Chartbook 391 How to predict a 1 in 1,000 year riot: the insurance expertise gearing up to commodify America's agony.

This weekend the US faces the prospect of historic confrontation between anti-Trump demonstrators and militarized police. According to the New York Times, “No Kings” demonstrations are planned in all 50 states, and organizers have estimated that there will be roughly 2,000 of them.

Trump, meanwhile, is geeing up the troops with an inflammatory address at the garrison at Fort Bragg.

These protests have the potential to bring together what are hitherto relatively scattered anti-ICE protests into a mass movement, but also to provoke a dramatic escalation of violence by the police and security forces.

The US may well be heading towards a summer of protest on the scale of the Black Lives Matter demonstrations in 2020, or larger.

Given the emergency, Cam and I took up the subject in the pod this week and I went looking for benchmarks with which to orientate ourselves in the escalation that may lie ahead.

In so doing, I stumbled across a truly jaw-dropping “literature” that seeks to quantify and predict civil disorder in the United States and the rest of the world, using statistical models.

In good time for the second Trump Presidency, the insurance industry and security firms have been readying data with which to prepare their customers for the turbulence and losses that may lie ahead. Ultimately, the purpose is to commodify the billion-dollar risks attendant on America’s polarization and mounting civil disorder manageable and thus to make them insurable. The purpose of the analysis is to gauge the extent and intensity of America’s civil disorder, for the purpose of insurance pricing.

In February 2021, in the aftermath of BLM, the World Economic Forum (Davos) published a blog post boldly declaring:

“How 2020 protests changed insurance forever”.

It was the scale of the BLM protests and the police violence that shocked the study of American civil disorder onto the agenda of analysts.

If there are two things the insurance industry really does not like it is #1 large losses and #2 unexpectedly correlated risks.

In recent years its experience with civil disorder risks in the US had ben highly localized, notably the Los Angeles riots of 1992. But 2020 was different, protests spread across the entire country, like a chain-reaction or wildfire. This is bad news for insurers.

The firms who keep track of this data are somewhat coy about the totals. The most widely cited data on civil disturbances and losses in modern America are compiled by an outfit called Property Claim Services (PCS), which, since 1950, has used insurance claims as its basic source. It uses damage of $ 25 million as a minimum cut off point for something it labels as “catastrophic”.

PCS is an affiliate of Verisk Analytics. The firm is not keen to release precise data to the public, since that is how it earns its fees. But, as PCS boss Tom Johansmeyer told Axios, what made BLM different was that:

All previous catastrophes — as classified by the insurance industry — happened in a particular city. This was the first that happened not just in multiple cities, but in 20 states.

"Not only is this the first, this is the first — kind of with a cymbal crash,"

What information we do have available clearly suggests that BLM 2020 was in a league of its own.

Source: Axios

The American riots have their roots in the racism of American society and the manifold issues around urban policing. But the scale of the explosion was not particular to the US. As Mohit Pande Chief Underwriting Officer PropertyView noted in March 2024.

It wasn’t long ago that no single SRCC event had exceeded USD 1 billion in losses, with the most significant instance on record having been the Los Angeles riot in 1992. In more recent times though, large scale SRCC events have accumulated to the extent that SRCC has become a prominent risk topic for (re)insurance industry CXOs and Board of Directors (see table).

The industry has been rocked by riots in France in 2023 and giant protests in Chile, South Africa and Colombia. Nor is it just the scale and severity of disorder but its frequency that worries the insurance experts. The Carnegie Endowment for International Peace’s Global Protest Tracker, records protests in more than 132 countries since 2017, with almost a quarter of them lasting for three months or more.

Source: Allianz

Insurance analysts have worked hard to identify the source of these spiraling losses. In part it is the scale and intensity of the protests that explains the scale of losses. But, if you look more closely at the data you find that losses are highly concentrated.

In the BLM protests in the US in 2020 that resulted in perhaps $ 2 billion in losses, roughly $ 700 m was attributable to just three retailers. In the enormous protests in Chile in 2019 which resulted in a staggering $3 billion in claims, a third came from a small group of retailers, with one business alone accounting for one fifth of the total loss.

The response of the insurance industry is to further localize and specify their models to identify the most likely sites of major losses.

Models start with general factor such as economic inequality and political disaffection, but those very general factors are then twinned with more specific analyses of urban geography, with a view to identifying potential sites of high risk and high property value exposure. The methodologies are well explained in a timely article that appeared on 10 June 2025, by two industry insiders Weimeng Yeo and Tim Brewer, entitled simply: “How to predict a riot: Developing risk-based loss models for social unrest”.

The insurance industry experts begin by surveying which sites might be most suitable for attracting media attention. They noted for instance that in 2020 the protests in Portland Oregon were concentrated around a Federal Courthouse, whereas in Washington DC they centered on Lafayette Square. Whatever might attract protestors now also attracts the attention of insurance experts.

All of these “risk factors” were then fed into “state of the art SRCC models, such as the Synthetik SRCC Quantum Tool” which can identify ‘target points’, such as government buildings, financial institutions, or religious sites. SRCC stands for strikes, riots and civil commotion.

On this basis, Synthetik identified 3% of Minneapolis - ground zero for the George Floyd protests in 2020 - as high-risk. This allows the software to identify both specific targets.

The map below shows 374 SRCC Target Points identified in Minneapolis (blue – economic, orange – religious, green – political)

Figure 1: Screenshots taken from the Synthetik SRCC Quantum Tool, demonstrating simulations of SRCC risk across various regions

Entire neighborhoods, can be identified as more or less high-risk.

SRCC Risk Map for Minneapolis

Proportion of the Minneapolis land area by SRCC risk classification
Only 3% of Minneapolis is at HIGH risk of an SRCC event

Sources: Screenshots from Synthetik SRCC Quantum Tool

The models are intelligent enough not only to identify the basic parameters of urban differentiation, but also to track the dynamics of protests themselves. The models recognize that protests do not spread simply in concentric circles around a “target point” but rather follow “linear pathways” along roads, from one symbolic target to the next.

With the basic parameters specified, the models can then be run over and over again to simulate the full range of possible outcomes in any given metropolitan area, allowing actual riots to be benchmarked against the “maximum likely damage” estimated by the model.

In the case of Minneapolis during BLM in 2020, a Synthetik Quantum SRCC model simulated over 1,000 civil unrest scenarios and predicted a maximum likely loss of $700 million affecting 1,773 properties^[13]. Actual losses in 2020 in Minneapolis were $500 million, impacting just over 1,000 properties, strongly supporting the basic conclusions of the model.

All of this expertise was compressed into the first-of-its-kind catastrophe model which Verisk (Nasdaq: VRSK) released to the market in April 30 2025. As Verisk explains, its model responds to the mounting tide of protests worldwide which it estimates has inflicted $10 billion in insurance losses, far smaller than the losses due to wildfires or hurricanes, but ten times larger than the $ 1 billion in damages caused by terrorism.

Verisk’s model is a finely calibrated model of risk assessment, or - one might also say - scaremongering.

As Verisk Maplecroft informs its potential clients: “A 1 in 1,000-year SRCC event could cause losses 10 times greater than those from the 2020 protests, while very low-probability SRCC tail events could potentially impact commercial and municipal properties at the ZIP code level nationwide, the majority of which are located in metropolitan areas.” The average US ZIP code (postal code) has roughly 9700 inhabitants. So this is quite a granular model of American society.

Zooming out to the level of cities, Verisk offers a chart showing the plot of SSRC risk against property values. All of America’s larger cities have elevated risks. But NYC is in a league of its own.

Source: Verisk Maplecroft

The model then allows Verisk to measure the likely impact of a “1 in 1,000 year event about ten times more intense than the 2020 riots in the United States”. In other words Verisk is helping insurance companies to price what would be the closest thing that the US has seen to a revolution or civil war in its recent history.

Rather than pausing, even for a second, to consider the import of the scenario they are consdering, Verisk blithely continues:

“The need for an ongoing rigorous assessment of SRCC and how it may impact your exposures and pricing is fundamental if you want to mitigate the impacts of the major loss events seen in Chile, the United States, South Africa, France, and New Caledonia in the last five years. The Global SRCC Predictive Scores are available now and benefit from an array of integration options including API, shapefile or CSV delivery, and Touchstone Geospatial.”

The future of the Republic may be at stake, but the insurance industry, at least, will be well-prepared. It is, in the most literal sense, true that it is easier to imagine the end of the American polity as we know it, than to imagine the end of property insurance.

---

Thank you for reading Chartbook Newsletter. It is rewarding to write. I love sending it out for free to readers around the world. But it takes a lot of work. What sustains the effort are voluntary subscriptions from paying supporters. If you are enjoying the newsletter and would like to join the group of supporters, click here: