C Formalizing the Theory
In Section 3.2.4, I briefly addressed the relationship between my argument and one leading formal model of commenter influence in agency rulemaking. Here, I offer an expanded version of that discussion.
My argument that lobbying strategies like pressure campaigns aim to create political information requires several crucial amendments to existing information-based models of rulemaking. Specifically, I argue that information about the political context in which policymakers operate can persuade them to make policy changes. Allowing policymakers to be persuaded in such a way allows public pressure campaigns to be incorporated into these models. Doing so also resolves some puzzling results of models that assume “fixed” political preferences. Allowing policymakers’ political priorities to be affected by political pressure from commenters (either directly indirectly through, for example, changing the behavior of members of Congress) creates uncertainties about policymakers’ political positions and incentives for lobbying organizations to attempt to affect them by providing political information.
It was not the aim of this dissertation to develop or test the implications of formal models. Rather I briefly review the necessary modifications to one leading formal model in order to illustrate the importance of political information to theories of policymaking. My aim is to illustrate a few of the implications of my argument for formal theory.
In the most sophisticated model of notice-and-comment rulemaking to date, Libgober (2018) posits a utility function for policymaker \(G\) as shown in Equation (3.1).
\[\begin{equation*} u_G(x_f) = \alpha_0 x_f^2 + \sum_{i=1}^N \alpha_i u_i (x_f) \end{equation*}\]
where \(x_f\) is the spatial location of the final policy, \(u_i\) is the preference of “potential commenter” \(i\), and \(\alpha\) is a vector of “allocational bias”—i.e., how much the agency decisionmakers care about their preferences \(\alpha_0\) relative to accommodating the preferences of others \(\alpha_{i=1:N}\). Bureaucrats balance their own understanding of their mission against their desire to be responsive. In Libgober’s model, \(\alpha_{1:N}\) is a fixed “taste” for responsiveness to each member of society (i.e., each potential commenter), so policy decisions simply depend on their answer to the question “what do people want?”
Incorporating insight about the power of technical information, we might interpret \(\alpha_0\) (the policymakers’ understanding of their own preferences) to be affected by technical information. Including political information in this model requires two additional parameters related to a second question “why would agency officials care?”
First, like other lobbying strategies, public attention and pressure may shift the strategic environment, leading policymakers to strategically shift their allocation in favor of some groups and away from others. Let this strategic shift in allocation be a vector \(\alpha_s\). For example, interest groups may mobilize elected officials to support their lobbying efforts. If these elected officials can sanction or reward agency officials or veto the agency’s policy, their involvement may reshape agency officials’ strategic calculations. Agency officials may then strategically adjust their policy.
Second, campaigns may directly persuade agency officials to adjust their allocational bias, for example, by supporting claims about the number of people an organization represents or the intensity or legitimacy of their policy demands. Let this direct shift in allocations by \(\alpha_d\). This parameter captures persuasion on normative grounds and beliefs about which segments of society (i.e., which potential commenters) deserve the benefits or government policy or should be responsible for bearing the costs. Policies allocate costs and benefits across groups. If an organization makes a persuasive argument for distributional justice or shows that it represents a large segment of the public, officials may update their beliefs and biases about how to allocate costs and benefits.
Let policymakers’ original, immutable taste now be \(\alpha_t\). Having decomposed the policymaker’s allocative bias into three parts (their fixed tastes \(\alpha_t\), shifting strategic environment \(\alpha_s\), and potential to be convinced \(alpha_d\)), the policymaker’s utility function is now Equation (3.2).
\[\begin{equation*} u_G(x_f) = (\alpha_{t0} + \alpha_{s0} + \alpha_{d0}) x_f^2 + \sum_{i=1}^N (\alpha_{ti} + \alpha_{si} + \alpha_{di}) u_i (x_f) \end{equation*}\]
If, after the comment period, the strategic environment is unchanged and officials remain unpersuaded to change their beliefs about which segments of society deserve favor, \(\alpha_s\) and \(\alpha_d\) are 0, and the model collapses to the original information game based on fixed tastes. This outcome is less plausible when groups go public and expand the scope of conflict.
Adding these parameters resolves a puzzling result of Libgober’s model. Empirically, rules that receive comments do not always change. This result is impossible in a model where bureaucrats only have known fixed tastes and potential commenters only seek policy changes. For policy-seeking organizations to lobby but fail to influence policy requires that they are either uncertain or wrong about an agency’s allocative bias or their ability to shift it. Incorporating political information allows change and thus uncertainty in an agency’s biases.
Even if we assume that policymakers’ allocative preferences are known, allowing them to be influenced by commenters means that one commenter’s incentives to comment now depend on other commenters’ lobbying strategies. This characterization of rulemaking aligns much more closely with empirical studies that show organizations providing technical information as a means of persuading policymakers. Likewise, this characterization of rulemaking aligns with my theory that groups mobilize public pressure campaigns to generate political information that could persuade agency officials to change their position for political reasons.
Incorporating political information allows us to begin formalizing intuitions about mechanisms of influence and thus the motivations for commenting. For example, Libgober (2018) asks, “What proportion of commenting activity can be characterized as informing regulators about public preferences versus attempting to attract the attention of other political principals?” (p. 29). Adding political information to the model allows us to formalize this question: Under what conditions do the decision to comment depend on an organization’s beliefs about \(\alpha_t\) versus beliefs about \(\alpha_s\)? Empirically, we may often be able to infer that the difference in commenting can be attributed to group \(i\)’s beliefs about \(\alpha_{si}\) if the behavior of political principals varies but other observed parameter values are similar across rules at a given agency.
Rational-choice explanations of why organizations comment on proposed rules build on an intuition that potential commenters will comment only when the benefits exceed the costs of doing so. This intuition ought to apply to other lobbying strategies such as public pressure campaigns as well. Adding public pressure campaigns as a lobbying strategy to Libgoeber’s model is straightforward. In this model, a potential commenter has negative quadratic preferences centered on their ideal policy \(p_i\) and \(u_i = -(x_f - p_i)^2\) where \(x_f\) is the final policy chosen by the agency decisionmakers. An organization will comment if the cost of doing so is less than the difference between their utility when agency decisionmakers select a policy having been informed about the organization’s ideal point \(p_i\) versus when they select a policy after having made a guess about the organization’s ideal point, \(z_i\). If \(c_i\) is organization \(i\)’s cost of commenting, then \(i\) will comment if it expects to be better off providing information than abstaining:
\[\begin{equation} E[u_i | p_i] > E[u_i | z_i] + c_i \tag{C.1} \end{equation}\]
Similarly, an organization will go public when it expects that the cost of sponsoring a pressure campaign to be less than the difference in utility when agency officials select a policy having been informed about the intensity of broader public preferences \(p_{public}\) versus when agency officials select a policy having made a guess about the intensity of the attentive public’s preferences, \(z_{public}\). While organizations often make dubious claims to represent broad segments of the public, a petition or mass comment campaign may provide information about \(p_{public}\) that agency officials see as more credible. If \(c_{campaign, i}\) is organization \(i\)’s cost of running a mass mobilization campaign, then \(i\) will launch a campaign if
\[\begin{equation} E[u_i | p_{public}] > E[u_i | z_{public}] + c_{campaign, i} \tag{C.2} \end{equation}\]
This suggests that public pressure tactics should be more common when agency officials are either poorly informed or distant from public opinion and potentially influenced by the types of political information created by public pressure campaigns.
In addition to informing agencies about public preferences, pressure campaigns may alter the strategic decision environment for agencies. The extent to which changes to the decision environment help or harm an organization’s cause may affect their decision to sponsor a public pressure campaign. Public pressure campaigns may shift the strategic environment in at least two ways. First, the general level of public attention may “politicize” a rulemaking. That is, it may make political factors more salient and technocratic factors less salient, perhaps by attracting the attention of political appointees, the White House, or members of Congress. Some organizations will do better and others worse in a more political decision environment. Second, the specific level of public support for an organization’s lobbying coalition may affect bureaucrats’ decisions to different degrees, depending on how politicized the rulemaking is.
To formalize these two intuitions, let \(\beta_i\) be the effect of the level politicization \(\gamma\) of rule \(j\) on organization \(i\)’s utility, \(E[u_{ij}| \gamma_j]\). Second, let \(\delta\) represent a general increase in utility for any organization \(i\) for an additional unit of public support on rule \(j\) given the rule’s level of politicization. Let \(\omega_{ij}\) represent a one-unit increase in support (e.g. an additional petition signature or form letter) for organization \(i\) on rule \(j\). More public support may only matter in more politicized decision environments. In more technocratic environments, support may be disregarded. This has several implications:
- Organizations that gain from politicization \(\beta_i>0\) may be double-rewarded for mobilizing pressure because they benefit both from how their campaign increases general politicization \(beta_i\) and the specific support \(\delta\) for their position.
- Organizations that receive negative utility from politicization \(\beta_i\) will not sponsor campaigns at low levels of overall politicization.
At some level of politicization, \(\gamma\), organizations facing an opposing public pressure campaign may have more to gain by counter-mobilizing than they stand to lose by further politicizing the policy process.
Additionally, an organization may comment or run a mass mobilization campaign if it benefits in ways that are independent of policy outcomes. Strategies such as “going down fighting” can be incorporated by adding exogenous benefit parameters to the utility function of the potential commenter/mobilizer. Let \(v_i\) be the benefit of commenting, independent of its effect on the policy outcome, such as pleasing members or reserving the right to sue. Let \(w_i\) be the benefit of running a mass mobilization campaign independent of its effect on the outcome of the policy at hand, such as fulfilling expectations of existing members or recruiting new members. An organizations utility function would then be
\[\begin{equation} u_i = -(x - p_i)^2 + v_i + w_i \tag{C.3} \end{equation}\]
Again, the observed behavior of commenting without policy change becomes a possible result if commenters are allowed a the strategy of “credit claiming” or “going down fighting” and incentives to do so.