Introduction
The nonprofit sector plays a vital role in contemporary governance by addressing social needs, supporting vulnerable populations, and complementing government services (Enjolras et al., Reference Enjolras, Salamon, Sivesind and Zimmer2018; Li et al., Reference Li, Lu, Zhao, Wang and Chen2026; Osborne, Reference Osborne2010). To pursue these objectives, nonprofits seek financial support from government agencies, private foundations, corporations, and individual donors (Green et al., Reference Green, Ritchie, Bradley and Parry2021; Young, Reference Young2007).Footnote 1 Much of this support arrives with funder-imposed conditions that limit how resources may be used, restricting funds to specified purposes, activities, or time periods (Barman, Reference Barman2008; Heinzel et al., Reference Heinzel, Cormier and Reinsberg2023; Hung, Reference Hung2021; Yermack, Reference Yermack2017). For example, a foundation may restrict a grant to a youth mental health initiative, while a corporate donation may be earmarked for purchasing laptops for an after-school tutoring program. Restricted funding therefore matters because it shapes how flexibly nonprofits can allocate resources and respond to changing needs.
Existing research shows that restricted funding is not neutral in its organizational consequences. Prior studies link it to lower financial discretion, greater administrative burden, mission drift, and, in some settings, weaker operating performance (Clerkin & Quinn, Reference Clerkin and Quinn2019; Henderson & Lambert, Reference Henderson and Lambert2018; Hung, Reference Hung2021; Shon et al., Reference Shon, Hamidullah and McDougle2019; Wiepking & de Wit, Reference Wiepking and de Wit2020), although some work also points to benefits such as stronger discipline or improved cost efficiency (Li et al., Reference Li, McDowell and Hu2012; Mensah & Werner, Reference Mensah and Werner2003; Yermack, Reference Yermack2017). If restricted funding affects how nonprofits budget, adapt, and deliver services, then understanding why some organizations face higher levels of restricted funding than others becomes an important question in its own right. However, most existing studies focus on outcomes rather than antecedents. This leaves limited understanding of the organizational conditions under which funders attach tighter or looser constraints to revenue. The question is also empirically important because variation is substantial. Among the UK’s top 100 charity fundraisers, for example, restricted income ranged from 0% to 94% of total income in one year (Xledger, 2019).
We focus on organizational antecedents because funders often make decisions under incomplete information and therefore rely on observable characteristics of recipient organizations. Drawing on agency and signaling perspectives, we view information asymmetry as the common problem linking the two theories (Connelly et al., Reference Connelly, Certo, Ireland and Reutzel2011; Eisenhardt, Reference Eisenhardt1989; Jensen & Meckling, Reference Jensen and Meckling1976; Spence, Reference Spence1973). Restrictions can serve as a governance response to uncertainty about how resources will be used, while organizational features such as financial position, staffing, and program scope provide signals about risk, capacity, and stewardship. These recipient-side characteristics are therefore theoretically important because they may shape how funders assess organizations and structure funding conditions. Examining them allows us to ask whether nonprofits systematically experience different levels of restriction based on observable organizational attributes. This organization-side explanation complements, rather than substitutes for, funder preferences and relational dynamics as other important sources of variation in restricted funding.
The UK charity sector offers a useful setting for this question. It is a large and diverse field in which charities operate across health, education, social care, and environmental protection, and reporting under the Statement of Recommended Practice provides detailed information on restricted and unrestricted funding (Tabassum & Fern, Reference Tabassum and Fern2023). At the same time, UK funding practices have long emphasized accountability and demonstrable results, while recent debates have also highlighted the value of flexibility (Mills, Reference Mills2023). This combination makes the UK context well suited for examining how funders structure support and why restriction levels vary across organizations.
This study addresses that gap by examining organizational antecedents of restricted funding in charities in England and Wales from 2007 to 2023. We ask: What organizational factors shape the level of restricted funding? Our analysis shows that charities with higher financial reserves face lower levels of restricted funding, whereas greater service-area breadth and wider beneficiary breadth are associated with higher levels of restriction. Administrative cost levels are not a significant predictor. These findings contribute to the literature in three ways. First, they shift attention from the consequences of restricted funding to the recipient-side conditions associated with nonprofits’ exposure to funding restrictions. Second, they connect agency and signaling perspectives through information asymmetry by identifying observable organizational characteristics associated with tighter or looser funding conditions. Third, they extend a literature dominated by U.S.-based evidence by providing findings from the UK context, although these findings should be interpreted in light of the institutional and reporting features of that setting.
The remainder of the paper proceeds as follows. Section “Literature review” reviews research on restricted funding. Section “Theoretical framework” develops hypotheses grounded in agency and signaling perspectives. Section “Research context and methodology” describes the UK context, data, and research design. Section “Results” reports the results. Section “Discussion and conclusion” discusses theoretical and practical implications and concludes.
Literature review
Restricted funding refers to financial contributions provided to nonprofits with specific conditions on usage, as stipulated by the donor or funding entity (Hung, Reference Hung2021; Wiepking & de Wit, Reference Wiepking and de Wit2024; Yermack, Reference Yermack2017). Unlike unrestricted funds, which nonprofits can allocate freely across their operations, restricted funds are earmarked for particular projects, activities, or purposes aligned with the funder’s priorities (Prentice & Clerkin, Reference Prentice and Clerkin2023; Thornton, Reference Thornton2010; Wiepking & de Wit, Reference Wiepking and de Wit2020). This funding model is widespread in the nonprofit sector, where donations and grants from government agencies, foundations, and individual donors frequently come with conditions intended to ensure accountability, align with funder objectives, maintain oversight, and assess impact (Barman, Reference Barman2008; Elbers & Arts, Reference Elbers and Arts2011; Heinzel et al., Reference Heinzel, Cormier and Reinsberg2023). As a result, restricted funding has become a notable aspect of nonprofit financial structures, reflecting the broader emphasis on transparency and accountability within the nonprofit sector.
Restricted funding can offer benefits by channeling resources toward specific activities that align with both the nonprofit’s mission and the funder’s goals. This targeted approach may enhance program efficiency, enforce operational discipline, and promote transparency, ultimately building donor confidence (Li et al., Reference Li, McDowell and Hu2012; Mensah & Werner, Reference Mensah and Werner2003). These benefits are more likely when designated funds support programs already central to the mission, which can reinforce priorities and facilitate clearer planning and reporting. However, restrictions can also create challenges when donor-designated priorities do not fully match broader organizational needs. Prior studies show that restricted funding can limit coverage of essential administrative costs, contributing to a “nonprofit starvation cycle” (Gregory & Howard, Reference Gregory and Howard2009). Managing multiple restricted grants can also add administrative burden and divert attention from mission-focused work (Henderson & Lambert, Reference Henderson and Lambert2018). Over time, strict adherence to donor priorities may reduce adaptability and innovation, as organizations feel pressure to prioritize external goals over local or emerging needs (Clerkin & Quinn, Reference Clerkin and Quinn2019; Elbers & Arts, Reference Elbers and Arts2011).
Over the past two decades, a small but growing body of research has examined the organizational outcomes of restricted funding. Yermack (Reference Yermack2017) found that while restricted donations offer financial stability for U.S. art museums, they also constrain autonomy and are associated with lower profit margins. Hung (Reference Hung2021) similarly observed that restricted funding in U.S. arts and cultural nonprofits limits financial discretion and negatively affects service delivery. In the U.K., Henderson and Lambert (Reference Henderson and Lambert2018) noted that while grant reliance provides financial stability, restrictions can introduce vulnerabilities, challenge autonomy, and contribute to mission drift. Elbers and Arts (Reference Elbers and Arts2011) reported that restrictive funding often pushes international development nonprofits in India and Ghana to prioritize donor objectives over local needs, leading to inefficiencies in program delivery. Conversely, Mensah and Werner (Reference Mensah and Werner2003) found that financial restrictions may improve cost efficiency in U.S. nonprofit higher education institutions.
These findings suggest that restricted funding can meaningfully shape nonprofit behavior, discretion, and performance. If restriction levels affect how flexibly nonprofits allocate resources, manage trade-offs, and respond to community needs, then understanding why some organizations experience more restricted funding than others is theoretically and practically important. Studying antecedents extends the literature from documenting the consequences of restricted funding to explaining the organizational conditions associated with tighter or looser funding constraints. More specifically, organizational antecedents capture recipient-side characteristics available to funders when assessing prospective or existing grantees. Examining these characteristics can reveal whether exposure to funding restrictions varies systematically with observable differences among nonprofits, while complementing explanations centered on funder characteristics and relational dynamics.
Comparatively, few studies examine the organizational antecedents of restricted funding, and those that do are largely situated in the U.S. context. Barman (Reference Barman2008) found that nonprofits often allow donors to impose restrictions on their contributions in response to donors’ demand for control. Prentice and Clerkin (Reference Prentice and Clerkin2023) compared financing and operational factors associated with restricted versus unrestricted donations in higher education nonprofits and identified distinct determinants of each type. Building on this work, we focus specifically on the recipient side of the funding relationship and ask whether observable organizational characteristics systematically correspond to the degree of restriction organizations experience. This perspective highlights an organization-level layer of explanation that complements funder-side and relational accounts. Drawing on agency and signaling theories, we examine how organizational characteristics may provide information about risk, capacity, and stewardship and, in turn, be associated with funders’ use of restrictions. We extend prior research by testing these relationships in the UK charity context.
Theoretical framework
Agency theory and signaling theory are linked by a common problem: information asymmetry. Funders provide resources to nonprofits but cannot fully observe how those resources will be managed once granted, whether spending will remain aligned with intended purposes, or whether the organization has the capacity to deliver promised outputs. Under these conditions, restrictions can function as an ex ante governance response to uncertainty. Agency theory highlights why funders may impose such controls when they perceive greater risks of misallocation, goal drift, or weak alignment between their objectives and nonprofit behavior (Eisenhardt, Reference Eisenhardt1989; Jensen & Meckling, Reference Jensen and Meckling1976; Steinberg, Reference Steinberg, Hopt and Von Hippel2010).
Signaling theory complements this perspective by explaining how nonprofits’ observable characteristics can either reduce or intensify that information problem. Because funders cannot directly observe all dimensions of organizational quality, they rely on visible cues to infer capacity, stewardship, and reliability (Connelly et al., Reference Connelly, Certo, Ireland and Reutzel2011; Spence, Reference Spence1973). Under this combined framework, restrictions are more likely when observable features suggest higher risk or weaker organizational capacity, and less likely when those features signal stability, competence, and alignment. Our study therefore examines how financial and organizational characteristics shape funders’ use of restrictions by influencing the degree of uncertainty they face.
This unified perspective guides our hypotheses on financial reserves, administrative costs, service-area breadth, and beneficiary breadth. In each case, the core question is whether the characteristic reduces or intensifies information asymmetry and, in turn, affects funders’ demand for tighter conditions.
Financial reserves
Financial reserves represent surplus resources that nonprofits hold to address unforeseen challenges or sustain operations during revenue shortfalls (Bourgeois, Reference Bourgeois1981; Irvin & Furneaux, Reference Irvin and Furneaux2022). Prior studies underscore their stabilizing role across varied conditions. For example, Calabrese (Reference Calabrese2018) found that operating reserves reduce expense volatility during downturns, while Kim and Mason (Reference Kim and Mason2020) observed that nonprofits with larger reserves were less likely to face program delivery or financing difficulties during the early stages of the COVID-19 pandemic. Similarly, Lee and Woronkowicz (Reference Lee and Woronkowicz2019) noted that nonprofits with greater cash holdings allocate more toward program spending, enhancing their ability to expand services and meet community needs.
In the presence of information asymmetry, reserve levels can shape how funders assess recipient risk. Lower reserves increase the likelihood that organizations will face cash shortfalls and resource pressures, which may raise funders’ concerns about misallocation or disrupted delivery. At the same time, reserve levels serve as observable cues of stability, stewardship, and planning capacity. Strong reserves suggest an ability to absorb volatility and sustain operations, whereas weak reserves may signal greater vulnerability and execution risk (Irvin & Furneaux, Reference Irvin and Furneaux2022; Lee & Woronkowicz, Reference Lee and Woronkowicz2019). Taken together, these mechanisms imply that lower reserves intensify uncertainty and make restrictions more likely. Thus, we posit:
H1: Nonprofits with higher financial reserves are less likely to experience higher levels of restricted funding.
Administrative costs
Administrative costs are expenses for managing and supporting operations, including salaries, rent, utilities, and accounting. These costs are often viewed as diverting resources from direct outputs that donors and funders value most (Hung et al., Reference Hung, Lu and Hung2025; Weisbrod & Dominguez, Reference Weisbrod and Dominguez1986). The broader literature on overhead aversion and the nonprofit starvation cycle shows that funders are often reluctant to support administrative spending and prefer program-specific funding that yields visible outputs (Gregory & Howard, Reference Gregory and Howard2009).
Our interest is not in whether disclosure of administrative costs affects restriction, but in whether the level of administrative costs shapes funders’ uncertainty about resource use. Even when administrative cost information is disclosed, funders may still be unsure whether such spending is necessary, proportionate, or effectively tied to program delivery. Higher administrative cost ratios may therefore raise concern about resource allocation and prompt closer control. They may also function as a visible cue of stewardship and spending priorities, even if they are not a direct measure of efficiency (Coupet & Berrett, Reference Coupet and Berrett2019; Lu & Zhao, Reference Lu and Zhao2019). Prior studies show that donors often dislike administrative costs and prefer program spending (Gneezy et al., Reference Gneezy, Keenan and Gneezy2014), and that higher overhead can depress support (Tinkelman & Mankaney, Reference Tinkelman and Mankaney2007; Wong & Ortmann, Reference Wong and Ortmann2016). Taken together, these considerations suggest that higher administrative costs may intensify uncertainty and increase the likelihood of tighter earmarks. Thus, we posit:
H2: Nonprofits with higher levels of administrative costs are more likely to have higher levels of restricted funding.
Service-area breadth
Many nonprofits operate as multiservice organizations, delivering programs across several domains rather than specializing in one area (Oster, Reference Oster, Seaman and Young2010). We refer to the number of domains an organization serves as service-area breadth. Broader portfolios can widen reach, attract diverse funders, and support survival through cross-subsidization across programs (James, Reference James1983; Kistruck et al., Reference Kistruck, Qureshi and Beamish2013; Mendoza-Abarca & Gras, Reference Mendoza-Abarca and Gras2019; Park et al., Reference Park, Lu and Shon2022). However, greater breadth may also raise the likelihood of restricted funding because funders often target support to activities that match their priorities.
Under information asymmetry, broader service portfolios can make it harder for funders to assess how resources will be allocated across programs and whether support will remain focused on intended activities. Managing multiple service areas increases coordination and monitoring complexity, while also making it less clear what the organization’s core expertise is. Evidence on earmarked funding highlights its role as a control device in more complex settings (Heinzel et al., Reference Heinzel, Cormier and Reinsberg2023), and nonprofit research notes the added burden of managing diverse program and revenue portfolios (Gronbjerg, Reference Gronbjerg1993). At the same time, specialization often provides a clearer cue of competence, whereas broader scope may create category-spanning ambiguity and weaken audience evaluations (Hsu, Reference Hsu2006; Zuckerman, Reference Zuckerman1999). Taken together, greater service-area breadth can intensify uncertainty and make restrictions more attractive to funders. Thus, we posit:
H3: Nonprofits with greater service-area breadth are more likely to have higher levels of restricted funding.
Beneficiary breadth
Nonprofits also vary in the breadth of populations they serve, from narrowly defined groups to broad, heterogeneous communities (Anheier, Reference Anheier2014). A focused beneficiary scope can clarify needs, outcomes, and alignment with funder priorities, which facilitates assessment and reporting (Benjamin, Reference Benjamin2013; Ebrahim, Reference Ebrahim2005). By contrast, serving many beneficiary groups can expand reach but also complicate program design, resource allocation, and performance assessment. These trade-offs make beneficiary breadth a meaningful organizational feature for understanding how restrictions arise.
Under information asymmetry, greater beneficiary breadth can make it more difficult for funders to assess who will benefit from their support and how clearly outcomes can be attributed to particular groups. Broader beneficiary scope increases coordination and monitoring complexity and may raise concern that resources will be spread too thinly or redirected away from intended populations. At the same time, a clearly defined target population provides a more focused signal of expertise, mission clarity, and measurable impact, whereas broader beneficiary sets can blur those cues. Prior work on targeted giving similarly suggests that donors often attach conditions to channel resources toward preferred groups and purposes (Barman, Reference Barman2008; Wilsker & Young, Reference Wilsker and Young2010). Taken together, these mechanisms imply that wider beneficiary breadth can increase funders’ demand for restrictions. Thus, we posit:
H4: Nonprofits with greater beneficiary breadth are more likely to have higher levels of restricted funding.
Research context and methodology
We situate the study in the UK charity sector, a large and diverse field that delivers services across health, education, social care, and environmental protection, contributing to the country’s social and economic well-being (Alcock, Reference Alcock2010; Calò et al., Reference Calò, Montgomery and Baglioni2022). The sector has grown substantially in recent decades as charities partner with public bodies and private funders to address complex needs. In 2020/21, more than 160,000 voluntary organizations employed about 925,000 staff and contributed roughly £18 billion to the economy (Tabassum & Fern, Reference Tabassum and Fern2023). Beyond economic impact, charities also play important civic roles by fostering social cohesion, amplifying community voice, and advocating for marginalized groups (Carmel & Harlock, Reference Carmel and Harlock2008; Clifford, Reference Clifford2012).
Since the 1980s, New Public Management reforms heightened emphasis on accountability, monitoring, and performance evaluation in public discourse. UK institutional funders mirrored these shifts, prioritizing control and alignment with stated objectives. In this context, restricted grantmaking became a commonly used model, with conditions specifying purpose or timing to assure compliance and demonstrable results (Mills, Reference Mills2023). Under UK trust law, such conditions bind trustees to use funds strictly as stipulated, reinforcing the sector’s focus on transparency and stewardship. During the COVID-19 pandemic, many funders adopted more flexible practices, including rapid-response grants, lighter reporting, and more unrestricted or multi-year support (Cairns et al., Reference Cairns, Firth and Hopgood2021). More than 140 funders have since joined the Institute for Voluntary Action Research’s “Open and Trusting” community to sustain these efforts.Footnote 2 At the same time, restricted funding remains a core mechanism through which funders structure support and pursue alignment and verifiable impact (Mills, Reference Mills2023). In short, the UK context offers a relevant setting for examining how funders structure support and how organizations manage the associated constraints.
Data
This study draws on data from the Charity Commission for England and Wales, the regulatory body overseeing registered charities in the region. The Charity Commission maintains a comprehensive public register with annual returns and financial statements prepared under the Statement of Recommended Practice (SORP). As the statutory regulator, the Commission enforces filing requirements and reviews submissions, and this oversight supports accuracy and comparability, providing a robust foundation for our empirical analysis.
After sourcing data from the Charity Commission’s Register of Charities (https://register-of-charities.charitycommission.gov.uk/en/), we conducted a series of data cleaning procedures to ensure the consistency and accuracy of the panel dataset, which spans 2007 to 2023 and initially contained 51,237 observations. First, we removed 1,554 returns whose reporting periods were longer or shorter than 12 months to maintain year-to-year comparability. Second, we excluded 1,703 records with data issues.Footnote 3 Third, to support models with lagged variables, we then retained only organizations observed in at least two consecutive years, which removed 10,069 single-year observations. We then winsorized all variables used in the models at the 1st and 99th percentiles to limit the influence of extreme values while retaining observations. Finally, all monetary variables were adjusted for inflation using the Consumer Price Index (CPI), with 2007 as the base year. The final dataset includes 37,911 observations from 8,698 charities.
Dependent variable
The dependent variable in this study is restricted funding. According to the SORP for charities, restricted funding in the UK refers to donations or other income designated for specific purposes or time periods by the donor or funding body, which can include restricted donations, grants, and sponsorships (Charity Commission and OSCR, 2019). To capture the extent of financial constraint imposed by funders, restricted funding in this study is measured as the proportion of restricted funding to total funding (the total amount of money available to a charity), representing the share of a charity’s financial resources subject to these limitations.
Independent variables
For organizations’ financial characteristics, financial reserves refer to unrestricted funds set aside for general use, excluding endowments and restricted funds, to support stability, address unforeseen needs, and enable planning (Charity Commission and OSCR, 2019). The Charity Commission requires charities to maintain a reserves policy that specifies the amount of reserves held, their purpose, and their intended use. This policy must be disclosed in annual reports, promoting transparency and accountability to funders, beneficiaries, and the public. Following established literature (e.g., Irvin & Furneaux, Reference Irvin and Furneaux2022; Pizzini & Sterin, Reference Pizzini and Sterin2025), the reserves level is measured using the reserves ratio, calculated as the proportion of reserves to annual total expenditures. Administrative costs are defined as the proportion of resources allocated to general operational expenses. This variable is measured as the ratio of administrative costs to total expenditures, a commonly used metric in the literature (e.g., Hung & Berrett, Reference Hung and Berrett2021).
For organizational characteristics, we measured a charity’s service-area breadth and beneficiary breadth using classifications provided by the Charity Commission. The Commission categorizes charities’ activities (“what”) into 17 service areas, such as Accommodation/Housing, Arts/Culture/Heritage/Science, Disability, Education/Training, and Environment/Conservation/Heritage. It also identifies seven target groups served (“who”), such as Children/Young People, Elderly/Old People, and People of a Particular Ethnic or Racial Origin. Service-area breadth is determined by counting the service categories reported by each charity, while beneficiary breadth reflects the number of target groups reported. Charities may report multiple service categories and multiple beneficiary groups.
Control variables
To account for potential confounding factors, the analysis incorporates several control variables. Government funding, measured as the proportion of government contracts and grants to total income, may influence restricted funding levels in two ways. On one hand, government funding typically includes usage restrictions. On the other hand, it may signal trustworthiness and credibility (Lu, Reference Lu2016), potentially shaping the behavior of other funders and the likelihood of funding restrictions. Fundraising efficiency, defined as the ratio of total contributions to fundraising expenditures, captures the cost-effectiveness of solicitation efforts and can also signal performance to donors and funders. Prior work links fundraising expense ratios to giving (e.g., Tinkelman & Mankaney, Reference Tinkelman and Mankaney2007), and recent evidence ties fundraising efficiency to restricted donations (Prentice & Clerkin, Reference Prentice and Clerkin2023).Footnote 4
Workforce composition is included as a control because the balance between volunteers and paid staff may shape funders’ perceptions of capacity and oversight, which in turn may influence their willingness to attach conditions. We proxy this with the ratio of volunteers to full-time employees, using figures from public filings and consistent with prior work that uses staffing composition as a measure of workforce professionalization (e.g., Guan et al., Reference Guan, Dong and Lu2025; Hung & Berrett, Reference Hung and Berrett2023; Sanzo-Pérez et al., Reference Sanzo-Pérez, Rey-Garcia and Álvarez-González2017). Higher values indicate greater reliance on volunteers relative to paid staff. We interpret this variable as a staffing-composition proxy rather than a direct measure of organizational professionalism or overall capacity. Organizational size, measured as the natural logarithm of total assets, reflects an organization’s capacity and financial resources (e.g., Calabrese, Reference Calabrese2018; Prentice & Clerkin, Reference Prentice and Clerkin2023; Shon et al., Reference Shon, Hamidullah and McDougle2019), while organizational age captures the length of time a charity has been in operation.
Model specification
We employed a random-effects model to examine the drivers of restricted funding for two reasons.Footnote 5 First, two primary independent variables, service-area breadth and beneficiary breadth, are time-invariant, and fixed effects would difference them out.Footnote 6 Second, our panel has many organizations (8,698) and few years per organization (average of three), for which random effects are more efficient (Wooldridge, Reference Wooldridge2010). A Hausman test did not reject the random-effects specification. The random-effects model is structured as follows:
where
$ {Y}_{it} $
is the dependent variable indicating the restricted funding of organization i in year t,
$ {\beta}_0 $
is the intercept, β indicates the coefficients for the independent variables,
$ {X}_{it-1} $
, which includes variables we are interested in and other control variables for organization i in year t-1,
$ {\alpha}_i $
, is the time-invariance factor for each organization i,
$ {\mu}_t $
indicates year dummy variables, and
$ {\unicode{x025B}}_{it} $
is the error term for organization i in year t.
Results
Table 1 presents the descriptive statistics of the variables used in the analysis, with their correlations provided in Appendix A. The descriptive statistics of the dependent variable are particularly noteworthy. Among the 37,911 observations, 7,870 (20.76%) report no restricted funding, while 30,041 (79.24%) have some level of restricted funding. This highlights the prevalence of restricted funding among UK charities. Across all observations, restricted funding has a mean value of 0.195 and a standard deviation of 0.260, indicating that, on average, nearly 20% of charities’ funding is restricted. However, this proportion varies significantly across organizations, with some relying on restricted funding for nearly all their funding (up to 97.5%). Interestingly, this modest average level of restricted funding, coupled with substantial variability, aligns closely with findings from studies on U.S. nonprofits (e.g., Hung, Reference Hung2021; Surysekar et al., Reference Surysekar, Turner and Wheatley2015).
Descriptive statistics
Table 1. Long description
The table contains six columns: Variable, N, Mean, S D, Min, and Max. It is divided into three sections.
1. Dependent variable:
- Restricted funding: N 37,911, Mean 0.195, S D 0.260, Min 0, Max 0.975.
2. Independent variables:
- Financial reserves: N 37,911, Mean 0.803, S D 1.670, Min 0, Max 12.62.
- Administrative costs: N 37,911, Mean 0.189, S D 0.199, Min 0, Max 1.
- Service-area breadth: N 37,911, Mean 2.953, S D 1.968, Min 1, Max 10.
- Beneficiary breadth: N 37,911, Mean 2.424, S D 1.587, Min 1, Max 7.
3. Control variables:
- Government funding: N 37,911, Mean 0.225, S D 0.300, Min 0, Max 0.991.
- Fundraising efficiency: N 36,069, Mean 2.684, S D 13.52, Min 0, Max 113.8.
- Workforce professionalization: N 37,656, Mean 4.550, S D 14.27, Min 0, Max 111.1.
- Organization size: N 37,908, Mean 14.39, S D 1.571, Min 11.41, Max 18.6.
- Organization age: N 37,911, Mean 21.97, S D 15.01, Min 1, Max 59.
On the independent variable side, financial reserves indicate that charities in our sample maintain an average reserve level of 0.8, suggesting that these organizations typically set aside resources equivalent to approximately 9.6 months of operating expenses (80% of annual expenditures). However, the distribution is highly dispersed, with a standard deviation twice the mean, reflecting significant variation across organizations. In contrast, administrative costs exhibit a more consistent pattern, with a mean of 18.9% and a standard deviation of a similar magnitude, suggesting that most charities allocate roughly 20% of their total expenditures to administrative expenses. Charities also display considerable diversity in their service provision, operating across an average of nearly three distinct service areas (mean = 2.95) and serving just over two beneficiary groups (mean = 2.42). These figures indicate that many organizations adopt a multiservice approach and cater to diverse populations.
Table 2 presents the regression results. Columns 1 and 2 display the results with service-area breadth and beneficiary breadth included separately due to concerns about their potential high correlation (0.49). Column 3 includes both variables. Overall, the findings are highly consistent across specifications, so we treat Column 3 as the main specification. Financial reserves are negatively associated with restricted funding at the 1% level. Specifically, a 10-percentage-point increase in the reserves ratio corresponds, on average, to about a 0.097-percentage-point decrease in the restricted-funding share. In contrast, administrative costs do not have a statistically significant effect on restricted funding (p > 0.1), suggesting limited salience of overhead for restriction decisions.
Regression results
Table 2. Long description
The table presents regression results for Restricted funding across three numbered columns: 1, 2, and 3.
Key variables and their coefficients (with standard errors in parentheses) include:
* Financial reserves (t minus 1): Consistent negative coefficient of approximately minus 0.0097 (p < 0.01) across all models.
* Administrative costs (t minus 1): Negative coefficients around minus 0.009, not statistically significant.
* Service-area breadth: Positive coefficient of 0.0140 (p < 0.01) in model 1 and 0.0110 (p < 0.01) in model 3.
* Beneficiary breadth: Positive coefficient of 0.0142 (p < 0.01) in model 2 and 0.00756 (p < 0.01) in model 3.
* Government funding (t minus 1): Positive coefficients between 0.00269 and 0.00361, not statistically significant.
* Fundraising efficiency (t minus 1): Negative coefficient of minus 0.00026 (p < 0.01) across all models.
* Workforce professionalization (t minus 1): Positive coefficient of approximately 0.00033 (p < 0.05) across all models.
* Organization size (t minus 1): Negative coefficients between minus 0.00616 and minus 0.00654 (p < 0.01).
* Organization age (t minus 1): Negative coefficients between minus 0.000486 and minus 0.000715 (p < 0.01).
* Constant: Positive values between 0.215 and 0.232 (p < 0.01).
All models include Year fixed effects. The total number of observations is 26,985 across 8,693 organization I D entries. Significance levels are indicated by asterisks: three for p < 0.01, two for p < 0.05, and one for p < 0.1.
Note: Robust standard errors in parentheses; ***p < 0.01, **p < 0.05, *p < 0.1.
Both service-area breadth and beneficiary breadth are positively and significantly associated with restricted funding at the 1% level, indicating that organizations operating across more domains or serving more groups face higher restriction levels. Specifically, each additional service area corresponds to about a 1.10-percentage-point increase in the restricted-funding share, and each additional beneficiary group corresponds to about a 0.756-percentage-point increase, on average.
The control variables provide additional information. Government funding does not appear to influence restricted funding, as its coefficients are not statistically significant across specifications. Conversely, higher fundraising efficiency is associated with lower levels of restricted funding, with coefficients significant at the 1% level. Workforce professionalization also shows a positive and statistically significant relationship (5% level), indicating that organizations with higher ratios of volunteers to full-time employees (i.e., lower levels of professionalization) are likely to experience more restricted funding. Both organizational size and age have strong negative effects on restricted funding, also significant at the 1% level, indicating that larger and more established organizations are likely to receive less restricted funding.
We also conducted several sensitivity analyses to assess whether the main findings are driven by alternative treatments of the data or model specification. First, we re-estimated the models using 5th/95th percentile winsorization (instead of 1st/99th) to assess sensitivity to extreme values; the results are reported in Appendix B, Table B1, Column 1. Second, to address skewness in variables, we estimated log-transformed specifications on complete cases (dropping zero/negative values) and provided the results in Appendix B, Table B1, Column 2. Third, as noted earlier, 20.76% of observations in our sample have no restricted funding, while 79.24% have restricted funding. To assess whether the main explanatory variables influence funders’ decisions to impose restrictions (yes or no), we conducted a robustness check, with the results presented in Appendix B, Table B1, Column 3. Fourth, under the SORP, restricted funding in the UK includes endowment funding, as these funds are subject to donor-imposed restrictions. However, studies in other countries do not always include endowment funding in their measurements. As a robustness check, we excluded endowment funding and remeasured restricted funding as the proportion of restricted income to total income, following prior studies (e.g., Hung, Reference Hung2021). The results for this analysis are reported in Appendix B, Table B1, Column 4. Across these checks, the substantive pattern of the main results remains largely consistent in direction and significance. These checks increase confidence that the findings are not artifacts of a particular specification, though they do not remove broader limitations related to measurement and model-based inference.
Discussion and conclusion
Restricted funding remains a frequently debated topic in nonprofit finance. Rather than revisiting organizational outcomes, we shift attention to the organizational antecedents. Guided by agency and signaling theories, we examine how observable recipient characteristics are associated with variation in funding restrictions under conditions of information asymmetry. Using a panel dataset of UK charities, this study complements a literature largely focused on U.S. nonprofits and adds a comparative perspective. The results offer a more nuanced view of restricted funding practices by identifying recipient-side organizational patterns associated with variation in funding flexibility across charities.
Theoretically, the findings highlight an organization-level layer of explanation for restricted funding. Agency and signaling theories are connected through the common problem of information asymmetry: funders cannot fully observe how recipient organizations will use the resources and therefore may rely on observable organizational characteristics when assessing risk, capacity, and stewardship. The observed relationships are consistent with this account. Financial reserves may serve as signals of financial stability that reduce uncertainty, whereas broader service areas and wider beneficiary scope may increase ambiguity about resource allocation and monitoring. These findings suggest that organizational characteristics constitute an important recipient-side source of variation in funding conditions, complementing rather than replacing explanations centered on funder preferences or funder-nonprofit relationships. More broadly, the study identifies concrete organizational characteristics through which information asymmetry may be associated with tighter or looser funding conditions and provides testable pathways for future research on how funders interpret these signals in actual funding decisions.
More specifically, the results show that lower financial reserves are associated with higher levels of restricted funding, a pattern consistent with the argument that weaker financial buffers may increase uncertainty about organizational stability and resource use. In agency terms, thin reserves may raise concerns about an organization’s ability to manage financial shocks and sustain programs without additional controls. From a signaling perspective, they may communicate more limited capacity to absorb volatility and meet commitments. This pattern aligns with and complements research on reserves as indicators of stability and competence (e.g., Calabrese, Reference Calabrese2018; Kim & Mason, Reference Kim and Mason2020). Practically, organizations may reduce funder concern by building and maintaining adequate reserves, documenting drawdown and replenishment rules, and clearly reporting reserve activity. Clear policies and consistent practices can reassure funders and potentially support more flexible funding aligned with strategy.
In addition, administrative cost levels are not a significant predictor of restriction levels in our models. This finding suggests overhead may be less central to funders’ restriction decisions than often assumed, at least relative to the other organizational characteristics examined here. The point is not that overhead is irrelevant, but that its salience for restriction decisions appears weaker. This result aligns with recent work showing mixed or limited overhead effects (e.g., Qu & Daniel, Reference Qu and Daniel2021; Schubert & Boenigk, Reference Schubert and Boenigk2019) and tempers the view that funders are uniformly averse to administrative costs. Practically, nonprofits may not need to prioritize overhead minimization solely to avoid restrictions. Instead, they may focus on clear budgeting, credible cost allocation, and transparent reporting that links administrative capacity to program quality and outcomes (Lu et al., Reference Lu, Shon and Zhang2020; Wing & Hager, Reference Wing and Hager2004).
Moreover, nonprofits with broader service portfolios and wider beneficiary reach tend to face higher levels of restricted funding. This pattern highlights the potential importance of coordination and monitoring complexity in multiservice settings and extends the small literature on service diversification (Kistruck et al., Reference Kistruck, Qureshi and Beamish2013; Mendoza-Abarca & Gras, Reference Mendoza-Abarca and Gras2019). While expanding service areas can support survival and growth (James, Reference James1983), it can also make resource allocation and goal alignment more difficult for external funders to assess, which may be associated with tighter use conditions. A similar logic applies to beneficiary breadth: addressing varied populations may make it more difficult to determine how resources are distributed across intended groups, complementing work on accountability and beneficiary engagement (Benjamin, Reference Benjamin2013). Practically, nonprofit managers can reduce ambiguity by clarifying how programs map to funder priorities, explaining allocation rules across areas and populations, and presenting program-level outcomes. These steps may help address funder concerns and support more flexible funding.
This study has several limitations. First, while our analysis identifies important associations between financial and organizational characteristics and levels of restricted funding, these findings should not be interpreted as causal. Second, our focus on organizational antecedents captures only one recipient-side layer of explanation. Funder preferences, grant characteristics, and relational dynamics between funders and nonprofits may also shape restriction levels or condition how organizational characteristics are interpreted. Future studies combining organizational and funder-level information could examine how recipient characteristics interact with funder preferences and prior relationships. Third, our dependent variable captures aggregate restricted funding and does not distinguish among restriction types or revenue sources. This aggregation may mask heterogeneous mechanisms and could attenuate or blur some relationships if different forms of restricted funding respond to different organizational factors. Future research should disaggregate restricted funding by type and source to test whether these dynamics vary across revenue streams and funding purposes.
Fourth, the dataset may underrepresent smaller UK charities, as these organizations are not required to report detailed financial and organizational information to the Charity Commission. This may limit our understanding of restricted funding across the broader nonprofit landscape. Lastly, because the study focuses on charities in England and Wales, the findings may not fully generalize to other national contexts, where regulatory, cultural, and funding environments differ. Future research could address these limitations by using more detailed and representative data and extending the analysis to nonprofits in other national settings.
In conclusion, this study examines organizational antecedents of restricted funding, an underexplored aspect of nonprofit finance. By analyzing how financial reserves, administrative costs, service-area breadth, and beneficiary breadth are associated with variation in restriction levels, we identify a recipient-side layer of explanation for why nonprofits experience different degrees of funding flexibility. The results advance the literature on funder-nonprofit relationships by showing that observable organizational characteristics are systematically associated with funding conditions, while complementing rather than displacing funder-side and relational explanations. They also offer practical implications for nonprofits navigating funding constraints. As restricted funding continues to shape the sector’s financial landscape, the evidence points to strategies that may help balance funder requirements with organizational flexibility.
Data availability statement
The data that support the findings of this study are available in the Register of Charities maintained by The Charity Commission for England and Wales at https://register-of-charities.charitycommission.gov.uk/en/.
Competing interests
The authors have no conflict of interest to declare.
Appendix A. Correlation table
Correlation table
Table A1. Long description
A lower-triangular correlation matrix with 10 rows and 10 columns. The variables are: 1. Restricted funding, 2. Service-area breadth, 3. Beneficiary breadth, 4. Financial reserves, 5. Administrative costs, 6. Workforce professionalization, 7. Fundraising efficiency, 8. Government funding, 9. Organization size, and 10. Organization age.
Key correlations include:
* Variable 1 (Restricted funding) correlates positively with 2 (0.115) and 3 (0.095), and negatively with 4 (-0.160), 9 (-0.075), and 10 (-0.057).
* Variable 2 (Service-area breadth) has a strong positive correlation with 3 (0.491) and a negative correlation with 10 (-0.146).
* Variable 4 (Financial reserves) correlates positively with 9 (0.295).
* Variable 8 (Government funding) correlates negatively with 4 (-0.155), 6 (-0.108), and 9 (-0.161).
* Variable 9 (Organization size) correlates positively with 10 (0.315).
Statistical significance is indicated by asterisks where three asterisks mean p < 0.01 and two asterisks mean p < 0.05. Diagonal values are all 1.000.
Note: ***p < 0.01, **p < 0.05, *p < 0.1.
Appendix B. Robustness checks
Robustness check results
Table B1. Long description
The table presents regression results for four models, labeled 1 through 4, where the dependent variable is Restricted funding.
Variables and their coefficients (standard errors in parentheses):
* Financial reserves (t minus 1): Model 1 is -0.0290*** (0.00220); Model 2 is -0.0356*** (0.00642); Model 3 is -0.0123*** (0.00176); Model 4 is -0.00366*** (0.00102).
* Administrative costs (t minus 1): Model 1 is -0.00357 (0.00856); Model 2 is -0.0179 (0.0478); Model 3 is 0.00689 (0.0122); Model 4 is 0.00113 (0.00519).
* Service-area breadth: Model 1 is 0.0124*** (0.00164); Model 2 is 0.106*** (0.0138); Model 3 is 0.0114*** (0.00224); Model 4 is 0.00640*** (0.00117).
* Beneficiary breadth: Model 1 is 0.00684*** (0.00188); Model 2 is 0.106*** (0.0168); Model 3 is 0.0123*** (0.00267); Model 4 is 0.00563*** (0.00146).
Summary Statistics:
* Control variables: YES for all models.
* Year fixed effects: YES for all models.
* Observations: Model 1 is 26,985; Model 2 is 26,832; Model 3 is 26,985; Model 4 is 23,720.
* Number of organization I D: Model 1 is 8,693; Model 2 is 8,646; Model 3 is 8,693; Model 4 is 7,873.
Asterisks denote p-values: *** p < 0.01, ** p < 0.05, * p < 0.1.
Note: Model (1) applies 5th/95th percentile winsorization to all variables. Model (2) estimates log-transformed specifications on complete cases. Model (3) uses a binary dependent variable indicating any restricted funding (yes = 1, no = 0). Model (4) sets the dependent variable to the proportion of restricted income to total income. In all models, the remaining covariates match those in Table 2. Robust standard errors in parentheses. ***p < 0.01, **p < 0.05, *p < 0.1.
Open access
