Transcript Pro-poor growth? - The Economics Network

Growth and Poverty Reduction:
Pro-poor growth?
Growth and Poverty Reduction: Pro-poor growth?
Lecture Outline
(i)
What is pro-poor growth?
(ii) What are the Theoretical Under-pinning of Pro-Poor
Growth?
(iii) Methodology
(iv) Evidence of pro-poor growth?
(v) Policies for pro-poor growth: 1970s vs. present?
Growth and Poverty Reduction: Propoor growth?
(i) What is pro-poor growth?
Definition of Pro-Poor Growth:
“…growth that leads to significant reductions in (absolute)
poverty”
(OECD 2001, and UN 2000) – italics added in brackets.
Too broad for economists since what definition of poverty do
researchers use? Kraay (2004) makes this point in his World
Bank Working Paper No. 3225, “When is Growth Pro-Poor?”
Growth and Poverty Reduction: Propoor growth?
A basic idea from the works of White and Anderson
(2001) and Kakwani and Pernia (2000) is that any
increase in growth should benefit the poor more than
the rich.
This really is “inequality-reducing” growth rather than
pro-poor growth – is concerned with relative poverty.
The question is, “Should new growth benefit the
poor more, thus increasing their incomes and thus
reducing inequality, whilst the rest of society sees
little income improvement?”
Growth and Poverty Reduction: Propoor growth?
If, ‘Yes’ to this then could have the issue of national
income increasing by 5% but income of the poor
increasing by 7%: whereas there could be a possibility
of growth for the poor of 7% when national income
increased by 10%.
The poor in the second scenario are absolutely better off,
but are relatively worse-off compared to the non-poor:
In the first scenario the poor are absolutely worse-off
compared to scenario 2, but are relatively better off.
Growth and Poverty Reduction: Propoor growth?
(Q)
So which one is better?
(A) In poor countries better absolute improvements preferred
to relative improvements, at least initially…..
Problem with the above ‘inequality-reducing’ scenario is that
we do not know whether following an inequality-reducing
growth plan will result in lower growth or higher growth.
So we want to have improvements in both absolute levels of
income (absolute poverty tackled) and relative levels of income
(relative income of poor improves and income inequality
declines?).
Growth and Poverty Reduction: Propoor growth?
Formalising these issues has been undertaken by White and
Anderson (2001)
White and Anderson (2001): 3 definitions of pro-poor
growth
(1) The poor’s share of incremental income exceeds their
current share.
This means that the incremental increase in the level of income
to the poor>incremental increase in the level of income for all of
society,
Growth and Poverty Reduction: Propoor growth?
p
p
(Yt  Yt 1) /(Yt  Yt 1)  t 1
Where the numerator represents change in income of
the poor, the denominator is the change in income of
society and t 1 represents the income share of the poor
in the last time period, t-1.
Growth and Poverty Reduction: Propoor growth?
So,
p
p
(Yt  Yt 1) /(Yt  Yt 1)  t , and thus t t1
There is a relative improvement of the poor.
This can be shown diagrammatically,
Growth and Poverty Reduction: Pro-poor growth?
Changes in Y, between t
and t-1
change in Y for non-poor
between t-1 and t
Y, T growth
change in Y for the poor
between t-1 and t:
If t  t 1 then pro-poor
growth.
t-1
t
Time
Growth and Poverty Reduction: Propoor growth?
The problem with definition (1) is that the poor’s share
can increase slightly and the richest 10% or 15% can
still cream off much more and this is pro-poor
growth.
E.g. If poorest 20% have 5% of income and the richest
20% have 40% of income then if in every extra $1 the
poorest get 10cents and the richest 30cents then still have
PPG.
If in the above t  t 1 then White and Anderson
(2001, pp. 269) coin the phrase anti-poor growth (APG).
Growth and Poverty Reduction: Propoor growth?
(2) The second definition brings into the equation the share of
people in a country who are defined as ‘poor’. Formally,
p
p
o
p
(Yt  Yt 1) /(Yt  Yt 1)  P  N / N
Where the P term represents the share of poor
people/households in the country.
What the equation is saying is that the increase in the share of the
poor’s income>share of the number of poor in the country’s
population.
Growth and Poverty Reduction: Propoor growth?
Can be re-arranged so that,
p
p
(Yt / N p )  (Yt / N )  (Yt 1 / N p )  (Yt 1 / N )
These represent average income levels of the poor and of
society, and can be further simplified to,
p
p
Yt  Yt  Yt 1  Yt 1, or
Y p  Y
Growth and Poverty Reduction: Propoor growth?
The change in average income of poor>change
in average income of society.
Problem with definition 2 is that it is too
restrictive and under this definition few
countries actually have pro-poor growth.
Growth and Poverty Reduction: Propoor growth?
Definition 3
Take an ‘international’ norm of median income shares
of the bottom 20% and 40% (can choose any %).
Issue here is that “if the poor’s share currently exceeds
the international norm then their share of incremental
income can be less than their current share and thus
qualify as PPG” (ibid, pp. 269).
Growth and Poverty Reduction: Propoor growth?
Another Definition of Pro-Poor Growth:
“…focuses on accelerating the rate of income growth of the poor and thus increase the rate of poverty
reduction”
(Ravallion and Chen, 2003)
Pro-Poor Growth = F(GDP growth)
Changes in income equality have an ambiguous effect on pro-poor growth since can
impact on GDP growth.
Thus, if pro-poor growth is to accelerate then need to accelerate growth but also need
to enhance and make poor households aware of the opportunities growth generates.
Hence there is no one agreed definition of what PPG actually is….hence a huge debate
as to whether PPG has occurred or not!!
Growth and Poverty Reduction: Pro-poor
growth?
(ii) What are the Theoretical Under-pinning of Pro-Poor
Growth?
Gunnar Myrdal in 1920s and 1930s India argued that initial
income inequality was an important factor in improving the
quality of people and hence productivity.
Ravallion and Datt (2002) report larger absolute povertyhousehold income elasticities in countries with lower gini
indices.
%PHH / %YHH  3 if giniindexis 0.25.
%PHH / %YHH  1if giniindexis 0.6.
Growth and Poverty Reduction: Pro-poor
growth?
(ii) What are the Theoretical Under-pinning of Pro-Poor Growth?
The idea here is that any growth that does occur is likely to benefit more
people if income inequality is low in the first place.
(Q)
So what then determines income inequality?
Assets – particularly land in LDCs
Education
Networks
Rural-Urban
Property Rights, Legal System
Growth and Poverty Reduction: Propoor growth?
(ii)
What are the Theoretical Under-pinnings of Pro-Poor Growth? Cont…
The 1970s…
In the 1970s the ‘Redistribution with Growth’ development economists believed in the
inverted-U hypothesis of Kuznets.
Kuznets inverted-U of growth and inequality:
Stage 1:
low per capita income level, low income inequality.
Stage 2:
rises.
per capita income increases with development, income inequality
Stage 3:
gets to a point where per capita income increases with continued
development, income inequality declines.
Growth and Poverty Reduction: Propoor growth?
Growth
Growth
Leads to
Inequality
Evidence from Ahluwalia (1976) supports the Kuznets hypothesis,
by simply regressing inequality onto income and income-squared:
Inequality  1Y   2Y 2
He found that as economic growth increased so income inequality
increased (1  ve ) but at a decreasing rate (  2  ve ): however
unsure where the turning point is!
Growth and Poverty Reduction: Propoor growth?
The Redistribution with Growth economists
argued (in line with Kaldor’s growth model) that
inequality caused growth since the rich had a
higher marginal propensity to save.
Also Lewis’s model of economic development
with unlimited labour supply was consistent with
rising income inequality through profits of
entrepreneurs growing more quickly than wages.
Growth and Poverty Reduction: Propoor growth?
Empirical Rejection of Kuznets, 1980s-1990s…
Rejected in Bruno, Ravallion and Squire, (1996) since much of
the empirical evidence of the 1970s and 1980s was flawed, and
studies were actually capturing between-country effects not withincountry effects.
By using panel estimates country-specific effects, time effects
and joint country and time effects are captured and the Kuznets
U-shaped curve disappears.
Studies find that between-country effects are causing the
inverted-U shape and that for some countries (e.g. India) the
relationship between inequality and income is simply U-shaped:
Inequality high, then low, then high again as growth increases.
Growth and Poverty Reduction: Propoor growth?
Currently whilst the theoretical debate continues as to whether economic
growth causes income inequality to change, the empirical evidence is stacking
up against a correlation in the first place.
E.g. Deininger and Squire (1996), Chen and Ravallion (1997), Easterly (1999),
Dollar and Kraay (2002) and Deaton (2005). According to Fields (1989,
2001),
Method should be looking at country-specific analysis – when Fields (1989)
looked at 70 growth spells across 20 countries he found that inequality rose in
10 countries, decreased in 11 and remained unchanged in 1.
“…income inequality increased in about half the growth spells and declined in the other
half.”
(World Bank, 2005, pp.17)
Growth and Poverty Reduction: Pro-poor growth?
New Theories of Pro-Poor Growth?
High Initial Inequality
Low Growth
Initial income inequality feeds into poor growth or greater
income equality positively affects growth rates.
Based largely on the conflicting progress of East Asia (e.g. South
Korea and Taiwan) and Latin America.
Latin America: notoriously unequal in income distribution (Brazil
regularly found to have the highest Gini coefficient).
Growth and Poverty Reduction: Pro-poor growth?
This means that any growth benefits the rich only and
tends to be skills-biased and capital intensive, thus the
poor have no chance of getting a piece of the
expanding pie: relative poverty increasing.
The income inequality also means that consumption is
relatively low since the rich have low MPC which
negatively effects AD and growth.
Growth and Poverty Reduction: Pro-poor growth?
East-Asia growth of the 1960s and beyond has seen no conflict between growth
and income distribution, meaning income inequality remains constant as growth increased.
(Q) Is there a theory that can explain this?
(1) The mechanism given for this ‘income inequality-neutral’ path is that low
initial income inequality results in more evenly distributed economic growth.
The reason is that consumption expenditure patterns are similar amongst the poor
for goods which they themselves produce hence generating demand for these
labour-intensive products: hence mass consumption takes off rather than
consumption being driven by the minority.
(2) Also the case that savings of the poor can be channelled if appropriate
investment opportunities are in evidence.
Growth and Poverty Reduction: Pro-poor growth?
(iii) Methodology
Dominant Method - Cross-Country Studies (Time
Series and Panels).
Has growth affected absolute poverty, relative poverty
and been pro-poor or not.
Growth and Poverty Reduction: Propoor growth?
Method I – Cross-Country Studies
White and Anderson (2001): find a negative relationship
between growth and income growth of the poor: i.e.
growth negatively effects the portion/share of income the
poorest of the population get.
Growth and Poverty Reduction: Propoor growth?
Back to the pro-poor question and for this see White and Anderson (2001)
Regress changes in income of the poor as a share of changes in total income of country,
p
p
(Yt  Yt 1) /(Yt  Yt 1)
and changes in share of income of the poor,
p
p
(Yt / Yt )  (Yt 1 / Yt 1)
onto a number of regressors that include, change in GDP per capita, change in trade openness dummy,
change government expenditure as share of GDP, change in political rights and civil liberties, and change
in life expectancy.
Why 2 dependent variables? Dependent variable 1 can be affected by outliers represented by large
changes in incomes of the poorest groups when total income for the country has increased. Changes in
the poor’s share of income gets around this issue.
Growth and Poverty Reduction: Propoor growth?
White and Anderson (2001) – (cont…)
Find that variations in the poor’s share of incremental income
(Dependent variable 1) is very large for growth
rates<4%..........some incremental shares are negative and very
large.
Confirms our expectations.
However using dependent variable 2 find that,
Growth and Poverty Reduction: Pro-poor growth?
Regression results for Change in Share of Income (poorest 40% and poorest
20% - t-tests in brackets
DQ20
DQ40
Constant
Growth (if improves)
Change in Political Freedom (if
worsens)
Trade Openness
Change in Urban (more
urbanisation)
Change in life expectancy (greater
life expectancy)
-0.001
0.000
(-0.76)
(-0.40)
-0.056
-0.011
(-2.03)
(-1.03)
-0.006
-0.002
(-1.97)
(-1.53)
-0.001
-0.001
(-0.94)
(-1.30)
0.003
0.002
(2.41)
(3.41)
-0.006
-0.002
(-1.85)
(-1.57)
Growth and Poverty Reduction: Propoor growth?
Note:
(1) Positive coefficient means an increase in the explanatory variable will
have a positive impact on the poor’s share of GDP, i.e. improve the poor’s
share of GDP
Find that growth negatively impacts on the poor – only the poorest 40%
though. Implication is there is a trade-off between growth and distribution
which contradicts World Bank thinking.
Less political freedom bad for the poor.
Increased urbanisation good for the poor
Growth and Poverty Reduction: Propoor growth?
Dollar and Kraay (2002, 2004): find a positive relationship between growth and
growth of incomes of the poor.
The Model
The key thing is the coefficient on yct . This represents the elasticity of income
of the poorest quintile with respect to mean income.
Control for 4 policy interventions that are likely to positively contribute to PPG: (1)
primary educational attainment (2) public spending on health and education, (3)
labour productivity in agriculture relative to rest of economy and (4) formal
democratic institutions.
Find none of these factors impact on PPG in their cross-country survey.
Growth and Poverty Reduction: Pro-poor
growth?
That the coefficient on log GDP of the country is
NOT significantly different from 1…..NO evidence
that % change in GDP is different from % change in
GDP of poorest 20%.
They cannot “reject the hypothesis that incomes of the
poor on average rise equiproportionately with average
incomes” (pp. 198), See next slide for Table 5 that
confirms this.
Growth and Poverty Reduction: Pro-poor growth?
Growth and Poverty Reduction: Pro-poor growth?
Figure 4, Dollar and Kraay (2004).
Growth and Poverty Reduction: Propoor growth?
Finds that changes in inequality and changes in income
are not correlated.
(Word of warning given by authors and by others is that crosscountry comparisons are subject to a lot of measurement error and
that country-specific studies are required for a clearer picture –
What the World Bank has done).
Dollar and Kraay (2004), Economic Journal paper in
special edition of linkages between trade, development
and poverty.
Growth and Poverty Reduction: Propoor growth?
Whilst not explicitly looking at income inequality Dollar
and Kraay findings are consistent with headcount
poverty declining more in those countries with more
equitable distributions of income.
Example from Klasen (2003).
Since Dollar and Kraay are looking at proportionate
income changes of the average and the poorest quintile
then clearly if income inequality is lower in a country
then the proportionate increase in income will lead to a
greater reduction in absolute poverty.
Growth and Poverty Reduction: Pro-poor growth?
Barro (2000) tests the Kuznets hypothesis again with a panel
of countries.
Regresses growth rate per capita output onto variables that
theoretically are predicted to determine growth, e.g. log of per
capita GDP, rule-of-law index, democracy (or freedom)
index, investment/GDP, years of schooling.
Then includes a gini coefficient index.
Growth and Poverty Reduction: Pro-poor growth?
Model 1
Model 2
Model 3
Growth rate is the dependent variable here.
When the gini index is introduced linearly there is no relationship with growth (0.000) – Model 1.
When the gini index is interacted with log(GDP) – a proxy for economic development – then we
see a negative relationship between income inequality and growth (-0.328) but that when
log(GDP) is higher the relationship is actually positive (0.043) – Model 2.
Implication is that at lower levels of log(GDP) income inequality does significantly impact on
growth rates.
Growth and Poverty Reduction: Pro-poor growth?
Barro also finds evidence that Kuznets curve is alive and
kicking. Income inequality first increases and then
decreases with economic development (log(GDP)).
However log(GDP) is not explaining
the majority of the variation in income
inequality across countries or across
time.
Growth and Poverty Reduction: Pro-poor growth?
Summary
Still no consensus.
Some find that changes in income are not correlated with changes in inequality means
that any growth in income does not appear to have any impact on inequality and
impacts positively on absolute poverty, (see Fields, 1989 and 2001).
Dollar and Kraay (2002) find that growth of the country positively effects growth of
income of the poor – implication is that growth is thus crucial for reducing absolute
poverty and the number of absolutely poor.
White and Anderson (2001) find that when dependent variable takes a ‘relative’ form
that growth negatively effects the share the poorest in a country have of GDP.
Barro (2000) finds that income inequality can impact on growth rates of very poor
countries only.
Growth and Poverty Reduction: Propoor growth?
(V)
Policies for pro-poor growth?
Killick (2002) mentions a wish list that would enhance pro-poor growth:
(1) Land reform – NOT land grab.
(2) Improved access by the poor to education and health - public sector, or publicprivate partnership?
(3) Micro-credit schemes targeted on the poor – charity or public sector since is too
risky for private sector to invest in!
(4) Adoption of labour intensive techniques in production
(5) Agricultural and rural development – very broad
(6) Government expenditure on education and health - public sector?
(7) Avoidance of macroeconomic crises – external factors that cannot be avoided, e.g.
sustained food price increases caused by emerging economies changing their diet.
(8) Investment in rural infrastructure – public sector, or public-private partnership?
(9) Labour-intensive industrialisation
Growth and Poverty Reduction: Propoor growth?
World Bank Report, pp.74.
For Agriculture:
–
–
–
–
–
Investments in infrastructure to connect the poor, e.g.
telecommunications, roads, public transport.
Strengthen property rights notably of women particularly regarding
land
Create incentive frameworks that do NOT discriminate against those
economic activities the poor are already undertaking
Improve technology for food-producers so can protect crop. Essential
given urban food demand increases.
Help poor households reduce and cope with risk which could
encourage greater risk with more high-yielding crops – (Q) Are poor
households risk-averse in gambling when times are good and risktaking when times are bad?
Growth and Poverty Reduction: Propoor growth?
•
•
•
•
For Non-agricultural poor:
Designing labour market rules and regulations that “balances workers’
needs with employers’ needs”, (ibid, pp. 75) – is an issue in many Latin
American countries where trade unions are strong, also the case in South
Africa.
Access to secondary and girls’ education important for poor households
given the growing skill bias in non-agricultural employment – “falling
fertility rates and rising female labor market participation is essential in a
pro-poor growth strategy” (ibid, pp. 75).
Quality of investment climate (assumed to be determined by macro and
trade environment, as well as degree of labour market regulation)
determine quantity and quality of employment.
Improved infrastructure for the poor.
Growth and Poverty Reduction: Propoor growth?
Factors that affect the success/failure of these World Bank policies
include:
•
•
•
•
•
Initial income and asset inequality – if high then can have negative impact
on pro-poor growth.
Must re-distribute…but how? Land reform not land grab, transfer payments in the form
of state pensions to the poor.
Importance of agriculture to the poor. Need to improve the efficiency of
agriculture (e.g. technology, co-operatives, training, access to financial
markets).
Since agriculture is affected by climactic conditions there needs to be a risk
management structure in place to protect poor farmers but which do
encourage more risk-taking – (Q) Are poor farmers likely to be risk-takers?
Delivery of services and capacity of institutions to deliver to the poor –
issue of corruption.
Growth and Poverty Reduction: Pro-poor growth?
References
Dollar, D., and Kraay, A., (2002), “Growth is Good for the Poor”, Journal of Economic Growth, Vol 7, pp. 195-225.
Fields, G., (1989) “Changes in Poverty and Inequality in Developing Countries”, World Bank Research Observer 4(2), 16785.
Fields, G., (2001) Distribution and Development: A New Look at the Developing World, MIT Press.
Kakwani and Pernia (2000) “What is pro-poor growth?”, Asian Development Review, Vol 18(1), 1-16.
OECD 2001 “Rising to the Global Challenge: Partnership for Reducing World Poverty.” Statement by the DAC High Level
Meeting, April 25-26, Paris
Ravallion and Chen (2003) “Measuring Pro-Poor Growth.” Economic Letters Vol 78(1), p3-99.
White and Anderson (2001) “Growth vs Redistribution: Does the pattern of growth matter?.”, Development Policy Review,
Vol 19(3), 167-289.
World Bank, (2005), Pro-Poor Growth in the 1990s Lessons and Insights from 14 Countries.