Growth elasticity of poverty

Download Report

Transcript Growth elasticity of poverty

Poverty-Growth Links
Applied Inclusive Growth Analytics
Kenneth Simler and Roy Katayama (PRMPR)
June 30, 2009
Outline
1) Why look at poverty with growth?
2) Website: “Measuring growth-poverty links”
3) Five tools for measuring poverty-growth
relationships
4) Summary
Why look at poverty?

General consensus that:
Poverty reduction is meaningful goal of development
 Growth is necessary for sustainable poverty
reduction


However, the extent to which growth translates
into poverty reduction varies across countries.
Benefits of growth may not be reaching the poor
 Distributional changes can offset growth effects

Growth and poverty reduction
Annual change in poverty headcount (%)
10
Romania
Zambia
Indonesia
-3
Bolivia
Brazil
Ghana
Burkina Faso
Senegal
Bangladesh
India
Tunisia
Uganda
El Salvador
6
Vietnam
-10
Annual GDP per capita growth, 1990s (%)
Source: Pro Poor Growth in the 1990s. Country Case studies
Growth spells and poverty reduction
Source: Bourguignon (2002)
Poverty-growth-inequality triangle
Source: Bourguignon (2004)



Poverty reduction= f (growth, Δdistribution)
What are effects of growth on distribution?
What are effects of inequality on rate and
pattern of growth?
Poverty-growth-inequality triangle
Source: Bourguignon (2004)

Ex-post analysis of this relationship can:
Inform ex-ante analysis of poverty and
distributional impacts of policies
 Help policymakers in evaluating policy options

Looking beyond averages

Inclusive growth analysis requires:
Good understanding of growth at the mean,
 …but also the incidence of growth across the
distribution,
 ... and changes to the distribution and poverty.


Review of ESW indicated:

Many could have been strengthened by utilizing
existing tools on growth-poverty links.
WEBSITE: “MEASURING THE
GROWTH-POVERTY LINK”
Overview of website and contents
Useful growth-poverty tools

Website: Measuring the Growth-Poverty Link
(http://go.worldbank.org/J70VTQSAK0)


Purpose: Make tools that explore growth-poverty links more
accessible and results easier to understand
5 existing tools to explore growth, distribution, and poverty





Growth elasticity of poverty
Growth incidence curve
Rate of pro-poor growth
Growth-Inequality decomposition of poverty
Sectoral decomposition of poverty
Overview of each tool on website



Definitions and Concepts
Limitations and Extensions
Quick Results




Annotated examples



Data requirements
Stata/ ADePT options
Helpful tips
Stata commands
Interpretation of results
References / Related Papers
FIVE TOOLS
With examples from Uganda case
1. Growth elasticity of poverty

Indicates how effectively growth has
translated into poverty reduction.
% _ change _ in _ poverty

% _ change _ in _ pcGDP

Misnomer:


Should be GDP elasticity of poverty
Initial conditions matter:


Location of poverty line (initial poverty levels)
Shape of the distribution (initial inequality)
Uganda: Growth elasticity of poverty
1993
2003
2006
0.56
0.39
0.31
270,267
375,829
399,978
Gini
0.37
0.43
0.41
Percent change
1993-2003
2003-2006
in poverty headcount
-31.2%
-19.8%
in per capita GDP
39.1%
6.4%
Growth elasticity of poverty
-0.8
-3.1
Percentage point change in poverty headcount
-0.18
-0.08
Growth semi-elasticity of poverty
-0.5
-1.2
Poverty headcount
Per capita GDP (constant LCU)
2. Growth incidence curves



Illustrates growth rate of income (expenditure)
for each percentile of a distribution.
Gives equal weight to people…rather than to
dollars
Refers to anonymous percentiles

Individual at 10th percentile at t0 is not necessarily
same individual at 10th percentile at t1
Uganda: GICs
1992-2002
Growth rate in mean
Mean percentile growth rate
Headcount poverty (1992)
Rate of pro-poor growth
2002-2005
=4.09
=3.26
=56.43
=2.90
Growth rate in mean
Mean percentile growth rate
Headcount poverty (2002)
Rate of pro-poor growth
=3.61
=4.73
=38.82
=4.44
Growth incidence curves -- example
Pov line (2005/06)
8
7
Decreasing inequality
6
5
4
Distribution neutral
3
2
Increasing inequality
1
0
0
20
31
40
60
80
Income percentiles - Poorest to Richest
100
3. Rate of pro-poor growth

Represents the mean growth rate of the poor



Not to be confused with growth rate in the mean
of the poor
Related to GIC: Area under GIC up to poverty
line (also equals the change in the Watts index)
General definition:
P.R.(actual)
RPPG 
* Growth _ Rate
P.R.(if _ distributionally _ neutral )
<
4.Growth-inequality decomposition
Quantifies the relative contribution of economic growth and
redistribution to changes in poverty.
Pt n  Pt 0
Change
in poverty
=
G (t0 , tn ; r )
Growth
component
+ D(t 0 , t n ; r )
Redistribution
component
+ R(t0 , tn ; r )
Residual
Uganda: Growth-inequality decomp.
Density
56.427
38.819
b) Poverty
rate (P0)
-----------------------------------------------------------------------------17.608
-17.608
-17.608
c) Change in
P0
-----------------------------------------------------------------------------25.134
-26.211
-25.672
d) Growth
component
----------------------------------------------------------------------------8.602
7.526
8.064
e)
Redistributio
n component
----------------------------------------------------------------------------f) Interaction
component
-1.076
-1.076
0.000
-----------------------------------------------------------------------------------------------------------------------------------
1992 as reference (base year 1)
.6
g) Average
effect
-----------------------------------------------------------------------------
.4
Base year 2
.2
Base year 1
0
a)
.8
Uganda: 1992-2002
-4
-2
0
2
ln(per capita expenditure/poverty line)
4
1992
2002
growth with 1992 distribution constant
redistribution with 1992 mean constant
kernel = epanechnikov, bandwidth = 0.0880
6
5. Sectoral decomposition of poverty

Quantifies relative contributions to changes in
aggregate poverty of:


changes in poverty within sectors and
inter-sectoral population shifts
Pt n  Pt 0
Change
in poverty

= k (st k )( Pt k  Pt k ) +  (st k  st k )( Pt k ) +  (st k  st k )( Pt k  Pt k )
0
n
0
n
0
Intra-sectoral
component
n
0
k
k
Inter-sectoral
component
Interaction
component
Typical sectors for decomposition:



Urban/rural
Regions
Economic sectors
0
n
0
Urban-Rural Sectoral Decomposition
(Uganda 1992—2002)
Rural
Urban
Total
Pop share (1) Pop share (2)
0.8758
0.8624
0.1242
0.1376
1.0
1.0
Poverty (1)
60.35
28.77
56.43
Rural
Urban
TOTAL intra-sectoral
population shift
interaction
-15.4404
-1.79096
-17.2313
-0.42317
0.043014
87.7%
10.2%
97.8%
2.4%
-0.2%
Total change (Headcount)
-17.6115
100.0%
Poverty (2)
42.72
14.35
38.82
Uganda: Rural / urban decomposition (1992
– 2002)
a) Poverty in 1992 (headcount)
56.4
b) Poverty in 2002 (headcount)
38.8
Sector
Pop’n share
(period 1)
Absolute
change
Contribution
(%)
c) Rural
87.6
-15.5
87.7
d) Urban
12.4
-1.8
10.2
e) Total intra-sectoral
-17.2
97.8
f) Population shift effect
-0.4
2.4
g) Interaction effect
0.04
-0.2
h) Change in poverty (HC)
-17.6
100.0
Summary



Website: Measuring the Growth-Poverty Link
(http://go.worldbank.org/J70VTQSAK0)
These tools provide an initial look beyond
averages at the poverty and distributional
impacts of growth.
However, integration with growth story is
necessary to get a fuller economic picture.