Photo credit: Flickr/Berit Watkin

Members of Parliament accurately predict who becomes Prime Minister

Vote intention polls performed poorly at predicting the 2015 British General Election.  What are the alternatives?

Instead of asking citizens how they intend to vote, we might want to ask them who they think will win.  I explore the accuracy of this approach at the LSE General Election 2015 Blog.

Similarly, we also might want to measure what Members of Parliament think the chances are that their party leader becomes Prime Minister.  I explore the accuracy of this approach in a recent article in Research & Politics.

Why should MPs be good at forecasting who wins?

MPs have the means, motive, and opportunity to select a party leader with the highest electoral appeal.

The MPs know the leadership candidates like nobody else does (“means”).  The candidates are MPs themselves and so selectors and candidates observe each other on a daily basis.

The MP’s future career depends on selecting a party leader with the highest electoral appeal (“motive”).  A weak party leader is likely to lead to fewer votes at the next general election, putting the MP’s job at risk.

Finally, since 1963 both the Conservatives and Labour allow MPs to formally vote in the party leadership election (“opportunity”).  The number of votes for the selected leader should indicate her electoral appeal.

The Party Leadership Model

To translate the voting behaviour of MPs at party leadership elections into forecasts of general elections, I calculate the difference in vote shares between the selected leader and her main contender (“performance”).  I then compare the performance of Conservative and Labour party leaders, and predict the leader with the better performance to become Prime Minister.  I term this process the “Party Leadership Model”.

Are MPs actually good at forecasting?

The Party Leadership Model accurately predicts who becomes Prime Minister.  The model picked the right candidate in 8 out of 10 past elections between 1966 and 2010.  It correctly predicted, for instance, the victories of Thatcher in 1979 and of Blair in 1997.

According to a Bayesian analysis, there is a 95 per cent probability that having the larger margin in party leadership elections increases the chances of winning the General Election.

These accurate forecasts were made a long time in advance.  The model forecasts with a lead time of about four years on average.

Overall, these findings suggest that MPs as a group are very good at forecasting who becomes Prime Minister.

Are MPs better than vote intention polls?

MPs even achieve a higher forecasting accuracy with less information than vote intention polls collected around the time of the party leadership election.

The most accurate “poll of polls” correctly forecasted five out of seven general elections, using the vote intentions of about 24,000 respondents on average.

In contrast, the Party Leadership Model correctly forecasted six of these seven general elections using the votes of about 600 MPs on average.

This finding suggests that the vote of one MP in a party leadership election is more informative than the vote intention of forty survey respondents.

How well did the Party Leadership Model predict the 2015 General Election?

The above results were all based on ex post forecasts.  How well did the Party Leadership Model perform in its first ex ante forecast, the 2015 British General Election?

It correctly forecasted that David Cameron would become Prime Minister again.

In 2005, David Cameron was elected as Conservative party leader, 16.7 percentage points ahead of his main contender, David Davies.  In 2010, Ed Miliband was elected as Labour party leader, 6.8 percentage points behind his brother, David Miliband.  Because David Cameron (16.7 per cent) was more popular than Ed Miliband (-6.8 per cent), the Party Leadership Model correctly predicted a re-election of David Cameron.

By using leadership contests the Party Leadership Model made this forecast with data from four-and-a-half years before the event.

This finding suggests that the two party leadership elections from more than four years ago were more informative than the vote intention polls published in the days before the 2015 General Election.

About Andreas Murr

Andreas Murr

Andreas is Lecturer in Quantitative Methods in Political Science at the Department of Politics and International Relations. Andreas specialises in quantitative methods, particularly in Bayesian statistics and hierarchical models. His substantive research focuses on electoral behaviour, including models of decision making and election forecasting.

Andreas Murr @ University of Oxford

Check Also

Catalonia’s quest for independence: Plain dead or alive and kicking?

The outcome of the recent regional election in Catalonia (which was held on September 27th) ...

Leave a Reply