Thoughts on politics and life from a liberal perspective

Thursday, 23 July 2009

How low can we go?

General elections in the UK used to be about two parties. Labour and the Conservatives. They used to get most of the votes between them. A few percent would go to the third party and "others" but the vast majority voted red or blue.

In this sort of situation, where you effectively have a two party system, the First Past the Post electoral system was perceived as working quite well. It often ensured that one of the parties had a majority (even though this was often on a minority of the vote) but it was often only a few percent difference and usually the party that won the election in terms of seats, got the most votes (although not always - February 1974 being a good example of this).

In 1951, 93.1% of voters plumped for red or blue. In 1955 it was 96.1%. The problem is that since the 1950s, the proportion of people who vote for one of the two main parties has declined. Here is a table showing the shares each party got for each election from 1951 - 2005 and the combined total (source - collated from Wikipedia pages on general election results):

As you can see, we have gone from 90 odd percent voting for Labour or Conservative in the 1950s to only around two thirds voting for both of thoise parties in 2005. Here is the data in graph form which makes the decline a bit clearer:

Now, before anyone accuses me of hypocrisy, I am well aware that I have complained before about graphs that don't start from 0 but there is a reason I have done it like this. Frankly, if support for the two main parties combined at a general election ever dropped below 50% then it would be game over for First Past the Post irrespective of whether (or probably even because) one of the two parties could form a majority government. This is a psychological floor to my argument which is why I have made it the bottom of the graph.

As can be seen from this graph, there is a bumpy but steady decline. You could draw a line of best fit through this which if continued would suggest that in another 20 years or so we could well hit that 50% mark.

As if to prove my point, looking at the latest "General Election Prediction" from Electoral Calculus which analyses opinion polls and electoral geography to come up with their best bet of how the votes will go at the next election their current prediction (from 7th July 2009) is:

So a combined predicted total of 62.34%. This would follow the line of best fit from the above graph perfectly showing a continuation of this trend of decline in support for the two major parties.

So what conclusions can we draw from this? I would say that we are becoming increasingly pluralistic as an electorate. We have got used to voting for smaller parties and those often described as "others". I think the vote share of the two main parties will continue to decline and it will get harder and harder for them to justify the present system.

At the moment, those in favour of First Past the Post have the power of incumbency with it being the existing system and inertia will generally militate against a change. However recent opinion polls have shown that people are generally in favour of a more proportional system and as that line of best fit comes down and down over the next few years I think things will flip. Proponents of FPTP will be pressured to justify why such a manifestly unfair system that disproportionately benefits Labour and the Conservatives should be allowed to continue.

UPDATE: Tom Freeman has done an excellent follow-up to my post here where he hypothesises what could happen if the vote shares for Tory and Labour gradually came down and the others gradually went up. He's got pie-charts and everything!


Matthew Huntbach said...

Even when it was almost a straight two-party system FPTP had the flaw of distorting geographical representation. There were few or no representatives of party A voters in areas of party B dominance and few or no representatives of party B voters in areas of party A dominance.

As a result, Labour came to be seen overwhelmingly as northern and inner-urban, the Conservatives as overwhelmingly southern and rural/suburban. Both parties were damaged by this, it created an image of the nation as more divided than it really was, and it left big groups of people in the country without a real voice.

FPTP sort of worked here because there was quite an even distribution of the votes. It produced disastrous distortions in other countries to where we exported it. The role of FPTP in wrecking African democracy at its birth really ought to be written up, starting with the damage done by over-representing the Nkrumahists and under-representing the Danquahists in Ghana's elections as it gained independence.

James Graham (Quaequam Blog!) said...

There is an interesting discussion about this phenomena in academic circles, which is summarised in Helen Margetts' chapter in Unlocking Democracy: 20 Years of Charter 88. In GB, the "Effective Number of Parties" was 3.4 in terms of votes and 2.3 in terms of seats. As the former number rises, which it looks set to do, the chances of us ending up with a hung parliament also increases. What's more the whole system begins to lose all credibility. The prediction is that when the ENP in terms of votes reaches 4, the current system will be said to have effectively broken down entirely.

Canada it seems is somewhat ahead of us and now seems to be stuck in rut of permanent minority government despite using FPTP. I like to think that if this happens here, our own politicians may be somewhat quicker on the uptake than the Canadians.

Unknown said...

surely it is only the conservatives that are committed completely to retaining FPTP

Duncan Stott said...

I do get excited when a blogpost has a graph in it. This certainly didn't disappoint.

There's absolutely nothing wrong with graphs starting from not 0.
What your criticism of Fraser Nelson's graph rightly did was show how misleading it is to plot two things with completely different scales on the same graph.

FloTom said...

I am afraid I do not see the premise of your argument. When the old Liberal party was the main Party instead of the Labour Party and it was going into a decline did they immediately change the voting system? Did democracy come to an end because they didn’t?

The FPTP system is not there to prop up one of the main parties or the other and if that were so we would still have a main party being Liberal. Your work is from a limited historical perspective and is therefore suspect.

James Graham (Quaequam Blog!) said...


The politics of the 1920s and 1930s are not comparable to the current situation. But yes, party politics did pretty much fall apart during that period, as did the economy.

Fundamentally though, while we saw the Liberals being gradually replaced by Labour during that period, we aren't seeing that now. We aren't moving from one system of two-party politics to another, we're moving towards genuine multi-party politics.

alice casey said...

An interesting example of how citizens think of voting systems and electoral reform more widely can be seen in the work of the Ontario Citizens Assembly. This was a public deliberation on reform across Ontario, Canada which engaged a number of ordinary citizens in deliberation on the issues. They came up with a number of very rational ideas about how to modify the system as had been able to take time to fully understand the significance and need for such reform. However, when a referendum then took place amongst the wider public who had not engaged with the issue, the motion for reform was defeated. Amongst the general public, there isn't much awareness of the implications of the current failing system, and understandably, there isn't much interest either! Yet when people have a chance to grapple with the issues, they do demand change, as the Ontario Assembly shows. Until more active demand from the wider public can be galvanized, we won't see reform this side of 2020.

FloTom said...

Really James on what basis do you say we are moving towards mulit party politics?

How do you know that we are not seeing the old partys namely Labour and Tory being replaced by others as you say happened with the Liberals. It begs the question did people think they were seeing the end of the two party system then?

There is no knowing also how long that "multi party" politcs that you claim we are moving into, withourt any evidence I might suggest, may last. Is it a response to just the expenses etc.

Even if we are moving towards multi party poliltics I fail to see how that should automatically mean a change in the voting system. Far from it if FPTP as you suggest can deliver multi party democracy why should we change it at all.

Arden Forester said...

It's entirely possible for the Conservatives and Labour to garner less than 50% between them, yet one of them "win" the election.

Tony Blair "won" with only 20% of the support of the total electorate. when the likes of Jack straw have this pointed out to them, they wince for a moment then forget about it.

Democratic opinion has changed in the UK. I am not sure I care for an Italian-style scramble for government legitimacy, but I find I can no longer accept a government that speaks for only a few and tries to ride roughshod over the many!

Harry Hayfield said...

Past general election performances (Con swings / incoming Con government)

1951: Lib -7% (3 losses from 9)
1955: Lib n/c (No change)
1959: Lib +3% (No change)
1970: Lib -1% (6 losses from 12)
1979: Lib -5% (1 loss from 13)
1983: Lib n/c (5 gains from 12)
2001: Lib +2% (6 gains from 46)
2005: Lib +4% (10 gains from 52)

LondonStatto said...

General elections in the UK used to be about two parties. Labour and the Conservatives.

And so will the next one, at least in terms of forming a government.

The question the electorate will ask is "Cameron or Brown?".

Tom Freeman said...

Hi Mark, nice post. I decided to take your line of thought and run with it - possibly a bit too far, but never mind.

James Graham (Quaequam Blog!) said...


Really James on what basis do you say we are moving towards mulit party politics?

On the basis of the 50 year trend that Mark outlines above - the transitionary period of the 1920s and 1930s by contrast happened relatively quickly.

On the basis that the decline in Labour/Tory support is not matched by an increase in Lib Dem support.

On the basis that the SNP and Plaid Cymru are currently in government in Scotland and Wales respectively.

On the basis that where we have PR elections in the UK, the Lib Dems don't do particularly well in them whereas minor parties such as the Greens and UKIP (and now the BNP) do.

Is that enough? I think I could go on.

London Statto:
The question the electorate will ask is "Cameron or Brown?".

By definition the "electorate" is not asking itself that question, or it wouldn't be voting for so many other parties - or not at all.

People seem to forget that the "electorate" is in fact millions of individuals who all have wildly differing opinions. To argue against the statistical evidence by claiming to know the hive mind of the collective electorate is just plain daft.

MatGB said...

Mark, the decline is more pronounced if you include the NatLibs with the Tory vote in 1951 as is done in 1955.

Definitely worth noting that the 1950/1951 elections were the first fought under modern FPTP--until then, many constituencies used what is I'm told technically referred to as "block voting" (multi-member FPTP). Admittedly 2 seats were the most common but still.

When the system changes (as happened with the 1948 ROPA), the party system within Parliament changes to accommodate--single member seats encourage two strong parties--within each district, the two parties might be different but generally it's true.

FloTom--the problem with FPTP as implemented by Labour in the 1940s is it accentuates geographic divides and creates large numbers of 'safe' areas where the other party just doesn't contest. It also encourages a large number of voters to choose between the top two candidates within each district, further distorting true preferences.

Was, for example, Michael Heseltine being more honest when he ran for election as a National Liberal or later on when he ran as part of the merged Conservative and Unionist party ticket? Wouldn't a system that allows candidates to be more honest with their labels yet still form effective Govts drawn from across the country be more useful to us?

Arden--that 20% stat is largely bogus. Sure, turnout fell, but it overwhelming fell in the 'safe' seats that were uncontested by any party--why bother voting in a safe seat, you know who your MP is going to be. Turnout'll be up next time, probably not up to 1992 levels, but up. The result is likely to be even more distorted though.