Chapter 5 Conclusion

The legitimacy of bureaucratic policymaking is said to depend on the premise that rulemaking provides for public voice Rosenbloom (2003). Yet, to date, normative debates about the value of public comment processes have had little systematic empirical knowledge to draw upon. If input solicited from ordinary people has little effect on policy outcomes, directly or indirectly, it may be best understood as providing a veneer of democratic legitimacy on an essentially technocratic or elite-driven process. Worse, if public pressure campaigns are largely astroturf—misleading impressions of public support sponsored by narrow private interests—it could further tilt the scales of the policy process toward concentrated economic power.

In this chapter, I first discuss the implications of my theory and findings for the study of bureaucratic policymaking. Section 5.2 then suggests some avenues for future research. Section 5.3.3 concludes with implications for current debates over how to think about the value of public participation and what those values—combined with the findings of this dissertation—imply for policy reforms.

5.1 Findings

This dissertation has made several steps toward better understanding public engagement in bureaucratic policymaking.

Chapter 2 set out a theory of why groups participate and offer the most comprehensive assessment to date of participation in bureaucratic policymaking. In contrast to much of the debate over the value that the lay public may or may not provide as individuals, I argue that engaging in federal policymaking is almost exclusively a collective affair. Supporting this argument, I find that most civic participation in bureaucratic policymaking results from pressure campaigns from advocacy groups. Given the barriers to even knowing about, much less engaging in bureaucratic policymaking (Kerwin and Furlong 2011), it is unsurprising that people who participate are almost rarely alone.

My theory and evidence for mobilized public engagement have implications for scholarship. First, we learn little by comparing form letters to comments written by lawyers (Seifter 2016, pg. 1313); instead, we must study these as two tactics used by the same organizations and lobbying coalitions. Second, public participation is mediated through organizations; studying civic participation means studying the organizations that mobilize people. While I focus on public pressure campaigns targeting bureaucratic policymaking, many of the theories and methods that I advance should be well-suited to study other situations in which organizations mobilize public pressure.

Public pressure campaigns targeting agency rulemaking in the U.S. are overwhelmingly sponsored by public interest groups. Compared to business groups, I theorize that public interest groups more often have incentives to expand the scope of the conflict, and I show that they more often do. However, mobilizing is also a tactic that requires resources, leading the vast majority of mobilizing to be led by a few national advocacy organizations. The resources required to lobbying in national policy means that a small set of large organizations often mobilize other organizations with less capacity for policy advocacy (as documented by Nelson and Yackee (2012), English (2019b), and M. A. Dwidar (2021)). I show that these same large advocacy organizations mobilize almost all public participation. The implication is that lobbying coalitions are the proper unit of analysis for most studies of public participation or interest group influence in bureaucratic policymaking—not individuals or even organizations. While studying lobbying coalitions is more difficult than studying individual organizations, this dissertation has developed theories and methods to systematically study the composition and influence of lobbying coalitions across institutions and over time.

Chapter 3 builds on my theory of why groups opt to mobilize public pressure to include both direct and indirect pathways of policy influence. Both pathways are contingent on a number of institutional features related to how bureaucratic institutions process incoming information (more on this below). It is difficult to assess the direct relationship between public pressure and lobbying success because the policies that groups target are correlated with their likelihood of success. My theory of why groups mobilize, set out in Chapter 2, predicts that private interest groups are more likely to mobilize public pressure only when it is likely to help their cause. And indeed, the results presented in 3.4 show that pressure campaigns sponsored by private interest groups have a higher correlation with lobbying success. This could be evidence that public pressure campaigns sponsored by narrow private interests are more influential—perhaps because they are more surprising, causing agency officials to update their beliefs about public opinion. However, the confounded and endogenous relationship between who chooses to launch a pressure campaign and who wins makes it difficult to draw firm conclusions about the direct effects of public pressure campaigns from these data.

My analysis in Chapter 3 suggests that public pressure campaigns are most likely to influence bureaucratic policymaking indirectly by attracting the attention of members of Congress. The scale of public attention and pressure campaigns, in particular, are highly correlated with involvement from members of Congress, and coalitions with more members of Congress are more likely to win. While both pressure campaigns and legislator engagement may be driven by the salience of the policy process, increasing the salience of the policy process is often the main goal and a plausible effect of pressure campaigns (more on this in Appendix C). Thus, while members of Congress may also be driven to engage in rulemaking by related factors like media coverage, public pressure campaigns are almost certainly very intentionally pushing those other levers of power while simultaneously mobilizing public participation in the policy process. If we view mass comment campaigns as a proxy for a coalition’s broader lobbying efforts to raise awareness and expand the scope of the conflict, we can more confidently say that pressure campaigns likely affect the behavior of members of Congress.

Chapter 4 investigates another potential mechanism of direct influence: claims about distributive justice. My analysis suggests that both pressure campaigns and minority activism can affect the discourse of policymaking, including the preambles to policy documents where agencies justify their proposed and final policies. When commenters raise distributive justice concerns, agencies are much more likely to address these concerns. The more people who raise such concerns, the more likely it is that agencies address them. Raising the overall level of public attention can have a similar effect as raising specific concerns. Agencies are more likely to add a discussion of how policies address environmental justice when comments raise justice concerns or they receive large numbers of comments. This may be related to the fact that courts are more likely to uphold policies with high levels of participation (Judge-Lord 2016). In contrast, when an agency already addressed environmental justice in their proposed rule, rules with more comments were less likely to change, supporting the hypotheses that more salient rules may be harder to change because the agency has anticipated public scrutiny. Their position stated in the draft is more likely to be the position of the final rule.

If I had found that most public comments were the result of astroturf campaigns, the implications would be clear: public comment periods would be merely a veneer for economic power. Conversely, if I had found that public comments were mobilized by a nationally representative set of organizations and that they were highly effective in reducing the bias toward businesses in rulemaking, the implications would also be clear: public comment periods would democratize policy outcomes and should be expanded.

Instead, my findings are mixed. Public pressure campaigns do systematically counter the narrow private interests that usually dominate bureaucratic policymaking, but it is not clear how effective they are. Specifically, it is not clear that they are more effective than pursuing representation more indirectly through the legislative branch. The number of allies in Congress that a coalition mobilizes is the only clear predictor of lobbying success that lobbying coalitions have any control over.

The few organizations that dominate public pressure mobilizing are—for the most part—not the “usual suspects” (few of them lobby as frequently as the Chamber of Commerce or the American Petroleum Institute). However, the breadth of the constituency represented by those organizations is unclear. On the one hand, they regularly mobilize millions of genuine supporters. On the other hand, the interests are concentrated on certain policy issues (especially environmental issues) and not others (for example, transportation safety). Even within the policy areas they operate, it is not clear how well they represent the people for whom they claim to speak.

These findings raise as many questions as they answer. The next section briefly discusses a few of the many opportunities for future research to build on my analysis.

5.2 Future research

5.2.1 Interest Group Representation

To the extent that public pressure campaigns shape agency decisions, a broader research program is needed to investigate who exactly these campaigns mobilize and represent. My analysis of second-order representation in comments raising environmental justice concerns shows that the large national advocacy groups that dominate public pressure organizing often claim to represent groups that tend to be excluded from policymaking. Sometimes this is done in collaboration with disenfranchised groups, and sometimes it is not. Much more research is needed to assess the quality of second-order representation in coalition lobbying.

Unlike classic pluralist assumptions about how interest group representation operates—where an organization represents a defined class or membership—pressure campaigns mobilize an open-ended list of potential supporters that theoretically includes most of the public. Indeed, a key piece of political information that campaigns generate is signals about the potential for a movement to grow and further expand the scope of conflict. For example, a large public pressure campaign may signal impending letter-writing campaigns targeting members of Congress. Pressure campaigns are also associated with protests and increased media attention. Expanding the scope of conflict often comes with a threat to further expand the scope of conflict. Policymakers know that their policy decisions shape the political conflicts in which they are embedded.

5.2.2 Networks and Lobbying Coalitions

In Chapter 2, I show that most public pressure campaigns are coordinated by a relatively small number of organizations that repeatedly lobby both with and without pressure campaigns. Lobbying coalitions range from two organizations to hundreds. These data are flush with opportunities for network analysis using organizations as nodes and coalitions as edges. Simple measures like node centrality may tell us a great deal about the structure of advocacy coalitions and how they change over time and across policy areas. Because every comment is linked to a specific policy and has a timestamp, there may be opportunities to study the evolution of networks over time. For example, do small, local, or specialized groups comment first and then recruit more general national advocacy group allies to amplify their message? Or do large national advocacy organizations more often recruit smaller groups to grow their coalition’s size and diversity? One could observe trends over time both within a given comment period on one proposed rule and over decades across rules.

Anecdotes from my data pose questions for network research. For example, large environmental organizations increasingly use environmental justice rhetoric; is this related to the number of ties they have to frontline activist groups? If so, did the ties or the rhetoric come first? By observing the same organizations lobbying in different coalitions over time, these data allow time-variant network analysis. Another example: environmental groups often supported the Obama administration’s policies but occasionally opposed them. When they did, one of their key progressive allies in mobilizing efforts, Organizing For America (formerly President Obama’s campaign organization, Obama For America), was conspicuously absent from the coalition. These data allow network analysis with lobbying coalition and policy-level covariates (e.g., supporting or opposing a given president’s agenda).

Organizations’ use of different social aggregation technologies offers one specific opportunity for studying issue networks and lobbying networks. Public comments are often generated through third-party nonprofit advocacy platforms that usually serve either the political left (e.g., Care2, MandateMedia, DemandProgress, and Daily Kos) or political right (Americans for Prosperity). Additionally, there are for-profit petition platforms like change.org and companies that serve a variety of political and corporate causes (e.g., SalesForce and VoterVoice). VoterVoice, for example, is a product of the government relations firm FiscalNote. FiscalNote specializes in lobbying on legislation, while VoterVoice specializes in “grassroots” advocacy services. VoterVoice states on their webpage:

Founded in 2000, VoterVoice, our flagship advocacy solution, was designed to fill a gap in the market for a robust tool that would provide high value to users and keep innovating to meet the needs of modern digital communication. As the market leader, we continue to set the trends and prioritize the features that drive results for you. Our vision remains simple: Deliver a seamless, powerful platform that lets you inspire advocates to action, so you can impact policy through actionable insights. We strive every day to honor our commitment to the 2,400 associations, nonprofits, and corporations who rely on us." (votervoice 2021).

Finally, there are a large number of front groups for corporate lobbying campaigns, many of which are run by government relations firms under multi-year contracts from industry associations. For example, I discuss Energy Citizens in 2.3.1.

The constellation of nonprofit and for-profit advocacy tools that different organizations use provides a distinct layer of nodes and edges (in addition to the policies on which they are lobbying and the coalitions they lobby with). For example, many of the national progressive advocacy organizations seemed to pay less attention to some issues (e.g., immigration) than others (e.g., climate change). The result was that immigrant advocates more often generated comments through change.org petitions rather than the nonprofit social aggregators that other progressive causes tended to use. These patterns of lobbying, mobilizing, and organizing behavior offer rich information about evolving issue structures and political alliances.

5.2.3 Public Pressure and Congressional Oversight

Members of Congress engage in agency policymaking in a variety of ways, only some of which are systematically captured in the data used to estimate the relationship between public pressure and legislator behavior in Section 3.4. Comparing the dataset on congressional behavior that I have assembled through this dissertation to other data on congressional engagement in bureaucratic policymaking could yield insights about both legislative behavior and bureaucratic policymaking. For example, when members of Congress submit official “comments” on an agency rule or forward their constituent comments (often, legislators do both), they are almost certainly logged in the public record on regulations.gov. However, if legislators do not want their communications to appear in the public record, they may contact a different office (for example, the secretary’s office rather than the program office issuing the rule). Judge-Lord, Grimmer, and Powell (2018) find systematic variation in how members contact agencies and sub-bureaus, and Powell, Judge-Lord, and Grimmer (2020) show that legislators often advocate on behalf of campaign donors.

5.2.4 Endogeneity

Assessing the impact of pressure campaigns is difficult because an organization’s decision is endogenous to the organization’s perceived probability of success and correlated with other factors we care about (for example, whether the organization is a business or nonprofit). Future research might disentangle different motivations for launching campaigns by measuring the amount of effort that organizations put into sophisticated lobbying. While different types of organizations have different capacities for sophisticated lobbying (e.g., the ability to hire lawyers and scientists), organizations also vary in the amount of effort they put into such lobbying. Small investments in sophisticated lobbying relative to the scale of public organizing may indicate that the organization was mobilizing supporters for reasons other than affecting policy. (It could also indicate that the organization perceives the policy fight to be more political than technical). This approach does not fully address the problem that lobbying behavior depends on the expected probability of success, but it may at least capture the extreme cases where organizations mobilize primarily for reasons other than influencing the policy at hand.

5.2.5 Mechanisms of Influence

As theorized in 2.2 and 3.2, the causal process potentially linking public pressure and policy decisions depends on how institutions filter and process the information they receive. When agencies receive large amounts of public comments, they often hire private-sector consultants to process and summarize public input. The instructions that agencies give to their consultants and the summaries that consultants produce offer a rich and systematic way to study the processing of political information. Researchers should be able to obtain these through Freedom of Information Act (FOIA) requests.

Another way to study how agencies process comments and the political information they contain would be to study the information that agencies (or their consultants) do and do not record about comments. For example, when a commenter fails to identify their name or organization when submitting comments, agencies occasionally fill this information in with “Unknown.” Agencies do this for mass comments as well, tagging a comment “Mass comment - organization unknown.” Comments that my methods identified as nearly identical were sometimes tagged as mass comments and sometimes not. This variation may offer a way to observe the extent to which an agency is processing—and thus capable of reacting to—political information. For example, recall the astroturf campaigns sponsored by the American Petroleum Institute (API) in 2.3.1.1. While all of these form letters resembled API press releases, some were identified as being associated with API-sponsored campaigns while others were not, suggesting that the agency may have been aware that API was mobilizing in some cases and not in others. This variation may provide leverage to assess whether public pressure campaigns are more likely to increase lobbying success (e.g., of API) if the agency explicitly associates them with an organization or coalition.

5.2.6 Methods of Measuring Policy Influence

Chapters 3 and 4 offer three methods to assess the impact of public pressure on bureaucratic policymaking: large-scale hand-coded lobbying success, detailed case studies, and computational methods that measure change between draft and final policy documents. Each of these methods could be improved upon.

The measures of lobbying success, position, and opposition used in Chapter 3 do not distinguish different dimensions of conflict within each policy. For example, tribal governments and advocacy organizations often lobby in coalitions with environmental groups against extractive industries on public lands but raise distinct objections related to tribal sovereignty. While I distinguish tribal sovereignty interests as a distinct and overlapping lobbying coalition, I do not distinguish which of their multiple interests are opposed. My data currently reduce support and opposition to a single spatial dimension, meaning that all reasons for supporting a policy are considered to be “opposed” if another coalition opposes the rule (even if for a different reason). This coding could be refined by disaggregating policies into multiple issues (dimensions of conflict) and assessing coalition lobbying dynamics on each dimension of conflict.

While I do not explicitly engage the vast literature on “process tracing” methods in the qualitative parts of Chapter 4, these methods informed my approach. Because notice-and-comment rulemaking under the Administrative Procedure Act is an unusually structured policy process, it is an excellent candidate for these methods. Compared to almost any other political phenomenon, rulemaking has clear inputs, outputs, and critical junctures that can be aligned and compared across multiple observations. Cases where rules span multiple administrations (like the Mercury rule discussed in 4) and rules that reverse rules published under previous administrations may be especially good candidates for process tracing. While atypical and inconvenient for the present research, the fact that the Trump administration published new rules on many of the same rulemaking dockets on which the Obama administration published rules presents clear criteria for selecting rules that are obviously related to previous rules.

There are opportunities to improve computational measures of policy change and lobbying success (D. Carpenter et al. 2020). Using the same methods that I use to group comments into coalitions (outlined in 2.3.2) and assess change in how policy documents discussed environmental justice (4.3.3), we can systematically capture which words (and thus the percent of words) changed between draft and final policy text. The challenge is linking these changes to commenter demands, which may also be done through my text reuse methods or a combination of other text similarity measures (Rashin 2017).

We may also combine hand-coding and computational text analysis tools. I will focus on two of the many opportunities on this front: using hand-coded data as training data for machine learning tools and using computational methods to analyze hand-selected parts of the text. First, hand-coded lobbying coalitions provide a corpus of related texts that may be used to train classifiers to automatically detect lobbying coalitions. Trained classifiers may perform much better than unsupervised clustering methods (models not trained on coded data).

Second, human coders are particularly adept at identifying and classifying important parts of texts. For example, humans can easily identify sentences where commenters suggest policy changes. Once identified, computational tools that rely on similar words and phrases to link comments with changes in policy text may provide much better measures of lobbying success than they would if using the entire text of a comment. Additionally, hand-coded metadata, such as the pro/con position of a comment, may help avoid challenges that natural language processing tools have with sematic features like negation. For example, we will be much more confident in whether an organization is asking an agency to make a change or not to make a change. Finally, hand-coded lobbying coalitions can allow inferences about lobbying success to leverage larger pools of texts that we know are asking for similar policy changes.

5.3 Re-thinking Public Participation in Bureaucratic Policymaking

One of the main contributions of this dissertation has been to clarify a way of thinking about and studying public comments. In light of the empirical evidence in the previous chapters, I now revisit some of the dominant ways of thinking about public comments among scholars and practitioners who are currently debating reforms to participatory processes. I start with a sketch of the various positions staked out by administrative law scholars (who have done nearly all of the work on this topic), each rooted in different theories of democracy. I then discuss several specific challenges and proposed reforms to the policy process.

I conclude that reforms underway in U.S. federal institutions that solicit, collect, and process comments are already heading toward enhancing participation and the quality of political information that it generates. However, certain additional reforms, particularly those that call for more transparency of mobilizing groups, would further improve the political information available to agencies and outside observers. I have argued that we should see civic participation in bureaucratic policymaking as a collective activity mediated by mobilizing organizations. This perspective focuses our attention on advocacy organizations: who they are, who they represent, who they are mobilizing, and how they are mobilizing—all potential targets for research and policy reform. Specifically, this perspective emphasizes reforms that require transparency from lobbying organizations.

5.3.1 Public Pressure Campaigns as Petitions

Understanding mass commenting as a form of petitioning, as I have argued, runs against how scholars often discuss it. With a few notable exceptions (e.g., Cuéllar 2005; Mendelson 2011; Rauch 2016), legal scholars often imply that mass commenting is misguided and potentially harmful spam of the technocratic process. In a report submitted as a comment on “Public Forum on the eRulemaking Initiative’s www.Regulations.gov Web site and the Development of the Federal Docket Management System” about reforming the public comment process, Stewart Shulman summarizes the concern:

Many fear a surging wave of electronic mass submissions will overwhelm and thus delay agencies with limited resources. Furthermore, some warn us that electronic rulemaking may already have instantiated a sense that rulemaking decisions are akin to a plebiscite. (EPA-HQ-OEI-2004-0002-0015)

In response to public pressure campaigns, agency officials and observers often assert that comments are not votes and rulemaking is not a plebiscite (Mendelson 2011). At the same time, agencies frequently do tally up comments on each side. In a 2008 rule regulating border crossing by private aircraft, The U.S. Customs and Border Patrol (CBP) cited the scale of public engagement as providing information on “public sentiment.”

based on the ample number of comments received by the end of the original comment period, CBP believed that public sentiment was accurately captured. (CBP-E8-26621/p-61)

In the preamble to the 2015 Waters of the United States Rule, EPA evoked both majoritarian and pluralist justifications:

This final rule reflects the over 1 million public comments on the proposal, the substantial majority of which supported the proposed rule, as well as input provided through the agencies’ extensive public outreach effort, which included over 400 meetings nationwide with states, small businesses, farmers, academics, miners, energy companies, counties, municipalities, environmental organizations, other federal agencies, and many others. (2015-13435/p-67)

EPA Administrator Gina McCarthy testified that EPA “received over one million comments, and 87.1 percent of those comments…are supportive of this rule” (Lipton and Davenport 2015). When the EPA issued a new rule rolling back the definition of Waters of the United States under the Trump administration, the preamble noted that there were “a number of comments” in favor of the new definition but “a significant number of comments” in favor of the 2015 definition (2020-02500/p-231). While not going so far as to explicitly state that a supermajority of commenters opposed the new policy, it is still remarkable that the agency would choose to note quantitative opposition to the policy. For some reason, this political information merited attention.

Agencies also regularly refer to form comments as petitions. When the Bureau of Land Management (BLM) received 6,948 identical emails that did not cite an organization, BLM officials (or consultants) labeled these comments a “Pro Fracking Petition” (LM-2013-0002-5720). A nearly identical comment was signed by “Oil and Gas industry individuals” (BLM-2013-0002-5688). Another batch of 1,289 nearly identical emails submitted through the social networking website Care2.org was labeled “Anti Fracking Petition” (BLM-2013-0002-5721)

The organizations running public pressure campaigns are even more explicit that they consider mass comments to be petition signatures, frequently citing mass comments as “Letters in support of [our organization’s] Comments” (NOAA-NMFS-2012-0059-0023) or a “petition from World Wildlife Fund-US (WWF-US), signed by 271,893 WWF-US supporters” (BOEM-2016-0003-2556). The files containing comments collected by organizations are often named “signatures.pdf.” Clearly, agencies and organizations already consider mass comment campaigns to be a form of petitioning.

How does this understanding of public pressure campaigns as petitions affect the debate over their value? Does a focus on political information affect the debate over the value of public participation in bureaucratic policymaking? In the next subsection, I sketch out some of the normative positions in debates over how to reform the rulemaking process. I then comment on several proposed reforms in light of the theory and findings presented in this dissertation.

5.3.2 The Roots of the Debate

Normative scholars vary significantly in both the value they place on political information and the extent to which they believe that meaningfully engaging large numbers of people is possible. I group those concerned with reforming the public comment process into five camps: (1) Participatory democracy optimists, (2) Regulatory democracy reformers, (3) Pluralist reformers, (4) Rational pluralists, and (5) Skeptics.

Participatory democracy optimists

Optimists see notice and comment as the “purest example of participatory democracy in actual American governance” (Herz 2016, 1). The process is technically open to anyone with an opinion to offer, and agencies are, to an extent, required to respond to substantive ideas. This optimism is rooted in a voluntarist idea of democracy and a value for inclusive and substantive discourse. In its optimism, these observers often assume that those who “ought” to be included do not have large barriers beyond the formal rules of the institution. It also assumes that meaningful discourse among those who participate is possible and likely to occur.

Regulatory democracy reformers

A related camp is slightly less rosy about the current institutions and practices but is nevertheless optimistic that reforms can at least improve the quality of discourse, and thus policy. For these scholars, democracy is more a function of procedure than inclusion. Cuéllar (2014) argues that rulemaking could be more discursive. Mendelson (2011) found that agencies often discard non-technical comments but argues that they should not because mass comments contain valuable information. For this camp, the quality of the public debate is more important than the total number of people, their affiliations, or their biases. They emphasize the transformative power of discourse.

Pluralist reformers

A different brand of reformer focuses less on discourse and more on interest group representation. Reform-oriented scholars building on pluralist ideas of representation argue that lobbying organizations should be required to disclose their membership, funding, and decision-making processes (Seifter 2016). In this view, organized groups—not individuals—are the central actors in public comment processes. However, because agencies often lack information about groups, it is difficult to know how well they represent the people they claim to. Interest groups’ faithful representation of their members is crucial to pluralist ideas of democracy.

Rational pluralists

Another group of scholars focuses on the benefits of organized interest groups and experts that can provide credible technical information. In this view, random or self-selecting members of the general public are not helpful or appropriate participants in the policy process. “The goal of e-rulemaking is to more fully capture such credible, specific, and relevant information, not to solicit the views of random, self-nominating members of the public” (Herz 2016). Speaking on the topic of mass comment campaigns, Oliver Sherouse, a regulatory economist at the Small Business Administration, expressed the same sentiment:

It’s worth remembering why we have a public comment process in the first place, which is that the public has knowledge that regulators do not have and that they need to do their jobs well…how do the poor quality mass comments affect small businesses in the comment process? The most obvious problem would be if legitimate small business concerns are just lost in the flood" (GSA 2019, 31–32).

Sharehouse raised an additional concern that mass-comment campaigns might distract people from writing their own, more informative comments. While recognizing that “not everyone who does not sign onto that kind of [mass] comment would be willing to write a high-quality one,” the tradeoff of a few “quality” comments for many “poor quality” comments is worthwhile. (GSA 2019, 32) This perspective is open to limited public participation, but only certain concerns are legitimate, for example, the opinions of client industries and new technical information on which to base rational analyses.

Skeptics

The most skeptical camp goes even further, arguing that open solicitations to the general public should be abolished. From the observation that the lay public usually has little to offer a technocratic process, they conclude that it is a waste of resources to solicit opinions that won’t be counted. More importantly, this camp holds civic republican ideas about the proper relationship between citizens and the state. It defines a democratic process as one that follows as straightly as possible from the elected branches. The goal is to de-politicize bureaucratic policymaking.

5.3.3 Reforming the Policy Process

The theory and findings in this dissertation are most compatible with what I have called the “pluralist reformer” project. While the discursive ideals of “regulatory democracy” optimists and reformers may be possible (at least at a small scale), their goal of including the lay public in a more discursive policy process is very far from how I have shown that civic participation in bureaucratic policymaking currently occurs. The rational pluralist camp either ignores the value of what I call political information or dismisses it as a normative ill. 

The “pluralist reformer” perspective aligns with the theory and results presented in Chapter 2. Interest group organizations are key intermediaries between the public and policymakers. Some make dubious claims about whom they represent. Other organizations organize mass comment campaigns, providing more concrete evidence of a supportive constituency.

Transparency More challenging is the concern that narrow, concentrated interests that typically dominate rulemaking will sponsor astroturf campaigns, thus creating a false impression of public support. This is a real concern. If dark money can secretly create impressions of support for a policy, the case for expanding participation is weaker. My analysis suggests that this is less common than genuine participation, but I did uncover examples of astroturf and fraud.

As with any form of public participation, fraud merits attention. Organizations submitting comments on behalf of people who do not exist or did not consent is akin to fraudulent petition signatures or fake constituent letters being sent to members of Congress.

Reforms requiring more transparency in the process by which organizations represent segments of society, like those proposed by Seifter (2016), would help guard against astroturf and fraud. Transparency also clarifies political information, making it more useful both to agencies and outside observers. To the extent that participation is mobilized by organizations, agencies can ask these organizations to be transparent about their sources of funding.

However, reforms requiring groups to disclose information about their funding and membership only go partway to address groups’ representational claims, which often extend far beyond their membership. Such claims are more difficult to assess. However, similar reforms could require disclosure of the funding and methods by which organizations gather petition signatures. Organizations or individuals acting in their professional capacity as organizations (not as individuals) could be required to disclose the source of the money that paid them to produce their comments. Likewise, organizations running campaigns could be required to disclose the sources of funding for the campaigns and how much they spent mobilizing public support. These numbers will provide a helpful denominator for agencies to gauge public enthusiasm and other kinds of political information that public pressure campaigns provide. More information about petition signers could help clarify which segments of the public are participating. As discussed in Chapter 4, the people that organizations claim to represent are not always the people who sign their petitions.

Authentication. If comments are allowed to be collected by third parties like the Sierra Club (recall Figure 2.4), the government can encourage but may not be able to enforce authentication tools like CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart). Organizations that upload large batches of comments already go through a verification process, but this does not address the veracity of their petition signatures. Requiring confirmation or authentication by email or phone alone would be a barrier to participation.

The most secure approach that allows third parties to gather comments may be to authenticate a sample of comments by asking each commenter to provide an email address or phone number at which to receive a message asking them to confirm the text of their comment. Asking for a second interaction with the agency raises the cost of participation but may also make participation a slightly richer experience. People may be more likely to feel that their voices were heard. Of course, many people will fail to authenticate their comments, but like mass comments themselves, the number of people willing to take the time to authenticate offers information on the intensity of genuine support. Unauthenticated comments need not be discounted, but authenticated comments may indicate stronger preferences.

More importantly, authentication guards against identity theft—fake political information. It is difficult to provide a large number of fake phone numbers, and real phone numbers used fraudulently may generate complaints, alerting the agency that the campaign may be engaging in fraud. Again, fraud complaints need not discount an entire campaign. Opponents of a campaign by the Sierra Club, for example, could submit fraudulent comments through their system in order to complain when asked to authenticate and thus cast doubt over the campaign. However, a legitimate campaign will also have a portion that will authenticate, allowing rough estimates of the scale of true support versus fraud.

Some means of authentication, such as email, may be slightly more open to fraud than others, like texts, but the benefits of lower barriers to participation may outweigh the greater risk of fraud. Because mass comment campaigns include a fairly large number of individuals, especially if the barriers to participation are lower, there are many opportunities to identify fraud of any significant scale, and the stakes of a few fraudulent comments slipping through the cracks are low.

Lower Barriers to Participation A second facet of reform may center around increasing access, both for organizations and the people they mobilize to participate.

Is it possible to broaden civic participation in bureaucratic policymaking? Is it worthwhile? Many assume not, but my analysis offers some hope. As I show in Chapter 2, more participation generally means more diverse and public-spirited voices at the table. As I show in Chapter 3, more public participation is related to more oversight from members of Congress (in part because they receive copies of public comments from their constituents). As I show in Chapter 4, agencies respond when people raise new concerns, especially when a large number of people raise them. At the same time, there is little evidence that mobilizing more people has an impact on the outcomes of particular policy fights. The coalition with the most comments is not necessarily more likely to win. It is possible that narrow private interests are able to strengthen their cases by showing evidence of broader public support. However, the correlation between pressure campaigns backing business interests and lobbying success could also simply be the result of selection effects: public interest groups have incentives to mobilize on policy fights even when they are likely to lose, whereas private interest groups do not.

Critics may argue that opening the door wider to petition campaigns lets in biases. Will this information be biased? Yes, but all information provided by lobbying organizations has a bias. Indeed we know that people who opt to participate are disproportionately privileged. An unequal society leads to unequal participation, but higher barriers to entry only make this worse.

Several recent reform proposals align with viewing mass comments as a form of petitioning. Rauch (2016) suggests making comment periods more like polls. Nationally representative polling may be a good investment in political information for some rules, but we know from polling research that individuals do not always have clearly formed opinions, and much depends on question-wording. The reality is that American politics is animated through groups—people rely on organizations they trust to keep them informed and engaged. Raso and Kraus (2020) suggest a similar reform whereby people could “upvote” comments with which they agree. Raso and Kraus (2020) lament the “mail from the public dumped on the agency’s doorstep” and claim that “upvoting” would “make rulemaking more interactive. The happy result would bring us closer to the deliberative national town meeting.” In contrast, I argue that, empirically, mass comment campaigns already act as a poll or upvote. An upvoting feature would thus codify the existing dynamic, not fundamentally depart from it. Either way, the possibilities for “deliberation” with the mass public likely exist in the historical push and pull of pressure groups and government institutions over time, not among millions of people upvoting their policy position in a single rulemaking process.

Reducing the Cost of Providing and Processing Information Reformers often highlight the value of linking comments to particular questions or sections of a draft rule. Technology can certainly improve sophisticated debates about a rule’s technical provisions. Indeed, sophisticated commenters have adopted track-changes technology to provide marked-up versions of draft rules. New technologies may provide an even more important role in gleaning political information from the “torrents of email” generated by mass comment campaigns.

Parsing the questions Agencies often specifically ask questions and solicit comments on specific topics in a proposed rule. These topics offer an initial structure to allow commenters to self-identify the topics of their comments. The American Bar Association’s Section of Administrative Law and Regulatory Practice recommends “technology that would allow agencies to identify categories that commenters could select when submitting comments.” To stay open to new ideas, lobbying organizations should be allowed to petition the agency to add topics to the menu or ask additional questions.

For example, agencies may encourage groups to mobilize in support of the groups’ answers to the questions that agencies ask. Given past trends, this will likely take the form of petition campaigns, and agencies should have mechanisms to receive those comments as such.

Offering Answers Agencies could also lower the bar to participation by posting answers to their questions from different organizations and allowing people to co-sign or disagree with them. This is both easier for participants and requires less work for agencies to interpret where the commenter stands.

5.4 Final Remarks

Can the masses provide useful information to policymakers? If the theory set out in this dissertation is correct, they can, but the extent to which U.S. federal agencies currently have the capacity to process and use this information is unclear. Reforms could make biases more transparent and broaden the number and diversity of groups that are able to participate.