Keith Rankin: A Minuscule Landslide
Keith Rankin, 14 June 2001
Last week I suggested that the result of the British election would be closer than most people expected. I was right,
the gap between Labour and the Conservatives was only nine percentage points. What I did not realise, however, is just
how much the first-past-the-post system distorts British elections. So I did some sums, tried out the BBC's "Virtual Vote" tool, and made some comparisons with New Zealand.
Firstly I should note that there are different interpretations of the low turnout. New Foreign Secretary Jack Straw
claimed that the low vote was evidence of additional support for Labour. I cannot accept this. British voters in safe
Labour seats have turned out in the past, knowing that their vote makes no difference to the final result. It was in
many of these seats that the turnout was very low this time, suggesting that huge numbers of former Labour voters were
rejecting Labour while refusing to endorse anyone else as being any better.
The actual results, shown below (and in this graph), show that Labour was only marginally more popular in 2001 than in 1987 when Margaret Thatcher ruled the roost. In
1987, the margin between Labour and the winning Conservatives was 11.5%. In other words, with essentially the same vote
in 1987, Labour was at the losing end of a bigger landslide than the hapless Tories were in 2001. In 1992, Labour got
more votes than in 2001, despite suffering a massive defeat.
British election results (United Kingdom excl. Northern Ireland)
1983
1987
1992
1997
2001 Labour
8,456,934
10,029,270
11,560,484
13,517,911
10,740,168 Conservative
13,012,316
13,760,935
14,093,007
9,600,940
8,355,267 LibDem
7,780,949
7,341,651
5,999,384
5,243,440
4,815,249
total votes
29,906,212
31,800,052
32,828,981
30,497,107
25,554,809 eligible voters
41,144,233
42,091,593
42,150,416
42,609,051
43,194,105
72.7%
75.5%
77.9%
71.6%
59.2%
Labour did not win this year's British election. It's just that Labour was the least unsuccessful loser. Democracy was
the loser on the day.
The table below (and in this graph) shows that less than one-quarter of eligible voters registered support for the Blair government. Landslide?
British election results: percent support of eligible voters by party
1983
1987
1992
1997
2001 Labour
20.6%
23.8%
27.4%
31.7%
24.9% Conservative
31.6%
32.7%
33.4%
22.5%
19.3% LibDem
18.9%
17.4%
14.2%
12.3%
11.1%
The Liberal-Democrats keep getting more and more seats, despite getting fewer and fewer votes. Nonetheless, they remain
comprehensively under-represented.
How does the support of Labour in Britain compare to that of Labour in New Zealand? Because New Zealand has different
voter registration procedures, the simplest way of making a valid comparison is to use each party's support as a percent
of the total population of the respective countries. The assumption is that the same proportion of each country's
population was eligible to register as a voter.
It turns out that Labour in New Zealand, which governs only as a part of a minority coalition, was, at the last
election, significantly more popular than Labour is in Britain at present. I don't recall anyone calling the 1999 New
Zealand election result a landslide victory for Labour.
New Zealand
Great Britain
1996
1999
1997
2001 population
3,755,400
3,829,800
57,000,000
58,000,000 Labour votes
584,159
800,199
13,517,911
10,740,168
15.6%
20.9%
23.7%
18.5% Nat/Con votes
701,315
629,932
9,600,940
8,355,267
18.7%
16.4%
16.8%
14.4%
We might also note that Labour in New Zealand in 1996 was significantly less popular than the British Tories who were
trounced in the following year. Yet despite a level of support barely greater than that for the British Conservatives
this month, NZ Labour, peeved that it did not become the Government in 1996, contributed little to the parliament of
1996-99. National in 1996 was actually more popular than Blair's New Labour was in this month's British election.
Despite getting less than a quarter of eligible voters' votes, Blair's Labour government could actually have retained
power with less than half of the support they actually gained. The first-past-the-post voting system is very uneven in
its favours.
When I entered the actual vote percentages into Virtual Vote (which uses 1997 results to predict electorates won) I got:
estimated % vote share
predicted seats
actual seats Labour
40.7
402
413 Tory
31.7
180
166 LibDem
18.3
49
52
other
9.3
28
28
So the Virtual Vote algorithm is quite accurate. It is reasonable to assume that the 14 excess seats gained by Labour
and the Liberal Democrats represent the effect of tactical voting in certain strategic constituencies.
When I reversed the Labour and Tory votes, Virtual Vote computed:
estimated % vote share
predicted seats Labour
31.7
288 Tory
40.7
307 LibDem
18.3
33
other
9.3
31
So if the Tories had gained a similar "landslide" margin in the popular vote to that gained by Labour, they would not
even have won the election. Labour and the LibDems, with a combined vote of 49%, would have formed a de jure or de facto coalition.
Next I tried it with Labour coming third, behind the LibDems and the Conservatives. This result could have happened if
most of the non-voters had voted for change.
estimated % vote share
predicted seats Labour
30
330 Tory
31
189 LibDem
33
109
other
6
31
It turns out that Labour would have "won" the election with an overall majority of the 659 seats, despite coming only
third in the popular vote. This possibility make's the 2000 Florida anomalies seem almost democratic.
Next I tried to see how far Labour's vote might have to fall in order to get a 3-way tie in terms of seats won. I got:
estimated % vote share
predicted seats Labour
20
208 Tory
32
210 LibDem
39
208
other
9
33
The Liberal-Democrats would have needed twice as many votes as Labour in order to become the senior partner in a
LibDemLab coalition. This is caused by the fact that LibDem voters (and to a lesser extent Conservative voters) are more
evenly spread than Labour voters throughout Britain. Labour voters are concentrated in the larger cities.
In my final experiment, I tried to find how few votes Labour could have got, and yet still be returned to the Treasury
benches. It turns out that Labour only needed 16% of the total vote to win the election; just 10% of the eligible vote.
It works like this:
estimated % vote share
predicted seats Labour
16.0
159 Tory
38.0
317 LibDem
36.5
145
other
9.5
38
The Tories could expect the support of just the 11 UUP and DUP MPs from Northern Ireland, a total of 328. The remaining
331 MPs would have most naturally joined or supported a Labour-led government.
It is no wonder that Blair's New Labour has gone soft on proportional representation.
Would Labour in New Zealand benefit from a return to a British electoral system that has proved itself even more
ridiculous than that of the United States? The answer is no. Support for Labour in New Zealand is more evenly spread
through the country than it is in the United Kingdom. In New Zealand, first-past-the-post has always distorted the vote
in favour of National: eg 1993, 1981, 1978.
So long as Labour is both a major party and a centre party, under MMP it will be the dominant party in government in New
Zealand, for most of the time.
Meanwhile, we must wait a few more years to see how the Brits react to Labour winning comfortably (in terms of seats)
while gaining fewer votes than the main alternative party. It will happen, but maybe not for a while given that, in the
hearts and minds of the British people, the Tories and the Liber-Democrats are regressing almost as rapidly as Labour
is. Only when a result as absurd as those I have shown actually occurs will they fully appreciate just how tenuous their
country's democracy is.
© 2001 Keith Rankin
email: keithr@pl.net
website: http://pl.net/~keithr/