How does insurance respond in times of social and political upheaval?

With much of the Arab world going through political and social upheaval, oil and commodity prices rising sharply, and sky-rocketing food prices putting much of the world’s population back into extreme poverty, the risk of increased civil unrest occurring across the globe is a real possibility. In anticipation of such an event, multinational companies need to ensure that they not only have appropriate risk management and response procedures in place to protect their employees, but that they also have adequate insurance to protect their balance sheets. Similarly, (re)insurers need to carefully review their underwriting exposures, particularly regional accumulations and aggregations, to maintain the functionality on which most of global business depends.


The types of insurance that an insured would expect to apply during periods of unrest include fire and property insurance, business interruption insurance, contingent insurance, employees and general liability insurances, cargo insurance and trade, and credit insurances, to name a few. However, it is important that the insureds and local insurers appreciate that, as a general rule, these insurances will contain war and civil risks exclusions, which are intended to exclude damage caused by specific perils that are often the result of much damage during civil unrest and political upheaval.These perils include (but are not limited to):

  • civil war;
  • civil commotion assuming the proportion of or amounting to an uprising;
  • insurrection;
  • rebellion;
  • revolution;
  • usurped power; and
  • terrorism.

Importantly, some also exclude damage caused by riots – although buy-back endorsements for riots, civil commotion and malicious damage can be purchased for a premium.

Judicial analysis into the meaning of many of these perils has been provided previously by Mustill J (as he then was) in the leading decision of Spinney (1948) & ors v Royal Insurance Co Ltd[1980], which considered physical damage caused during the political strife in Lebanon during the mid-1970s. In his analysis, he addressed the subtle distinctions between many of the terms. A summary of his judicial guidance given on some specific terms is set out below.

Civil war

‘A civil war is a war that has the special characteristics of being civil, ie internal rather than external. Nevertheless, a civil war is still a war. The words do not simply denote a violent internal conflict on a large scale.’

The characteristics of a war include armed conflict between two (or more) sides, each of which owes an allegiance to its respective side, and each side must be capable of being identified by reference to a community of leaders, objectives and administration (although the objective to seize powers of state alone is not sufficient to constitute a war while the absence of such a motive, where all other requirements are present, would not negate against there being a civil war). The number of combatants, strength of armaments, size of territory held, and the effect the conflict has on the population and public administration, will all be factors to be considered.

Usurped power

An archaic term whereby there is an arrogation of the state’s power by a group of people having universality of purpose and using violent means, constituting constructive treason.


Organised resistance to the state with the purpose of supplanting the state’s power (in part or whole).


A limited rebellion or a rebellion having less degree of organisation.

Civil commotion assuming the proportion of uprising

A narrower technical meaning is applied to the term ‘civil commotions’ in the context of the civil risks exclusions. In this case a ‘civil commotion’ is a rising of a considerable body of people for the purpose of general mischief, being something more cohesive than a mindless mob, but not amounting to a rebellion, with the element of tumult being present. As to a ‘popular uprising’ it ‘means the same as insurrection’, with ‘greater spontaneity and looser organisation’, and there must be some unanimity of purpose between a substantial portion of the population involving the displacement of the government and state.


Mustill J did not have to consider this issue, but English statute (Public Order Act 1986) sets out the requirements for riot. A riot exists where 12 or more persons together use or threaten unlawful violence (not necessarily all or all at the same time) for a common purpose (which can be inferred from conduct), which would cause a person at the scene to fear for their personal safety. US state law appears to adopt a looser view than earlier English common law of what should constitute a riot, without a requirement to either a fixed number of people or common purpose, other than requiring the existence of a ‘local disturbance’ (see Pan American Worldwide Airways Inc v The Aetna Casualty & Surety Co [1974]).


As is evident from the above summaries, the difference between some of the perils is subtle and the evidence necessary to prove the existence of an excluded peril is likely to be difficult to obtain. Usually, where an exclusion is relied on, the party seeking to exclude cover (the insurers) has the burden of proving the existence of the excluded peril and that the excluded peril caused the damage. However, because of the significant exposures that arise from war and civil risk perils, and the obvious difficulties in obtaining sufficient evidence to prove the existence of the excluded peril, many of the civil risk exclusions contain a reverse burden clause that shifts the onus onto the claimant. The insured therefore has the responsibility of proving the absence of both the excluded civil risk peril and its causative role if the (re)insurer raises one of the excluded perils as a defence to cover. The reasonableness of such a clause is accepted under English common law, although it is arguable that an English court would only allow (re)insurers to rely on the reverse burden clause once they have produced an arguable case that the excepted peril caused the damage (see Spinney and Grell-Taurel Ltd v Caribbean Home Insurance Company Ltd & ors [2002]).

Importantly, the reverse burden clause, if clearly drafted, can also vary the usual requirement that there must be a direct causal connection between the excluded peril and the damage for the excluded clause to have effect, which will allow all damages indirectly resulting from the excluded peril(s) to be excluded. In that instance, to succeed in reversing the burden, the insured must be able to show that it was more probable than not that the excluded peril(s) did not cause the damage. Proving a negative is never easy, especially when access to information is hampered by civil unrest.

One excluded peril that has proven problematic is that of ‘terrorist’ or the ‘act of terrorism’. Obviously, insurers are not expected to indemnify for damage caused by a terrorist or an act of terrorism, but there are a myriad of drafts in the market with varying degrees of application. For example, the NMA 2918 War and Terrorism Exclusion Endorsement does not exclude damage from riots, but its definition of an ‘act of terrorism’ is broad enough to encapsulate the peril of riot as defined by English statute, which would be a strange anomaly. While it has been stated that ‘in the modern sense, a terrorist is someone who kills, maims or destroys indiscriminately for a public cause’ (see Michael Miller’s Marine War Risks (3rd edition), p195, clause 18.8), the definition of ‘terrorist’ has been incapable of agreement at the United Nations since its inception, as political motivations drive the description. Today’s terrorist can be tomorrow’s freedom fighter and next week’s government. The Pakistan riots following Benazir Bhutto’s assassination and the Thailand disturbances in 2010 are recent examples of incidents in which the prevalence of different definitions of ‘act of terrorism’ or ‘terrorist’ within insurance exclusion clauses have resulted in varying responses from insurers and reinsurers.


For insureds it is imperative that a detailed analysis of the insurance wording is undertaken with regard to the geographical location of business exposures and balance sheet risks to ensure risk management processes properly protect those exposures identified as not being part of the risk appetite. Where there is non-alignment exists, take steps to acquire the relevant endorsements of cover. Insurers likewise need to be aware of their regional and global accumulations and aggregations, and to ensure any facultative reinsurances respond on as broad a basis as the cover provided to the insured(s).