US elections, and their amazing effect on UK shares by Richard Dyson

Long-term investors are unlikely to alter course simply because an American election comes along.

Of course, some sectors will suffer ahead of a poll if an outcome is expected – or achieved – that threatens to damage that sector.

But aside from that arguably rational market response, there are the inexplicable patterns in share price movement that appear to occur around major, regular events (or indeed seasons), including elections.
“In 82pc of US election years, the UK stock market rises – by an average 33pc”Stephen Eckett, The Stock Market Almanac

These are the anomalies that seem to confound the theory of efficient markets.

In Britain the bible of such market anomalies is Stephen Eckett’s annually updated Stock Market Almanac.

It gleefully dishes up statistics to suggest that markets, far from efficiently responding to known information, seem to take little account of such information and are instead reliably seasonal or event-driven.

What history tells us…

American elections fall every four years and occur in early November.

Mr Eckett has taken daily FTSE All Share data relating to American elections since 1972 (Richard Nixon’s second victory). He obtained returns for the four days ahead of the election day, the day itself, and the four after – and identified a pattern.

[chart removed]

The average returns for days one to four (before the poll) were all positive, with the lowest average being one third of one per cent.

The average return on day five (polling day) was a positive 0.64pc. After the elections, though, returns turned negative in three of the four days. Only one (day eight) averaged a positive return, and then of a mere 0.29pc.

The consistency is striking and, as Mr Eckett observed, suggests that the British market “trades stronger as the election approaches and then tails off”.

There is a peculiar pattern for the British market during American election years as a whole, too.

Mr Eckett has found that in 14 out of 17 elections years since 1948 the markets here have risen, “with a rather extraordinary average annual return in those years of 32.7pc”.

That is extraordinary, given that the average annual gain of British shares over a century or so – generally regarded as the longest span of reliable data – is a mere 9pc.

Mr Eckett offers no explanation, and that’s not really the point of his book.

But what you can see if you look more closely at the data is that some absolutely stonking gains were made in some years, such as Nixon’s win (1968); Reagan’s 1980 and 1984 wins, and Bill Clinton’s wins of 1992 and 1996.

If this year is going to conform to the pattern it will need to get going. Since January 4, the All Share is up by just 7pc.

And what the market foretells …

In a different vein, others are looking to the market to predict what the outcome in November will be.

Pimco, the American bond fund manager, has looked at the performance of the S&P 500 index for three-month periods ahead of every election since 1928.

It finds that in 19 of these 22 elections, returns have correctly predicted whether the incumbent party will remain in power or be ousted.

Positive returns for that period indicate that the incumbent stays put. That should not be especially cheering news for Democrat Hillary Clinton as the S&P 500 was at 2,183 on September 8 and is now lower at 2,140.

Nine weeks to go …

By other predictive measures, however, her prospects are brighter.

Pimco said that “for the past 50 years if the sitting president had an approval rating of 50 or above, the party in power has remained in power.”

Mr Obama’s rating is 54.