The Nonprofit Starvation Cycle: Empirical Evidence From a German Context

Abstract

The nonprofit starvation cycle describes a phenomenon in which nonprofit organizations continuously underinvest in their organizational infrastructure in response to external expectations for low overhead expenditure. In this study, we draw on nonprofit financial data from 2006 to 2015 to investigate whether the German nonprofit sector is affected by this phenomenon, specifically in the form of falling overhead ratios over time. We find reported overhead ratios to have significantly decreased among organizations without government funding and that the decrease originates from cuts in fundraising expenses—two results that are in contrast to previous findings from the U.S. nonprofit sector. With this study, we contribute to nonprofit literature by engaging in a discussion around the starvation cycle’s generalizability across contexts.

Keywords

nonprofit starvation cycle overhead financial efficiency ratio management generalization

Introduction

In recent years, scholars and practitioners have engaged in a debate about the nonprofit starvation cycle. The dynamics of the starvation cycle were first described by Wing and Hager (2004), who undertook the Nonprofit Overhead Cost Project in collaboration with their colleagues Pollak and Rooney. The term was later coined by Gregory and Howard (2009), who popularized the concept within the scientific community. The cycle suggests that many nonprofit organizations continuously curtail their spending on administration and fundraising, jointly called “overhead.” The authors’ argument goes as follows: Funders expect nonprofit organizations to spend an unrealistically low share of their income on overhead. As a result, organizations are pressured to invest too little in their organizational infrastructure, such as staff compensation and training, information technology, or fundraising activities. However, when organizations report lower overhead expenditure, funders expect them to spend even less on overhead in the future. According to Gregory and Howard (2009, p. 50), the overall outcome is a vicious cycle with a “chicken-and-egg-like quality” between funders’ expectations and organizational behavior, leading to falling overhead ratios over time and investment gaps that impair organizational effectiveness.

In the past years, several studies have focused on specific causal mechanisms underlying the starvation cycle (Garven, Hofmann, & McSwain, 2016). For example, some studies have analyzed the role of overhead ratios (i.e., overhead expenditure to total expenditure) in giving decisions. This has specifically been investigated in the contexts of individual giving (Gneezy, Keenan, & Gneezy, 2014; Meer, 2014) and governmental grant allocations (Ashley & Van Slyke, 2012). Other studies have used surveys to capture what precise overhead ratio thresholds are deemed legitimate among governmental agencies (Berlin, Masaoka, & Schumann, 2017), foundations (Eckhart-Queenan, Etzel, & Prasad, 2016; McCray, 2014), and donors (Grey Matter Research, 2018; YouGov, 2014). To assess the degree to which nonprofit organizations respond to external expectations, recent research has also focused on ratio management practices by nonprofit organizations (Kitching, Roberts, & Smith, 2012), including an analysis on factors driving such behavior (Berlin et al., 2017; Parsons, Pryor, & Roberts, 2017). “Ratio management” occurs when organizations actively try to adjust financial ratios, by either manipulating accounting information or by making resource reallocation decisions (Garven et al., 2016).

However, overall evidence of the existence of a starvation cycle that foremost includes decreased overhead spending over time is scarce. To our knowledge, the work of Lecy and Searing (2015) is the only attempt in existing nonprofit literature to verify the hypothesis of falling overhead ratios. Their longitudinal analysis of archival financial data indeed showed a 2.6 percentage point drop (from 20.9% to 18.3%) in the average reported overhead ratio among U.S. nonprofit organizations between 1985 and 2007.

Most strikingly with respect to the scarcity of empirical evidence is that no studies have tested the starvation cycle model in contexts outside the United States. This is surprising, as we presume that the perpetuating dynamics between expectations of low overhead spending and ratio management behavior exist in other countries as well. Empirical findings can only be considered generalizable and externally valid if they hold up when tested against other populations, settings, and variables (Campbell, 1957). Until now, however, nonprofit literature has remained silent on the starvation cycle’s external validity and scope conditions. Indeed, an overall lack of generalization studies across national boundaries has been articulated for nonprofit research in general (Helmig, Spraul, & Tremp, 2012) and for the discourse on nonprofit overhead in particular (Meer, 2017).

In light of this research gap, the goal of our study is to test for the existence of the starvation cycle phenomenon in Germany. We argue that Germany provides a particularly suitable context for two reasons. First, the fundamental mechanisms of the starvation cycle find circumstantial evidence, and second, the German nonprofit sector differs from the U.S. sector in several contextual conditions. With the shift to a new context, we aim to contribute to nonprofit literature by assessing the generalizability of the finding of falling overhead ratios over time that Lecy and Searing (2015) note for the United States.

We therefore conduct a study that, according to Tsang and Kwan (1999), can be categorized as an empirical generalization. In other words, we draw on a different population to assess nonprofit overhead spending over time (from 2006 to 2015 in Germany), while largely complying with the methodological approach of Lecy and Searing (2015). Although research has argued that under varying contextual conditions, generalization of empirical findings is difficult (Helmig et al., 2012), we follow the argument of Tsang and Kwan (1999) that “the more imprecise the replication, the greater the benefit to the external validity of the original finding, if its results support the finding” (p. 768).

A recent report by the Dorothy A. Johnson Center for Philanthropy identifies the discourse around capacity building and infrastructure investments as a key trend in philanthropy research and practice (Behrens et al., 2018). Our study therefore addresses a topic that is highly relevant for and sensitive to nonprofit financial management. Gaining a better understanding of expenditure patterns can help nonprofit leaders detect areas of underinvestment in their organizational infrastructure and better reflect on the consequences of financial reporting decisions. Our study can further incite donors and institutional grant makers to consider the role of organizational capacity in their expectations toward appropriate levels of overhead expenses (Lecy & Searing, 2015). Finally, our study is relevant for charity watchdogs, who should be aware that their overhead benchmarks and evaluation methodologies could equally affect nonprofits’ investment decisions.

The article first elaborates on the theoretical foundations of the starvation cycle phenomenon. We then discuss the different contextual environments between the German and U.S. nonprofit sectors and their potential effects on starvation cycle outcomes. Afterward, we conduct our empirical analysis of overhead spending over time, drawing on a unique sample of German nonprofit organizations. We conclude with a discussion of implications and an outlook for future research.

Theoretical Foundations of the Starvation Cycle Phenomenon

The starvation cycle describes a process consisting of three elements. First, funders hold expectations for overhead spending of nonprofit organizations, where expectations refer to the numerical threshold for an acceptable overhead ratio. These expectations are considered unrealistic when compared with overhead benchmarks in other industries. Second, nonprofit organizations perceive pressure to conform to these funder expectations, as they need to secure resources in increasingly competitive markets. Third, organizations react by forgoing infrastructure investments and by misrepresenting their expenditure data, two measures both aimed at reducing reported overhead ratios. As reporting lower overhead ratios influences expectations for an acceptable benchmark, these three elements are likely to create a vicious cycle with two outcomes: (a) continuously falling overhead ratios over time and (b) structural investment gaps that ultimately impair organizational effectiveness (Gregory & Howard, 2009; Hager, Pollak, Wing, & Rooney, 2004; Wing & Hager, 2004). Figure 1 depicts the starvation cycle model.

Figure 1.

The nonprofit starvation cycle.

Previous research has drawn on two theories to explain these dynamics. Wing and Hager (2004) originally conceptualized the starvation cycle through the lens of institutional theory. According to institutional theory, organizations seek compliance with expectations and social norms from their stakeholders (Meyer & Rowan, 1977). Organizations have an incentive to reduce overhead spending, because this can signal “symbolic performance” (Krishnan & Yetman, 2011, p. 1007), leading to more positive stakeholder evaluations. The reciprocal influence of stakeholder norms and organizational behavior can then explain why overhead spending declines over time.

Second, Parsons et al. (2017) explain organizations’ willingness to engage in ratio management through resource dependence theory (Pfeffer & Salancik, 1978), arguing that organizational survival critically depends on the ability to secure resources in increasingly competitive donor markets. Through the lens of resource dependency theory, reducing overhead spending is a strategy to gain a competitive advantage, if giving decisions are based on efficiency criteria.

Toward an Empirical Generalization Across Contexts

In recent years, nonprofit literature has called for more research aimed at generalizing empirical findings across different country contexts (Helmig et al., 2012; Meer, 2017). Empirical findings can be considered generalizable, if they hold up when tested against other populations or setting (Campbell, 1957).

So far, most research on nonprofit overhead has been conducted within a U.S. context, although controversies around the topic exist in other high-income countries as well. We argue that Germany provides a particularly suitable new context to assess the starvation cycle, as we observe several similarities with the U.S. nonprofit sector that contribute to the fundamental dynamics of the starvation cycle. For example, governments often shift grants from institutional funding to project funding (Priller et al., 2012), implying tighter restrictions on overhead expenditure as part of their grant making. Private donors also display an interest in funding programs directly and actively articulate a desire for low overhead ratios (YouGov, 2014). As a result, ratio management as an organizational response is regularly addressed as an issue in media reports. Last, the German nonprofit sector is also characterized by an increasingly high degree of competition (Zimmer, 2015), which is a key driver for ratio management practices in the United States (Lecy & Searing, 2015).

Against this backdrop, we base our study on the assumption that despite a lack of systemic empirical evidence, the German nonprofit sector is equally subject to the mechanisms of the starvation cycle. We follow the resource dependency perspective assuming that increasing levels of competition in the donation market may trigger a starvation cycle:

Hypothesis 1: Overhead ratios among German nonprofit organizations have declined over time.

A challenge for theory-testing generalization studies is that a specific set of contextual differences may affect the mechanisms underlying a theory (Tsang & Kwan, 1999). Following this argument, we first engage in a discussion around contextual conditions that differ between the German and the U.S. context (which is the basis of most of the research on overheads) and that might affect starvation cycle dynamics in Germany. For our analysis, we draw on research on comparative welfare regimes (Salamon & Anheier, 1998), comparative corporate governance (von Hippel, 2010), and recent surveys on the nonprofit sector (Priemer, Labigne, & Krimmer, 2015; Priller et al., 2012). As elaborated subsequently, this analysis yields four contextual differences as potentially relevant in the context of the starvation cycle: differences in nonprofit sector financing (Priemer et al., 2015; Salamon, Sokolowski, & List, 2003), differences in transparency regimes (Anheier, Hass, & Beller, 2013), differences in professionalism (Langer & Schröer, 2011), and differences in use of funds regulations (von Hippel, 2010; Zimmer, 2015).

Differences in Nonprofit Sector Financing

A first notable difference between the German and the U.S. nonprofit sectors is their financing structures. Most strikingly, nonprofit organizations in Germany rely more heavily on contributions from the government than organizations in the United States. Empirically, these differences in revenue composition have been captured by the Johns Hopkins Comparative Nonprofit Sector Project: In 2003, governments contributed 64% to the overall civil society organization revenue in Germany, compared with 31% in the United States (Salamon et al., 2003). Although the numbers have remained rather constant for the United States according to the National Center for Charitable Statistics (McKeever, 2015), more recent studies on the German nonprofit sector (Priemer et al., 2015; Priller et al., 2012), though subject to methodological differences, indicate a decrease in government contributions. Comparative nonprofit research has predominantly explained financing differences through social origins theory, suggesting that they are deeply rooted in welfare regime traditions (Salamon & Anheier, 1998).

We assume that financing differences will likely affect perceived ratio management pressure in German nonprofit organizations. We follow the argument of the starvation cycle that pressure arises from a desire to comply with funders’ expectations (Gregory & Howard, 2009). Empirical findings indicate that government dependency contributes more to perceived ratio management pressure than donor dependency (Berlin et al., 2017; Parsons et al., 2017). In light of budgetary constraints, governments often exert pressure by being particularly strict when it comes to overhead ratios granted as part of their grant making (Berlin et al., 2017). Such pressures further increase if nonprofit organizations bear costs of professionalization and increase their administrative capacity to meet administrative grant procedures and reporting requirements (Frumkin & Kim, 2002).

Differences in Transparency Regimes

A second relevant difference between the U.S. and the German nonprofit sectors is the prevailing transparency regimes. We view a transparency regime as the entire set of standards and practices of making information about the organization available to stakeholders (Rey-Garcia, Martin-Cavanna, & Alvarez-Gonzalez, 2012), which can be either formal (i.e., enforced legally through compulsory disclosure requirements) or informal (i.e., promoted voluntarily or through self-regulation initiatives) (Hale, 2013).

Although financial data of most U.S. nonprofit organizations is publicly disclosed in the form of tax filings to the Internal Revenue Service (IRS), financial information of German nonprofit organizations generally falls within the realm of tax secrecy and is therefore not disclosed by the tax authority (von Hippel, 2010). Although tax secrecy builds the core of the regulatory system, varying reporting requirements exist for specific fields of activity (e.g., educational institutions) or legal forms (e.g., nonprofit corporations with limited liability status), resulting in a highly fragmented system (Anheier et al., 2013).

Beyond statutory requirements, voluntary disclosure of annual reports has become an organizational practice for the segment of organizations that are active on the donation market, as indicated by the growing self-regulatory Initiative for a Transparent Civil Society. However, even this segment of the nonprofit sector still lacks a consolidated reporting standard with similar scope as the IRS form 990 in the United States that would allow comparability of financial data among organizations.

In the starvation cycle context, public availability and comparability of financial data can be key drivers for increased competition and ratio management pressure. Although the empirical literature on giving behavior is skeptical about whether many donors actively research financial information themselves, experiments indicate that when organizations can be easily compared, financial information becomes a contributing factor in donors’ decision making (Van der Heijden, 2013).

Differences in Professionalism

A third difference is the degree of professionalism of the two nonprofit sectors. We follow Suarez (2010, p. 311), who defines professionalization as “the increasing presence of specialized expertise in an organization and . . . shifts from volunteer labor to paid staff.” Empirically, the U.S. nonprofit sector arguably displays a higher degree of professionalism, as indicated by the greater prevalence of educational programs. In 2007, there were 284 nonprofit management programs at 179 educational institutions (Mirabella, 2007). In Germany, nonprofit management–specific programs are less common (Mirabella, Gemelli, Malcolm, & Berger, 2007), though the number of programs incorporating nonprofit-related components into their curricula has increased in the past years (Langer & Schröer, 2011). Research has attributed the rise of professionalism at an earlier stage in the U.S. context to the overall higher economic weight of the sector resulting from the liberal welfare regime logic (Langer & Schröer, 2011). This is reflected in the higher share of the nonprofit sectors’ contributions to their respective gross domestic product, which is estimated to be at 6.6% in the United States (Salamon, Sokolowski, Haddock, & Tice, 2013) versus 4.1% in Germany (Rosenski, 2012).

Generally, professionalism makes organizations more receptive to external expectations and pressures (Hwang & Powell, 2009). However, in the specific context of nonprofit finance, Parsons et al. (2017) found managerial sophistication (measured through organizational and personal characteristics) to reduce perceived external pressures and ratio management. The authors argue that sophisticated managers have the skills to provide alternative signals for organizational effectiveness (other than financial ratios).

Differences in Use of Funds Regulations

As the fourth noteworthy factor to consider, German and U.S. nonprofit organizations operate within different legal contexts with varying requirements for the use of funds. A key difference in this domain is the requirement of timely disbursement: German tax law ties the charitable status of nonprofit organizations to the principle of timely disbursement of funds, meaning that most of the annual income must be spent within the next two fiscal years. In the United States, such a requirement does not exist, granting organizations more flexibility in smoothing expenditure in response to changes in the environment, such as economic downturns (Calabrese, 2013; Grizzle, Sloan, & Kim, 2015).

The ability to build operating reserves is an important factor in the overhead context, as certain investments in organizational infrastructure likely depend on this ability. Empirically, Grizzle et al. (2015) noted a positive correlation between reserves and administrative expenditure. Mitchell (2017) argued that social norms toward lean reserves and overhead spending limit the ability of organizations to manage their finances more strategically. Timely disbursement arguably further aggravates such pressures, which is why it has been considered a major impediment toward more modern financial management by German nonprofit scholars (Zimmer, 2015).

In light of the four outlined contextual differences, we draw the preliminary conclusion that the starvation cycle phenomenon likely exists in a German context as well. Although not explicitly tested in the following empirical analysis, we identify higher dependency on government funding, a lower degree of professionalism, and highly restricted use of fund regulations as factors conducive to such dynamics. Table 1 provides a summary of the four identified sets of differences and potential effects on starvation cycle dynamics.

Table 1.

Overview of Contextual Differences.

Factor	Compared with the United States	Presumed effect on starvation cycle	Mechanism
Differences in financing	Higher government dependency	Positive	Government dependency increases ratio management pressures (Berlin, Masaoka, & Schumann, 2017; Parsons, Pryor, & Roberts, 2017).
Differences in transparency	Fragmented reporting system; lacking standards of financial disclosure	Negative	Financial transparency increases competition and ratio management pressure (Lecy & Searing, 2015).
Differences in professionalism	Lower degree of professionalism	Positive	Less managerial sophistication leads to higher likelihood for ratio management (Parsons et al., 2017).
Differences in use of fund regulations	Tighter restrictions on reserve building	Positive	Timely disbursement of funds impairs more professional financial management (Zimmer, 2015).

Method

Data and Sampling

We draw on data from annual reports of German nonprofit organizations over the 2006-2015 period. Table 2 shows the sample size for each year of our study period with 171 organizations in 2006 and 215 organizations in 2015. The data came from the German Central Institute for Social Issues (DZI), an agency that collects and evaluates annual reports as part of a seal accreditation process. DZI performs the role of what Bies (2010) termed “compliance self-regulation,” meaning that the agency serves to protect donor investments and assists donors in selecting which organizations to give to. Note that the number of organizations in our sample varies over time, because each year new organizations apply for certification and other organizations choose not to recertify. Despite this variation, 162 organizations held the seal of approval in each year over the entire period from 2006 to 2015. The year 2006 provides a suitable starting point for our analysis, because in that year DZI made changes to its cost reporting guidelines, limiting comparability with earlier years; 2015 is the most recent year for which data were fully available.

Table 2.

Overview of Study Sample.

	Year
	2006	2007	2008	2009	2010	2011	2012	2013	2014	2015
Total DZI sample	171	185	194	203	206	215	222	225	227	215
Expenditure (in €1,000)
Total	16,343	14,669	14,372	13,807	16,886	16,421	16,804	18,292	18,841	19,554
Fundraising	1,219	1,092	1,051	1,031	1,130	1,075	1,045	1,086	1,096	1,241
Admin	1,031	1,018	1,036	924	1,053	1,077	1,068	1,133	1,116	1,079
Total overhead	2,251	2,110	2,086	1,955	2,183	2,152	2,113	2,219	2,212	2,320
Revenue (in €1,000)
Total	16,271	14,818	15,071	14,660	18,896	16,985	17,012	19,258	19,456	20,293
Donation	8,317	7,603	7,517	7,178	9,019	7,991	7,093	7,815	7,575	8,264
Government	2,795	2,279	2,398	2,460	2,727	2,698	3,410	3,583	3,828	4,007

Note. All expenditure and revenue data are mean values. They were adjusted for inflation and are expressed in real values of 2015. DZI = German Central Institute for Social Issues.

By drawing on a sample of certified organizations only, we aim to enhance the quality of our data in two respects. First, DZI requires organizations to follow a standardized financial reporting scheme detailed in its cost reporting guidelines. We thus acknowledge recent criticism that very different computation schemes of financial ratios and expense allocation conventions find application in practice (Berlin et al., 2017; Eckhart-Queenan et al., 2016). Second, most of the data from annual reports in our sample have been subject to an external audit, which is an accreditation requirement for organizations with annual revenue of more than €500,000 in two consecutive fiscal years. In 2015, 75.8% of the organizations in our sample conducted an external audit. In the U.S. context, external audits significantly improve financial reporting accuracy (Keating, Parsons, & Roberts, 2008; Krishnan, Yetman, & Yetman, 2006).

Table 2 presents an overview of our sample characteristics for each year of our study period, and the appendix provides a more detailed account of the distributions of our key variables.

Representativeness of a sample is given if its characteristics accurately reflect the characteristics of a population. The latest study with representative data for the entire German nonprofit sector is the Civil Society in Numbers (ZiviZ) survey. For instance, ZiviZ data show that the nonprofit sector is predominantly composed of voluntary associations (95% of the approximately 630,000 organizations in total) and that on average organizations are small (66% with annual revenue of less than €20,000) (Priemer, Krimmer, & Labigne, 2017; Priemer et al., 2015). Our sample is based on a specific segment of the sector, namely, charitable organizations that actively engage in fundraising on the donation market. As no representative data exist for this specific segment of organizations and existing knowledge from various data sources is fragmented (Mews & Boenigk, 2015), all studies conducted in this field have to label their study as nonrepresentative. However, DZI data are commonly considered one of the most reliable data sources within the segment of charitable organizations (Boenigk et al., 2017).

Measurement

In line with prior research, we view overhead as containing both fundraising and administrative expenses (Bowman, 2006). Please note that we rely on reported financial data, which might not always entirely accurately reflect true spending due to accounting discretion. Our measurement of these expense categories follows the DZI cost reporting guidelines, which detail how various cost items are to be classified. Similar to IRS form 990 reporting, DZI guidelines demand functional reporting (i.e., a proportional allocation of expenses to certain expense categories). However, differences exist in the precise definitions of expense categories, as detailed next.

Fundraising and general public relations expenses

This category comprises all expenses associated with (a) raising contributions, including donations, and government funds and (b) general public relations activities (e.g., brand marketing) or project reporting. Exceptions are public relation activities that directly address the organization’s mission, such as advocacy campaigns, which instead are counted toward program expenditure. However, by incorporating nonmission-related public relations activities, this expense category is based on a broader definition than the “fundraising” expense category in IRS forms.

Administrative expenses

Conversely, administrative expenses follow a narrower definition than “Management and General Expenses” in IRS forms. Administrative expenses according to DZI guidelines refer to general operations of the organization, including expenses for basic infrastructure, accounting, board of directors, and personnel administration. Unlike the IRS calculation method, these expenses do not include any outreach activities (e.g., annual reports, websites).

Overhead expense ratio

Following previous research, we measure this ratio as the sum of administrative and fundraising expenses as a percentage of total expenditure (Bowman, 2006). Despite the outlined differences in the fundraising and administrative expense categories, the sum of both categories (i.e., total overhead) should again largely deliver comparable numbers.

Control variables

Previous research indicates that overhead ratios naturally vary depending on a number of organizational characteristics (see, for example, Ashley & Faulk, 2010; Hager, Pollak, & Rooney, 2001; Lecy & Searing, 2015). First, we account for organizational size, commonly measured by total revenue. Second, we consider funding differences by including donation dependency (the share of donations by total revenue) and government dependency (the share of government grants by total revenue), as these shares may affect perceived pressure and organizational needs for fundraising capacity. Third, we consider nonprofit subsectors. DZI uses a classification scheme of multiple tags (25 different fields of activity in total) for certified organizations. We recoded this classification scheme to comply with the National Taxonomy of Exempt Entities classification. However, due to the nature of our sample, this procedure delivered only three categories: human services, international, and a residual category (other). Despite this shortcoming, we still deem this an important control variable, as Lecy and Searing (2015) particularly found decreasing overhead ratios for the human service subsector.

Results

Findings on Overall Trend in Overhead Ratios

Figure 2 shows the distribution of all reported overhead ratios (across years). The distribution indicates bunching just below the 10% and 20% benchmarks. Note that DZI classifies certified organizations by their overhead ratio into one of three categories: overhead of less than 10% (considered “low”; n = 933), between 10% and 20% (considered “appropriate”; n = 753), and between 20% and 30% (considered “acceptable”; n = 376). Figure 2 therefore provides a first indication that certified organizations might be inclined to manage their overhead ratio around these thresholds.

Figure 2.

Distribution of reported overhead ratios.

An interesting question is whether bunching has increased over time, implying that a downward trend in reported overhead ratios might reflect increased attention to norms set by the accreditation agency. To control for increased bunching, we compared the numbers of organizations reporting overhead ratios within a two-percentage point range below and above the two main overhead thresholds (10% and 20%). We do not find that the share of organizations below the two thresholds has systematically increased over time.

Regarding the overall trend in the average reported overhead ratios, we find a slight decline from 13.03% to 12.07% over the course of our study period from 2006 to 2015 (see Figure 3). We will measure the statistical significance of this decline later on in our regression analysis. We repeated this analysis only considering the organizations that held certification in every year between 2006 and 2015 (n = 162), to rule out that year-over-year differences in our sample composition were driving our results. Although we find a slightly less pronounced decline with the reduced sample (12.96% in 2006 and 12.20% in 2015), the overall trend is similar.

Figure 3.

Trends in overhead expense ratios by expense categories.

In terms of our two main overhead-related expense categories administration and fundraising, we find that the decline in the reported overhead ratio can primarily be attributed to decreased average fundraising expenditure (see Figure 3). Specifically, our data show a decline in reported fundraising expenses (divided by total expenditure) from 6.29% in 2006 to 5.67% in 2015 and a decline from 6.74% to 6.39% in the reported administrative cost ratio.

As part of their financial reporting requirements, the organizations in our sample further distinguish all expenses incurred within a specific category between either personnel expenses or material expenses. Although total expenses have decreased (see Figure 3), we find an opposed trend within the fundraising category: personnel expenses for fundraising increased from 1.56% of total expenditure in 2006 to 1.98% in 2015, while material expenses for fundraising decreased sharply from 4.73% to 3.70% (see Figure 4). Administrative expenses for personnel and material do not show strong fluctuation over the study period. Taken together, total overhead-associated personnel expenditure was 4.73% of total expenditure in 2006 and 5.08% in 2015, whereas total overhead-associated material expenditure declined from 8.30% in 2006 to 6.99% in 2015.

Figure 4.

Trends in overhead expense ratios by expenditure types.

Findings on Overhead Trends by Revenue Structure

Previous research has noted a positive correlation between organizational size and overhead (Bowman, 2006; Hager et al., 2001; Lecy & Searing, 2015). The argument behind this relationship is that when organizations begin growing, they move from volunteer labor to a higher reliance on professional staff. This relationship was further identified to be inverse for very large organizations, which might benefit from economies of scale (Hager et al., 2001; Lecy & Searing, 2015). Our correlation matrix in Table 3 confirms a positive correlation between size and reported overhead ratios.

Table 3.

Pooled Correlation Matrix (N = 2,062).

	Variable	1	2	3	4	5	6
1	Overhead ratio	—
2	Admin expense ratio	.747**	—
3	Fundraising expense ratio	.800**	.198**	—
4	Total revenue (log)	.174**	.002	.255**	—
5	Donation dependency	−.013	−.146**	.112**	−.271**	—
6	Government dependency	−.104**	−.019	−.136**	.175**	−.654**	—

p(t) < .05. **p(t) < .01 (two-tailed).

Figure 5 shows the relationship between size and overhead and also provides some indication of the noted inverse relationship for the largest organizations in our sample.

Figure 5.

Organizational size and overhead ratios.

As Table 2 shows, average total revenue increased over the course of our study period. With the positive correlation between size and overhead, we would expect increasing revenue also to increase overhead ratios. As we find the inverse to be true, changes in organizational size do not sufficiently explain the downward overhead ratio trend. A concern might be that the downward trend in reported overhead ratios is driven by a relative increase in organizations within the largest revenue category, and hence by economies of scale instead of real organizational behavior. To control for this, we compared the relative share of organizations within the largest revenue category across years. We do not find this share to have significantly increased over time (9.94% in 2006 and 10.23% in 2015). We also investigated whether our result of a downward trend in reported overhead ratios (Figure 3) changes, if we exclude organizations within the largest revenue category all together. However, with the reduced sample we find a similar decline from 13.00% to 12.07%.

In a similar fashion to expense categories, we can distinguish revenue by donation dependency and government dependency. On a closer look at Table 2, critics could infer that increased government funding accounts for the observed downward trend in overhead ratios because additional government grants have benefited program spending (overhead spending in absolute terms has remained stable over the years). To better understand the dynamics between overhead and revenue sources, Figure 6 shows trends in reported overhead ratios over time segmented by total revenue and government funding. We split the sample using the median of total revenue and further distinguished between organizations with positive and zero government funding. Contrary to the first impression from Table 2, the graphs indicate that organizations without government funding are affected by the trend toward decreased overhead ratios. This tendency appears to be consistent across both large and small organizations. Taken together, the average reported overhead ratio among organizations without government funding decreased from 13.12% in 2006 to 10.14% in 2015, while it increased from 12.69% to 13.83% among organizations with government funding.

Figure 6.

Trends in overhead ratios by revenue.

Findings on Overhead Trends by Nonprofit Subsector

For our three different subsector categories, we find a strong decline in the average reported overhead ratio for human service organizations, which is in line with Lecy and Searing’s (2015) findings. Organizations classified as international oftentimes act as grant makers, while human service providers more often act operationally (i.e., run programs themselves). The latter group should thus have higher overhead requirements, which is only consistent, however, for the years 2006-2010 (see Figure 7).

Figure 7.

Trends in overhead ratios by nonprofit subsector.

Regression Results

As a final step, we tested whether the noted downward trend in overhead ratios is statistically significant. To test for changes over time, we introduced a count variable (“year”), which starts with the value 0 in 2006 and ends with 9 in 2015. We employed two different types of regression models. First, following Lecy and Searing (2015), we ran an ordinary least squares (OLS) regression without controlling for unobserved heterogeneity within individual organizations. Second, we ran fixed effect regression models, given that the observations across our panel data are likely not independent from one another. Fixed effects regression controls for organization-specific effects that are time-invariant. A significant Hausman test revealed fixed effects to be preferred over a random effects approach.

Models 1 and 2 show the results of both approaches for our full sample (see Table 4). For both models, we do not find a statistically significant negative effect over time. However, previous analysis revealed potential differences by funding structure. We therefore ran similar regression models based on two subsamples: only considering organizations with positive government funding (Models 3 and 4) and only without any government funding (Models 5 and 6). For both subsamples, we find statistically significant effects over time, which are positive for organizations with government funding and negative for organizations without.

Table 4.

Regression Results With the Overhead Ratio as Dependent Variable.

	Full sample		Only organizations with government funding		Only organizations without government funding
	Model 1 (pooled OLS)	Model 2 (fixed effects)	Model 3 (pooled OLS)	Model 4 (fixed effects)	Model 5 (pooled OLS)	Model 6 (fixed effects)
Year	−0.0010 (0.086)	0.0002 (0.553)	0.0018* (0.013)	0.0017*** (0.000)	−0.0031*** (0.000)	−0.0010* (0.027)
Total revenue (log)	0.0072*** (0.000)	−0.0312*** (0.000)	0.0003 (0.977)	−0.0428*** (0.000)	0.0095*** (0.000)	−0.0286*** (0.000)
Government dependency	−0.0584*** (0.000)	−0.0511*** (0.000)	−0.0689*** (0.000)	−0.0662*** (0.000)
Donation dependency	−0.0179* (0.020)	0.0009 (0.915)	0.0327** (0.002)	0.0115 (0.347)	−0.0226* (0.049)	0.0043 (0.761)
Subsector: Human service	0.0073 (0.090)		0.0247*** (0.000)		0.0136* (0.020)
Subsector: Other	0.0610*** (0.000)		0.0478*** (0.000)		0.0717*** (0.000)
Organization-specific fixed effects	No	Yes	No	Yes	No	Yes
Intercept	0.0349* (0.019)	0.5771*** (0.000)	0.1164*** (0.000)	0.7839*** (0.000)	0.0017 (0.940)	0.501*** (0.000)
N	2,062	2,062	1,061	1,061	996	996
R²/ $R_{adj}^{2}$	.123/.121	.819/.796	.179/.174	.812/.780	.166/.162	.844/.815
F	48.16/p(F) < .001	35.74/p(F) < .001	38.31/p(F) < .001	33.67/p(F) < .001	39.39/p(F) < .001	29.43/p(F) < .001

Note. Models 2, 4, and 6 account for organization-specific fixed effects. Subsectors are omitted in these models as they are time-invariant. OLS = ordinary least squares.

p < .05. **p < .01. ***p < .001.

Against this backdrop, the data only partially supports our hypothesis that overhead ratios have fallen among German nonprofit organizations over time. The decrease is only statistically significant for organizations that do not receive government funding.

Discussion and Conclusion

The goal of this study was to test for the existence of a nonprofit starvation cycle in Germany, ultimately contributing to nonprofit literature by assessing the generalizability of the phenomenon in light of contextual differences. We find a slight decline in the average reported overhead ratio from 13.03% to 12.07% between the years 2006 and 2015, though this change does not find statistical significance in regression analysis. However, we do find a statistically significant decline among organizations without government funding, where the average reported overhead ratio decreased by 2.98 percentage points.

Our results differ from findings in a U.S. context in several respects. First, Lecy and Searing (2015) found that declining overhead ratios were due to cuts in administrative expenses, while the fundraising expense ratio had in fact increased over the course of their study period. We instead find an opposing trend, with decreasing fundraising expenses accounting for falling overhead ratios. Second, in the United States, falling overhead ratios were associated with cuts in staff wages. In our case, staff expenses have increased, while material-related expenses, notably in the domain of fundraising, have shown a decrease. Last, Parsons et al. (2017) identified higher government dependency as a key driver of ratio management in the United States. According to our data, it appears that in Germany organizations without government funding face higher pressure than those in collaboration with government, although this will have to be further explored.

Based on our findings, we consider three arguments potentially relevant to explain the observed expenditure trends: We here call them starvation, steady state, and structure. The starvation argument, in line with the theoretical framing of our study, is that organizations indeed reduce overhead in response to external public pressure. The ongoing process of professionalization, as indicated by the increase in organizational size in our sample, could have made organizations more receptive to external expectations (Hwang & Powell, 2009). As a consequence, organizations might be inclined to skimp on their material-related expenses when staff-related expenses increase, as they face trade-offs between staff compensation and infrastructure investment.

The steady state argument is that organizations, instead of constantly reducing overhead, seek to manage ratios around certain acceptable thresholds in the sector (Kitching et al., 2012; Ling & Roberts, 2017). Our finding of bunching around the 10% and 20% thresholds is very much in line with this argument and the most direct indicator of ratio management in our study. However, we do not find bunching to have increased over time so that this argument does not sufficiently explain the overall decline in overhead ratios.

The structure argument holds that the observed expenditure changes reflect organizational change that is unrelated to ratio management. Increasing fundraising-associated personnel expenses might reflect that organizations give higher weight to fundraising as a core organizational task, indicating an increased relevance of the fundraiser’s profession. In line with this argument, lower material-related expenses imply that organizations to a lesser extent rely on external fundraising agencies, because costs of external services fall within the category of material expenses. In addition, decreasing material-related fundraising expenses could reflect that solicitations and campaigning have become more cost-effective through digitization and social media. However, while the structure argument can explain expense shifting between different subcategories, it also does not sufficiently explain a decrease in the overall overhead ratio.

Our study is subject to several limitations. First, we cannot claim representativeness of our sample for the German nonprofit sector or the population of charitable organizations. The lack of reporting standards and public disclosure makes it largely infeasible to conduct comparable research that can claim higher representativeness of the entire population of nonprofit organizations. Second, our analysis relies on reported financial data, which has been highlighted as a limitation before (Lecy & Searing, 2015). We do not know for certain whether expenditure changes occurred because of actual spending decisions, differences in functional reporting practices over time, or changing degrees in intentional misreporting. We tried to increase reporting sophistication by exclusively drawing on data from certified organizations. Third, we lack detailed data on changes in fundraising personnel of the organizations, making it difficult to draw more nuanced conclusions about professionalization effects. Specifically, we do not know whether increased personnel expenses are a result of higher wages or more staff.

Finally, our study provides avenues for future research. In line with our goal of generalization, future studies should investigate the existence of a starvation cycle in other countries, to gain a better understanding of the concept’s scope conditions and the relevance of contextual differences. For example, research could assess the effects of transparency differences on resulting ratio management pressure. The mechanisms of the cycle themselves also need further analysis. This could entail an exploration of expectations toward overhead expenses from various internal and external stakeholders to determine drivers of organizational pressure and ratio management. From an organizational perspective, additional research could shed light on whether organizations face pressure to continuously decrease overhead ratios or rather to manage ratios around specific acceptable thresholds. Management scholars should also engage in a discussion around reasonable investments into organizational infrastructure to better understand in what circumstances infrastructure can be considered adequate and when it is not. Last, a clear link between overhead expenses and organizational performance has yet to be established. Despite methodological challenges to assess this relationship, Marwell and Calabrese’s (2014) empirical study on the relationship may provide a valuable starting point for further research in this direction.

Footnotes

Appendix

Distribution of Variables in Study Sample for Year 2015.

Variables	2015 (n = 215)
Variables	Minimum	Median	M	Maximum	SD
Revenue (in €1,000)
Total	18	1,549	20,293	944,780	79,807
Donation	18	754	8,264	136,768	22,067
Government	0	5	4,007	156,514	18,960
Expenditure (in €1,000)
Total	17	1,372	19,554	882,611	75,735
Fundraising	0	55	1,240	35,640	4,205
Admin	0	80	1,079	83,110	5,998
Total overhead	0	155	2,320	115,666	9,278

Note. Across all years, the sample comprises 28 data points with zero overhead and 132 data points with zero fundraising (total n = 2,062). Previous research found that this can reflect poor reporting quality (Krishnan, Yetman, & Yetman, 2006). However, the trends we find in Figure 3 and Table 4 do not change, if these data points are excluded from the analysis. In addition, the German Central Institute for Social Issues evaluates the plausibility of reported fundraising expenses and activities, limiting the scope of misreporting.

Acknowledgements

We would like to thank the German Central Institute for Social Issues (DZI) for providing the data for this study and for their continued support throughout the project. We also thank Elizabeth Searing and Mark Hager for their valuable feedback at the 2017 ARNOVA conference and for comments received at the 2018 ISTR conference. Last, we are grateful for the constructive feedback from two anonymous reviewers.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Peter Schubert is a PhD student and research assistant at Universität Hamburg. His research interests include nonprofit management, accountability, and finance. His current research projects focus on overhead costs and the nonprofit starvation cycle.

Silke Boenigk is a professor for nonprofit management at Universität Hamburg. Her research interests include nonprofit strategy, fundraising, charitable giving, blood donor management, and refugee integration.

References

Anheier

H. K.

Hass

Beller

(2013). Accountability and transparency in the German nonprofit sector: A paradox? International Review of Public Administration, 18(3), 69-84.

Ashley

S. R.

Faulk

(2010). Nonprofit competition in the grants marketplace. Nonprofit Management and Leadership, 21, 43-57.

Ashley

S. R.

Van Slyke

D. M.

(2012). The influence of administrative cost ratios on state government grant allocations to nonprofits. Public Administration Review, 72, 47-56.

Behrens

Franklin

Caldwell

Moody

Downey

Olivarez

Peterson

(2018). 11 trends in philanthropy for 2018. Grand Rapids, MI: Dorothy A. Johnson Center for Philanthropy.

Berlin

Masaoka

Schumann

(2017). Two-Legged Stool: New Findings from California on Nonprofits and Overhead. Nonprofit Policy Forum, 8(2), 165-181.

Bies

A. L.

(2010). Evolution of nonprofit self-regulation in Europe. Nonprofit and Voluntary Sector Quarterly, 39, 1057-1086.

Boenigk

Hölz

Mildenberger

Schrötgens

Vahlpahl

Wilke

(2017). Research on giving in Germany. In Hoolwerf

Schuyt

(Eds.), Giving in Europe: The state of research on giving in 20 European countries (pp. 1-20). Amsterdam, The Netherlands: European Research Network on Philanthropy. Retrieved from http://ernop.eu/wp-content/uploads/2017/10/Giving-in-Europe-country-report-Germany.pdf

Bowman

(2006). Should donors care about overhead costs? Do they care? Nonprofit and Voluntary Sector Quarterly, 35, 288-310.

Calabrese

(2013). Running on empty: The operating reserves of U.S. nonprofit organizations. Nonprofit Management & Leadership, 23, 281-302.

10.

Campbell

D. T.

(1957). Factors relevant to the validity of experiments in social settings. Psychological Bulletin, 54, 297-312.

11.

Eckhart-Queenan

Etzel

Prasad

(2016). Pay-what-it-takes philanthropy. Stanford Social Innovation Review, 14(3), 37-41.

12.

Frumkin

Kim

(2002). The effect of government funding on nonprofit administrative efficiency. Cambridge, MA: John F. Kennedy School of Government.

13.

Garven

S. A.

Hofmann

M. A.

McSwain

D. N.

(2016). Playing the numbers game. Nonprofit Management & Leadership, 26, 401-416.

14.

Gneezy

Keenan

E. A.

Gneezy

(2014). Avoiding overhead aversion in charity. Science, 346, 632-635.

15.

Gregory

A. G.

Howard

(2009). The nonprofit starvation cycle. Stanford Social Innovation Review, 7(4), 49-53.

16.

Grey Matter Research. (2018). The Donor Mindset Study VI. Phoenix, AZ: Author. Retrieved from https://www.greymatterresearch.com/index_files/Generosity.htm

17.

Grizzle

Sloan

M. F.

Kim

(2015). Financial factors that influence the size of nonprofit operating reserves. Journal of Public Budgeting, Accounting & Financial Management, 27, 67-97.

18.

Hager

M. A.

Pollak

Rooney

(2001). Variations in overhead and fundraising efficiency measures: The influence of size, age, and subsector. Washington, DC: Urban Institute.

19.

Hager

M. A.

Pollak

Wing

Rooney

P. M.

(2004). Getting what we pay for: Low overhead limits nonprofit effectiveness (Nonprofit Overhead Cost Project, Brief No. 3). Washington, DC: Urban Institute.

20.

Hale

(2013). Understanding nonprofit transparency: The limits of formal regulation in the American nonprofit sector. International Review of Public Administration, 18(3), 31-49.

21.

Helmig

Spraul

Tremp

(2012). Replication studies in nonprofit research: A generalization and extension of findings regarding the media publicity of nonprofit organizations. Nonprofit and Voluntary Sector Quarterly, 41, 360-385.

22.

Hwang

Powell

W. W.

(2009). The rationalization of charity: The influences of professionalism in the nonprofit sector. Administrative Science Quarterly, 54, 268-298.

23.

Keating

E. K.

Parsons

L. M.

Roberts

A. A.

(2008). Misreporting fundraising: How do nonprofit organizations account for telemarketing campaigns? The Accounting Review, 83, 417-446.

24.

Kitching

K. A.

Roberts

A. A.

Smith

P. C.

(2012). Nonprofit resource allocation decisions: A study of marginal versus average spending. Journal of Governmental & Nonprofit Accounting, 1(1), A1-A19.

25.

Krishnan

Yetman

M. H.

(2011). Institutional drivers of reporting decisions in nonprofit hospitals. Journal of Accounting Research, 49, 1001-1039.

26.

Krishnan

Yetman

M. H.

Yetman

R. J.

(2006). Expense misreporting in nonprofit organizations. The Accounting Review, 81, 399-420.

27.

Langer

Schröer

(2011). Professionalisierung im Nonprofit Management [Professionalization in nonprofit management]. Wiesbaden, Germany: Verlag für Sozialwissenschaften.

28.

Lecy

J. D.

Searing

E. A.

(2015). Anatomy of the nonprofit starvation cycle: An analysis of falling overhead ratios in the nonprofit sector. Nonprofit and Voluntary Sector Quarterly, 44, 539-563.

29.

Ling

Roberts

A. A.

(2017). Identical program ratios: A red flag of ratio management. Retrieved from https://ssrn.com/abstract=2929649

30.

Marwell

N. P.

Calabrese

(2014). A deficit model of collaborative governance: Government–nonprofit fiscal relations in the provision of child welfare services. Journal of Public Administration Research and Theory, 25, 1031-1058.

31.

McCray

(2014). Is grantmaking getting smarter? A national study of philanthropic practice. Washington, DC: Grantmakers for Effective Organizations. Retrieved from https://ncg.org/sites/default/files/resources/GEO_IsGrantmakingGettingSmarter_2014_field_study.pdf

32.

McKeever

B. S.

(2015). The nonprofit sector in brief 2015: Public charities, giving and volunteering. Washington, DC: Urban Institute.

33.

Meer

(2014). Effects of the price of charitable giving: Evidence from an online crowdfunding platform. Journal of Economic Behavior & Organization, 103, 113-124.

34.

Meer

(2017). Are overhead costs a good guide for charitable giving? IZA World of Labor, 2017, 1-10. Retrieved from https://wol.iza.org/uploads/articles/329/pdfs/are-overhead-costs-good-guide-for-charitable-giving.pdf?v=1

35.

Mews

Boenigk

(2015). Giving in Germany: Toward systematic information on a fragmented nonprofit sector. In Wiebking

Handy

(Eds.), The Palgrave handbook of global philanthropy (pp. 170-189). London, England: Palgrave Macmillan.

36.

Meyer

J. W.

Rowan

(1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83, 340-363.

37.

Mirabella

R. M.

(2007). University-based educational programs in nonprofit management and philanthropic studies: A 10-year review and projections of future trends. Nonprofit and Voluntary Sector Quarterly, 36(4), 11-27.

38.

Mirabella

R. M.

Gemelli

Malcolm

M. J.

Berger

(2007). Nonprofit and philanthropic studies: International overview of the field in Africa, Canada, Latin America, Asia, the Pacific, and Europe. Nonprofit and Voluntary Sector Quarterly, 36, 110-135.

39.

Mitchell

G. E.

(2017). Fiscal leanness and fiscal responsiveness: Exploring the normative limits of strategic nonprofit financial management. Administration & Society, 49, 1272-1296.

40.

Parsons

L. M.

Pryor

Roberts

A. A.

(2017). Pressure to manage ratios and willingness to do so: Evidence from nonprofit managers. Nonprofit and Voluntary Sector Quarterly, 46, 705-724.

41.

Pfeffer

Salancik

G. R.

(1978). The external control of organizations: A resource dependence perspective. New York: Harper & Row.

42.

Priemer

Krimmer

Labigne

(2017). ZIVIZ-Survey 2017. Essen, Germany: Stifterverband für die Deutsche Wissenschaft.

43.

Priemer

Labigne

Krimmer

(2015). ZIVIZ-Finanzierungsstudie 2015 [ZIVIZ-financing study 2015]. Essen, Germany: Stifterverband für die Deutsche Wissenschaft.

44.

Priller

Alscher

Droß

P. J.

Paul

Poldrack

C. J.

Schmeißer

Waitkus

(2012). Dritte-Sektor-Organisationen heute: Eigene Ansprüche und ökonomische Herausforderungen; Ergebnisse einer Organisationsbefragung [Third sector organizations today: Own aspirations and economic challenges. Results of an organization survey] (WZB discussion paper). Berlin: WZB. Retrieved from https://www.wzb.eu/system/files/docs/ende/zeng/dso_gesamt_finale_23-05-2013_online.pdf

45.

Rey-Garcia

Martin-Cavanna

Alvarez-Gonzalez

L. I.

(2012). Assessing and advancing foundation transparency: Corporate foundations as a case study. The Foundation Review, 4(3), 77-89.

46.

Rosenski

(2012). Die wirtschaftliche Bedeutung des Dritten Sektors [The economic weight of the third sector]. In Bundesamt

Statistisches

(Ed.), Wiesbaden, Germany: Wirtschaft und Statistik (pp. 209-217).

47.

Salamon

L. M.

Anheier

H. K.

(1998). Social origins of civil society: Explaining the nonprofit sector cross-nationally. VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations, 9, 213-248.

48.

Salamon

L. M.

Sokolowski

S. W.

Haddock

M. A.

Tice

H. S.

(2013, March). The state of global civil society and volunteering: Latest findings from the implementation of the UN nonprofit handbook (Center for Civil Society Studies Working Paper No. 49). Baltimore, MD: Johns Hopkins Center for Civil Society Studies.

49.

Salamon

L. M.

Sokolowski

S. W.

List

(2003). Global civil society. Baltimore, MD: Johns Hopkins Center for Civil Society Studies.

50.

Suarez

D. F.

(2010). Collaboration and professionalization: The contours of public sector funding for nonprofit organizations. Journal of Public Administration Research and Theory, 21, 307-326.

51.

Tsang

E. W.

Kwan

K. M.

(1999). Replication and theory development in organizational science: A critical realist perspective. Academy of Management Review, 24, 759-780.

52.

Van der Heijden

. (2013). Charities in competition: Effects of accounting information on donating adjustments. Behavioral Research in Accounting, 25, 1-13.

53.

von Hippel

. (2010). Nonprofit organizations in Germany. In Hopt

K. J.

von Hippel

(Eds.), Comparative corporate governance of non-profit organizations (pp. 197-227). New York, NY: Cambridge University Press.

54.

Wing

Hager

M. A.

(2004, November). Who feels pressures to contain overhead costs? Paper presented at ARNOVA Annual Conference, Los Angeles, CA.

55.

YouGov. (2014). Die Spender der Zukunft [The donors of the future]. Retrieved from https://yougov.de/loesungen/reports/studien/spender-zukunft/

56.

Zimmer

(2015). Germany’s nonprofit organizations: Continuity and change. Sociologia e politiche sociali, 18(3), 9-26.