Why didn’t we see the Russian invasion coming?
Some thoughts on the analytical prelude to war
At a briefing on Wednesday afternoon in London to discuss the outlook in emerging markets, executives at Ashmore, an investment house specialising in developing countries, kicked off by addressing “the elephant in the room” — Russia.
The country had just recognised the independence of two regions in eastern Ukraine and the US was warning of an imminent broader land grab.
The situation, Ashmore pointed out, was fluid. It demanded a “very, very careful” approach from fund managers. But the core view was that the Russian president would not send troops towards Kyiv. There would be no full invasion and instead, “a good chance that soon we won’t be talking about it”. A few hours later, Russia invaded its neighbour on multiple fronts.
The above, from the FT, is the starkest demonstration that the political risk industry - and the investment and political analysis worlds more generally - did not predict Russia invading Ukraine.
According to the forecasting site Metaculus, the chance of Russia invading Ukraine in 2022 was hovering around 50% for much of January and early February.
We can debate later whether, based on what we knew at the time, that 50% was the correct assessment. What is curious to me is why 50% may have been the correct assessment? Why was the biggest European conflict in decades so difficult to forecast only days before it happened?
Was it mis-predicted?
It’s always hard to know when something unexpected happens whether it was truly missed or correctly forecast as low probability (but then that low probability happened).
The best example of this is the 2016 election. Those who said that Donald Trump had no chance of winning made the wrong call; those who said that Trump was unlikely to win were correct (based on polling and other information available).
So what was the Russian invasion? The unexpected or the unforeseen?
It largely seems that it was unforeseen.
Eurasia Group’s 2021 top risks did not include Russia-Ukraine.
Russia wasn’t in Oxford Analytica’s top ten questions for 2022.
The EIU’s ten risk scenarios for 2022 only addressed Russia in the context of cyberwarfare (which also addressed China and Iran).
In fact, there is no clearer example of this threat being misread than this very public mea culpa from a Russian analyst.
Why was it missed?
So we have the Russian invasion not only being misjudged in its probability, but completely omitted from the global political risk landscape for the past year. This was despite ominous warnings from US intelligence over the last few months and clear signs of Russian mobilization. Why?
One indication comes from Eurasia Group’s 2022 risks, published in January.
President Vladimir Putin is trying to force the West to address Russian objections to the eastward expansion of NATO. He amassed at least 70,000 troops on the border with Ukraine to gain leverage on two issues: the nature of an agreement with Kyiv and the rules of the game on Russia's relations with its western neighbors. Putin has put his credibility on the line; if he doesn't get concessions from the US-led West, he is likely to act, either with some form of military operation in Ukraine or a dramatic action elsewhere.
Diplomacy will probably avert military confrontation by focusing on areas where compromise is possible. Ukraine has no prospect of getting into NATO, but the United States and its NATO allies aren't going to say that publicly. Yet the sides could reach a tacit agreement to that effect. Agreement is also possible on other issues including limits on weapons transfers to Ukraine and on military exercises near the Russian border.
But it's a very close call. Even if issues such as these are agreed on it may not be enough to satisfy Putin. If Putin directly invades Ukraine, Russia will at the least face a ban on US persons trading in the secondary market for new Russian sovereign debt, and NATO forces will move their forward positions closer to Russia's border, raising tensions to a degree not seen since the collapse of the Soviet Union.
EG read the situation as revolving around Ukrainian membership in NATO. It now seems clear the invasion stemmed instead from Putin’s belief that Ukraine should not be an independent country and desire to subordinate it to Russia like he has with Belarus. His speech preceding the conflict and his 2021 essay on Ukrainian history made that clear.
Yet EG wasn’t necessarily wrong to think that. In a straight news article from The Guardian on Feb 14, Putin’s position was framed as exclusively based around Ukrainian membership in NATO:
Putin has voiced concerns that if Ukraine grows its ties with the alliance further, it could be made a launchpad for Nato missiles targeted at Russia. He had said Russia needed to lay down “red lines” to prevent that.
He has also claimed that accepting Ukraine into Nato would lead to military action by Nato to reclaim the Crimea peninsula, annexed by Russia in 2014.
“Do you want France to go to war with Russia?” an angry Putin asked a French journalist during the press conference after his recent Kremlin meeting with the French president, Emmanuel Macron. “That’s what will happen!”
The first problem was that the risk world saw Putin’s essay and his remarks about NATO expansion and saw the first as a red herring and the second as the heart of the matter. It turns out that the opposite was true. The second problem was that the risk world didn’t think Putin would act on this idea.
The two are connected. As Zach Beauchamp at Vox wrote:
Yet the speech helps shed light on a critical aspect of the crisis: why Putin is doing this. Fail to understand the sense of historical grievance, and the way that it shapes his view of the “threat” from Ukraine, and you fail to understand why he’s willing to play such a high-stakes game today.
How to prevent this from happening again
On the one hand, this seems like a huge failure from the political risk world. On the other hand, it seems like not even many in the Russian government saw this coming.
Putin, by many accounts, has changed over the last two years. From NY Mag:
The New York Times has broached the awkward question that has been on the minds of anyone who has paid even passing attention to Vladimir Putin recently: Has he lost his marbles? The Times reports that people have noticed that Putin “has fundamentally changed amid the pandemic, a shift that may have left him more paranoid, more aggrieved, and more reckless.” Two years since the onset of COVID, the Russian leader remains severely isolated, interacting with cabinet officials largely via video and keeping trips abroad to a minimum. When he does have to meet people face-to-face in Moscow, whether it’s Foreign Minister Sergey Lavrov or French president Emmanuel Macron, they must first pass through a “disinfection tunnel” and then sit at a social distance of Olympian proportions, at tables so long that they have become a physical manifestation of Putin’s remoteness from the rest of the world.
Should the risk world have noticed this? Maybe. But Putin was not the most scrutinable leader before the pandemic. And we don’t assume that someone’s personality will change dramatically at 67 years old. Thinking that Putin in 2022 will act much like Putin in 2020 is a reasonable assumption.
So it’s hard to fault analysts for not forecasting this. It’s also hard to fault them for not thinking about it in early 2021. Russia attacking Kyiv was low risk and the precipitating event (Putin appears to sink into delusion during COVID) wasn’t apparent. Perhaps we should have placed a probability on it, but it likely would have been so low that it wouldn’t make it onto anyone’s top ten lists.
Canaries in the coal mines
The real error seems to have been not taking US intelligence warnings seriously.
Analysts shouldn’t simply parrot US intelligence, in the same way that we shouldn’t pass on any government’s announcements without adding our own judgement.
But the volume and intensity of intelligence alarms should have been a warning sign that something had changed. We’re used to a certain tempo and tone of intelligence assessments. In late 2021, those changed. They may have been right or they may have been wrong. They were, certainly, different than before.
For highly unlikely, not-even-being-considered risks, perhaps one of the best ways we can approach them is to consider in advance what would be a similar warning sign. We may not personally know Putin or be able to observe his state of mind, so who would be (in this case, Western intelligence agencies and heads of state who meet him). In that case, for that risk, those are the canaries that we pay particular attention to. For others, we’ll need to find their canaries before it’s too late and we see the risk being realized on the news.