![Sample our Area Studies journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5922/TOC-Subject-Sample-2021-Area-Studies.jpg"})
ABSTRACT
Recent anecdotes suggest that shareholdings of for-profit firms by nonprofit organizations engender agency conflicts between donors and managers. Analyzing a unique dataset of nonprofits in Korea, we find evidence that donors of the share-affiliated nonprofits rely less on the program ratio, a primary accounting performance metric in nonprofits, in making subsequent donations than those in unaffiliated nonprofits. We also report that the affiliation effect is primarily driven by individual donors but effectively attenuated by government monitoring. Our results highlight a novel source of agency problems in nonprofits.
Data Availability
Data are available from the public sources cited in the text.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. Stitchting Ingka Foundation became the ultimate parent of IKEA after Mr. Kamprad contributed all his shares to the foundation in 1982. While Mr. Kamprad does not own any shares of IKEA himself, he still retains his control over IKEA by being the chair of a five-person executive committee that runs the foundation and makes important decisions of IKEA, including dividend payments. As the foundation is tax-exempt, contributions and dividends to the foundation are not taxed. Refer to an article by the Economist, titled ‘IKEA: Flat-pack accounting’ in May 11th, 2006, for more details on IKEA’s organizational structure.
2. Chaebols are large conglomerate groups in Korea that grew rapidly after the Korean War with the generous support from the Korean government. The founding family largely controls chaebols and makes highly centralized decisions for the member firms, allowing member firms to share financial and non-financial resources with each other (Bae, Kang, and Kim Citation2002; Chang and Hong Citation2000). Chang and Choi (Citation1988) note that the principal shareholders (or the founding family members) of chaebols control all the affiliated companies through shares owned by themselves, their family members, and nonprofits established by themselves.
3. Such accusations on chaebols triggered the Korean Fair Trade Commission (KFTC) to initiate a full-scale investigation into this matter. Refer to an article by Reuters for more details on Samsung’s inheritance of shares through nonprofits (https://www.reuters.com/article/us-samsung-group-moves/samsung-heir-takes-charge-of-foundations-in-succession-step-idUSKBN0NZ2LJ20150515), and an article by Bloomberg for more information on the initiation of KFTC’s investigation in 2017 (https://www.bloomberg.com/news/articles/2017-11-16/korean-giants-are-under-the-microscope-again-now-for-their-giving).
4. This approach is analogous to that in prior studies that examine the implication of agency problems in shaping investors’ reaction to a firm’s accounting information in financial markets (e.g., Fan and Wong Citation2002; Yeo et al. Citation2002; Petra Citation2007; Habib and Azim Citation2008).
5. Nonprofits engaging in religion activities are not included in our sample since they are exempt from public disclosure.
6. We note that another related and equally important question is whether and to what extent share affiliation itself facilitates subsequent donations (hereafter the ‘direct impact’ of affiliations on donations). We predict that affiliated nonprofits would receive smaller contributions from outside donors concerned about potential agency conflicts, while enjoying greater contributions from affiliated for-profits or common individuals. We attempt to separate the effect of outside donors from that of affiliated ones in subsequent analyses.
7. We treat major shareholders, related parties, and executives of for-profit firms equally, assuming their incentives are well-aligned with each other.
8. In some cases, the nonprofit is explicitly disclosed as a related entity of the for-profit firm’s major shareholder. We classify such instances as having affiliations since prior studies suggest that principal shareholders of chaebols tend to control all the affiliated companies through shares owned by nonprofits established by the shareholder herself (Chang and Choi Citation1988).
9. In defining affiliated nonprofits, we do not impose a threshold for the number or percentage of shares held by the nonprofit, because the nonprofit need not exert significant influence over the affiliated for-profit by itself in our argument. Rather, the common individual may exert influence over the for-profit firm with the shares held collectively by itself, the nonprofit, and other firms under its control. Nevertheless, we find our results remain qualitatively similar when we additionally require a nonprofit to hold more than 5% or 10% of the outstanding shares of the for-profit to be classified as affiliated with a for-profit.
In defining affiliated nonprofits, we do not impose a threshold for the number or percentage of shares held by the nonprofit, because the nonprofit need not exert significant influence over the affiliated for-profit by itself in our argument. Rather, the common individual may exert influence over the for-profit firm with the shares held collectively by itself, the nonprofit, and other firms under its control. Nevertheless, we find our results remain qualitatively similar when we additionally require a nonprofit to hold more than 5% or 10% of the outstanding shares of the for-profit to be classified as affiliated with a for-profit.
10. Following Yetman and Yetman (Citation2013), we use two variables to control for nonprofit size: we include SIZE to control for scale effects, and SMALL to control for potential non-linear effects on donations to unsophisticated nonprofits. SIZE is the natural logarithm of total assets in millions of KRW, and SMALL is an indicator for nonprofits with total assets less than KRW 100 million. Our results are robust to using SMALL based on the alternative threshold of KRW 200 million or KRW 500 million, as well as to excluding SMALL from the model.
11. Our results are robust to controlling for the level of contemporaneous donations.
12. For example, we find significant variation in PROGRATIO across nonprofit types, such that the mean value of the program ratio is highest in research & scholarship nonprofits (0.639) and lowest in hospital nonprofits (0.248) (untabulated). Nonprofits’ reliance on individual versus corporate donations also differs depending on nonprofit type, such that education and social welfare nonprofits rely heavily on individual donations, whereas culture & arts and other nonprofits rely more on corporate donations. To control for such differences, we include nonprofit-type fixed effects in our multivariate analyses. In untabulated analysis, we find that the sensitivity of donations to the program ratio is consistently observed for all types of nonprofits.
13. Information on nonprofits is available at the NTS website (https://www.hometax.go.kr/websquare/websquare.wq?w2xPath=/ui/pp/index_pp.xml&tmIdx=20).
14. The mean value of PROGRATIO is 47.6% in our study, which is lower than the 79% reported in Yetman and Yetman (Citation2013). One potential reason for the lower program ratio is that the program expenses of Korean nonprofits include only the direct expenses spent in pursuing programs, but not indirect expenses allocated from administrative activities to the programs. Despite the lower average program ratio, Korean donors also view the program ratio as an important determinant in making donations, as observed by the positive association between DONATION and PROGRATIO in our empirical analyses.
15. We also confirm that the variance inflation factors for our test variables are lower than 3.4 in all regression models, which is well below the threshold of 10 suggested by Kennedy (Citation2008).
16. Due to the lack of data on donor-level information, we cannot directly trace whether corporate donations come from affiliated for-profits or unaffiliated ones. However, prior studies provide reasonable support for our claim. For example, affiliated for-profits make greater donations than do unaffiliated ones (Kim, Pae, and Yoo Citation2019). Moreover, two out of three for-profit firms contributing to charities make major donations to CEO-affiliated charities (Masulis and Reza Citation2014).
17. 17.0% of the identified private foundations hold affiliated stocks, whereas only 3.5% of the other nonprofits (i.e., nonprofits other than private foundations) hold affiliated stocks, which is consistent with the concern that private foundations are more likely to be affiliated with for-profits.
18. We note that the poor information environment of private firms may cause the coefficient difference to be insignificant. Specifically, obtaining information about private firms may be costly and thus, donors may not be fully aware of nonprofits’ affiliation with private firms. In this case, the coefficient on PROGRATIO*AFFNPO_UNLIST would be underestimated, causing the coefficient difference to be insignificant.
19. Specifically, related to H2, the difference in the coefficients on PROGRATIO for affiliated and unaffiliated nonprofits is statistically significant for individual donations (t-statistic = 8.70), while insignificant for corporate donations (t-statistic = 0.01). Related to H3, the difference in the coefficients on PROGRATIO for affiliated and unaffiliated nonprofits is insignificant when the nonprofit receives government grants (t-statistic = 0.24), while statistically significant when the nonprofit does not receive government grants (t-statistic = 6.87).
20. For example, affiliated nonprofits may organize programs aimed at benefiting affiliated entities rather than at creating public value; or they may restrict beneficiaries of the programs to their affiliated parties rather than the public. Donors may also suspect that affiliated nonprofits classify ancillary expenditure, such as expenses spent for excessive manager compensation, empire building, or any transactions that do not help increase public value, as program expenses. Unfortunately, we are not able to empirically prove the above examples, because of the unavailability of detailed information on nonprofit programs, such as the contents, beneficiaries, or components of program expenses.
21. This analysis is conducted only for nonprofits affiliated to listed firms since data on the largest shareholder is not available for unlisted firms.