US Elections And Stock Market Performance

With the US presidential elections a little over two months away, what can we infer from the existing body of literature around US presidential elections and stock market performance? I’ve distilled some of the thoughts of academics, influential finance experts and leading political research organisations below.

Presidential Cycles

The US presidential cycle refers to a four-year period during which stock prices tend to fall in the first half of the presidency, reach the lowest point in the second year, rise again during the second half of the second year, and reach the peak in the third and fourth years (Gartner and Wellershoff, 1995).

Chart 1 uses data from the S&P 500 and DJIA going back to 1789. On the left (red) we see evidence that shows a consistent pattern in which the first two years of a presidential term have tended to produce below-average returns, while the last two years have been above average.

Looking over a shorter time span, and focusing on just DJIA data, does not remove the presence of a presidential cycle. Since 1833, the Dow Jones industrial average has gained an average of 10.4% in the year before a presidential election, and nearly 6%, on average, in the election year. By contrast, the first and second years of a president’s term see average gains of 2.5% and 4.2%, respectively (Kiplinger, 2016).

Indeed, Nickles (2004) says the following: ‘buying on 1 October of the second year of the presidential election term and selling out on December 31 of year four is a strategy that would have sidestepped practically all down markets for the last 60 years.’

Source: Fidelity

Chart 1 also includes average returns over the first two years (dark blue) and the full four-year presidential term (light blue). The statistics have been subdivided by partisan control. There appears to be a higher stock market return for Republicans in the first two years after the election, but little difference can be gleaned across the full four-year term.

While these descriptive statistics are informative, many other factors can influence stock returns during presidential cycles. There are four seminal academic research papers that attempt to control for spurious results attributed to higher systematic risk, time periods, business cycles and more.

Two find that Democratic administrations have the edge over the presidential cycle. Hensel and Ziemba (1995) look at investment returns between 1928-1993 and find that small-cap stocks had significantly higher returns during Democratic administrations. Meanwhile large-cap stocks had statistically identical returns under both Republican and Democratic election wins. As such, a simple investment strategy of investing in small-cap stocks during Democratic administrations and large-cap stocks during Republican administrations produced higher mean returns than did investing in large-cap stocks throughout the period.

The second paper is that of Santa-Clara and Valkanov (2003). This exceptionally thorough and well-cited research looks at CRSP data from 1927-1998 to examine stock returns across the four-year presidential terms. They find that average excess return of the value-weighted CRSP index over the three-month Treasury bill rate has been about 2% under Republican and 11% under Democratic presidents. This extends to 16% difference for an equal-weighted portfolio. Perhaps most interestingly, the difference in returns is attributable to surprises in the economic policies of the parties, rather than a ‘Democratic risk premium’.

Snowberg, Wolfers, and Zitzewitz (2007a) present evidence in favour of a Republican return premium. They analyse high-frequency financial fluctuations following the release of flawed exit poll data in 2004. They find that markets anticipated higher equity prices, interest rates and oil prices and a stronger dollar under a Bush presidency than under Kerry. Prediction market-based analyses of all presidential elections since 1880 also reveals a similar pattern of partisan impacts, suggesting that electing a Republican president raises equity valuations by 2-3%, and that since Reagan, Republican presidents have tended to raise bond yields.

Finally, we turn to Sy and Al Zaman (2011). They attempt to put to rest the ‘Presidential Puzzle’ – the suggestion that Democratic administrations command a premium because of markets mispricing policy decisions (found by the two papers above and as many as ten other academic papers). Covering the time period spanning 1926-2007, they suggest that there is no statistically significant difference in stock returns between Democratic and Republican presidencies after crucially controlling for differences in systematic risk across presidencies. In their opinion, the higher stock returns during Democratic presidents are related to higher market and default risk premiums during these presidencies.

US Election Year – Party Affiliation and Stock Returns

Next we turn to financial market reactions on US election year. Table 1 below looks back at stock market returns on election years going back to 1928.

Source: Morgan Stanley

The results point towards better stock returns on election years. Morgan Stanley make the following observations: there have been 23 elections since the S&P 500 index began. In these election years:

• 19 of the 23 years (83%) provided positive performances, with an average return of 11.3%.
• When a Republican was elected, the average return that year was 15.3%.
• When a Democrat was elected, the average return that year was 7.6%.

As was the case across the full presidential cycle, the magnitude of the equity increase is conditional on party affiliation. The early academic literature pointed to a Republican premium on election year. Niederhoffer et al. (1970) only had 18 presidential elections to work with since the beginning of the modern Dow-Jones. They find that the market is ebullient the week before election day: on 12 occasions the market rose, while it fell on five others (2.45% average). On the day following the election, markets rose on eight of the nine occasions a Republican won (1.12% average) and just four of the nine occasions a Democrat won (-0.81% average).

10 years later, Riley and Luksetich (1980) also demonstrate that presidential elections strongly affect stock markets and that the markets react positively to the victory of a Republican candidate and negatively to a Democratic president in the short run after the election. They show this by aggregating price movements during the 17-week period surrounding each of the 20 elections since 1900.

More recently, Knight (2006) and Jayachandran (2006) find evidence that presidential party affiliation or the party that controls the Senate has an impact on specific firms. The former tracks daily data across 70 firms during the six months leading up to the 2000 US presidential election. They find Bush-favoured firms, such as tobacco firms, rose on average in value by 9% more than Gore-favoured firms, such as Microsoft, following the Republican win. Jayachandran (2006) finds that in May 2001, when Senator Jim Jeffords switched from the Republican Party to the Democratic Party, public firms lost on average 0.8% of the market capitalization for every $250,000 contributed to Republicans in the previous election cycle.

Finally, Jones & Banning (2009) investigate possible relations between stock market performance and various occurrences in US elections using monthly market returns over a period of 104 years. They find that ‘market returns do not appear to vary based on partisan control of the government’, concluding that ‘neither election results nor the election cycle appears to offer much help in predicting stock returns’. They also suggest that different control combinations of the White House and/or the US Congress have no systematic effect on the stock market.

What Could We Expect from Another Trump Win?

With President Trump officially re-nominated for the 2020 election, it is perhaps informative to understand how markets responded to his surprise 2016 election win.

On the election day in November 2016, the US stock market plunged as futures contracts on the Dow Jones Industrial Average retreated 506 points, or roughly 4%, in reaction to Trump’s anti-free trade views. The effect was not confined to US markets, it was global. In London, the stock market plunged 2% following futures losses amounting to 4%. The Hang Seng index in Hong Kong fell 2.7%, the South Korean Kospi dropped 2.5% and the Japanese Nikkei 225 went down 5.1%.

The following day, however, markets bounced back. Bouoiyour and Selmi (2017a) make the observation that ‘collapsing stock returns around the election result is reversed by positive abnormal return on the next day, except some cases where we note negative responses following the vote count’. In their paper, they conduct an event study methodology that assesses the abnormal returns (AR) and cumulative abnormal returns (CAR) behaviours for several sectors of the S&P 500 stock market around the day of the US presidential election result (8 November 2016). Table 2 below summarises their findings.

Source: Page 174 of ‘The financial costs of political uncertainty’

Some sectors exhibited positive responses, including consumer discretionary, energy, industrials, materials, aerospace and defence, and real estate. These sectors reacted significantly on the event day and most responses do not last 2-20 trading days.

The way to read the table is as follows: the cumulative abnormal returns of energy sector firms were positively impacted by 1.42% in the 239 days prior to the November 2016 election. Around the day of the event, abnormal returns were -1.11%. The CAR were 1.16% (with a t statistic of 1.84) after two days of the election event, and continued to witness a CAR of 2.46% (with a t statistic of 4.19) after 20 days. The paper suggests that these findings can be explained by Trump’s deregulatory energy plan aimed at repealing the Clean Power Plan and enabling more drilling for shale oil and natural gas on federal lands.

They also find that six sectors adversely responded to the election result. These include consumer staples, financials, health care, information technology, communication services, and utilities. Health care, for example, witnessed an AR of -2.04% over two days after the Trump’s win in the elections, and -4.12% after 20 days.

Another paper finds similar sectoral differences, but on days leading up to the election. Pham et al. (2018) examines 47 events, starting with Trump’s announcement that he would contest the presidential election, using the event study methodology and asset pricing models. They observe that the 2016 presidential election affected the US stock market and that the market was highly responsive when Trump secured the Republican nomination. The life insurance sector was one of the most negatively affected sectors due to Trump’s intention to replace Obamacare. The industrial engineering sector was the largest beneficiary on election day, posting 2.9% abnormal returns.

One implication from these papers is that winners and losers are selected based on the policies of the incoming president. For example, the winners won because of plans to rebuild the infrastructure, renegotiate trade agreements, cut taxes, ‘reform’ labour laws, boost military spending and roll back Obamacare. The same sectors may not necessarily respond in a similar fashion this time around.

How Does Election Uncertainty Influence Equity Returns?

Next we look at how greater political uncertainty can influence equity returns on election year. Political uncertainty can drive stock price movements if new information exerts significant influence on the country’s macroeconomic, fiscal, and monetary policies (Niederhoffer et al., 1970). The US presidential election is a political event that involves an uncertain result with sizable economic implications. As an application of Brown et al.’s (1988) Uncertain Information Hypothesis (UIH), election year uncertainty has the potential to influence prices.

I’ve presented evidence that stock returns tend to be higher on election years. The literature suggests that the scale of the win will depend on the level of uncertainty regarding the results. For example, financial markets did not much respond to the re-election of Presidents Ronald Reagan or Bill Clinton, since both were heavy favourites by election night (83% and 93%, respectively). In contrast, President Harry Truman’s re-election contained a lot of news, since his pre-election odds of winning were much lower (11%).

There are a number of seminal works that deal with the effect of political uncertainty surrounding US presidential elections on the US stock market. Three papers in particular focus on the delayed results of the 2000 presidential election and find that greater election uncertainty puts downward pressure on stock prices.

Nippani and Medlin (2002) use regressional analysis to show that the impact of the delay in the declaration of a winner in the 2000 US presidential election negatively affected US and international stock markets. He, Lin, Wu, & Dufrene (2009) investigate the information cost of stock trading during the 2000 presidential election and find that election uncertainty induced information asymmetry of politically sensitive firms under the Bush/Gore platforms. They concluded that election uncertainty negatively affected stock markets.

The impact of political uncertainty is also not limited to US markets. Nippani and Arize (2005) find evidence indicating that US elections negatively affect both Mexican and Canadian stock markets, concluding that they are closely integrated with their American counterparts and that the two markets follow the US presidential elections as closely as US markets do.

Furthermore, Bouoiyour and Selmi (2017b) examine the effects of the surprise 2016 election win on BRICS (Brazil, Russia, India, China and South Africa) stock markets. Their results reveal that although some markets emerged as losers, others benefited from Trump’s victory. Unsurprisingly, they find that the Chinese market took the biggest hit, while the damage inflicted on India and South Africa was much more limited.

Spanning more than just one election, Li and Born (2006) use polling data on the US presidential elections from 1964 through 2000 to examine the effects of election-induced uncertainty on stock returns and volatility. They find that if the election does not have a candidate with a dominant lead, stock market volatility and average returns rise.

They also find that stock prices increase before presidential elections when the outcome of the election is uncertain. This again provides evidence that the heightened uncertainty surrounding the US presidential elections is sharply reflected in the behaviour of stock prices.

Using a slightly different approach, Goodell and Bodey (2012) investigate price-earnings ratios of the S&P 500 firms around US presidential elections. They document that price-earnings ratios are negatively associated with the lessening of election uncertainty around US presidential elections, suggesting that decreasing uncertainty about the electoral outcome leads to a decrease in stock market valuations.

How Does Election Uncertainty Impact Market Volatility?

While several studies have examined the association between US presidential elections and stock returns, surprisingly little attention has been given to the effects of elections and election-induced uncertainty on stock market volatility. From the evidence that is available, greater election uncertainty can lead to a higher degree of stock market volatility.

Outside of the US, Gemmill (1992) focuses on the British parliamentary election of 1987. The paper documents that stock market volatility, as measured by the implied volatility of the FTSE 100 index, increased substantially before the election at the same time as opinion polls indicated with increasing probability a victory for the Conservative Party.

Using a broader set of 27 OECD countries, Bialkowski et al. (2008) investigate stock market uncertainty around national parliamentary and presidential elections. They document that national elections induce periods of increased stock market volatility and argue that the increased volatility indicates that election outcomes generally surprise investors.

Beyond the work of Li and Born (2006), mentioned in the previous section, there are two US-specific pieces on market volatility. First, Goodell & Vahamaa (2013) study the impact of political uncertainty and the political process on option-implied stock market volatility during US presidential election cycles. The empirical findings indicate that political uncertainty around US presidential elections affects option-implied stock market volatility. The results indicate that the presidential election process engenders market anxiety as investors form expectations regarding changes in macroeconomic policy.

Mnasri and Essaddam’s (2020) study supports the finding above. Using data from the last seven US elections (1992 to 2016), they find that US presidential elections induce higher stock market volatility. The estimation results show that a 1% change in the probability of success of the candidate of a political party other than that of the current president causes an increase of almost 0.1 in %VIX.

Across the board, the empirical findings suggest that greater uncertainty around the US presidential winner can lead to a higher degree of stock market volatility.

The Bottom Line

With fewer than 100 days left until the 2020 presidential election, investors cannot be blamed for trying to guess what’s in store for the next three months. Based on data going as a far back as the 1700s, election years appear to be associated with better equity performance. If this is the case, then we may expect the recent market rallies to continue. Furthermore, if polls point towards Biden as a firm-favourite, markets may also be less volatile.

Upon election day, in a historical context, a Biden win may not be as fruitful for financial markets as a Trump re-election. In these unprecedented times, however, it is hard to tell who the markets will prefer.