Diversity and Charitable Giving James Andreoni, UCSD, A. Abigail Payne, McMaster University Justin Smith, Wilfrid Laurier University David Karp (McMaster) 1
Facts and Issues Most Developed Countries are becoming more diverse Diversity represents many things: Skin colour, religion, culture, beliefs, income/education Key Issue: How do we measure diversity? What is the metric? Focus on visible minority & religious status What is the formula? Simply not shares of a population But the combination of those shares Idea is to identify the composition of one community versus another community How define a community? A city seems too big We use an area defined as forward sortation area : ~5000 households 2
Ethnic Diversity Winnipeg 2006 Kildonan North End Charleswood River Heights Grant Park Downtown Osborne Village Fort Rouge Kingston Crescent Transcona St. Boniface Community Community (Urban or Rural) Diversity Quintile No Data 1 2 3 4 5 Water St. Vital W N E 10 0 10 20 Kilometers S Change in Ethnic Diversity Winnipeg 2006 Kildonan North End Downtown Transcona Osborne Village River Heights Fort Rouge St. Boniface Charleswood Grant Park Kingston Crescent St. Vital Community Community (Urban or Rural) Diversity Change Quintile No Data 1 2 3 4 5 Water 10 0 10 20 Kilometers W N S E 3
Ethnic Diversity Golden Horseshoe 2006 Bolton Aurora King City Richmond Hill Markham Vaughan Whitby Oshawa Bowmanville Pickering Brampton Georgetown Toronto Mississauga Milton Oakville Burlington Hamilton Stoney Creek Grimsby Thorold Niagara Falls Community Major City (Population > 100,000) Minor City (Population < 100,000) Town or Village Diversity Quintile No Data 1 2 3 4 5 Water W N E 40 0 40 80 Kilometers S Change in Ethnic Diversity Golden Horseshoe 1996-2006 Bolton Aurora King City Richmond Hill Markham Vaughan Whitby Oshawa Pickering Bowmanville Brampton Georgetown Toronto Mississauga Milton Oakville Burlington Hamilton Stoney Creek Grimsby Thorold Niagara Falls Community Major City (Population > 100,000) Minor City (Population < 100,000) Town or Village Diversity Change Quintile No Data 1 2 3 4 5 Water W N E 4 40 0 40 80 Kilometers S
Motivation There are numerous positive benefits from having diverse communities But with respect to publicly funded goods Support for these goods may be diluted Alesina and La Ferrara (2005); Stichnoth and Van der Straeten (2009) Why? Basic idea: if preferences over goods differ It is more difficult to reach a consensus over the level of those goods 5
Previous research focusing on publicly provided goods Less on schools Alesina, Baqir, & Easterly (1999, 2000) Poterba (1997) Goldin & Katz (1999) Less on roads/hospitals Alesina et al (1999, 2000) Less on social programs Alesina, Glaeser and Sacerdote (2001) Okten and Osili (2004) Lower completion rates on census Vigdor (2004) 6
This Paper Does community diversity decrease support for private provision of public goods? Alesina and La Ferrara (2000): model of group affiliation Unclear prediction for private provision If people sort into groups w/ commonality (ethnicity, religion, etc) Might increase support If charities perceived to help ones outside of a group Might decrease support 7
Previous Research on Giving & Diversity Hungerman (2009) Spending by all-white churches More sensitive to increases in % black in a community Hungerman (2008) Government spending less likely to crowd out church spending by United Methodist churches in ethnically diverse communities Fong and Luttmer (2011) Giving to Hurricane Katrina victims higher when giver and recipient have same racial background Fong and Luttmer (2010) Participants viewed Blacks less worthy of charity But did not give less to blacks 8
This paper Can draw sharper conclusions than prior studies Data from individual tax returns Reported donations to charities Aggregated to neighbourhood level ~5000 households Captures average individual giving And community diversity Ethnicity: 1996-2006 More than just black/white Religion: 1991-2001 comprehensive Vigdor (2002, 2004) supports aggregation 9
Measures of Diversity After controlling for individual shares, other neighbourhood characteristics Focus on compositional effect: Fragmentation 10
Data & Measures Average donation per adult Total donations / adult population (15+) Fraction of households who donate Urban areas Source of most inflows Approximately 1,416 neighbourhoods 18,944 neighbourhood-year observations 11
Table III - Visible Minority and Religion Groups Over Time 1996 Percentage Point Change 2006-1996 % of FSAs with Increasing Share 1996-2006 Visible Minority Not Visible Minority 85.47% -5.64 12.50% East/Southeast Asian 6.43% 2.05 73.45% Chinese 3.97% 1.01 64.05% Filipino 1.07% 0.58 60.95% Southeast Asian 0.80% 0.17 55.93% Korean 0.30% 0.28 52.37% Japanese 0.30% 0.02 41.61% South Asian 3.02% 1.88 66.97% Black 2.58% 0.51 69.53% Other 1.40% 0.61 76.37% Latin American 0.82% 0.41 71.99% Other 0.59% 0.21 60.49% Arab/West Asian 1.10% 0.59 63.96% 12
1991 Percentage Point Change 2001-1991 % of FSAs with Increasing Share 1991-2001 Religion Catholic 45.05% -1.90 42.47% Christian 35.76% -4.68 17.15% United 10.26% -1.73 18.25% Anglican 7.91% -1.38 19.74% Other Christian 5.34% 0.38 56.93% Presbyterian 2.44% -1.05 6.98% Baptist 2.19% 0.05 53.24% Lutheran 2.29% -0.30 28.22% Orthodox 1.78% 0.10 49.25% Pentecostal 1.42% -0.31 29.81% Jehovah's Witness 0.60% -0.08 40.78% Mennonite 0.50% -0.08 25.12% Reform 0.38% -0.15 14.86% Mormon 0.33% -0.02 40.38% Salvation Army 0.33% -0.11 20.04% No Religion 13.72% 3.88 91.03% Jewish 1.67% -0.10 44.87% Muslim 1.26% 1.39 74.78% Other 2.54% 1.42 65.50% Buddhist 0.82% 0.53 67.40% Hindu 0.76% 0.55 51.25% Sikh 0.69% 0.44 39.78% Other 0.26% -0.11 22.23% 13
Summary Stats Panel A: Overall Mean SD 10 25 50 Average Donation 205.18 199.59 55.22 98.78 164.67 Fraction of Households Donating 0.474 0.141 0.300 0.375 0.463 Visible Minority Fragmentation Index 0.254 0.218 0.024 0.066 0.180 Religion Fragmentation Index 0.553 0.198 0.167 0.525 0.629 14
Regression: Total Donations (1) (2) (3) Fragmentation Index 876.53-138.49-271.02 (121.02) (65.03) (109.03) % Black -1308.54 103.59 1133.85 (288.25) (162.17) (364.11) % East Asian -503.83 43.82 682.17 (103.38) (100.94) (179.55) % South Asian -706.69 106.52 1106.51 (163.70) (113.45) (149.00) % Arab West Asian -1345.70-456.98 182.38 (416.53) (340.91) (332.50) % Other -2568.65 506.08 1339.44 (551.67) (340.13) (388.64) Household Income 123.82 79.74 (9.40) (14.93) Census Controls No Yes Yes FSA Effects No No Yes Year Effects Yes Yes Yes 15
Effect of Ethnic Diversity Increased diversity Reduces charitable giving Interpretation: increasing diversity by 0.1 Lowers donation by $27 A decline equal to 14% Between 1996 and 2006: Diversity increased by 0.06 points Charitable giving decreased by 8% Negligible effect on % households giving 16
Effect of Religious Diversity Also reduces charitable giving Not as robust as ethnic diversity Some evidence of difference in behaviour based on type of religion 17
What does this mean? Community diversity is increasing But diversity can lead to a divergence in preference for public and charitable goods Which leads to reduction in the provision of these goods As we think about how we can best support charities and encourage private giving The question should be what is needed to combat these effects How do we structure information about the services provided by charity? How do we structure charities? How do we think about public and private support of charity? 18
Future Directions in Research This paper Focuses on charitable giving as a whole Future work Focus on giving to charities By type: social welfare, health, arts, etc. Study effects of changes in neighbourhood in which charity is located on donations, other revenues, charity behaviour 19