hosted by lse works: case the relationship between ...€¦ · policy needs to simultaneously...
TRANSCRIPT
Hosted by LSE Works: CASE
The Relationship between Inequality and Poverty: mechanisms and policy optionsDr Eleni Karagiannaki Research Fellow, CASE, LSE
Hashtag for Twitter users: #LSEworks
Dr Abigail McKnightAssociate Professorial Research Fellow and Associate Director, CASE, LSEChris Goulden
Deputy Director, Policy and Research, Joseph Rowntree Foundation
Professor Stephen MachinChair, LSE
Dr Chiara MariottiInequality Policy Manager, Oxfam
Understanding the relationship between poverty and inequality
Eleni Karagiannaki and Abigail McKnightLSE Works Public Lecture
8 February 2017
Motivation• Well documented upward trend in inequality in high and middle income countries since 1970s; although trends are not uniform across countries
• Growing concern about harmful effects of inequality on societies including the role inequality played in the lead up to the financial crisis
• Recent shift in thinking away from the assumption that policy can successfully target poverty reduction in rich and middle income countries without addressing income inequalities
• Big players ‐World Bank, United Nations, World Economic Forum, OECD, Oxfam, etc – setting twin goals and outlining recommendations that policy needs to simultaneously tackle poverty and inequality in rich as well as poor countries
• … but knowledge and evidence gaps
CASE research programme
• Joseph Rowntree Foundation funded a three year programme of research which is part of a wider partnership with the LSE’s International Inequalities Institute – Improving the evidence base for understanding the links between inequalities and poverty
• Oxfam funded a rapid Review of the evidence on the relationship between economic inequality and poverty
Approach• Examining the conceptual basis
• Documenting measurement issues
• Extending the empirical evidence base
• Understanding the mechanisms
• Exploring potential policy responses
Measurement issues• Measures of income inequality and poverty are summary statistics calculated from the same distribution (household income), therefore we would expect these measures will be linked in a ‘mechanical’ sense
• The strength of the relationship between inequality and poverty will depend on the extent to which any inequality measure is sensitive to dispersion of income in the lower half of the income distribution
• Theoretically it is possible to have: (1) no relative income poverty (income < 60% median income) but high inequality (high concentration of income among a small group of very rich households); high relative income poverty but low inequality (very low dispersion of income above the median) but in practise this is not what we observe
• We are interested in identifying what mechanisms underlie distributions of income (eg) where there is high inequality/poverty versus low inequality/poverty
UK income distribution 2014/15
HBAI 2016, BHC
The empirical relationship between inequality and poverty in rich and middle income countries: Evidence from the EU Income and Living Conditions Database
Eleni KaragiannakiLSE works
February 2017
Plan of the presentation
Using comparative distributional statistics from the Eurostat Income and Living Conditions database I will present evidence on:
(1) the extent to which higher levels of inequality are associated with higher levels of poverty and
(2) whether increasing levels of income inequality across a number of European countries have been associated with increasing poverty.
Key findings• Levels of inequality and poverty are highly correlated
• This correlation is stronger for inequality measures that summarize the degree of inequality at the bottom of the distribution and stronger when poverty is measured by poverty rates than poverty gaps
• A positive (albeit slightly weaker) correlation is estimated examining the relationship between changes in inequality and changes in the incidence and the depth relative income poverty as well as between changes in inequality and in the incidence of anchored poverty
• Despite the positive correlation between poverty and inequality trends there is substantial degree of heterogeneity across countries in how poverty and inequality evolved over this period: there are countries where inequality and poverty have moved in different directions
I. Differences in the level of inequality and poverty across different European countries in
2014
Levels of income inequality and relative income poverty are strongly correlated
latvia
malta germany
spain
polandluxembourg
czechrepublic
france
unitedkingdom
greece
austria
portugal
finlanddenmark
hungary
iceland
lithuaniaitaly
estonia
cyprusswedenbelgium
slovakianorwaynetherlands
ireland
510
1520
25Rel
ative
pove
rty ri
sk (%
)
20 25 30 35 40Gini
latvia
maltagermany
spain
polandluxembourg
czechrepublic
france
unitedkingdom
greece
austria
portugal
finlanddenmark
hungary
iceland
lithuaniaitaly
estonia
cyprusswedenbelgium
slovakianorwaynetherlands
ireland
510
1520
25Rel
ative
pove
rty ri
sk (%
)
2 3 4 5 6P90:P10
latvia
maltagermany
spain
polandluxembourg
czechrepublic
france
unitedkingdom
greece
austria
portugal
finlanddenmark
hungary
iceland
lithuaniaitaly
estonia
cyprusswedenbelgium
slovakianorway netherlands
ireland
510
1520
25Rel
ative
pove
rty ri
sk (%
)
1.6 1.8 2 2.2 2.4P90:P50
latvia
malta germany
spain
polandluxembourg
czechrepublic
france
unitedkingdom
greece
austria
portugal
finlanddenmark
hungary
iceland
lithuania italyestonia
cyprusswedenbelgium
slovakianorwaynetherlands
ireland
510
1520
25Rel
ative
pove
rty ri
sk (%
)
1.6 1.8 2 2.2 2.4 2.6P50:P10
r=0.87*** r=0.95***
r=0.81*** r=0.94***
Inequality and relative income poverty risk in 2014 for 26 European countries
Levels of income inequality also tend to be highly correlated with the depth of income poverty but the relationship is weaker
slovakia
bulgaria
croatia
romania
austria
latvia
spain
denmarkiceland
belgium
portugal
germany
czechrepublic
hungaryunitedkingdom
netherlands
finlandfrance
malta
fyrom
swedennorwaypoland
serbia
lithuania
luxembourg
greece
estonia
irelandcyprus
italy
1020
3040
Pov
erty
gap
ratio
20 25 30 35 40Gini
slovakia
bulgaria
croatia
romania
austrialatvia
spain
denmarkiceland
belgium
portugal
germany
czechrepublic
hungaryunitedkingdom
netherlandsfinland
francemalta
fyrom
swedennorwaypoland
serbia
lithuania
luxembourg
greece
estonia
irelandcyprus
italy
1020
3040
50Pov
erty
gap
ratio
3 4 5 6 7P90:P10
slovakia
bulgaria
croatia
romania
austria
latvia
spain
denmarkiceland
belgium
portugal
germany
czechrepublic
hungary
unitedkingdom
netherlands
finland
francemalta
fyrom
swedennorwaypoland
serbia
lithuania
luxembourg
greece
estonia
irelandcyprus
italy
1520
2530
3540
Pov
erty
gap
ratio
1.6 1.8 2 2.2 2.4P90:P50
slovakia
bulgaria
croatia
romania
austrialatvia
spain
denmarkiceland
belgium
portugal
germany
czechrepublic
hungaryunitedkingdom
netherlandsfinland
francemalta
fyrom
swedennorwaypoland
serbia
lithuania
luxembourg
greece
estonia
irelandcyprus
italy
1020
3040
50Pov
erty
gap
ratio
1.5 2 2.5 3 3.5P50:P10
r=0.54** r=0.71**
r=0.43* r=0.82***
Inequality and poverty gap ratio in 2014 for 26 European countries
II. Changes in the level of inequality and poverty across European countries in 2005‐14
Changes in income inequality are positively correlated with changes in the incidence of relative income poverty
swedendenmark
iceland
germanyslovakia
austria
finlandczechrepublicmalta
netherlands
luxembourg
norway belgiumfrance
hungary
cyprus
ireland
spainitaly
greece
estoniaunitedkingdom
polandlithuania
latvia
portugal
-.6-.4
-.20
.2.4
% c
hang
e in
pov
erty
risk
-.2 -.1 0 .1 .2% change in Gini
swedendenmark
iceland
germanyslovakia
austria
finlandczechrepublicmalta
netherlands
luxembourg
norwaybelgium
france
hungary
cyprus
ireland
spainitaly
greece
estoniaunitedkingdom
polandlithuania
latvia
portugal
-.50
.5%
cha
nge
in p
over
ty ri
sk
-.3 -.2 -.1 0 .1 .2% change in P90:P10
swedendenmark
iceland
germanyslovakia
austria
finlandczechrepublicmalta
netherlands
luxembourg
norwaybelgium
france
hungary
cyprus
ireland
spainitaly
greece
estoniaunitedkingdom
polandlithuania
latvia
portugal
-.50
.5%
cha
nge
in p
over
ty ri
sk
-.15 -.1 -.05 0 .05 .1% change in P90:P50
swedendenmark
iceland
germanyslovakia
austria
finlandczechrepublic
malta
netherlands
luxembourg
norwaybelgium
france
hungary
cyprus
ireland
spainitaly
greece
estoniaunitedkingdom
polandlithuania
latvia
portugal
-.4-.2
0.2
.4%
cha
nge
in p
over
ty ri
sk
-.2 -.1 0 .1 .2% change in P50:P10
r= 0.55** r= 0.74***
r= 0.26 r= 0.81***
% change in inequality and relative poverty risk
Changes in income inequality are also positively correlated with changes in the depth of poverty
% change in inequality and in the poverty gap ratio
sweden
denmark
iceland
germany
slovakia
austriafinland
czechrepublic
maltanetherlands
luxembourg
norway
belgium france
hungary
cyprus
ireland
spain
italy
greeceestonia
unitedkingdompoland
lithuania
latvia
portugal
-.20
.2.4
.6%
cha
nge
in p
over
ty g
ap ra
tio
-.2 -.1 0 .1 .2% change in Gini
sweden
denmark
iceland
germany
slovakia
austriafinland
czechrepublic
maltanetherlands
luxembourg
norway
belgiumfrance
hungary
cyprus
ireland
spain
italy
greeceestonia
unitedkingdompoland
lithuania
latvia
portugal
-.4-.2
0.2
.4%
cha
nge
in re
lativ
e po
verty
risk
-.3 -.2 -.1 0 .1 .2% change in P90:P10
sweden
denmark
iceland
germany
slovakia
austriafinland
czechrepublic
maltanetherlands
luxembourg
norway
belgiumfrance
hungary
cyprus
ireland
spain
italy
greeceestonia
unitedkingdompoland
lithuania
latvia
portugal
-.20
.2.4
.6%
cha
nge
in p
over
ty g
ap ra
tio
-.15 -.1 -.05 0 .05 .1% change in P90:P50
sweden
denmark
iceland
germany
slovakia
austriafinland
czechrepublic
maltanetherlands
luxembourg
norway
belgiumfrance
hungary
cyprus
ireland
spain
italy
greeceestonia
unitedkingdompoland
lithuania
latvia
portugal-.4
-.20
.2.4
% c
hang
e in
pov
erty
gap
ratio
-.2 -.1 0 .1 .2% change in P50:P10
r=0.42 r=0.66***
r=0.15 r=0.77***
A positive (though weaker) correlation is also estimated when the poverty line is anchored at 2005 levels
swedendenmarkiceland
germany
slovakia
austria
finland
czechrepublic malta
netherlands
luxembourg
norway
belgiumhungary
cyprus
irelandspain
italy
greece
estonia
unitedkingdom
poland lithuania
latvia
portugal
-1.5
-1-.5
0.5
1%
cha
nge
in a
ncho
red
pove
rty ra
te
-.2 -.1 0 .1 .2% change in Gini
swedendenmarkiceland
germany
slovakia
austria
finland
czechrepublicmalta
netherlands
luxembourg
norway
belgiumhungary
cyprus
irelandspain
italy
greece
estonia
unitedkingdom
poland lithuania
latvia
portugal
-2-1
01
% c
hang
e in
anc
hore
d po
verty
-.3 -.2 -.1 0 .1 .2% change in P90:P10
swedendenmarkiceland
germany
slovakia
austria
finland
czechrepublicmalta
netherlands
luxembourg
norway
belgiumhungary
cyprus
irelandspain
italy
greece
estonia
unitedkingdom
poland lithuania
latvia
portugal
-1.5
-1-.5
0.5
1%
cha
nge
in a
ncho
red
pove
rty
-.15 -.1 -.05 0 .05 .1% change in P90:P50
swedendenmarkiceland
germany
slovakia
austria
finland
czechrepublicmalta
netherlands
luxembourg
norway
belgiumhungary
cyprus
irelandspain
italy
greece
estonia
unitedkingdom
poland lithuania
latvia
portugal
-1.5
-1-.5
0.5
1%
cha
nge
in a
bs. p
over
ty
-.2 -.1 0 .1 .2% change in P50:P10
r=0.41* r=0.51***
r=0.19 r=0.55**
% change in inequality and in the anchored poverty risk
Conclusions • Levels of income inequality and poverty display a very strong positive correlation
• This positive correlation is stronger for inequality measures that summarize the degree of inequality at the bottom of the distribution and stronger when poverty is measured by poverty rates than poverty gaps
• A positive (albeit slightly weaker) correlation is estimated between changes in inequality and changes in the incidence and the depth relative income poverty as well as changes in the incidence of anchored poverty
• Despite the positive correlation between poverty and inequality trends, the analysis also identified the varying experiences across countries in how inequality and poverty evolved: there were countries inequality and poverty trends have moved in different directions
⇒ policy and institutions matter
Mechanisms
Mechanisms• Economic mechanisms
• Fundamental drivers – distribution of abilities and rates of return – acting through the labour market
• Resource constraints (Scale/the ‘race’ between the state and the rich)
• Political mechanisms• Self interest of rich and powerful elite• Reinforcement mechanisms
• Social and cultural mechanisms• Values, attitudes and beliefs• Fear – punitive (and impoverishing) reactions to crime
Policy plays a key role in ameliorating or exacerbating the extent to which these mechanisms relate inequality to poverty
Economic mechanisms – the labour marketSkill biased technological change, globalisation and weakening of labour market institutions are thought to be the main drivers behind changes in labour market inequality over the last few decades. These changes provide an explanation for both increases in income inequality and poverty risk.
• Demand shift in favour of high skilled workers and a weakening in the wage bargaining power of low skilled workers increases the risk of unemployment, low pay and precarious employment for lower skilled workers and increases wage inequality between skill levels;
• The relationship between individual labour market outcomes (pay/unemployment) and household level outcomes of income inequality and poverty risk is complex;
• Household formation, household composition, cash transfers and direct taxes all play a key role in defining this relationship.
Economic mechanism – the labour marketIndividual
y = ‐23.416x + 24.592R² = 0.073
0
5
10
15
20
25
30
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
At risk of p
overty ra
te (%
< 60%
med
ian income)
Individual gross annual earnings inequality (Gini)
y = ‐0.0718x + 0.3242R² = 0.0067
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Househ
old eq
uivalised
income
ineq
uality (Gini)
Individual gross annual earnings inequality (Gini)
32 EU‐SILC countries (2013) – working age population
Poverty
Ineq
uality
Economic mechanism – the labour marketHousehold
y = 88.579x ‐ 18.028R² = 0.5
0
5
10
15
20
25
30
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
At risk of p
overty ra
te (%
<60%
med
ian income)
Household annual total gross earnings inequality (Gini)
y = 79.487x ‐ 13.962R² = 0.6332
0
5
10
15
20
25
30
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5At risk of p
overty ra
te (%
<60%
med
ian
income)
Household equivalised annual gross earnings inequality (Gini)
32 EU‐SILC countries (2013) – working age population
Poverty
Poverty
Economic mechanism – the labour marketHousehold equivalised
y = 0.8694x ‐ 0.0265R² = 0.7376
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Hou
seho
ld equ
ivalised
income ineq
uality (Gini)
Household annual equivalised earnings inequality (Gini)
Less redistribution
Ireland
Slovenia
Bulgaria
32 EU‐SILC countries (2013) – working age population
More redistribution
Political mechanisms• Rise of rich and powerful elite (Stiglitz, 2012; Gilens and Page, 2014; Piketty, 2014)Influence government policy (opportunity hoarding and the role of donors to political campaigns and political parties)Lower income individuals withdraw from the voting boothsPolitical parties focus on policies that favour the voting electorate (median voter has higher than median income)
• Electoral systems and their propensity for redistribution (Iversen and Soskice, 2006)The electoral system in which individuals cast their votes plays a key role in shaping political parties, the composition of governing coalitions, and the likelihood of redistributionCentre‐right governments tend to dominate in majoritarian systems whereas centre‐left governments tend to dominate in PR systems
• “Winner‐take‐all politics into loser‐take‐all poverty and inequality” (Hacker and Pierson, 2010)
Social and Cultural mechanisms• Commonly held belief that inequality is too high but people tend to underestimate the true level of inequality – some evidence suggests that this is influenced by various types of segregation (geographical/media/jobs/social networks/schools)
• Political economy models (Meltzer and Richard, 1981) predict that in democracies an increase in inequality will lead to an increase in redistribution, but the literature shows that this prediction doesn’t always hold due to a number of factors shaping individuals redistributive preferences:
• Own income• Expectations of upward/downward mobility• Values and beliefs (why some people are poor/rich)• Under‐estimate of the level of inequality and overestimates of social mobility• Beliefs on the effectiveness/impact of certain policies (eg social security and work incentives)• Persistently high inequality can influence social norms
• Increases in income inequality correlated with increases in public punitiveness(Côté‐Lussier, 2016) and rates of incarceration. “Those with no capital get the punishment” (Sim, 2009)
Summary• Measurement: The statistical measures we commonly use to assess levels of inequality and poverty give rise to correlations between these two concepts;
• Empirical evidence: The positive correlation between income poverty and inequality is stronger for inequality measures that summarise the degree of inequality at the bottom of the distribution, and stronger for headcount measures than poverty gaps;
• Mechanisms: The literature identifies a number of mechanisms which help to explain the shape of the income distribution and why increases in inequality can lead to increases in poverty;
• Policy: The evidence suggests that tackling poverty without addressing inequality will be ineffective in the long‐run unless the mechanisms that link the two are broken.
Hosted by LSE Works: CASE
The Relationship between Inequality and Poverty: mechanisms and policy optionsDr Eleni Karagiannaki Research Fellow, CASE, LSE
Hashtag for Twitter users: #LSEworks
Dr Abigail McKnightAssociate Professorial Research Fellow and Associate Director, CASE, LSEChris Goulden
Deputy Director, Policy and Research, Joseph Rowntree Foundation
Professor Stephen MachinChair, LSE
Dr Chiara MariottiInequality Policy Manager, Oxfam