Can Labour Win?
| Date | 01 April 2010 |
| Author | Paul Whiteley |
| Published date | 01 April 2010 |
| DOI | 10.1111/j.2041-9066.2010.00005.x |
| Subject Matter | Feature |
As is well known, winning a general election in Brit-
ain does not mean capturing a majority of the votes, but
rather a majority of the seats in the House of Commons.
trivial though this may seem, on two separate occa-
sions since the second world war the party that won the
most votes did not win the most seats. This happened
in the election of 1951, when Labour won more votes
than the Conservatives but gained fewer seats and so
the Conservative Winston Churchill returned to power.
The second occasion occurred in February 1974 when
Edward Heath’s Conservatives won more votes than Ha-
rold Wilson’s Labour Party, but since the latter won most
seats it formed the government. For this reason fore-
casting models concentrate on identifying the number
of seats likely to be won by the different parties in the
House of Commons.
The election forecasting model is based on a relation-
ship between votes and seats fi rst identifi ed in 1909 in
evidence given to the Royal Commission on Systems
of Elections. It was introduced by James Parker Smith,
Can Labour Win?
The Conservative Party set a post-war record by
winning four successive elections from 1979 to
1997. This achievement raises the question of
whether or not Labour can equal that feat by winning
again in 2010, or whether Gordon Brown will be re-
placed as prime minister by David Cameron. Predicting
elections is a lot like forecasting the weather – there
are a huge number of variables to account for but even
then the possibility of an ‘act of God’ can never truly
be discounted. Election forecasting is something that
interests many people, particularly MPs for obvious
reasons, but just as many are sceptical about whether
it is any use. Like meteorology, there are many different
ways and models to forecast elections, but the most
reliable method, Seats–Votes, has proved a very useful
bellwether for voting patterns in recent general elec-
tions in the UK. So what does the Seats–Votes method
predict this time around? Will the upcoming election
decimate Labour in the same way that 1997 almost
wiped out the Conservatives? Or is there hope yet for
embattled premier Gordon Brown?
Types of Model
Forecasting models come in three types: fi rst, there is
election night forecasting, sponsored by television pro-
grammes. This work is driven by the desire to predict
the outcome of an election on the basis of a few early
returns, and is really about using small samples of con-
stituency results often supplemented by exit polls to in-
fer the result. The second type is based on opinion polls,
which look at the relationship between political support
for different parties and other measures such as the state
of the economy over time. In this case, predictions are
based on measures such as the economic ‘feel-good’
factor – if the economy is doing well then the governing
party is generally rewarded by increased support, and if
it is doing badly the opposite happens. These relation-
ships can be used to forecast political support and are
often referred to as ‘popularity function models’. The
third approach, which is the one discussed in this article,
uses voting data from actual elections measured over
time to estimate future voting trends. The last of these
is the least common type of approach and is referred to
as the ‘vote function model’.
For months opinion polls have been predicting a Conservative victory in the next general
election. But is David Cameron certain to be the next prime minister or could Gordon Brown
snatch victory from the jaws of defeat? Paul Whiteley forecasts the upcoming election – and
comes up with some rather surprising results.
The Seats-
Vote method
has proved
a reliable
bellwether for
recent general
elections –
what does
it predict
for 2010?
Pool/Reuters/Corbis
5April 2010
To continue reading
Request your trial