Insurrection in insurance law

The concept of an ‘insurrection’ in the sense of an uprising against the state is probably as ancient as states themselves.  The root of the word is the Latin insurrection, meaning to ‘rise up’. Cicero’s charge against Lucius Catalina in 65BC was that he caused a ‘general insurrection to be raised throughout all Italy … that Cataline should put himself at the head of the troops; the Rome should be set on fire at once and that a general massacre should be made of all the senate … [etc].’

‘Insurrection’ has been an aspect of English criminal law since at least the Treason Act 1351 which, incidentally, is still in force.  Insurrection is not specifically mentioned in that Act but came to be regarded as a matter covered by the act through case law. For example one textbook from 1788 says: ‘those who make an insurrection in order to redress a public grievance … and of their own authority attempt with force to redress it, are said to levy way against the king, although they have no direct design against his person, in as much as they incidentally invade his prerogative’. A 1735 textbook notes simply that insurrection means ‘tumult against the King.’  And of course in English law the monarchy is the embodiment of the state, so the concepts of treason against the King or Queen and treason against the State (the Crown) were inseparable.

The framers of the US Declaration of Independence, many of whom were English lawyers, accused King George III of ‘exciting domestic insurrections amongst us’, essentially accusing the King of treason against his own states and citizens in North America.

By the mid 19^th century non-marine insurers were specifically naming insurrection as an excluded peril in property insurance, at least in North America. In 1899 Lloyd’s marine insurers agreed to exclude war risks from standard policies but, it seems, did not specifically mention insurrection in those exclusions until the mid-1940s. In 1938, in light of losses incurred in the Spanish Civil War, London Market non-marine insurers agreed a standard form of political risk and political violence exclusion, the NMA 464, which excluded:

‘Loss or damage directly or indirectly occasioned by, happening through or in consequence of war, invasion, acts of foreign enemies, hostilities (whether war be declared or not), civil war, rebellion, revolution, insurrection, military or usurped power, or confiscation or nationalisation or requisition or destruction of or damage to property by or under the order of any government or public or local authority.’

The NMA 464 remained the market’s standard exclusion until the 11 September attacks on New York, after which ‘terrorism’ was defined and added to new London Market standard exclusion clauses. These standard exclusions been exported worldwide on the back of reinsurance. One can therefore expect to see ‘insurrection’ excluded in virtually all standard marine and non-marine policies all around the world. On the other hand, the gap in cover created by the standard exclusions gave rise to the war risks / political violence insurance markets from which it is possible to buy cover back specifically for one or more of the usual exclusions. For example the LSW 667, sold in the London Political Violence insurance market,  expressly provides cover for ‘directly caused by war, civil war, revolution, rebellion, insurrection, or civil strife arising therefrom or any hostile act by or against a belligerent power.’

Despite the fact that ‘insurrection’ appears in every policy its meaning is discussed in only handful of cases, as follows.

This case before the High Court in London concerned riot damage and looting to the insured’s supermarket in Lebanon in 1976 against a background of very complex political, ethnic and religious unrest, the details of which are set out at length in the judgment. The policy excluded losses directly or indirectly caused by ‘rebellion or insurrection’, which the judge considered together, as follows:

‘As regards ‘rebellion’ I adopt the definition in the Oxford English Dictionary: ‘Organised resistance to the ruler government of one’s country; insurrection, revolt.’. To this I would add that the purpose of the resistance must be to supplant the existing rulers or at least to deprive them of authority over part of their territory.

The dictionary defines ‘insurrections’ in a similar manner, but also suggest the notion of an incipient or limited rebellion. I believe that this reflects the distinction between the two exceptions as they are used in the present case subject to the rider that a lesser degree of organisation may also mark of an insurrection from a rebellion. But with each exception there must be action against the government with a view to supplanting it.’

On the facts of the case, none of the various factions involved intended to force a change of government, and so the court found that looting did not fall within the rebellion / insurrection exclusion.  (The court did however find that that looting did fall within the exclusion for losses directly or indirect caused by ‘civil commotion assuming the proportions of a popular rising’).

This case concerned an early terrorism insurance policy that expressly covered ‘loss or damage caused by … 1.1.2 any terrorist or any person acting from political motive’, but which  contained an exclusion in the following terms:  ‘In no case shall this insurance cover … loss damage or expense caused by war, civil war, revolution, rebellion, insurrection or civil strife arising therefrom …’

In 1982 to 1983 members of the Mozambique National Resistance (Renamo) blew up a pipeline in Mozambique. Insurers accepted that the losses fell within the cover afforded by insuring clause 1.1.2 set out above. So, although Renamo was a terrorist group, the question for the court was whether ‘it was engaged in civil war, rebellion or insurrection in Mozambique and carried out attacks on the pipeline and tank farm as part of such engagement.’ If so, the loss would be excluded.

The defending insurer’s argument was that, notwithstanding that Renamo could be characterised as terrorists, they were terrorists engaged in an insurrection and so the loss should be excluded.  The insured argued was that Renamo was a ‘fake black liberation organisation’ which was really just a puppet of the Rhodesian secret services and had no real or popular base.  Therefore, insurers said, its activities were not those of a spontaneous internal uprising, and could not therefore amount to ‘insurrection’.

The court noted the description of insurrection in Spinney’s and held that it bore its ‘ordinary business meaning’ and added that:

‘ ‘Rebellion’ and ‘insurrection’ have somewhat similar meaning to each other. To my mind each means an organised and violent internal uprising in a country with, as a main purpose, the object of trying to overthrow or supplant the government of that, though ‘insurrection’ denotes a lesser degree of organisation and size than ‘rebellion’.’

The English court noted in passing, the detailed discussion of ‘insurrection’ in the US Court of Appeals case of Home Insurance Co. of New York v Davila (1954):

‘… it is said: ‘An `insurrection' refers to a revolt, rebellion, or seditious uprising against the government.’ Sometimes the word ‘insurrection’ is used to characterize an outbreak or disturbance more limited in its objective than the forcible overthrow of the government — for instance, where the civil authority in a community has been defied and temporarily rendered impotent in consequence of a labor struggle … But we are dealing here with the meaning of ‘insurrection’ in an insurance policy which expressly covers fires set in consequence of a ‘riot’; and which contains no exclusion of fires caused by a ‘civil commotion’, a category which insurers sometimes include in a clause of excepted perils … Therefore, we think .. there must have been a movement accompanied by action specifically intended to overthrow the constituted government and to take possession of the inherent powers thereof. An insurrection aimed to accomplish the overthrow of the constituted government is no less an insurrection because the chances of success are forlorn. It is no less an insurrection because after it was suppressed twelve reasonable men on a jury might conclude that the uprising was foredoomed to failure from the start. At the time of its breaking out, an insurrection may not necessarily look impressive either in numbers, equipment, or organization.’

The court in Sturge embarked upon a detailed examination of Renamo’s history and activities and found that, notwithstanding that it was strongly supported by Rhodesia it had independent aims and ambitions of its own, hated the government of Mozambique and wished to overthrow it.

Accordingly, the court concluded that the losses were: ‘caused by an insurrection within the meaning of the policy. Renamo … was an organisation of Mozambicans who, with Rhodesian and South African encouragement and practical assistance had risen up with the intent of overthrowing the Mozambique Frelimo government and who carried out the attacks for that purpose.’ The loss was therefore excluded by the policy.

This decision of the Court of Appeal of Trinidad & Tobago concerned the coverage of riot and looting losses occurring in the aftermath of the invasion of the legislature by a small group.

The insurance policy specifically covered Riot but also contained the standard exclusion for ‘loss or damage occasioned by or through or in consequence directly or indirectly, of … (b) insurrection … (c) Acts of Terrorism.’

On 27 July 1990 in Port of Spain Trinidad a sentry at police headquarters was shot and killed, a small number of people then set of an explosive device in the courtyard causing the building to be destroyed by fire. 42 members of the Muslimeen stormed into parliament (the Red House) whilst it was in session, discharging bullets. The Prime Minister was taken hostage. 70 Muslimeen took over the television station after which their leader appeared on national television and announced that ‘the government of Trinidad & Tobago has been overthrown.’ On 1 August the participants surrendered and were charged with treason. Immediately after the television broadcast looting and rioting broke out in Port of Spain.

The insurers and insured agreed that: the immediate cause of loss was ‘riot’ (a covered peril) and that actions of the Muslimeen were acts of insurrection and terrorism. The question was whether the riot damage to the insured’s premises was excluded because it was ‘directly or indirectly’ caused by the insurrection.

The court considered the purpose of the policy was to provide only qualified cover for riot. Cover did not extend to all kinds of riots; it excluded riots losses that were both directly and even only ‘indirectly’ caused by insurrection. There was no evidence before the court of a direct link between the riot losses and the insurrection but the court had no difficulty in inferring an indirect link between them. The losses were therefore excluded.

At 10:45 on 19 May 2000 George Speight and several armed confederated seized the Parliament and took the Prime Minister and other government members hostage. A breakdown of law and order in Suva followed and at 1249 a riot began at the claimant’s shop that a mob looted and damaged.

The claimant’s premises were insured under a Lloyd’s policy for all risks of physical loss or damage. The policy contained the usual exclusion for: ‘loss or damage directly or indirectly caused by or resulting from: war … insurrection [etc]’.

The Supreme Court of Fiji, which included a British and Australian judge, surveyed all of the cases mentioned above and gave the following very clear summary of the propositions they established:

‘(1) an ‘insurrection’ is an attempt by force to overthrow the established government. This depends on the objective of those involved, particularly the leaders. It does not depend on their prospects of success.

(2) The numbers taking part need not be large, and there need not be a high level of planning. An ‘insurrection’ may be a loosely organised affair.

(3) The exclusions in this policy are not limited to damage proximately or directly caused by an excluded event.

(4) The exclusions cover damage ‘indirectly’ caused by an excluded event. How far the chain of indirect causation extends in time and space is a question of fact and degree calling for judgment.

(5) The fact that the looting was for personal gain does not establish that the ‘insurrection’ was not an indirect cause.

(6) The fact that the insurgents were not involved in the looting and did not instigate it does not establish that the ‘insurrection’ was not an ‘indirect’ cause.’

There was no evidence that Mr Speight and his confederates planned or instigated the looting or arson. Accordingly the looters were not part of the insurrection and it was not part of the direct cause of the loss or damage suffered by the claimants. The coup was however the catalyst of the breakdown of law and order. Insurers were able to show that when the rioting broke out at the claimants’ shop the insurrection had not become merely ‘a matter of history’ but was causative and close enough in time and space to be an indirect cause of the petitioner’s loss and damage. Accordingly, the loss was excluded.

Insurrection is a term of great antiquity in English criminal law but it is only relatively recently that it has been specifically considered in the context of insurance law.  The meaning is now clear, and is very well summarised in the Tappoo judgment mentioned above. In our experience disputes tend not to focus so much on the meaning of ‘insurrection’ but on the evidence and whether there is a sufficient causal link between the insurrection (or other excluded political violence peril) and the particular loss to bring it into play.

Spinney’s, Tappoo and Grell-Taurel are good illustrations of common situations, where the direct cause of damage to property is rioting and looting by persons unknown against the backdrop of on ongoing or recent excluded political violence peril, such as war, insurrection, or rebellion etc. The policy covers riot or looting on the one hand but excludes, for example, losses indirectly caused by insurrection on the other.  How are the clauses to be applied? The common mistake when adjusting claims is to try somehow to assess if the proximate cause of property damage looks ‘more like a riot’ or an ‘insurrection’. That is, however, not usually the correct approach and will often lead to the wrong conclusion. As the courts have made clear, the conjunction of cover for riot on the one hand, and exclusions for losses directly or indirectly caused by political violence such as insurrection etc on the other means that the policy does not cover all kind of riot. Rather, it only covers losses directly caused by riot where that riot is not, even indirectly, caused by an excluded peril.

On other words, the use of the concept of ‘indirect causation’ in the exclusion displaces the usual proximate cause test. That means that losses will still be excluded if they are directly (proximately) caused by a covered peril but indirectly caused by an excluded one.

How far back in the chain of causation an ‘indirect’ cause of damage must be a question of fact and degree calling for judgment and in the cases show that, where there is evidence of a background of political violence, the courts are quite prepared to infer, essentially using common sense, a connection between that and particular rioting and looting events, even in the absence of evidence of any firm connection. Moreover the fact that particular rioters and looters may have been motivated by personal gain (as is often the case) does not itself negate the effect of the exclusion.

In short, the standard market exclusions were intended to insulate standard property insurers from losses even indirectly connected with political violence perils, including war, civil war, rebellion, insurrection and so on, and the courts have shown that they will give effect to them.