Overhead costs in Dutch charitable organizations

1

Overhead costs in Dutch

charitable organizations

Bachelor Thesis Economics and Business

Amber Detering

Student number: 10985700

Track: Economics and Finance

Supervisor: Stephan Jagau

Date: 26

th

June 2018

2

Statement of originality

This document is written by Student Amber Detering who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Abstract

This study analyses the relation between the overhead costs and the size of Dutch

charitable organizations. Also, the relation between the executive remuneration and the size of charities is examined. Overhead costs in charitable organizations have been attacked in the last years, since these costs were called excessive. Prior literature showed that the public is willing to donate more to charities with lower overhead costs. Studies which focused on these overhead costs showed that relatively bigger charities, face relatively lower

overhead costs meaning that charities get more efficient. The results in this research, show that the elasticity of overhead with respect to assets is below 1-to-1, which also suggests that relatively bigger charities, face relatively lower overhead costs. This can be explained by economies of scale that arises when a charity grows. Furthermore, the elasticity of executive remuneration with respect to the size of a charitable organization is significantly different than zero, meaning that an increase in size of a charitable organization, also increases the remuneration of the executive.

3

Table of contents

1)

Introduction. ... 4

2)

Literature review. ... 5

2.1)

Charitable giving… ... 5

2.2)

The relation between size and overhead costs. ... 6

2.3)

The relation between size and executive remuneration. ... 7

3)

Research Method ... 9

3.1)

Data. ... 9

3.2)

Methodology. ... 9

3.3)

Hypotheses. ... 11

4)

Results ... 12

4.1)

Descriptive statistics ... 12

4.2)

Regression with respect to overhead costs……….…….12

4.3)

Regression with respect to executive remuneration ... 13

5)

Discussion of results. ... 16

6)

Conclusion. ... 18

7)

References. ... 19

4

1.Introduction

According to a research done by Bekkers, Schuyt, & Gouwenberg (2017), it is found that around 81% of the households in the Netherlands donate money to charity. This can be done by collection, paying a fixed amount per period, legacy or by several other ways. In 2015 more than €5.7 billion was spent on charity in the Netherlands. This is around 0.85% of Dutch GDP.

The amount of money given to charitable organizations rose in the last years (CBS, 2015). In the ‘90 the average money spent on charity was around €250,- per Dutch

household. This amount increased to about €400,- in the year 2012.

Since the last couple of years doubts were formed about the expenditure of charities. Some salaries of CEO’s of charitable organizations were called excessive (Volkskrant, 2004). These charities have been attacked for the excessive salaries given to their CEO’s. For example the CEO of the Dutch charitable organization Hartstichting. The CEO of this charity earned a salary of €170,000 per year.

Also, discussion rose about the money spend on the goal of the organization

(Volkskrant, 2015). With high annual costs, donors criticize that too much money is spent on the building of the organization and the organization itself instead of on the main goal.

Moreover, due to an increasing competition amongst charitable organizations, charities are having higher advertising expenses (Rose-Ackerman, 1982). Charitable organizations spend a high amount of money on fundraising to get as much donations as possible. Fundraising costs lower the expenses that are spent on mission related goals. Therefore, excessive fundraising expenses by charities are being questioned.

According to CBF, about 180 million euro has been spent on administration cost and control and more than 288 million on fundraising expenses in 2016 (n.d.). Furthermore, 11.76% of income of all charities had been received by the health sector and 2.23% by the sector religion in 2016 (CBF, n.d.).

These discussions lead to the aim of this thesis. This research focusses on the overhead costs in charitable organizations and its relation with the size, based on the height of the assets. In particular, it leads to the following research question: What is the influence of the size of a Dutch charitable organization, based on total assets, on the money spend on overhead cost?

In this document, first more literature on this subject is being discussed, focusing on charitable giving and the costs in charitable organizations. After the literature, the research method will be explained, which is divided into data, methodology and hypothesis. In section 4, ordinary least squares regressions are run and results are formed. In the discussion of results, the results and being analyzed and discussed. Lastly, an overall conclusion will be formed and the research question will be answered.

5

2. Literature review

A vast amount of studies focused on charitable giving and the costs of charitable organizations. These studies are discussed in this literature review. This section will be divided in three components: charitable giving, the relation between size and overhead costs of charitable organizations and the relation between size and executive remuneration.

2.1 Charitable giving

Donors prefer donating money to charities that have lower overhead expenses. Overhead costs are seen as the money that does not go to the missions of the charity, which makes donating to a charity with high overhead expenses less attractive. Charitable

organizations which spends a lower percentage on overhead costs, therefore receive more money from donors and get more support (Greenlee, & Brown, 1999; Tinkelman &

Mankaney, 2007).

In a field experiment done by Gneezy et al. (2014), it has been shown that donations are lower when overhead costs increase. In this experiment people were told that the overhead costs were already covered and that their donation would go completely to the mission related goal. The money donated increased by 75% when people were told that their donation would be an overhead-free donation. Money donated and overhead costs are therefore negatively related.

A similar research have been done by Portillo, & Stinn (2018). In this study, they examined whether there were preferences in charitable giving amongst different sorts of overhead costs. The overhead costs they included in this research were fundraising costs and salaries. The results of this research suggested that this preference was also related to an outside option. If people were offered an overhead-free donation, 70-80% of the donors would donate to that charity. When no overhead-free donation was offered, about two-third of the people tend to have a strong preference for their donation to go to fundraising costs instead of money which is spend on salaries.

Caviola et al. (2014) found that an evaluability bias evolves when only a single charitable organization is presented to people. This evaluability bias occurs when people do not have all information about the charitable organization and therefor prefer the

organizations with the lowest overhead cost instead of the organization which is most cost efficient, meaning that an organization uses the money donated more efficient. When only a single charity has been shown, people, in this research, chose for the charitable organization which had the lowest overhead costs. This is due to the fact that people can more easily observe the overhead costs than the cost efficiency. When two charitable organizations are presented and the differences in overhead cost and cost efficiency can be evaluated, people prefer the charitable organization which is most cost efficient. This shows that people prefer cost efficient charitable organizations, but choose for low overhead cost organizations when only a single charity is presented. This can be explained by the evaluability bias, since the low overhead costs are more easily evaluated.

Chen (2009) studied the effect of different factors on charitable giving. From his research it could be concluded that when a nonprofit organization gets bigger in size, measured by number of full-time employees, its public support also gets higher.

Furthermore, a positive relation was suggested between public support and the height of the fundraising expenses. Likewise could be concluded, that when a nonprofits organizations board of directors increases, the donation revenue increases too (Chen, 2009; Aggarwal, Evans, & Nanda, 2012). The study of Chen (2009) also showed that different sectors of

6

nonprofit organizations receive different amount of public support. This study found that the nonprofit organizations in health got a lower support than all other types of nonprofit

organizations, except these operating in law and public advocacy services.

In the profit sector, people are just interested in the quality of the corporations product or service instead of the money earned by an executive. In nonprofit organizations, people are interested in the height of executive remuneration since high levels of

remuneration can be seen as fraud or waste and therefor reduce the demand for services of the nonprofit (Hansmann, 1980). Herzlinger (1994) suggests that, according to the public, an executives of a nonprofit cannot earn the same or more than an executive of a for-profit executive.

What can be concluded from this prior literature, is that people tend to donate less to charities with higher overhead costs. This is in line with the discussions about the excessive overhead expenses of charitable organizations, which were attacked the last couple of years. Also, people tend to disapprove high executive remuneration more than fundraising costs. Furthermore, when a charitable organizations gets bigger, measured by FTE’s and number of directors in the board, its public support increases too.

2.2 The relation between size and overhead costs

In for-profit firms, firms tend to be more efficient when they get bigger (Jovanovic, 1982). This is explained by the fact that they have more market power and can benefit from economies of scale. Smaller firms, on the contrary, have a larger possibility to grow but also larger probability to fail. Also, similar researches in nonprofit organization have been

performed to get a better insight in the relation between size and expenses.

The expenses of charitable organizations are divided into three components, namely, expenses to mission related goals, administration expenses and fundraising expenses. The executive remuneration is allocated over these three expenses since it is seen as an indirect costs. These costs are investigated in a study done by Heijden (2012). This research

examined whether there are differences between small and large Dutch fundraising charities regarding the efficiency of the three costs. This sample included 1196 different Dutch

charitable organizations with differences sizes. The size of these organizations are measured by the height of income. This research found that income has a weak positive relationship with administration expenses which implies that there are no significant differences between small and large charities. This weak relationship means that it cannot be justified as statistically significant. This result is different than the studies performed by Wise (1997), Mourão and Enes (2017), Kahler and Sargeant (2002), and Hyndman and McKillop (1999). In the research done by Wise, 25 small, 25 medium and 25 large charitable organizations in the united Kingdom based on the Hendersons’ Top 2000 Charities 1994, have been used. This study found a negative relation between the ACE ratio, the ratio of administration costs to total expenditures, and total income. This result supported the expectation that larger charities can benefit from economies of scale in administration costs. This implies that when a charity gets relatively larger, based on income, the administration costs of that charity gets relatively lower. Mourão, & Enes did a research on administration cost in Portuguese non-profit organizations. This study likewise proved that larger institutions profit from economies of scale in administration costs. Kahler, & Sergeant also did research on the effect of size on overhead costs of charities. The database they used was the Top 500 Fundraising Charities. However, they only used the charities that were in this top 500 the whole 5 years. This leads to a survivorship bias since the unsuccessful charitable organizations will drop out of this dataset. Yet, their conclusions did also meet the results

7

found by Wise. Smaller charities face proportionally higher administration costs with respect to total expenditure than larger charitable organizations and therefore have lower benefit from economies of scale. Kahler and Sargeant created an inverse function that can be applied when comparing the ACE ratios between charitable organizations. Hyndman and McKillop used the same database and came to the conclusion that there was clear evidence for economies of scale while examining the relation between administration costs as a percentage of total expenses and the size of a charitable organization.

Also, studies on fundraising costs with respect to the size of a charity have been done. Yi (2010) examined the fundraising efficiency in charities. In this research fundraising efficiency is compared with size. The results which were obtained from this research implies that size and fundraising efficiency are positively correlated. This could again be explained by economies of scale. Another cause might be that larger charities could benefit from the preference of donors. Donors tend to see size as a measure for the quality of a charitable organization, and therefore prefer to support and to donate money to larger charities (Olson, 2000; Callen et al., 2003; Chen, 2009). This is because larger charities are seen as more mature, meaning that these organizations already have their position recognized in the market. From this research it is not clear which effect dominates but what can be concluded is that larger public charities tend to be more efficient than smaller public charities (Yi, 2010). Also, the fundraising expenses ratio, which is formulated as the money that an organization spends on fundraising specific materials per dollar spend on wages and salaries on

fundraising specific labor, is positively related to the fundraising efficiency. This suggests that labor intensive charities are the ones that are most efficient.

Sergeant and Kahler (1999) also did research on fundraising expenses. They, again, used the Top 500 Fundraising Charities to form conclusions about the fundraising expenses. These results suggested that there is no significant relation between the fundraising ratio, formulated as the fundraising expenses as a percentage of fundraising revenues, and the size of a charitable organization, except while focusing on corporate fundraising expenses. The relation between corporate fundraising expenses and the size of the charity, could be explained by the fact that large corporations would rather be identified with charities which have a high reputation. Since only the Top 500 charities are used, this dataset might not be completely representative.

The results from the study of Hyndman and McKillop (1999) suggested that the fundraising ratio was positively related to the size of the charities. This positive relation was also concluded by Heijden (2012), meaning that smaller charities have lower fundraising ratios. This could be explained by disadvantages of economies of scale. This implies that it gets more costly to do fundraising activities.

All in all, administration costs seem to increase when size increases, but at a slower pace. This implies that, according to the literature found, a charity gets more efficient, regarding administration costs, when the charity gets bigger. The explanation for this is based on economies of scale. Furthermore, the literature on fundraising costs show different conclusions. Studies shows positive relations, no significant relations and also negative relations with the size of a charity. No clear expectation about the effect of size on fundraising costs can be drawn out of this literature.

2.3 The relation between size and executive remuneration

Several studies on executive remuneration have been performed for the for-profit firms. These studies show that executive remuneration increases significantly when firm size

8

increases (Larson, & Morris, 2014; Tosi et al., 2000; Firth et al., 1996; Singh, & Yavuz, 2015; Nourayi, & Mintz, 2008; Finkelstein, & Hambrick, 1989; David et al., 1998).

Finkelstein and Hambrick (1989) performed a research on 63 American firms in the ‘leisure’ industry. These results showed that when assets increase by 1%, the remuneration of the CEO significantly increased by 0.49%. In another research, done by David, Kochhar, & Levitas (1998), data of 125 firms of the 200 largest U.S. corporations in four different years have been used. This study suggested that a 1% change in assets, increased the level of compensation of the CEO significantly by 0.17%. Firth et al (1999) used data from listed companies in Hong Kong. When the executive earned less than 1 million dollar, this observation has been deleted from the sample. The restricted dataset included 351 observations. Based on this restricted dataset, research has shown that there is a positive relation between size and CEO payment, meaning that a 1% increase in assets, increases the remuneration of the executive significantly by 0.324%.

Notable is, that there is a large variety in how size effects the remuneration of the executive. This could be explained by the differences in datasets that are used. When a sample includes remuneration costs which do not vary a lot amongst the observations, it is a logical effect that size does not has a great influence. When excluding, for example, the observations with executive remuneration lower than 1 million, this creates a measurement bias. Also, when using only data of the largest U.S. corporations, this creates a bias.

All studies show a positive effect of size on executive remuneration. This could be explained by the increased complexity of the tasks or extra responsibility that an executive gets when a firm gets bigger. Also, higher profits are obtained which makes it possible to give a higher compensation to executives.

In a study performed by Oster, it is suggested that the size of a nonprofit organization is an important factor in the compensation of executives (1998). When the size of a non-profit organization increases, the compensation of the executives will rise too. Likewise, executive compensation is different between sectors. When an organization has a higher affiliation with religion, executive compensation will be held down.

However, executive remuneration in nonprofit organizations creates tension (Oster, 1998). Boards need to find a good balance on how to motivate executives and how to reward them. Charitable organizations have continual pressure of the public, who do not approve excessive remuneration costs.

9

3.Research Method

In this section the research method will be discussed. The research method is divided into different components: data, methodology and hypothesis.

3.1 Data

The data, used in this sample, is retrieved from Centraal Bureau Fondsenwerving (CBF). The CBF collects information of charitable organizations which are located in the Netherlands. The same information can be found in the annual reports of the charities. In these annual reports, extra information about the organization is presented. In total, data of 85 charitable organizations has been collected, 54 charitable organizations in the health sector and 31 in the sector religion. The charitable organizations in the dataset are all licensed by CBF. This means that they meet strict quality requirements which also includes the transparency of the charitable organizations. The charitable organizations used in this sample can be found in Appendix 1. Likewise, all the information of the charities used in this sample, can be found there.

This sample includes 8 charitable organizations which spend zero costs on

fundraising. This data will be excluded from the dataset since it can generate a bias in the outcome. When deleting the charitable organizations with zero fundraising costs, 77 charities will be left.

Also, this sample includes an outlier which can be observed in the boxplot, added in Appendix 2. Barnett formulated an outlier as an observation which shows abnormal deviation from the sample (1974). The outlier in this sample occurred because the fundraising costs are extremely low compared to other charities. The annual report of this charitable

organization, Gipsy Mission, explains that they barely spend money on fundraising costs and if these costs are made, that these are sponsored. SPSS produces two different sorts of outliers in their boxplots, extreme and mild ones. The outlier which occurred in this sample, is a mild one. An outlier is mild when the outcome is one and a half times as large as the difference between the first and the third percentile. When excluding the charitable

organizations with zero fundraising costs and the charitable organization Gipsy Mission, 76 charities are left. So the restricted dataset now exists of 76 charitable organizations. 48 of these charities are operating in the health sector and 28 in the sector religion.

Mostly smaller charitable organizations do not have an executive or do not give remuneration to their executive. These organizations are mainly based on volunteers which drives the remuneration costs to 0. These charities are only deleted from the sample when running the regression on the relationship between the percentage of assets spend on remuneration and the total assets. A restricted dataset for the executive remuneration exists of 56 observations. This restricted dataset will be used to run the regressions for executive remuneration.

3.2 Methodology

In this paper, it will be investigated whether the size of a Dutch charitable

organization, based on total assets, influences the money spend on overhead costs. The expenses of charitable organizations are divided into 3 major components: mission-related expenses, administration costs and fundraising costs. The overhead costs, in this research, are being formulated as the expenses that are not spend on the mission. This will be the administration costs plus the fundraising costs of the charitable organization.

10

The expectation will be that the elasticity of overhead with respect to assets shows a below 1-to-1 relation. This could be explained by the fact that when a charitable organization increases, the variable costs will increase but the fixed costs will stay the same. This is the benefit of economies of scale that larger charities obtain.

To answer the research question, an ordinary least squares (OLS) regression will be run. An OLS regression gives the best fitted line for a dataset. It does this by minimizing the sum of square residuals. An OLS regression is used to get an insight in the relation between the dependent and independent variables. After these regressions, a T-test will be done, to show whether the independent variables have a significant effect on the dependent

variables. The dependent variables, the logarithms of overhead, administration and

fundraising costs, will be tested against the null hypothesis. The null hypothesis will be that the elasticity of these dependent variables with respect to size, shows a 1-to-1 relation. This implies that a percentage increase in assets, increases the dependent variable by the same percentage.

The dependent variable will be the logarithm of the overhead costs. The overhead cost will be divided into administration costs and fundraising costs, to have a look at how these non-mission expenses contribute to the total overhead costs. The assumption which is made for administration costs, is that it will be seen mostly as fixed costs. This means that when a charitable organization increases in size, it can spread the fixed costs over the higher assets. This means that the administration costs will increase when the size of a charity increases, but at a slower pace.

The results of previous studies on fundraising costs showed different results. In this research fundraising expenses are expected to have a below 1-to-1 elasticity with respect to the seize of a charity. This assumption is made since it will be easier for a charity to raise money if it already has a good reputation and therefore could spend less on this expense.

As independent variable the logarithm of assets will be used. Also, two control variables will be included. These control variables are related to the dependent variable. Control variables are used to remove their effect out of the dependent variable. The first control variable will be included to examine whether there are changes between the regions that a charitable organization operates in. This control variable is a dummy variable with Europe is 1 if the charitable organization is operating only in Europe; 0 when also other continents get their focus. The other dummy variable researches whether there are differences between the health and the religious sector.

First three regressions will be done without the control variables. These regressions will show the effect of the logarithm of assets on the logarithm of overhead costs,

administration costs and fundraising costs. These regressions are shown below. Regression 1: 𝑙𝑙𝑙𝑙𝑙𝑙 (𝑙𝑙𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑒𝑒𝑒𝑒) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝜀𝜀

Regression 2: 𝑙𝑙𝑙𝑙𝑙𝑙(𝑒𝑒𝑒𝑒𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑙𝑙𝑎𝑎) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝜀𝜀 Regression 3: 𝑙𝑙𝑙𝑙𝑙𝑙 (𝑓𝑓𝑓𝑓𝑎𝑎𝑒𝑒𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑙𝑙) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝜀𝜀

These regressions can be explained as a log-log regression. The coefficients in this model serve as the elasticity of the dependent variable with respect to the independent variable. This implies that a 1% change in the independent variable, results in a β1% change in the dependent variable. In this regression this would be that when assets change by 1%, the overhead costs will increase by 𝛽𝛽1%. Another reason why the logarithm of both variables are taken, is that the dataset is skewness to large values. This occurs when there is a large variety in the height of assets. This can be for example shown when the mean is a lot bigger than the median, which is applicable in this situation. The mean of the assets of the

11

After these regressions have been performed, the control variables are included to see if they change the dependent variables significantly and how the model reacts to these particular control variables. The following regressions will be done.

Regression 4: log (𝑙𝑙𝑜𝑜𝑜𝑜𝑜𝑜ℎ𝑜𝑜𝑒𝑒𝑒𝑒) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝛽𝛽2 ∗ 𝐸𝐸𝑓𝑓𝑜𝑜𝑙𝑙𝐸𝐸𝑜𝑜 + 𝛽𝛽3 ∗ 𝐻𝐻𝑜𝑜𝑒𝑒𝑙𝑙𝑎𝑎ℎ + 𝜀𝜀 Regression 5: log (𝑒𝑒𝑒𝑒𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑙𝑙𝑎𝑎) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝛽𝛽2 ∗ 𝐸𝐸𝑓𝑓𝑜𝑜𝑙𝑙𝐸𝐸𝑜𝑜 + 𝛽𝛽3 ∗ 𝐻𝐻𝑜𝑜𝑒𝑒𝑙𝑙𝑎𝑎ℎ + 𝜀𝜀 Regression 6: log (𝑓𝑓𝑓𝑓𝑎𝑎𝑒𝑒𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑙𝑙) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝛽𝛽2 ∗ 𝐸𝐸𝑓𝑓𝑜𝑜𝑙𝑙𝐸𝐸𝑜𝑜 + 𝛽𝛽3 ∗ 𝐻𝐻𝑜𝑜𝑒𝑒𝑙𝑙𝑎𝑎ℎ + 𝜀𝜀 With Europe is 1 if charitable organization is operating only on Europe; 0 otherwise. With Health is 1 if charitable organization is operating in the Health sector; 0 when operating in religious sector.

Also, a separate regression will be run for the total remuneration of the executive. The total remuneration includes annual income, gross salary, holiday allowance, the extra month, variable pay, social securities, employers part, taxable allowance, pension premium, employers part, pension compensation, other allowance, long term and payment in relation to beginning of end contract of contract. The assumption for remuneration is that it will be positively related to the size of a charity meaning that an increase in assets, increases the remuneration of the executive. This is based on the literature found in nonprofit

organizations, stating that when a for-profit organization increases in size, the remuneration of the executive will rise too. The regressions that will be used to measure the effect of size on executive remuneration are shown below. The first regression will be performed without the control variables but these will be included in the second regression.

Regression 7: log (𝑜𝑜𝑜𝑜𝑎𝑎𝑓𝑓𝑎𝑎𝑜𝑜𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑙𝑙𝑎𝑎) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝜀𝜀

Regression 8: log (𝑜𝑜𝑜𝑜𝑎𝑎𝑓𝑓𝑎𝑎𝑜𝑜𝑜𝑜𝑒𝑒𝑎𝑎𝑎𝑎𝑙𝑙𝑎𝑎) = 𝛽𝛽0 + 𝛽𝛽1 ∗ log(𝑒𝑒𝑎𝑎𝑎𝑎𝑜𝑜𝑎𝑎𝑎𝑎) + 𝛽𝛽2 ∗ 𝐸𝐸𝑓𝑓𝑜𝑜𝑙𝑙𝐸𝐸𝑜𝑜 + 𝛽𝛽3 ∗ 𝐻𝐻𝑜𝑜𝑒𝑒𝑙𝑙𝑎𝑎ℎ + 𝜀𝜀 The dependent variable executive remuneration, will be tested against the null hypothesis. The null hypothesis will be that executive remuneration is not dependent of the size of a charitable organization. This implies that a percentage increase in size, does not change the executive remuneration.

3.3 Hypotheses

The expectations regarding the research question are formulated in the hypotheses below. There are two hypotheses formed to examine the effect of size, based on total assets, on overhead costs and executive remuneration.

Hypothesis 1: The overhead costs of a charitable organization have a below 1-to-1 elasticity with respect to the size of a charity. This implies that relatively bigger charities, face relatively lower overhead costs.

Hypothesis 2: The executive remuneration of a charitable organizations is positively related to the size of a charity. This implies that a size increase, increases the remuneration of the executive.

12

4. Results

This section includes the results of several tests. These tests are used to answer the research question. Different models will be used to give more insight in this research.

4.1 Data descriptive

The dataset which is used for the regressions, is the restricted dataset. This means that the outlier is deleted from the test. Information about the mean, median, standard deviation, minimum, maximum and number of observations of the dependent and

independent variables are shown in Table 1 below. The data statistics including the outlier can be found in Appendix 3.

Table 1. Data statistics with outlier

Variables Mean Median Standard

deviation

Minimum Maximum Observations

Log(overhead) 5.517735 5.57843 0.786874 3.694254 7.353474 76 Log(administration) 5.063915 5.062341 0.709801 3.14145 6.538951 76 Log(fundraising) 5.240618 5.318051 0.92758 2.471292 7.281215 76 Log(remuneration) 4.951919 4.998143 0.264128 4.092931 5.364699 56 Log(assets) 6.327508 6.257221 0.826569 4.797842 8.563358 76 Dummy Europe 0.618421 1 0.489002 0 1 76 Dummy Health 0.631579 1 0.485582 0 1 76

4.2 Regressions with respect to overhead costs

The regressions, discussed in the methodology, are being performed. The outcomes of the first three regressions, excluding the outlier, are shown in Table 2. The first number represent the beta of the independent variable and the number between brackets imply the standard deviation of the independent variable. The stars, next to the standards deviation, indicate the significance level. This significance level shows whether the elasticity of the dependent variables with respect to assets is significantly different from 1-to-1. In Appendix 4 the regression outcomes with the outlier, Gipsy Mission, are presented. As expected, the results got more significant when the outlier is excluded.

13

Table 2. Regression with 1 independent variable

Dependent variables

Log(overhead) Log(administration) Log(fundraising)

Independent variable Log(assets) 0.848 (0.050)*** 0.708 (0.056)*** 0.956 (0.068)

*10% significance **5% significance ***1% significance

As can be seen in Table 2, the beta of the assets, and thus the elasticity of overhead with respect to assets, is significantly below 1-to-1. This implies that relatively bigger

charities, face relatively lower overhead costs. The relation is that a 1% increase in assets, suggests a 0.848% increase in overhead costs. Also, the independent variable assets, shows an elasticity which is significantly lower than 1-to-1 with respect to administration costs. This means that a 1% increase in assets causes the administration costs to rise with 0.708%. Again, relatively bigger charities, face relatively lower administration costs.

Differently, the assets of a charitable organization shows an elasticity of 1-to-1 with respect to the costs of fundraising, meaning that a percentage increase in assets, increases the fundraising costs by the same percentage. This implies that the fundraising costs change equally, relative to a change in size of a charitable organization.

In the second model, control variables are included. These control variables are related to the dependent variable and therefore included in the regression to remove their effect out of the equation. The results of the regressions, including the independent variable the logarithm of assets, dummy variable Europe and dummy variable Health, are shown in table 3.

Table 3. Regression including control variables

𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃 𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐃𝐃𝐯𝐯

*10% significance **5% significance ***1% significance

As can be concluded from the results in Table 3, the elasticity between assets and overhead and administration expenses, are significantly below 1-to-1, respectively with 5% and 1%. This suggests that relatively bigger charitable organizations, face relatively lower

Log(overhead) Log(administration) Log(fundraising)

Independent variables Log(assets) 0.875 (0.051)** 0.748 (0.057)*** 0.986 (0.071) Dummy Europe -0.086 (0.089) -0.088 (0.098) -0.047 (0.123) Dummy Health -0.153 (0.093) -0.228 (0.103)** -0.173 (0.129)

14

costs. The elasticity of fundraising with respect to assets has no significant difference to a 1-to-1 relation. This suggest that a percentage increase in size of a charity, increases the fundraising costs by the same percentage. The control variable health has only an effect on administrations costs with a significance level of 10%. The control variable Europe has no significant effect on these dependent variables. Neither dummy variable Europe nor dummy variable health has a significant effect on the fundraising costs.

When a charitable organization is operating in the health sector, administration costs will significantly decrease with 0.228%. This implies that operating in the health sector has a negative relation with the administration costs of a charity.

Furthermore, when a charitable organization increases assets by 1%, administration costs will rise with 0.748% and overhead costs will rise with 0.875%.

4.3 Regression with respect to executive remuneration

The last two regressions run, are the regressions with the logarithm of remuneration as the dependent variable. For these regressions the restricted dataset for executive

remuneration is used. The restricted dataset for executive remuneration exist of 56 observations. The first of the two regressions shows the relation between the total assets and the executive remuneration. The results can be seen in Table 4. The stars, next to the standard deviation, again, indicate the significance level of the independent variable. This significance level shows whether size has a significant effect on remuneration, so whether the elasticity is different than zero.

Table 4. Regression with 1 independent variable

Dependent variable

Log(remuneration) Independent variable Log(assets) 0.224 (0.037)***

*10% significance **5% significance ***1% significance

In Table 4, a positive effect of size on executive remuneration can be detected. The coefficient of the assets, 0.224, shows that an increase of 1 percentage in assets, changes the remuneration of the executive by 0.224%.

The second regression includes the control variables. These control variables are dummy variable Europe and dummy variable Health. These, again, are related with the dependent variable, the logarithm of assets.The control variables are included to remove their effect out of the equation. The results of this regression are shown in Table 5 below.

15

Table 5. Regression including control variables

Dependent variable Log(remuneration) Independent variables Log(assets) 0.191 (0.040)*** Dummy Europe 0.017 (0.060) Dummy Health 0.114 (0.068)*

*10% significance **5% significance ***1% significance

The results of this regression suggests that the size of a charitable organization has a positive significant effect on the remuneration of the executive. A 1% increase in assets, increases the remuneration by 0.191%.

The control variable health has a positive effect on the dependent variable. This implies that when a charity operates in the health sector, remuneration increases by 0.114% with a significance level of 10%. Lastly, what can be concluded from this regression is that the dummy Europe has no significant effect on the dependent variable.

16

5. Discussion of results

To measure the effect of size on overhead costs and executive remuneration, regressions are performed. The overhead costs are defined as the non-mission related expenses, also known as the administration costs and fundraising costs together. From Table 2 and 3 it can be concluded that the elasticity of overhead with respect to total assets is below 1-to-1. This implies that relatively bigger charitable organizations have relatively lower overhead costs. This is in line with the literature, suggesting that a bigger charity benefits from economies of scale.

To obtain a better idea on where this below 1-to-1 elasticity comes from, regressions with the size of a charity as independent variable, and administration costs and fundraising costs as dependent variables are performed. Studies done by Wise (1997), Mourão, & Enes (2017), Kahler, & Sargeant (2002), and Hyndman, & McKillop (1999), suggested that

economies of scale in administration costs occurs when charities get larger in size. This is due to the fact that administration costs mostly exists of fixed costs. Note that size is measured differently amongst these studies, namely based on total income and on total expenditure instead of on total assets.

When observing the results of the regressions with administration costs as

dependent variable, it shows that the elasticity of administration expenses with respect to assets is below 1-to-1. Meaning that a relatively bigger charity, has relatively lower

administration costs. This result is in line with the literature based on administration costs. Operating in the health sector has a negative influence on administration costs. This implies that when a charity is active in the sector health, the administration costs will go down by 0.228%. A possible explanation for this could be that charitable organizations in the religious sector have more volunteers and therefore lower personnel costs. But to derive conclusions like this, more research is needed.

The effect of size on fundraising costs, on the contrary, does not show the result that was expected. The expectation was that the elasticity of fundraising costs with respect to assets would show a relationship below 1-to-1. However, the results suggests that this elasticity is 1-to-1, meaning that a percentage change in assets, changes the fundraising costs by the same percentage. This might be possible due to fundraising which gets more costly when the charity grows or that charitable organizations need to keep their fundraising expenses high in order to keep the high profile image (Heijden, 2012). Also, increased competition could have driven these fundraising costs higher (Rose-Ackerman, 1982). But again, this cannot be concluded without further research.

The remuneration of the executive is being divided over 3 components of the total expenses, namely, administration, fundraising and mission-related expenses. This percentage to which this remuneration is allocated to the different components differs

amongstthe different charitable organizations. This implies that a part of the remuneration is seen as a mission related expense. This is being question by the public since these costs are indirect costs and therefore not directly related with the mission expenses. This

discussion about executive remuneration creates tension since there is not a clear, correct balance between how to motivate executives and how to reward them (Oster, 1998). This suggests that more research is needed to correctly specify overhead costs in charitable organization.

Moreover, excessive executive remuneration in charitable organizations have been attacked by the public. Therefore, an analysis on executive remuneration has been

17

remuneration of the executive, a significant positive relation can be observed. This implies that an increase in assets of 1%, increases the remuneration of the executive by 0.191%. This is in line with the second hypothesis made. Literature shows that executive

remuneration in for-profit firms significantly increase when the size of a firm increases. If we compare the literature with these results, we can see that both in for-profit firms and in nonprofit organizations, the executive remuneration increases when the organization increases in size. When looking at the numbers of the studies performed on for-profit firms, there is a large difference in the effect of size on executive remuneration. This could be explained by the different datasets which are used, which creates biases. The size effect in for-profit firms, based on the study of Finkelstein and Hambrick (1989), David, Kochhar, & Levitas (1998) and Firth et al. (1999), showed an effect ranging from 0.17 to 0.49. With a size effect of 0.191 on executive remuneration, based on this research, the effect is on the lower side of the interval. This suggest that the executive remuneration in charitable

organizations are on the lower side compared to the executive remuneration of the for-profit sector. Accordingly, remuneration costs seems not to be excessive compared to the for-profit sector, nonetheless, no clear conclusions can be derived.

Based on this research, it cannot be concluded whether it is correct that the

executive remuneration is labelled as excessive, compared to the for-profit industry. In order to form a conclusion for this matter, more specified research on this is needed. This could be done by a meta-analysis which focuses on both executive remuneration in for-profit

organizations and charities.

Also, when a charity operates in the health sector, this has a positive effect on the remuneration. This suggests that an executive gets a 0.114% higher remuneration while operating in the health sector. This matches the results found by Oster (1998) who

suggested that charitable organizations with a higher affiliation to religion, compensate their executives less than other sectors.

People prefer donating money to charities with lower overhead costs (Greenlee, & Brown, 1999; Tinkelman, & Mankaney, 2007; Gneezy et al., 2014; Portillo, & Stinn, 2018). However, when having information about the quality of a charitable organization, a donor prefers the charity which delivers the best quality of the money spend on mission related goals. This can be explained by the evaluability bias (Caviola et al., 2014). From this results of this research you could say that larger organizations are more efficient money wise, but to conclude that it is better to donate money to larger organizations, more specified research on the quality of Dutch charitable organizations is needed.

18

6. Conclusion

Literature show that the level of overhead costs of a charitable organization seems to have a significant effect on charitable giving. Therefore, a better understanding on how overhead costs are influenced by size, should be of great interest. In this research the effect between the non-mission related costs and the size of a charitable organization is examined.

Increased doubts about the overhead costs in charitable organizations led to great discussions. High annual costs of charities were attacked (Volkskrant, 2015), fundraising costs were high due to increased competition (Rose-Ackerman, 1982) and executive remuneration has been called excessive (Volkskrant, 2004).

Donors tend to see the size of a charity as a measure of quality, which causes preference of donating to larger charities. Also, donors prefer to donate to charitable organizations with the lowest overhead costs. Due to the discussions on overhead costs in charitable organizations, it is important for charities to keep the overhead costs low.

According to the literature, charities should become more efficient when it increases in size. This implies that when a charity increases, the overhead and remuneration costs will increase relatively less.

The analysis, based on 76 Dutch charitable organizations, revealed that relatively bigger firms, face relatively lower overhead costs. This is also the answer on the research question on how the size of a charitable organization, based on total assets, influences the money spend on overhead. The mathematical answer will be that a 1% change in assets suggests that the overhead costs change with 0.875%.

This relation is based on total overhead costs, which can be divided into

administration and fundraising costs. This change is mainly because of the administration costs. Administration costs increase by 0.748% when assets increase by 1 percent. This is due to economies of scale of larger charities (Wise, 1997; Mourão, & Enes, 2017; Kahler, & Sargeant, 2002; Hyndman, & McKillop, 1999). A percentage change in the fundraising costs, on the contrary, tend to cause a change in size with the same percentage.

The results of the regressions on executive remuneration in charitable organizations shows that the executive remuneration will increase, when the size of a charity increases,. The results suggest that a size change of 1%, changes the executive remuneration by 0.191%. Literature on executive remuneration for for-profit firms also shows that bigger firms give more remuneration to their executive. It also shows that the executive remuneration in charities is at the lower end of findings on executive remuneration in for-profit firms. Whether the remuneration costs of charities are excessive, cannot be concluded from this research. Therefore, a more careful investigation would be necessary.

To conclude, based on the outcomes of the regressions, we can accept the two hypotheses made on the effect of size on overhead costs and executive remuneration. This implies that relatively bigger firms, face relatively lower overhead costs. Also, an increase in size of a charitable organization, significantly increases the remuneration for the executive.

19

References

Aggarwal, Evans, & Nanda. (2012). Nonprofit boards: Size, performance and managerial incentives. Journal of Accounting and Economics, 53(1-2), 466-487.

Barnett, V., & Lewis, T. (1974). Outliers in statistical data. Wiley.

Bekkers, R., Schuyt, T., & Gouwenberg. B. (2017). Geven in Nederland 2017. Amsterdam, Nederland: Lenthe.

Callen, J., Klein, A., & Tinkelman, D. (2003). Board Composition, Committees, and Organizational Efficiency: The Case of Nonprofits. Nonprofit and Voluntary Sector Quarterly, 32(4), 493-520.

Caviola, L., Faulmüller, N., Everett, J., Savulescu, J., & Kahane, G. (2014). The evaluability bias in charitable giving: Saving administration costs or saving lives? Judgment and Decision Making, 9(4), 303-316.

CBS. (2015). Nederland steeds ruimhartiger voor goede doel. Retrieved from

https://www.cbs.nl/nl-nl/nieuws/2015/51/nederland-steeds-ruimhartiger-voor-goede-doel.

CBF. (n.d.). Retrieved from https://www.cbf.nl/sectorbrede-informatie/verdeling-sector/. CBF. (n.d.). Retrieved from https://www.cbf.nl/sectorbrede-informatie/baten-en-lasten/. Chen, G. (2009). Does meeting standards affect charitable giving? An empirical study of

New York metropolitan area charities. Nonprofit Management and Leadership, 19(3), 349-365.

D. Larson, P., & Morris, M. (2014). Sex and salary. Supply Chain Management: An International Journal, 19(4), 385-394.

David, P., Kochhar, R., & Levitas, E. (1998). The effect of institutional investors on the level and mix of CEO compensation. Academy of Management Journal, 41(2), 200-208. Finkelstein, S., & Hambrick, D. (1989). Chief executive compensation: A study of the

intersection of markets and political processes. Strategic Management Journal, 10(2), 121-134.

Firth, Tam, & Tang. (1999). The determinants of top management pay. Omega, 27(6), 617-635.

Gneezy, Uri, Keenan, Elizabeth A, & Gneezy, Ayelet. (2014). Behavioral economics. Avoiding overhead aversion in charity. Science (New York, N.Y.), 346(6209), 632-5. Hansmann, H. (1980). The Role of Nonprofit Enterprise. The Yale Law Journal, 89(5),

835-901.

Heijden, H. (2013). Small is beautiful? : Financial efficiency of small fundraising charities. The British Accounting Review : The Journal of the British Accounting Association, 45(1), 50-57.

Herzlinger, Regina E. (1994). Effective oversight: A guide for nonprofit directors. Harvard Business Review, 72(4), 52.

Hyndman, N., & McKillop, D. (1999). >Conversion Ratios in Charities in England and Wales: An Investigation of Economies of Scale. Financial Accountability & Management, 15(2), 135-153.

Jovanovic, B. (1982). Selection and the evolution of industry. Econometrica : Journal of the Econometric Society, an Internat. Society for the Advancement of Economic Theory in Its Relation to Statistics and Mathematics, 50(3), 649-670.

Kahler, J., & Sargeant, A. (2002). The size effect in the administration costs of charities. European Accounting Review, 11(2), 215–243.

20

Mourão, P., & Enes, C. (2017). Costs and Economies of Scale at Not-for-Profit

Organizations: The Case of the Santa Casa da Misericórdia de Barcelos Between 2002 and 2013. Social Indicators Research, 132(2), 821-840.

Nourayi, M., & Mintz, S. (2008). Tenure, firm's performance, and CEO's compensation. Managerial Finance, 34(8), 524-536.

Olson, M. (1974). The logic of collective action : Public goods and the theory of groups (4th pr ed., Harvard economic studies). Cambridge, Mass: Harvard University Press. Oster, S. M. (1998), Executive Compensation in the Nonprofit Sector. Nonprofit

Management and Leadership, 8: 207-221.

Portillo, & Stinn. (2018). Overhead aversion: Do some types of overhead matter more than others? Journal of Behavioral and Experimental Economics, 72, 40-50.

Rose-Ackerman, S. (1982). Charitable giving and "excessive" fundraising. The Quarterly Journal of Economics, 97(2), 193-212.

Sargeant, A., & Kähler, J. (1999). Returns on Fundraising Expenditures in the Voluntary Sector. Nonprofit Management and Leadership, 10(1), 5-19.

Singh, Minu, S., & Yavuz, Cigdem, Y. (2015). Firm Performance and CEO Compensation : Determinants of CEO Compensation.

Tosi, H., Werner, S., Katz, J., & Gomez-Mejia, L. (2000). How Much Does Performance Matter? A Meta-Analysis of CEO Pay Studies. Journal of Management, 26(2), 301-339.

Volkskrant. (2004). Beloning top maakt goede doel nerveus. Retrieved from https://www.volkskrant.nl/economie/beloning-top-maakt-goede-doel-nerveus~a696675/.

Volkskrant. (2015). Goede doelen willen af van discussie over kosten. Retrieved from https://www.volkskrant.nl/economie/goede-doelen-willen-af-van-discussie-over-kosten~a4145293/.

Wise, D. (1997). Size and Administration Costs in the Voluntary Sector: A Note. Financial Accountability & Management, 13(1), 81-88.

Yi, D. (2010). Determinants of fundraising efficiency of nonprofit organizations: Evidence from US public charitable organizations. Managerial and Decision Economics, 31(7), 465-475.

21

Appendix:

Appendix 1.

The boxplot of the logarithm of overhead as independent variable. As can be seen, observation number 3 is an outlier. This observation is Gipsy Mission.

Appendix 2: Data of the charitable organizations

Charitable organization Total executive remuneration* Expenses to fundraising Administration and control

Assets Region Sector

Kuren met Reuma 0 4,849 7,235 62,783 Europe Health

ME/CVS-Stichting Nederland

73,868 24,983 32,627 96,155 Europe Health

Gipsy Mission 0 8 1,184 98,985 Europe Religion

Sumbing Bibir 0 296 4,650 100,220 Other Health

Apollos 0 4,630 13,485 122,885 Other Religion

SWODB 0 2,531 3,580 133,230 Europe Health

In de Rechte Straat 0 15,042 19,372 176,906 Europe Religion

Vlinderkind 0 2,696 8,529 178,241 Europe Health

OGO 0 13,947 10,040 193,547 Europe Religion

Nationaal Huidfonds 89,768 67,525 116,587 199,059 Europe Health

Beat Batten! 0 36,572 6,199 223,796 Other Health

JDRF Nederland 0 140,716 80,479 232,624 Europe Health

Bijbelvereniging 0 29,705 29,181 234,270 Europe Religion

Antwoord 40,539 90,805 24,444 245,944 Other Religion

22

MissieNederland 65,876 55,634 68,743 307,017 Europe Religion

ADF Stichting 97,647 24,334 7,830 320,336 Europe Health

Artsen voor kinderen 25,755 20,019 40,498 346,052 Europe Health

Bible League 37,920 186,612 50,013 355,582 Other Religion

Hartekind 0 26,590 86,541 388,069 Europe Health

Spaanse Evangelische Zending

0 30,058 34,989 473,600 Other Religion

Trans World Radio 67,932 283,180 144,703 559,745 Other Religion

Haarwensen 0 41,806 32,821 580,673 Europe Health

Agapè 12,386 173,256 239,950 599,380 Other Religion

De Ondergrondse Kerk 79,691 103,683 109,555 726,376 Other Religion

Stophersentumoren.nl 0 1,565 36,872 737,432 Europe Health

Operatie Mobilisatie 14,433 219,120 462,809 780,642 Other Religion

Vrienden Beatrix Ziekenhuis

0 22,596 16,260 922,601 Europe Health

3xM 104,689 107,290 94,304 1,039,506 Other Religion

Dokters van de Wereld 93,313 389,540 110,240 1,104,393 Other Health

Gave 19,116 70,600 121,306 1,106,226 Europe Religion

Metakids 98,795 142,380 32,717 1,229,199 Europe Health

ADRA-Nederland 75,140 54,203 66,464 1,248,159 Other Health

Hulp Oost-Europa 0 170,208 44,851 1,255,065 Europe Religion

Open Doors 100,358 1,187,856 532,510 1,310,324 Other Religion

Interserve Nederland 60,055 128,622 64,183 1,358,454 Other Religion

VNB 0 36,260 88,891 1,361,367 Europe Religion Evangelische Hogeschool 88,377 167,791 94,072 1,440,607 Europe Religion Nederlandse Cystic Fibrosis Stichting 126,781 224,941 114,296 1,497,108 Other Health HGJB 93,432 124,636 213,471 2,183,682 Europe Religion

Christenen voor Israël 98,703 495,975 404,829 2,207,891 Europe Religion

Wycliffe Bijbelvertalers 79,122 381,319 168,679 2,636,837 Other Religion

Light for the World 41,419 412,215 77,790 2,654,417 Other Health

IZB 92,908 202,037 176,706 2,835,014 Europe Religion

Trombosestichting Nederland

83,053 342,523 61,523 3,057,089 Europe Health

Jeugd met een Opdracht

0 11,269 137,929 3,441,987 Other Religion

Nationaal MS Fonds 91,301 339,822 99,453 3,626,030 Europe Health

Oogfonds 121,469 236,326 142,566 3,872,919 Europe Health

MISSIO 0 133,911 29,775 4,376,085 Other Religion

Health Works 142,242 214,127 634,489 4,883,272 Other Health

Compassion Nederland 79,955 975,144 621,834 5,312,235 Other Religion

Steun Emma

Kinderziekenhuis AMC

32,212 89,761 38,952 5,577,742 Europe Health

Spieren voor Spieren 95,505 1,341,438 155,997 5,837,466 Europe Health

Fonds Psychische Gezondheid

105,914 218,820 284,818 5,898,469 Europe Health

Vrienden van de Hoop 102,080 536,304 165,157 6,467,070 Europe Health

23

De Gereformeerde Zendingsbond

108,070 626,847 370,489 9,096,745 Other Religion

HandicapNL 127,951 901,992 495,284 9,457,104 Europe Health

ALS Nederland 112,012 765,862 108,097 9,605,068 Europe Health

Leprastichting 135,000 815,000 476,000 9,650,000 Other Health

Dorcas Aid International

96,022 1,918,744 1,064,675 9,978,282 Other Health

Hersenstichting 109,110 1,862,241 726,972 11,428,039 Europe Health

Maag Lever Darm Stichting

149,435 1,419,858 767,050 12,925,914 Europe Health

Alzheimer Nederland 150,161 3,190,189 945,172 13,435,732 Europe Health

Diabetes Fonds 138,818 1,010,989 618,831 13,634,916 Europe Health

Nederlandse

Brandwonden Stichting

124,446 882,351 452,035 14,781,539 Europe Health

Amref Flying Doctors 108,343 1,755,903 947,350 15,792,805 Other Health

MS Research 121,609 338,324 257,416 15,796,367 Europe Health

NBG (Nederlands Bijbelgenootschap)

136,536 1,279,000 837,000 17,763,000 Other Religion

Longfonds 167,773 2,832,369 707,378 23,536,038 Europe Health

Nierstichting Nederland

163,037 4,182,000 1,337,000 30,439,000 Europe Health

Reumafonds 231,579 2,715,000 299,000 32,836,000 Other Health

Aidsfonds 129,306 2,125,000 1,488,000 39,586,000 Europe Health

Stichting Kinderen Kankervrij (KIKA) 166,615 5,444,741 166,326 57,371,203 Europe Health KNCV Tuberculosefonds 154,996 730,938 1,161,406 95,591,726 Other Health

Hartstichting 179,000 7,570,496 1,707,433 108,797,497 Europe Health

KWF Kankerbestrijding 183,740 19,108,000 3,459,000 365,896,000 Europe Health

24

Appendix 3: Data statistics with outlier, Gipsy Mission, included.

Table 1. Data statistics with outlier

Variables Mean Median Standard

deviation

Minimum Maximum Observations

Log(overhead) 5.486027 5.578345 0.82972 3.076276 7.353474 77 Log(administration) 5.038064 5.058031 0.740707 3.073352 6.538951 77 Log(fundraising) 5.184286 5.305431 1.045669 0.90309 7.281215 77 Log(remuneration) 4.951919 4.998143 0.264128 4.092931 5.364699 56 Log(assets) 6.31021 6.175253 0.835025 4.092931 8.563358 77 Dummy Europe 0.623377 1 0.487717 0 1 77 Dummy Health 0.623377 1 0.487717 0 1 77

25

Appendix 4: Regressions with outlier, Gipsy mission, included.

Table 1. Regression with 1 independent variable, including the outlier Dependent variables

Log(overhead) Log(administration) Log(fundraising)

Independent variable Log(assets) 0.881 (0.053)** 0.734 (0.057)*** 1.032 (0.082)

*10% significance **5% significance ***1% significance

Table 2. Regression including control variables, including the outlier 𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃𝐃 𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐯𝐃𝐃𝐯𝐯

*10% significance **5% significance ***1% significance

Log(overhead) Log(administration) Log(fundraising)

Independent variables Log(assets) 0.901 (0.055)* 0.769 (0.059)*** 1.045 (0.086) Dummy Europe -0.131 (0.095) -0.126 (0.102) -0.153 (0.150) Dummy Health -0.105 (0.100) -0.188 (0.106)* -0.060 (0.157)