The (Im?)possibility of Horizontal Accountability
By
Mark Jarvis and Karine Levasseur[1]
A paper presented to the Canadian Association for Programs in Public Administration (CAPPA) Fourth Annual Research Conference in Public Management and Public Policy on May 25 - 26, 2015 at Glendon College, York University in Toronto, ON.
Introduction
While not a new phenomena, contemporary public administration and governance theory continues to be marked by a sense that "the preponderance of critical management and policy issues have become horizontal rather than vertical” (Bakvis and Juillet 2004, 5). That is, that the policy issues and initiatives of the day -- and modern service delivery in particular -- demand alternative governance arrangements exemplified by structures, culture and human resource practices geared towards individuals and organizations working collaboratively, across organizational divisions, whether inside or outside of government or both (see Dickinson and Sullivan 2014, for example). This "goes beyond just government"; it “involves working across boundaries within the public sector or between the public sector and private or voluntary sector... [and]... focuses attention on a set of actors that are drawn from, but also beyond, the formal institutions of government” (Stoker 1998, 93). Theoretically, at least, these arrangements represent working relationships between equals, where "partners" share authority and responsibility, in contrast to superior-subordinate relationships associated with traditional governance relationships where authority may be delegated but responsibility and authority is ultimately held by superiors (Aucoin and Jarvis 2005). The key point of differentiation here is a move towards, if not fully achieving, joint decision-making between partners (Rodal and Mulder 1993, 36-37). This amounts to more than just a consultative exercise.
Amongst the institutional arrangements said to be essential in responding to the so-called 'horizontal challenge' is the establishment of accountability arrangements that reflect this horizontal approach. As the range and number of collaborative arrangements expands, so has the belief that traditional forms of hierarchical accountability are ill-equipped to cope with these new governance practices that celebrate increasingly delegated, dispersed and collaborative decision-making (for discussion see Behn 2001; Aucoin 2003; Schillemans 2008). It is argued that these new accountability practices must recognize the limited control that any single individual or entity can muster over decisions or outcomes in the context of these decentralized governance arrangements (Considine 2002; Behn 2001). Some have even gone so for as to suggest that hierarchy is anachronism (see Considine 2002).
Rather than rely on a relic of a bygone era of democratic governance and public administration dominated by command and control approaches, the parties to these collaborative arrangements may see themselves as accountable to one another (equal-to-equal) for their decisions, use of resources and/or achievement of results. Others have provided a more nuanced perspective. Instead of viewing horizontal accountability as simply displacing hierarchical accountability, these authors have recognized complementarity between the horizontal and hierarchical accountability approaches, even as the emphasis on collaborative arrangements and horizontal accountability increases. In theory, hierarchical accountability remains whereby partnersare accountable to their superior within their home organization; but, horizontal accountability now exists whereby partners are accountable to other partners for their actions in the partnership (MacDonald and Levasseur 2014; Howard and Phillips 2012; Schillemans 2009; Levasseur and Phillips 2004).
Yet, despite the enthusiasm, horizontal accountability remains somewhat a mirage. This raises the question why, despite the increased emphasis – not to mention the now considerable volume of writing on the topic! – has horizontal accountability remained so elusive conceptually and in practice? It is not our goal to confirm or dispel debates about the complementarity of horizontal and hierarchical accountability. We do not see any disjuncture between the two. As Klijn and Koppenjan (2014, 247) argue “the fact that tensions arise between the two forms of accountability is not so much a confirmation that governance networks and their horizontal accountability mechanisms are hard to combine with representational democracy, but rather an indication that adequate arrangements to combine both horizontal and vertical accountability are lacking”. Instead, this paper argues that richer conceptualization of horizontal accountability is needed to answer key questions: Who is accountable in horizontal accountability arrangements? To whom?For what? When? How?What is the nature of the obligation? With stronger conceptualization, there may exist the possibility that more robust mechanisms to secure horizontal accountability can emerge.
Defining Accountability
As a first step towards arriving at a more precise conceptualization of horizontal accountability, it is essential to conceptualize accountability itself. Accountability is best understood narrowly defined as a mechanism. Conceptualized as a mechanism (see Bovens 2010), accountability focuses on the relationships and institutional arrangements, whether formal or informal, that allow an actor or organization to be held to account by some other individual or organization (Day and Klein 1987, 5; Romzek and Dubnick 1998, 6; Lerner and Tetlock 1999, 255; McCandless 2001, 22; Pollit 2003, 89; Mulgan 2003; Aucoin and Jarvis 2005). In this tradition, accountability is a process in which an actor is obliged to render an account to another individual or body who, in turn, is empowered to scrutinize and debate the account provided and, finally, to form a judgment of the actors behavior, decisions or performance. The actor to which the account is due, is also able to impose rewards or sanctions, as well as corrective action, based on their assessment of the account(s) as they deem required (Aucoin and Jarvis 2005, 7). As Bovens (2010, 951) notes, accountability defined in this way operates as an external constraint, rather than an internal check adopted by a particular individual (see the classic debate between Friedrich 1940 and Finer 1941).
This conceptualization results in a different focus of inquiry than competing conceptualizations that treat accountability as a virtue.[2] Accountability investigated as a mechanism shifts focus to “the way in which these institutional arrangements operate… not whether the agents have acted in an accountable way, but whether they are or can be held accountable ex post facto by accountability forums” (Bovens 2010, 948). This leads to a series of basic questions at the root of understanding accountability mechanisms: Who is accountable to whom? For what? How? And, when? The answers to these questions, of course will vary from accountability arrangement to accountability arrangement.
While this mechanistic approach to understanding and studying accountability is often associated with hierarchy, this need not be the case. Hierarchical accountability refers specifically to those accountability relationships where a superior actor delegates authority to a subordinate and the “commensurate accountability from subordinate to superior…The relationship between the subordinate and the superior—the accountable actor and the forum to whom the account is due—speaks to who is accountable to whom and for what, while allowing for variations in the specific ways that account-giving and scrutiny occurs” (Jarvis 2014,410). What sets hierarchical accountability apart is the nature of what Bovens, Schillemans and Goodwin (2014, 12) refer to as the “nature of the obligation” to account, which “largely relates to the nature of the relationship between the [accountable] actor and the forum” to which the account is due. The “superior-subordinate relationship makes the obligation to provide an account clear—the accountability forum to whom the account is due is in a place of direct and immediate authority over the accountable actors for the authority they have conferred upon them (Jarvis 2014,410).” While hierarchy is not the only type of accountability where the accountable actor is formally compelled to provide an account (e.g., some legal and regulatory arrangements compel actors to account), it is perhaps the accountability mechanism with the most obvious power to enforce accountability arrangements, at least in theory.
Conceptualizing Horizontal Accountability
In turn, horizontal accountability is often defined by what it is not: hierarchical. For example, according to Schillemans (2011, 390) “horizontal accountability then in contrast refers to forms of accountability where the accountee is not hierarchically superior to the accountor.” Similarly, horizontal accountability is defined by Aucoin and Jarvis (2005, 36) as “parties involved [in] shared authority and responsibility . . . [considered] . . . accountable to one another for the discharge of their respective responsibilities in the collective undertaking.”
Horizontal accountability is generally associated with the development of more collaborative forms of governing. This emphasis on collaboration — or co-labouring as Wanna (2011, 3) suggests — reflects a style of governing whereby the lines between the public, private, and non-profit sectors have become blurred according to Stoker (1998, 17). Collaborative governance does not “rest on recourse to the authority and sanction of government” (Stoker 1998, 17), but rather treats government as a partner working alongside partners from other levels of government and from other sectors, notably the for-profit and non-profit sectors (Wanna 2011, 3). Working as partners, the emphasis is on the development of “constructive and supportive” relationships whereby “partners negotiate structures that allow open, frank, and respectful dialogue with a commitment to work consensually wherever possible” (MacDonald and Levasseur 2014, 100; see also Ansell and Gash 2008). With the emphasis on collaborative partnerships, partners become accountable to each other for their involvement in the partnership.
We should be cognizant, however, that there are different types of collaborative arrangements that can emerge. In Wanna’s (2011, 4) estimation, collaborative arrangements can look and feel different depending on the scale of the collaborative effort and motivation to collaborate. Scale and motivation leads to an “escalating ladder of commitment — from the lowest level of perfunctory collaboration to the highestand most elaborate level of integration…similar to Arnstein’s ladderof consultation” (Wanna 2011, 4). As outlined in Table 1, the lowest level is not really collaboration per se, but consultation. At this level, one partner — normally government — retains power over the decision-making process, but engages partners in some sort of consultative processes to obtain feedback on a program, service and/or policy. Comparatively, the highest level of collaboration involves a sharing of power, cooperation and joint decision-making amongst government and non-government partners and is thought to generate transformative change. The benefit of higher forms of collaboration according to Paquet (1999) is the opportunity for organizations to gain new or insightful knowledge from other partners who have expertise or front-line experience to better respond to public policy problems. As summarized in Table 1, between these two end-points are a range of collaborative arrangements with varying degrees of consultation and meaningful collaboration.
Understanding horizontal accountability fully means understanding how, explicitly, accountability actually works. Short of rejecting a mechanistic view of accountability, the basic questions of horizontal accountability still demand an answer: Who actually accounts to whom? For what? How? When? From where do partners gain the power to compel one another to be held to account? Answers to these questions will depend, of course, on the type of collaborative arrangement in play. While specific accountability arrangements will differ from partnership arrangement to partnership arrangement, we attempt to provide a more theoretical answer to each of these questions below by applying Wanna’s continuum of collaboration arrangements. We employ Wanna’s continuum because it provides a helpful framework that can concretely speak to the differences in partnerships. By examining how partnership arrangements vary, we can hypothesize about the different implications for horizontal accountability as a result of varying forms of collaborative arrangements. In applying Wanna’s typology, we collapse the three medium levels of collaboration Wanna outlines for straightforwardness sake.[3]
Table 1: Wanna’s (2011) scale of collaboration
Degree of collaboration / What is involved – activitiesHighest level / Interaction between partners that generates transformative change with substantial engagement, empowerment and power sharing. Decisionsare made by consensus and cooperation.
Medium–high level / Significant engagement of partners in the decision-making process (both policy formulation and implementation). Decision-making is devolved to clients.
Medium level / Commitment to consultation and collaboration that may involve joint programming and joint funding strategies.
Medium–low level / Consultation with clients, but also emphasis on co-production whereby clients are directly involved to improve service delivery. Change generally leads to technical improvements in service delivery.
Lowest level / Akin to consultative exercises at the operational level where the state retains power over the decision-making process generally leading to incremental changes.
Source: Wanna 2011: 4.
‘Who’ is accountable and ‘to whom’?
The first questions are of course, who is actually accountable?andto whom are they accountable? in the context of a partnership arrangement (see table 2 below). The simplest answer here is that all partners may well consider the other partners they collaborate withto be accountable to them,whether formally or informally, for how they discharge some aspect of their respective roles and responsibilities in the arrangement, regardless of how limited the scale of the collaboration is.
The lowest levels of collaboration can be partially formalized to help frame who is accountable and to whom. Recalling that the lowest level of collaboration is really akin to a consultation exercise, there is a noticeable power differential. In consultation processes, government generally asks for feedback related to a specific issue or problem. In these instances, there is no sharing of power, nor is there the same level of risk or reward that occurs at higher levels of collaboration. So rather than speak of ‘partners’, the language used for these levels of collaboration is ‘stakeholders’. In these instances, government solicits the feedback of stakeholders and retains power to decide how the feedback is used — or not used — to develop public policy. Given itsmonopoly on power in these relationships, it is perhaps not surprising that government is, at least in theory, held to account for its role in developing and implementing consultation exercises. Some governments may devise government-wide or department-specific documents to guide the consultation process, but such documents may also indicate who within government is accountable and their roles and responsibilities.
Such documents are often established by central government agencies, but their implementation falls to line departments and as such there may still be variability in adherence to the central agency protocol. One such example is the Government of the United Kingdom’s (2012) consultation document. While lacking in legal or statutory authority, this guide is intended to shape the norms and behaviours associated with consultation exercises. In terms of ‘who is accountable’, this document explicitly holds “departments and other public bodies” accountable for the design and implementation of consultation exercises at the outset of the document: “This guidance sets out the principles that Government departments and other public bodies should adopt for engaging stakeholders when developing policy and legislation” (Government of the UK, 2012, 1).
While government is clearly accountable, it is presumably accountable to the broad range of stakeholders. In the consultation document advanced by the Government of the United Kingdom (2012) described above, departments and other government bodies are held accountable to stakeholders. Stakeholders themselves, however, may also be accountable, but in less formal ways. Even those stakeholders who are simply being consulted by government may be deemed accountable for respecting the rules under which the consultative exercise takes place or how they manage and utilize information that they become privy to through the process, especially if it is confidential.
At mid-levels of collaboration, discussions of who is accountable and to whom vary. In the instance that partners engage in joint programming or shared funding, but no involvement in public policy development, there may be formalized mechanisms vis-à-vis shared agreement to indicate which partners are accountable and to whom they are accountable. In the instance where one partner — likely government — provides contributory funding then there may not be a shared agreement per se, but a contractual or legal document to provide said funding to hold the partner to account for outputs, outcomes and plans.
At higher levels of collaboration, it is important to remember that partnerships are not simply ad hoc arrangements. Rather, partnerships at this level are durable with a deep commitment to working together over the long term (O’Flynn 2008). As such, collaboration at the this level is generally more formalized “which involves creating new structures within which to embed authority, developing a common mission, engaging in comprehensive and shared planning, and in which formal communication across multiple levels occurs. Collaboration includes pooling and jointly acquiring resources, sharing rewards, but also [sharing] increased risk” (O’Flynn 2008, 185). Working in such a joint manner, with shared risks and rewards, we might expect that as a result of these complex arrangements, accountability relationships will be formalized and specified through documents like terms of references or roles and responsibilities, indicating who is accountable to whom, and for what (which we turn our attention to in the following section). An important question to ask here is: how is accountability formalized to indicate ‘who’ is accountable and ‘to whom’? One option is the development of a signed agreement by the partners and could be considered as a framework or a compact or an accord. In such a document, which is not a legal document or contract,but a document that outlines how the partnership will work, a section may exist that outlines the partners to whom each is accountable. By way of example, the Voluntary Sector Initiative (2000-2002) was a $97 million initiative prompted by the federal government to improve its relationship with, and the capacity of, the non-profit and voluntary organizations. One of the by-products of this initiative was the development of ‘Voluntary Sector Initiative Accord’ (2001). The Accord is a shared framework agreement between the Government of Canada and the voluntary sector. It outlines the principles and values by which the partners – federal government and the voluntary sector – will build their relationship. While the Accord is a framework agreement, Susan Phillips (2001, 3) argues it much more than this:
“An accord is both a framework document and a process. As a framework, an accord cannot be understood simply as another form of a legally binding contract…but [established] on the creation of a shared vision and principles and a commitment to mutual undertakings based on responsibilities of equal partners”.