Indecision 2020: The Potential for a 2000 Election Night Reboot

During election cycles, historical return analysis under different presidential and congressional outcomes is as common as political advertisements dominating the airwaves. While the latter serves as an effective way for candidates to communicate policy messages, the former is more anecdotal than statistically significant.¹

Although the information in those return time series is interesting, it is difficult to use it to extrapolate future trends. That’s especially true for an election as uncertain as this year’s presidential race. However, one election just may provide some insight into the 2020 cycle: the 2000 presidential recount election.²

Running back a recount?

For those who may not remember — or were still in preschool — the outcome of the 2000 election was delayed by 36 days as numerous hanging chads on Florida voter ballots were recounted and then litigated in a court of law – and in the court of public opinion. After numerous filings, injunctions and motions, the US Supreme Court ruled in favor of George W. Bush, and on December 13, former Vice President Al Gore conceded the election on national television.

The logistics of holding an election during a pandemic is the one reason that we may have a 2000-era delayed result reboot. Unlike past elections, the 2020 election could be determined by more votes cast by mail-in ballots as opposed to in person. Estimates indicate that 64% of all votes may be cast by mail in this election³  – up from 22% in the 2016 election.⁴

In addition to the sizable increase in number of mail-in ballots, different states have different rules on the cutoff date for votes to be counted. In certain states, votes must be received by Election Day; in other states, votes must be postmarked no later than Election Day.⁵  These state-by-state differences combined with the uptick in mail-in ballot usage have led the US Postal Service (USPS) to warn 46 states and the District of Columbia that it cannot guarantee that all ballots cast by mail for the presidential election will arrive in time to be counted.⁶  All of this is occurring while USPS resources are limited, as due to the removal of sorting machines, the USPS estimates that it is now operating with a 10% loss in capacity.⁷  This current state of USPS operations has already led to a lawsuit related to the 2020 election voting processes,⁸ and it could be the first of many if we do have an uncertain election night outcome.

Mail time

Voting patterns and preferences add to the likelihood that we will see partial results on election night, followed by an extended delay in declaring a winner. Republicans have made 26% of the mail-in ballot requests, versus 43% made by Democrats.⁹ Additionally, according to a Wall St. Journal/NBC News Poll, 66% of Republicans prefer to vote in person, as opposed to 26% of Democrats. As mail-in votes may not be counted as quickly as in-person votes, the early indications on election night could make it look as if President Trump has won a second term. If that is the case, markets may move in the direction of assets that would be supported by this result.

But with the majority of the votes potentially being cast by mail, it’s highly unlikely that the election will be confidently called in certain states on election night – some estimates indicate that it could take weeks to count all the votes.¹⁰  As a result, we could see knee-jerk reactions in the overnight markets as news and votes trickle in – stoking more volatility while the expectations for outsized moves are already high.¹¹

And if there is a drawn-out ballot count, what may assets do? The 2000 performance may offer a clue – even if the sample size is just one period.

Back in 2000, risk assets fell across the board, as shown below. Not only did US equities – both large and small – fall, but so did non-US stocks. The riskier small-cap equities fell the most, declining by over 7% over that 36-day period; at one point, they were down by double digits.

More defensive assets that are less correlated with stocks performed quite well, however. As shown below, broad core US aggregate bonds were positive over this time period. But their rally was driven by US Treasuries, and in particular, long-term US Treasuries. Meanwhile, credit lagged. Gold, another defensive asset, also performed well during this time period, posting a gain of almost 2%. These trends illustrate that defense was sought as a result of the uptick in policy uncertainty.

Taking this one period analysis a bit further, we see how even within equities, the market preferred defense rather than offense. As shown below, by using quantitative factor baskets of low-volatility, high-volatility, high-quality and low-quality stocks, we can see how risk preference further evolved. The concentrated baskets of low volatility and high quality vastly outperformed their corresponding counterparts — by 15.6% and 8.2%, respectively. They outperformed the market by 7.9% and 6.1%, respectively, as well.

Positioning for an inconclusive election night

I equate using historical return data to frame out current election market opportunities with sports commentators suggesting that the 25-year win/loss record between two teams could help predict the winner of the current matchup. Giving the winningest team the edge in today’s competition completely overlooks the exogenous variables of changes in coaching staff, field personnel, stadiums, and style of play in those 25 years.

Still, the probability of a 2000-style delay in the 2020 results is high and worth examining, as the market doesn’t like uncertainty — especially when it comes to the presidency. Adding to this year’s uncertainty is the composition of the US Supreme Court — if an election lawsuit reaches their doorstep, as it did in 2000. With the passing of Justice Ruth Bader Ginsburg, the US Supreme Court now has just eight seats filled. If the eight-person court splits 4-4 in any case, the decision reverts back to the lower court’s ruling. While a new justice is likely to be nominated soon, confirmation by the Senate by Election Day (40 days away) would far outpace the average number of days it typically takes to confirm a new US Supreme Court Justice (67 days).¹²  It is, therefore, possible that potential election lawsuits could coincide with the confirmation of a new justice, who could rule on the election’s outcome — stoking uncertainty in the markets for longer than 2000’s 36 days. After all, unlike in 2000, there likely will be more than just one state with an election outcome in doubt.

Back in 2000, the 36 days of election suspense favored defensive positions, such as long-term Treasuries, gold and low-volatility and high-quality stocks. Without a decision on November 3, the 2020 return pattern may be similar.

With that in mind, investors should consider bolstering portfolios with some of the aforementioned risk-off positions, seeking to mitigate the potential for a 2020 election outcome delay to be a drag on returns.