Examining the Relationship between Different Types of Information Disclosure of Foundations and Chinese Donations

1 Introduction

Prior to 1978 or the economic reform of China, non-governmental organizations (NGOs) in public welfare were non-existent. After four decades of rapid yet steady economic development, almost 5000 foundations were in place at the end of 2015. The total amount of donations received was about 34.9 billion RMB (5.7 billion USD) with 104 billion RMB (16.9 billion USD) in total net assets. However, compared to the well-established philanthropy industry in the western countries, the philanthropy industry in China continues to lag behind its western counterparts, and is suffering from strong setbacks due to the lack of transparency. Although the Chinese government has passed several regulations that require foundations to disclose information, they are not mandatory and only provide guidance without a monitoring presence and disciplining process, which reduce inclination to disclose information (Dang 2015). In 2011, the notorious “Guo Meimei” scandal greatly impacted the reputation of the Chinese philanthropy industry as the incident created doubt on the information transparency of donation money use.^[1]

Anecdotal stories report that the Chinese are less inclined to donate if they suspect that a charitable organization is hiding important information.^[2] Some research has provided empirical evidence showing that Chinese donations are positively associate with the information disclosure of foundations. However, as foundations are disclosing all kinds of information, what type of information disclosure is more associated with Chinese donations? This is an interesting topic, but the evidence is still largely absent. In this research, we aim to fill in this research gap.

Many studies in the literature have focused on a list of related variables obtained from released financial reports that may affect charitable donation behaviors. They include but are not limited to the: price of the donation (Hyndman 1991), program ratio (Baber, Roberts, and Visvanathan 2001), administrative ratio (Greenlee and Brown 1999), adequacy of equity (Trussel and Parsons 2004, 2007), revenue concentration (Greenlee and Trussel 2000), operating margin (Greenlee and Trussel 2000), level of administrative costs (Greenlee and Brown 1999), other revenues (Posnett and Sandler 1989), fundraising expense (Weisbrod and Dominguez 1986), organization age (DiLorenzo and Bennett 1994), organizational size (Tinkelman 1999), and government grant support (Roberts, Smith, and Taranto 2003). The variables in these studies, most of which are based on philanthropy practices in the U.S., enhance current understanding on how different financial variables affect charitable donations. However, these financial variables cannot represent the transparency of foundations.

One of the related strands in the literature shows that ratings of charitable organizations provided by a third party have a significant impact on donations (Chhaochharia and Ghosh 2008; Gordon, Knock, and Neely 2009; Grant 2010; Harris and Neely 2016; Sloan 2009). However, this strand of studies focuses on the third party evaluation instead of examining how donors respond to the disclosure behavior of foundations.

Another strand of literature has directly studied the relationship between the donations and foundation information disclosure. Many papers have confirmed the positive relationship between the donations and the general information disclosure of foundations (Haski-Leventhal and Foot 2016), or the web information disclosure (Blouin et al. 2018; Saxton, Neely, and Guo 2014), or the accounting information disclosure (Buchheit and Parsons 2006), or the financial information disclosure (Blouin, Lee, and Erickson 2018; Parsons 2007). The consensus in the literature is that more disclosed information does bring more donation. However, as in reality foundations are revealing all kinds of information such as basic information, financial information, project information, and their own governance information, etc., one interesting follow-up question would be what type of disclosed information might be more effective in affecting people’s behavior. Nevertheless this question is not fully addressed as most of literature more focus on general or specific type of information disclosure, and lacks of examining and comparing how different type of information disclosure is associated with charitable donation.

In the literature, there are few studies on how the transparency of foundations affects donations in China. Deng, Lu, and Huang (2015) use survey data on 800 NGOs in human services and find that their donation income is positively related to their level of transparency. However, the scope of the survey data that they use is limited as they only focus on a subsample of the organizations in the charitable fund industry in China, and neglect many influential charitable fund organizations in China such as the Chinese Red Cross Foundation, China Soong Ching Ling Foundation, etc. Nie, Liu and Cheng (2016) examine how various factors might affect the information disclosure behaviors of foundations. They find that less government support and more donations will compel Chinese foundations to disclose more information. Ni and Zhan (2017) show that embedded government control might help foundations to receive more donations. Wei (2017) also explores factors that influence donation revenue to Chinese foundations from the perspective of government control. However, none of them distinguishes the types of information, and compares and examines how different type of information disclosure is associated with Chinese charitable donation.

In this research, we aim to fill in this research gap by exploiting a unique dataset constructed by the China Foundation Center (CFC).^[3] This dataset offers a unique measurement: the Foundation Transparency Index (FTI) which reflects the transparency of Chinese philanthropic foundations. There are 41 items for consideration, which reflects the transparency of four types of information: the foundation’s basic information (e. g., address, mission statement, etc.), the financial information (e. g., revenue, assets, etc.), the project information (e. g., project location, expense, etc) and the governance information (e. g., corporate governance rules, asset management policies, etc.).^[4]

A large sample of Chinese foundations is used in this study and the results first confirm the finding in the literature that there is a significant and positively robust relationship between the general transparency score of foundations and donations in China. More importantly, we examine and compare how different types of information disclosure affects charitable donation and find that Chinese are more sensitive to the disclosure of foundation’s governance information.

This study makes two contributions to the literature as follows. First, the work contributes to the growing body of literature on the Chinese philanthropy industry. Although this industry in China is growing at a steady pace, most of current research is still based on practices in the western countries. The current research will help to provide some evidence on the behaviors of Chinese foundations. Specifically, this research shows that among all types of information disclosure, Chinese donations seem to be more sensitive to one specific type of information disclosure: foundation’s governance information disclosure.

Second, the current research work has important policy implications for both policy makers as well as practitioners in the philanthropy industry. We argue that Chinese authorities should consider more mandatory policies that demand more transparency and information disclosure by Chinese foundations, especially the information related to the fund governance, hopefully to increase the participation of the Chinese in charitable activities and encourage their donations to promote the continual advancement of the philanthropy industry in China.

The remainder of this paper is organized as follows. Section 2 provides a brief introduction on the development of the Chinese philanthropy industry and the CFC which provides the FTI. Section 3 describes the research methodology. Section 4 provides the dataset. Section 5 presents the study results and Section 6 concludes.

2 Industry Background

2.1 Rapid Development of Philanthropy Industry in Post-reform China

After the establishment of People’s Republic of China in 1949, the Chinese government provided universal employment, basic welfare and disaster relief to all members of society, and completely eliminated non-governmental public welfare organizations. The economic reform and opening-up policy in 1978 however changed this situation, so that foundations and charitable activities emerged in response to help those who are in need as the government no longer provides a universal welfare system.

Many non-governmental foundations have emerged since 1980. Foundations are now playing an increasingly more important role in social development in China. Chinese social philanthropy efforts highlight social charity infused with love and vitality, and as a result, the sector has made steady progress in a period of less than 40 years. Figure 1 shows the rapid growth in the number of Chinese foundations since 1981 when charitable services received legal recognition. As of December 31, 2015, the total number of charitable foundations in China is 4954, with an average annual growth rate of 18% since 2004.

Figure 1:

Growth trend of number of foundations in China. Source: China Foundation Center. Note: Number of foundations increasing at a rate of 18% since 1981. At end of 2015, the number of foundations in China was 4954.

3 Regression Model

A market where one party (seller) has more or better information about products over another party (buyer) creates information asymmetry (Akerlof 1978). More information disclosed or available in the market may mitigate asymmetric information problems and reduce transaction costs. Signaling is a widely used means to mitigate uncertainty and information asymmetry (Spence 1978).

In the context of charitable donations, charities disclose information and send signals to the public to attract donations. Harris and Neely (2016) argue that donors use disclosed information as creditable signals of charity success and conclude that charities that openly disclose information receive more donations than those that do not openly disclose details.

The literature has provided much empirical evidence which confirms this relationship. This study aims to push the research beyond the current frontier by examining and comparing the relationship between the behaviors of Chinese donors and different types of information disclosure of Chinese foundations. We decompose the FTI scores into four sub-dimensions: which are the transparency scores for foundation’s basic information, financial information, project information and governance information. These four variables will be selected as the main independent variables in the regression model.

Control variables selected for the regression are based on findings in the literature. Financial report related variables may affect charitable donations as per Trussel and Parsons (2004, 2007). Therefore, we follow Trussel and Parsons (2004, 2007) and apply their control variables which include: fundraising expenses, fundraising ratio, adequacy of equity, operating margin, revenue concentration, other revenue, organizational size, organization age, program ratio, administrative ratio, grant revenue and program revenue. The definition of each control variable is provided in Table 1.

To examine the relationship of donation with different types of information disclosure, the following OLS regression is adopted:

where Don is revenue from donations in the regression model, i represents the different foundations, t represents information in 2015, and t−1 represents information in 2014. The reason for the years 2014 and 2015 is to reflect the two sets of data used here: the financial reports of all the charitable foundations registered with the CFC in 2014 and 2015. BASIC_t−1 is the FTI score for disclosing foundation’s basic information in the previous year; FINANCIAL_t−1 is the score for disclosing foundation’s financial information; PROJECT_t−1 is the score for disclosing foundation’s project information and lastly, GOVERNANCE_t−1 is the score for disclosing foundation’s governance information in the previous year.

The key regression result of interest is β₁–β₄. If the coefficient is significantly positive, it means that the level of transparency for the corresponding type of information shown by the FTI score is positively associated with Chinese donations. We are interested in which type of information disclosure is associated with Chinese donations.

4 Data

As stated earlier, two sets of data are used in this study. One set includes the financial reports of all of the charitable foundations registered with the CFC in 2014 and 2015, which are used to calculate the control variables for the regression analysis and were downloaded from the CFC official website. The financial reports include a statement of financial position, profit and loss statement, and statement of cash flow. The original data contain an abundance of information, including annual revenue; revenue from donations, investments, programs, government grants and other; annual, program, general and administrative, fundraising, and other expenses; total and net assets; and so on and so forth. Descriptive information on the key variables is presented in Table 2.

The other set is information that demonstrates the transparency of 3626 foundations in 2014, which includes their FTI, e.g. total score and the score for each of the four components of the FTI: basic, financial, project and governance information. The details on the information for the 41 indicators are also included.

The descriptive information on the key variables for 2015 is presented in Table 3. The two sets of data were merged by using an object identifier (OID) to test the hypothesis.

Table 3 reports the summarized statistics of the key variables: the donation variables, the FTI and the transparency score of four components: basic, financial, project and governance information. The mean of FTI_t−1 is 56.56 and the median of the FTI is 52.8. This could be taken to mean that the current information disclosure of foundations in China is still lacking, and according to the scores, about half of the information has not been disclosed by the foundations. For each category of information disclosure, there is also significant variation. For instance, the score of governance information disclosure has a standard deviation of 5.7 with the maximum value of 23.6 and minimum value of 0.

Table 3 also reports the summary statistics for the donation value. The mean value is 11,300,000 which is much greater than the median value of 1,300,000. This suggests that the distribution donations is highly left-skewed due to the relatively few large donations. In the regression, we take the log-transformed value of the donations. Figure 2 plots the histogram of the log-transformed donations. The distribution is much more symmetric and resembles the normal distribution.

Figure 2:

The histogram of the log-transformed donation.

5 Results

Table 4 shows the correlation of all of the independent variables which are used in the regression analysis. The FTI score of the previous period and the donations received in the current period are relatively highly correlated. This suggests that there might be a trend in the impact on the donations. Therefore, in the regression, the donations in the previous period are also controlled.

Table 5 presents the baseline regression results. First, the regression results show that the FTI score of the governance information in the previous year is significantly positively related to donation revenue in the current year, and the coefficient value is 0.0376 at the 0.001 significance level. Meanwhile, the regression coefficients of the basic information, financial information and the project information are all insignificant. All these suggest that among all four types of information disclosure, Chinese donations are more sensitive to the disclosure of the foundation’s governance information.

The regression results in Table 5 also show some other interesting information. First, the donations in the previous period are significantly correlated with the donations in the current period, which suggests that there may be an impact of the trend. However, even if donations in the previous year are controlled, the coefficient of the FTI score on current donations is still significantly positive. Table 5 also shows that the 12 control variables have a complex relationship with current year donations. Some are insignificant, while others may have an ambiguous relationship with donations. Actually, these complex and mixed results are expected as pointed out by Trussel and Parsons (2004, 2007).^[6]

The main focus of this study is to examine and compare the relationship between different types of information disclosure of the foundations and Chinese donations. Baseline regression results show that only the coefficient of GOVERNANCE is positive. To examine the stability of the result, especially to address the concern whether the coefficients are affected by the inclusion of the 12 control variables, a robustness check was carried out by running a series of regressions and the control variables were added consecutively.

Table 6 shows the results for this series of regressions. The main dependent variable is still the donation amount received in the current year. The BASIC, FINANCIAL, PROJECT and GOVERNANCE are the key variables of interest. The first column shows the result of the first regression in which only these four variables are included as the regressors. Then, the other control variables are added consecutively to examine the robustness of the estimation results of β₁–β₄ against the inclusion of the control variables. Altogether, 14 regressions were run and the results are shown in Table 6.

First, the regression results in Table 6 show that the estimation results of β₁–β₄ is robust. The inclusion of more control variables does not change the sign of the estimation results of β₁–β₄. The estimation results of β₁–β₃ are always insignificant. Meanwhile, the estimation result of β₄ always remains significantly positive. While the values of β₄ change slightly, most remain near 0.0375. Therefore, this robustness check of the regression results lends strong support to the conclusion of the baseline regression.

We further examine the heterogeneity of the relationship by dividing the sample into private foundations and public foundations. In China, there are two types of foundations: public foundations and private. We are interested in whether the baseline results still hold when we examine specific type of foundations. Table 7 reports the regression results. The first column shows the regression results for the private foundations and the second column shows the regression results for the public foundations with similar control variables as in the baseline regression. The results are also similar compared with the baseline regression results. The estimation results of β₁ to β₃ are both insignificant. Meanwhile, the estimation result of β₄ always remains significantly positive. This suggests that the foundation type does not play an important role in Chinese donation behavior.

The robustness of the results was further examined by adopting the Tobit Model. Since there are substantial proportion of the foundations that did not receive donations in 2015, we adopt the standard Tobit model to address the issue that some foundations received no donation in 2015. The function formula of the Tobit model follows the baseline regression function. In the Tobit model, the censored dependent variable is observable if the donation of the foundation is greater than zero in 2015; otherwise, the value of the censored dependent variable is not observed if the donation of the foundation is zero. Table 8 reports the regression results. The results show that β₁–β₃ are insignificant; and the estimation result of β₄ is significantly positive, which is consistent with the baseline regression results.

In sum, the results of above regressions suggest that Chinese donations are not significantly correlated with the disclosure of the basic, financial and project information of the foundations, but highly correlated with the disclosure of foundation’s governance information.

6 Conclusions

The philanthropy industry in China is becoming increasingly more prominent in public benefit activities as a supplementary means to the efforts of the state or local governments, in areas such as healthcare, education, natural disaster relief, rights of women and children, etc. Both the policy makers and the practitioners have also realized the importance of foundation information disclosure on the donation behaviors.

The literature has fruitful findings in showing that foundation information disclosure has strong impact on people’s donation. The consensus in the literature is that more disclosed information does bring more donation. Then, a more interesting and important question, which should be answered, is what type of disclosed information might be more effective in affecting people’s behavior, given the fact that foundations are revealing all kinds of information such as basic information, financial information, project information, and their own governance information, etc. In another word, current literature more focus on general information disclosure, and lacks of examining and comparing how different type of information disclosure is associated with charitable donation.

In this study, the aim is to fill in this research gap and to empirically investigate and compare how different types of information disclosure of Chinese foundations are correlated with Chinese donation behavior by examining a unique dataset of the CFC. By utilizing a large sample of Chinese foundations, the analysis results show a significant and robust result that Chinese donors are more sensitive to the disclosure of the governance information of foundations, and donations are not significantly correlated with the disclosure of the basic, financial and project information of foundations.

The evidence provided in this study also has meaningful policy implications for practitioners as well as the policy makers in the charitable industry. For instance, from industrial practitioner’s perspective, this finding gives them suggestive direction on what type of foundation information they should focus on to disclose. As for policy makers, it is recommended that they amend related laws or regulations so that more governance information about foundations can be released.

One of the limitations of this study is the relatively short time period of the data. Data that span a longer timeframe will be more robust in explaining for the donation behaviors towards Chinese foundations. Another possible research question might be to examine the impact of foundation transparency on the donation source diversification. However, due to the limitation of the data availability, current research cannot address this question. This interesting topic will be left for future study.