The Administrative Hunch: Is Cass Sunstein a Tocquevillian?
Stephen Sawyer’s post in January on the democratic paradoxes of the administrative state got me thinking, somewhat tangentially, about the Harvard law professor Cass Sunstein.
Like many non-lawyers, I first came across Sunstein during the Nudge craze of the late 2000s, which he has kept up with more recent books like, The Ethics of Influence: Government in the Age of Behavioral Science (2016) and Trusting Nudges: Toward A Bill of Rights for Nudging (with Lucia A. Reisch, 2019). By now, Sunstein’s thesis is familiar: politicians should take more cues from behavioral economics. If social science proves that people err in predictable ways, we could all use a “nudge”—bolstered by the latest psychological research—in the right direction. Good data can help nudge us not just to opt for better choices but also to make the choices we ourselves would have liked to make, had we possessed full information and self-control. In Sunstein’s preferred system, vices aren’t banned outright, but policy makers (“choice architects”) design our tax returns, cafeterias, electric bills, and retirement plans so that wiser selections become more intuitive (See here for The Tocqueville Review’s special issue on nudging in 2016).
My hunch is that the popular enthusiasm for nudging has mostly passed. Yes, Sunstein’s Nudge co-author, Richard Thaler, won the Nobel Prize in Economic Sciences in 2017, and the field of behavioral economics continues apace. But the post-partisan politics that Nudge heralded never arrived. Conservatives were naturally suspicious of an Ivy League professor prodding them toward the salad bar. And now, with anxieties over Cambridge Analytica and algorithms run amok, Sunstein’s center-left audience is less sanguine about the power of subliminal messaging— whether in the hands of the state or big tech companies. In order to figure out how people usually make and revise everyday decisions, we need “Big Data,” and the gathering of Big Data is looking a lot less neutral than it did seven or eight years ago.
Yet even if Nudge does turn out to be just another social science fad, the administrative impulse behind it seems a persistent fact of mass democracy. As Sawyer reminds us, Tocqueville himself was profoundly critical of state absolutism, while he nevertheless took an active interest in public law and modern regulatory powers. Sunstein has taken to calling himself a “libertarian paternalist,” and perhaps this oxymoronic label captures Tocqueville’s view of administration rather well.
In his more bureaucratic moments, Tocqueville trusted administration as a vehicle for equality. Similarly, one of Sunstein’s reasons for nudging us toward better choices in finance and health is his worry that inequality burdens our basic capacity to navigate mass society. Like Tocqueville the administrator, Sunstein maintains that the state can recognize utility and act efficiently on behalf of the public welfare. Although I doubt any of the 2020 presidential candidates will run on a “libertarian paternalist” platform, Sunstein does describe an attitude toward administrative governance that has been with us since the nineteenth century. Tocqueville certainly had his libertarian paternalist moments, and the tendency is even more pronounced in John Stuart Mill.
Sunstein’s project also blurs the line between the social and the political. Most often, it sounds as if the people doing the nudging will be well-intentioned designers or employees in corporate HR. But Sunstein praises federal initiatives, like food and safety guidelines and fuel efficiency standards, and leaves open the possibility that congressional action could nudge the nudgers. This notion has made more headway in states like Britain and the Netherlands, and some of Sunstein’s most enthusiastic supporters are EU researchers.
Perhaps the most striking aspect of Sunstein is his desire to present these administrative commitments as democratic. In earlier work on judicial “minimalism,” Sunstein cautioned judges against issuing over-ambitious decisions, lest they infringe on the “deliberative democracy” that is best left to legislators and local rule. But, more recently, he has tried to locate a version of deliberative democracy within the executive branch.
Sunstein served as Administrator of the Office of Information and Regulatory Affairs (OIRA) during Obama’s first term, and he has gone so far as to characterize OIRA as a facilitator of “deliberative democracy.” An OIRA conference call seems a far cry from a New England town meeting. However, Sunstein draws on the language of recent literature that stresses deliberative democracy’s “epistemic” value. This defense of deliberation, following John Dewey, argues that deliberative exchange is valuable, first and foremost, because it’s the only way for citizens to cultivate knowledge. We don’t deliberate because it lends legitimacy or because it feels good but, rather, because it’s how democratic citizens learn. For Sunstein, OIRA functions as an “information aggregator,” on the grounds that it collects diffuse information from across the executive. “On purely epistemic grounds,” Sunstein says, “there is special reason for deference to the decisions of the most knowledgeable branch.”
Note how deliberation morphs into deference. Thanks to the exchange of information across government agencies and federally-facilitated “public comment” periods, the executive branch, Sunstein suggests, can replicate certain democratic practices enough that the demos ought to cede control to administrators. OIRA has a reputation for being reasonably anti-regulatory and more efficiency-driven than other arms of the executive branch, and no doubt most of its employees have good intentions and are willing to share ideas. But does this really amount to democratic knowledge? Can a government agency “learn,” as Sunstein says OIRA frequently does?
In The Public and its Problems, Dewey worried that “A class of experts is inevitably so removed from common interests as to become a class with private interests and private powers, which in social matters is not knowledge at all.” I share Dewey’s concern. OIRA administrators might honestly exchange information in a manner that appears deliberative but that nevertheless fails to track social reality. By Dewey’s lights, this falls short of “knowledge.” I would be willing to depart from Dewey and, from a more straightforwardly utilitarian perspective, concede that OIRA administrators sometimes offer a helpful information-collecting function. But should we call this function democratic? Do we cheapen democracy when economists working in minor federal offices become our best examples of healthy political debate?
Elsewhere, Sunstein makes a similar category mistake in his discussion of prediction markets. He posits that betting forums like Iowa Electronic Markets or the Hollywood Stock Exchange may actually overcome “deliberative failure.” In normal deliberation, participants tend to coalesce around a group consensus and fail to account for outliers. In contrast, prediction markets can aggregate participants’ preferences while isolating group biases. It’s true that online forums aggregate “many minds” and greatly increase accuracy for certain types of predictions. But again, do we really want to equate large-scale data with democratic knowledge?
The nagging problem is that democratic knowledge is normally quite inefficient. On any given day, simple number-crunching might make more sense than Dewey’s trial-by-error approach. Furthermore, democracies often forget or repress the knowledge they do attain. The United States has elected some excellent leaders and, well, some very poor ones. Our learning curve is jagged. Responsible administration might help facilitate the democratic process, but it also risks becoming the refuge of those fed up with democracy’s timeline. Both versions of administration are present in Sunstein. His disappointment with the former may explain his nudges into the latter. However, this administrative tension isn’t unique to Sunstein. Just ask Tocqueville.
Photo Credit: DARPA via Wikimedia commons, “Big Data” (US-PD)