ratio

• The range is simply the difference between the
  highest and lowest observations.
• salary difference between highest and the lowest
  earner
• Easy to understand and compute
• But ignores about distribution (outliers)
• The Range Ratio is computed by dividing a value
  at one predetermined percentile (higher) by the
  value at a lower predetermined percentile (95/5
  percentile)
• Easy to calculate and understand
• Outliers are taken care
• But this measure focuses just on 2 observations

9
The mean absolute deviation

• This measure takes the advantage of the entire
  income distribution.
• Concept inequality is proportional to distance
  from the mean income.
• We take all income distances from the average
  income, add them up and divide it by total income
  to express the average deviation as a fraction of
  total income

10
Coefficient of Variation

• The Coefficient of Variation () of a set of
  values is calculated as
• 100(Standard Deviation)/(mean value of set)
• For a dataset that is closely bunched around the
  mean, the peak will be high, and the coefficient
  of variation small.
• Data that is more dispersed will have a shorter
  peak and a higher coefficient of variation.
• Ceteris paribus, the smaller the coefficient of
  variation, the more equitable the distribution.

11
CV of height of various regions
12
Theils T statistic

• where n is the number of individuals in the
  population
• Yp is the income of the person indexed by p
• and µy is the populations average income.
• If every individual has exactly the same income,
  T will be zero this represents perfect equality
  and is the minimum value of Theils T.
• If one individual has all of the income, T will
  equal ln n this represents utmost inequality and
  is the maximum value of Theils T statistic.

13
Between group and within group

• Theils T statistic is made up of two components,
  the between group element (Tg) and the within
  group element (Twg).
• T Tg Twg
• When aggregated data is available instead of
  individual data, Tg can be used as a lower bound
  for the populations value of Theils T
  statistic.

14
100
80
Line of complete equality
60
Percentage share of national income (cumulative)
A
40
B
Lorenz curve
20
O
O
100
60
80
20
40
Percentage of population
15
100
Gini coefficient A / (A B)
80
Line of complete equality
60
Percentage share of national income (cumulative)
A
40
B
Lorenz curve
20
O
O
100
60
80
20
40
Percentage of population
16
The Gini-Coefficient

• Divide A by the sum of A B to get the Gini
  coefficient
• If the Lorenz curve is on the 45 deg. Line, the
  Gini coefficient would be 0
• Range 0-1
• Limitations

17
100
Gini coefficient A / (A B)
80
60
Percentage share of national income (cumulative)
L1
40
L2
Lorenz curve
20
O
O
100
60
80
20
40
Percentage of population
18
Infant Mortality Residual vs. Gini Coefficient
(Epstein et al)
19
Relation between poverty and inequality

• Inequality focuses on the distribution of
  attributes, such as income or consumption, across
  the whole population.
• In the context of poverty analysis, inequality
  requires examination if one believes that the
  welfare of an individual depends on their
  economic position relative to others in society.
• Vulnerability is defined as the risk of falling
  into poverty in the future, even if the person is
  not necessarily poor now it is often associated
  with the effects of shocks such as a drought, a
  drop in farm prices, or a financial crisis.

20
Historical data
(Strauss and Thomas) Health nutrition and
economic development
21
Poverty and development
22
Human development index

• Life expectancy at birth (this will indirectly
  reflect infant and child mortality)
• Educational attainment of the society which is a
  composite index (It takes weighted average of
  adult literacy (with weight of 2/3) and a
  combination of enrolment rates in primary,
  secondary and tertiary education (with weight
  of 1/3).
• Per capita income (adjusted)
• HDI is calculated using average of the above
  three indicators.

23
Comparison of per capita income and HDI, 1998
24
Poverty Monetary dimensions

• Monetary measures of poverty can be measured
  using income or consumption.
• Consumption will be a better indicator for
  poverty measurement than income
• Consumption is a better outcome indicator than
  income
• Consumption may be better measured than income
• Consumption may better reflect a households
  ability to meet basic needs

25
Non-monetary dimensions of poverty

• Health and nutrition poverty
• Education poverty
• Composite indices of wealth
• One can combine the information on different
  aspects of poverty that creates a measure which
  takes income, health, assets and education into
  account.
• Limitation of composite indices It is not
  possible to define a poverty line. Analysis by
  quintile or other percentile remains possible.

26
Important issues poverty

• Absolute or relative?
• Absolute poverty measures the number of people
  living below a certain income threshold or the
  number of households unable to afford certain
  basic goods and services.
• Relative poverty measures the extent to which a
  household's financial resources falls below an
  average income threshold for the economy..
• Temporary or chronic
• Fluctuations in income and consumption (famine)
• Seasonal differences in food availability and
  consumption
• Household or individual expenditures are reliable
  measures to assess chronic poverty

27
Important issues poverty (2)

• Households or individuals
• Most of the measures neglect the allocation of
  resources within households
• Most of the potential victims are female and
  elderly
• Poverty line
• Is it possible to have a fixed notion of poverty?
• Poverty lines are approximations to a threshold
  that is unclear as the effects of sustained
  deprivation are often felt at a later point in
  time

28
Poverty

• Rural and urban poverty
• Land holding
• Nutrition
• Intake
• Calories
• Intra-household allocation
• Female headed households

29
Rural-urban disparity
30
Data

• http//www.iisg.nl/hpw/data.html
• http//www.measuredhs.com/aboutsurveys/dhs/start.c
  fm
• http//utip.gov.utexas.edu/data.html
• http//pwt.econ.upenn.edu/php_site/pwt_index.php
• http//www3.who.int/whosis/menu.cfm
• http//www.wider.unu.edu/wiid/wiid.htm
• http//chinadatacenter.org/newcdc/
• http//chinadataonline.org/
• http//sedac.ciesin.columbia.edu/china/popuhealth/
  popagri/census90.html

31
Thank You