Politics of Redistribution & Inequality Flashcards Preview

CPE > Politics of Redistribution & Inequality > Flashcards

Flashcards in Politics of Redistribution & Inequality Deck (84)
Loading flashcards...
1

Evidence for relationship between partisanship and welfare state generosity?

Huber and Stephens (2001)

(i) Welfare generosity strongly positively associated w/cumulative left control

2

Evidence that income negatively associated with support for redistribution?

Finseraas (2009)

3

What is the Robin Hood paradox?

Perotti (1996)

Cross-sectional negative association between income inequality and redistribution (more equal countries redistribute more, not less)

4

Evidence for interaction between welfare state generosity and voter turnout?

Franzese (2002)

5

Empirical evidence that overall support for generous welfare states explained by correlation between labour market risk and low income levels?

Rehm, Hacker and Schlesinger (2012)

6

Partisanship determined to large extent by coalition dynamics associated with electoral systems

Iversen and Soskice (2006)

7

Evidence that PR electoral systems associated with different level and composition of public spending

Funk and Gathmann (2013)

SWISS CANTONAL ELECTORAL REFORM

1. Data:
(i) novel’ evidence of electoral reform in Swiss cantonal parliaments
(ii) Over last century or so, Swiss cantons at different times moved from plurality to PR electoral systems
(iii) Common history reduces problems of unquantifiable historical + institutional differences inherent in cross-sectional studies

2. Results:
(i) PR benefits left parties, leading to increased social spending on welfare and education
(ii) PR results in lower 'targeted' spending (e.g. roads)

8

Pre-existing strength of left-wing may have influenced initial choice of electoral system

Boix (1999)

9

Novel evidence that electoral success of left primary a consequence (not cause) of electoral reform?

Funk and Gatmann (2013)

1. Evidence from Swiss Cantonal parliaments
2. Over last century or so, Swiss cantons at different times moved from plurality to PR electoral systems
3. Common history reduces problems of unquantifiable historical + institutional differences inherent in cross-sectional studies

10

1. Right/centre-right controlled .....% of parliamentary representation in almost all ..... countries in last .... years and yet ....

2. Which study?

1. Right/centre-right controlled >50% of parliamentary representation in almost all PR countries in last 100 years and yet there have been almost no attempts to adopt majoritarian systems

2. Iversen, Soskice and Cusack (2007)

11

Empirical evidence that increased redistribution positively associated with income inequality

Kenworthy and Pontusson (2005)

Within-country association between growth of inequality and increases in redistribution 1975-2000

12

Within-country association between growth of inequality and increases in redistribution 1975-2000

Kenworthy and Pontusson (2005)

13

3 possible explanations for pattern that countries w/majoritarian electoral systems elect more right-wing governments? Which find empirical support?

Manow and Doring (2017):

1. Empirical support for A and B:

A. Voting behaviour (Iversen/Soskice argument)
B. Electoral geography (regional distribution of votes biases vote-seat translation against left in majoritarian systems, due to wide margins by which left wins urban districts)
C. Party fragmentation (if right more fragmented than left in PR countries, less chance of winning)

14

Evidence about changes preferences for redistribution in OECD in 80s/90s?

Kenworthy and McCall (2007)

1. Empirical evidence from 8 OECD countries in 80s/90s
2. Little/no change in preferences for redistribution

15

Funk and Gathmann (2013)

SWISS CANTONAL ELECTORAL REFORM

1. Data:
(i) novel’ evidence of electoral reform in Swiss cantonal parliaments
(ii) Over last century or so, Swiss cantons at different times moved from plurality to PR electoral systems
(iii) Common history reduces problems of unquantifiable historical + institutional differences inherent in cross-sectional studies

2. Results:
(i) Electoral success of left primary a consequence (not cause) of electoral reform
(ii) Pre-existing left strength didn’t predict electoral reform
(iii) PR benefits left parties, leading to increased spending on welfare and education
(iv) PR leads to lower 'targeted' spending (e.g. roads)
(v) No clear evidence that PR changes total overall level of government spending

16

Huber and Stephens (2001)

PARTISANSHIP AND WELFARE GENEROSITY

Welfare generosity strongly positively associated w/cumulative left control

17

Finseraas (2009)

INCOME AND REDISTRIBUTION PREFERENCES

Income negatively associated with support for redistribution

18

Franzese (2002)

TURNOUT AND WELFARE GENEROSITY

Positive interaction between welfare state generosity and voter turnout

19

Rehm, Hacker and Schlesinger (2012)

DETERMINANTS OF WELFARE STATE SUPPORT

1a. Overall support for generous welfare states explained by correlation between labour market risk and low income levels
1b. Strong empirical support cross-nationally using unemployment benefits and across policy domains within the USA

2. Strong empirical support for both income and unemployment risk as motivations for individual-level welfare state support

20

Boix (1999)

POTENTIAL ENDOGENEITY OF ELECTORAL INSTITUTIONS

Pre-existing strength of left-wing may have influenced initial choice of electoral system

21

Kenworthy and Pontusson (2005)

INEQUALITY AND REDISTRIBUTION PREFERENCES

1. Within-country positive association between growth of inequality and increases in redistribution 1975-2000 (due to 2 'automatic' mechanisms)

A. Increased unemployment:
(i) Example (Finland) - % of working-age households receiving unemployment benefit more than tripled from 80s to 90s
B. Concentrated gains at top + progressive tax systems

2. Low turnout = compelling reason why US welfare state much less responsive to rising market inequality compare to other countries

22

Manow and Doring (2017)

WHY ELECTORAL SYSTEMS PRODUCE LEFT/RIGHT BIAS

1. Empirical support for A and B:

A. Voting behaviour (Iversen/Soskice argument)
B. Electoral geography (regional distribution of votes biases vote-seat translation against left in majoritarian systems, due to wide margins by which left wins urban districts)
C. Party fragmentation (if right more fragmented than left in PR countries, less chance of winning)

23

Kenworthy and McCall (2007)

REDISTRIBUTION PREFERENCES

Empirical evidence from 8 OECD countries in 80s/90s

1. Common trend of significant increases in income inequality
2. Little/no change in preferences for redistribution
3. Perceptions of income inequality not significantly changed

24

Evidence of increasing income inequality in OECD

1. OECD avg. Gini coefficient increased ~10% from mid 80s to late 2000s (from ~0.29 to ~0.32)

2. Inequality increased in 17/22 countries, including in egalitarian Nordic countries

25

1. OECD avg. Gini coefficient increased ..... from the mid 80s to late 2000s (from ..... to .....)

2. Inequality increased in ..... out of ..... countries, including .....

1. OECD avg. Gini coefficient increased ~10% from mid 80s to late 2000s (from ~0.29 to ~0.32)

2. Inequality increased in 17 out of 22 countries, including in egalitarian Nordic countries

26

Meltzer and Richard (1981)

1. Decisive median voter prefers more redistribution as inequality increases
2. If income of median voter less than average, they have more to gain from redistribution

27

Mechanisms by which redistribution has increased due to increased inequality in recent decades, despite little/no change in preferences for redistribution

Kenworthy and Pontusson (2005)

1a. Increased unemployment:
(i) Shifts to post-industrial economic structures caused increased income inequality and unemployment
(ii) Increased unemployment led to sharp increase in claims of unemployment + related benefits
(iii) Such claims mostly made by low-income earners, resulting in ‘automatic’ increase in redistribution from rich to poor


1b. Concentrated gains at top:
(i) Increased income inequality largely due to concentrated gains at top
(ii) Progressive nature of OECD tax systems means this has led to automatic increase in redistribution

28

Evidence that income inequality largely driven by concentrated gains at top

Hacker and Pierson (2010)

In USA, from 70s to ~2000, % of income going to:
(i) Top 1% doubled
(ii) Top 0.1% quadrupled

29

Hacker and Pierson (2010)

In ....., from 70s to ~2000, the % of income going to:
(i) Top .....% ..... (from .....% to .....%)
(ii) Top .....% ...... (from .....% to .....%)

Hacker and Pierson (2010)


In USA, from 70s to ~2000, % of income going to:
(i) Top 1% doubled
(ii) Top 0.1% quadrupled

30

Gimpelson and Treisman (2017)

INEQUALITY MISPERCEPTIONS

1. Demand for redistribution varies not w/actual inequality, but perceptions of it (which are often wrong)

2. Individuals tend to place themselves in centre of income distribution

3. Given misperceptions about inequality and place in income distribution, unclear voters even know whether they would benefit from redistribution or not

31

Demand for redistribution varies not w/actual inequality, but perceptions of it (which are often wrong)

Gimpelson and Treisman (2017)

32

Individuals tend to place themselves in centre of income distribution

Gimpelson and Treisman (2017)

33

Given misperceptions about inequality and place in income distribution, unclear voters even know whether they would benefit from redistribution or not

Gimpelson and Treisman (2017)

34

Possible explanations of why voters' perceptions of inequality not significantly increased, despite common trend of increased market income inequality

1. Reference group theory (Gimpelson and Treisman 2017)

2. Perceptions based on post-tax/transfer inequality (Kenworthy and McCall 2007)

35

Verba et al (1995)

VOTING CORRELATED WITH INCOME

1. Probability of voting strongly correlated w/income

2. Low-income people much less likely to vote

36

1. Probability of voting strongly correlated w/income

2. Low-income people much less likely to vote

Verba et al (1995)

37

Example of policy 'drift' that has facilitated rising income inequality in the USA

Hacker and Pierson (2010)

1. non-decisions’ can have significant effects, like the continued taxation of mega-rich hedge-fund managers at low capital gains rates, due to out-of-date historic IRS rules

38

Hacker and Pierson (2010)

1. Distinctive pattern of winner-take-all inequality in the US, w/highly concentrated gains at top

2. Evidence – from 70s to ~2000, % of income going to:
(i) Top 1% doubled
(ii) Top 0.1% quadrupled

3. Organised interests – policy-makers respond to powerful organised interests, not disorganised voters

4. Policy ‘drift’ – ‘non-decisions’ can have significant effects, like the continued taxation of mega-rich hedge-fund managers at low capital gains rates, due to out-of-date historic IRS rules

39

..... and ..... (2005)

Example (Finland) - the % of working-age households receiving unemployment benefit ..... from ..... to .....

Kenworthy and Pontusson (2005)

Example (Finland) - the % of working-age households receiving unemployment benefit more than tripled from 80s to 90s

40

Rueda (2008)

EFFECT OF PARTISANSHIP MEDIATED BY INSTITUTIONS

3 key ways in which partisanship may influence wage inequality:
(A) Minimum wage legislation
(B) Welfare state generosity
(C) Public sector employment

A. Minimum wage not strongly linked to inequality when corporatism high because corporatist institutions constrain ability of governments to influence wage distribution (by making actions of social partners more influential on inequality)

B. Insider/outsider divide may cause left governments to protect insiders' interests, rather than increasing welfare generosity (as conventionally assumed)

C. Standard partisanship prediction that left governments promote public sector employment

41

Problem with total social spending as % of GDP as measure of welfare state generosity

1. Social spending might increase for reasons not related to generosity (e.g. recession, ageing population)
2. Example – social spending as % of GDP in Finland varied significantly in the 90s, but fluctuations largely due to economic crisis and unrelated to changes in generosity

42

Perotti (1996)

ROBIN HOOD PARADOX

1. Cross-sectional negative association between income inequality and redistribution (more equal countries redistribute more, not less)

43

Rehm (2011)

Individuals w/greater risk of losing job more likely to demand higher unemployment benefits

44

Individuals w/greater risk of losing job more likely to demand higher unemployment benefits

Rehm (2011)

45

Iversen and Soskice (2009)

1. Individuals w/greater labour market risk likely to support increased social spending and welfare generosity

2. Welfare states = insurance systems that accompany different skill regimes

46

Evidence that labour market risk related to welfare state support

1. Rehm (2011)
(i) Individuals w/greater risk of losing job more likely to demand higher unemployment benefits

2. Iversen and Soskice (2009)
(i) Individuals w/greater labour market risk likely to support increased social spending and welfare generosity

47

Clark, Golder and Golder (2009)

IMPACT OF ELECTORAL INSTITUTIONS

1. Initial choice may have been influenced by left strength
2. But this doesn't mean that now, since introduced, PR hasn't independently reinforced left strength

48

Lupu and Pontusson (2011)

STRUCTURE OF INCOME DISTRIBUTION AND REDISTRIBUTION PREFERENCES

1a. Social distance = key determinant of redistribution preferences
1b. Income = good proxy for social distance in absence of cross-cutting ethnic cleavages

2. Theory – social distance/affinity impacts ability to empathise with (and hence support for redistribution to) certain groups

3. Empirical evidence - middle-income voters empathise w/poor and support redistribution when income distance between middle/poor small relative to distance between middle/rich (w/o cross-cutting ethnic cleavages)

49

Evidence that structure of the income distribution influences redistribution preferences

Lupu and Pontusson (2011)

1a. Social distance = key determinant of redistribution preferences
1b. Income = good proxy for social distance in absence of cross-cutting ethnic cleavages

2. Theory – social distance/affinity impacts ability to empathise with (and hence support for redistribution to) certain groups

3. Empirical evidence - middle-income voters empathise w/poor and support redistribution when income distance between middle/poor small relative to distance between middle/rich (w/o cross-cutting ethnic cleavages)

50

Alesina and Glaeser (2004)

When minority ethnic groups perceived to be disproportionately poor, support for redistribution depressed

51

Evidence that perceptions of minority groups impact support for redistribution

Alesina and Glaeser (2004)

When minority ethnic groups perceived to be disproportionately poor, support for redistribution depressed

52

When minority ethnic groups perceived to be disproportionately poor, support for redistribution depressed

Alesina and Glaeser (2004)

53

Eger (2010)

Sweden:
1. Clear evidence that ethnic heterogeneity negatively affects support for social welfare spending
2. Recent immigration had negative impact on support for universal welfare spending

54

Evidence that racial heterogeneity impacts support for welfare states

Eger (2010)

Sweden - clear evidence that ethnic heterogeneity negatively affects support for social welfare spending

55

Dimick, Rueda and Stegmueller (2018)

ALTRUISM AND SUPPORT FOR REDISTRIBUTION

1. Test several approaches in which support for welfare/redistribution determined by altruism related to income inequality

2. Finding – support for ‘income-dependent altruism’ hypothesis, which predicts:
(1) as inequality level increases, both rich and poor have stronger pro-redistribution preferences
(2) rich more responsive to changes in inequality

3. Main determinant of welfare state support = individual utility and self-interested reasons for welfare state support remain most important

56

Beramendi and Rehm (2016)

Policy feedback - progressivity and redistribution preferences

1. Strength of association between income and support for redistribution varies significantly cross-nationally

2. Progressivity of tax/transfer system = major determinant of this variation
(a) more progressive - income more strongly predicts redistribution preferences because 'who gets what' conflict increases

57

Evidence for policy feedback as explanation of cross-national variation in welfare state/redistribution preferences

1. Beramendi and Rehm (2016)
(a) progressivity of tax/transfer system = major determinant of predictive power of income on redistribution preferences across different countries

2. Gingrich and Ansell (2012)
(a) individual risk much less important determinant of welfare state preferences in countries with:
(i) high employment protection
(ii) welfare benefits not dependent on employment status

58

Gingrich and Ansell (2012)

Policy feedback - employment protection/welfare benefits and individual risk

1. Individual risk much less important determinant of welfare state preferences in countries with:
(i) high employment protection
(ii) welfare benefits not dependent on employment status

59

Why do Beramendi and Rehm (2016) argue that cross-national variation in strength with which income predicts redistribution support explained by progressivity of tax system?

a) more progressive - income more strongly predicts redistribution preferences because 'who gets what' conflict increases

(b) less progressive - more overlap between tax contributions and receipts, so redistribution struggles less salient

60

Trumpy and Cavaille (2014)

MULTI-DIMENSIONALITY OF REDISTRIBUTION PREFERENCES

1. Dichotomy of self-interests vs altruistic motivations for redistribution attitudes = wrong

2. Strong support for multi-dimensional framework
(i) Self-interested motive – redistribution support as beneficiary ‘from rich’
(ii) Other-oriented motive – redistribution support as contributor ‘to poor’

61

Evidence for multi-dimensionality of redistribution preferences

1. Trumpy and Cavaille (2014)

A. Strong support for multi-dimensional framework
(i) Self-interested motive – redistribution support as beneficiary ‘from rich’
(ii) Other-oriented motive – redistribution support as contributor ‘to poor’

2. Rueda (2017)

A. Material self-interest = most important factor shaping poor's redistribution preferences (who consistently support redistribution)
B. Rich can 'afford to be altruistic' and support redistribution, conditional on identity of the poor (due to social identity theory, in/out groups)

62

Evidence that altruism explains some support for welfare states/redistribution

1. Dimick, Rueda and Stegmueller (2018)

A. Finding – support for ‘income-dependent altruism’ hypothesis, which predicts:
(i) as inequality level increases, both rich and poor have stronger pro-redistribution preferences
(ii) rich more responsive to changes in inequality
(iii) Main determinant of welfare state support = self-interest

2. Rueda (2017)

A. Material self-interest = most important factor shaping poor's redistribution preferences (who consistently support redistribution)
B. Rich can 'afford to be altruistic' and support redistribution, conditional on identity of the poor (due to social identity theory, in/out groups)

63

Is self-interest or altruism a stronger motivation for welfare state support?

Dimick, Rueda and Stegmueller (2018)

Main determinant of welfare state support = individual utility and self-interested reasons for welfare state support remain most important

64

Policy feedback - employment protection/welfare benefits and individual risk

1. Individual risk much less important determinant of welfare state preferences in countries with:
(i) high employment protection
(ii) welfare benefits not dependent on employment status

Gingrich and Ansell (2012)

65

Policy feedback - progressivity and redistribution preferences

1. Strength of association between income and support for redistribution varies significantly cross-nationally

2. Progressivity of tax/transfer system = major determinant of this variation
(a) more progressive - income more strongly predicts redistribution preferences because 'who gets what' conflict increases

Beramendi and Rehm (2016)

66

IMPACT OF ELECTORAL INSTITUTIONS

1. Initial choice may have been influenced by left strength
2. But this doesn't mean that now, since introduced, PR hasn't independently reinforced left strength

Clark, Golder and Golder (2009)

67

Individuals w/greater labour market risk likely to support increased social spending and welfare generosity

Iversen and Soskice (2009)

68

WHY ELECTORAL SYSTEMS PRODUCE LEFT/RIGHT BIAS

1. Empirical support for A and B:

A. Voting behaviour (Iversen/Soskice argument)
B. Electoral geography (regional distribution of votes biases vote-seat translation against left in majoritarian systems, due to wide margins by which left wins urban districts)
C. Party fragmentation (if right more fragmented than left in PR countries, less chance of winning)

Manow and Doing (2017)

69

Positive interaction between welfare state generosity and voter turnout

Franzese (2002)

70

SWISS CANTONAL ELECTORAL REFORM

1. Data:
(i) novel’ evidence of electoral reform in Swiss cantonal parliaments
(ii) Over last century or so, Swiss cantons at different times moved from plurality to PR electoral systems
(iii) Common history reduces problems of unquantifiable historical + institutional differences inherent in cross-sectional studies

2. Results:
(i) Electoral success of left primary a consequence (not cause) of electoral reform
(ii) Pre-existing left strength didn’t predict electoral reform
(iii) PR benefits left parties, leading to increased spending on welfare and education
(iv) PR leads to lower 'targeted' spending (e.g. roads)
(v) No clear evidence that PR changes total overall level of government spending

Funk and Gathmann (2013)

71

Evidence on impact of electoral system on overall level of government spending?

Mixed evidence

72

Why do Funk and Gathmann (2013) find that PR leads to lower 'targeted' spending on things like roads?

In PR systems w/national districts, geographically targeted spending makes no electoral sense

73

Why is 'targeted' spending (e.g. on roads) higher in majoritarian systems?

Given each district chooses 1 representative, makes electoral sense to pursue geographically targeted spending ('pork barrel')

74

Kwon and Pontusson (2003)

PARTISANSHIP AND REDISTRIBUTION

1. Government partisanship helps explain cross-national differences in redistribution

2a. Strength of partisan effects rose from 70s to early 90s, but then declined
2b. Explanation - constraints associated with integration of international capital markets and liberalisation of capital flows

75

PARTISANSHIP AND REDISTRIBUTION

1. Government partisanship helps explain cross-national differences in redistribution

2a. Strength of partisan effects rose from 70s to early 90s, but then declined
2b. Explanation - constraints associated with integration of international capital markets and liberalisation of capital flows

Kwon and Pontusson (2003)

76

Evidence that government partisanship helps explain cross-national differences in redistribution

Kwon and Pontusson (2003)

PARTISANSHIP AND REDISTRIBUTION

1. Government partisanship helps explain cross-national differences in redistribution

2a. Strength of partisan effects rose from 70s to early 90s, but then declined
2b. Explanation - constraints associated with integration of international capital markets and liberalisation of capital flows

77

Evidence of changing strength of effects of partisanship on redistribution over time

Kwon and Pontusson (2003)

PARTISANSHIP AND REDISTRIBUTION

1. Government partisanship helps explain cross-national differences in redistribution

2a. Strength of partisan effects rose from 70s to early 90s, but then declined
2b. Explanation - constraints associated with integration of international capital markets and liberalisation of capital flows

78

1. Counter-example to Iversen and Soskice's (2006) electoral incentives argument?

2. Explanation?

1. Germany - primarily centre-right governments, despite influential PR component in mixed electoral system

2. Role of Christian democracy

79

How do electoral rules influence economic outcomes?

Electoral systems influence:

1. Range of opinions represented in legislature
2. Governing parties
3. Party-voter linkages and legislators' incentives to cultivate a personal vote

80

Pontusson and Rueda (2010)

LEFT PARTIES ONLY RESPOND TO INCREASED INEQUALITY WHEN LOW-INCOME VOTERS POLITICALLY MOBILISED

1. Puzzle - common trend of increased income inequality, yet substantial cross-national differences in policy responses

2. Finding - when voter turnout above average, inequality associated with left parties adopting positions further to left, relative to median voter

3. Explanation:
(i) Core left voters generally = low-income earners
(ii) Higher inequality increases distance between income of these voters and average income
(iii) So greater inequality increases pressure from core constituency on left parties to redistribute more
(iv) Turnout among low-income voters significant factor shaping response of left parties
(v) Key finding - LEFT PARTIES ONLY RESPOND TO INCREASED INEQUALITY WHEN LOW-INCOME VOTERS POLITICALLY MOBILISED

81

LEFT PARTIES ONLY RESPOND TO INCREASED INEQUALITY WHEN LOW-INCOME VOTERS POLITICALLY MOBILISED

1. Puzzle - common trend of increased income inequality, yet substantial cross-national differences in policy responses

2. Finding - when voter turnout above average, inequality associated with left parties adopting positions further to left, relative to median voter

3. Explanation:
(i) Core left voters generally = low-income earners
(ii) Higher inequality increases distance between income of these voters and average income
(iii) So greater inequality increases pressure from core constituency on left parties to redistribute more
(iv) Turnout among low-income voters significant factor shaping response of left parties
(v) Key finding - LEFT PARTIES ONLY RESPOND TO INCREASED INEQUALITY WHEN LOW-INCOME VOTERS POLITICALLY MOBILISED

Pontusson and Rueda (2010)

82

Evidence that political mobilisation conditions policy response to rising inequality

Pontusson and Rueda (2010)

1. Puzzle - common trend of increased income inequality, yet substantial cross-national differences in policy responses

2. Finding - when voter turnout above average, inequality associated with left parties adopting positions further to left, relative to median voter

3. Explanation:
(i) Core left voters generally = low-income earners
(ii) Higher inequality increases distance between income of these voters and average income
(iii) So greater inequality increases pressure from core constituency on left parties to redistribute more
(iv) Turnout among low-income voters significant factor shaping response of left parties
(v) Key finding - LEFT PARTIES ONLY RESPOND TO INCREASED INEQUALITY WHEN LOW-INCOME VOTERS POLITICALLY MOBILISED

83

Rueda (2017)

MATERIAL-SELF INTEREST + OTHER-REGARDING REDISTRIBUTION PREFERENCES

1. Material self-interest = most important factor shaping poor's redistribution preferences (who consistently support redistribution)

2. Rich can 'afford to be altruistic' and support redistribution, conditional on identity of the poor (due to social identity theory, in/out groups)

3a. Group homogeneity magnifies importance of altruism for rich
3b. Group heterogeneity limits importance of altruism for rich

84

MATERIAL-SELF INTEREST + OTHER-REGARDING REDISTRIBUTION PREFERENCES

1. Material self-interest = most important factor shaping poor's redistribution preferences (who consistently support redistribution)

2. Rich can 'afford to be altruistic' and support redistribution, conditional on identity of the poor (due to social identity theory, in/out groups)

3a. Group homogeneity magnifies importance of altruism for rich
3b. Group heterogeneity limits importance of altruism for rich

Rueda (2017)