This may not be news to you, but in November of this year Americans go to the polls to elect a new president. The first presidential election in the United States was held in 1789; there were no political parties then but the election pitted Federalists against Anti-Federalists (which somewhat echoes current divisions in the European Union) and there was even a candidate named Clinton involved. George Washington was elected president in that first election and was re-elected three years later in the 1792 election. Since that second election, presidential elections have been held every four years.
Anyway, enough of the trivia, why should investors be interested in the timetable of US presidential elections? In 1980 an academic paper† announced that the US economy “has been managed to expand prior to an election and contract thereafter”. To our modern, cynical ears the authors may seem rather naive in finding this worthwhile reporting – ie that politicians tweak the economy for electoral advantage. But it was obviously news at the time. The paper found that stock prices increased strongly in the two years leading up to the election, and performed weakly in the following two years. With this discovery of the four-year election cycle of stock prices the authors boldly claimed that “stock prices are not random”.
A number of studies followed that supported this initial finding. One such study found that the annual excess return of the S&P 500 was almost 10 per cent higher during the last two years of the presidential cycle than during the first two years. Another paper observed that most of the bear markets of the 20th century happened in the first year or two of a presidential term, and proposed the strategy of investing in the stock market in October of the second year of the presidential term and selling out in December of year four.
The research was further refined by identifying year two of a presidential term as being especially weak (often with negative returns) and year three as particularly strong – possibly the result of a bounceback from the weak previous year.
So, it would seem that there exists a four-year presidential election cycle in the stock market. The significance of all this for investors in the UK stock market is that as the monthly correlation of the UK and US markets is very high, phenomena that affect the US market tend to end up influencing the UK market as well.
In which case, is the presidential election cycle observable in the UK market? The chart below shows the average annual returns for the FTSE All-Share index for each of the four years of the US presidential election cycle (PEC). Two periods are studied (1920-2012 and 1948-2012) to check the consistency of the result.
From the chart we can certainly see that the second year has been the weakest and the third year the strongest. This is consistent with the PEC studies cited above. But the fourth year has been marginally weaker than the first, which contradicts the PEC theory. The third year strength is impressive; since WWII the FTSE All-Share has only fallen once in the third year of the PEC – that was in 2011.
The next chart plots the FTSE All-Share index (on a log scale) from 1956 with the dates of the US presidential elections indicated with vertical bars. From this chart it is possible to discern that the elections have sometimes coincided with significant turning points in the market.
Reasons for the PEC
The explanation for the PEC effect on shares in the US is that the president forces through difficult, and quite possibly unpopular, legislation in the first couple of years of his term. Then the president’s party traditionally loses the mid-term elections, after which the president is concerned with regaining popularity. Monetary and fiscal policies are eased to promote economic growth and lower unemployment. And research has found that the Federal Reserve policy has been more accommodative during the third year of a president’s term.
But this seemingly reasonable explanation is not unchallenged. In 2011 a study found that the presidential election cycle in stock returns and the government’s economic policy influence on stock returns are two separate phenomena. And that, if anything, it was more likely that stock returns were influencing economic policy than the other way around. And last year another paper supported this view that the PEC effect could not be explained by politicians employing their economic influence to remain in power.
Small Democrat and Big Republican
An obvious question to ask is whether the stock market is affected by whether a Democrat or Republican is in the White House. A widely cited paper in 2003 found that the excess return in the stock market was significantly higher under Democratic than Republican presidencies. A possible reason for this was offered in a 2011 paper that argued that this can be explained when risk is properly taken into account: Democrat presidencies are associated with higher market and default risk premiums than their Republican counterparts. However, another paper in the same year had the more prosaic argument that during Democratic presidencies companies with high government exposure experience higher cash flows and stock returns, while the opposite happens during Republican presidencies.
The chart below illustrates whether an equivalent ‘Democrat premium’ effect exists in the UK market. The chart compares the performance of two portfolios, each of which invests in the FTSE All-Share only when a Democrat (or Republican) is in the White House and is in cash for the rest of the time. Starting in 1948 the Democrat portfolio would today have a value about twice that of the Republican portfolio, which would support the Democrat premium theory. It should be noted, though, that the performance of the two portfolios was roughly equal in 2000, and then George W arrived (poor chap, not his fault perhaps but his term was bracketed by the dot-com crash and then the credit crunch).
His presidency wasn’t the worst over this period for investors, though: that accolade goes to Richard Nixon.
The Democrat premium theory isn’t accepted by everyone, or at least some people argue it needs to be qualified. Several studies found no significant difference in stock market performance between Republican and Democratic presidential administrations for large-cap stocks. However, the story is very different for small-caps, which performed significantly better under Democrat administrations – enjoying an annual return difference of more than 20 per cent. Conversely, it was found that the debt market performs significantly better during Republican administrations. This led to the development of a strategy proposed in a 1995 paper of investing in small-cap stocks during Democratic administrations and either intermediate-term government bonds or large-cap stocks during Republican ones.
In 2010 fund manager Mebane Faber suggested a strategy to combine features of the presidential election cycle and the ‘January effect’ (this latter effect holds that small-caps outperform large-caps in January). He wrote that historically across the 48 months in the four-year cycle the January of year three is the single biggest outperformer, with median returns since 1927 of nearly 8 per cent a month for small-caps. Small-cap ETFs could be used to exploit this or, if you wanted to turbo-boost that strategy, according to Faber small-caps that are significantly down from their highs in mid-December respond the best to the January effect.
A Congressional Effect
There is another effect that is interesting to mention. As far as I know it was first discussed in a 1997 paper which reported that almost the entire advance in the US stock market since 1897 was made during periods when Congress was in recess. Average daily returns when Congress was not meeting were almost 13 times greater than when Congress was in session. The research was updated in a new paper in 2006, confirming that stock returns are lower and volatility higher when Congress is in session. This latter paper concluded: “This is consistent with a mood-based explanation that sees Congress as ‘depressing’ the average investor.” This is a sentiment that might be shared by UK investors when the House of Commons sits.
†Fred C Allvine, Daniel E O’Neill, ‘Stock Market Returns and the Presidential Election Cycle: Implications for Market Efficiency’ (1980)
The outlook for 2016
This will be the fourth year in the presidential election cycle. Supposedly, this year is part of the strong second half of the presidential term, although as we see in the chart on page 32, the average annual return for the FTSE All-Share index since 1920 has been 5 per cent – not nothing, but less than years one and three. And, for completeness, the index has seen positive returns in 61 per cent of those PEC fourth years since 1920. So, a mildly positive but not stellar historic precedent.
At the stock level, there are just three FTSE 350 stocks that have risen in every presidential election year since 1996, they are: Capita (CPI), Provident Financial (PFG), and Amlin (AML). All three saw average annual returns in these years of more than 20 per cent.
Of course, there’s no guarantee that any of the above will hold true if there is a president in the White House called Trump
Free 2016 Almanacs
We have 10 FREE copies of Stephen Eckett’s UK Stock Market Almanac 2016 to give away. We’ll be drawing the names out of a hat so simply send an email to rosie.carr@ft.com with the word Almanac in the subject line for a chance to win a copy.
Readers can also purchase a copy of the Almanac 2016 by Steve Eckett for the SPECIAL PRICE of £20 INCLUDING UK P&P for the print edition.
SImply enter the following promo code at the checkout to receive your discount – ICALMANAC16. To order, visit http://www.harriman-house.com/book/view/5911/investing/stephen-eckett/the-uk-stock-market-almanac-2016/.
