The US election shows why ‘equal boundaries’ will not fix the bias in the electoral system

Like, I suspect, most election geeks I was glued to the election coverage on Tuesday night as Barack Obama defeated Mitt Romney apparently winning with 332 Electoral Votes to 206.

US Presidents are elected by a system called the electoral college whereby every state gets as many ‘electoral votes’ as its number of representatives in Congress. For instance, California has 53 members of the House of Representatives and 2 members of the Senate, hence it has 55 EVs. Wyoming has 1 member of the House and 2 Senators and has 3 EVs. Washington DC also has 3 EVs, despite not being a state. EVs are awarded to the winner of the state on a bloc basis. So this year Obama won the most votes in California, and thus won all 55 of its EVs.

This system is similar to FPTP in many ways – there are safe states on both sides and key ‘swing states’ which essentially decide the election.

In 2000 the election was famously won by George W Bush with 271 electoral votes to 266 for Al Gore even though Al Gore won 48.4% of the popular vote to 47.9% for Bush. The electoral system has the capability to produce winners who have not actually won the electoral vote therefore, something also true of the British electoral system (though this has not happened since February 1974).

In the UK the electoral system suffers from electoral bias in favour of the Labour Party. In the 2010 the Conservatives won 7.1% more votes than Labour and secured 306 seats. If Labour had won 7.1% more votes than the Conservatives they would have won 354 seats, enough for a very solid majority government (indeed this would be a drop of only 2 seats from the last election).

This bias gives Labour a decisive electoral advantage.

Analysis by Nate Silver of FiveThirtyEight, the statistician and polling analyst whose model was proved extremely successful on election night, suggests that a similar issue may be emerging in the US. His analysis suggests that Mitt Romney may have had to win the popular vote by as much as 3% in order to win the electoral vote.

The reason is simple – demographic shifts in many states have advantaged the Democrats. Virginia, once a solid Republican leaning state has turned into a swing state thanks to many liberal, highly educated voters moving to the suburbs of Washington DC in the North of the state. States like Colorado have shifted towards the Democrats due to large influxes of Latino voters.

In the UK the government (though the Lib Dems have now said they’ll vote against the boundaries) has tried to fix the electoral bias through its plan for redrawing the constituency boundaries. They argue that the issue is that the seats are of an unequal size and that this benefits the Labour Party. However, equalising seat boundaries is far from the main cause of the electoral bias. According to analysis of the provisional boundaries by Anthony Wells of YouGov Labour would need a lead of 4.3% (compared to 3% now) to win an absolute majority under the new boundaries. The Conservatives would need a lead of around 7.4% (compared to 11% now). Some of this reduction in bias is also down to the new boundaries simply being more recent than the old ones.

The government’s plan for fixing the boundaries is deeply disruptive, with government minister Baroness Warsi criticising the constituency map as “mad and insane”. This disruption does not even achieve what it sets out to do because the plan does not appreciate that whilst some electoral bias does come from unequal boundaries it is far from the only source, and one issue is a demographic issue of our own.

Firstly, Labour voters tend to be from demographics (ethnic minority, young, poorly educated, urban based) that turn out less than Conservative leaning demographics. Hence Labour seats are won with less voters. Secondly, Labour seats tend to be won with smaller majorities than Conservative ones. Conservative seats – especially those in the South East tend to be won by huge margins, whereas Labour’s vote is better spread out (often described by political scientists as having a ‘more efficient voter spread’).

In the US, the boundaries cannot be changed – they are set according to the state boundaries, but any electoral system that uses boundaries will produce bias. In both cases, the simplest way to reduce bias would be electoral reform.

Our American partner organisation FairVote advocates the National Popular Vote Interstate Compact. Technically the states decide how their electoral votes are cast, and so the Compact, if passed, would have states award their electoral votes to the national popular vote winner, essentially turning the election into a single national contest. At the time of writing the Compact is around halfway to adoption.

In the UK, with our parliamentary system, the best way to reduce bias is to simply make elections more proportional. Electoral bias and disproportionality are not the same thing, but reducing the latter in this case will reduce the former as it is harder to create such bias across a larger constituency.

Our 2005 publication – Conservatives and the Electoral System makes this very point.