This post originally appeared on Politics in Spires.
Although their official report isn’t due to be published until the 23rd of April, on Wednesday the Guardian reported that the Joint Committee on Lords Reform has decided to back an ‘open preferential voting system’ rather than the government’s preferred Single Transferable Vote (STV) for electing members to a reformed House of Lords. Open preferential voting is a hybrid electoral system which allows voters to either express a rank ordering of individual candidates in the same manner as STV, or to select a list of candidates chosen by a political party, much like a closed list proportional representation system. In Australia, where the method is used to elect the Senate, this is known as voting ‘below’ and ‘above’ the line respectively due to the design of the ballot paper.
The Joint Committee’s decision has been strongly criticised by the Electoral Reform Society (ERS), who have campaigned for STV since 1884, as being ‘a dog’s breakfast’. The ERS’s main objection is that the introduction of above the line voting is that closed list PR systems would increase the power of political parties to influence who gets elected, which, in the words of their chief executive Katie Ghose, would result in the Lords being ‘a mini-replica of the commons, full of party-politicians with pre-set loyalties’ whilst ‘STV would help create a chamber free to provide objective scrutiny’. This argument is based on evidence from Australia that, given the choice, the vast majority (around 95%) of voters choose to vote above the line and so it is a de facto closed list PR system.
At first glance, research suggests that the ERS is correct in their argument: central party control over candidate selection and position on party lists tends to increase party unity in legislative behaviour. Further scrutiny however suggests that the picture is complicated by the fact that the proposed reforms would elect members to the Lords for single, non-renewable, 15 year terms. Research into party discipline, such as that by Christopher Kam, suggests that party discipline is driven by a combination of socialisation, positive incentives offered by party leaders for toeing the party line, and disciplinary measures party leaders can impose for breaking party discipline. One of the strongest incentives party leaders can offer in a closed list PR electoral system is a high place on the party list (and a corresponding better chance of re-election) and conversely leaders can punish rebellious MPs by moving them down or off the list. It is this aspect of closed list PR systems that explain why they tend to have high levels of party discipline. That members of a reformed Lords would be unable to stand for re-election effectively removes that effect from the equation, and there is no evidence to suggest that Lords elected under an open preferential system would be any more likely to follow the party leadership than those elected under STV.
The evidence of the very high use of above the line voting in Australia is also more complicated than it first appears. The electoral rules in Australia require that if you vote below the line you are required to give a preference ordering to every candidate on the list. At the 2010 election, there were 84 candidates for the Senate in New South Wales, so perhaps we should not be surprised that 97.76% of voters in NSW voted above the line – requiring a preference ordering of 84 candidates is a very high demand to make of voters. The evidence also suggests that the number people voting above the line is a direct consequence of the number of candidates standing in the election, at the 2010 election Tasmania and the Australian Capital Territory had only 24 and 9 candidates respectively, and the rates of above the line voter were much lower at 79.82% and 75.93%.
This raises two questions: how many candidates are likely to stand for election in each Lords electorate and how many preferences will voters be required to express? The draft reform bill proposes electing the 240 elected Lords in thirds, with 80 seats contested in each election, in constituencies of around 5-7 seats, with roughly one member for every 570,000 voters. Conveniently, this yields constituencies of roughly similar size to the number of voters in the Australian states of Victoria and NSW which elect 6 Senators each (except in double dissolution elections when they elect 12). At the last election these states fielded 60 and 84 candidates respectively. It would be rash to say that an elected Lords will definitely have an equally high number of candidates, but multimember constituencies with large electorates tend to have high numbers of candidates. Even if the constituencies of a reformed Lords had only a third of the candidates as their Australian counterparts, a full preference ordering would require voters to assess between 20 and 28 candidates.
The current proposal in the draft reform bill avoids this difficulty by only requiring voters to express as many preferences as they want to. Though this may seem like a reasonable suggestion and is the norm in STV elections, it is important to understand the consequences this has on voting behaviour. Evidence, again from Australia, though this time from state elections (where full preference listing is not required) suggests that the majority of voters will only cast a single preference. The result of this is a high number of votes being ‘exhausted’ (having no preferences left after the voter’s preferred candidate(s) have been eliminated) before the final count. In layman’s terms: they needn’t have bothered casting their vote at all.
If, as advocates of STV tend to be, we are interested in taking into account the preferences of as many people as possible in choosing representatives this is clearly problematic. An obvious solution to this is to encourage voters to express a certain number of preferences, or as is the case in Australia, require them to do so. However, this makes higher demands of voters than simply casting a single vote, which in turn leads voters making mistakes on their ballot paper. The introduction of above the line voting for Australian Senate elections was designed to overcome just this problem. Prior to its introduction in 1984, nearly 1 in 10 votes cast in Australian senate elections were counted as ‘informal’ (Australian political parlance for an incorrectly filled in vote that is then discarded). This was halved at the first election following the introduction of above the line voting, and at the last election only 3.75% votes were counted as informal.
It is not clear yet whether the Joint Committee will recommend some form of requirement to express a certain number of preferences, though given their advocacy for open preferential voting it would not be surprising if they did. Although it has problems of its own, as a method of overcoming some of the difficulties of STV it is worth serious consideration, and not simply dismissed as it has been by the ERS. Both of these options have costs but as students of electoral systems have known since Kenneth Arrow’s famous theorem, no electoral system is perfect. Ultimately the government will have to choose between large numbers of voters choosing pre-determined preference orderings by voting above the line or large numbers of voters having their preferences ignored because they did not express them properly on the ballot paper.
For a response from the Electoral Reform Society see here.