The COVID-19 pandemic and its consequences are the largest shock to the geopolitical order in at least a decade.
It’s reasonable to take a moment to see whether the political risk industry - which is dedicated to forecasting such shocks - saw it coming and what we should do if they didn’t.
It is early, certainly, to be conducting an after-action report, so take this as the first draft of the first stage of the crisis. Further assessments will be made every few months until a major report, similar to the Stimson Center’s look back at the Arab Spring, can be written.
Defining success and failure
One of the first difficulties in assessing the industry is defining success and failure. In a sector where our predictions are almost all probabilistic, how do we classify a single event?
The approach I’m taking is to determine if this event was a black swan or a gray swan. Political risk can be forgiven for missing black swans but should be picking up on the gray ones.
What’s the difference between swans?
As Philip Tetlock opined in 2016 about Donald Trump’s performance in the Republican primary:
His success has been so astounding that, as Jack Shafer wrote in Politico, it looks a lot like what Nassim Nicholas Taleb famously dubbed a “black swan” — an enormously consequential event that is unpredictable but seems foreseeable, even obvious, in hindsight.
He continues:
The prospect of a President Trump is closer to a gray swan than a black one — and it offers a valuable opportunity for learning just how much we can and cannot know about the future… Trump fits into the comparison class of system-destabilizing populists — from Huey Long to Ross Perot — pretty well. Just because politics is a complex system doesn’t mean we can’t make (and improve) political predictions.
Lots of people thought that Donald Trump was a black swan event - the kind of thing no one could predict. In hindsight, we can argue about the role of The Apprentice in burnishing his image as a decision-maker, the importance of birtherism, and the role of free media. But before he announced his nomination, no forecaster thought he could win.
Here’s how it was treated by Jerry Seinfeld at Saturday Night Live’s 40th anniversary show in February 2015.
Seinfeld: Next question. Yes, Tina?
Palin: No, no it's Sarah. Sarah Palin.
Seinfeld: Yes, sorry, my God, sorry!
Palin: I'm just curious, Jerry, how much Lorne Michaels would pay me if I were to run in 2016?
Seinfeld: Run for president? Sarah, I don't think there's a number too big.
Palin: Okay, just hypothetically then, what if I were to choose Donald Trump as my running mate?
Seinfeld: Sarah, you're teasing us! That's not nice!
Yes, that was Sarah Palin asking a question about Donald Trump being her running mate. And it was treated as a joke so ludicrous it made a Seinfeld bit on SNL.
Even Nate Silver predicted that “Trump’s campaign will fail by one means or another.”
Tetlock argued that Trump winning was unlikely but not unforeseeable, a gray swan. There were precedents for his type of politics.
I would go further than Tetlock. At the time of Palin’s joke with Seinfeld, Trump winning could be seen as the kind of unpredictable event no one could forecast - a true black swan.
But as Trump entered the race and continued to lead in the polls, it became more realistic. A candidate leading in the polls for months with universal name recognition often wins. By December 2015, it was a gray swan and should have been treated as such.
The Trump election shows that the difference between a black swan and a gray swan can be timing. A political black swan can become less so as time passes, slowly and steadily becoming more recognizably gray.
Timing the swans
I would argue that COVID-19 started as a black swan.
We know pandemics are possible, but it is virtually impossible to identify which month or year one will emerge. A pandemic, like a major cyberattack, nuclear explosion, or terrorist strike, is a risk that we know may occur but the when and where are not easily predictable with any level of certainty by the industry.
But eventually, the warning signs of a coming pandemic began to grow brighter.
On January 22, the former Ebola response coordinator Ronald Klain wrote in the Washington Post:
With one confirmed case on U.S. soil, more likely already here and 8,000 visitors from China arriving every day, it is already too late to avoid multiple cases of the dangerous new coronavirus in the United States. We are past the “if” question and squarely facing the “how bad will it be” phase of the response.
On January 30, the WHO declared the coronavirus a Public Health Emergency of International Concern.
On February 23, Italian towns in Lombardy were locked down. On February 24, markets around the world melted and continued to drop.
On March 11, the NBA suspended games and Trump gave a national address.
These facts lead to my assessment of how to determine which color swan COVID-19 was and when:
I would argue that before January 30, the emerging coronavirus epidemic in China could be considered a black swan event, and would be unknown outside of specialists.
Between January 30 and February 24, the epidemic was now a pandemic and became a gray swan concern. Risk analysts should have been tracking it.
After February 24, the business world was fully cognizant of COVID-19. Risk analysts who caught on at that point were well behind the ball.
After March 11, anyone who dismissed it was ignoring the reality around them.
How did political risk fare?
Using the twitter feeds of four prominent political risk firms as a proxy for the industry’s attention, we find that coronavirus was first mentioned on January 20, with all firms tweeting about it by January 24.
This is a good sign for the industry’s attention to COVID-19. All anticipated the WHO’s declaration and, therefore, were ahead of the black swan deadline.
However, almost all coverage of the outbreak was confined to its impact on Chinese politics and economics until February 5.
Even during the ‘gray swan’ period of February, coronavirus was mentioned in only 16% of tweets on the average day, compared to 38% in March.
One prominent political risk analyst even spoke at a conference on February 28 downplaying the risk of coronavirus, comparing it to the seasonal flu.
Even up until February 24, assessments were focused on China, which was mentioned in the vast majority of tweets discussing coronavirus.
Grading political risk
The political risk industry can be praised for paying significant attention to the coronavirus before much of the market.
This is in part because the industry has a global remit, and because China’s actions could have had a major impact on the global economy.
However, it would be difficult to argue that the political risk industry clearly foresaw the global societal and economic impact of the pandemic.
It was not until well after equities markets began to fall that there was a clear shift in focus. The political risk industry certainly jumped on the issue before the broader population, but largely missed the gray swan period (of course, this is a broad assessment, and certain analysts and firms did better).
Why didn’t political risk see this coming?
There are several mitigating reasons for the industry’s sub-optimal performance.
First is what the risk world had to cover. In the United States, late January and early February was dominated by impeachment and the first presidential primaries. What was happening in Wuhan was important, but not as immediate to a Western audience. It is understandable at the time to give the emerging outbreak a back seat.
Second is the impact of cognitive biases and the nature of global business. As psychologists and behavioral economists have found, people tend to discount information that does not fit pre-existing mental models. Because the expectations of flu-like illnesses in Asia were shaped by the history of SARS and H1N1, which were damaging but contained, the novel coronavirus was predicted to follow a similar path. That would explain the early focus on China, rather than the globe.
Third are informational silos. Political risk deals with politics, economic, and business trends. Epidemiology is not often in a risk analyst’s daily reading. An unfamiliarity with the science likely contributed to a delay in grappling with the severity of the problem. For COVID-19 to filter through to risk analysts, it had to get into mainstream media first.
A final reason is that risk analysts often work in a Bayesian fashion, updating prior expectations gradually with new information. That approach has been found in academic studies to be superior to more rapid shifts in forecasts. However, one situation where that approach fails is when change occurs exponentially. Truly novel and fast-breaking situations require imaginative systems thinking. In other words, being late to recognize the severity of unprecedented events is an inherent limitation of methods that maximize accuracy in normal times.
How did Two Lanterns fare?
In the interests of full disclosure, I should note that I was similar to many in not foreseeing the global impact of COVID-19 during the black swan period. However, I did appreciate the gravity of what was happening sooner than many in the industry, as black turned to gray.
In January, I was certainly aware of what was happening in Asia. I thought that it could trigger a recession in the US and Europe by disrupting supply chains and dampening Chinese demand.
I started to quickly shift towards greater worry about the global economy when COVID-19 began to spread in Italy, especially when towns went into lockdown on February 21.
At that point, the historical comparison to swine flu was becoming more tenuous. It also had spread to areas that had considerable connections with the rest of Europe and the United States. I can’t say that I predicted the extent of social distancing and the depth of the economic crisis then, but I was starting to place my Bayesian thinking on a shelf.
The overwhelming sense back then was uncertainty. As the stock market tumbled the week of February 24, I remember thinking that I didn’t know where the floor might be. Usually there’s some awareness of the baseline projection and the worst-case scenario. But as commentators on CNBC kept drawing lines of where they predicted the stock market to bottom out, I realized they were just guessing (or relying on now-useless heuristics).
In a sense, though, that uncertainty was itself useful.
Too often, analysts become stuck in their positions. When that veteran risk analyst dismissed COVID-19 as no more dangerous than the flu on February 28, the problem wasn’t that he was objectively wrong, but that his thinking was a week out of date.
Now, I certainly don’t claim any special predictive powers into COVID-19. I am not an epidemiologist, and my analysis was based on what I was reading in the media. The only reason why I think I mentally got a jump on the markets, and then on the general public, is that my work has always centered on dealing with uncertainty.
It’s the core of what I teach in my training sessions and making forecasts in uncertainty is the foundation of many of my products. So much was changing on a daily basis that a willingness to keep an open mind was the best skill someone could have.
Given how much we don’t know about the coming months, it’s a skill we’ll all need to keep working on.
----------
Thanks for reading.
Chris
About Two Lanterns
Two Lanterns Advisory is a political risk consultancy based in Boston, Massachusetts. For information on training courses in political risk, hiring a consultant, or commissioning reports, check us out at http://www.twolanterns.co.