As the closing days of the United Kingdom election campaign have focused on the economy, I thought I would repeat the analysis that Theodore Gray did on Dow Jones returns under United States presidential parties—but using UK data.

I started by going to an interactive Mathematica Demonstration that Theodore wrote. Like all Demonstrations, it doesn’t just present information, it encodes the analysis, so by downloading the source code, I was able to re-deploy it on UK data quite quickly. The data was a little more difficult (detailed at the end of this post).

So what did I find?

Well at first glance, it is a strong win for the Conservative party. If you started with £10,000 just after the last Liberal government in the 1920s and only invested in Conservative government years, you would end with £397,000, compared to £59,000 if you invested only in Labour years.

Unlike in the US case, modest assumptions about the lag time before policies take effect do not fundamentally change the conclusion. If anything they make the difference greater by putting the blame for the problems in the early 1970s onto Labour’s Harold Wilson.

You need to assume a lag of nine years before Labour edges ahead in returns.

In the UK case, though, it is not a steady story. Little of the returns are in the early years while the UK struggled with post-war reconstruction of its economy, right through to the 1970s’ times of inflation and industrial action. The big returns start in the 1980s. So in the end, this is largely a comparison of returns during the Thatcher/Major Conservative governments against the Blair/Brown Labour governments.

And just as in Theodore’s analysis of US markets, the best return has been to invest steadily regardless of who is in charge.

Notes on the data used:

Unfortunately, the FTSE 100, the UK equivalent of the Dow Jones index, only goes back as far as 1984, so the rest has been filled in from “‘Excess Volatility’ on the London Stock Market, 1870–1990” (De Long and Grossman, 1993). Having learned the lessons of Theodore’s analysis, I have only looked at real returns (taking into account inflation), including dividend payments. This is particularly important for the older data, when dividends were pretty much the whole reason for owning shares and source of all the returns.

The remaining problem was that I only had annual data, so rather than rewrite Theodore’s code I took the approximating step of generating monthly data by assuming steady returns throughout the year. This means that very sharp shocks to the markets have been smoothed over the year. This should not be important unless the shock happened around the end of a government (allowing for the delay that you have chosen).