![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
Despite increased attention to ethnic differences in political behaviour, there is little research on ethnic minorities as political donors and almost none outside the United States. We draw on an administrative dataset of contributions to candidates, which we augment with donors’ ethnicity. Focusing on the 2015 Canadian election, we find ethnic minorities are generally less likely to donate than other Canadians, but South Asian Canadians donate at astonishingly high rates. Contrary to previous research, there are only modest differences in the size of donations across ethnic groups. Linking donation data to candidate characteristics and census data reveals substantial co-ethnic affinity effects among Chinese and South Asian Canadians. Even in the absence of co-ethnic candidates, however, South Asians donate at a substantial rate. The proportion of donations to out-of-district and weaker candidates is also quite high, which could signal symbolic considerations are especially important to ethnic minority donors. The substantial heterogeneity between ethnic groups and the different effects on rates versus size of donations add important nuance to our knowledge of ethnicity and political behaviour.
Acknowledgements
For capable research assistance, we thank Sheeriza Azeez, Anna Johnson, Martin Wong, and especially, Rupinder Liddar and Rahul Gupta. Earlier drafts of this paper were presented at the 2018 meetings of the Canadian Political Science Association and the Midwest Political Science Association.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Our data show 727,497 donations of all kinds to the major parties. According to Statistics Canada, the 2015 Canadian population over the age of 18 was 27,816,114. There are some complications (e.g. minor or provincial parties, and people as young as 14 years of age can legally donate), but in any case, the real donation rate seems far lower than the reported survey rate.
2 In the pilot testing phase, we included some other groups, such as Black, Jewish, Southeast Asian, and West Asian, but even with additional training, these categories could not be reliably coded, so they have been collapsed. In addition, the surnames of Indigenous peoples, descendants of enslaved peoples, and Caribbean immigrants are often the same as those with European origins. Therefore, we suspect many Indigenous and some Black donors have been coded as European. This is clearly problematic as it subsumes colonised and enslaved peoples within the same category as Europeans. Other approaches are needed to understand donor behaviour among Indigenous peoples and Black Canadians. Substantively, this feature of our method means European-origin donors are likely even more over-represented than shown.
3 Exact addresses are not included in the database; in their absence, we cannot determine whether one ‘John Smith’ is the same as another. However, when we merge name and postal code into a quasi-unique identifier, 91% of donations are from a unique name/postal code. The real number of unique donors could be even higher, since multiple people with the same name potentially live in the same postal code.
4 Statistics Canada data do not provide an easy answer to the question ‘how many Canadians are European-origin?’ The ethnic origin question, which does include various European origins, allows respondents to enter multiple origins and thus double-counts. We therefore draw on the Census question that asks respondents whether they identify with a visible minority catagory. Visible minorities are individuals who are non-white and non-Caucasian in race or colour); the category does not include Indigenous peoples. Therefore, to estimate the proportion of the Canadian population that is European-origin, we use the ‘not a visible minority’ category and subtract 4.9%, which is the proportion of the population identifying as Indigenous.
5 In this analysis, we only include candidates who received as least one donation. We believe some candidates simply did not file a return (despite legal requirements, some candidates do not seem to file returns even several years after the election), and therefore coding them as having zero donations would induce error.
6 All models with candidate ethnicity include Indigenous as a category of the candidate ethnicity variable to avoid including them as European candidates. However, since this is not a primary category of interest here, they are not included in the predicted value plots. Interested readers can examine the detailed model results in the appendix. Unfortunately, we do not have data on indigeneity of donors.
7 We were unable to reliably pinpoint specific ethnicities for non-Chinese and non-South Asian minority candidates. We therefore do not know their precise co-ethnic donation status, so we have excluded them from this analysis.
8 In fact, given the prominence of Sikh politicians, we suspect Sikhs may be driving a portion of the South Asian effect; even further disaggregation would be beneficial, but we do not have reliable data on donor religion, and it is difficult to derive this from existing sources.