![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
We assess the determinants of private donations across a large sample of US based NGOs with foreign aid activities. Our results show that donations do not depend in the expected way on publicly available information on NGO characteristics that reveal an efficient and targeted use of funds, notably the efficiency price of NGO aid and the degree of specialisation. Private donors rather rely on the frequently offered option to designate donations to preferred purposes – even though this is unlikely to tie the NGOs' hands.
Acknowledgements
We are most grateful to two anonymous reviewers for constructive criticism and helpful advice. We would also like to thank Stefan Keitel, Alexander Leinemann, and Sandra Schmidt for carefully checking hundreds of NGO websites on possible ways to designate donations to specific purposes. Michaela Rank provided valuable assistance with respect to USAID data. Financial support from the German Research Foundation (DFG GZ: DR 640/2-1) is gratefully acknowledged.
Notes
1. See also Werker and Ahmed (Citation2008: ).
2. As stressed by Aldashev and Verdier (Citation2010), fundraising efforts by one particular NGO have two effects: diverting away donations from other NGOs and increasing the overall pool of donations (by ‘awakening’ or ‘activating’ potential new donors).
3. The ‘price of giving’ is another term widely used in the literature on the economics of altruism. Similar to the efficiency price as defined above, the focus is on the cost to the donor of increasing the NGO's charitable output by one unit. In contrast to the efficiency price, however, the price of giving refers to the after tax cost of donations (see, for example, Khanna and Sandler, Citation2000: 1545).
4. Expenditure items that are not directly related to the NGO's charitable programmes and projects comprise the costs of administration and management as well as expenses for fundraising. Note also that, throughout this article, ‘private donors’ include both individuals and institutions. The data we use do not allow us to distinguish between different types of private donors, even though this would be desirable as donations from individuals and institutions may react differently to NGO characteristics.
5. However, the efficiency price as defined by these authors does not account for administration and management costs. Callen (Citation1994) finds that more efficient organisations in a sample of 72 Canadian NGOs in the health sector receive higher donations.
6. Just 13 per cent of respondents state that they focus on an NGO's reputation. In an earlier survey, more than 80 per cent of respondents rated as important or very important that NGOs spend ‘an adequate amount … for program’ (Hager et al., Citation2001: 3).
7. As long as all NGOs underreport to the same extent OLS regressions generate a biased constant term, but the coefficients of the variables of interest would be unaffected. It cannot be ruled out, however, that NGOs with higher costs have stronger incentives to underreport.
8. See Appendix C in the Online Appendix to this article.
9. The type of activity relates to ‘sectors’ of aid, for example, social services such as education or health. We use the term ‘sectoral dimension’ in the following.
10. See Bilodeau (Citation1992) for an exact and formal account of the meaning of ‘enough’ in this context: ‘Essentially, individuals must earmark more money in aggregate to each good than the monopoly [NGO] would from the undesignated funds it has’ (Bilodeau and Slivinski, Citation1997: 461).
11. According to Bilodeau and Slivinski (Citation1997: 451), a dominant NGO using the pool of undesignated donations to offset the effects of any earmarks provides ‘a mirror image’ of Gary Becker's Rotten Kid Theorem. A household head with sufficient income may neutralise actions by other family members that would otherwise alter the distribution of income within the family from the distribution preferred by the household head.
12. Even if the share of undesignated funds were known and small enough, it remains open to question whether NGOs could credibly commit themselves to adhere to donor designations ex post.
13. Andrés-Alonso et al. (Citation2006) show that large official financiers (notably the Spanish Agency for International Cooperation) improve the efficiency of Spanish NGOs engaged in international development cooperation through effective monitoring. Private donors may rely on official monitoring and, thus, increase donations.
14. According to Andreoni and Payne (Citation2011: 339) a marginal dollar spent on fundraising yields about five dollars in new donations.
15. The reasoning refers to the direct effects of fundraising on private donations; see above for indirect effects, that is, fundraising expenditures increasing the efficiency price of NGO aid.
16. We are particularly grateful to an anonymous referee for alerting us to this point. A preferred option would be to include the number of country offices and sister organisations of US based NGOs, instead of the dummy variable. However, this information is missing for various NGOs in our sample. Note that NGO characteristics and some other more specific control variables are introduced and described in Section 3.
17. See: http://pvo.usaid.gov/usaid/index.html; accessed February 2011.
18. The data on revenue and expenditure items used in the empirical analysis come from the 2009 VolAg Report: Report of Voluntary Agencies Engaged in Overseas Relief and Development (http://www.usaid.gov/our_work/cross-cutting_programs/private_voluntary_cooperation/volag2009.pdf), which lists 559 NGOs.
19. See: http://www.pvo.net/usaid/index.html; it is not reported, however, how much the NGOs spend in particular sectors and countries.
20. For details, see – in the Online Appendix.
21. The average share of overseas programmes is almost 80 per cent. Domestic programmes within the United States account for the rest.
22. The variable Designation option: others captures a heterogeneous variety of options to designate. These options range from choices of (just a few or a fairly large number of) specified projects to the possibility to enter self-defined preferred activities in a free text field.
23. We experimented with the instrumentation of Countries active and Official funds in an earlier working paper version of this article; for details see: http://www.ifw-members.ifw-kiel.de/publications/donations-to-us-based-ngos-in-international-development-cooperation-how-un-informed-are-private-donors/kwp-1680. However, an anonymous referee alerted us to several weaknesses of the instruments used there.
24. The weighting with population is only applied to the fundraising overlap along the geographical dimension. Comparable weights are not available for the overlap along the sectoral dimension. Nevertheless, the interpretation is straightforward and comparable between both dimensions. Both overlaps are defined in logs so that their coefficients reflect elasticities.
25. On the other hand, the dummy variable is set equal to zero for those NGOs in our sample that are not listed as international organisations in the Yearbook and for listed NGOs that are classified as internationally oriented national (in our case US) organisations.
26. Indeed, Countries active proved to be insignificant at conventional levels in unreported 2SLS estimations with the average number of countries in the closer neighbourhood used as the instrument (http://www.ifw-members.ifw-kiel.de/publications/donations-to-us-based-ngos-in-international-development-cooperation-how-un-informed-are-private-donors/kwp-1680).
27. Yet the negative coefficient of the dummy variable for online donations is fairly surprising. The large majority of NGOs in our sample offers this option. The explanation could be that those NGOs not offering the online option receive particularly large donations from just one or a few private donors (for example, foundations financed by wealthy philanthropists). Importantly, our major results are hardly affected when replicating the estimation for a reduced sample that excludes NGOs not offering the online option; see column (2) of .