Concern that Congress is unable to fulfill its basic constitutional responsibilities has reached an unprecedented degree of mainstream acceptance. Writers on the left and the right bemoan diminished Congressional capacity, the body’s general disinterest in defending its own institutional prerogatives or reaching bipartisan agreements on looming policy issues (e.g., Social Security solvency, deficit reduction, etc.). There is no shortage of explanations for—surely overdetermined—Congressional dysfunction and gridlock, from increasing elite polarization to bicameral disagreements to the centralization of power in party leadership in Congress and so on.
One common argument blames increases in the transparency of Congress’s internal workings for the lack of bipartisan compromise on major policy issues. According to this argument, Congress’s committees for most of the institution’s history were allowed to negotiate and vote in secret, hiding their activities not only from the public but from lobbyists, other members of Congress, and the media. Even the rules of some committees were kept under wraps. In these backrooms, partisans of different stripes would build interpersonal, cross-partisan connections and ultimately deliver policy compromises. The secrecy of these committees enabled compromise in the process of partisan negotiation.
Then, over the course of the twentieth century, following the Second World War, Congress eroded the secrecy of committee hearings through a series of enactments and rule changes, most importantly the Legislative Reorganization Act of 1970. This act, along with subsequent rule changes, removed the secrecy of committee roll call votes and shifted away from secret convenings as the norm in the committee system. Adding to these rule changes, beginning in 1979, C-SPAN cameras were introduced for floor debates in the House and later in the Senate. C-SPAN coverage further expanded to cover many committee hearings as well, illuminating previously undisclosed committee negotiations.
Taken together, all this new sunshine allowed lobbyists for special interest groups, members of Congress outside a given committee, the media, and the mass public to grapple directly with the policymaking process. This increase in transparency coincided with the transformation of Congress’s committees. Committee hearings that were once the nexus of policy development and information acquisition became stages for partisan kabuki theatre. Moreover, power in the policymaking process has shifted away from once-powerful committee chairs to party leadership offices. Proponents of the pro-secrecy story point the finger at these changes as a cause for broader legislative dysfunction.
The case against transparency is facially plausible. Formal, game-theoretic models of bargaining often separately consider public and private negotiations, given the different incentives the players face when negotiating in front of an audience. This is particularly true when the audience is your boss, the electorate, and other actors who can directly influence your ability to keep your job (e.g., special interest groups who make campaign donations). And while most of the public doesn’t pay attention to the fine details of Congress’s inner workings, outside groups and partisan actors can easily rewind the tape of committee hearings and replay selective, out-of-context moments of electoral interest for members’ constituents. It would be very surprising indeed if fully exposing previously secret policy negotiations to public view had no effect on the committees’ ability to reach a compromise, or the nature of the resulting policy outputs.
So should we once again close Congress’s committees, turn off the C-SPAN cameras, and expect a rapid return to Congress’s golden age of committee bargaining? If only it were so simple. A subtle and difficult challenge for secrecy as a reform mechanism is the threat of leaks.
A recent article explores the problem of leaks in depth, asking what we can expect from committees negotiating in three different institutional settings: in public, in secret, and in secret but with the possibility of leaks. Using a formal model, the authors demonstrate two key results. First, committees negotiating in total secrecy (no threat of leaks) achieve the best outcomes, and committees negotiating in public exhibit status-quo bias. Second, and counterintuitively, committees negotiating in public reach the same outcomes as committees negotiating in secret but under the threat of leaks. That is to say, committees that operate in secret but under the threat of leaks are predicted to fail, in equilibrium, to produce better outcomes than committees negotiating in full public view.
How could this be the case? The basic strategic problem that underlies this result is called dynamic inconsistency. At the beginning of a bipartisan committee negotiation, assume both parties want to reach a policy compromise, and enacting a durable compromise is their best-case scenario. However, once the negotiations commence, both parties will be tempted to leak information from the committee proceedings to damage their opponent. For example, imagine a bipartisan committee negotiating Social Security reform. This committee would need to consider reforms that, on their own, are highly unpopular and could be damaging if attributed to one party alone (e.g., raising the retirement age, increasing payroll taxes, etc.). Members of either party would be tempted during the negotiation to leak that the other side floated an unpopular reform to gain an electoral advantage, thus damaging the prospect that the committee will reach a compromise.
Nothing’s to stop a rogue committee member or staffer from calling up a reporter after a secret meeting and spilling the beans.
This dynamic inconsistency problem, where the parties agree at time 1 that compromise is good but can’t commit to not double-crossing their opponent at time 2, undermines the efficacy of secret committee governance. In the authors’ model, both negotiating parties are incentivized to hold back information that could lead to compromise for fear of leakage. If somehow the two parties could tie themselves to the mast and commit to each other in advance that they won’t leak, then it would be possible to reach a compromise. But left to their own devices, the temptation to deviate during the negotiations and attack their opponent is too strong, overriding their interest in reaching a compromise.
To be sure, game theoretic models of politics are stylized, reducing the strategic environment in ways that are somewhat unrealistic. In this case, we might worry, for example, that leaks of verifiable and non-verifiable information might induce different strategic behavior, or that interpersonal factors like trust could affect the probability of a successful compromise (as some former members of Congress suggest). But the fundamental intuition and strategic tradeoff in this case is clear enough, and it is easy to find examples of secret committee negotiations that failed due, at least in part, to leaks. Moreover, a subtle implication of the model is that when the threat of leaks is present, we might expect bipartisan negotiations not to begin in the first place. Why participate in a bipartisan negotiation at all if you believe your counterparty will betray you in public at the first opportunity? There is also technological variation over time that contributes to the threat of leaks. In the age of the Internet, it is easier than ever to leak documents, recordings, etc.
Of course, the policy being debated also matters. In some areas, the parties can “expand the pie” to reach an agreement, for example, by increasing both defense and nondefense discretionary spending. This is sometimes called integrative bargaining, where the two parties try to reach an agreement that incorporates aspects both parties care about, typically over multiple dimensions. In other policy domains, bargaining is necessarily zero-sum (i.e., if I win, you must lose). These are often situations with a binary outcome: a judge is confirmed or not, a trade agreement is approved or not, etc. In the spending case, if the federal government had a balanced budget requirement, then any new spending would come at the immediate expense of other spending, or new taxes. In lieu of such an amendment, members of Congress often happily borrow to cover new spending, converting a zero-sum case into an integrative one. In theory, secrecy is useful in the integrative, multidimensional case—where members can imagine a range of mutually beneficial compromises—and less helpful for zero-sum negotiations, where at the end of the day, someone loses.
This model also helps explain why so much integrative policy development happens outside of Capitol Hill in the myriad think tanks and non-profit organizations around Washington, DC. These institutions serve a vital role in the policymaking process by granting policymakers (more often their proxies) a forum for developing compromise proposals—or searching for integrative solutions—that is largely inaccessible to the public. Leaks from these forums would be of little electoral or political consequence given the degrees of separation between these outside deliberations and actual elected officials.
Whether or not this ecosystem of third places for policy negotiation is a net public good is hard to know. Policy non-profits are themselves funded and operated with money from large grant-giving organizations and special interest groups. One might reasonably worry that these financial interests affect the policy development that goes on in these places. For what it’s worth, in the modern Congress, special interest groups and donors have direct access to the policymaking process as well. Counterintuitively, it’s possible that negotiations that happen off the Hill may in some sense decrease the policy influence of special interests, since electoral accountability in these spaces is out of reach.
While leaks undermine secret committee negotiations, employing secret ballots for more votes in Congress is worth considering. Many votes in both chambers are carried out by voice vote, but under the Constitution, final roll call votes on the floor of either chamber must be recorded if one-fifth of present members demand it. However, there is no such constitutional provision for committees. Regardless, many committees follow the one-fifth rule for disclosing roll calls. Even more extreme, in some committees, a single member can request a recorded vote. We could convert some portion of votes, say all committee-level and below, either to total secrecy or to a teller system where individual members’ votes are not recorded (i.e., only the count is disclosed). Secret ballots are already employed for some other votes in Congress. For example, Senate Republicans elected Majority Leader John Thune by secret ballot last year.
If Congress’s committees were to again turn away from public view, they might also consider creating a small institution to enforce information security. This would hardly be unheard of, given Congress’s proclivity to create non-partisan support offices and agencies like the Senate Parliamentarian or the Congressional Budget Office. As we’ve already established, members from both parties would often be made better off if they could somehow commit to not leaking. A third party, call them the “Congressional Paper Shredder,” with preferences orthogonal to partisan politics, could credibly play such a role.
Even still, nothing’s to stop a rogue committee member or staffer from calling up a reporter after a secret meeting and spilling the beans. If negotiations involve the president, which on major topics they almost always eventually do, then the executive branch becomes yet another potential source of leaks. It’s difficult to imagine institutional tweaks that could address these avenues for leaks, making the continued substitution to outside groups or other venues for policy development and compromise even more likely. On the other hand, making non-floor votes secret is entirely feasible in principle. One can’t leak information that doesn’t exist!