Documente Academic
Documente Profesional
Documente Cultură
A. Wismadi
14. Evaluation of Design
Cases
ACCESS
36
Criteria for Equity Measures
Four criteria for good measurement:
1. Anonymity-permutations of incomes among people should not matter for inequality judgments.
2. Scale independence principle (Relative Income Principle)– only relative incomes shares matter.
3. Population independence principle- population size doesn’t matter, i.e.- all that matters are
proportions of the population that earn different levels of income.
4. Dalton principle– if one income distribution can be achieved from another by constructing a
sequence of regressive transfers, then the former distribution must be deemed more unequal
than the latter.
37
Curve distribution comparison
no intersecting intersecting ?
=
Problem in Lorenz Curve:
No information on level of prosperity
• Lorenz Curve from group of people below and above poverty level
may be identical
poverty line
CDF (Cumulative Distribution Functions)
Deficit: A+C
The area beneath the poverty incidence curve
For distribution F(y) this is A + C at poverty line z1
A Severity:C
The area beneath the poverty deficit curve
y income level
z the poverty line
(Demery, 2007)
TIP Curve
• The Three Is of Poverty Curve (Jenkins and Lambert 1997)
• Incidence Intensity Inequality
• Rank individuals from poorest to richest, cumulating their average
• poverty gaps: g( y i , z) max z y i ,0
g(y i , z) z yi
• or normalized gaps ( y i , z ) max ,0
z z
• Plot the per capita cumulative gaps against cumulative population
share
TIP Curve problem :
overlooked the non-poor-group
• As we move from poorest
individuals upwards, the
per capita cumulative gap
increases
• When y is above z the TIP
curve is flat.
• If every poor person had
the same income, the TIP
curve would be linear
• If income unequal
amongst poor, the TIP
curve will be bowed
inwards
(Demery, 2007)
(Demery, 2007)
Introducing New Measure:
W-Curve!
• Existing measure does not satisfy pro-poor growth policy evaluation, new
measure is required.
• Approach for developing W-Curve:
• Including properties of Equity (defined ‘floor’, Poverty Line, minimum line to provided
opportunity)
• Including non-poor (rich) group properties (prosperity)
• Intuitive and quantitative analysis (combining graphical and quantitative evaluation)
• Enable to compare two distribution curves
• Comprehensive measurement including (almost) all properties of pro-poor growth
measurement
Introducing New Measure:
W-Curve Properties
Inequality (E)
HIGH :~1
LOW :0
Poor Non-Poor
Step-by-step to build W-Curve Start
Estimate HH
Income*
Calculate Absolute
Number of HH in
Value of Gap (-)
each village
and Surplus (+)
Share of Income (Y) Share of HH (X)
Cumulative
find Plot
sort Distribution of Y=Y+minY W-Curve
min(y) (X,Y)
(X,Y)
find
find (y’) find (y’’)
1st (x’) calculate
on on last
where Gini Index
1st (x) (x)
y=0
Degree of
Degree of
Poverty Prosperity
Severity (S)= Inequality (E)
Incidence (I) (P)=
cot(y’/x’)
*: Income as example for explaining the measurement method cot(y’’/1-x’)
Resource Allocation