Who polices the administrative state?
Oversight is political. Deciding what (and how vigorously) Congress should investigate are inseparable from preferences about leadership and policy. Those are not universally shared. This seems as clear now as ever, with Democrats pitching a return to majority status in the House. By the conventional narrative, Republicans are holding back a torrent of document requests, subpoenas, and public hearings—the essential ammunition for negative headlines, scandal, and (according to the President and a few Democrats) impeachment. Most recently, Democratic leaders have promised to use their projected majority to put Trump “under the microscope.”
This does not sit well with everyone. Some might acknowledge oversight is informed by politics, but there is something wrong with oversight in service of politics. Henry Waxman (D-CA) recently cautioned Democrats to focus on commonly held principles like waste and effective governance. Elise Bean, who heads a non-partisan program to train congressional staffers, argues that “investigations aimed solely at damaging the presidency would be a grave mistake” that would “generat[e] more noise than light.” Meaning, if the conventional wisdom that oversight is politically motivated is correct, it is bad for the public at large.
Unfortunately, most social science suggests politics influences when and how much Congress polices the executive branch. For example, in their recent book, Douglas Kriner (Cornell University) and Eric Schickler (UC Berkeley) find that from 1898 to 2014, divided party control of the House and the presidency led to an extra 44 days of investigations compared with unified control. The political gain from “throwing mud” on the other party’s ability to govern incentivizes this disparity. Justin Peck (Wesleyan University) and I found an especially illustrative case-in-point: prior to the popular election of Senators, the divided-unified government difference did not exist in the Senate. Once incumbent Senators began campaigning like members of the House, investigations took on a House-like pattern.
This begs an obvious question: is there any hope for those who laud the pursuit of shared (and possibly non-partisan) values like accountability and good governance? Is oversight always driven by politics?
Maybe not. For one thing, past research does not say that oversight comes to a standstill when Congress and the president share partisan affiliation. For another, there may be other tools of oversight that do not exhibit the same blatantly political pattern. In a recent study, titled “Who Polices the Administrative State?” (ungated draft) – I found just that.
I collected thousands of records of correspondence between members of Congress and the executive branch. Agencies keep these records, so they have to be FOIA’d one agency at a time. (If there are any executive branch officials reading this who have been on the receiving end of my requests: thank you very much and I hope the result is worthy of your time.) Correspondence between Congress and agencies runs the gamut of topics: from everyday constituent problems and grant support letters, to comments on pending regulations and information requests about programs. In total, my study includes around 50,000 cases of individual members of Congress contacting 16 different agencies from 2007-2011.
That time period is critical because it includes the transition from George W. Bush to Barack Obama. Presidential transitions, of course, lead to dramatic changes in agency personnel that shape the direction of policy. These shifts in agency ideology have been quantified using the political contributions of agency officials, thanks to innovative work by Adam Bonica (Stanford University), Jowei Chen (University of Michigan), and Tim Johnson (Williamette University). Some agencies shift more than others based on Senate confirmation and presidential priorities. If the conventional wisdom about oversight is true, we should see members of Congress more vigorously overseeing agencies who move away from their preferences. In the case of the Obama transition: Democrats should go easier on the agencies that become more liberal, whereas Republicans should flood them with calls, letters, and document requests.
I find nothing of the sort. In the data, the unconditional probability of a member contacting an agency in each Congress is about 36%. But for a typical (i.e. one standard deviation) increase in the ideological distance between agency and member of Congress, the change in the probability of oversight is essentially zero. By that, I mean it is not statistically distinguishable from zero (by convention), and the largest effect within a 95% confidence interval is an increase of 0.7%. Instead, oversight is driven by other factors that will not surprise: being on the relevant oversight committee and poor program performance. This is a far cry from the dramatic differences in other forms of oversight.
What explains the difference? My hunch is the incentives created by public oversight. When members of Congress contact the executive branch informally, few know about it. In fact, to get someone to pay attention, congressional offices sometimes send their letters to journalists—hoping it makes its way into a story. But most correspondence is nothing like this. To offices on both sides of our separate powers, this contact is a daily part of governance. No grandstanding, no public shaming, and no soundbites. In other words, oversight out of public view may provide a credible opportunity for Congress to perform its duty in good faith.
No wonder, then, why the Trump administration’s initial instruction to agencies to ignore requests from Democrats in Congress drew rare, bipartisan condemnation. It is possible that Republicans have no incentive to silence the oversight of their colleagues because they themselves benefit from their colleagues sharing the burden of policing the administrative state. Many concerns—like making sure veterans can make timely medical appointments, or ensuring nuclear power plants run safely—are non-partisan.
This argument is a “hunch” because I have no direct evidence beyond what is provided here and in the study. And scholars have just begun to examine vast data on communication between Congress and the executive branch; my study is only one example of the fruit of that labor (others are here, here, and here). But what is clear, given contemporary debates about oversight, is that there are open questions about how Congress can most effectively perform this function—and it is up to researchers to generate the evidence to answer them.