Small Wars Journal

Civil Unrest and Natural Disasters: Debunking a Myth That’s Been Around Too Long

Wed, 01/19/2022 - 8:22pm

Civil Unrest and Natural Disasters: Debunking a Myth That’s Been Around Too Long

by Tom Johansmeyer

When there’s a natural disaster, according to conventional wisdom, civil unrest is likely to follow. That line of thinking is as seductive as it is intuitive. However, there isn’t much data to support it. The occasional half-hearted inquiry forced by the lack of empirical evidence generally leaves you as unsatisfied as your first self-cooked Thanksgiving dinner. The notion that natural disaster increases the risk of civil unrest taps into the dark sense that people will revert to survival instincts when the norm is threatened. When it comes to non-terror political violence following natural catastrophes, though, the historical experience just isn’t there. The occasional instance of it is an anomaly at best – and a debatable one at that – with almost no foundation of 70 years of catastrophe data.

The available research on the impact of natural disaster events on civil unrest is notably thin. According to Philp Nel and Marjolein Righarts, “[T]here are surprisingly few studies that systematically explore how natural disasters affect the patterns of politics and/or conflict.” Their research showed an increased risk of civil unrest but was thin on examples, given how few there have been. Others, such as Kettlewell et al, examined the Aceh tsunami of 2004, but the example is complicated by the fact that conflict was already ongoing. A similar issue arises in Ide et al, in a study of flood-related political unrest. The authors did identify 11 countries with flood-related political unrest, although again, questions remain about the role of the natural disaster relative to ongoing political violence or instability. From a broad review of the literature it’s clear that quantification is an important problem; there has been little opportunity to measure the impact of natural disasters consistently, let alone link them to subsequent civil unrest (as opposed to opportunistic crime or other socially unacceptable activity).

The team I lead at PCS, a Verisk business, estimates the industry-wide insured losses from natural and manmade catastrophe events using a proprietary methodology that has been widely accepted by academia (for example here, here, and here). Using data reported by affected re/insurers, independent claim adjusting companies, and other catastrophe response stakeholders, PCS forms an industrywide insured loss estimate for the event. Given the significant spike in civil unrest catastrophes over the past three years, PCS has found it increasingly important to debunk the myths that persist around all forms of political violence to make it easier for risk-bearers and other affected entities to understand the threat better.

To understand the potential interrelationships between natural disasters and civil unrest, both need clear definitions. We used our own definition of “catastrophe,” which requires an event to have an industrywide insured loss above a certain threshold (see table below) and affect a significant number of insurers and insureds.

Catastrophe Reporting Thresholds and Activity by Region


Reporting Threshold

Year of First PCS Event

Number of Events


United States

US$25 million


> 2,000

Thirteen SRCC* events, none related to natural catastrophe activity


C$25 million


> 100

No qualifying  SRCC events


MX$300 million


> 20

No qualifying SRCC events


TRY30 million


> 10

One SRCC event not related to natural catastrophe activity


US$2 billion


> 10

Built primary as a risk-transfer index

Latin America, non-U.S. Caribbean

US$500 million


> 15

Built primary as a risk-transfer index

Source: PCS, a Verisk business
*Strike, riot, and civil commotion – synonymous with civil unrest and riot and civil disorder

Although there are some differences in approach between what is done in the United States, Canada, Mexico, and Turkey and the other regions represented, the foundation is still the same, in that we engage affected re/insurers and other entities to provide their view of the projected ultimate loss of the event (with the former cohort providing more granularity).

PCS views civil unrest as requiring a large amount of people gathered for the purpose of protest, riot, or demonstration in a manner that ultimately involves the destruction of property and other forms of violence. While there may not be a central purpose or agenda to those present for the event, there’s at least an underlying reason or driver that brings the large group together. The acts taken (such as looting) may not further a sociopolitical agenda, but they do result from the original cause that brought out those involved in civil unrest. Opportunistic looting after a hurricane, for example, is not the same as a wave of lootings across all areas affected by a storm because a municipality was slow to provide aid or other relief.

For the United States, PCS has data on 13 catastrophe events going back to 1949. There is no natural catastrophe event in that record that includes among its listed perils “riot and civil disorder.” In the other regions reported by PCS, the findings are no different. Although the historical records don’t go as far back, PCS has no natural disaster events with riot and civil disorder as a secondary peril in Canada, Mexico, or Turkey, not to mention the new regions that have the streamlined methodology: the Asia-Pacific region and Latin America. The lack of riot and civil disorder as a secondary peril in relation to a natural catastrophe event doesn’t result from a lack of reporting capabilities. PCS has reported standalone political violence catastrophe events in the United States, Chile, and Turkey, and we’ve developed informal loss estimates (outside our normal scopes of reporting) for recent civil unrest in France, Hong Kong, Colombia, and South Africa. We’ve seen natural catastrophes and riot and civil disorder –but only separately.

Outside the United States, a review of data in Swiss Re’s sigma publication series shows no cases of civil unrest as a result of natural disasters. In general, natural disasters represent a fairly small portion of global insured losses. In 2018, for example, all economic losses from major man-made events accounted for only US$9 billion, compared to US$155 million in economic losses worldwide from natural disaster events, according to data from Swiss Re sigma.

The only two natural catastrophes where a civil unrest/political violence component could be seen were after the Aceh tsunami in 2004 and the Thai floods of 2010. For Aceh, the tsunami is widely perceived has having helped end a civil war. Further, PCS’s report on the insured loss from the natural catastrophe shows no civil unrest component. The natural catastrophe in Thailand occurred within an environment of civil unrest, making it impossible to link any post-catastrophe civil unrest to the catastrophe itself. The fighting that followed the floods, essentially, was not caused by the floods.

The notion that desperate people will act desperately provides a sense of proportionality that can be extended to any situation where people face threats to their safety and comfort. Natural catastrophe events are no exception. Major disasters have caused extensive damage, loss of wealth, and even loss of life. However, the belief that such profound instances of loss will push people to civil unrest has little in the way of supporting data. Rather than contemplate the possibility of civil unrest after a natural disaster, planning, policy, and relief bodies should focus on new ideas for pre-event resilience and post-event remediation. That could provide real protection for people facing the natural catastrophe risks that could profoundly change their lives.  


Hendrix, Cullen. 2013. Do Natural Disasters Fuel Unrest? Political Violence at a Glance. 19 November. [Accessed 10 December 2021].

Ide, Tobias; Anders Kristensen; and Henrikas Bartusevicius. 2020. First comes the river, then comes the conflict? A qualitative comparative analysis of flood-related political unrest. Journal of Peace Research. 581(1), pp. 83-97.

Kettlewell, Nathan; Fruhling Rijsdijk; Sisira Siribaddana; Athula Sumathpala; Agnieszka Tymula; Helena Zavos; and Nichlas Glozier. 2018. Civil War, Natural Disaster and Risk Preferences: Evidence from Sri Lankan Twins. IZA Discussion Paper Series. 11901.

Nel, Philip and Marjolein Righarts. 2008. Natural Disasters and the Risk of Violent Civil Conflict. International Studies Quarterly. 52(1), pp. 159-185.

Property Claim Services (PCS), a Verisk business. 2021. Everything You Need to Know about PCS: A Full Guide to Catastrophe and Noncatastrophe Insurance Industry Loss Reporting.

Society for Threatened Peoples. 2010. Thailand: Flood disaster exacerbates distress of civil population – peace efforts must be stepped up. ReliefWeb. 5 November. [Accessed 10 December 2021].

Strangio, Sebastian. 2014. Aceh Ten Years After the Tsunami. The Diplomat. 23 December. [Accessed 10 December 2021].

Swiss Re Institute. 2019. Natural catastrophes and man-made disasters in 2018: “secondary” perils on the frontline. sigma. 2.

About the Author(s)

Tom Johansmeyer is head of PCS, a Verisk business, which estimates the industry-wide insured losses from disaster events around the world. He writes and speaks regularly on natural catastrophes, cyber attacks, and political violence events.