This is the second in a series of essays about Jonathan Rauch’s new book, The Constitution of Knowledge. This series is sponsored by Claremont McKenna’s Salvatori Center for the Study of Individual Freedom.
Adam J. White is a senior fellow at the American Enterprise Institute and co-executive director of George Mason University’s Gray Center for the Study of the Administrative State
In his urgent new book, Jonathan Rauch shows that Western civilization’s capacity for ascertaining, testing, and sharing the truth depends in no small part on four key institutional communities: expert researchers, mainstream journalists, government agencies, and the legal community. These are the institutions of his “Constitution of Knowledge.” Their rules and norms, he writes, create a system of “mechanisms of public decision-making and social adjudication,” which “resembles the U.S. Constitution.”
The key word is “resembles.” Rauch takes care not to overdo the constitutional analogy, and so should we. His Constitution of Knowledge is a framework for knowing the world; the U.S. Constitution, by contrast, is a framework for shaping the world. This is a fundamental difference, even when the goal of both constitutions is the advancement of an objective standard: for one constitution, the truth; for the other, the public good. Still, the analogy helps to illuminate many aspects of the problems facing us today.
But recall how the U.S. Constitution’s institutions were supposed to advance the public good: less by redundancy than by specialization. Congress is vested with legislative powers, the President with executive power, and the courts with judicial power—the branches of government differ in what they are meant to do and how to do it. While we often talk loosely of “checks and balances,” our constitutional system requires not just “checks and balances” generally but the three institutions specifically, checking and balancing one another in inherently different ways.
And this is why institutional drift among the three branches of government is such a problem for our constitutional system. The more that a given branch of our government confuses its purpose for that of others—the more the executive branch becomes our legislature, or the more the legislature becomes an adjunct of administration, or the more judges become politicians—then the more that the constitutional system as a whole suffers, even when the adrift branch intended to do good.
So in considering Rauch’s Constitution of Knowledge, we need to focus not just on the general purpose of the entire system, but rather on the specific purpose of each part of the system. For each of his four main informational institutions—professional research, mainstream journalism, government agencies, and the legal community—we can see the superficial differences among them. But can we safely say that each of these four institutions has a distinctive purpose that defines its work, its norms, and its limits?
We can easily spot superficial differences between, say, academics, journalists, federal agency leaders, and public interest lawyers. Yet we often see them doing very similar things. Look no further than cable news shows or newspaper op-ed pages, where you’ll find a jumble of each group, all making completely interchangeable contributions.
So what actually distinguishes the four groups’ contributions to Rauch’s Constitution of Knowledge? Do these institutions still have well-defined purposes that shape, limit, and channel their members toward each institution’s particular contributions to knowledge? Or, more likely, does each suffer from varying degrees of institutional drift?
Of the Constitution of Knowledge’s four institutions, journalism probably faces the best-known problems: from market incentives driving major news organizations to pursue ever more polarized subscriber bases, to the ideological fervor roiling newsrooms. But it is not the only informational institution with such problems. The community of expert researchers, for example, faces unprecedented uncertainty—especially on college campuses—about its proper purpose, powers, and limits. Too often, we see researchers trying to treat questions of value judgment as questions of “science” or “fact.”
But here let us focus especially on government agencies. Their inclusion in Rauch’s discussion of knowledge formation is itself a contribution on his part, for he reminds us of government’s crucial and historical role in the processes of information. This role dates back to the very beginning of our republic, to the expeditions that built our nation but also built our stock of knowledge.
As Rauch explains, federal agencies’ relationships to the Constitution of Knowledge actually vary: some agencies gather discrete facts; some conduct general research and compile statistics; some use information to create regulatory policies.
But many do all of the above and at the same time, finding facts and making policy in a unified process. An agency identifies a general policy goal, then proposes a specifically regulatory initiative, then invites the public to submit information and arguments, and then finalizes the policy with justifications rooted in the record that has been built along the way. The fact-value distinction is thin enough in theory; in administrative practice it’s practically nonexistent.
Rauch sees things somewhat differently. He credits the Administrative Procedure Act of 1946 for requiring agencies to “subject [their proposed regulations] to public comment and criticism, justify them with objective evidence and analysis, and make sure the evidence and analysis can stand up in court.” But the APA, now marking its 75th anniversary, does this only gently and malleably, if at all. It nominally forbids agencies from making regulations that are “arbitrary” or “capricious,” but this prevents only the most poorly reasoned policies; similarly, the APA sometimes requires agencies to support their policies with “substantial evidence,” a standard that might seem onerous but actually requires only “more than a mere scintilla” of evidence to clear the bar.
Such standards do little to incentivize informational rigor; if anything, they disincentivize it. Instead of “evidence-based policymaking,” we get too much “policy-based evidence making,” as Oren Cass aptly put it. And even when policymakers try to be as factually rigorous as possible while pursuing policy aims, they risk undermining the public credibility of their factual findings by intertwining them with their newly proposed policies.
Upon winning the Nobel Prize in Economics in 2018, Paul Romer urged Congress to “protect the independence of the federal government’s scientific advisers, who must be uncompromising in their pursuit of facts,” by “set[ting] them apart from regulators, who rely on facts but must compromise to balance competing interests when they write regulations.”
This strikes me as correct and urgently needed. Yet I must confess that I have yet to see how the line might be perfectly drawn in modern agencies when Congress has built them to be both factfinders and policymakers.
Perhaps the Madisonian approach in such circumstances would be to strive for the separation of fact and policy, even if the ideal is unattainable, by first looking for the particular institution’s ideal contribution to the pursuit of truth, and then asking how we can practically structure the institution to achieve it.
Such an approach could take inspiration from the founding era’s great government-supported expeditions, which built both the nation and the nation’s stock of scientific knowledge. It could also take inspiration from the early informational agencies, like the Census Bureau and the U.S. Geological Survey, and from more recent examples like the Energy Information Administration (created in 1977) and the White House’s own Office of Information and Regulatory Affairs (1980). All of these are reminders of the nation’s capacity for building credible informational institutions, even in difficult times or on controversial subjects.
Again, the goal is to remove as much policymaking power from the regulatory factfinders as practically possible, and to remove as much factfinding power from the regulatory policymakers. The policymakers would be entitled to their own opinions, but not to their own facts; and, by the same token, government factfinders must not become de facto policymakers themselves.
The work of institutional design would also require thinking hard about whether the informational agencies should remain concentrated in the executive branch and independent agencies, or if more of them should be reformed as legislative commissions.
Again, these considerations are just facets of the more fundamental constitutional question: how do government agencies either contribute to knowledge or rely on it, and how do we shape them to best undertake those particular constitutional roles? We can and should ask similar questions about Rauch’s other informational institutions—research, journalism, and law—in order to reform each of them to advance their respective institutional purposes in the Constitution of Knowledge.