This article is adapted from the last chapter of the author's book Speak Softly and Carry a Big Stick: How Local TV Broadcasters Exert Political Power (New York: iUniverse, 2005).

The founders of the United States were deeply concerned about the corrupting influence of power. They understood that, given the chance, elected officials would seek to preserve and enhance their power, even at the expense of democratic institutions. Accordingly, they designed a government based on separation of powers, where "ambition" would "counteract ambition." This entailed an elected president with veto power over legislation, an independent court with the ability to declare legislation unconstitutional, a legislature in which a two-thirds majority can override a presidential veto, and bicameralism in which legislation must pass both houses of the legislature.

The resulting system, mimicked by state and local governments, has worked remarkably well in ensuring that one branch of government cannot successfully usurp the power of another branch. But experience has proven it to have one major flaw: inability to reign in the power of elected incumbents within the most powerful branch of government, the legislature. This is manifested in the high reelection rate of incumbents. For example, in every election since 1996, the reelection rate among incumbent members of the U.S. House of Representatives has been over 98 percent.

The same pro-incumbent bias is evident in the states, especially in the big states with professional legislators. In California, every single incumbent legislator up for election in both the Assembly and Senate won reelection in 2004. In elections for state legislatures nationwide, only 61 percent of seats were even contested.

Elected officials have achieved such electoral success by creating barriers to entry for potential challengers. These barriers to entry manifest themselves in thousands of little details of democratic design but are generally listed in broad categories such as legislative ethics, campaign finance, redistricting, legislative records, and voting systems.

Of course, deny it as they might, the underlying reason for the pro-incumbent bias of our current democratic system is that elected officials have an intrinsic conflict of interest in instituting democratic reforms that might reduce their own chance of reelection. In effect, the fox (elected officials) has been put in charge of guarding the chicken coop (democratic competition) while pretending to be a good shepherd (guardian of democracy).

That elected officials act on their conflict of interest does not make them bad, anymore than a mouse that flees a cat so it can survive. Winning office is incredibly hard and, like life itself, not to be casually discarded.

The Constitution makes no mention that legislators might have a conflict of interest when designing institutions with a direct bearing on their own reelection. The major implicit exception is the First Amendment, which limits the ability of elected officials to turn the press into a PR vehicle for their own reelection.

The founders' blind spot may be attributed to the fact that within the conceptual confines of elected representative democracy such a conflict is insolvable. That is, if elected officials are the only possible type of democratic representative, then democratic reform free of conflict of interest is impossible. Of course, direct democracy would have been an alternative mechanism to deal with the conflict-of-interest problem, but the founders -- rightfully, in my opinion -- had a profound distrust of continental scale direct democracy, which is why they set up a representative democracy in the first place.

The Solution

That elected officials have a conflict of interest in instituting democratic reforms has certainly not gone unnoticed. For example, Harvard Professor Dennis Thompson, an expert on legislative ethics, concludes: "How can ethics committees claim to judge an individual conflict of interest when they themselves stand in a position of institutional conflict of interest? . . . No matter how much the ethics committees are strengthened and their procedures improved, the institutional conflict of interest remains." Similarly, Columbia University Professor Samuel Issacharoff, an expert on redistricting, says: "The redistricting process will not be reformed from within. No politician has incentive to change a system by which he or she obtained office and that dramatically enhances the prospects of remaining there."

In response to this blatant conflict-of-interest problem, a seemingly infinite array of "independent" entities have been proposed. But what these independent entities have in common is that they work within the confines of our system of representative democracy. At a minimum, elected officials or the political parties they control pick the members (or member) of the independent entity, which, as a result, isn't completely independent. In addition, members are often in some way accountable to those who select them. To the extent the members are truly independent, then another problem is introduced: they are unaccountable to the general public. We may hope that such truly independent members have good character and thus act in the general public interest. But if they don't, there is no way to hold them to account.

The fact that independent entities aren't completely independent does not imply that they are worthless.

When an elected official delegates a task (democratic reform) to a third party (the independent entity), the official's cost of influencing the outcome may be increased. Still, a high rate of incumbent reelection demonstrates how hard it is for these so-called independent entities to act truly independently.

Consider two of the most popular types: the bipartisan commission and use of the judiciary. The problem with the bipartisan committee is that it may reduce partisan bias but not alleviate the common interest of the political parties in protecting the incumbent elected officials who control the parties. The problem with using the judiciary is a combination of politicization and lack of accountability. The appeal of judges is that they are relatively disinterested parties, but giving them a vital political function provides elected officials with an incentive to politicize the judiciary even more than it is now. To the extent this problem is alleviated -- for example, by using a panel of retired judges -- the problem of unaccountability comes to the fore.

The solution I propose to this democratic dilemma is creation of a citizens jury on candidate information and electoral systems (a "citizens electoral jury"). The three essential features of this proposal are that (1) the jury be constituted of a random selection of voters, large enough to be representative of the general population; (2) the jury's jurisdiction be limited to candidate information and election rules; and (3) the jury be embedded in the checks-and-balances system of government.

Random selection of members gives the jury democratic legitimacy, so long as the sample of voters, like a poll, is adequately large to be highly representative of the preferences of the population covered by the legislature.

The narrow jurisdiction of the jury targets the area where conventional representative democracy, which is based on delegating decision making to elected representatives, breaks down.

Embedding the jury in the formal checks-and balances system of government makes it difficult for elected officials to ignore or otherwise override the jury's decisions when the jury fulfills its checking function.

The citizens electoral jury, with its large-scale and randomly selected citizen membership, has some of the same attributes as the "deliberative opinion polls" advocated by, for example, Jim Fishkin, Ned Crosby, and Robert Dahl. In both cases, random selection of individuals who are used to stand in for the general populace is intended to solve the motivation and information problems that bedevil the population at large when considering a complex issue. As with delegating decisions to elected representatives, there is a huge gain in efficiency from delegating a decision to citizen jury representatives because it is vastly more efficient for five hundred people than for two hundred million people to deliberate. The free-rider problems that plague large democratic bodies are also substantially mitigated.

The citizens electoral jury differs from a deliberative opinion poll in its second and third essential elements: a narrowly targeted jurisdiction and tight integration into the legal machinery of government. As with a regular opinion poll, any issue can be the subject of a deliberative opinion poll. But a citizens electoral jury is designed to deal only with a narrow type of issue where elected officials have a direct conflict of interest. Similarly, like a regular opinion poll, a deliberative opinion poll need not be integrated into the legal machinery of government. But such integration is an essential feature of a citizens electoral jury.

The Citizens' Assembly on Electoral Reform in British Columbia

To date, the closest implementation of a citizens electoral jury has been the recent Citizens' Assembly on Electoral Reform in British Columbia.

In 1996, the Liberal party of British Columbia won more votes than the second place New Democratic party. But thanks to the first-past-the-post electoral system, the New Democrats won control of the government. In 2001, the head of the Liberal party, Gordon Campbell, promised that if the Liberal party won the 2001 elections it would create a citizens' assembly on electoral reform to recommend ways of improving the existing electoral system. The Liberals won the election with an overwhelming majority -- 57.6 percent of the popular vote and seventy-seven of the seventy-nine electoral districts in British Columbia. Newly elected Premier Campbell then went about fulfilling his promise to the electorate.

On April 30, 2003, the legislature passed a motion to create the Citizens' Assembly. The premier then proposed and the legislature ratified the chair of the assembly. In the following months, 160 individuals were randomly selected in a multistage stratified sample to serve as the assembly's members. In addition, the chair of the assembly was given a vote, bringing the total to 161. The 160 were stratified so that two came from each of the seventy-nine political districts and the remaining two came at-large from the Native American (what in Canada are called "First Nations") community. Each political district was to have one male and female representative.

Initially, more than twenty-six thousand individuals, approximately two hundred from each political district, were selected in a stratified random sample so that each gender and age bracket in the sample reflected that of the general population. Only a small fraction of those thus selected agreed to participate. They in turn were stratified to maintain age and gender balance. From this group of approximately 1,000, the 160 were randomly selected.

The Citizens' Assembly deliberated in three stages. The first, from January through March 2004, was called the learning stage. Assembly members met over six weekends to learn about various electoral systems. The second, from May through June, was called the hearings stage. Subsets of assembly members participated in fifty hearings throughout the province to gather public comments. The last stage, from September through November, was called the deliberation stage. Assembly members met over five weekends and deliberated on the strengths and weaknesses of various proposals before selecting an electoral system based on the single transferable vote. This recommendation was then placed on the ballot for a referendum to take place on May 17, 2005. To pass, the ballot item needed 60 percent of the vote and acceptance in forty-eight of the seventy-nine electoral districts. Regular ballot items only needed 50 percent to pass, but the legislature wanted a supermajority for this item because it involved fundamentally changing British Columbia's system of democracy. The ballot item read: "Should British Columbia change to the BC-STV electoral system as recommended by the Citizens' Assembly on Electoral Reform?"

The May referendum received 57.4 percent of the vote and won seventy-seven of the seventy-nine electoral districts. Since it didn't reach the 60 percent threshold, it did not pass. Whether or not one considers this vote an affirmation of the Citizens' Assembly process depends a lot on one's perspective. In itself, the majority vote is remarkable. But given research showing that when voters are confused by a ballot item they vote the status quo, I interpret the vote count as an underestimate of latent support for the Citizens' Assembly concept. About half of the electorate had not heard of the referendum by election day and were thus predisposed to oppose it when they saw it for the first time in the voting booth. The assembly was a historically unprecedented type of democratic body that proposed an unfamiliar and relatively complex electoral system -- the single transferable vote (STV) -- yet had only a miniscule budget to promote itself and its recommendations. The assembly was for the most part internally focused during its deliberations and then disbanded five months before the vote. The Liberal party did not promote the recommendations. One reason was that they had promised before the assembly started not to take a position. Another reason may have been that once it was in power its supporters had little incentive to modify the electoral system in a way that would reduce barriers to entry for political competition. More generally, no political parties had an interest in promoting STV because this system of voting is candidate-centered and thus would undermine their power.

Members were paid $150 per day plus all transportation and lodging expenses for each day they participated in the assembly. Only one assembly member dropped out during the course of the year, and attendance at meetings was consistently above 90 percent.

The assembly maintained a sophisticated Website for internal deliberations and formal submissions from the public.

Other Variations on the Idea of a Citizens Electoral Jury

Although the Citizens' Assembly strikingly demonstrates that a citizens electoral jury can work, it is merely one variation of this basic concept. Using the Citizens' Assembly as a reference point, I suggest a number of modifications.

Expanded Jurisdiction

The Citizens' Assembly dealt only with electoral reform. But as described earlier there are other issues where elected officials have an intrinsic conflict of interest. I'd include all of them in the jurisdiction of a citizens electoral jury: campaign finance, legislative ethics, legislative transparency, redistricting, and voting systems.

The Problem of the U.S. Constitution

The Citizens' Assembly involved a referendum. The U.S. Constitution makes no provision for a referendum at the national level of government. Of course, it would in theory be possible to pass a Constitutional amendment allowing referendums. But that creates a probably insurmountable hurdle for implementing this idea. I suggest instead that the citizens electoral jury be a standing committee of Congress called the Citizens' Committee on Candidate Information and Electoral Systems or, as a shorthand, "Citizens Electoral Committee." This would not be an advisory committee but a new type of congressional committee with powers similar to any other standing congressional committee. In addition, like a conference committee, it would serve both branches of Congress. The citizens electoral committee would deliberate on issues within its jurisdiction and then send bills to the floor of both houses of Congress for a vote. This type of committee power might seem modest, but there is reason to believe that in fact the recommendations of the citizens electoral committee would rarely be overturned by the full Congress.

The difficult part of passing popular democratic reform legislation is usually getting it to the floor in Congress for a roll-call vote because, once there, it usually passes with an overwhelming majority. For example, lobbyist disclosure law took decades to reach the floor of the U.S. Senate and House of Representatives for a vote. But when it did in 1995, it passed the Senate 98