March 3, 2015


Unlocking Hidden Consensus in Legislatures

A legislature is a small group with a big impact. Even for people who will never be part of one, the mechanics of a legislature matter — when they work well, we all benefit, and when they work poorly, we all lose out.

At the same time, with several hundred participants, legislatures are large enough to face some of the same collective action problems and risks that happen in larger groups.

As an example, consider the U.S. Congress. Many members might prefer to reduce agricultural subsidies (a view shared by most Americans, in polls) but few are eager to stick their necks out and attract the concentrated ire of the agricultural lobby. Most Americans care more about other things, and since large agricultural firms lobby and contribute heavily to protect the subsidies, individual members face a high political cost for acting on their own (and their constituents’) views on the issue. As another case in point, nearly all Congressional Republicans have signed Grover Norquist’s Taxpayer Protection Pledge, promising never to raise taxes at all (not even minimally, in exchange for massive spending cuts). But every plausible path to a balanced budget involves a combination of spending cuts and tax increases. Some Republicans may privately want to put a tax increase on the table, as they negotiate with Democrats to balance the budget — in which case there may be majority support for a civically fruitful compromise involving both tax increases and spending cuts. But any one Republican who defects from the pledge will face well-financed retaliation — because, and only because, with few or no defectors, Norquist can afford to respond to each one.

These are classic collective action problems. What if, as in other areas of life, we could build an online platform to help solve them?

Many of the most important new online platforms of recent years share a similar story. They unlock a surprising amount of new value by making it easier for people to coordinate. The simplest examples, like Groupon, LivingSocial, or Gilt, let consumers flock together to get better pricing for the things they want to buy. “Crowdfunding” sites like Kickstarter have harnessed a similar logic to more powerful ends, unleashing new creativity and innovation: Kickstarter attracted almost a hundred million dollars in pledges last year and funded more than 10,000 projects, ranging from documentary films to a new model of electric bicycle. The Kickstarter model has also inspired a number of civic sites, including, Citizenvestor, and Patronhood.

These existing sites have often used thresholds or triggers to make sure that no one has to go first: No one has to commit before critical mass is achieved. A user who pledges support on Kickstarter providers her payment details, but knows she will only be charged if enough other users also sign up to fund the project’s whole budget. And when a merchant offers a coupon on LivingSocial, she knows that she will only have to provide the coupon if a  threshold number of people sign up to buy it (causing the deal to “tip” and become activated).

The right online platform might be able to work similar magic inside a legislature. [1] The users would be the legislators, each of whom could propose new bills or high-level policy commitments. Members could anonymously sign up to support a legislative proposal, with the proviso that their names will only become public if a threshold level of support is reached. If too few sign on, the supporter list stays private. If enough sign on, the supporter list becomes public — all at once. With no first mover, this could be a great way to unlock hidden consensus inside of the legislature. Members commit their political capital, in the same way as Kickstarter users commit their funds.

This kind of conditionally anonymous voting wouldn’t be natural to do on paper. The closest historical model in U.S. federal politics may be the House of Representatives’ discharge petition process, which until 1993 allowed members to add their names anonymously to a petition to consider a bill that had been bottled up in committee – and made the names public only if a majority of the House signed on. However, even then, the petition’s sponsor knew who all the signatories were, subject to an imperfect gag rule, and members could not be confident that their signatures would actually remain anonymous. [2]

With a computerized system, it is at least possible to appoint a neutral intermediary to keep the system up and running, and to configure that system so that individual initiatives can be considered without anyone, including a proposal’s sponsor, knowing who stands where. Using advanced cryptography it might be possible to design such a system so that even the system’s operator would be incapable of violating members’ anonymity. The more confidence members can have that their anonymity will stay protected (even if the petition process lasts a long time) the better the chances will be that the system can unlock hidden consensus.

There are a number of different ways to explore this idea. The online platform might feed directly into the legislative process, or might operate in parallel as a kind of sounding board. The platform’s policy proposals (and the eventual named lists of supporters) might be globally reachable, or might be publicized only within the legislature itself. Creators of new proposals might be kept anonymous, be required to identify themselves, or be given a choice. Members might, or might not, be given an option to change their minds after signing up to support a proposal. In any event, it would be important for the underlying technology to be secure — legislators would need to trust, and would need to be well-justified in trusting, that their support decisions would remain private until the threshold of support is reached.

A system like this would likely reduce the power of organized interest groups, since coordinated, simultaneous action by legislators could be much harder for outsiders to regulate. One way to look at this would be to say that legislators would become less accountable, since pressure groups would be less able to hold them to account. But it’s ultimately citizens themselves, rather than advocacy groups, that hold the franchise to keep legislators accountable. I’d argue that the kind of system I’m proposing here would create a bit more freedom of action for legislators — giving them slightly more room to exercise independent judgment on particular policy questions — and would increase the power of elections, and of citizens themselves, at the expense of the professionalized advocacy community.

Different legislatures undoubtedly have different norms and legal requirements — this idea would need to be tailored to fit them. But it might be piloted at any level of government, anywhere the world — as long as a legislature has more than a handful of members, these mechanisms have a shot at working well.

Given the overall level of frustration with legislative processes — and given the surprising magic that similar platforms have been able to create for larger groups — I think the idea is well worth exploring. A legislature willing to try it out could be a real pioneer.

[1] There is also an exciting family of new technologies for gathering public feedback on policy proposals — MADISON, We the People, and others among them. But this would be different: a tool targeted specifically at legislators themselves.

[2] For example, the very last time the anonymous discharge process was used, ironically to end the anonymity of future petitions, the sponsor eventually violated the spirit of the gag rule by revealing the names of the 223 House members who had not signed, effectively deanonymizing everyone in the chamber.


  1. avatar Jonathan Hall says:

    This is an interesting idea, but I’m not sure the implementation needs to be a web app (and, in fact, the nagging doubt about the security of a database presumably operated by some think-tank-like entity might be actively harmful.) This sort of thing could easily be implemented on paper, particularly in countries (I’m thinking ones with a Westminster system, but there may be others) where the administration of legislatures is non-partisan – rather than the sponsor keeping the list of names, the Speaker’s office (in Westminster systems, the Speaker is an elected MP, but renounces party affilliation on becoming speaker, and their non-partisanship is generally taken seriously) keeps the list (and maybe keeps the sponsor, or perhaps all members, or the public) updated on the numbers of supporters.

    This doesn’t provide guaranteed anonymity, but I’d predict most MPs in a Westminster system would prefer to trust the Speaker than the security of a web app.

    • As below, I’d say that the question of whether paper is feasible comes down to whether or not you’re using strong crypto to avoid the need for a trusted human handler of the anonymous votes. The cryptographically ambitious approaches to this problem would need a computer somewhere.

  2. avatar Jonathan Hall says:

    On reflection, I’m not sure this sort of thing would actually be good for democracy.

    Hiding legislators’ intentions from voters until they’re a fait accompli seems an intuitively bad idea – it’s depriving voters of the best possible information to decide who to vote for. After all, if an idea is unpopular with the majority of a legislator’s constituents, we probably ought to discourage that legislator from supporting it.

    Hiding legislators’ intentions from potentially-corrupting donors until they’re a fait accompli is arguably beneficial, but it seems to me like a pretty poor partial workaround, ignoring the actual problem, which is that these donors have the power to corrupt legislators in the first place.

    • Inevitably, voters’ decisions are all things considered — they take into account many factors, not just one vote cast by the incumbent (or pledged by the challenger). Lobbyists, on the other hand, often do have laser-beam focus on a particular issue.

      Part of the thought here is that partially obscuring policymaker intentions from lobbyists would better empower lawmakers to reflect the views of citizens. Widespread citizen frustration with Congress, reflected in the body’s record low approval ratings (which are somewhere south of twenty percent) suggests that reducing the effective ability of lobbyists to police legislator behavior would be a significant benefit, at least in the U.S. context.

      This idea doesn’t ignore the actual problem of special interest influence — instead, it takes full account of the fact that alternative efforts to limit donors’ influence have failed to pass constitutional muster.

      Some versions of this approach could indeed be implemented on paper. But any version that relies on strong crypto to do away with the need for a trusted custodian would have to be digital at its core.

    • “pretty poor partial workaround, ignoring the actual problem”

      I am with Johnathan on this one. This solution while it sounds great, it is only a bandaid trying to fix only the symptom of bleeding (without addressing the sharp objects that keep on cutting).

      I have been working on a series of “True Reforms” for politics. A “True Reform” is defined as political policy or enactment of law that correctly addresses the “cause” of the problem (and not just the symptom).

      Such True Reforms, dig deep into what is causing the problem. Incidentally one of these reforms I have been working on is True Healthcare Reform, which I won’t go into as it is off topic. But, I use the healthcare system as an example to what I mean between the differences between True Reforms and False Reforms.

      “A man walks into the doctor’s office complaining of a headache, a cough, an feeling fatigued. The doctor (without so much as checking the patient) prescribes the following treatment. Tylenol to reduce the headache, a cough suppressant syrup to reduce the coughing, and plenty bed rest to rejuvenate the fatigued body.”

      That is an example of a “False Reform” or one that recognizes there is a “problem” but simply wants to treat the symptoms of it.

      A True Reform in our example would be like the doctor doing some diagnostic checks, perhaps checks on the blood, temperature of the body, bacteria cultures, etc. to determine that what is causing these three symptoms may actually be a case of acute bronchitis. Which would have a significantly different treatment.

      The idea of anonymity of legislatures’ commitment or lack thereof to policies is a great idea to reduce the corruption caused by the entire lobbying industry. However, it doesn’t solve the problem with the lobbying in the first place.

      A “True Reform” would be outlawing lobbying in general. Make legislatures accountable only to their constituents. In fact, this is one point I have in my long line of True Ethics Reforms.

  3. avatar Jeff Tignor says:

    I have a similar concern to the one Jonathan raised. Elected officials are accountable to their constituents on a day-to-day basis, not just on election day. Elected officials are meeting with constituents at meet-and-greets, communicating via paper newsletters, e-mail newsletters, Twitter, etc. If a Congressperson has taken a position against tax hikes, that person is accountable for that position every day. And is likely reiterating that position frequently. If one day that Congressperson votes to raise taxes, even as part of a large majority, it is going to be shocking to constituents. It seems to me that most voters do not agree with 100% of the positions that a particular politician takes, even when they do vote for that person. Depriving constituents of the true position of their representative, especially on a major issue, seems like it would anger voters even more than a Congressperson taking an ultimately unpopular position in a transparent manner.

    • To the extent something would anger voters, aren’t elected legislators less likely to do it?

      I imagine that they would employ this mechanism for cases in which their constituents either don’t much care about an issue (as is the case for most members with respect to agricultural subsidies), or where many of their constituents (perhaps a majority) would view the change favorably.

      • I think that’s optimistic. Look, for example, at all of the fairly stupid ideas in the tech sphere (SOPA, PIPA) that had fairly strong support in congress until constituents started screaming their heads off.

        What you’re missing, I think, is that well-funded lobbying types will still be able to determine legislators’ intent on particular policy issues, because they still have the time and money to go down and meet with the legislators or their staffs every day. (Sure, the legislators or their staffs could lie to the lobbyists, or change their positions, but that’s the case however you arrange matters.) Ordinary citizens, meanwhile, would know only that an issue was under discussion (I’m assuming at least that much transparency is left in an app) and would have to try and separate the signal from the noise of all the proposals that would be floated without any intent of gathering major support.

        And there’s another side of it as well: by limiting the information about what measures a legislature is seriously considering, you’re also limiting legislators’ access to the opinions of the people they represent. If I have no way of telling (without a lot of work) what bills are really up for passage and what my representative thinks about them, then I’m in a position of shouting into a void, and much less likely to offer the opinons a constituent should.

  4. avatar Chris Camp says:

    While I share some of the concerns mentioned by others, I still believe this idea to be worthy of further study.

    As I see it, the strength of the idea rests on the following assumption: more information => better outcomes. At base this is about revealing information that, for various reasons, is currently kept hidden.

    Everyone commenting on this board likely agrees on these two key points:
    1. Democracies (and, in turn, legislatures and legislators) should aim to Maximize General Welfare: promoting the general welfare is foundational – it’s mentioned both in the preamble and at the very start of Section 8 (the section on Congressional powers and purposes) of the Constitution. Ideally we’d have a system that allows legislators, though a robust process of democratic engagement and education, to settle on policy positions that they truly believe in. In this idealized world, politicians voices and votes would reflect their true beliefs about which policies are in the public interest. The current situation, as everyone should admit, does not match this ideal. While we don’t know how often this occurs because there’s no good data on it, we can safely assume that politicians personal beliefs about a policy are not perfectly reflected in their public comments or voting record. The system proposed here would allow for more information on the size and geography of this gap. The long term effects on the gap are obviously unknown but I do share David’s hopes that more honesty, even if via anonymity (really some degree of pseudonymity since we would know, at minimum, that any assertion was made be a member of Congress (or the relevant legislature)), would be useful in our goal to optimize the general welfare.

    2. Democratic Process. Democracy is all about process. Democracy runs a broad spectrum from full consensus direct democracy to varieties of representative democracy, but at core there is some tie to the citizenry – some element of input. In our current system we value the ability for citizens to participate at every step throughout the legislative process. Depending on the details, a system for capturing the true* positions of legislators wouldn’t undermine the ability of citizens or corporations to continue to engage with legislators. So long as there is still an opportunity for a full debate on any policy proposal or bill, I don’t see how this damages the democratic process in any meaningful way. I think our core interests are protected so long as anonymous/pseudonymous assertions aren’t the final word, so long the anon/pseudo-communication system is just a method for overcoming first mover (and other) costs, and not a plan to bypass a robust democratic debate.

    In a related note, I’ve often thought it would be interesting to have a parallel shadow vote for legislatures, where they simultaneously vote openly and anonymously. How different would the votes be? and what would come of airing this difference in public?

    Interesting thoughts…


    *it’s probably naive to believe that every legislator would use this tool to communicate their conscience.