Introduction

Accountability is a key concern receiving wide attention in the nonprofit literature and is especially critical considering the rise of the third sector in recent years. As governments privatize and de-centralize state functions to third parties (Agrawal and Ribot 1999; Kettl 2000; Milward and Provan 2000; Sorens 2009), there has been a significant increase in the numbers of nonprofit organizations (NPO), their relative power and influence, the amounts and origins of the funds to which they have access, the size of their tax exemptions, and the range of service areas in which they are involved (Ebrahim 2005; Ebrahim and Weisband 2007; Jordan and Van Tuijl 2006; Weisbrod 1997). As the influence of NPOs increases, so have incidences of scandals and corruption (Carman 2010; Gibelman and Gelman 2001; Gibelman and Gelman 2004; Greenlee et al. 2007). Accountability concerns have increased scrutiny and study of NPOs from governments, watchdog groups, academia, and media (Carman and Fredericks 2010).

The idea of accountability is ubiquitous in our society and yet is noted for its complexity, ambiguousness, and context dependency (Ebrahim 2003a, b; Ebrahim and Weisband 2007; O’Dwyer and Unerman 2008). Even in the face of much research across public, business, and nonprofit sectors, accountability is a “multifaceted concept fraught with ambiguity” (Salamon 2002, p. 524) and “…has become a cliché and, like all clichés, is a substitute for thinking” (Behn 2001, p. 6). Accountability has such widely varying interpretations and a high degree of abstraction that common understanding of its detailed implications is often elusive (Pollitt and Hupe 2011). Consequently, the disparities and ambiguities concealed by differently interpreted terminology and logically inconsistent frameworks inhibit theory building. In order to understand the detailed nature of accountability and its application in various contexts and to provide conceptual clarity, a consistent set of frameworks and definitions is needed among a particular set of accountability stakeholders. This article aims to uncover the sources of ambiguity in accountability literature and proposes an alternative framework to resolve these concerns.

In the first section of this article, we analyze key accountability frameworks from the public, private, and nonprofit literatures, and begin to pinpoint the sources of ambiguity. By synthesizing the frameworks in Table 1, we analyze the accountability loci—the particular route by which accountability is achieved and the stakeholders involved in the relationship. We consider the relationship between legitimacy and accountability, arguing that accountability is the mechanism by which legitimacy is created in practice. In the second section, we overcome the ambiguity critique by developing a holistic accountability framework (HAF) that facilitates defining, measuring, and implementing accountability in complex, multi-stakeholder environments, by providing a means to operationalize commonly encountered but ambiguous accountability goals such as transparency, controllability, and responsibility. Finally, we suggest limitations of the approach and future research needed.

Table 1 Summary of key accountability frameworks reviewed in the public, private, and nonprofit literatures
Full size table

This study is important because the issue of accountability is likely to become increasingly challenging. The boundaries between sectors are progressively more blurred and many nonprofits are intertwined in a network of private donors, government contracts, and private profit making spin offs (Brandsen et al. 2005; Eisenberg 2005; James 2003). As economic pressures increasingly stress government and donors, the need to compete for funding, contracts, and in some cases clientele often endows NPOs with the same operating logic as commercial organizations (Abeles 2008; Avina 1993; Jordan and Van Tuijl 2006). Furthermore, considering the amount of scholarly attention on accountability, the extent to which theoretical literature has produced practically oriented advice to tackle these issues is limited (Knutsen and Brower 2010). Our conclusions are applicable for both scholars and practitioners, and we expect the findings to be utilizable immediately by senior managers to develop accountability strategy and to understand the conflicts between different forms of accountability and the varying needs of stakeholders.

Conceptualizing Accountability

In this section, we summarize and analyze key accountability concepts and frameworks across the public, nonprofit, and business sector literatures, considering also the general trends in understanding accountability. Before turning to the review, however, we must briefly describe the literature search and analysis methodology.

This article is based on an extensive literature search from journals in online full text collections of Wiley Interscience, SpringerLink, Sage, Elsevier, and Emerald Insight, followed by the academic databases ABI-INFORM, JSTOR, and Web of Knowledge. Searches were limited to the journal categories: nonprofit sector, voluntary sector, third sector, public administration, public management, management, business, political science, and social sciences interdisciplinary. Searches employed the keywords: accountability or accountability + legitimacy and at least one of framework, model, concept, and theory. As many studies covered accountability in the context of discussion around another concept (e.g., democracy) or examined a specific aspect of accountability in a detailed manner (e.g., implementation of new regulation for nonprofits), we refined the search to include only studies that looked at the issue of accountability in totality or explicitly claimed to develop a new framework. The timeframe was restricted from 1980 to 2011 as the majority of theoretical development on accountability in the nonprofit sector is quite recent (Ospina et al. 2002). Using the references from this sample, we identified books on the subject of NPO accountability and other relevant studies.

Once the sample set of studies was obtained, we set about to list the frameworks and authors, examining whether they created an entirely new approach or either used or developed others’ approaches. Taking those studies that offered a new approach, we applied a standard lens used to define theory, typologies, and models (see, for example, Jaccard and Jacoby 2010; Miner 2005) to examine how variables, constructs, linking propositions, hypotheses, boundaries, and assumptions were presented in each study. We also examined the levels of analysis of each framework and its domain of applicability. We intended our literature review to trace the development of understanding of accountability rather than present a detailed examination of all frameworks. Therefore, while we did conduct the theoretical analysis described on all frameworks, we chose to present only those frameworks that were used by five or more subsequent studies.

Given the vast literature on accountability, we focused on academic articles in peer-reviewed journals and omitted dissertations and articles not in English. Concerning the nonprofit literature, a potential bias may be present as the majority of the articles in our sample were from US-based institutions and authors. There were many articles in our first search sweep that were European in origin, yet these mainly focused on detailed cases of accountability such as specific implications of local legislation, rather than broader issues in accountability.

Accountability in the Public Sector

In the context of public sector, Behn (2001) summarizes the traditional reasons for accountability, as developed in the study of public administration: to ensure that an organization operates fairly; to ensure responsible spending in accordance with defined rules; to guard against the use or abuse of power; to ensure performance such that organizations or individuals meet targets; and finally, to ensure that individuals entrusted with public money act in accordance with societal standards and values. In essence, these five reasons reflect fundamental principles of legitimate, fair, and equitable democracy; however, as with many other aspects of democratic systems, the actual realization of such ideals is far from simple.

From the late 1800s, early scholars in public administration have considered the practical implementation of such fundamental principles. Woodrow Wilson and others argued that by separating politics from administration, government bureaucracies would be given clear orders derived from political deliberations, and accountability would be retained at the political level (Goodnow 1900; Wilson 1887/1987). With the rapid rise in the scope and complexity of government functions at the turn of the century, however, the politics–administration dichotomy broke down forcing scholars to consider the issues of accountability for administrators. In a classic debate, Friedrich (1940) and Finer (1941) argued over specific methods to ensure bureaucratic accountability by either controls internal to an organization such as professional codes of conduct or by constraints imposed externally to the organization through political or legal means.

This early debate developed the notions of establishing instrumental accountability either through external rules or through socially created internal standards. External means encompass representative accountability relationships in political theory and principal–agent relationships for legal or bureaucratic control. Socially derived means encompass professional, moral, or cultural rules that are created through mutual social processes. As in either case, accountability is a relational concept that describes some agreement between two or more parties; at this point, we must differentiate two components of accountability: explicitly identifying the stakeholders in the accountability relation such as “regulatory agencies” and “agency leader,” and identifying the locus of accountability in order to understand the context of that relationship such as “legal” and “bureaucratic.” Examples are given in detail in the second column (Accountability to Whom?/Stakeholders) and third column (Accountability Loci) of Table 1.

A more recent and widely used framework expanding on the notion of external versus internal control through different accountability loci is from Romzek and Dubnick’s (1987) paper on NASA’s role in the Challenger disaster. They define accountability as “the means by which public agencies and their workers manage the diverse expectations generated within and outside the organization” (p. 228). Their framework incorporates methods of control internal and external to an organization by focusing on four systems of public accountability: external legal means, external political means, internal bureaucratic means, and internal professional means. These accountability systems are created through the interplay of defining and controlling expectations, and the amount of discretion given in defining those expectations. Romzek and Dubnick’s study represents one of the first attempts to conflate external and internal accountability loci in a single framework and, most importantly, demonstrates tension between them. As a result of NASA’s technical and managerial problems, the pursuit of political means distracted from professional accountability. Subsequent studies using this framework have yielded similar conclusions about the tensions inherent between the various methods of accountability in the public sector (Romzek and Ingraham 2000).

Another widely used framework in public administration is from Koppell’s (2005) study on the Internet Corporation for Assigned Names and Numbers. In an effort to resolve “conceptual fuzziness,” Koppell presents five distinct dimensions of accountability: transparency, liability, controllability, responsibility, and responsiveness. Rather than denoting specific mechanisms, methods, or accountability loci, these dimensions focus on the end product or goal of accountability in order to define its meaning. At this point, we highlight the “Accountability Goals” column of Table 1 which draws attention to the purpose or outcome of accountability. Lending further testament to ambiguity, widely varying descriptions and abstractions of accountability goals were encountered in the various models reviewed.

There is growing recognition that accountability is not only instrumental and objective. Scholars in public administration have noted that accountability depends on the particular perspective of involved stakeholders (Bergsteiner and Avery 2008; Dubnick 2005; Hanberger 2009), while others have understood accountability as a social phenomenon constructed by stakeholders in the system (Ebrahim and Weisband 2007). In research about accountability in a government funded program, O’Connell (2005) moves away from the instrumental conception by arguing that accountability inspired by agency theory cannot apply in a context of a multiple actor network. O’Connell combines two approaches to accountability: new institutionalism (DiMaggio and Powell 1991) in which organizations or networks comprised interactions between themselves and others in the organization field; and Kearns’ accountability environment (Kearns 1994) in which a multitude of forces place pressure on organizations and actors to engage in certain behaviors and dissuade actors from engaging in undesirable behavior. Thus, accountability is “the multidimensional product of many forces operating in the accountability environment…, best described as emerging from their multi-stranded interactions” (O’Connell 2005, p. 86).

Behn (2001) echoes this sentiment by calling for a move from adversarial accountability to collective responsibility, arguing that a “360-degree” accountability system feedback model should replace the standard, unidirectional, subordinate-superior hierarchical methods. He envisions an organization in which people are not only accountable to their immediate superior, but also to everyone in their “accountability environment,” which includes “peers, team members, customers and suppliers” (p. 199). Denhardt and Denhardt (2007) agree that accountability is a complex issue involving a balance of external, internal, and normative social controls. They demonstrate the flaws of oversimplifying democratic accountability using a narrow set of performance measures or by trying to imitate accountability processes from the private sector. The nature of democracy, the role of citizenship, and a public service dedicated to serving citizens in the public interest cannot be ignored. Furthermore, the key to resolving accountability issues is to engage in deliberative dialogue with all stakeholders.

Accountability in the Private Profit Making Sector

Given its strong relationship to democracy, the study of accountability has been pursued traditionally from the perspective of the public sector. As civil action in both national and international contexts is increasingly occurring through the use of profit making private actors, there is impetus to examine the idea of accountability in the private sector with more circumspection. Unlike in the public sector where government process and policy results are subject to complex and often conflicting accountability requirements, in the profit making commercial world, the basic notion of accountability focuses on ensuring that the company operates within the law and that profit is gained. Given the executive hierarchical nature of business organizations, accountability is a corporate governance issue—a responsibility falling primarily with the board of directors (OECD 2004). The specific accountability arrangements depend on whether a company is “public,” such that shares can be traded freely, or “private,” where the number of shareholders and their rights to transfer shares are limited (Mulgan 2000).

In recent decades, the concept of accountability in the private, for-profit sector has been expanded, with business ethics scholars speaking of stakeholder models of accountability and of corporate social responsibility (CSR) instead (Stoll 2008). Four categories of CSR approaches have been identified (Garriga and Melé 2004). Instrumental approaches assert that a corporation is solely an instrument for wealth creation; hence only by maximizing profits will private companies truly benefit the public good. Political approaches to CSR emphasize the social power of corporations and the responsibility this bestows upon leadership. This responsibility must be effectively exercised to avoid the risk of losing social power and consequently business. These political and instrumental approaches are limiting in that CSR is identified only as a means for profit making, with some even viewing CSR as a thinly veiled attempt to improve reputation and an activity that undermines productivity (Schrage 2005). Furthermore, it is questionable whether chief executive officers have a right to make donations to charities with money that belongs to stockholders not to corporate leadership.

Two other approaches take a more expansive view of CSR. Ethical approaches such as sustainable development and stakeholder models argue that a company has a moral obligation to act ethically with due process and respect for the environment and wider community of stakeholders, often balancing through charitable giving and corporate citizenship any harm it may do in the community (Campbell 2005; Swift 2001). Integrative approaches extend the stakeholder view even further, replacing the concept of CSR with corporate social integration or the “fourth sector” which recognizes that a healthy society ultimately creates expanding demand for business, thus to ignore opportunities where corporations can enact social good may hinder success in the long term (Porter and Kramer 2006; Sabeti 2009).

The challenges inherent in CSR and corporate social integration are in some senses analogous to the challenging task faced by governments to strive for democratic values of justice, equity, and fairness: while these are laudatory principles, they invoke difficult tensions with other important private sector values of efficiency and effectiveness. While corporate governance is increasingly concerned with CSR given the recent cases of corporate malfeasance, these new requirements represent an expansion of the governance role (Strandberg 2005; Van den Berghe and Louche 2005). Furthermore, the underlying socio-political assumptions of CSR are not neutral and require additional mechanisms to establish legitimacy for an organization (Gjølberg 2010). This suggests that in the private sector, CSR cannot be directly equated with accountability in the public sector. In a democratic system, private organizations are typically endowed with certain rights that guarantee freedom of action within certain legal or regulatory constraints; a condition that establishes “normative” legitimacy (Koppell 2008). Yet in the absence of legally provided normative legitimacy, how can private organizations establish what Koppell (2008) calls “cognitive” and “pragmatic” legitimacy?

These complex questions are outside the scope of this article, however, several key conclusions should be noted. CSR recognizes that in order for a private business to establish legitimacy beyond that bestowed by legal governmental rules, wider groups of stakeholders rather than only customers and shareholders must be considered through a process of shared dialogue (Brennan and Solomon 2008). Furthermore, to varying degrees, CSR concepts recognize some form of accountability to this wider community of stakeholders. Consequently, the variety of mechanisms created to implement CSR mirror, to a striking extent, the accountability mechanisms of NPOs and government organizations (Hohnen 2007; Ward et al. 2010).

Accountability in the Nonprofit Sector

Nonprofit organizations in all their potential forms and service areas exist in a challenging position in regard to accountability. In some respects, in order to survive they must operate in a manner analogous to profit making organizations by keeping strategies private, by ensuring that funding is not passed on to competitors and by maintaining financial solvency. In other respects, they operate like quasi-governmental democratic organizations by ensuring accountability, fairness, equality, and participation in decision making (Ackerman 2004). The balance between quasi-private and quasi-public modes depends, to a large extent, on the particular context, the size of the organization, the service area, reputation, operational capacity, financial solvency of both organization and donors, and the availability of volunteer staff (Dees and Anderson 2003).

There have been various attempts to develop accountability frameworks in the nonprofit literature and many have borrowed aspects from the field of public administration (Ospina et al. 2002). In commonality with nonprofits, public administration shares the pursuit of the public interest, and, as a result, accountability has been a widely studied concept since the inception of nonprofit studies. We briefly cover several of the key frameworks that are relevant to the discussion.

Scholars have highlighted the importance of taking a stakeholder-based approach to accountability (Edwards and Hulme 1996; Najam 1996). Christensen and Ebrahim (2006, p. 196) define accountability as “being answerable to stakeholders for the actions of the organization, whether by internal or external initiation…(t)hus, NPO can be accountable on multiple levels: upward, lateral, and downward.” The definition of accountability is broadened to consider being held responsible by others (upward accountability), taking responsibility to staff and volunteers (lateral accountability), and being responsible to the needs of clients (downward accountability), which also includes the public trust. Both practitioners and scholars have noted the difficulty in reconciling these multiple accountabilities as, in many situations, they are in conflict with one another (Edwards and Hulme 1996; Smith and Lipsky 1993). Each level of accountability requires different accountability mechanisms, formal and informal; however, upward and downward accountability can be better achieved in the presence of strong lateral accountability mechanisms because ultimately, accountability should create a culture of trust between stakeholders.

In their study on accountability in Amnesty International, O’Dwyer and Unerman (2008) expand the stakeholder perspective to differentiate between hierarchical accountability and holistic accountability. Hierarchical accountability is “narrowly functional, short-term in orientation and favors accountability to those stakeholders who control access to key resources for both resource use and immediate (campaign) impacts” (p. 803). Also called functional accountability in the literature, this view focuses on control and external oversight of NPOs and considers only those most influential and powerful stakeholders. Holistic accountability “expands the concept of ‘performance’ articulated within hierarchical accountability to embrace quantitative and qualitative mechanisms concerned with signifying the long term achievement of organizational mission and the impact of this achievement in bringing about structural change” (p. 804). By expanding the perspective to consider a long-term view, they expand the categories of stakeholders to which an NPO must be accountable and emphasize downward accountability. In order to avoid the potentially damaging effects of focusing on control and justification aspects of accountability, they urge senior leadership of NPOs to consider the balance between holistic and hierarchical accountability mechanisms.

Although the stakeholder and holistic views described above were a notable advancement, they still consider accountability as primarily instrumental and based on multiple principal–agent relationships. In common with recent thinking in the public administration literature, many NPO scholars argue that the nature of accountability must be rethought. Ebrahim (2005, p. 82) considers accountability as “a relational, rather than absolute concept: It is about relationships of power among multiple organizational actors…a system of multidirectional and contingent relations” rather than as a “collection of independent binary links.” Similar concepts are also being developed for the governance of NPOs (Stone and Ostrower 2007) and have been recognized in the public administration literature (Bergsteiner and Avery 2008; Goldsmith and Eggers 2004). This conception of accountability departs from the instrumental view that is dominant within public administration literature and much of the NPO literature.

In another departure from the instrumental view, Brown and Moore (2001) treat accountability “not as an abstract, fixed, moral ideal but instead as a strategic idea to be formulated and acted on by…[NPOs]…with the goal of better understanding and achieving their strategic purposes” (p. 570). They recognize that accountability is never universal but, instead, is contingent on the particular environment. In the context of large NPOs, there may be multiple stakeholders involved and it may not be entirely clear how individual accountability requirements—if perceived as a binary principal–agent relationship based on some financial or legal basis—should be prioritized, especially when stakeholders have differing aims and expectations. As accountability involves ensuring that the various expectations of stakeholders are met, developing an NPO’s mission and strategy must include a decision about structuring its accountability relationships.

Brown and Moore (2001) focus on the “strategic triangle” of value, support and legitimacy, and operational capacity, as “three crucial calculations leaders must make if their organizations are to survive, produce socially valuable results, and successfully adapt to changing circumstances” (p. 575). Each element of this framework corresponds to a particular group of stakeholders and a particular type of accountability. They note that no single accountability structure is right for all types of organization and depends in particular on whether the NPO’s purpose is in service delivery, capacity building, or policy influence.

In common with the public administration and private sector, our analysis of the nonprofit literature demonstrates that the concept of accountability has been expanded from a purely instrumental view to a concept that is context dependent and derived from mutual social deliberation between organizations and stakeholders affected by the actions of the organization. Although the nonprofit literature is comprehensive in its coverage of various mechanisms to ensure operative accountability, there is little that deals with the accountability in relation to the fundamental legitimacy of organizations.

Without doubt, NPOs have realized great achievements in civil society; however, there are fundamental unresolved questions about the true legitimacy of such organizations (Abeles 2008; Jordan and van Tuijl 2000). This legitimacy dilemma derives from the fact that most NPOs straddle the boundary between public and private. As private entities, organizations are legally guaranteed the right to association and to work toward whatever aims the founders and funders desire, within certain regulatory constraints (Reich et al. 2009). In another sense, many NPOs are providing public services and must consider that the recipients of their service have at least some stake in determining the nature of the service.

Brown (2008) offers a way to begin understanding these issues, arguing that legitimacy and accountability should be considered by NPOs from the point at which organizational strategy is created, during detailed planning and implementation, and through to the final evaluation of implemented programs. Legitimacy is defined as “perceptions by key stakeholders and wider publics that the existence, activities, and impacts of (NPOs) are justified and appropriate in terms of central social rules and regulations, values and norms, and widely held expectations” (p. 34). Brown’s central tenet is that rather than relying on or waiting for external agents, governments, or professional bodies to define what is credible and legitimate, NPOs should, as part of their core strategy, “actively construct legitimacy arguments and accountability systems” (p. 11). He suggests that this construction follows the process of defining stakeholders, negotiating indicators, understanding the wider consequences of the NPO’s activities, and building organizational support systems. Furthermore, these constructions should occur at all levels: individual, organizational, and domain or sector. The key point of Brown’s model is that the accountability system is the mechanism by which NPO legitimacy is operationalized into practice.

Although from the perspective of an organizational user, Brown’s model is comprehensive and places accountability at the center of an organization’s strategic planning, operations, and results, it does not offer specific processes to achieve these constructions, nor does it highlight—as we will in the next section—some of the difficulties commonly encountered in accountability frameworks.

Unpacking Accountability

The literature review demonstrates many different ways of viewing accountability and many different frameworks used to organize and describe the concept. In this section, we first pinpoint the aspects of each framework that cause the often stated ambiguity in accountability: inconsistent theoretical construction and failure to consider levels of analysis. Second, we develop a proposal—the HAF—to resolve the ambiguities.

All the accountability frameworks reviewed are built from two distinct types of components: variables and constructs, defined in Table 2. Variables are “tangible” physical entities that exist in the world, are observable, and have multiple values (Miner 2005). These include: stakeholders; the organization or NPO itself; and specific accountability mechanisms such as reports, laws, evaluations, or professional standards. Constructs are abstractions of phenomena created to facilitate understanding. These include accountability outcomes such as “transparency” and organizational groupings such as “political.” A problem inherent in many of the accountability frameworks that contributes to their ambiguity is that they do not distinguish between constructs and variables and, in most cases, blend them together in descriptions or typological frameworks.

Table 2 Components of accountability frameworks
Full size table

For example, Koppell (2005) describes accountability by five abstract constructs: transparency, liability, controllability, responsibility, and responsiveness. Within the description of each construct, Koppell mixes real variables such as specific mechanisms and stakeholders, with further abstract constructs—organizational categorizations such as the undefined “professional norms.” Although other organizational categorizations such as political groupings or government bureaucracies resemble “real” entities, in terms of accountability they represent only abstractions, as to be accountable to a “bureaucracy” only means something in the context of specific mechanisms and specific stakeholder groups within that organization.

Consider Romzek and Dubnick’s (1987) 2-dimensional framework that classifies accountability by the source of control (external and internal) and the degree of scrutiny (high or low). This framework produces four aspects of accountability described as political, hierarchical, professional, and legal. These are not particularly useful, as they do not allow actual accountability processes to be specified in detail. The state of being “politically” accountable can only be defined, in real terms, by referring to specific variables: groups of stakeholders such as congressional committees and interest groups. In reality, the particular types of mechanisms may vary drastically between these two groups, meaning that “political” is only an abstraction. We categorize these constructs as accountability organizational paths, or accountability loci, as they simply organize collections of stakeholders and mechanisms into general groupings in the context of a particular organizational arrangement. Examples and a definition are presented in Table 2.

From the various frameworks analyzed, we can determine several components that are “invariant” to the context; that is, they are required as part of any fundamental definition of accountability. In this description, we clearly delineate between real variables and abstract constructs. In common with the views proposed by Ebrahim, Lee, and others (Christensen and Ebrahim 2006; Ebrahim 2003a, b, 2005; Ebrahim and Weisband 2007; Lee 2004), it is evident that the fundamental meaning of accountability is entirely contingent on the stakeholders and the particular types of mechanisms through which they expect to “see” accountability. Therefore, the variable stakeholders must be the starting point in any attempt to design a framework (our understanding of this term is explained in Table 2). As Edwards and Hulme (1996) and Ebrahim (2003a, b, 2005) demonstrate, it is beneficial to classify stakeholders into upward, downward, and lateral groupings; however, these groupings are, for the most part, irrelevant to the actual mechanisms of accountability that are employed. Although these groupings were groundbreaking in the nonprofit literature, as they explicitly included all stakeholder groups for the first time, in practice they are simply an organizing construct to guide thinking.

The second variable that must be considered as part of any framework is the mechanism, which defines specifically how accountability will happen. For example, Ebrahim (2003a) states five types of mechanisms: disclosures and report tools; performance assessment and evaluation tools; participation processes; self-regulation processes, and social auditing tools. His typology involves assigning certain characteristics to each mechanism (e.g., the extent to which the mechanism is internally or externally induced). While these characteristics are useful to know, again, they are largely irrelevant to the actual mechanisms. In another example with useful—but ultimately extraneous characteristics—described by Avina (1993), Edwards and Hulme (1996) and Najam (1996), mechanisms are categorized by their functional accountability (accounting for resources or resource use) or their strategic accountability (accounting for the wider societal impacts caused by an NPO).

In addition to the “tangible” variables of stakeholders and mechanisms, there is one fundamental “abstract” construct required: the purpose or goals of accountability (see the “definition” column of Table 2). Koppell’s (2005) framework presents a familiar and concise way of describing accountability in terms of its overall purposes of transparency, liability, controllability, responsibility, and responsiveness. In practical terms, when NPO leadership asks the question—what is the functional organizational purpose of accountability?—it may be these five abstract terms that come to mind.

A common problem besetting many accountability frameworks, especially in the public administration literature, is the tendency to stipulate specific stakeholders, mechanisms, and organizational pathways as a definitional part of the framework. Sociological theorists often caution against the employment of universal, objective typologies as they tend toward instrumental control (see, for example, Fay 1975). Many of the authors covered in the literature review noted that the instrumental conception of accountability was contradictory to its true intent and instead argue that accountability is a social phenomenon (Bourgon 2007; Denhardt and Denhardt 2007; Ebrahim and Weisband 2007; O’Connell 2005). Consequently, any attempt to reformulate a framework must begin with general categories of variable and constructs, rather than employing specific values.

In the quest to build a framework that is both theoretically sound and practically useful, we propose using the variables of stakeholders and mechanisms essentially to operationalize essentially the constructs—accountability goals—in a single unified structure. This structure, however, does not depend on the values of any variable or definition of any construct. The key point is that for each specific situation, NPO leadership must perform the exercise of identifying the stakeholders, and then engage in a deliberative dialogue with them to construct the meanings of accountability goals through the combinations of specific mechanisms and stakeholders.

Another point of ambiguity in accountability frameworks stems from the lack of identification of levels of analysis. Weisband and Ebrahim (2007) identify four levels of accountability analysis: individual, organizational, network, and structural. A review of the various frameworks presented in Table 1 reveals that often these levels of analysis are conflated within the same category or model component. For example, O’Dwyer and Unerman (2008, p. 804) define holistic accountability as: “broader forms of accountability for the impacts that an NGO’s actions have, or can have on a broader range of other organizations, individuals, and the environment.” Although the definition conveys the general idea of holistic accountability and the authors give several examples of specific organizational mechanisms, as a conceptual framework, it is clear that more specificity is needed to convey the complexities of various relationships, the nature of which vary considerably between levels of analysis. We suggest that each single mechanism of accountability needs to be understood in the context of a particular level of analysis.

The Holistic Accountability Framework

Brown’s (2008) framework calls for the “negotiated” construction of sector and organizational legitimacy with other stakeholders through the development of accountability mechanisms. The HAF provides an operationalization of this process by combining the variables of stakeholders and mechanisms, to construct the meaning of accountability goals. This allows the identified ambiguities inherent in many accountability frameworks to be resolved by clearly distinguishing between real variables and abstract constructs, and specifying unique stakeholder-mechanism combination forces a consideration of levels of analysis.

On the “Stakeholders” column in Fig. 1, stakeholders are listed vertically. For further insight, it is useful, but not essential, to group them according to the upward/downward/lateral categories. The “Accountability Goals” column lists commonly encountered accountability outcomes. While we suggest Koppell’s (2005) typology of goals as a starting point, in practice, the NPO leader could add or delete as necessary. We suggest that adding organizational learning should be a key accountability goal, noting the judicious suggestions made by scholars on this issue (Carman and Fredericks 2010; Ebrahim 2005). In the “Strategic Conditions” column, we list the goals of value, support, and legitimacy in the “strategic triangle” postulated by Brown and Moore (2001); however, we suggest that these are not goals of accountability, but overall organizational goals and therefore we consider them strategic conditions for accountability. Nevertheless, these higher level strategic conditions could be defined by a combination of lower level accountability goals and mechanisms, or by separate combinations of mechanisms.

Fig. 1
figure 1

The holistic accountability framework: demonstrating how “Liability” can be defined for the NPO board by the combination of several legal and bureaucratic means of accountability

Full size image

The “Accountability Loci” column lists organizational and contextual groupings of mechanisms. As a starting point, we use the political/legal/hierarchical/professional groupings from Romzek and Dubnick’s (1987) framework, and the market and public reputation groupings from Grant and Keohane (2005). Although these loci may be confused with stakeholder categories, we note that they are simply a way of grouping mechanisms. For example, NPO donors may define a goal of “transparency” through the specific accountability mechanism of publishing of quarterly evaluation reports online. If the donors are national government, then this mechanism can be grouped under the “political” locus; however, this mechanism may also satisfy the professional community in which the NPO belongs and thus can simultaneously be grouped under the “professional” locus. Any particular mechanism could satisfy any number of stakeholder/locus combinations, depending on whether this is relevant for that particular grouping of stakeholders in defining what an accountability goal means.

Thus, by constructing accountability goals in dialogue with stakeholders, by the combination of various packages of mechanisms, the NPO leaders would in effect, define the meaning of the various accountability goals (i.e., transparency, liability, controllability, etc.) in the context of a particular situation. The advantage of doing so is twofold. Firstly, in the case of NPOs for which nation-state-based political and institutional theories are not well specified, accountability can only be a socially defined and contingent concept. As Weisband (2007, pp. 311–312, emphasis in original) notes:

This approach allows us to speak of accountability interpretivism….Such a method calls on those who study accountability to become consistently engaged in the development of self-reflective understandings of the doctrines, laws, or values used to justify and validate accountabilities’ schemes and practices….Accountability interpretivism would hold that accountability norms and practices exist not as political facts or practices alone, but rather as “interpretive” facts reflective of the scheme of doctrines that justifies political practices within the political community.

The HAF allows NPO leaders to understand the structural accountability requirements in the system (i.e., specific disciplinary or instrumental legal requirements), then through the process of deliberative dialogue to construct the goals and to understand the normative accountability requirements which are driven by the values of stakeholders.

The framework is denoted as “holistic” as by specifying the complex combination of accountability requirements to each stakeholder, NPO leadership is forced to understand the system relationships. This is paramount as accountability cannot be understood as a series of binary principal–agent relationships, but as a network (Slaughter 2004). Furthermore, by specifying the package of mechanisms relevant to each stakeholder, the HAF will compel NPO leadership and stakeholders to consider the various tensions between accountability requirements.

Limitations of the Approach

Although using the HAF allows NPOs to operationalize ambiguous accountability goals and to structure deliberative dialogue, some practical and theoretical concerns remain. Practically, there is a feasibility limit to the holistic aspect of the HAF: large NPOs providing a range of services across many locations would likely have many stakeholders and organizational leaders would have to conduct significant research to understand their own accountability environment. Furthermore, the actual process of negotiation and deliberation would require substantial time and effort to achieve. Perhaps this is simply a question of resources; nevertheless, we claim such a process will become ever more necessary in the future, for the reasons identified in the introduction.

Several theoretical issues require further research to address. Firstly, without some form of external coercion, the HAF cannot prevent abuse of power or ensure that the weakest stakeholders concerns are heard; however, this problem may be addressed if such an approach was voluntarily adopted by the community as best practice (Gugerty 2009). This is part of a wider issue of credibility, rights, and legitimacy: to what extent should NPOs involve the beneficiaries of their services in designing strategy and defining results? Secondly, although many scholars have investigated the various tensions between different forms of accountability, to use the HAF effectively may require these tensions to be made explicit and linked to the specific construct of the framework.

Conclusions

We have described the trend in the literature toward understanding accountability as a constructed concept that combines both instrumental, objective elements, and socially constructed, negotiated elements. Furthermore, we have pointed out the key ambiguities present in many descriptions of accountability, suggesting that the only meaningful way to operationalize ambiguous terms is through a social process of deliberative dialogue organized by the HAF. Within the boundaries of a nation-state, the HAF offers a method for NPO leadership to construct meaningful definitions of key accountability terms, through deliberative dialogue with stakeholders. In a transnational setting, the HAF provides NPO leadership with a schema to develop an understanding of accountability in an environment in which there is no geographical locus within a nation-state for donors, nonprofit ownership and control, government control and regulation, citizens, and clients. In both cases, the definitions of accountability in terms of mechanisms relevant to stakeholders are constructed with a view to the overall organization strategic goals: public value, legitimacy, support, and operational capacity.

The concept of accountability emerges naturally from the idea of democracy as a constraint on power temporarily delegated to politicians from citizens (Hanberger 2009; Peruzzotti 2006). Politicians are elected on the basis of a promise to take particular policy actions or represent interests on behalf of the citizenry, for which they are subsequently held to account for achieving (Tilly 2007). Thus, the general principle of accountability is a logical consequence of this system, even if details of its practical implementation are not.

Although scholars have successfully refuted critiques about the lack of representativeness of NPOs, arguing that the analogy between accountability in political systems and accountability in NPOs is misleading (Charnovitz 2006; Peruzzotti 2006), the critique remains that the mechanisms by which an NPO’s targeted beneficiaries can influence the founding idea, the mission, and the strategy of NPOs are still unclear. For the multitude of NPOs that provide public services or advocate certain causes, under what circumstances should the recipients of the service or beneficiaries of the cause have a voice in the foundational aims and mission of the NPO in a way that is analogous to citizens in a democratic polity? How does this conflict with the basic rights of association held by an NPO? How can the mission of an NPO be judged as legitimate? Although the particular theoretical resolution to these arguments is outside the scope of this article, we suggest that understanding accountability via the use of the HAF is a way to frame and grasp these difficult, contentious issues.

There are two research avenues derivable from the HAF. First, the HAF can be used to understand accountability challenges in practice and as a diagnostic tool for organizations. We expect that the process of defining accountability goals from the specific accountability mechanisms will reveal conflicts and tensions between the various accountability goals and between the various network arrangements of the overall system of stakeholders. The framework will thus facilitate a structured and systematic investigation into these issues. Second, the empirical question of the actual utility of the framework in practice will need answering. In this investigation, a researcher would wish to capture data on organizations using the HAF and determine the impact on strategy and operations caused by its use. A particular case of interest is the application of the HAF to transnational NPOs, which often face complex and competing accountability claims. The diversity of involved actors and the uncertain position of third sector civil society among traditional state-based democracy where organizations often originate, and the weak or corrupt states in which they operate, create challenging accountability situations (Benjamin 2009; Nelson and Dorsey 2008).