Liberty Tree partner FairVote organized a group of U.S. citizens to observe the Scottish elections last month, where they used a proportional representation voting system for the first time. Below is their full report.

On May 3rd, Scotland held groundbreaking elections for its regional parliament and for local government, using two different proportional voting methods. As a result of these new, fairer methods, the Scottish National Party (SNP) ousted the Labor Party from power in the parliamentary vote and with other opposition parties gained major ground in local elections. At the same time, however, a sharp rise in invalid ballots and delays in the count caused a storm of controversy.

I was part of a 25-member delegation of civic leaders, city councilors and election officials organized by FairVote and the British Electoral Reform Society that observed the elections and attended pre-election and post-election briefings on redistricting and election administration in Britain. We need more such delegations, as there is much we can learn from the experiences of other advanced democracies as they work to reform their election practices.

A lot went right in Scotland. Since the first elections for the newly-created regional parliament in 1999, the government has been elected by a “mixed member” proportional voting method pioneered in Germany and recently selected by a citizen’s assembly in Ontario, Canada for a referendum this fall. Voters cast two votes: one for a district representative, like our Members of Congress, and one for the party they would like to see in government. Within local regions, the party votes are used to elect candidates to balance the results from the districts to provide overall proportionality. Suppose a region has 20 total seats, with 12 districts and 8 list seats. If a party wins 7 district seats and 40% of the list vote, it will win one list seat to end up with a total of 8 seats (40%) from the region. If a party wins no districts and 10% of the list vote, it wins two seats (for a total of 10%) from its party list.

The result has been fair representation of Scotland’s political diversity, with four parties regularly winning more than 15% of the vote and smaller parties like the Greens winning significant shares of seats and votes. The system helps promote women, who comprised more than 40% of the initial parliament in 1999. Government has been steady, with the Labor Party forming coalitions with a natural ally, the Liberal Democratic Party, after the 1999 and 2003 election. In 2007, voters made a sharp turn toward the Scottish National Party, giving it a narrow plurality of both the district vote and the list vote. Even with less than 30% of the list vote, however, the Labor Party would have won an absolute majority of district seats. Proportional voting provided more accurate representation and ultimately put the SNP into power. Required to share power with others, the SNP will not be able to impose independence, however, reflecting voters’ skepticism about such a move.

The elections for the 32 local governments drew particular interest from our delegation. Covering all of Scotland, although often connected to certain cities, these local governments have relatively large numbers of representatives elected on a partisan basis. Before 2007, they were elected by a winner-take-all, plurality voting method that gave a big advantage to the largest party. In Glasgow, for example, the Labor Party in 2003 won 71 of 79 seats. Still, voters’ support for different parties had resulted in a number of divided governments, with 19 localities forming coalitions after the 2003 elections. But the increase in the number of Scotland’s viable parties drew attention to how to provide fair representation, and two separate commissions recommended adoption of a candidate-based system of proportional voting called choice voting in the United States and the single transferable vote in Scotland. As with instant runoff voting, voters rank candidates in order of choice. With three or four seats being elected in every district, it takes about 25% strong support of voters to win a seat. Seeking to have as many voters as possible have their one vote help elect a candidate, a voter’s ballot will move from counting a first choice to a second choice if that first choice can’t win with a voter’s vote.

The results were highly representative of voter intent. Reflecting its general weakening in support and loss of “seat inflation” it had gained with winner-take-all, the Labor Party declined from 509 seats to 348 seats. The SNP doubled from 181 seats to 363 seats and the Conservative and Liberal Democrat parties both gained slightly to a total of 309 seats. Independents and smaller parties actually dropped from 234 seats to 202 seats, with a drop in independents (often disgruntled members of major parties) and an increase in minor party winners. In Glasgow, the Labor Party dropped down from 71 seats to 45 seats, about the share they would have won in 2003 under a proportional system.

Parties adapted to the system cautiously, often choosing to nominate only as many candidates as they thought might win (parties have complete control over nominations through private means, as is true in most nations). By accommodating increased voter choice, ranked voting methods should encourage more candidates, and parties that leaned toward “over-nomination” generally fared better by giving voters more chances to support their candidates. Because they often only ran one or two candidates in a district, parties didn’t nominate as many women as was hoped, but women candidates were 10% more likely to win than under the old plurality voting system.

Elections are important both for the kinds of campaigns and opportunities for dialogue they produce and for the representation and governance they provide. Scotland did well in both these measures, with lively dialogue boosting turnout over 50% (while far short of France’s recent 85% turnout, it was significantly higher than turnout in the midterm elections for our far more powerful U.S. House of Representatives) and with representation accurately reflecting voter preferences.

What drew initial media focus was the sharp rise in voter error and a delay in releasing results. The scene was all too familiar to our American delegation. Scotland – and all of Britain – handles election administration admirably. Mandatory for all citizens, voter registration is estimated to be well over 90% of registered voters, far higher than our 71%. Election officials are explicitly charged with activities designed to boost turnout rather than only administer the vote, while a national commission monitors and promotes best practices, funds elections and voter education campaigns and carefully reviews election administration to learn how to improve it in the future. Even in the most highly contested elections, election officials are nearly always accepted as absolutely non-partisan, in sharp contrast to the finger-pointing that follows many close elections in the United States. Voting was quick and efficient, with a long polling day (7 am to 10 pm), sufficient polling places and a short voting process.

But Scotland made some serious mistakes, and FairVote and the Electoral Reform Society joined an early call for a full inquiry into the root of these problems. The biggest surface problem was an tenfold increase in invalid votes in the parliamentary elections – a range of error initially reported as potentially as high as 10%, although ultimately determined to be about 4% in the district vote , 3% in the party list vote and less than 2% in the local elections. The reasons for that increase showed that Scotland and the United Kingdom as a whole still have important progress to make along a path similar to changes underway in the United States.

The root of the problem was changes in ballot design and instructions to voters. The situation was similar to that in Florida’s 2000 presidential contest, where new ballot designs such as the infamous “butterfly ballot” lead to high rates of voter error. In Scotland some major party leaders were concerned that the use of two separate ballot papers in the parliamentary elections, had helped small parties win more party list votes. They pushed to combine the district and list vote on one page, with the list vote first. The new ballot read at the top in big type “you have two votes,” with smaller type above each column indicating one vote to be cast for that column – a confusing set of instructions exacerbated by the fact that on ballots in some major jurisdictions, an important arrow to the “one vote in this column” line was not provided. At the same time, as previously mentioned, Scotland was using the new choice voting method for local elections for the first time. Adding to the confusion, Scotland had also just moved to a ballot designed for electronic counting. Some four million pounds was spent by the Scottish electoral commission to promote voter participation and to educate voters about voting procedures, but that education message wasn’t focused as much on change in the parliamentary ballot paper as other changes.

Our delegation identified the following areas as particularly problematic, all of which tied to problems with our own elections:

• Public review and testing of ballot design changes: The change from two ballots (where voter error was less than 0.5%) to a single ballot was reviewed by some voters in 2006, but apparently focused more on what they liked about it than on how well they used it. The change was not tested in a meaningful way with the actual candidates and parties. It should be a standard practice to test any such ballot changes thoroughly and seek public comment. Maintaining a version of the old ballot design on two pages without doubt would have kept invalid ballots down to previous low levels.

• Immediate overvote notification and ballot recording at the polls: Although Scotland moved to an optical scan counting system for the first time, it counted those ballots at a central location. Voters were not alerted to their overvotes, a problem the United States has now corrected for in-person voting in federal elections due to the Help America Vote Act. If Scotland maintains an optical scan system, it should invest in enough machines to allow voters an opportunity to fix errors and to record those ballots immediately at the polls, which also would avoid potential tampering with uncounted ballots before they arrive at the central counting place.

• Provisional ballots and voter registration changes: Scotland does not provide voters with the right to cast provisional ballots and the deadline for registration is several weeks before the election. While Scotland’s voter registration system produces much higher rates of registration than ours, it is still prone to some error, and voters should have the right to cast provisional ballots and also register closer to Election Day.

• More consistent standards: Scotland and the United Kingdom as a whole have a history of election officials for each parliamentary constituency having great leeway in how they register voters, organize Election Day and count ballots. One result is inconsistent standards across districts, and with Scotland’s new electronic counting procedures, we also saw inconsistent ways of judging the validity of ballots within the same central counting area. Election officials were under great pressure to produce a rapid result, but that involved scanning thousands of ballots and reviewing possible errors as quickly as possible, with teams of two counters reviewing the electronic record of each ballot. Their review was visible to observers on a screen behind them, but it happened very quickly and without clear coordination among different teams of counters. If Scotland does not invest in polling place scanners that can record ballots during the day, we would suggest slowing down the process of evaluating potential errors and having absolutely clear standards for what is and is not a valid vote that is the same across Scotland.

• Standards for commercial vendors: As in the United States, a for-profit, private firm received the contract to administer the electronic counting machines. But it was unclear how the company’s processes were evaluated and software certified, and it surprised our delegation to see so many company employees playing a direct role in operating the scanning equipment and participating in the evaluation of ballot validity. I would recommend testing procedures that gain full voter confidence and absolute public control of the counting process.

• Improved absentee voting procedures: A snafu with the printing and mailing of ballots led to thousands of voters not receiving their absentee ballots until far too close to Election Day, but they were not allowed to vote in person or cast provisional ballots. Better benchmarks for when to print and mail ballots are important any time absentee voting is allowed; interestingly, France received very high turnout without any absentee voting in its recent elections.

Our delegation has returned from the trip with far greater insight into ways of running elections better and providing fair representation – and how other nations can struggle with the same issues confronting Americans in our decentralized, often under-funded electoral process. Protection of the right to vote is absolutely fundamental to representative democracy, and we all have much to learn – and much to improve.

More info:

Areas of Focus:

Tags:

User Comments