Theorizing Regulatory Intermediaries

The basic three-party model

We begin by assuming a unidirectional flow of intention and action, in which a regulator (public or private; domestic, international, or transnational) adopts regulations and chooses intermediaries to change the behavior of particular targets. The arrows in the R→T and R→I→T notations represent this unidirectional flow. The unidirectional assumption corresponds to the traditional, hierarchical view of regulation depicted in principal-agent models, where regulation is prescriptive and compliance is based primarily on deterrence and sanctions. From this baseline, however, we see that both targets and intermediaries may reverse the directional arrows in pursuit of their goals. Feedback and other forms of intermediary involvement can be highly beneficial, but they can also create significant problems for regulation, including capture.

We focus on regulation that takes the form of rules. Those rules could be “hard,” as in highly legalized domestic regulation; they could be “soft,” as in much international regulation, where legalization is more limited (Abbott and Snidal 2000). In fact, regulation at both levels is frequently a mix of hard and soft rules, in part because it is typically impossible to write a complete set of hard rules without softer supporting arrangements regarding implementation. Where nonstate actors fill the role of regulator, the rules they prescribe, often described as “standards,” are typically “voluntary,” not legally binding. They may, however, share other properties of legal rules (e.g., precision) and may be backed by significant incentives for targets to adhere and comply.

An example of a “hard” regulation is a legally binding environmental rule, adopted by a national regulatory agency, to govern the production processes of business firms. The rule may impose precise requirements such as maximum emission levels, or may entail more general requirements, such as adoption of an environmental management plan, that may require interpretation by courts. An example of “soft” regulation is an environmental performance standard promulgated by an NGO or multistakeholder organization such as MSC, which lacks authority to adopt legally binding rules but promotes voluntary adherence and compliance by targets. Governmental bodies also use soft regulation at both the domestic and international levels.

In analyzing both types of regulation, the two-party R→T model of regulation is too simple in many respects. It excludes important actors, and therefore limits our understanding of complex processes such as the capture of regulators or intermediaries. The two-party model also makes strong assumptions regarding the regulator’s capacities to formulate, implement, and enforce regulatory policies. Of particular concern are the implicit asumptions that R has direct access to T, direct means of influence that can effectively change T’s behavior, and direct means to gather information on compliance and the impact of the rules. In many circumstances, regulators lack direct access to their targets, means of influence or other capabilities necessary to regulate them, and sufficient channels for information gathering. A regulator suffering from these shortcomings will often seek to overcome them by working through a third party, which we label the “intermediary,” designated “I” in the R→I→T formulation. For example, a global banking regulator such as the Basel Committee on Banking Supervision lacks direct regulatory access to domestic banks, and so relies on regional networks of national regulatory agencies to increase the impact of its standards on these targets (Jordana, this volume).

We have generalized the RIT model from the model of orchestration (O-I-T), which depicts how an orchestrator (which may be a regulator) uses “soft” techniques to engage intermediaries that utilize their own capabilities to shape the behavior of targets. ⁴ But a similar situation often underlies principal-agent (P-A) or delegation models of regulation, where the principal delegates authority to an agent to affect the behavior of a target. In both cases, RIT presents regulation as indirect, operating through an intermediary. The P-A model is normally seen as involving a direct relationship, because it concentrates on interactions between the principal and the agent (implicitly treating the agent’s ability to manage the target as unproblematic). By contrast, the RIT model encompasses the full range of relationships between the principal and the target as mediated by the agent (i. e., P-A-T). Our focus, however, is on the role of the intermediary, with respect to both regulator and target.

Returning to our examples of regulation, hard RIT regulation might involve a national regulatory agency (such as the FDA) adopting legally binding production rules, while using external (private) intermediaries, as well as its own employees, to monitor compliance with these rules. Soft RIT regulation might involve a private standard-setting body such as the MSC or FLO using independent auditors to certify compliance by targets. The RIT model thus helps us to better understand both intragovernmental and extragovernmental relationships within regulatory regimes. In both examples, the intermediary is chosen because it has specialized expertise and other capacities that allow it to interpret the rules, supervise their implementation, and monitor compliance more effectively than can the regulator.

Intermediaries and their properties

We can define an intermediary in the basic RIT model as any actor that acts directly or indirectly in conjunction with a regulator to affect the behavior of a target. The intermediary is a go-between, whose presence necessarily makes some aspects of regulation indirect, as the intermediary stands between the regulator and its target. Its role can be formal or informal; its participation can be driven by functional or power considerations; and it can serve public or private interests. The inclusion of one or more intermediaries in a regulatory regime represents the kind of move toward complex forms of governance that are captured by the notions of regulatory governance and decentred governance (Levi-Faur 2011a; Black 2008). Note again that our interest in regulation leads us to focus only on regulatory intermediaries; we exclude a wide variety of intermediaries, such as banks, brokers, and matchmakers, which operate in nonregulated contexts. While much of our analysis may apply to such cases, we tailor it here to the specific case of regulatory intermediaries.

The R-I and I-T relationships may both be explicit or tacit. The regulator may formally engage the services of the intermediary, as FLO contracts with the independent organization FLOCERT to conduct certification audits. The target may also formally engage an intermediary, such as an external auditor, to meet a regulator’s requirements. In other cases, there may be no formal arrangement at all—as when a regulator crafts a rule in anticipation that a third-party intermediary will respond in certain ways vis-à-vis the target, but without directly engaging that party. For example, a regulator without the ability to monitor compliance directly might adopt a rule knowing that NGOs will independently monitor target behavior. Similarly, a regulator might require targets to make certain disclosures, knowing that these will empower and provoke activists to pressure the targets to comply with regulatory rules. A regulator might also adopt a rule anticipating that other organizations will incorporate it into their own standards (Marx and Wouters, this volume). Between the extremes of formal and tacit relationships, a regulator may orchestrate independent organizations, enlisting and supporting them to monitor and pressure targets to comply (De Silva, this volume).

The goals of intermediaries vary widely. Private sector intermediaries such as auditing firms or credit rating agencies need have no concern for the issue at hand, acting primarily for payment; some may share the interests of private sector targets, attempting to turn regulation to their benefit. NGOs and other civil society intermediaries may share the social values of the regulator (or even go beyond them) and act to help achieve them. Most intermediaries are organizations, so their organizational missions, interests, and cultures also shape their willingness and ability to participate as intermediaries. Some intermediaries may have strong substantive or organizational interests in shaping the regulatory rules themselves. They may join other interest groups in lobbying at the rule adoption stage; rule-makers may also seek their expertise. Here we emphasize the ability of intermediaries to provide feedback to the regulator at later stages, based on their activities and experience.

These diverse goals make particular actors more or less suitable as intermediaries in specific situations; they may also produce conflicts of interest among regulators, intermediaries, and targets that make regulation difficult. Yet the “individual” interests of potential intermediaries may be mediated or exacerbated by cooperation and competition among them, by their bargaining power vis-à-vis regulators and targets (Mattli and Seddon 2015), and by opportunities to collaborate with regulators or targets. To understand regulatory outcomes, such interactions need to be taken into account.

The principal reason for regulators to incorporate intermediaries into the regulatory process is that intermediaries possess capacities relevant to regulation that regulators themselves lack, or that intermediaries can provide more effectively or at lower cost. In this article, we discuss these capacities in terms of what intermediaries actually do; that is, the functions and roles they perform. As a prelude to that discussion, it is useful to enumerate four broad categories of capacities that intermediaries bring to regulatory governance: ⁵

These categories are neither exhaustive nor exclusive. Nonetheless, they provide a useful starting point in considering why a regulator might engage intermediaries.

Most of these capacities must be assessed relative to a specific context or situation: an intermediary might have high capacity in one setting but not in another. Capacities such as access to targets, expertise about a problem or target behavior, independence from the regulator, and legitimacy almost always depend on the specific regulatory activity at stake. Similarly, the same intermediary may have different capacities vis-à-vis different regulators, targets, or regulatory issues.

Finally, regulators can only use intermediaries if appropriate actors with the necessary capacities are available. Intermediaries are frequently preexisting rather than created anew. A regulator may, for example, search the landscape of actors and identify NGOs active on an issue; it can then activate them as intermediaries or orchestrate their participation in regulatory governance. In other cases, intermediaries may present themselves; firms with expertise in auditing and certification, such as SGS or Bureau Veritas (Galland, this volume), often actively seek to assist regulators in return for payment.

In other cases, however, no appropriate intermediary may exist. However, this does not necessarily pose an insuperable barrier to intermediation, as the regulator may be able to create suitable intermediaries for the task at hand. A national regulatory agency may create an expert body to facilitate implementation by targets; a supranational body such as the EU may create a network of national regulators to assist it in translating rules to national contexts; an international organization such as the UN Environment Programme (UNEP) may help to create a multistakeholder organization (such as the Global Reporting Initiative) to adopt and implement voluntary regulatory standards (Abbott and Snidal 2010).

Often the role of intermediary depends on acceptance and perhaps authorization or empowerment by the regulator. To be sure, some intermediaries—such as NGOs that monitor target behavior—operate independently of the regulator and guard their independence. But even for intermediaries with independent capabilities and roles, working with a regulator typically magnifies their impact. For example, credit rating agencies have access and expert authority independent of the state; they may even have independent moral authority because of their role as neutral arbiters. Public acceptance of their activities may be enough to make them effective gatekeepers in credit markets (Kraakman 1986; Partnoy 2007). Nonetheless, working with a regulator confers state endorsement and may even give their decisions the force of law (Kruck, this volume). In other cases, the regulator actually creates the intermediaries’ roles, by requiring audits by independent monitors to certify target performance, or accreditation of monitors by independent bodies (Auld and Renckens, this volume; Loconto, this volume; Lytton, this volume). The resources and authority that the regulator provides may give it some degree of control over intermediaries, and may protect it from capture by targets, but the need for intermediary independence may constrain such control (Maggetti, Ewert, and Trein, this volume).

Implementation

Although targets must ultimately implement regulations, intermediaries facilitate many aspects of implementation. First, they interpret or clarify regulatory rules, and then elaborate and adapt them for the circumstances of particular firms, industries, states, or other targets. This may entail “translating” regulations into practical forms, such as by creating guides or elaborating best practices. We hypothesize, then, that intermediaries will be more often utilized to perform this function where regulatory rules are broad or ambiguous, or where they apply to diverse sectors or other target groups. Second, intermediaries can provide training and advice on implementation, including technical assistance, and may even provide some material support, for example, for developing country firms or small to medium-size enterprises. Third, intermediaries can evaluate the costs and impacts of alternative means of implementation, helping targets and regulators to pursue the most efficient ways of achieving regulatory goals.

Implementation draws on a range of intermediary capacities. Expertise and operational capacities, including access to targets, are of paramount importance; independence may be central to certain regulatory tasks, such as evaluation.

Monitoring and enforcement

For regulation to be effective, the target’s performance must typically be monitored, through “police patrol” or “fire alarm” procedures; the rule must then be enforced if performance deviates from that which is prescribed. ⁶ Intermediaries are valuable as monitors when they possess the necessary expertise and operational capacity, including access to targets, as well as sufficient independence to ensure credibility. Intermediaries also play important roles in activities that lie between monitoring and enforcement, such as dispute resolution and rule interpretation. Intermediaries are sometimes important as enforcers for privatized forms of domestic regulation, and even more so for international and private regulation, where regulators’ enforcement authority is more limited. The actions of regulatory intermediaries—such as public disclosure of noncompliance or revoking authorization to use a product label—may generate substantial pressure to comply, effectively hardening otherwise soft regulation.

These roles are often combined in a single intermediary that monitors and evaluates target behavior and enforces rules. For example, third-party certifiers may monitor target behavior against established rules; provide guidance for how to bring behavior into compliance; and enforce the rules through positive incentives, such as the market benefits of product labels, or negative sanctions, such as adverse publicity. Indeed, their activities can go far beyond these simple carrots and sticks to include developing communities of compliance; promoting dialogue among targets, regulators, and others; and changing the organizational culture of targets. In the FLO system, for example, the intermediary FLOCERT conducts certification and verification audits of targets, interprets FLO standards, and resolves disputes related to product labeling; it also facilitates implementation through guidance and consulting and provides formal feedback to FLO and targets (Loconto, this volume).

As noted above, an important issue in this context is how the intermediary is financed, especially who pays for monitoring and enforcement. If targets pay for audits and other forms of monitoring, as they do in many private regulatory systems (including FLO) and in some privatized governmental regulation (van der Heijden, this volume), that raises questions of credibility, which can sometimes be offset by the monitor’s reputation for independence or by institutional arrangements such as accreditation. If regulators or other parties pay for such services, that may provide a point of control for them and undermine intermediary independence.

Agenda-setting and negotiation: Dynamic rule development

The basic R→I→T model assumes that the regulator alone promulgates the relevant rules (presumably after input from concerned interests), so it effectively treats those rules as exogenous to the actions of intermediaries and as static. In fact, regulations rarely remain unchanged, and feedback from intermediaries (as well as from targets, beneficiaries, and other evaluative audiences) can contribute importantly to their evolution over time (Auld and Renckens, this volume). While this dynamic view of the RIT relationship might be seen as an extension of the basic model, it is so central to the roles of regulatory intermediaries that we include it here under the headings of agenda-setting and negotiation, which are typically seen as the initial stages of the regulatory process.

Intermediaries are often important in the development of rules over time. Feedback from intermediaries may be based on their own experience with the interpretation and implementation of rules: they can provide first-hand information on areas of ambiguity, problems of maladaptation to local conditions, difficulties of implementation and burdensome demands, barriers to effective monitoring, more efficient alternatives and similar matters, and how rules might be revised, rescinded, or expanded. Where intermediaries are close to targets, they can pass on the targets’ views, suggesting how rules might be amended to lessen regulatory burdens while achieving regulatory goals; they can also help to build communities of compliance among targets and other stakeholders in society. Perhaps idealistically, intermediaries can be facilitators and mediators, enhancing trust and strengthening “regulatory dialogues” between regulators and targets, and helping them to collaborate in improving the effectiveness of regulation while lowering its costs and adverse implications.

Strong intermediary feedback may be unanticipated (and even unwanted) by the regulator, but smart regulators will anticipate, encourage, and plan for it. For example, the EU uses national agencies to help it formulate and implement its rules, and UN organizations accredit certain NGOs as observers to meetings and as ongoing partners for similar reasons. Intermediaries, though, may see this as an opportunity to shape regulatory rules to meet their own goals, and even to capture regulatory authority.

Institutionalizing the role of intermediaries

While our model of intermediation is actor-centered, intermediaries do not operate in a vacuum. Like other actors, they are embedded in institutional environments that shape their roles, opportunities for action, capabilities, and legitimacy (North 1990; DiMaggio and Powell 1991; Hall and Taylor 1996). If successful, moreover, regulatory innovations, by design or by trial and error, are institutionalized, become norms, and come to be taken for granted. Intermediation practices and architectures too can be institutionalized, changing the architecture and balance of power within a regulatory regime. Institutions diffuse intermediation practices and architectures (by emulation, learning, coercion, or incentives) and localize them for use by particular actors. In other words, the successful operation of intermediaries requires institutionalization and localization (Czarniawska-Joerges and Sevón 1996). Institutional analysis also allows for path dependency as an explanation for the structure and purposes of regulatory regimes and the power and autonomy of intermediaries within them (Streeck and Thelen 2005; Kruck, this volume).

Regulatory chains

A regulatory system need not include only a single regulator or intermediary; such systems are frequently concatenated. The degree of support from political principals—such as legislatures or ministries—influences the regulator’s need for intermediaries and its ability to control them (Maggetti, Ewert, and Trein, this volume). Among regulators, legislatures frequently delegate broad rule-making authority to administrative agencies. Courts and quasi-judicial bodies, likewise, may be granted such extensive authority to interpret broad constitutional, statutory, or treaty provisions that they become, at least in part, regulators in their own right (De Silva, this volume). ⁷ Even where an intermediary’s main role is to assist in implementation, it may be required to interpret, translate, and adapt rules to such a degree that, intentionally or not, it functions as a secondary regulator (Auld and Renckens, this volume). And meta-regulators such as the Global Food Safety Initiative set standards for private regulatory schemes (Havinga and Verbruggen, this volume).

Regulatory systems often include chains of intermediaries. An important example is the separation between auditors, building inspectors, and similar actors, which monitor target compliance, and accreditation bodies, which assess the competency of auditors. Such chains are common in public (Lytton, this volume), private (Loconto, this volume), and privatized (van der Heijden, this volume) regulation. Risking an infinite regress, accreditation bodies may themselves be required to receive accreditation from a higher authority. Often the regulator initiates this chain of relationships: R → I₁ – I₂ – I₃ → T. In other cases, an intermediary may empower subintermediaries: R → I₁ → I₂ → T (De Silva, this volume). For example, in certain cases it may be important to focus not only on the organization with overall responsibility to facilitate implementation or monitor compliance but also on how individual employees, departments, or subsidiary companies carry out those activities (Auld and Renckens, this volume; Galland, this volume; Lytton, this volume).

Many regulatory systems include chains of targets. Typically, private standards address targets at the consumer-facing end of the supply chain, such as retailers. But these firms are often required to impose those standards on their suppliers, even requiring suppliers to obtain certification from independent auditors.

Finally, in a world of legal and regulatory pluralism (Berman 2012; Scott 2001), intermediaries often operate in parallel, while on a larger scale multiple regulatory chains operate in parallel within a regulatory regime. It is not uncommon for business firms (e.g., food producers) to be the target of multiple national regulatory systems, subnational regulatory systems, industry self-regulatory systems, and private standards systems—each with its own chain of secondary regulators and intermediaries. The same organizations may even play similar intermediary roles across multiple regulatory systems, perhaps with homogenizing effects.

Multiple roles and ambiguous boundaries

The same actor may perform multiple functions within a single role in the regulatory process—as when an intermediary serves as facilitator, monitor, and enforcer of target compliance. While the RIT model treats the three main actors as clearly defined and distinct, their roles are not always clear-cut, such as when intermediaries also perform regulatory functions or are the targets of regulations (Havinga and Verbruggen, this volume). While the distinctions between actors may blur in practice, the theoretical model is useful for understanding the overall relationships.

We have already noted that intermediaries may operate as secondary regulators, when they are granted significant authority to make, interpret, and adapt rules, or when they incorporate other regulators’ rules into their own standards. Intermediaries may also overlap regulators more directly. In multistakeholder bodies such as the Forest Stewardship Council (FSC), for example, certification bodies participate as members in framing and adopting rules.

The intermediary and target may also overlap in diverse ways. For example, the intermediary may be an industry association, made up of the firms that are the targets of regulation. Such an association could be viewed as identical to the targets, or at least as “belonging” to them. ⁸ The association may, however, have somewhat different interests (e.g., acting for the industry, not for any individual firm) and develop different capacities, including some autonomy from its members (R → T* → T). In other cases the intermediary could be a subset (t) of the target group (R → t → T where t ∈ T). This approach resembles co-optation, enlisting a key element of the target population to pursue a regulatory agenda vis-à-vis the entire population (Selznick 1949). Where the intermediary is a multistakeholder organization, such as the FSC, its members will usually include targets. Some intermediaries (and regulators) follow a strategy of “embedding” targets deeply in their operations, which may facilitate learning and continuous improvement (Marx and Wouters, this volume).

In similar ways, regulators may overlap with targets. In a self-regulatory system, such as the chemical industry’s Responsible Care program, an industry association is the regulator, with its own members as its targets. Such a regulator may still employ diverse intermediaries, such as independent monitors. In a multistakeholder organization, such as the FSC, the regulator normally includes targets in its governance structure. The extent to which these overlaps influence the quality of regulation, favorably or unfavorably, depends on both formal structures and informal relationships, and must be assessed through careful contextual analysis.

A dynamic RIT model

In a three-party regulatory system, feedback flows through multiple channels. Targets provide feedback directly to regulators; intermediaries provide feedback based on their own experiences; and intermediaries receive information and advice from targets and pass those on to regulators. These connections introduce a more dynamic model of regulation, which can be represented as

OPEN IN VIEWER

with the arrows representing feedback. When the target provides feedback directly to the regulator, the sequence can be expressed R-I-T-R; when it provides feedback through an intermediary, as R-I-T-I-R. In the latter sequence, the intermediary may evaluate and filter T’s feedback, making it more useful to R while also insulating R from direct pressure by T. This need not reduce the effectiveness of T’s feedback; it may enhance it if I’s independence gives greater credence and legitimacy to T’s message. But it also provides an opportunity for I to modify or add to T’s feedback in its own interest. As the cycle of feedback and rule revision continues, these sequences become iterative and increasingly lengthy.

Feedback processes like these are essential to understanding the development of regulation over time; neither P-A nor orchestration models typically pay sufficient attention to them. Of course, including all forms of feedback over long time periods can lead to overly complicated analyses. For that reason we stick with the simplest model (RIT), bringing in such complications only when needed to gain additional insights.

Beneficiaries as intermediaries and targets

A particularly important extension of the RIT model is considering the beneficiaries of regulation within the model (Koenig-Archibugi and Macdonald, this volume). Beneficiaries are always relevant to regulation. The analytical question is whether they should be treated as active participants in the RIT process; the normative question is whether including them (formally or informally) improves the quality of regulation. The RIT model can help to clarify these relationships by bringing in beneficiaries as more than just passive actors where appropriate.

Beneficiaries figure in regulatory processes in several roles.

Beneficiaries are also significant at the dynamic stages of agenda-setting and rule reformulation. The regulator may simply consider their interests as a rationale for regulation, or they may actively advocate for regulation. Beneficiaries are thus of concern for both normative and positive reasons. When the regulator is accountable to beneficiaries (e.g., as political regulators are to voters), they can reward or punish it for its actions (R → I → T → B → R).

Classical regulatory capture

Classical capture occurs when the target dominates the regulator. A captured regulator loses its independence from the target, and is therefore diverted from pursuing regulation in the public interest. The target may also proactively persuade a regulator to adopt rules that benefit it (T → R). For example, firms often lobby regulators to limit competition—as when sports team owners seek antitrust exemptions or stores lobby for laws restricting hours of operation (Stigler 1971). Or the regulator may initially establish a rule that furthers public policy goals, but the rule creates incentives for the target to penetrate or influence the regulator to mute the regulation’s impact or, in the extreme case, to modify it to serve the target’s (not the beneficiaries’ or the public’s) interests (Bernstein 1955; Carpenter and Moss 2013).

While intermediaries might appear irrelevant in this setting, their involvement can inhibit capture. If a regulator engages third parties to act within the regulatory process, or even to act tacitly, intermediaries can provide checks on capture by monitoring the behavior of targets and regulators, by acting as whistle-blowers, and holding regulators and targets accountable (e.g., through public disclosure of misconduct). Intermediaries are most likely to play these roles when they are motivated by the regulatory goals themselves, as is frequently the case for beneficiaries.

Capture of the intermediary

In other cases, intermediaries may not provide obstacles to capture by targets, but instead the means for capture to occur. Targets often lobby for the inclusion of intermediaries to perform tasks such as interpreting the rules, assisting in implementation, or monitoring. This could be a sincere effort to increase regulatory efficiency by expanding expertise and other capacities, while reducing compliance costs. For example, if a regulator lacks sufficient monitoring capacity to accurately differentiate compliant from noncompliant firms (Ayres and Braithwaite 1992), it may be forced to apply uniform regulations to all targets. By employing intermediaries with greater monitoring capacity, in contrast, the regulator can adopt more nuanced rules, and thereby reduce costs for itself and for compliant firms.

Alternatively, targets may use intermediaries as a pathway to capture. Intermediation provides two channels for capture. First, the target may use the intermediary to capture the regulator, by influencing the intermediary to recommend changes in the rules that favor the target (R → I → T → I → R). Second, the target may capture the intermediary itself, by influencing it to interpret or implement existing rules in ways that favor the target. Reliance on an intermediary reduces the cost of capture if the target has to capture only the weaker link (R or I) in the regulatory chain. Whether capture is likely also depends on the “capture potential” of the target group, which is based on its ability to act collectively and exert power, and on design decisions that insulate intermediaries, to a greater or lesser degree, from target influence (Maggetti, Ewert, and Trein, this volume).

In addition to dominating existing intermediaries, targets frequently establish friendly intermediaries (e.g., professional or industry associations), with interests similar to their own, which creates channels for capture. These intermediaries may be sufficiently responsive to regulatory intent to avoid stricter regulation. Alternatively, targets may bring in new intermediaries with the aim of influencing regulators directly (as when industry groups hire supposedly independent, expert organizations to advise regulators).

Multiple intermediaries can facilitate capture, especially if they have conflicting interests (e.g., alternative certification bodies that compete for business, a common feature of private standards schemes), and the target need capture only one of them (e.g., if the target may decide which certification body to employ). More broadly, if different intermediaries adopt divergent interpretations of a rule, targets may claim that compliance is too costly, or even impossible. Conversely, capture becomes more difficult if it requires capturing both the regulator and the intermediary. To the extent that the regulator and intermediary monitor one another, capture may become very difficult for the target.

Finally, the ease and locus of capture may be affected by design rules. In particular, if the target pays for the intermediary’s services—such as factory audits or financial instrument ratings—then the intermediary may be overly responsive to the target’s needs, especially if targets can engage in “intermediary shopping” (van der Heijden, this volume) and threaten to choose alternative intermediaries. This risk may be attenuated by professional or reputational considerations (as for accountants), or by institutional devices such as rotation or accreditation.

Capture of regulator by the intermediary

Intermediaries too may capture a regulator. The initial demand for regulation frequently emerges from beneficiaries and other interest groups; erstwhile intermediaries often participate in these lobbying efforts. Intermediaries play similar roles in the dynamic regulatory process, providing feedback to regulators and recommending modifications in rules or policies. Intermediaries often have unusual influence in these settings: their expertise, informational advantages, and experience put them in a better position than the regulator in terms of understanding what modifications are needed. In such cases, intermediaries become the leaders in regulation, with the ostensible rule-makers following them.

Intermediaries’ goals are not necessarily noble or in the public interest. For example, accountants, auditors, credit ratings agencies, standards bodies, lawyers, and other potential intermediaries may press for regulations that would increase demand for their services and products (and oppose those that would lessen it), while couching their advice in terms of their ability to strengthen interpretation, monitoring, and enforcement. Intermediaries also press for rules that play to their organizational strengths and reduce their costs. For example, inspection firms prefer “desk audits” over “field audits,” because they are less costly and can be performed more consistently; they therefore urge regulators to adopt rules or compliance indicators that are “auditable” in terms of documents or easily observed outputs, not on-the-ground impacts (Galland, this volume). Intermediaries may also informally interpret rules to make them more amenable (Bartley 2016).

Capture of the intermediary by the regulator

A fourth type of capture is domination of an intermediary by the regulator. Such domination may be beneficial to the public interest if it prevents slippage between regulatory intent and practice due to the divergent goals of the intermediary, or if it prevents capture by powerful targets (Maggetti, Ewert, and Trein, this volume). But it may instead be antithetical to the public interest. For example, a regulator may introduce an intermediary to create the illusion of action or to distract attention from its own inadequacies. ¹⁰

Even short of that, intermediation is often more than mere agency: it is also an assurance mechanism designed to create greater trust between target and regulator. The intermediary provides assurance through means such as adherence to professional norms and commitments, or institutionalized neutrality. In these settings, regulator domination of the intermediary is problematic. First, a captured intermediary may not (be allowed to) perform the desired regulatory functions. It may be unable to follow professional norms or act in a neutral fashion; it may even resist participation in the regulatory system to avoid regulator interference or takeover. Second, without assurance or trust-building, the target may resist the regulation and avoid complying with it, and/or perceive the system to be biased against it. In short, regulator domination may undermine the very purpose of using an intermediary.