Taiwanese society has many nonprofit charitable organizations (NPCO) that deliver a wide range of humanitarian services to residents. These NPCO rely mainly on the generosity of the public to fund their services. As a government incentive to encourage the general public to give to good causes, Taiwanese government taxation codes provide incentives via tax deductions for donations made to NPCO. Taiwanese people give significantly to nonprofit organizations. According to the 2004 Taiwan Social Change Survey made available on the Internet (Academia Sinica, 2004), nearly 38% of the surveyed participants reported an average cash donation of NT$7,969 (roughly US$250, in relation to a per capita income of US$13,982) to NPCO during the calendar year 2003. However, of the surveyed participants who contributed to charity, only 15% indicated they knew how the recipient organizations spent their donated money (Academia Sinica, 2004).

Current accounting rules in Taiwan require nonprofit entities to provide information about the stewardship of their organizations. Two types of accounting information are relevant: a) financial information disclosure to the donors by the recipient NPCO about the money received from donors, and b) financial accountability of the recipient NPCO to their donors. Accounting researchers examining these issues in other countries have documented that individual donors did take into consideration disclosure and accountability as factors influencing decisions to donate (Gordon & Khumawala, 1999; Greenlee & Brown, 1999; Parsons, 2007; Posnett & Sandler, 1989; Tinkelman, 1998; Weisbrod & Dominguez, 1986), and in a study carried out in the US Trussel and Parsons (2007) identified efficiency, stability, information available, and reputation of the organization as the four factors that influenced donors’ decision to give. However, Parsons (2003) reported that there is a lack of understanding of NPCO financial reporting in charitable giving in the US and called for research to investigate the value-relevance of financial reports and accounting information in donors’ giving behavior. The same situation may also apply in Taiwan.

However, another factor that strongly influences donations in western societies is religiosity, as is clearly demonstrated by the Giving USA Foundation estimate of giving by individual donors to religious congregations and religious organizations in 2008 as $106.89 billion, a figure that accounts for 35% of the total given to NPCO in the US (U.S. charitable giving estimated to be $306.39 billion in 2007, 2008). Nevertheless, despite a literature search we could not find any studies in which the effects of financial information on potential donors’ decision to give in conjunction with the effect of religiosity have been examined. Of course, other factors beside religiosity may also impact on a potential donor’s decision to give. Demographic factors found to be associated with charitable giving include age, gender, marital status, income and education level (Burgoyne, Young, & Walker, 2005; Lee & Chang, 2007). However, in our study we focused on the effects of religiosity on both the decision, and the amount, to give to charity in conjunction with variables pertaining to knowledge of financial information. Thus, our first research question was as follows:
How do religiosity and knowledge of financial information rank in importance in contributing to the individual’s charitable-giving behavior with respect to other demographic variables of gender, age, income, marital status, and self-reported socioeconomic status?

A search of extant literature revealed that the effect of religiosity has not been adequately analyzed in the context of human rationality. The question of whether giving behavior can be seen as an individual’s rational choice or simply an extension of one’s self-perceived religious obligation, in the broad sense, remains unanswered in the current literature. In the theory of planned behavior (TPB), Ajzen (1991) provides a feasible means of assessing whether or not charitable-giving behavior can be considered as a rational and planned behavior. If the TPB can be demonstrated to explain an individual’s giving behavior, then this behavior should be treated like any other rational human behavior, and differentiated from irrational behaviors such as gambling and substance abuse behaviors.

According to the TPB, a planned behavior is seen as being driven by the intention to perform such a behavior. However, an intention to perform a behavior is differentiated from the target behavior itself. The intention may or may not lead to the target behavior itself, depending on circumstantial conditions. Three antecedents – attitude, subjective norm, and perceived behavioral control – are linked to behavioral intention, which, in turn, contributes to the target behavior. According to the TPB, if an individual evaluates the target behavior as positive (attitude), if he or she thinks significant others would want him/her to perform the behavior (subjective norm), and if he/she perceives that the target behavior is within his or her control (perceived behavioral control), he/she will have a high intention to engage in the target behavior. However, since behavioral intention does not always lead to actual behavior because of circumstantial limitations, perceived behavioral control is postulated also to have a direct impact on the target behavior (I do it because I can). Religiosity may be viewed as an antecedent to developing an attitude toward giving. Religiosity may also influence the subjective norm for religious people because their significant others will most likely include individuals who are of the same religious persuasion.

Our second research question was:

Can charitable giving be viewed, according to the theory of planned behavior, as a rational behavior for both religious and nonreligious people?

From a sociological perspective, one might presume that more frequent religious participation would naturally translate into a greater tendency to give for charitable purposes. Does the relationship between religious participation and giving vary according to the religion to which an individual is affiliated? This question alludes to the moderating effect of religious beliefs on the relationship between religious participation and charitable-giving behavior, and gave rise to our third research question:
Does religious participation predict differently 1) the decision to give, and 2) the amount to give, depending on the individual’s religious affiliation?

Method

Instrument

A questionnaire was constructed for this study to measure Taiwanese adults’ attitudes toward charitable-giving behavior using the framework of the theory of planned behavior (TPB; Ajzen, 1991). Attitude toward charitable giving was measured in the form of a differential semantic scale by means of four pairs of adjectives: unnecessary-necessary, useless-useful, stupid-smart, impractical- practical. Subjective norm score was measured by three items that express the perceived social pressure regarding making a donation to NPCO from three sources: spouse, important relatives, and friends. Perceived behavioral control over giving was measured by three statements to which the range of possible responses varied from strongly disagree to strongly agree. The statements were: “I am financially capable of giving for charitable purposes”, “I have the final say in deciding to give or not to give”, and “I am in reasonably good financial circumstances”. All responses to attitude, subjective norm, and perceived control over giving items were expressed on a 5-point Likert scale from strongly disagree (1) to strongly agree (5) with subscale Cronbach’s α coefficients computed to be .948, .849, and .783, respectively. Giving behavior was assessed in two parts: the total amount of donations made to charity organizations in the last year and a question asking whether or not the respondent intended to give to charity in the next 12 months.

In addition to the TPB constructs, information on age, income, gender, self-perceived socioeconomic status, job, and religious affiliation were also requested in the questionnaire. Religious participation was assessed by the frequency of taking part in religious activities per week on a 5-point scale, with 1 indicating no participation at all and 5 indicating more than four times. The income question on the questionnaire referred to the average monthly disposable income measured in units of NT$100. Questions on both income and amount of giving were expressed in the fill-in-the-blank format and could reasonably be treated as possessing ratio scale qualities.

Eight items designed to measure the respondents’ perception of financial management of their donated money in relation to disclosure and accountability were presented in a checklist format. The eight questions were: “Does the NPCO disclose financial information publicly and/or periodically?”, “Do you seek to understand the financial management of your donated money?”, “Do you think the NPCO is under obligation to disclose financial information to donors?”, “Does the NPCO use donated money effectively?”, “Do you make an effort to inquire about the use of donated money?”, “Is the financial statement audited by a certified public accountant?”, and “Is the NPCO under obligation to have financial statements audited by a certified public accountant?” The first four questions measure the respondents’ perception of the financial information disclosure of their donated money (α = .808) and the last four questions measure their perception of the NPCO’s financial accountability to the donors (α = .716). The summed subscale scores ranged from 0 to 4 and were later transformed to a 5-point Likert scale ranging from not important at all (1) to very important (5).

Respondents

Each questionnaire was distributed to an income-earning adult family member by students enrolled in an undergraduate level psychology class during the 2008 fall and 2009 spring semesters. There were 423 questionnaires collected. After deleting those that were incomplete, the final sample contained 410 valid questionnaires (243 females and 167 males). The average age of the respondents was 39.78 (SD = 11.673), with the youngest respondent aged 20 and the oldest aged 78. The respondents’ occupations covered a wide range: public sector (n = 94), business (n = 42), private service-related jobs (n = 173), agriculture/fishing industry (n = 3), and others (n = 93). The self-reported religious affiliation was broken down into four groups: 112 (29.2%) respondents indicated they identified with no religion, 100 (26.1%) reported themselves as Buddhists, 116 (30.3%) as folk religion followers, and 55 (14.4%) as Christians. The religious affiliation of 27 respondents was coded as missing, including six I-Guan-Tao followers and various other unidentifiable religions. Thirty-one people did not answer the income question and 12 people did not respond to the total amount of giving question, so that data on their respective income and giving amount were coded as missing. Because of the missing values, effective sample size varied across different analyses.

Statistical Analyses

The relative importance of religiosity and financial information with respect to the amount of giving was assessed by entering the religiosity variable (average weekly frequency of participating in religious activities), financial information variables (disclosure and accountability), along with other demographic variables (gender, age, marital status, socioeconomic status, and income) into the stepwise regression procedures with .10 as the specified significance level to be retained in the model. The logarithmic transformations of age, income, and amount of giving were used instead of the original raw values.

A path analysis statistical procedure was performed by using the SIMPLIS command language (Jöreskog & Sörbom, 1993) to test the null hypothesis that charitable-giving behavior is a rational and planned behavior with the total amount of charitable donations made in calendar year 2008 as the target behavior. Unlike most inferential statistical procedures in which the rejection of a null hypothesis is desirable, we were hoping for a nonsignificant result because the null hypothesis stated that giving behavior is rational.

The effects of religion type and religious participation on charitable giving were subjected to a general linear modeling (GLM) procedure with religion type as a categorical variable and participation as a continuous variable. An interaction term was included to allow for the test of moderating effect of religion type on the relationship between religious participation and charitable giving. The relationship between charitable giving and religious participation was subjected to logistic regression analysis when giving behavior was viewed as dichotomous (give or not give) and the predictability of religious participation in predicting charitable-giving behavior was evaluated by odds ratios of two probabilities: probability of giving and probability of not giving.

Results

Descriptive Statistics

The monthly disposable income of respondents was distributed with a high degree of positive skewness (7.91) marked by a mean of NT$32,994 (US$1,030) and a median of NT$20,000 (US$625). The highest reported disposable income was NT$800,000 (US$25,000), and the lowest was nil. A quarter of the respondents reported disposable income of less than NT$10,000 (US$313). Out of 410 surveyed respondents, 270 (approximately 66%) made donations for charity purposes in the calendar year 2008. The original amount given was also distributed with a high degree of positive skewness (8.86) with a mean of NT$10,555 (US$ 330) and a median of NT$1,600 (US$50). A substantial number of the respondents (n = 147, 36%) gave nothing to charity in that year.

The age, income, and amount of giving variables were transformed by taking the natural logarithms of the original data. The transformed income data was distributed as slightly negatively skewed (skewness = -.315) with roughly equal mean and median (5.24 and 5.29, respectively) and a standard deviation of 1.465. The log transformed giving data was distributed as slightly positively skewed (skewness = .23) with roughly equal mean and median (3.93 and 3.91 respectively) and a standard deviation of 1.464. No significant differences in the amount of given were detected with respect to the effects of gender. No significant association between self-reported religious affiliation and other demographic variables (gender, job, and socioeconomic status) was detected when chi-square tests of independence were applied.

In regard to the financial management of their money donated to NPCO, over 50% of the respondents answered “Don’t know” to the question on whether financial statements issued by NPCO had been audited by certified public accountants. Almost 50% said they did not know whether the financial statements were disclosed publicly and/or periodically. We were interested, but not surprised to find that of the 278 people that indicated they could understand the operation of financial management of their donated money, only 53 people said that they made an effort to inquire about the use of donated funds.

The distributions of financial information scores appeared to be symmetrical for disclosure (skewness = .097, M = 3.556, SD = 1.263) and slightly positively skewed for accountability (skewness = .332, M = 3.064, SD = 1.120). The distributions of TPB construct scores for attitude and perceived control over giving were somewhat negatively skewed (skewness equal to -.495 and -.384, respectively). The means (and standard deviations) of attitude and perceived behavioral control were 3.721 (.981) and 3.536 (.912), respectively. These descriptive statistics indicated that, more often than not, Taiwanese adults’ attitude toward charitable giving was positive and they perceived that they were in financial circumstances that meant they were capable of giving.

Relative Importance of Variables

Gender, marital status, self-report socioeconomic status, and perceived financial accountability of the NPCO to the donors did not reach a level of significance sufficient to be included in the final model for the prediction of the amount of giving. Both forward and backward stepwise regression procedures produced identical results in terms of the set of variables selected into the final model. Age, income, religious participation, and perceived financial information disclosure were the significant predictors. The results of the final steps of the stepwise regression procedures are summarized in Table 1. As seen, religious participation was the most important predictor for the amount of giving. All significant variables positively predicted the amount of giving.

Planned Behavior

The TPB model was first applied to the whole sample and then separately according to the religious beliefs of each of the four groups. All path models were subjected to chi-square tests for goodness of fit with two degrees of freedom. The TPB model held good in all five SIMPLIS (Jöreskog & Sörbom, 1993) runs as indicated by nonsignificant chi-square values (ranging from 1.87 for Christians to 5.92 for the whole sample) and high adjusted goodness-of-fit indices (ranging from .86 for Buddhists to .96 for the whole sample). These results show that charitable-giving behavior can, indeed, be viewed as a rational and planned behavior for both religious and nonreligious people.

In the whole sample, except between subjective norm and intention, all path coefficients were highly significant in the direction we predicted. For the most part, these path coefficients showed a similar pattern when separately analyzed according to each religious grouping: attitude positively predicted intention to give and intention to give positively predicted actual giving behavior; perceived behavioral control exercised both direct and indirect effect on the actual giving behavior. None of the path coefficients in the Christian group proved to be statistically significant because of the lack of power associated with the small sample size.

Moderating Effect of Religion

A preliminary test showed no significant differences in income existed among the four religious groups (F(3, 345) = .68, p < .567). A highly significant interaction effect (F(3, 226) = 10.05, p < .000) in the GLM procedure demonstrated a strong moderating effect of religious affiliation on the relationship between religious participation and charitable giving. The effect of religious participation on the odds of giving to charity and the amount of giving for the four groups according to religious beliefs are summarized in Table 2.

When charitable giving was treated as dichotomous (to give or not to give), the odds of giving conditional on religious participation were significantly greater than were the odds for any religion type. A marginal significance existed in the no religion sample. That is to say, for every unit increase in religious participation the odds of an individual giving to charity would be increased by approximately 68%, even if that individual claimed to have no religious affiliation. The odds were increased by 146% if the person identified him/herself as a Buddhist and by 118% if the person identified him/herself as a Christian. When the charitable-giving variable was viewed as continuous (the amount of giving), the strongest relationship between charitable giving and religious participation was found in the Christian sample.

Discussion and Conclusion

The average total charitable giving of respondents to this survey was NT$10,555 (US$330) in the year 2008. This figure compares favorably to the average value of NT$7,981 (US$249) donated across all socioeconomic sectors in 2004 as reported by Chang (2007) from the Taiwan Social Change Survey. The main findings of this study were the following:

• Religiosity, as determined by weekly participation in religious events, had a stronger effect than all other demographic factors, on charitable giving behavior of the people we surveyed.
• Financial information disclosure had a modest impact on donors’ charitable-giving behavior, but this influence was far less significant than was the influence of religiosity.
• On the basis of TPB criteria, the charitable-giving behavior of the Taiwanese adults we surveyed was rational. This conclusion was valid for adherents of all religions, and also for those who had no religious beliefs.
• Religiousaffiliationmoderatedtherelationshipbetweenreligiousparticipation and charitable-giving behavior. Religious participation increased the odds of giving to charity by 68%, even for those who stated they had no religious beliefs or affiliation. The increase in odds was well over 100% for Buddhists and Christians. With respect to the amount of giving, the strongest positive relationship was detected in Christians and a weak yet significant relationship was detected for Buddhists. There was only a marginal relationship between religious participation and charitable giving for the followers of folk religion.
• Consistent with reports in the US (Parsons, 2003), approximately 50% of the Taiwanese respondents who gave for charitable purposes, irrespective of religious affiliation, had little knowledge about financial disclosure and accountability of the NPCO.

Skeptics may argue that if charitable giving is a rational behavior, then a motivational factor such as attitude toward giving and social factors such as subjective norms should have more influence on giving than do religious beliefs and participation. Indeed, when such hypotheses were subjected to empirical tests of the path coefficients under TPB, Smith and McSweeney (2007) reported that while attitude did significantly predict an individual’s intention to give, it was the past donating behavior that carried the most weight. People with religious faith donate to their own religious group, which increases their giving for general charitable purposes. In this regard, we were interested to note that 82% of those that participated in the study conducted by Smith and McSweeney reported that they identified with a particular religion and 99% of those were Christian. The findings of our study generally confirm their findings and extend it to the multi-faith society of Taiwan.

One point worth mentioning is the finding in our study of a nonsignificant subjective norm effect on charitable giving. There could be two reasons for this failure to find significance. Firstly, charitable giving is typically viewed as a private act and at the time of engaging in such act, individuals may not have an accurate perception of how their significant others would view their action. As a result, the subjective norms as defined in our study may not have as big an influence on charitable giving as predicted by TPB. Secondly, a high percentage of respondents reported a religious affiliation in our survey, and these respondents may have felt a need to distinguish themselves from their significant others because of the perceived religious obligation, especially if their significant others did not share their religious views. The latter speculation could find some support in the literature on the differentiation of religious moral norms from the usual subjective norm concepts in the applications of TPB (Manstead, 2000). Indeed, in more recent research it has been shown that it is moral norms rather than other types of subjective norm that have a direct impact on the intention variable (Smith & McSweeney, 2007; Warburton & Terry, 2000).

Lee and Chang (2007) examined factors affecting charitable giving in the Taiwanese context, without emphasizing the role of religion. They reported that although determinants affecting volunteering of time and services were mostly intrinsic, those for monetary donations were mostly extrinsic. If religious participation is viewed as an extrinsic indicator of religiosity, our results echoed their findings in that the frequency of participation in religious activities was the strongest among all other variables in predicting the charitable-giving behavior taking the form of monetary donations. In future studies of efficacy of religion, there is a need for refinement so that the existing measures of intrinsic or extrinsic religiosity can be more clearly identified, and in doing so this may facilitate a clearer understanding of the influence of religiosity on altruistic behaviors in general.

References

Academia Sinica. (2004). Taiwan Social Change Survey. Retrieved December 3, 2009 from http://srda.sinica.edu.tw/webpages/intTSCStopic.htm

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211.

Burgoyne, C. B., Young, B., & Walker, C. M. (2005). Deciding to give to charity: A focus group study in the context of the household economy. Journal of Community and Applied Social Psychology, 15(5), 383-405.

Gordon, T. P., & Khumawala, S. B. (1999). The demand for not-for-profit financial statements: A model for individual giving. Journal of Accounting Literature, 18, 31-56.

Greenlee, J. S., & Brown, K. L. (1999). The impact of accounting information on contributions to charitable organizations. Research in Accounting Regulation, 13, 111-126.

Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Chicago, IL: Scientific Software International.

Lee, Y.-K., & Chang, C.-T. (2007). Who gives what to charity? Characteristics affecting donation behavior. Social Behavior and Personality: An international journal, 35(9), 1173-1180.

Manstead, A. S. R. (2000). The role of moral norm in the attitude-behavior relation. In D. J. Terry & M. A. Hogg (Eds.), Attitudes, behavior, and social context: The role of norms and group membership (pp. 11-30). Mahwah, NJ: Erlbaum.

Parsons, L. M. (2003). Is accounting information from nonprofit organizations useful to donors? A review of charitable giving and value relevance. Journal of Accounting Literature, 22, 104-129.

Parsons, L. M. (2007). The impact of financial information and voluntary disclosures on contributions to not-for-profit organizations. Behavioral Research in Accounting, 19, 179-196.

Posnett, J., & Sandler, T. (1989). Demand for charity donations in private non-profit markets: The case of the UK. Journal of Public Economics, 40(2), 187-200.

Smith, J. R., & McSweeney, A. (2007). Charitable giving: The effectiveness of a revised theory of planned behavior model in predicting donating intentions and behavior. Journal of Community and Applied Social Psychology, 17(5), 363-386.

Tinkelman, D. (1998). Differences in sensitivity of financial statement users to join cost allocations: The case of nonprofit organizations. Journal of Accounting, Auditing, and Finance, 13, 377-393.

Trussel, J. M., & Parsons, L. M. (2007). Financial reporting factors affecting donations to charitable organizations. Advances in Accounting, 23, 263-285.

U.S. charitable giving estimated to be $306.39 billion in 2007. (2008). Giving USA Foundation. Retrieved October 26, 2009 from http://www.philanthropy.iupui.edu/News

Warburton, J., & Terry, D. J. (2000). Volunteer decision making by older people: A test of a revised theory of planned behavior. Basic and Applied Social Psychology, 22(3), 245-257.

Weisbrod, B. A., & Dominguez, N. D. (1986). Demand for collective goods in private nonprofit markets: Can fundraising expenditures help overcome free-rider behavior? Journal of Public Economics, 30(1), 83-96.

Academia Sinica. (2004). Taiwan Social Change Survey. Retrieved December 3, 2009 from http://srda.sinica.edu.tw/webpages/intTSCStopic.htm

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211.

Burgoyne, C. B., Young, B., & Walker, C. M. (2005). Deciding to give to charity: A focus group study in the context of the household economy. Journal of Community and Applied Social Psychology, 15(5), 383-405.

Gordon, T. P., & Khumawala, S. B. (1999). The demand for not-for-profit financial statements: A model for individual giving. Journal of Accounting Literature, 18, 31-56.

Greenlee, J. S., & Brown, K. L. (1999). The impact of accounting information on contributions to charitable organizations. Research in Accounting Regulation, 13, 111-126.

Jöreskog, K. G., & Sörbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Chicago, IL: Scientific Software International.

Lee, Y.-K., & Chang, C.-T. (2007). Who gives what to charity? Characteristics affecting donation behavior. Social Behavior and Personality: An international journal, 35(9), 1173-1180.

Manstead, A. S. R. (2000). The role of moral norm in the attitude-behavior relation. In D. J. Terry & M. A. Hogg (Eds.), Attitudes, behavior, and social context: The role of norms and group membership (pp. 11-30). Mahwah, NJ: Erlbaum.

Parsons, L. M. (2003). Is accounting information from nonprofit organizations useful to donors? A review of charitable giving and value relevance. Journal of Accounting Literature, 22, 104-129.

Parsons, L. M. (2007). The impact of financial information and voluntary disclosures on contributions to not-for-profit organizations. Behavioral Research in Accounting, 19, 179-196.

Posnett, J., & Sandler, T. (1989). Demand for charity donations in private non-profit markets: The case of the UK. Journal of Public Economics, 40(2), 187-200.

Smith, J. R., & McSweeney, A. (2007). Charitable giving: The effectiveness of a revised theory of planned behavior model in predicting donating intentions and behavior. Journal of Community and Applied Social Psychology, 17(5), 363-386.

Tinkelman, D. (1998). Differences in sensitivity of financial statement users to join cost allocations: The case of nonprofit organizations. Journal of Accounting, Auditing, and Finance, 13, 377-393.

Trussel, J. M., & Parsons, L. M. (2007). Financial reporting factors affecting donations to charitable organizations. Advances in Accounting, 23, 263-285.

U.S. charitable giving estimated to be $306.39 billion in 2007. (2008). Giving USA Foundation. Retrieved October 26, 2009 from http://www.philanthropy.iupui.edu/News

Warburton, J., & Terry, D. J. (2000). Volunteer decision making by older people: A test of a revised theory of planned behavior. Basic and Applied Social Psychology, 22(3), 245-257.

Weisbrod, B. A., & Dominguez, N. D. (1986). Demand for collective goods in private nonprofit markets: Can fundraising expenditures help overcome free-rider behavior? Journal of Public Economics, 30(1), 83-96.

Appreciation is due to anonymous reviewers.

Tungshan Chou, National Dong-Hwa University, No.1, Section 2, Dashiue Road, Hualien 97401, Taiwan, ROC. Email: [email protected]