Sunday, November 18, 2012

The polarisation of the US electorate

Something odd has been happening in US politics.

As presidential elections have become closer nationally, most individual states have become more and more firmly aligned with one or the other party. The Republican vote and the Democrat vote have been segregating, leaving fewer and fewer swing states.

I don’t know why this has been happening, nor whether it will continue, but here are the figures.
This chart shows how many of the 51 states (including DC) have voted the same way in the last four (or more) consecutive elections:


And this chart shows how many of the 538 electoral college votes are held by those states (based on each state’s current electoral vote):


Either way, it’s high: 80% of states, with 78% of the electoral vote, have not swung since 2000 or earlier. And, while the Republicans had a run of good results in the 1970s and 80s, today’s peak is made of a bipartisan mix. The recent rise in reliably Democrat states has come mostly by reducing the number of swing states rather than the reliably Republican ones.

If we tighten the criterion to look at states that have voted the same way for six elections or more, it’s much the same picture:


Going back through Obama, Bush and Clinton, past Romney, McCain, Kerry, Gore and Dole, 64% of states, with 65% of the electoral vote, have not changed sides.

And the winning margins are rising, too. This chart show how many states have given the same party a lead of 15% or more – averaged over four elections:


Here it is in electoral college votes:


But this is happening even as presidential elections have become closer nationally. This final chart shows the national margin of victory in the popular vote at each election and the average state margin of victory (for whichever party wins it):


To start with, the two lines go up and down together, as you might expect. In landslides like Johnson’s in 1964 or Nixon’s in 1972, lots of states voted heavily for the winner, and in closer elections like 1968 or 1976, more states were narrowly won (by both sides).

But since the late 1990s, this pattern has broken down. A series of narrow-to-middling popular vote leads has been matched by a steady but impressive rise in the average state margin of victory – shared, unlike previous highs, between both sides.

I don’t know why this has been happening or whether it will continue. But because election campaigns mostly focus on swing states, the polarisation of red states and blue states means that more and more Americans are getting less and less attention from their political leaders.

(Using data from here.)

Thursday, November 01, 2012

How much work are people getting?

How should we take part-time work into account when thinking about the employment rate?

The number of people in work is back to its pre-recession peak, which the government is unsurprisingly pleased with. But there’s been a shift from full-time jobs to part-time, which means that the total number of hours worked per week hasn’t risen so much:


Then, of course, you have to remember that the population is growing, including the large part of it that makes up the labour force (the people with jobs plus the people who are looking for jobs). So a simple rise in numbers isn’t the same as the rise in the employment rate – the rate has recovered a bit, but it’s still well short of its previous peak.

Putting these two thoughts together leads me to this chart. It shows the total number of hours worked divided by the total size of the labour force – in other words, how much work the average person who wants work is getting:


There has been pretty much no change since 2009. The rise in employment since the election averages out as an extra 2 minutes and 43 seconds of work per person per week. This compares with almost an hour and a half lost in the recession.

Data from the ONS.