Tax-Exempt Status and Associated Factors among Charitable Foundations in China

Abstract

Tax exemption plays an important role in the sustainability of charitable organizations (COs). The 2016 Charity Law of China provides stronger tax incentives for charity donations. Using 767 observations of Chinese charitable foundations (CFs) during 2010–2018 from the China Foundation Center database and manually collected tax-exempt status data, this study applies multivariate logistic regression analysis to examine the association between tax-exempt status and related key factors, such as transparency and donation dependency. This study found that a one-point increase in the transparency score of a CF is associated with a 3.9 percentage points higher likelihood of having at least one type of tax-exempt qualification (OR = 1.039, p < 0.01). There is in general a significantly positive association between tax-exempt status and donation dependency of CFs in China. After 2016, the CFs responded actively to the tax incentive provided by the Charity Law, which in return requires a higher level of transparency. These results suggest that taxation under the legal system may effectively function to promote the sustainability of charity foundations in China in the long run. Further studies are needed to explore in-depth why CFs with advanced tax-exempt qualifications concentrate in Beijing and Shanghai.

1. Introduction

A charitable organization (CO) refers to a social organization designed to benefit the general public or a specific group of the community with educational, humanitarian, or religious activities. CO activities go beyond giving relief to the indigent, extending to the promotion of happiness and the support of many worthy causes [1]. With social goals as parts of their mission, the activities of a CO go beyond giving relief to the indigent, as their actions can extend to include the promotion of happiness and the support of Sustainable Development Goals (SDGs) for society. These SDGs can be activities such as promoting health equity, employment opportunity, community security, responsible social practices, social equity, environmental protection, and natural resource conservation [2,3]. Due to wide social supporting roles and participation, COs are playing an increasingly important and active role during the decision-making and the implementation process for social sustainability, in both the public and private sectors, as well as at both the national and international level [4,5]. The contributions of COs, together with other nonprofit organizations, to sustainable social developments have been well recognized and acknowledged by both the academic community and international organizations [2,4,6].

COs are generally exempted from paying income or property taxes because these organizations perform functions of social welfare improvement, social equality enhancement, environmental protection, and other tasks for sustaining society that the government would normally have to perform [7]. Therefore, the government is willing to forgo the tax revenue in return for the public interest services rendered by COs [8,9,10,11,12]. In addition, the tax-exempt status is justified because COs follow a non-distribution constraint principle [13]. COs do not seek profit and can only maintain a surplus of revenue, if any, for work that is to provide a social benefit, rather than for distributing earnings to stakeholders, such as founders, members, employees, or anyone else in control [14,15].

Tax exemption includes, specifically, exemption from taxes (e.g., value-added tax, customs fees), tax reductions for charitable organizations, and tax reductions for donors (who can deduct donations from their tax base) [16]. To be qualified for tax exemption, a CO must continue to pursue its mission and follow the non-distribution constraint principle. Tax exemption plays a very important role in the governmental financial support of the COs. It is included among the key factors of the legal environment and financial viability monitored by the United States Agency for International Development (USAID) when the sustainability of civil society sectors are evaluated [17].

Charity foundations (CFs) in China are a type of charitable organization, having the functions of both raising funds and operating projects for their mission, rather than focusing on accessing grants [18]. Starting from the 1980s, through government financial subsidies, CFs in China have been growing at a rapid rate during the past decade with additional strong supports from social donations and private endowments [19]. The number of CFs has increased from 609 in 2003 to 7169 in 2020 [20]. While the governance and regulation of CFs in China is not yet well established, the occurrence and exposure of some major scandals about CF’s fundraising and management problems has severely harmed public trust and donation motivations. To address the public concerns, a higher level of transparency, accountability, and updated regulation of CFs was proposed and enacted in China. The first national Charity Law of China was passed and came into implementation in 2016. To encourage donations by individuals and organizations, the 2016 Charity Law of China strengthens the tax incentive by allowing deductible taxes to be carried forward for three years.

There are concerns that nonprofit organizations in China face additional barriers to gaining tax-exempt status or receiving benefits due to the incomplete provision of information and the low level of trust in the government [21,22]. These difficulties and barriers may be extraordinarily strong for grassroots nonprofit organizations [23]. Additionally, some scholars argued that nonprofit organizations in China often cannot classify income sources clearly; hence tax exemption and administration often cannot be implemented effectively [24]. During the past five years, there has been a volume of literature studying the transparency and associated factors of CFs in China [3,25,26,27,28]. More recently, there has been a surge of studies focusing on the political connections of Chinese foundations [29,30]. However, we have found no literature focusing on the empirical study of tax-exempt status of contemporary CFs in China.

Using a reliable dataset of 767 CF observations obtained from the China Foundation Center and manually collected tax-exempt status information, this study analyzed the tax-exempt status and associations along with the organization features among CFs. Section 2 of this study describes briefly the relevant regulation and legal environment of tax exemption for charitable organizations in China and the potential policy effects involved. Section 3 provides the theoretical framework and a literature review of tax incentives, agency problems, and the transparency of charitable organizations. Section 4 describes the research methodology, including the research hypothesis, data sources, and estimation models. A logistic regression method was adopted for the empirical analysis. Section 5 presents the estimated results from logistical regression analysis regarding three types of tax-exempt status. Moreover, Section 6 concludes that there are strong and significant positive associations between tax-exempt status, transparency, and donation dependency of CFs.

2. Institutional Background

In China, tax exemption for charitable organizations is implemented in accordance with Article 4 of the Enterprise Income Tax Law of the People’s Republic of China. There are basically two types of tax exemption qualifications; tax-exempt for the receiver of funds and tax-deductible for the giver of funds. A brief introduction of each tax exemption qualification is provided as follows.

According to Article 79 of the Charity Law of the People’s Republic of China (hereinafter referred to as the Charity Law) enacted in 2016, the income of charitable organizations is exempted from enterprise income tax, as long as it meets the stipulated conditions. Specifically, the qualified income of a qualified charity is tax-exempt income. The law excludes any income obtained from for-profit activities. Similarly, legal regulations of the National Taxation Administration and local Municipal Tax Service in China require that the income of charities must meet the following two conditions for tax exemption: First, the charitable organization must maintain its institutional tax-exempt status. Second, the income per se can be recognized as tax-exempt income. Specifically, the Enterprise Income Tax Law and Implementation Rules of China stipulate the requirements as follows: (1) The charity is established or registered in accordance with relevant national laws and regulations. (2) The charity is engaged in public welfare or nonprofit activities. (3) Except for the reasonable expenditures related to the organization, all income obtained should be used in accordance with the mission of the organization, or the previously registered and approved public welfare or nonprofit activities.

As for tax-deductible for contributors, natural persons, enterprises, or organizations in China that make donations to qualified charitable organizations qualify for a tax deduction [31]. A maximum of 12% of the total annual profits by organizational contributors [32] are tax-deductible. For individual contributors, a maximum of 30% of the taxable income declared is deductible [33]. To offer a stronger tax-deductible incentive to enterprises, the 2016 Charity Law of China made an amendment allowing deductible taxes exceeding 12% of the enterprise’s total annual profit to be carried forward for three years. For any donators in China to be qualified for tax deduction, they not only should donate to qualified charity activities (such as education, public welfare, poverty, or disaster aid), but also need to donate through qualified charitable organizations, that are officially qualified to receive donations.

To sum up, the tax-exempt status of a CF in China can be determined by a combination of multiple tax exemption qualifications. Specifically, it could be one of being tax-exempt only, tax-deductible only, or both tax-exempt and tax-deductible, or having neither qualification.

3. Theoretical Framework and Literature Review

3.1. Asymmetric Information and the Transparency of CFs

Historically, information about tax-exempt organizations has often been hard to obtain. Due to the asymmetric sharing of information, the public finds it difficult to get access to the internal information necessary to monitor CFs such that the agency problem has become a concern among charitable organizations too [34,35].

Transparency refers to the obligation or initiative of a charity to disclose key information about the organization, such as financial reporting, management practices, and effectiveness. To improve the governance of CFs and alleviate information asymmetries, there are mandatory and voluntary information disclosures to improve the transparency of CFs [6]. As for mandatory disclosures, government regulations require disclosures and related information to be provided to users of the charity [5]. For example, the Internal Revenue Service of the USA states: “A charitable organization should demonstrate credibility and transparency to its supporters through the release of comprehensive and accurate corporate mission, operation, financial and management information” [14,15]. The 1996 Tax Act of the USA requires a tax-exempt organization to provide a copy of its informational Form 990 to anyone who requests it and specifies penalties for non-compliance [35].

A system of voluntary information disclosure by charitable organizations can play an important role in improving their transparency. According to the signaling effect, institutions with good governance and success in fulfilling their missions can choose to distinguish their good quality through the voluntary disclosure of their internal information to the public [36]. They can strategically increase their public trustworthiness by creditably disclosing more information and enhancing their own public transparency [3,25,26]. By this means, a charitable organization with better transparency may collect more donations and receive more social resources so that it is able to achieve its mission in a sustainable fashion. Furthermore, there is a positive association between transparency and the donation revenue ratio among CFs in China [37].

3.2. Resource Dependence Theory

According to resource dependence theory, since CFs are not financially self-reliant and external resources are critical to their survival, CFs need to make efforts to accommodate the parties who are critical to or have control over the available resources [38,39,40,41].

A large number of empirical studies among CFs in western societies found supporting evidence for resource dependence [42,43,44]. There has been an increased demand for charities to disclose information as donors increase their giving [45,46,47]. Empirical studies about CFs in China also found that a higher level of online disclosure is associated with greater dependence on public donations [48].

3.3. Signaling Effect of Tax-Exempt Status

Since an organization must be legally qualified and its tax-exempt status must be evaluated and recognized by legal authorities, the current tax-exempt status hence serves as a positive signaling effect for the legitimacy and trustworthiness of the CFs [15]. Further, to maintain the tax-exempt status, the operations and financial statements of CFs are audited and examined by government authorities. This compliance, as a general rule, signals a higher level of credibility for financial reports and other reporting standards [49]. As a result, donors are willing to give more to the nonprofits that undergo higher quality audits [50].

Because tax exemption reduces the donors’ income tax, a financial incentive for donation, especially to CFs, is created. As such, the incentive to donate increases as the donor’s marginal tax rate increases [51]. Tax deductions have a positive effect on promoting the scale and amount of donations made by individuals and enterprises to CFs [52]. The major reason is that a tax deduction is considered to be reducing the cost of donations to donors [53,54]. In addition, allowing a tax deduction is much more attractive than giving tax credit, especially for high-income taxpayers, who are tax-sensitive. In short, offering a tax deduction may attract rich contributors to donate even more [55].

In emerging economies especially, tax-exempt status is regarded as a legal status and sign of government support [17]. Some literature documents that nonprofit organizations in China have faced additional barriers to gain official recognition for tax-exempt purposes, mainly due to some high-profile scandals in recent years [29].

3.4. Abuses of Tax-Exempt Status

For decades, there have been concerns that the tax-exempt status of some charitable organizations may have been used for tax avoidance or abused for other purposes [56,57] because agency problems have always been major concerns about their governance. CFs by legal requirement have no owner of profits, and the residual claimants are either unable to monitor efficiently or unwilling to do so [25,58].

There are primarily two categories of abuses undertaken by tax-exempt entities, specifically, charitable contribution deductions and the use of an exempt entity as an accommodation party in a transaction that will shift tax benefits [59]. For example, in the USA, abusive transactions may overvalue tangible or intangible property contributed to a charity when the organization claims the value, or undervalue the quid pro quo benefits a charity provides to a contributor in exchange for a contribution [59]. Omer and Yetman (2007) find that 19 percent of their pooled sample of CFs misreported their taxable income by overstating their taxable expenses by an average of 30 percent [56]. Shari Cohen (2005) states that hospital boards, both not-for-profit and for-profit, are more accountable and transparent in their governance and operation in order to justify tax-exempt status [60].

Regulations of tax-exempt status for COs were not well established in China in recent decades. Before the 2000s, COs used to be treated as a supplemental department of the government and did not need to pay tax. In addition, the tax authorities in China did not have legal documentation or relevant information about how to tax nonprofit organizations that were separately administrated by the Ministry of Civil Affairs. During the most recent two decades, almost all new COs were founded by non-government organizations. Due to a lack of registered information from the Ministry of Civil Affairs, most of these newly founded COs are not qualified for tax-exempt status [23].

Because the regulations in China did not distinguish clearly the taxable income and nontaxable income of a CO, there are concerns that some COs abused their tax-exempt status and took advantage of grey areas [24]. Additionally, there are also concerns that tax-exempt charitable giving is abused by corporations for tax avoidance purposes or other hidden motivations, such as a form of corporate political influence undetected by voters and subsidized by taxpayers [61]. Some corporate foundations in China might be established for the self-interest of a business, rather than charity or corporate social responsibility (CSR) motivations [62].

4. Materials and Methods

4.1. Research Hypothesis

According to the literature review and discussion above, it can be summarized that the tax-exempt status of a CF in China is a positive signal of the legitimacy, governance, and credibility of the CF. According to the theory of asymmetric information and signaling effect, we thus propose:

Hypothesis 1 (H1).

The tax-exempt status of a charitable foundation in China is positively associated with its transparency level.

According to the resource dependence theory, it is more likely for a CF to apply for tax-exempt status to attract more donations if its revenue has a higher-level dependence on external donations rather than from government subsidies or other operation revenues. Meanwhile, with the signaling effects of the tax-exempt status, a CF can increase its attractiveness and trustworthiness among potential donators. Together with the positive financial incentive provided by tax-exempt, we thus propose:

Hypothesis 2 (H2).

The tax-exempt status of charitable foundations in China is positively associated with the donation ratio of its income.

Compared to their counterparts in small towns or the countryside, foundations located in large cities and developed areas of China have better access to more social resources in terms of sophisticated supporters, professional managers, social networking, and more supports from government [49]. Meanwhile, the CFs in these areas may also face strong competition from established peers. We thus propose:

Hypothesis 3 (H3).

The charitable foundations in socially developed areas of China are more likely to obtain tax-exempt status.

As introduced in Section 2 of this paper, the 2016 Charity Law of China encourages social donations to CFs through the mechanism of a tax deduction with carry-forward. We thus propose:

Hypothesis 4 (H4).

The charitable foundations in China are more likely to obtain the tax deductible status after the year 2016, when the new Charity Law was implemented.

4.2. Data Source

The main data source of this study is the Chinese Foundations Center (CFC) database (www.foundationcenter.org.cn, with the latest accessed on 1 October 2020), which contains basic institutional, financial, and operational information for CFs in China [63]. We first randomly sampled 150 foundations from the database for the year 2014, which accounted for about ten percent of the foundations recorded in the database for that year. We then used the year 2014 at the middle point of time to extend from 2010 to 2018 to build a set of panel data. In this way, we can include as many foundations as possible. The most recent data available was from 2018, when the dataset was built in 2020. We then kept the CFs with complete variable information during the study period. The data were further cleaned by checking each variable for potential reporting errors. For example, some observations reported negative donation income, or a donation-revenue ratio greater than one. The final clean dataset keeps about 56.7% of the initial observations. It is an unbalanced panel dataset, containing 767 observations from 149 CFs during 2010–2018.

The tax-exempt status information of the foundations was manually collected from the annual reports posted on their homepages. Only a very few CFs had the tax-exempt status as “tax-exempt qualification only”. These CFs were excluded from the final sample, because their observation number was not sufficient for regression analysis. As a result, the CFs in the final sample in total had three types of tax-exempt status, specifically, both tax-deductible and tax-exempt, tax-deductible only, and without any tax-exempt status.

4.3. Statistical Method and Regression Model

This study performed multivariate logistic regression analysis, examining the key factors associated with the tax-exempt status of charitable organizations in China during 2010–2018. The full model of the estimation is as below:

Tax_exempt_status_it = β0 + β₁ Transparency_it + β₂ Donation_ratio_it + β₃ Fundraising_exp_it + β₄ln(total_income) + β₅Age_it + β₆Specialization_type_it+ Area_i + Year_t

(1)

where i refers to a charitable organization; t indexes any year during the period from 2010 to 2018; “Tax_exempt_status_it” is one of a dichotomy of tax-exempt qualifications for a CF_i in year t. It could be “with any tax exemption”, “tax-deductible only”, or “tax-deductible and -exempt” for a specific regression test; Donation_ratio_it is the percentage of donation income in the total revenue of a CF_i in year t; Fundraising_exp_it is the percentage of the fundraising expenses in the total donation income. ln(total_income)_it is the log of the total income; Age_it is the number of years since a CF_i’s establishment in year t; Specialization_type_it refers to the service range; Area_i contains a list of cities and provinces, indicating the head office location of a CF_i. Year_t is a vector of dummy variables, indicating years from 2011 to 2018 with 2010 as the base year.

The dependent variable is one of the three tax-exempt status dummy variables. The dummy variable “with any tax exemption” identifies those with at least one type of any tax exemption qualification. Further, the dummy variable “tax-deductible only” specifically indicates the CFs that have tax-deductible qualification only. The dummy variable “tax-deductible and -exempt ” indicates the CFs that have both tax-deductible and tax-exempt qualifications. The base group includes those with no tax exemption at all.

Transparency of the CF_i is measured by the China Foundation Transparency Index (FTI), which is regarded as among the best indicators of charitable organization transparency in China and has been widely adopted in recent studies [64]. FTI includes 40 sub-indicators from three categories: basic information (36.8 points), financial information (27.6 points), and project information (35.6 points). Each indicator is measured, evaluated, and then summed to get the final transparency score with 100 points as the full score.

Donation ratio is the ratio of donation income to the total revenue of an organization, measuring the degree that a CF relies on its donation. The funding sources for CFs in China include external donations, government subsidies, and investment income. For most CFs, donation income is still the most important financial source and few of them have investment income. Therefore, the donation ratio is sufficient to measure the resource dependency of a CO. The log transformation of this variable is adopted for a robustness check due to its non-normal distribution.

The fundraising expenses ratio is calculated by comparing the fundraising expenses and the total donation raised by a CO. This variable reflects the efficiency of the fundraising as well as the governance of a CF.

ln(total_income) is the log transformation of the total income of a CF and is included to control the size effect of the CFs. Since the total assets information of many CFs in the sample was not disclosed, total income (with donation revenue included) is an effective way to control for the size effect as well.

Age is the number of years since the establishment of a CF to the year studied. Usually, the longer the history a CF has, the better it is known by the people and typically the more popular it is while holding a higher public reputation [18,19,20].

Specialization type refers to the service range a CF may focus on. The CFs in this sample can largely be classified into three categories. The largest group includes those specializing in charity aid (including medical aid, poverty, and disaster relief). The other two groups are for educational development and culture promotion.

Area is a sector of dummy variables, indicating the head office location of a CF if it is in one of the most developed central cities or province (Beijing, Chongqing, Guangdong, Shanghai, or Zhejiang province). The other provinces in the base group (including Fujian, Guizhou, Hebei, Heilongjiang, Henan, Hunan, Inner Mongolia, Liaoning, Shandong, Shannxi, Shanxi, Sichuan, Tanjin, Tibet, and Yunnan) are mostly in the west or middle areas of China, which traditionally are less developed than Eastern China. This sample in total contains CFs from 21 provinces in China. About half of them are from the central cities or the most developed areas of China as identified by the dummy variables. The estimation of Area captures region-wide effects that are associated with the likelihood of a CF obtaining tax-exempt status.

Year is a sector of dummy variables, capturing year-related effects that may alter the likelihood of a CF obtaining tax-exempt status. Since the 2016 Charity Law of China was implemented and started to provide stronger tax-deductible incentives, it is expected that more CFs would apply and obtain tax-deductible qualification stipulated by the law.

5. Results

Table 1. Descriptive Analysis.

Panel A
Variables	Full Sample	With Any Tax Incentive (Sub-Sample)	Tax Deduction and Exempt (Sub-Sample)	Tax Deduction Only (Sub-Sample)
	Mean (SD)
Transparency	60.914 (16.442)	62.287 (17.349)	67.078 (13.993)	60.087 (18.288)
Donation ratio	0.734 (0.314)	0.770 (0.280)	0.775 (0.268)	0.769 (0.286)
ln (Donation ratio)	−0.570 (1.413)	−0.384 (0.716)	−0.397 (0.772)	−0.378 (0.690)
Fundraising expense	0.008 (0.024)	0.008 (0.022)	0.007 (0.016)	0.009 (0.024)
Ln(total income)	15.998 (1.636)	16.123 (1.618)	16.577 (1.804)	15.915 (1.480)
Age	11.6 (8.484)	10.498 (7.944)	12.514 (8.778)	9.573 (7.360)
N (Percentage)	767 (100%)	588 (76.66%)	185 (24.12%)	403 (52.54%)
Panel B
	Variables		N (Percentage)
Specialization	Charity aids		128 (16.69%)
	Education development		348 (45.37%)
	Culture		291 (37.94%)
Location	Beijing		45 (30.2%)
	Chongqing		7 (4.70%)
	Guangdong		6 (4.03%)
	Shanghai		3 (2.01%)
	Zhejiang		13 (8.72%)
	Other *		75 (50.34%)
Year	2010		98 (12.78%)
	2011		91 (11.86%)
	2012		106 (13.82%)
	2013		101 (13.17%)
	2014		122 (15.91%)
	2015		118 (15.38%)
	2016		43 (5.61%)
	2017		39 (5.08%)
	2018		49 (6.39%)

Note: * “other” includes Fujian, Guizhou, Hebei, Heilongjiang, Henan, Hunan, Inner Mongolia, Liaoning, Shandong, Shannxi, Shanxi, Sichuan, Tanjin, Tibet, and Yunnan Province in China.

presents descriptive statistics of the dataset. In Panel A, the mean and the standard deviation of the continuous variables in the full sample and sub-samples of various types of tax-exempt status are reported, respectively. About 76.7% of the CFs in this sample have at least one type of tax-exempt qualification. CFs with “tax-deductible and -exempt” account for about 24% of the sample, with a mean transparency of 67 points, which is much higher than the means of the full sample. Except for the age of organizations, there are no significant differences among other major variables of the CFs.

In Panel B of Table 1, the observation numbers and percentages of the category variables are reported. About 30% of the CFs are from Beijing, and about 50% are from other less developed areas in China.

The correlation analysis of the variables is as reported in

Table 2. Correlation analysis.

Variables	(1)	(2)	(3)	(4)
Transparency	1.000
Donation ratio	0.190 ***	1.000
	(0.000)
Fundraising expense	−0.041	−0.051	1.000
	(0.254)	(0.154)
Ln(total income)	0.360 ***	0.240 ***	0.022	1.000
	(0.000)	(0.000)	(0.547)

*** p < 0.01.

, and no collinear concerns were identified.

For the logistic regression analysis, first the dichotomy variable “with any tax exemption” was used as the dependent variable for a set of empirical estimations examining the full sample and the overall association with tax-exempt status. Subsamples of CFs with “tax-deductible only” or “tax-deductible and exempt” were then examined, with the base group including only those with no tax-exempt qualification. As a robustness check, a set of regression analyses with various specifications was tested.

The estimation results of the logistic regressions are reported in

Table 3. Results of regression on “with any tax exemption” (logistic).

	(1)	(2)	(3)	(4)	(5)
Variable	With Any Tax Exemption
Transparency	1.039 ***	1.035 ***	1.018 ***	1.031 ***	1.033 ***
	(0.008)	(0.008)	(0.006)	(0.007)	(0.007)
Donation ratio	2.169 ***		3.589 ***	1.968 **	3.190 ***
	(0.632)		(0.932)	(0.551)	(0.871)
Ln (Donation ratio)		1.382 ***
		(0.100)
Fundraising expense	0.050	0.016		0.083
	(0.186)	(0.061)		(0.298)
Ln(total income)	1.126 *	1.320 ***		1.114 *
	(0.072)	(0.098)		(0.070)
Age	0.915 ***	0.917 ***		0.927 ***
	(0.015)	(0.015)		(0.011)
Edu_development	0.695	0.582 **		0.689 *
	(0.159)	(0.140)		(0.145)
Charity_aids	3.727 ***	2.875 **		4.510 ***
	(1.897)	(1.487)		(2.266)
Beijing	0.748	0.873			0.352 ***
	(0.198)	(0.240)			(0.075)
Guangdong	0.747	0.764			1.065
	(0.386)	(0.396)			(0.515)
Shanghai	0.574	0.669			0.370
	(0.360)	(0.421)			(0.234)
Zhejiang	0.451 **	0.524 *			0.348 ***
	(0.173)	(0.204)			(0.128)
2011	1.051	1.091			0.981
	(0.394)	(0.426)			(0.354)
2012	1.384	1.400			1.154
	(0.510)	(0.535)			(0.407)
2013	1.460	1.431			1.126
	(0.549)	(0.555)			(0.400)
2014	1.411	1.402			1.019
	(0.507)	(0.520)			(0.344)
2015	1.435	1.338			0.977
	(0.517)	(0.496)			(0.331)
2016	2.722 **	2.342			1.359
	(1.367)	(1.272)			(0.630)
2017	2.396 *	2.339			1.227
	(1.267)	(1.358)			(0.596)
2018	2.688 **	2.161			1.181
	(1.319)	(1.119)			(0.518)
Constant	0.085 **	0.021 ***	0.454 **	0.168 *	0.334 **
	(0.085)	(0.024)	(0.163)	(0.154)	(0.155)
Observations	767	746	767	767	767
Pseudo R2	0.157	0.172	0.0502	0.140	0.0846

Note: odds ratios are reported in the table above. *** p < 0.01, ** p < 0.05, * p < 0.1.

,

Table 4. Results of regression on “tax deductible and exempt” (logistic).

Variables	(1)	(2)	(3)	(4)	(5)
	Tax Deductible& Exempt
Transparency	1.078 ***	1.065 ***	1.060 ***	1.051 ***	1.074 ***
	(0.016)	(0.016)	(0.010)	(0.012)	(0.013)
Donation ratio	1.888		2.493 **		2.788 ***
	(0.789)		(0.890)		(1.071)
ln(Donation ratio)		1.364 ***		1.289 ***
		(0.138)		(0.104)
Fundraising expense	8.691	0.725		0.000
	(49.192)	(4.194)		(0.003)
ln(total income)	1.097	1.311 **		1.417 ***
	(0.090)	(0.139)		(0.131)
Age	0.830 ***	0.834 ***		0.947 ***
	(0.027)	(0.028)		(0.016)
Edu_development	0.412 ***	0.325 ***		0.395 ***
	(0.141)	(0.115)		(0.117)
Charity_aids	4.318 **	3.322 **		2.480 *
	(2.598)	(2.031)		(1.369)
Beijing	14.057 ***	14.829 ***			1.399
	(7.598)	(8.232)			(0.395)
Guangdong	2.296	2.159			2.268
	(1.513)	(1.422)			(1.326)
Shanghai	10.300 ***	12.161 ***			2.078
	(7.348)	(8.708)			(1.400)
Zhejiang	0.206 **	0.232 **			0.117 ***
	(0.136)	(0.156)			(0.072)
2011	1.130	1.160			0.917
	(0.587)	(0.627)			(0.427)
2012	1.575	1.517			1.018
	(0.814)	(0.814)			(0.466)
2013	1.779	1.660			0.986
	(0.953)	(0.909)			(0.454)
2014	1.805	1.693			0.830
	(0.927)	(0.893)			(0.367)
2015	2.129	1.836			0.813
	(1.088)	(0.971)			(0.355)
2016	3.134	3.341			0.531
	(2.314)	(2.610)			(0.337)
2017	1.527	1.473			0.338
	(1.283)	(1.351)			(0.227)
2018	2.707	2.520			0.431
	(2.035)	(1.964)			(0.258)
Constant	0.004 ***	0.001 ***	0.015 ***	0.001 ***	0.007 ***
	(0.005)	(0.002)	(0.009)	(0.001)	(0.005)
Observations	364	351	364	351	364
Pseudo R2	0.320	0.342	0.126	0.235	0.178

Note: odds ratios are reported in the table above. *** p < 0.01, ** p < 0.05, * p < 0.1.

and

Table 5. Results of regression on “tax-deductible only” (logistic).

Variables	(1)	(2)	(3)	(4)	(5)
	Tax Deductible Only
Transparency	1.033 ***	1.029 ***	1.009 *	1.024 ***	1.028 ***
	(0.009)	(0.009)	(0.006)	(0.007)	(0.007)
Donation ratio	1.902 **		3.643 ***	2.077 **	2.580 ***
	(0.615)		(0.990)	(0.618)	(0.786)
ln(Donation ratio)		1.318 ***
		(0.103)
Fundraising expense	0.182	0.069		0.228
	(0.699)	(0.265)		(0.837)
ln(total income)	1.187 **	1.402 ***		1.121
	(0.087)	(0.121)		(0.079)
Age	0.918 ***	0.919 ***		0.915 ***
	(0.016)	(0.017)		(0.012)
Edu_development	0.731	0.621 *		0.739
	(0.190)	(0.167)		(0.171)
Charity_aids	3.274 **	2.568 *		4.118 ***
	(1.752)	(1.399)		(2.139)
Beijing	0.340 ***	0.394 ***			0.164 ***
	(0.101)	(0.121)			(0.040)
Guangdong	0.528	0.541			0.832
	(0.298)	(0.306)			(0.426)
Zhejiang	0.540	0.626			0.440 **
	(0.215)	(0.253)			(0.165)
2011	1.050	1.084			1.047
	(0.436)	(0.464)			(0.416)
2012	1.471	1.509			1.273
	(0.604)	(0.635)			(0.496)
2013	1.536	1.496			1.220
	(0.640)	(0.635)			(0.481)
2014	1.473	1.440			1.145
	(0.584)	(0.585)			(0.426)
2015	1.442	1.384			1.008
	(0.581)	(0.568)			(0.378)
2016	3.574 **	3.017 *			1.980
	(2.006)	(1.830)			(1.007)
2017	3.515 **	3.313 *			2.000
	(2.007)	(2.030)			(1.065)
2018	3.324 **	2.649 *			1.626
	(1.781)	(1.496)			(0.776)
Constant	0.045 ***	0.008 ***	0.527 *	0.178 *	0.383 *
	(0.051)	(0.011)	(0.192)	(0.183)	(0.192)
Observations	577	557	582	582	577
Pseudo R2	0.185	0.189	0.0402	0.147	0.120

Note: odds ratios are reported in the table above. *** p < 0.01, ** p < 0.05, * p < 0.1.

as odds ratios (OR). Calculated based on the original coefficients of logistic regressions, the OR reports the odds that an outcome will occur given a particular exposure or characteristic, compared to the odds of the outcome occurring in the absence of that exposure or characteristic. In this study, OR indicates the relative odds of a CF having a certain type of tax-exempt status, given its organizational features. If an OR is equal to one (OR = 1), the exposure/feature does not affect the odds of the outcome. If an OR is greater than one (OR > 1), the exposure/feature is associated with higher odds of an outcome. If an OR is less than one (OR < 1), it indicates an association with lower odds of an outcome. OR has been widely adopted in epidemiologic analysis and health-related studies [65,66]. OR is adopted in this study because it is more intuitive than the original coefficient of logistic regression.

Table 3 reports the estimation results using “with any tax exemption” as the dependent variable, which gives a general analysis of the CFs. The full sample was analyzed and the results of the full model are reported in Column (1) and other tests for a robustness check are reported in Column (2) to (5) of Table 3. As indicated in Column (1) of Table 3, if the transparency score of a CF increases by one point, the likelihood of this CF having any type of tax-exempt status increases by 3.9 percentage points (OR = 1.039, p < 0.01). When the donation ratio of a CF increases from zero to one, the likelihood of having any type of tax exemption qualification status increases by 116.9 percentage points (OR = 2.169, p < 0.01). In other words, the likelihood of having any tax exemption is 16.9 percentage points higher, on average, when a CO’s donation ratio is 10 percentage points higher. Results in Column (1) of Table 3 also indicate that CFs with larger total income (larger size) are more likely to obtain a type of tax-exempt status (the log total income, OR = 1.126, p < 0.1). CFs with more years in existence are slightly less likely to obtain a type of tax exemption qualification (OR = 0.915, p < 0.01).

When examining CFs of various specialization types, results in Column (1) of Table 3 indicate that the odds ratio of CFs focusing on education development is insignificant. However, CFs focusing on charity aids are about 2.7 times more likely to obtain a type of tax exemption qualification (OR = 3.727, p < 0.01). There is only a slight significant regional variation captured in this analysis, except for the issue that CFs in Zhejing Province are, on average, 55 percentage points less likely to have any type of tax exemption (OR = 0.451, p < 0.05). The year effects are significant or marginally significant since the year 2016. Compared with the base year of 2010, a CF is about 2.7 times (OR = 2.722, p < 0.05) more likely in year 2016, or 2.4 times (OR = 2.396, p < 0.1) more likely in 2017, or 2.7 times (OR = 2.688, p < 0.05) more likely in 2018, to report having at least one type of tax exemption qualification.

Column (2) of Table 3 reports the results of the robustness check including the log form of the donation ratio, which slightly deviates from a normal distribution. The results in Column (2) are overall consistent with those of the full model reported in Column (1), though with small differences in the size of the estimates. It is noted that a significant odds ratio for the type of “education development” (OR = 0.582, p < 0.05) is present under this specification. Additionally, year effects for 2016–2018 become insignificant.

As reported in Table 4, a subsample of the CFs that have both tax-deductible and -exempt qualification and those without any tax exemption were analyzed. A dummy variable “tax-deductible and -exempt” was used as the dependent variable, and the base group includes those without any tax exemption qualification.

The odds ratio of the donation ratio is insignificant in the full model reported in Column (1) of Table 4, but the log form remains significant as reported in Column (2) of Table 4 (OR = 1.364, p < 0.01). For this type of CFs, OR of education development type are highly significant (OR = 0.412, p < 0.01). OR of charity aids type are consistent (OR = 4.318, p < 0.05) and larger than results in Table 3. As for the region-fixed effects, large and significant differences were identified and are reported in Column (1) and (2) of Table 4. CFs in Beijing (OR = 14.057, p < 0.01) and Shanghai (OR = 10.300, p < 0.01) are about 14 times and 10 times as likely as CFs in the rest areas of China to have both types of tax-exempt qualification. Meanwhile, CFs in Zhejiang are only 20% as likely to have this qualification (OR = 0.206, p < 0.05). Those in Guangdong Province are not significantly different, despite Guangdong being one of the most economically developed areas in China.

Using a dichotomy variable of “tax-deductible only” as the dependent variable, a subsample of the CFs that have a tax-deductible qualification only and those without any tax-exempt status were analyzed. As reported in Table 5, the main results from this set of tests are similar as those using the full sample and reported in Table 3; Table 4. As for region effects, CFs in the Beijing area are significantly less likely by about 60 percentage points to have tax-deduction status only (OR = 0.340, p < 0.01). Shanghai and Chongqing were dropped from the regression due to insufficient observation numbers. It is noteworthy that, compared with the base year 2010, CFs during 2016 to 2018 are about 2.32 to 2.57 times more likely to have tax-deductible qualification (for 2016, OR = 3.574, p < 0.05; for 2017, OR = 3.515, p < 0.05; for 2018, OR = 3.324, p < 0.05).

6. Conclusions

Applying a model of logistic regression to analyze 767 observations of CFs in China from 2010 to 2018, this study found a generally significantly positive association between tax-exempt status and transparency as well as donation dependency of CFs in China. The estimations of transparency and donation ratio are consistent and robust when the full sample and subsamples were examined, respectively. These findings are consistent with Hypothesis 1 and Hypothesis 2 of this study. As predicted by Hypothesis 3 of this study, CFs with sophisticated tax exemption qualifications (both types of tax exemption) are highly concentrated in Beijing and Shanghai, the top megacities of China. After 2016, when the new Charity Law was implemented, charitable foundations in China were more likely to have the qualification of “tax-deductible only”, but were not more likely to have the more sophisticated type of both tax exemption qualifications since stronger incentives are provided for tax deductions only. Partially supporting Hypothesis 4 of this study, these findings provide detailed information about the change of CFs’ tax-exempt status in China.

These findings indicate that the CFs have responded actively to the tax deductible incentive provided by the 2016 Charity Law of China, which in turn requires a higher level of transparency from CFs. In this way, this study has provided empirical evidence that tax exemption is an effective mechanism to enhance the transparency of CFs in China as well. Hence, like peers in Western countries, taxation under the legal system in the long run may also have an effective function of promoting the transparency and sustainable development of charity foundations in China.

The major limitation of this study is survivor bias due to data availability of charity foundations in China. The survivor bias in this study is due to the fact that foundations with good governance and healthy financial status are more likely to disclose their information with accuracy and in complete ways. Data of CFs with incomplete variables or inaccuracy values eventually were dropped from the sample and worsened the survivor bias. The financial reports disclosed by charity foundations in China are not required by law to be audited by professional accountants and many reporting errors were identified during the data cleaning processes. For example, the donation income of some foundations was reported to be greater than the total income. In some cases, costs, expenses, or even donation income were recorded as negative values when obtained from the database. The final sample is biased toward the large foundations with acceptable governance and transparency.

Second, this study uses panel data models to examine the associations, rather than causal effects between tax-exempt status and the characteristics of charity foundations in China. Hence, there are concerns about self-selection bias, which means in this study that foundations with good governance, transparency or social resources are more likely to obtain the tax exemption qualification.

In addition, there are concerns about bias of unidentified variables. For example, the foundations in Beijing and Shanghai are very likely to obtain both tax-deductible and tax exemption qualifications. However, what are the fundamental reasons and characteristics associated with this situation? There have been some studies about social capital, networking effects, and political connections of foundations in China [30]. Due to limited data information, this study cannot explicitly identify and control for those unidentified variables behind these regional disparities.

In short, the results reported in this study are overall robust but should be generalized with caution. Due to potential survivor bias and self-selection bias as discussed above, the effects estimated in this study may be slightly larger than those if the true population of the foundations in China could be examined. In our future study, a propensity score matching method will be adopted to address the concern about self-selection bias and further study potential causal effects of tax exemption qualification on the governance of CFs in China.