This post by Brian Feinstein and Daniel Hemel is based on the authors’ draft article, Partisan Balance With Bite, which is forthcoming in the Columbia Law Review.

President Trump last week nominated two individuals to the Federal Energy Regulatory Commission: Kevin McIntyre and Richard Glick. McIntyre is an unsurprising choice: he is a partner at Trump’s “favorite law firm,” Jones Day, and he has donated thousands of dollars to Republican candidates. Glick, by contrast, does not at all fit the mold of the typical Trump nominee. He is a registered Democrat who now serves as general counsel to the Democrats on the Senate Energy and Natural Resources Committee, and he has a long record of donating to Democratic candidates.

Why did President Trump nominate a loyal Democrat to the Federal Energy Regulatory Commission instead of naming another Republican? A partial answer is that it would have been illegal for him to put another Republican on the commission: FERC’s organic statute says that no more than three of the commission’s five members can be members of the same party, and with McIntyre, Trump had reached his quota of three Republicans. But that is only a partial answer because the statute doesn’t preclude the president from nominating a registered independent or a DINO (Democrat in Name Only) to FERC once he has filled his three-Republican quota. And yet Glick, by all accounts, is a true blue Democrat.

FERC is just one among dozens of multimember boards and commissions that is subject to a partisan balance requirement—a rule that no more than a simple majority of members may hail from the same political party. Other agencies at the heart of the administrative state—including the SEC, FCC, and FTC—have partisan balance requirements in their organic statutes as well. Administrative law scholars and political scientists have long doubted whether these provisions meaningfully influence the ideological composition of federal agencies—and, as a result, have largely ignored their potential effects. In line with this scholarly skepticism, a 1976 report for  the Senate Commerce Committee asserted that neither Republican nor Democrat presidents had chosen “‘bona fide, honest-to-God’ members of the other party” to fill cross-party seats on two important commissions.

Despite the prevalence of partisan balance requirements and the widespread skepticism about their impact, no rigorous cross-agency study in the last 50 years has sought to determine whether partisan balance requirements actually affect the ideological composition of multimember agencies. And until we know that, it is difficult to say whether administrative law scholars should pay more attention to these provisions. Will cross-party appointees to multimember boards and commissions sound a “fire alarm” when majority party members stray from statutory directives—and thus make it easier for Congress, the courts, and the citizenry to monitor these agencies? Will cross-party appointees bring ideological diversity—and thus improve the prospects for reasoned deliberation by ensuring that a wide range of views are represented? Maybe, but before we can even begin to answer those questions, we need to know whether partisan balance requirements meaningfully constrain a president’s ability to appoint ideological allies to multimember agencies.

So we decided to investigate. With the help of two stellar research assistants, we collected data on 578 appointees to 23 agencies over six presidencies and 36 years (1979-2014). We then matched that information to the Database on Ideology, Money in Politics, and Elections (DIME), which includes estimates of the ideological preferences of nearly 15 million individuals based on their campaign contributions. Consistent with the general scholarly skepticism of partisan balance requirements, we expected to find presidents of both parties gaming the system. That is, we expected to find Democratic presidents appointing relatively liberal Republicans—and Republican presidents appointing relatively conservative Democrats—to cross-party seats. We didn’t expect to see a lot of Richard Glicks.

But what we actually found was more interesting. The figure below illustrates. The bars show the ideological distribution of each president’s appointees to agencies that are subject to partisan balance requirements—blue for appointees who registered as Democrats, red for registered Republicans. (Surprisingly, we found vanishingly few instances in which a president appointed a registered independent to one of these agencies.) Each bar extends one standard deviation from the mean, and the placement of the president’s head shows where he falls on the ideological spectrum. The most conservative DIME score possible is +2; the most liberal score is -2.

At the beginning of our period of study, we see a Democratic President (Carter) appointing a fair number of moderate and liberal Republicans to cross-party seats, and we see Republican Presidents (Reagan and George H.W. Bush) appointing relatively moderate-to-conservative Democrats. DINOs and RINOs abound. But starting with Clinton, we observe a change. Clinton’s Republican appointees are consistently conservative, and George W. Bush’s Democratic appointees are quite liberal. The trend continues under Obama, whose Republican cross-party appointees were, on average, even more conservative than George W. Bush’s Republican appointees.

In light of the partisan rancor that has crept into many areas of governance during the past several decades, this trend is surprising. With the breakdown of bipartisanship, we might expect that presidents would be even more motivated to fill seats with ideologically sympathetic appointees today than three decades ago. Instead, we see presidents choosing cross-party appointees who are increasingly unlike themselves ideologically. What’s going on?

One possibility is that during times of divided government, the Senate majority forces the president to make bona fide cross-party appointments. (Think Harry Reid making George W. Bush appoint a genuine Democrat, or Mitch McConnell making Barack Obama appoint a red-blooded Republican.) Yet this story doesn’t fit the data. Near the beginning of our period of study, we see Reagan and Bush appointing moderate and conservative Democrats to cross-party seats even when Democrats controlled the Senate. Near the end, we see Obama appointing “real” Republicans to cross-party seats even when Democrats enjoyed a filibuster-proof Senate majority. Overall, we find no consistent relationship between power in the Senate and appointment patterns.

A second possibility is that the launch of the Federal Election Commission’s online donor database in 1998 made it easier to sniff out DINOs and RINOs—and so harder for presidents to sneak “stealth” nominees by the Senate. But we don’t think this explanation is especially compelling either. Information on nominees’ campaign contributions was readily available to Senate staffers for almost our entire period of study, and we don’t see any jump or kink in the trend line around 1998.

We think the most plausible answer lies in a story of “partisan sort.” Over the course of the 1980s, 1990s, and 2000s, ideology and party affiliation grew ever more tightly linked, with conservative Southern Democrats switching over to the Republican Party and liberal/moderate Republicans in the Northeast and elsewhere either becoming Democrats or dying off. And as these trends persist, presidents have found it more difficult to identify competent cross-party appointees whose ideological preferences they share. Democratic presidents may still be on the hunt for RINOs to appoint to cross-party seats (and Republican presidents for DINOs), but with fewer of these creatures left in the wild, presidents increasingly are compelled to select their ideological opponents for cross-party positions.

Our results suggest a novel twist on the standard narrative of partisan polarization. The conventional wisdom holds that increased polarization has strained the architecture of government. At agencies subject to partisan balance requirements, however, the effects of partisan sort may be salutary; partisan sort may lead presidents to select bona fide cross-party members—who, in turn, may serve as in-house monitors and may counterbalance tendencies that would otherwise drive groups to go to extremes. At a time of heightened concerns regarding partisan polarization, that in itself is no mean feat.

We hope that our article can contribute to the administrative law literature methodologically and substantively. As for the former, we seek to show how DIME scores can shed light on longstanding puzzles in administrative law. As for the latter, we make the case that partisan balance requirements are a feature of the administrative state worthy of further study—and, moreover, that quantitative and qualitative analyses of partisan balance requirements must be sensitive to shifts in the ways that these provisions have operated over time. And while it’s too soon to say whether the trends we observe will continue under President Trump (with only a handful of cross-party nominations so far), what we can say is that by documenting past practices regarding appointments to agencies with partisan balance requirements, our study establishes a baseline that can facilitate future evaluations of appointment patterns in the Trump era and beyond.

Brian Feinstein is a Bigelow fellow and lecturer in law at the University of Chicago Law School. @BrianDFeinstein. Daniel Hemel is an assistant professor at the University of Chicago Law School. @DanielJHemel.