The outlook after the mutiny

Illustration by Laurel Molly, © All Rights Reserved

In late June, the Wagner Group, a Russian mercenary force led by oligarch Yevgeny Prigozhin, mutinied against the Russian military. Wagner forces captured the city of Rostov-on-Don and began rapidly advancing towards Moscow before turning back after negotiations.

Our forecasting group undertook an effort to make live probabilistic predictions about how events could unfold. In this article, we will review our forecasting judgments during the mutiny and also provide our group's current outlook on key questions facing Russia and Ukraine.

Our forecasts during the mutiny

Historically, judgemental forecasting requires time to be allocated for question writing and review, followed by the steady input of many forecasters over weeks and months. Live forecasting is therefore somewhat novel, and this was a first shot at this style of forecasting for us at Swift.

Additionally, at the Swift Centre, we place a lot of emphasis on conditional forecasting. As well as asking forecasters with incredibly strong track records what the chances of a given event is, we want to understand what the consequences of those events might be. This led to our forecasters generating various questions we thought may be consequential, and changing our focus as events unravelled.

Chances of success or continued conflict

Our initial, top-level question was on whether the Wagner ‘coup' would still be ongoing or have succeeded in 2 weeks after it had started. There was disagreement among our forecasters from the outset of whether we ought to label the event as a coup attempt and, in turn, how we should define a “successful” outcome for Wagner.

At one point, the group’s aggregate forecast on this was as high as 58%. However, the mutiny did not achieve its stated goals. Specifically, Prigozhin had demanded the handover of Defense Minister Sergei Shoigu and Chief of General Staff Valery Gerasimov, whom he blamed for military shortcomings and losses among Wagner forces. But both men remained in their positions after the mutiny ended, and the mutineers did not seize control of the Russian government or overthrow Putin. So, it seems clear that it was not successful — whether regarded as a mutiny or a coup attempt.

There was some internal debate about the success rate of a ‘typical’ coup (which itself is hard to define). One Swift forecaster calculated the average success rate of similar coups to be 28%, though others based their forecasts started from a figure of 50%, as is suggested by academic research. While many of our forecasters did not deviate far from this base rate — as can be wise when it is hard to trust the available information — others placed higher odds on this question resolving positively. This was, in part, because of the possibility of conflict persisting for several weeks, especially given the early indicators of a lack of conciliation, and also because Wagner’s revolt appeared to face very little resistance.

While some of our forecasters feel our forecasts strayed too high, particularly once it became apparent that the Russian leadership had not been isolated, most felt the group’s forecasts were well-justified given the information available. Even in hindsight, it is hard to say that the sequence of events were predictable.

Forecasting the consequences of an uncontained mutiny

We forecasted the chances of the mutiny succeeding, or continuing for 2 weeks, up until 18:36 UTC on June 24. Along with our top-level probabilities on that question, our forecasters provided their outlook for other critical questions along the way. Here is how things looked as news of a Moscow–Wagner deal broke:

Was the deal between Prigozhin and Putin upheld?

The exact details of the deal between Moscow and Wagner remain unclear, but Wagner ceased their advancement, and charges against Prigozhin were eventually dropped. In line with our understanding of the deal, Prigozhin now operates out of Belarus.

It is hard to determine whether the deal was ‘held’ in a strict sense, given reports of Prigozhin’s flights into Russia and Moscow’s delay in dropping charges since the deal was brokered. However, given that neither Moscow nor Wagner appear to have accused each other of reneging on the deal, it would appear it did indeed hold for two weeks, in line with our expectations on June 25:

Updated outlook

As well as providing quick-fire forecasts on the longevity of the Moscow–Wagner deal, our forecasters provided their outlook on broader questions facing Russia in our liveblog. One was whether Shoigu and Gerasimov — the army generals that Prigozhin wanted to unseat — would still be in power on 24 July. We narrowly expected they would be, putting the odds at 52%. By the end of June, as part of our internal forecasting, our confidence that they would remain in power increased to 80%.

As part of our liveblog, we considered the longer-term effects of the mutiny and provided our forecasts on Putin retaining power, Russia withdrawing their forces from Ukraine, and Ukraine successfully severing Russia’s territorial ‘land bridge’ this year. A month later, here is our updated perspective on these events and how they affect Putin's grip on power, along with the outlook for Prigozhin and future rebellion attempts.

For more detail on individual forecasters' reasoning, you can tap/hover over probability ranges to see their comments.

Ukraine severs Russia's territorial land bridge in 2023?

During the mutiny, our forecasters put greater weight on Ukraine successfully severing Russia's territorial land bridge — the region between the Azov Sea and the Dnipro River that links Crimea in the south with Donetsk and Luhansk in the east. This maneuver is a central objective of Ukraine's counteroffensive.

Since that weekend, however, our forecasters have reduced their probabilities, assigning a 33% chance that Ukraine severs the land bridge before year's end. As one explained, "Defence is fundamentally easier than offence, and Ukraine may be thinking that they want to use some of the equipment they have received elsewhere when they have more of a comprehensive force." Progress remains slow on this front.

If the land bridge is severed, our aggregate forecast gives an 82% chance to Putin remaining in power through 2023 (down from the 91% given to him in the median scenario). As one forecaster noted, "This would certainly be a significant setback for the Russian war efforts in Ukraine", and would therefore have to have some effect on Putin's chances of staying in post.

Chances of another Russian military rebellion or attempted coup this year?

The group sees a 10% probability of another Russian military rebellion or coup attempt before end of 2023. As one forecaster explained, "Looking at the history of coup attempts in authoritarian regimes, they do not tend to come in bunches ... especially when there’s a strong central government." The failed Wagner mutiny likely reduced the chances of near-term attempts.

Another noted, "In the aftermath of any failed uprising, you'd expect forces loyal to the incumbent to try to flush out further opposition to the incumbent. Given that there are early signs of this happening, with General Surovikin having been removed from his post, I expect the security forces loyal to Putin will be broadly successful at this and the chances of a military coup will be lessened over the next few months."

If another coup attempt did occur, our aggregate forecast puts 58% on Putin remaining in power until the end of the year. As one forecaster said, "If there's another attempt I think it will be serious, but coups are hard to pull off."

Will Vladimir Putin be President of Russia at the end of 2023?

Overall, the group sees a 91% probability that Putin remains Russia's president through 2023.

Many believe Putin has proven resilience and that his ouster does not appear imminent barring an unforeseen event. As one forecaster explained, "Putin has survived many challenges to his leadership over the past two decades and is probably still a relatively popular leader."

Some think the mutiny may have weakened Putin in the long-run, but are skeptical anyone can remove him in the near-term. As one noted shortly after the mutiny, "While the events of last weekend appear to have weakened Putin, I'm not convinced that that's the case. And, even if it has weakened him, I don't see a clear path to Putin not being President at the end of 2023."

Many believe Ukraine war developments will be key to Putin's outlook. As one explained, "If a collapse of the Russian positions happens in the next few months, then this will raise the chances of Putin leaving office." But most see this as unlikely near-term.

Prigozhin alive at the end of 2023?

As for the other main party involved in the mutiny, the aggregate of our forecasts places an 83% chance of Prigozhin remaining alive through 2023. As one forecaster explained, "For Putin, keeping him alive gives you options - he may want to use Wagner in the war in the future if things go worse, he may want to use them in Africa in a more active situation, he may want to signal that deals are capable of being made with him (to the West)."

Another noted, "It amazes me Prigozhin is still alive, but I think at this point Putin doesn't want to seem desperate or alienate Wagner mercenaries. I don't think Prigozhin longer-term prospects are very good, however."

If Putin is out of power, will Russian forces withdraw from Ukraine?

In the unlikely situation where Putin is out of power, how will things change? 9 out of 10 times, it would not result in Russia withdrawing from Ukraine — at least not before the end of the year.

As one forecaster explained, "I don't think there is any political appetite to do this even if Putin isn't President, so unless they are forced out by military means then I think this isn't happening." Another similar view: "I see very little chance of Russian forces withdrawing to pre-invasion positions whether or not Putin is the Russian President. The war effort in Ukraine has a great deal of support in Russia."

Overview

You can see the overview of our forecasts here (and on this standalone webpage):