Saturday, January 6, 2007

Jim Snider's Proposal for School Board Electoral Reform

Executive Summary
In my judgment, this proposal combines the best features of the Anne Arundel County School Board Nominating Convention and the best features of an elected school board. Specifically, it combines the higher quality democratic deliberation associated with a self-selected nominating convention and the more rigorous democratic accountability associated with an election.

Introduction
Anne Arundel County Executive John Leopold, House of Delegates Speaker Michael Busch, and many other leaders in Anne Arundel County have called for changing the current system of electing school board members. Although calls for changing the School Board's electoral system are not new, never before has there been such high level and widespread political support for such a change.

A consensus has emerged that the current electoral system for the School Board is unaccountable and undemocratic. However, a consensus has not emerged on how to address that problem.

Under the current system, Maryland's governor is granted the sole legal authority to appoint the School Board members. In doing so, he may consult from a list of nominees provided by the Anne Arundel County School Board Nominating Convention, a self-selected, all-volunteer, private caucus that nominates two individuals for each open seat. The Nominating Convention typically attracts about 100-150 volunteers called "delegates." The Nominating Convention meets on three different evenings in three different locations throughout the county. At these meetings, the delegates have the opportunity to do candidate Q&A. Typically, delegates only attend one of these Q&A meetings. Then, on a fourth evening, the delegates vote on the candidates.

During the last legislative session two bills were introduced to change the electoral system. One, the County Council Election Model, sought to create an elected school board modeled on the way the County Council is elected. The other, the Judge Election Model, sought to create an appointed school board modeled on the way judges are appointed.

An advantage of the County Council Election Model and the Judge Election Model is their familiarity. Although some of the details differ, they are both modeled on formal electoral systems currently in use in Maryland. This makes them relatively easy to explain and justify. Nevertheless, bills representing both models have repeatedly gone down to defeat, largely because their defects are well known.

Elected school boards are no panacea, as illustrated by neighboring Prince Georges County, where the Maryland legislature had to replace a dysfunctional elected school board. More than 95% of school boards in the U.S. are elected, and there is no consensus that such boards perform better than appointed ones. Indeed, there has been a trend in recent years for large school districts to move to appointed boards. Many scholars believe that school board elections are even more prone to special interest politics than presidential, congressional, gubernatorial, state legislative, mayoral, and county council elections. This is explained in part by the fact that only 20% of voters have children in public schools and much of the balance are notoriously ill-informed about school board elections. The press typically provide less and lower quality coverage of school board elections than those for more than a dozen other elected offices, including president, governor, mayor, U.S. senator, U.S. representative, state senator, state representative, and city councilor. Elected school boards tend to work best in communities far smaller than Anne Arundel County. When school board members have to raise large amounts of money and volunteer resources to run in a large district, they invariably have large special interest obligations to pay off.

Similarly, appointed boards are no panacea, as illustrated by the current debate over the school board electoral system in Anne Arundel County. The general rap against appointed boards is that they may provide valuable expertise and advice to elected officials, but they are inherently undemocratic so should be given minimal legislative powers. Problems of ignorance and undemocratic incentives may be aggravated when statewide politicians are asked to select a board to act on behalf of local interests. The specific type of appointed board now being proposed for Anne Arundel County would give statewide politicians the right to appoint individuals to a fifteen member nominating body. A major criterion for choosing the members of the nominating body would be their political power and the debts politicians owed them. The nominating body would nominate two candidates to the governor, and the governor would then be legally bound to select from among the two.

There are many arguments that have been raised for or against these two electoral models that I haven't raised here. There are also many other important electoral options and issues that haven't entered the debate at all. Even a 500 page book could barely cover all the options and arguments that might be relevant to selecting a school board electoral system for Anne Arundel County. Let it suffice to say that, as someone trained as a political scientist, I have tried to weigh as many of these electoral factors as possible before coming to a judgment about the best electoral system for selecting Anne Arundel County's school board representatives.

Please note also that there is no one right electoral system. Even in Anne Arundel County the range of variation in electoral system variables is large. Examples of variation include the number of members in the elected body, the number of members in the electoral district, the proportion of at-large vs. local district representatives in the elected body, the type of designation next to a name on the ballot; and the use of competitive vs. approval voting for candidates. Worldwide, of course, the range of electoral diversity is far greater. The goal in designing an electoral system is to weigh all the variables and design an overall system right for a unique political situation.

The Basic Proposal
The basic type of electoral system I propose is what political scientists call a "Citizens Assembly." The gist of the proposal is to 1) replace the current self-selected membership of the Nominating Convention with a randomly selected membership that has greater democratic legitimacy, and 2) give this democratically representative body formal, legally enforceable electoral powers, something that cannot be done with the self-selected Nominating Convention. When the governor is not bound by the decisions of the Nominating Convention, he is given an incentive to introduce backroom power politics into the electoral process, which harms both public and candidate participation in the Nominating Convention.

Specifically, I propose creating a "Citizens Assembly on School Board Candidates" ("Citizens Assembly" for short). This Citizens Assembly would consists of approximately 150 randomly selected citizens stratified by geography and gender. Geographic stratification would be based on the boundaries of each of the twelve high school feeder systems within the County. Gender stratification would result in an equal number of males and females being selected.

Once constituted, the Citizens Assembly would act a lot like the School Board Nominating Convention. It would interview the candidates for several nights and then submit two recommendations to the governor.

However, there would be a number of significant differences. It would be more democratically representative. Membership in the School Board Nominating Convention can be easily stacked by advocates of a particular candidate and, because membership is self-selected, tends not to be representative of the County. The potential for democratic failure is exacerbated by an administrative committee for the Nominating Convention that isn't subject to standard governmental due process rules and checks & balances.

Unlike the Nominating Convention, the Citizens Assembly could be given formal government powers. In a democracy, it is not reasonable to give formal governmental powers to an unrepresentative body such as the current School Board Nominating Convention. It must therefore remain a private entity whose decisions are not binding on the governor. However, a democratically representative Citizens Assembly would suffer no such legitimacy deficit, thus allowing it to be granted powers not otherwise consistent with democratic principles.

The Citizens Assembly also has similarities to three other widely known democratic institutions. In a presidential debate with a randomly selected audience, a private organization such as the Gallup Poll, under the supervision of the Presidential Commission on Debates, randomly selects several hundred individuals who then have an equal chance of asking a question to the presidential candidates before a national audience. This debate format, although extraordinarily expensive, has proved very popular with the American public. If it wasn't so expensive (e.g., bringing hundreds of people from across a wide geographic expanse to a single location is very expensive), it would probably be much more widely used. Unlike this presidential debate format, the Citizens Assembly would draw from a highly local community and thus have much lower transportation costs. It would also have formal electoral powers rather than serving as merely a vehicle to inform the general public.

Judicial juries are perhaps the best known democratic institution based on random selection. Jurors are randomly selected from the general pool of citizens and take on a huge responsibility on behalf of the general public. Unlike the judicial jury, the Citizens Assembly is much larger and therefore more representative. The maximum time commitment is much less, participation is voluntary, and interested parties cannot disqualify potential members.

Local and statewide presidential party caucuses are perhaps the best known democratic institution where a sample of the general population is given formal powers to nominate candidates on behalf of the general public. Unlike citizen assemblies, caucus members are restricted to members of a state recognized party and are not randomly selected. If a party doesn't win a certain percentage of votes or other indicator of widespread political support, it cannot automatically get its candidate's name on the general election ballot. In addition to caucus nominees, independent candidates with a sufficient number of petition signatures can also usually get their names on the ballot.

The Detailed Proposal
Here is a more detailed model of my proposal. Note that some of these recommendations could be incorporated into the current school board nominating convention.

A judge from the Anne Arundel County Court System would be asked to moderate the deliberations of the Citizens Assembly/Candidate Jury. This is akin to the chief justice of the Supreme Court moderating a presidential impeachment trial in Congress.

The County's jury office would be responsible for conducting the initial stratified random selection, sending out the invitations, processing the returns, and conducting a follow-up stratified random sample. Stratification involves dividing a population into homogeneous subgroups. A stratified random sample involves taking a simple random sample in each subgroup. The County's jury office is already set up to do this type of work, with the major exception that its subcontractor's computer would have to be reprogrammed to do the type of stratified random sample recommended here.

Selection of the Citizens Assembly membership would take place in several phases. An initial round of invitations would be sent to a stratified random sample about fifteen times larger than the final, expected size of the citizens assembly. From those that expressed interest in participating, a second round of stratified random invitations would be sent. Those that remained at the end of this round would constitute the Citizen Assembly├ó€™s membership. Additional details of the selection process could be modeled after the successful citizen assemblies recently conducted in British Columbia, Canada; Ontario, Canada; and the Netherlands (see www.jhsnider.net/citizensassembly/ for more information). The use of stratification and multi-round sampling is desirable because the response rate of those invited to participate in the Citizens Assembly is expected to be fairly low. The use of stratification ensures that the final sample is representative of the larger population in key characteristics, including geographic distribution and gender.

Members of the Citizens Assembly, like members of traditional juries, could not be personally lobbied. All communications would have to go through the public website. The judge would be responsible for dealing with improper attempts to influence members of the Citizens Assembly. Shielding participants from personal lobbying both saves them time and preserves the integrity of the process. An alternative would be to allow personal lobbying but require a report to be filed on a public website concerning the substance, date, and participants in the information exchange. This is the method taken by the Administrative Procedures Act, which regulates lobbying of federal agency personnel. But it is more cumbersome than the outright ban I propose for in person lobbying of the Citizens Assembly's members.

The Citizens Assembly would be televised, including delivery via webcast and public access cable TV. This would enhance both the prestige and accountability of the Citizens Assembly as well as provide valuable information to the public about school board candidates and issues facing the school system.

The deliberations would be stored on the County's website (or public access video server) and available via the Internet for a minimum of twenty years. This is now very inexpensive to do and would enhance the accountability of the school board as well as provide a valuable historical record.

A public website with discussion forum would be set up to facilitate public comment. See http://www.aacstudents.org/ for a possible model. This could be administered under the auspices of the countywide CAC or County Board of Elections. The members of the Citizens Assembly should not be deprived of the valuable information possessed by the general public, but it is vital that, as with a regular jury, the information is equally available to all so that incentives for lobbying and corruption are minimized.

A similar private website with discussion forum--akin to a jury room--would be set up for private communication among the citizens assembly members. This would probably be best handled by the County Board of Elections.

The candidate forums would take place in the County Council chamber. The chamber is centrally located, infrequently used, has all the requisite facilities, and would add prestige and legitimacy to the process. Like Anne Arundel County jurors, all Citizens Assembly members would be given free parking on the days of their service.

The Citizens Assembly would meet a total of three times: twice to interview the candidates and a third time to deliberate internally and vote. This is one less meeting than the current Nominating Convention. As the burden placed on members of the Citizens Assembly goes up, participation would likely go down, so great effort should be made to minimize that burden.

The Citizens Assembly would use preference voting (also known as the Single Transferable Vote or Instant Runoff Voting) rather than first-past-the-post voting. In preference voting, voters rank their preferences (i.e., 1, 2, 3...). In first-past-the-post voting, voters give candidates a simple yes-no vote. While harder to tally, preference voting eliminates incentives for strategic voting and the need for time consuming runoff votes. Strategic voting occurs when a voter misrepresents his or her sincere preferences in order to gain a more favorable outcome. An example would be a voter preferring a third party candidate over the two major party candidates but voting for one of the major party candidates so as to not waste a vote on a candidate perceived to have no chance of winning.

All seats would be at-large. It is more efficient to ensure representative selection via a stratified random sample than via the currently muddled mandate to artificially restrict candidate selection by geography and then expect the winners to represent countywide interests.

The County's Board of Elections would be responsible for administering the voting process. This is an activity where they already have a lot of expertise and a reputation for impartiality.

Future Enhancements
The initial goal should be to model the Citizens Assembly as closely as possible after the Nominating Convention. A major reason for this is that the more an electoral system is altered, the harder it is to both explain and implement. However, I would also recommend a number of additional modifications after the Citizens Assembly was established and proved workable:

The role of the governor in the electoral process would be eliminated. The Citizens Assembly would make the final candidate selections rather than just nominate candidates. If the Citizens Assembly proved competent and democratically representative, there would be no reason for the governor to be involved in the process.

The term of election would be reduced from the current 5 years to the 4 year cycle more common in Maryland. A four year term would increase accountability without creating an undue burden on those seeking re-election.

The number of elections would be reduced to two, staggered on two year cycles. During one election four members would be selected; at the other, three members. Fewer elections makes each one more important, which motivates citizens assembly participation, media coverage, and public participation. It also opens up the possibility of proportional representation (see next item).

The at-large elections would be based on proportional rather than winner-take-all voting. The great majority of democracies in the world (although not in the U.S. and most former British colonies) use proportional voting, which is more efficient in translating voter preferences into legislative seats so that a legislature represents the diversity of public preferences, including minority ethnic and geographic preferences, within a community. With winner-take-all elections, a substantial interest that is nevertheless a minority may be completely excluded from a legislative body. At a local level, a variety of U.S. communities and school systems have recently switched to proportional voting because it is perceived to be fairer. See Fairvote.org for details.

Conclusion
Did the ancient Athenian Solon perfect Athenian democracy or the Revolutionary Era Madison American democracy? The answer is no. Democracy is a work in progress requiring the same creativity and flexibility in design as building any other great institution. I apologize for burdening my readers with all this complexity, but electoral design is like composing a piece of great music. Getting even one important note wrong can destroy the whole project.