Abstract
The multiple principal problem refers to multiple collective action problems that organizations face when they must balance (competing) interests of multiple stakeholders under joint service delivery. It negatively affects different types of organization, yet we know little about how organizations (can) mitigate it. We expand a framework based on principal–agent theory, review the literature, and consider implications for effective governance of joint service delivery in the public sector. We observe that joint service delivery can lead to free‐riding and duplication in monitoring, lobbying by principals, and increased autonomy for agents, leading to inefficiency. We build a research agenda and tentatively suggest, based on the literature, that an interface approach, where an elected unitary actor is placed in a middle tier between politics and service delivery, might best mitigate the multiple principal problem, which is currently not dealt with effectively in public management.
1 INTRODUCTION
Is joint service delivery beneficial for public sector organizations? We know from the literature on inter‐municipal cooperation that knowledge sharing and scale economies frequently make joint service delivery efficient (Hulst et al. 2009; Bel et al. 2014; Bel and Warner 2015; Silvestre et al. 2017; Voorn et al. 2017). However, joint service delivery is also prone to failure, due to multiple possible collective action problems (Feiock 2009, 2013). For inter‐municipal service delivery, Garrone et al. (2013) find that conflict over objectives can trickle down to insufficient or incoherent directives to boards; Sørensen (2007) finds insufficient monitoring due to free‐riding. We observe similar inefficiencies for parliaments and agencies (McCubbins et al. 1987; Moe 1987; Wood and Waterman 1991; Hammond and Knott 1996; Koppell 2005; Dehousse 2008; Schillemans and Bovens 2015), public–private partnerships (Bognetti and Robotti 2007; Da Cruz and Marques 2012), and private firms with dispersed ownership (Young et al. 2002; Su et al. 2007; Ward and Filatotchev 2010; Lin and Chuang 2011).
Theory has focused on understanding these collective action problems in joint service delivery separately. Models of Gailmard (2009), Khalil et al. (2007), and Varian (1989) establish monitoring complexities; Dixit (2002), Martimort (1992), and Stole (1997) demonstrate an incentive‐setting problem; and Hammond and Knott (1996), McCubbins et al. (1987), and Schillemans and Bovens (2015) show difficulties of steering and accountability. Yet, the literature remains dispersed, policy recommendations for joint service delivery remain complex and incoherent, and predictive powers of models of joint service delivery remain weak. Some authors have begun collocating the collective action problems inherent to joint service delivery under a single term,1 hoping to find coherence to this multiple principal problem (e.g., Waterman and Meier 1998; Miller 2005; Dehousse 2008).
In this article, we expand a framework grounded in principal–agent theory to understand the multiple principal problem and build a research agenda for governance. In section 2, we describe our study’s background and explicate our approach. In section 3, we derive the framework and form hypotheses. In section 4, we review and synthesize the extant literature on the multiple principal problem, test the framework, and highlight implications for joint service delivery. In section 5, we consider typical organizational responses to the multiple principal problem, and highlight problems these responses face. In section 6, we suggest a solution to the multiple principal problem, based on an interface approach and grounded in median voter theorem. In section 7, we summarize, emphasize limitations, offer potential avenues for research, and conclude.
2 BACKGROUND AND APPROACH
Joint service delivery is widespread, for good reasons. When organizations cooperate, they can engage in risk sharing, specialize, and capture scale economies (Bel and Warner 2015; Silvestre et al. 2017). Accordingly, we observe that many firms and foundations have multiple shareholders (e.g., Maury and Pajuste 2005), that governments and agencies, local councils, and political parties frequently cooperate (e.g., Bouckaert et al. 2016), that public–private partnerships are used more frequently (Hodge and Greve 2007), and that cooperation between the public sector and the non‐profit sector is rising (Warner and Hebdon 2001). Such cooperation is called collaborative or joint service delivery when it exists for the delivery of (public) services (Hulst et al. 2009; Hilvert and Swindell 2013); alternative terms, such as ‘joined‐up government’ (Hood 2005; Silvestre et al. 2017) and ‘collaborative governance’ (Ansell and Gash 2008), are also used.
Empirical research demonstrates that joint service delivery is often beneficial, albeit not problem‐free (Bel and Warner 2015). Problems occur because cooperation requires (partial) externalization of decision‐making (Argento et al. 2009), which brings complexities. Collective action problems occur when the interests of individual stakeholders conflict with the collective interest of all stakeholders in the collaboration, leading to sub‐optimal outcomes. Stakeholders face transaction costs to mitigate those problems, including regulation and monitoring costs. Collective action problems and transaction costs together affect joint service delivery through ex‐post inefficiency, where collaboration is less efficient than it could have been, or ex‐ante inefficiency, when otherwise profitable cooperation does not occur. We know less about equity effects: how inefficiencies affect some stakeholders more than others (recent work by Cäker and Nyland 2017 and Spicer 2017 are welcome exceptions).
Many academics have focused on individual collective action problems and solutions to them (McCubbins et al. 1987; Varian 1989; Martimort 1992; Hammond and Knott 1996; Stole 1997; Dixit 2002; Khalil et al. 2007; Gailmard 2009; Schillemans and Bovens 2015). While this endeavour has been fruitful, individual treatment of these collective action problems has brought weak predictive powers for models of joint service delivery. Building a framework to predict outcomes under multiple principals is crucial for understanding joint service delivery (Dehousse 2008) and is this article’s ultimate aim.
We illustrate our theoretical case with the practical example of inter‐municipal service delivery to demonstrate how our framework emulates actual practice. Inter‐municipal service delivery reflects benefits and problems of joint service delivery, as current research shows that it is frequently quite successful (Dijkgraaf and Gradus 2007; Bel and Warner 2015; Teles and Swianiewicz 2018), while many problems can be linked to collective action problems (Bognetti and Robotti 2007; Sørensen 2007; Garrone et al. 2013; Bel and Warner 2015). We will refer to this practical example throughout the text. We emphasize that the framework applies to joint service delivery overall, and not just in the context of our illustration; to show that our framework is applicable in a variety of settings, we incorporate a literature review that explores multiple types of service organizations.
3 FRAMEWORK AND HYPOTHESES
3.1 Framework
Consider a municipality that operates a municipally owned corporation and hires an expert director to manage it, as frequently happens in Europe (Voorn, Van Thiel et al. 2018). The resulting relationship between the municipality and director is a relationship between principal and agent, where the agent acts on behalf of the principal. Principal–agent theory emphasizes the benefits of such relationships, which allow more expert service delivery. However, it also emphasizes problems of asymmetric information and moral hazard (Fama and Jensen 1983): the director may have objectives different from the municipality, and the municipality cannot take the director’s commitment for granted. Therefore, the municipality needs to introduce governance to align the director’s interest with its own.
Under principal–agent theory, steering and monitoring are key governance mechanisms that can encourage agents to act in principals’ interest (Fama and Jensen 1983). For our hypothetical director‐agent, steering protocols help: clear municipal directives build awareness of expectations, and give the municipality criteria to audit. Similarly, incentives such as variable pay can align the director’s interest with that of the municipality. Monitoring the director is also practical, presuming that it is not too costly. These steering and monitoring mechanisms increase the agent’s accountability.
Typically, principal–agent theorists consider the dyadic case of one principal, one agent, and one task. This is a simplification of reality. In organizations, relationships typically involve multiple actors. Consider local refuse collection: while some large cities singly own refuse collection firms, many refuse collection organizations in Europe are owned inter‐municipally, that is, by multiple municipalities, or by public–private partnerships (Dijkgraaf and Gradus 2007; Sørensen 2007; Da Cruz and Marques 2012; Garrone et al. 2013; Bel et al. 2014; Blåka 2017; Giacomini et al. 2018; Soukopová and Vaceková 2018; Voorn, Van Genugten et al. 2018; Gradus and Budding 2018). In the United States, inter‐municipal contracting is more frequent (Warner and Hebdon 2001; Marvel and Marvel 2008; Hefetz et al. 2012; Bel and Warner 2015). For other types of joint service delivery, such as firms with dispersed ownership, parliaments and agencies, the dyadic model is similarly inaccurate.
Making the dyadic model more inaccurate is the observation that principals’ interests frequently diverge. In refuse collection organizations owned by multiple municipalities, interests between municipalities diverge in pick‐up frequency, pick‐up times, attitude towards recycling, pricing strategies, and others (Van Genugten 2008). In inter‐municipal theatres, interests between municipalities diverge in the prime targeted demographic, focus on local or national culture, types of cultural programmes offered, or focus on tourism. Parliaments, agencies, public–private partnerships, and private firms with dispersed ownership face similar diverging interests among stakeholders. Crucially, since there is asymmetric information between the principals, where principals are not necessarily aware of each other’s behaviour, this creates a collective action problem for governance.
Once multiple principals are introduced, we can no longer presume that it is in individual principals’ best interest to steer or monitor the agent. Worse yet, individual principals can lobby the agent to pursue their interests in lieu of those of other
Once multiple principals are introduced, we can no longer presume that it is in individual principals’ best interest to steer or monitor the agent. Worse yet, individual principals can lobby the agent to pursue their interests in lieu of those of other principals.
.
