Decentering in American Governance

As conceived by British legal scholar Julia Black, “decentering” refers to regulatory strategies characterized by “complexity, fragmentation, interdependencies, ungovernability, and the rejection of a clear distinction between public and private” (Black 2002:254). The term evokes new governance forms that have emerged across the wealthy industrialized world (Carrigan and Coglianese 2011; Parker and Nielsen 2009), in which top-down state controls are being supplemented, or supplanted, by diffuse “webs of influence” (Black 2001:103)

Although coined to describe a contemporary worldwide trend, “decentered” also evokes an enduring “style of American statecraft … [characterized] by extensive delegation and formal complexity” (Mayrl and Quinn 2017:58). In this article, I use “decentered” to summarize two related features of this American way of governance. One is a pervasive policy strategy, characterized by the persistent delegation of functions to nongovernmental entities (Balogh 2015; Campbell and Morgan 2011; Clemens 2006; Hacker 2002; Mayrl and Quinn 2017; Mayrl and Quinn 2016; Quinn 2019). In Europe and other wealthy industrialized regions, decentered regulatory strategies may represent novel innovations (Black 2001; Parker and Nielsen 2009). In the United States, however, decentered strategies are not “new governance” at all, but a time-honored “antibureaucratic strategy of state-building” (Skocpol and Finegold 1983: 262).

The second dimension of decentering, which extends the term’s original meaning, refers to the sprawling structure of American state institutions. The growth of the administrative state in the U.S. was constrained by constitutional separation of powers, federalist dispersal of authority, and decentralist anti-state ideologies (Hamilton and Sutton 1989; Short 2011; Skocpol and Finegold 1982; Skowronek 1982). Born and raised in these inhospitable institutional circumstances, the U.S. federal bureaucracy is today observed to be less capable, less comprehensive, and less coherent than its counterparts abroad (Campbell and Morgan 2011; Rourke 2020; Skocpol and Finegold 1982). There is a direct relationship between central state incapacity and decentered strategies: American policymakers pursue these strategies when they are unwilling or unable to expand the size and power of federal bureaucracy (Balogh 2015; Campbell and Morgan 2011; Clemens 2006; Hacker 2002; Mayrl and Quinn 2017; Mayrl and Quinn 2016; Quinn 2019).

The concept of decentered governance provides a useful conceptual umbrella for related dimensions of American policy that were, at one time, often summarized under variations on the term “weak state.” This term has since fallen out of favor, as scholars have pointed to overwhelming evidence of the effectiveness of the American state (King and Lieberman 2017; Novak 2008). The demise of the “weak state” label has created a linguistic gap, into which political sociologists and others have inserted partial synonyms: the “hidden state” (Howard 1999), the “associational state” (Balogh 2015), the “Rube Goldberg state” (Clemens 2006), among others. “Decentered governance” can be used to fill this wider conceptual gap, and in a way that implies, not the absence of state power but rather its dispersal.

The American regulatory state exemplifies decentered governance. Its apparatus is fragmented across a patchwork of overlapping jurisdictions (Schiller 2016), and its authority is sharply constrained—by Congress, by the courts, and by arduous procedures imposed by regulatory foes (Axelrad, Kagan, and others 2000; Kagan 2007; McGarity 1991). Agencies depend for their resources on the fluctuating support of presidential administrations and Congress. This makes them perpetually vulnerable to having their capacity reduced by opponents of regulation, as occurred repeatedly during the Reagan and Trump administrations (McGarity 1986; Rein and Tran 2017).

To achieve regulatory objectives in the face of these challenging circumstances, U.S. lawmakers and regulators have both long made ample use of decentered strategies. A well-known example is the use of private lawsuits to define and enforce compliance (Farhang 2010; Melnick 2005) and exemplified by the case of equal employment opportunity in civil rights law (Dobbin 2009). Yet reliance on private litigation is only one of a menu of decentered governance strategies that are prominent in American regulation. Another common strategy is to rely on regulatory intermediaries (Abbott, Levi-Faur, and Snidal 2017): private organizations, such as audit firms, accreditors, and certifiers, that assume governance functions and that are funded, not by taxpayers, but by regulated organizations (Fransen and LeBaron 2019; Lytton 2017). Still another decentered strategy is to delegate governance functions to regulated organizations themselves, in systems of “enforced self-regulation” (Ayres and Braithwaite 1992; Coglianese and Lazer 2003). Enforced self-regulation offloads much of the cost of routine oversight to the regulated organizations and their specialized compliance offices (Bamberger and Mulligan 2015; Nelson 2021).

The Hazards of Decentered Accountability

This article is about what happens when decentered regulatory regimes impose the discipline of accountability. Accountability mechanisms can be quite effective at achieving regulatory objectives. Studies show that enforcement that holds organizations to predetermined standards can improve compliance outcomes (Kagan, Gunningham, and Thornton 2003; Short and Toffel 2010). Regulatory audits appear to be more effective than lawsuits at improving diversity in hiring (Dobbin, Schrage, and Kalev 2015; Kalev and Dobbin 2006). In this study, I do not address the efficacy of accountability for achieving regulatory objectives: the “outcome” in this particular case—ethical research—is notoriously difficult to measure (Abbott and Grady 2011).

Instead, the focus here is on the perverse side-effects of decentered accountability, which have been largely overlooked in sociological literature. An extensive literature on equal employment opportunity suggests that decentered regimes are prone to the hazard of symbolic compliance. The exemplary case is U.S. law prohibiting employment discrimination. Although the agency charged with combating discrimination was “toothless, unorganized, and broke” (Pedriana and Stryker 2004:747), civil rights law enabled private lawsuits (Farhang 2010). Rather than relying on its scant resources and authority, the agency and its civil society allies used private litigation to pressure employers to comply and to expand the definition of discrimination (Dobbin 2009; Pedriana and Stryker 2004). Judges in anti-discrimination lawsuits accepted evidence of employers’ preferred practices, such as diversity training programs, as indicators of substantive compliance. Employer defendants were therefore able to placate external authorities with symbolic compliance, or “legal structures designed to signal attention to law and thus confer legitimacy” (Edelman et al. 2011; Krieger, Best, and Edelman 2015:846). Symbolic compliance, in this account, is a buffering strategy through which organizations protect themselves from external pressures (Meyer and Rowan 1977).

Accountability, however, is designed to prevent mere symbolic compliance by penetrating and disrupting organizations’ most valued activities, or “technical core” (Meyer and Rowan 1977; Spillane, Parise, and Sherer 2011). Rather than being rewarded for following their own preferred practices, organizations are held to standards that were predefined by external authorities. Organizations must continuously narrate their internal workings to these authorities, who assess performance, administer penalties, and make demands for reform. Organizations may respond to accountability by emphasizing the production of performance indicators, rather than the goals these indicators were meant to promote (Braithwaite, Makkai, and Braithwaite 2007), which might be described as a form of “means-ends decoupling” (Bromley and Powell 2012). Yet the prioritization of means over ends is not an organizational strategy as much as it is a systemic feature of accountability, which is organized around the assessment of means rather than the achievement of ultimate goals (Power 1997).

Studies of accountability argue that this imperative has unintended consequences that are distinct from the symbolic compliance observed in new institutional studies (Espeland and Sauder 2016; Power 1996, 1997; Reich 2012; Strathern 2003). In sociology, a number of studies have focused on the consequences for individuals, such as epistemic distress (Hallett 2010), self-discipline (Sauder and Espeland 2009), and the internalization of accountability standards (Reich 2012). In this study, I am more concerned with the broader impact of accountability on organizations, and most specifically on the cost of accountability production. For accountability is never definitively achieved, but must be continuously produced: metrics must be generated, reports completed, forms completed, routines followed, audits endured, and so on. Following Heimer and Gazley (2012), I refer to such accountability production activities as “technical compliance.” Technical compliance may result in two kinds of costs to the organization: the indirect cost of disrupting valuable core activities; and the financial expense of paying for specialized accountability work.

Decentered governance regimes in the United States may be especially prone to high technical compliance costs. Comparative legal scholars suggest that, because U.S. regulators are mistrusted and constrained by Congress and the courts, they are often held to a standardizing logic, which requires them to uphold “uniform, detailed, and stringent rules” (Bardach and Kagan 2017:66). In this way, decentered policy institutions encourage a hyper-formalized version of accountability that is unsoftened by flexibility in interpretation. Standardization is passed down to regulated organizations as a meticulous “by the book” approach that punishes organizations for failing to attend closely to the minutiae of technical compliance (see Axelrad et al. 2000; Bardach and Kagan 2017; Kagan 2000, 2019). Even under systems of enforced self-regulation, in which organizations are expected to develop their own policies and procedures, organizations may in practice adhere closely to regulators’ detailed guidelines (Bardach and Kagan 2017:235).

As a result, technical compliance with U.S. regulations can be unusually costly. International corporations operating in different national environments report that U.S. regulatory paperwork is more time-consuming than that of other governments (Axelrad et al. 2000). A comparative study of nursing home regulation in the U.S., U.K., and Australia finds that American facilities devote far more resources to the production of standardized compliance documentation (Braithwaite et al. 2007). Nursing homes emphasize documentation because regulators focus on documentation is in audits and enforcement actions. Regulators, in turn, emphasize documentation because they lack the authority to make discretionary judgments of more substantive outcomes (Bardach and Kagan 2017; Kagan 2019).

The underlying logic of this regulatory style broadly parallels Porter’s (1995) accounts of “mechanical objectivity” being used to curb professional discretion. However, whereas Porter’s account emphasizes quantification—the distilling of performance into numbers to facilitate external control—some of the most burdensome forms of standardization reported in American regulation are those that cannot be quantified. More specifically, significant burdens appear to be created by procedural compliance, prescribing sequences of actions to be performed and then assessed in audits of comprehensive records. Procedural rules are experienced by the regulated organization as the obligation to continuously mass-produce auditable records, demonstrating that each procedure was followed to the letter in every relevant case. For example, workers in American nursing homes are required to record every time they turn a resident suffering from pressure sores—among a host of other everyday procedures—to create a meticulous paper trail to be presented to auditing regulators (Braithwaite et al. 2007:232). Guided by the motto, “if it’s not recorded, it didn’t happen,” nursing homes must devote large amounts of staff time and money—time and money that might otherwise be devoted to patient care—to the ongoing production of compliance documentation (Braithwaite et al. 2007:54).

Background and Research Methods

Passed by the U.S. Congress in 1974, the National Research Act authorized regulations to prevent ethical abuses in research. The law was a response to reports of the horrifying biomedical research scandals of the 1970s, the best known of which was the Tuskegee syphilis study. Importantly, the law did not contain a clause enabling private litigation, and lawsuits would not play a role in its evolving interpretation. The regulations required that federally funded research involving human subjects be reviewed by local committees, which became known as Institutional Review Boards (IRBs) (Frankel 1976; Stark 2012).

The IRB system was delegated and diffuse, exemplifying decentered policy structures and strategies. Local boards were given final authority over the ethics of research studies: there was no higher body to establish precedents or to receive appeals. Liberal lawmakers had tried to authorize a more centralized system, but were thwarted by the combined opposition of the NIH, conservatives, and the biomedical research community (Frankel 1976; Halpern 2008). Federal authority over the system was fragmented (see Figure 1). The original agency overseeing IRBs was the Office for Protection from Research Risks (OPRR), charged with regulating research funded by NIH, and later—with the promulgation of revised regulations known as the “Common Rule”—the research of other federal agencies that signed on to the Rule. In 2000, this agency was reorganized, relocated, and renamed the Office for Human Research Protections (OHRP). In addition to OHRP, the U.S. Food and Drug Administration (FDA) oversaw a similar (but not identical) set of regulations for privately-sponsored studies, under the jurisdiction of three separate offices dealing, respectively, with research on drugs, devices, and biologics (United States General Accounting Office 1996).

Figure 1

U.S. Regulatory Framework for Protecting Human Research Subjects, 2001.

Adapted from National Bioethics Advisory Commission (2001).

The agency charged with overseeing the Common Rule was small and underfunded (Greenberg 2007:133–34). It lacked authority to set official precedents and had no say over ethical decisions, which were entirely left to local boards. Instead, the agency’s main function was to hold research institutions accountable for compliance with prescriptive, mostly procedural, rules—especially regarding how IRB decisions were to be made, and by whom. Following a logic of enforced self-regulation, IRBs were also required to develop and follow their own local policies and procedures. To allow regulators to assess compliance with both federal and local rules, IRBs were expected to maintain comprehensive documentation of their activities (McCarthy 2001).

Until the second half of the 1990s, however, these accountability structures had little impact. Under the leadership of a director who favored an educational over a punitive strategy, the agency conducted little enforcement (McCarthy 2001). IRBs’ technical compliance obligations, although stated in the regulations, were easily overlooked by boards run by faculty volunteers, more concerned with research ethics than with following the letter of the law (Babb 2020). The impact of federal regulations would not be felt until the mid-1990s, when a fresh outbreak of biomedical research scandals triggered an unprecedented wave of federal enforcement.

This article analyzes how biomedical research institutions responded to this wave of enforcement, which brought the discipline of accountability into the highly decentered IRB system. The period of analysis begins in the late 1990s and concludes in 2018, the year in which new federal IRB regulations were published. The mode of analysis is the “theory-building” variant of “process-tracing”—using the structured analysis of a single case to build a plausible argument for more general causal processes (Bennett and Checkel 2015).

The analysis presented here draws on a larger qualitative study based on both interview data and a wide variety of documentary sources (Babb 2020). Among the documentary sources are articles that appeared in the trade journal IRB Advisor (“your practical guide to institutional review board management”) between 2001 and 2018. To understand how individuals experienced and acted upon large-scale social forces, I also conducted qualitative interviews with informants working in and around the IRB world in the 1990s and early 2000s at research institutions across the country.¹ I refer to my informants by pseudonym. Both documents and interviews were coded and analyzed inductively with the assistance of qualitative data analysis software.

It is important to acknowledge, at the outset, two limitations of this case study. First, the focus here is on the regulation of IRBs at research institutions: universities and academic medical centers, conducting studies mostly (although by no means exclusively), under the jurisdiction of the agency known since 2000 as the Office for Human Research Protections (OHRP). There exists a separate ecosystem of organizations providing research services to private biopharmaceutical companies, answering almost exclusively to Food and Drug Administration (FDA) regulations. These have evolved along a different trajectory (see Fisher 2008; Mirowski and Van Horn 2005), and are touched on only briefly in this article.

Second, my analysis focuses on the effects of accountability on biomedical institutions and researchers. It was primarily to address biomedical misdeeds that the first IRB regulations were adopted, and biomedical developments have primarily driven the system’s subsequent evolution. Readers of this article may be most familiar with the system’s considerable collateral impact on social and humanities research, which was swept along as regulators responded to biomedical incidents (Babb, Birk, and Carfagna 2017; Bledsoe et al. 2007; Katz 2007; Schrag 2010). Because this subplot in the IRB saga is complex in its own right, and only tangential to my argument, I do not narrate it here.

Findings

In 1996, a college student named Nicole Wan died in a study on the effects of smoking and air pollution. The young volunteer had received a lethal dose of lidocaine and it was soon discovered that the study’s IRB protocol had failed to specify the maximum dosage (IRB Advisor 2002; Rosenthal 1996). This shocking death was just one in a series of biomedical research scandals that unfolded over the course of the 1990s (Hilts 1994; Shalala 2000; Stolberg 2000).

Under pressure from lawmakers and the public, regulators launched an attention-grabbing enforcement campaign, using tools well-suited to their limited resources. Rather than making expensive site visits, regulators almost always conducted investigations remotely, through an assessment of research institutions’ documentation (Borror et al. 2003).² The lead agency in the crackdown was OHRP, which, although small and underfunded, had the power to suspend research institutions’ federally funded research (Koski 2003).

The following two sections trace the unfolding consequences of federal discipline for research institutions. The first section examines how federal discipline triggered the infamous “blizzard of paperwork”—a surge in IRBs’ technical compliance activities—and shows how the refraction of accountability through decentered institutions amplified these burdens. The second section shows how escalating accountability costs led research institutions to reorganize the labor of technical compliance for efficiency—most notably, by outsourcing ethics reviews to for-profit IRBs.

Discussion

This article has explored the origins of the blizzard of paperwork for which IRBs became famous around the beginning of the 21st century. By engaging in widespread enforcement, regulators unwittingly triggered the accountability mechanism that had long lain dormant in the heart of the regulations, guided by the aphorism, “if it wasn’t documented, it didn’t happen.” For the first time, biomedical research institutions were required to adhere to the letter of the rules and to provide meticulous documentation that they had done so. The findings of this study resonate with legal studies that draw a connection between limited U.S. federal bureaucratic capacity, by-the-book regulation, and heavy accountability burdens (Axelrad et al. 2000; Bardach and Kagan 2017; Kagan 2019). This study has also revealed two additional features of decentered American governance that may exacerbate technical compliance burdens: ambiguous mandates, which can encourage overcompliance; and fragmented authority, which can refract accountability duties, rendering them more complex, time-consuming, and demanding of expert attention.

This is not to say that the American IRB system is entirely different from its counterparts abroad. Across the wealthy industrialized world, governments delegate decisions about human research ethics to committees, typically containing a mixture of expert and lay members (Druml et al. 2009; Fitzgerald and Phillips 2006; Hedgecoe 2020; Hedgecoe et al. 2006; Shelley-Egan 2016, 2016). It is difficult to imagine a system in which the ethics of clinical trials were adjudicated by central government bureaucrats. In addition to being impractical—given the size of the biomedical research enterprise—a top-down system of human research ethics would surely lack the necessary level of specialized biomedical knowledge. Sometimes decentered governance is the most effective policy strategy (Black 2001).

In other ways, however, the American system stands out in sharp contrast. Particularly unusual is its heavy reliance on for-profit ethics review. Nothing comparable exists within the European Union (European Network of Research Ethics Committees 2022). Although Canada allows review by U.S.-based for-profit boards (and has given rise to one of its own), the practice has been far more limited and appears to be a concession to the practical realities of collaborating with American sponsors (Goldstein 2018).

There are at least two likely reasons for the rarity of commercialized human research review around the world. One is that it is considered ethically problematic. A second is that in most other national systems, more capable state institutions impose lighter accountability burdens, obviating the need for such privatized solutions. In support of this point, it is revealing to compare the American IRB system and its counterpart in the UK, made up of Research Ethics Committees (RECs). Like their American counterparts, British RECs are composed mostly of research experts, operating under the framework of government regulations (Hedgecoe 2020; Hedgecoe et al. 2006). Yet in the UK, these boards are administered not by local institutions, but by the National Health Service (NHS), which is solely responsible for regulating clinical trials (Allen 2016).

The centralization of the British system is a relatively recent development. Until the 1990s, British RECs were, like American IRBs, entirely reliant on the decisions of local authorities. Like their American counterparts, British researchers complained about the burdens and delays resulting from disparate decisions in multi-site studies (Hedgecoe 2020; Hedgecoe et al. 2006). In contrast to the U.S., however, authorities in the UK were able to consolidate REC governance under the umbrella of the strong, pre-existing NHS bureaucracy (Hedgecoe et al. 2006; Salman et al. 2014).

There is evidence that centralization has lessened technical compliance burdens. Since the REC system was reformed, biomedical investigators in the UK have reported a significant decrease in red tape (Hedgecoe 2020:193–96). Today, rather than requiring each board to devise its own policies, the NHS publishes and annually updates a single set of detailed standard operating procedures (United Kingdom. Health Research Authority 2022b). Because they are administered by NHS, UK committees have direct access to authoritative advice about the meaning of government regulations, eliminating the incentive to engage in overcompliance. Multi-site studies are not outsourced, but go through a single NHS online portal, where they are assigned to be reviewed by a single committee (United Kingdom. Health Research Authority 2022a, 2022b). There are no “commercial RECs” specializing in cutting through red tape and delays: none are needed in a system in which centralized ethics review is provided by a government agency.

Whereas for-profit ethics review is rare around the world, commercialized compliance services appear to be quite common in the United States, as is revealed by a quick internet search using the words “compliance vendor.” Research on this fascinating phenomenon is thin and scattered. Based on this current study, however, we might hypothesize that for-profit compliance service providers are especially likely to thrive in areas of regulation where decentered accountability has generated unsustainably high technical compliance costs.

An especially compelling example, and one paralleling the IRB story, can be found in the the world of U.S. financial services regulation. Both Sarbanes-Oxley (2002) and the subsequent Dodd-Frank Act (2010) caused a ballooning in financial firms’ technical compliance obligations, and a dramatic expansion of compliance staffing. By 2019, Citi reported having around 30,000 risk, regulatory, and compliance staff, and JP Morgan around 43,000 (English and Hammond 2019; The Economist 2019). By 2017, the cost of compliance was estimated at $270 billion per year, industry-wide, or 10 percent of operating resources, most of it spent on maintaining large cadres of compliance administrators (Farley 2017). These high costs induced financial firms to meet regulators’ demands more efficiently. Today, all depend on powerful software known as “regtech” (short for “regulatory technology”), which assumes the technical compliance functions that would otherwise be accomplished, albeit more slowly and imperfectly, by salaried employees (Bamberger 2010; Packin 2019). Many financial firms also rely on outsourcing to “compliance management vendors”—the equivalent of independent IRBs (English and Hammond 2019; Hammond and Cowan 2020).

Conclusion

This study has proposed the concept of “decentering” as a way of summarizing a well-known “style of American statecraft” (Mayrl and Quinn 2017:58). By working around fractured and limited bureaucratic capacity, and by mobilizing private actors and resources, the apparently weak American state can have a powerful impact (Balogh 2015; Campbell and Morgan 2011; Clemens 2006; Farhang 2010; Melnick 2005; Novak 2008; Quinn 2019). In this article, we have seen how a tiny federal agency with sharply limited authority was able to induce thousands of biomedical research institutions to overcomply with federal rules and to invest millions (if not billions) of dollars in administering their IRBs.

At the same time, this article underscores the ongoing relevance of state capacity for explaining policy outcomes. Decentered regulatory strategies that work around central state incapacity may achieve mandates at the cost of creating extensive perverse side-effects. One hazard, described in new institutional studies of employment law, is symbolic compliance (Edelman 1992; Edelman et al. 2011, 1999; Kalev and Dobbin 2006; Kalev, Dobbin, and Kelly 2006; Krieger et al. 2015) This article has addressed other hazards that arise when decentered regimes impose the discipline of accountability—relying not on private litigation for enforcement, but on rigid, prescriptive rules assessed in meticulous audits and inspections.

I have argued that decentered accountability elicits not symbolic compliance, but an obsessive focus on technical compliance, experienced as a blizzard of paperwork. I have also identified a long-run hazard of decentered accountability that merits future study: namely, reorganization of technical compliance to lower costs, and the consequent nurturing of commercial industries peddling compliance with efficiency, exemplified by independent IRBs. Such firms help regulated organizations to manage high technical compliance costs, but in ways that potentially undermine regulatory mandates and create further layers of complexity.

An obvious limitation of this article is that it is based on the case of a single regulatory regime. Within the universe of American regulation, there appear to be many parallel cases of decentered accountability, where compliance is defined by inspections and audits by fragmented regulatory authorities with limited capacity. Examples include financial services (described in the discussion above), nursing homes (Braithwaite et al. 2007), hospital patient safety (van de Ruit and Bosk 2021), and food safety (Lytton 2017). There is a need for more comparative research to identify which of the adaptations to decentered-accountable structures seen in the IRB story are universal, and which are more variable. There is also an enormous gap in our knowledge of the role and impact of for-profit compliance industries.

Moreover, very recent legal transformations will soon provide an unprecedented opportunity for researchers to study the consequences of diminished federal capacity in American regulatory governance. In the summer of 2024, the US Supreme Court struck down the Chevron doctrine, making it easier for businesses to challenge regulations in court, and even more difficult for agencies to exercise discretion in interpreting regulatory mandates. This attack on the American regulatory state, it is predicted, will embolden corporations to test the limits of regulators’ diminished powers (The Economist 2024).

This study, however, suggests a second possibility: that by further eroding agency capacity, the demise of the Chevron doctrine may inadvertently amplify the hazards of decentered accountability. It may induce agencies to rely more heavily on an inflexible, by-the-book style of enforcement that protects them from legal challenges; rely more on cryptic, informal guidance (Kalen 2008); and perhaps rely even more on private regulatory intermediaries, whose voluntary standards cannot be challenged in court. Moreover, by weakening federal regulatory authority, the decision may add more layers of complexity to already fragmented regulatory environments (Gaskin 2024). Paradoxically, an anti-regulationist Supreme Court decision may end up increasing regulatory burdens with more procedures, more record-keeping, more specialized work for compliance offices, and more business for third-party compliance vendors: in short, a new chapter in the ongoing saga of decentered accountability in American governance.

Competing Interests

The authors declare that they have no competing interests.