This is the third in Risk.net’s series of crowdsourced scenario generation exercises. You can view the results of the first and second surveys here.

It’s Monday, November 9, a week after the polls have closed on the most divisive US election in living memory. With all votes counted – and recounted multiple times in the largest swing states – Joe Biden has won. But Donald Trump says he’s going nowhere, calling fraud on the postal votes that swung the key states Biden’s way, and is unleashing a lengthy, bitter legal fight.

It’s an outcome financial markets have long feared – not necessarily for the legislative paralysis that would likely result, but rather because investor behaviour in the aftermath would be near impossible to predict. An infinite array of uncertainties collide with hard mechanics: who gets to decide the result – a divided Congress or a conservative-biased Supreme Court? Does the public accept the outcome, or does the nation erupt in flames? All the while, the coronavirus rages, heaping pain on an already ravaged economy.

Over the past month, Risk.net has been asking readers what they think would happen to financial markets in a range of election outcomes, including the prospect of a contested result. The responses were processed by RiskThinking.ai to generate a set of distributions and accompanying crowdsourced stress scenarios.

The results aren’t pretty. Under the scenario deemed most likely following a contested election result, the S&P 500 collapses to 2,100 by late January, while yields on 10-year US Treasuries reach a record low of 0%. Ordinarily, huge demand for US government bonds would drive the US dollar higher, but in this scenario, a collapse in oil prices to $10 a barrel sees the greenback tumbling against the euro to a post-crisis low of $1.40. US unemployment reaches 13%.

Anticipating the events of the next few months will be a challenge for all market participants, according to Evan Sekeris, who previously oversaw operational risk at the US Federal Reserve.

“There’s a perfect storm building up, with a very contested US election arriving at the same time as we are struggling to control the Covid pandemic. In terms of what this means from a stress-testing point of view, I think the current situation has highlighted a few weaknesses in the frameworks we’re using. It relates to how these idiosyncratic operational risk-type events affect and interact with macroeconomic scenarios that we have in our stress-testing frameworks,” he said.

Sekeris was speaking during a Risk.net webinar on US election stress-testing on October 23 (see video below).

On the question of how the US populace – already polarised, already on edge – would respond to a contested election, the chief risk officer (CRO) at one of the largest US asset managers puts his firm’s dilemma succinctly.

“Thinking that through and what it means ... How do you capture the mood of a country, and therefore the risk of that for financial markets? I think that will be an ongoing question for all of us,” he says.

Regardless, even in the event of a clear election outcome and prompt transfer of power, “social unrest is something that will be with us for quite some time”, he adds.

Results

The survey asked respondents to give their views on a set of financial indicators under four different regimes. In regime one, Donald Trump is re-elected president, while the Republicans keep control of the Senate, and win the House of Representatives – a clean sweep. Regime two is the reverse, with Joe Biden elected president and the Democrats controlling both chambers of Congress.

In regime three, respondents were asked to consider an unclear or contested result, with no conclusive resolution until early in 2021. Under regime four, there is continued legislative deadlock: it doesn’t matter who is president, because Congress remains divided, and major policy initiatives are impossible to enact.

The second of these regimes – a Biden win and Democrat sweep – was seen as the most likely by all respondents, with deadlock seen as the next most likely by US respondents, and a Republican win the second most popular choice of respondents in the rest of the world.

The point of the exercise, though, is not to forecast the outcome of the election, but to explore the grim financial possibilities under each of these regimes.

This survey’s risk factors are different to the ones used in its predecessors, which both focused on the impact of the coronavirus. Rather than being asked to forecast a vaccine timeline, US GDP and investment-grade credit spreads, respondents had to predict the change in US unemployment and the WTI oil benchmark between the date of the election and late-January 2021. They were also asked to call the next-decade impact on global carbon emissions, under each regime. The remaining risk factors – yields on 10-year US Treasuries, the euro/US dollar exchange rate, and the level of the S&P 500 – are the same as in the previous survey.

Perhaps unsurprisingly, expectations on the outlook for US stocks show the widest divergence of all financial factors polled across the different regimes. Respondents predicted sharp gains for stocks following a Republican sweep, with a majority expecting the index to rise versus its value on September 11, 2020, of 3,440. At the upper extreme, almost 10% of readers predicted it to break 4,000.

Conversely, falls are predicted following a Democrat win – to a median of 3,240 – with almost one-third of respondents tipping the index to fall below the 3,000 level, possibly a reflection of Biden’s plans to reverse at least half of the Trump regime’s generous cuts to headline corporate tax rates, should he gain the White House.

A contested outcome produces the most dramatic falls. The median forecast slips to 2,860, while at the lower bound, some respondents foresaw the blue-chip index halving to 1,500 – a level not seen since the start of Barack Obama’s second term.

Fabrice Fiol, deputy head of enterprise risk management for the Americas at Societe Generale, said his firm was wrestling with precisely this dynamic in its own attempts to build a credible baseline scenario for a contested outcome.

“We have a couple of outcomes for a contested election: results contested for a couple of days could create a situation where you have risk premiums going up, and volatility reacting to that. Then you could have another scenario which is more protracted, where you have the Supreme Court involved, and it goes beyond the December 14 safe harbour date, which could lead to a severe crisis around the contested election. Clearly, that’s something which could create a government shutdown and social unrest,” he said, also during last week’s webinar.

Respondents’ views suggest relative stability for the S&P 500 under the legislative deadlock scenario, with the median view a slight decline to 3,120, and limited optimism on the upside, with the majority of views falling into a fairly tight range between 3,000 and 3,500.

Paradoxically, despite the wide dispersion of views on the S&P 500 between the regimes, there is a surprising lack of a corresponding heterogeneity when it comes to forecasts on US Treasuries. For instance, the shape of the distribution of expected change in yields – relative to the 0.71% seen on September 11, 2020 – is very similar under three of the four election outcomes. In the case of a Republican win, Democrat win, or deadlock, the respective median predictions are 0.6%, 0.8% and 0.7%.

Social unrest is something that will be with us for quite some time

Chief risk officer at one of the largest US asset managers

Median expectations are only slightly lower for a contested result – although here, some respondents see a negative print for the 10-year, with a cluster of readers forecasting yields plummeting to -2%.

The US asset manager’s CRO points out that the seeming agreement visible in the distributions for the US 10-year could be deceptive, however, arguing an element of “anchor bias” might be at play. While US Treasuries are usually heavily negatively correlated to the performance of the S&P – and many predicted explosive S&P rises under the first regime, for instance – comparatively few participants voted for the former value to drop below zero.

On the one hand, this could indicate respondents anticipating the kind of dramatic correlation breaks that occurred in March, when the value of both Treasuries and stocks plummeted in lockstep, before the US Federal Reserve stepped in to calm markets. More prosaically, it could simply suggest a reluctance on the part of respondents to predict a negative print on the 10-year – because it hasn’t happened before.

Sekeris said: “When we think of tail events such as the pandemic, such as potential contested outcomes in the election, the question becomes: ‘Do we have historical frames of reference to help us quantify those? Do we understand the drivers of these events? How do we go about using all the information at our disposal to come up with estimates that are reasonable and accepted by people?’”

Similar agreement is seen for change in the EUR/USD exchange rate, relative to the closing price of $1.19 on September 11, 2020. Under all regimes, respondents expect the rate to rise. Once again, a contested result bucks the trend – albeit more gently this time, with respondents predicting a median rise to $1.24 with long tails in both directions. This relative stability may reflect the historical behaviour of the dollar, a currency historically seen as a safe haven during times of market stress; respondents may be anticipating that investors will horde US Treasuries in the event of a contested election. The median predictions for Republican win, Democrat win and legislative deadlock are all $1.20, albeit with extremes that go out as far as $0.75 in the Republican-sweep case and $1.50 in the event of Biden winning.

The scenarios

RiskThinking.ai used respondents’ views to construct 64 stress scenarios for each of the four regimes. By design, these are a composite of the more extreme forecasts in the survey.

A selection of the scenarios are shown in the table below; to download the full selection in a CSV format, click here. By default, these are grouped by likelihood – a function of the joint probability distributions of each risk factor relative to its modal consensus value, multiplied together (click here for the full methodology).

For a Democrat win – which the regime participants see as the most likely outcome overall – the highest-scoring scenario has a likelihood of 2.87%. In that scenario: the US 10-year yield falls to -0.7%; the EUR/USD sees a substantial gain to $1.30; the S&P 500 drops dramatically to 2,080; US unemployment climbs to 14%; WTI oil reaches a price of $54; and carbon emissions are projected to fall sharply, by -20%.Under a Republican sweep, the most likely scenario generated has a probability score of 3.39%. In that scenario: the US 10-year yield rises to 1.9%; the EUR/USD exchange rate holds relatively firm at $1.20; the S&P 500 jumps to an index level of 4,010; US unemployment falls to 5%; WTI oil leaps to $65.50; and carbon emissions increase by 20%.

The third regime, a contested result, has the scenario with the highest probability score – the one described at the start of this article. While it has the highest probability score of any of the 256 combinations, though, it’s worth keeping in mind that a contested result was the least-predicted outcome overall. Still, market participants should bear in mind: the point of the exercise is to catch black swans, rather than establish consensus.

Under the legislative deadlock regime, the most likely scenario has a probability score of 4.61%. Here, the US 10-year yield is 0.1%; the EUR/USD exchange rate is $1.30; the S&P 500 falls to 2,110; US unemployment moves up to 14.5%; WTI drops to $20; and longer-term carbon emissions increase by 70%.