Stress Testing the US Presidential Election: A framework for quantifying market reaction to an unpredictable outcome

With the US presidential election now less than seven weeks away, many clients have been asking us how to forecast market returns ahead of this pivotal event. It’s a fair question, but a challenging one, to be sure. The fact is, the outcome of the election itself is basically a toss-up at this point, and it is even more challenging to sort the likely effects on major financial variables. Indeed, there are so many caveats in an exercise like this that any results must be taken with more than a few grains of salt. That said, the Applied Research team is not one to shy away from a tough challenge. So we asked ourselves — why not try to model the potential impact of the US presidential election outcome on the STOXX® Global 1800 index?

The following is offered not so much as a definitive answer to the question posed by our clients, but as an example of a framework for thinking about and tackling this issue from a stress-test perspective. As always, do not hesitate to reach out directly to us with questions and comments, as we welcome discussions on this challenging subject.

We started with some background research[1] on the performance of the US stock market during election years to frame our stress test. First, the long-term average annual return for the S&P 500 index is about 9%, so a shock of about 10% on that index seems historically appropriate. Second, according to Strategas Research Partners[2], if the market is up in the three months preceding the election, the incumbent generally wins, and the market keeps going up in the three months after the election. Conversely, if the market is down in the three months prior to the election, the challenger tends to win, but the market still climbs for the next three months. So, it looks like US Presidential elections are always good for the stock market and no stress testing is necessary.

Well, not exactly. These numbers represent averages. In every election year, the returns have either been below or above that 9% level, sometimes substantially so. And the predictability of election results based on the stock market performance prior to the election is not always accurate, nor would we be able to rely on that prediction until the day before the election itself; a bit too late for comfort. So maybe modeling a few scenarios and running some stress tests on our index portfolio might be useful after all.

As always with stress testing, we started with the obvious—what if history repeats itself? Since we have had many US Presidential elections, let us see what would happen to today’s STOXX Global 1800 portfolio if we replayed history on today’s index portfolio. We used the Obama reelection of 2012 and the 2016 Trump election, taking the periods from November 1, 2012 and 2016, to January 20, 2013 and 2017, respectively, as our historical windows. These gave us a higher predicted gain of +8.9% for the Obama scenario versus +7.5% for the Trump scenario (but do not tell Trump I said that!). However, Trump’s results were built on top of eight years of economic and stock market growth. In that context, Trump’s gain actually looks pretty good.

Next we tried to get a bit more creative with our scenarios3, using a transitive stress test to shock the S&P 500 index returns, the USD (dollar index), and the 10-Year USTB yield, then calibrating the correlation between those two price series and the holdings in our global portfolio to generate some expected returns.

Most electoral math we have seen says Democrats will retain the majority in the House, but without the two-thirds they would need to override the Senate. This leaves four possible scenarios to look at:

  • Biden President / Republican Senate (BP/RS)
  • Biden President / Democrat Senate (BP/DS)
  • Trump President / Republican Senate (TP/RS)
  • Trump President / Democrat Senate (TP/DS)

Next comes the more ‘artistic’ part of stress testing, including defining economic scenarios for each outcome and their corresponding sign and size of potential shocks to our three factors (i.e., what does each scenario mean for the direction of US equities, the USD, and US interest rates). This is made more difficult this year given the very small wiggle room interest rates have under the current QE program.

Here is what we think we know so far. Both parties want to rebuild the infrastructure and stimulate (domestic) economic growth, but they differ in how they intend to pay for that new spending; one with taxes, the other by borrowing even more from foreign investors. One candidate wants to isolate the US, the other wants to mend relationships with (at the very least) our allies. One candidate (and party) seeks policies to ease social unrest, while the other seems to favor law and order.

Based on this, we attempted to predict how the market would react to these four possible outcomes and came up with the following shocks for our three factors: the yield of the 10-year USTB, the Dollar Index, and the S&P500 index. The table below summarizes our directional shocks to these three factors, and the date ranges used to calibrate the correlations between those factors and all the assets in our global index portfolio.

Sources: 10-year USTB, Dollar Index, S&P 500 Index

The narrative for our four scenarios goes something like this:

  • BP/RS – probably the best case for markets. This scenario will see higher government spending, no new taxes, less social unrest, and a more diplomatic approach to foreign policy, and reflects the maxim that Americans prefer a divided government. For this scenario, we used the three months after Obama’s 2012 reelection to calibrate our correlations.
  • BP/DS – this scenario introduces the possibility of higher taxes, even if only on the top 1% of earners, which is never popular with markets. Big Tech will face more regulatory oversight, too. Social unrest is reduced, and we also will see a more diplomatic approach to foreign policy under this outcome (e.g. the US becomes less unpopular in France). This scenario is reminiscent of Obama’s first term in office and the Obamacare passage through congress, we therefore chose the post-Obamacare implementation period of July to Dec 2010 to calibrate our correlations.
  • TP/RS – this scenario is the status quo. Trump remains divisive and unpredictable but pro-business, anti-taxes, and anti-regulation. Social unrest and geopolitics cause flare-ups now and then but the market has learned to live with Trump. For this scenario we used the full 2019 period to calibrate our correlations.
  • TP/DS – this scenario basically sees Trump pushed into a corner by the Democratic Senate. More executive orders, more veto threats, a stalled government process, and increased international tensions. This scenario is reminiscent of the type of blame game and stalling that was seen during the two government shutdowns of Trump’s first term in office[3]. For this reason, we chose the period around the second shutdown of December 2018 to January 2019 to calibrate our correlations.

Putting the current holdings of the STOXX Global 1800 through each of these scenarios gives us the results in the table below. These are expressed in terms of expected change in present value (%) and represent the expected gain/loss in the 60-days following the US election. The top row is the result for the index and the rows below that represent the standalone results for each of the GICS sectors (i.e., they do not add-up to the index returns).

Source: Axioma Risk

It is important to note that this just reflects what could happen in the first few months of the old or new administration and congressional makeup, and is based solely on our rough estimates of the election’s impact on major traded financial variables. Different investors will probably have quite different views of the potential impact of these scenarios. Our goal with this note was to set up a framework that describes how one might think about this issue.

As difficult as it is to predict the outcome of the upcoming US election, it is even more challenging to devise its expected impact on major financial variables. The flexibility of Axioma Risk allows investors to easily revise their views and quickly determine their impact. Investors should never rely on a single stress test and keep in mind that stress testing is a journey, not a destination. Revise your forecasts as often as necessary by incoming new information and rerun your stress tests monitoring the changes in impact to gain insights and perspectives.


[1] Fancy term for “Googling” it…

[2] Strategist Dan Clifton of Strategas Research Partners

3 As always, there are numerous scenarios to pursue, in terms of factors to shock, by how much, in which direction, etc. Please contact the Applied Research team at appliedresearch@qontigo.com to discuss the best ways of meeting your goals.

[3] January 2018 and Dec 2018 – Jan 2019.

Olivier d'Assier

Olivier d'Assier is Head of Applied Research, APAC, for Axioma and is responsible for generating unique regional insights into risk trends by leveraging and analyzing Axioma's vast data on market and portfolio risk. d'Assier's research helps clients and prospects better understand and adapt to the evolving risk environment in the Asia Pacific.