Catastrophe risk, insurance and terrorism

This article was downloaded by: [The UC Irvine Libraries]On: 21 October 2014, At: 20:37Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Economy and SocietyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/reso20

Catastrophe risk, insurance and terrorismRichard Ericson a & Aaron Doyle ba University of Oxfordb Carleton UniversityPublished online: 04 Jun 2010.

To cite this article: Richard Ericson & Aaron Doyle (2004) Catastrophe risk, insurance and terrorism, Economy and Society,33:2, 135-173, DOI: 10.1080/03085140410001677102

To link to this article: http://dx.doi.org/10.1080/03085140410001677102

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/loi/reso20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/03085140410001677102

http://dx.doi.org/10.1080/03085140410001677102

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

Economy and Society Volume 33 Number 2 May 2004: 135–173

Copyright © 2004 Taylor & Francis LtdISSN 0308-5147 print/1469-5766 onlineDOI: 10.1080/03085140410001677102

Catastrophe risk, insurance and terrorism

Richard V. Ericson and Aaron Doyle

Abstract

This article empirically investigates how the terrorist activity of September 11, 2001,was addressed by the insurance industry and government in the United States. Itshows that the insurance system worked reasonably well in compensating lossessuffered, albeit with various tribulations. It also demonstrates that the insuranceindustry, along with government as the ultimate risk manager, imaginatively reconfig-ured markets to continue terrorism insurance coverage in many contexts. Thefindings challenge many of Ulrich Beck’s contentions about catastrophe risks andinsurability. At the same time, they indicate the fragility of the insurance system.Insurers’ perceptions and decisions about uncertainty – with potential for windfallprofits as well as catastrophic losses – create crises in insurance availability andpromote new forms of inequality and exclusion. Hence, while the insurance industryis a central bulwark against uncertainty, insurers can also play a key role in fostering it.

Keywords: risk; uncertainty; catastrophe; terrorism; insurance; governance.

Catastrophe risks and insurability

Debates about risk society have focused on new catastrophe risks and whetherthey are insurable (Beck 1992a, 1992b, 1992c, 1994, 1999, 2002; Giddens 1990,1994, 1998; Dean 1999; Ericson

et al

. 2000, 2003; Ericson and Doyle 2004;Bougen 2003; O’Malley 2003). These debates originate with Beck, and we shallinitially focus on his arguments and the problems with them.

In imagining ‘world risk society’, Beck (1999) distinguishes between firstmodernity (modern, industrial society), in which risks were deemed calculableand subject to management, and second modernity (post-industrial, post-risk-calculation society), in which there are dangers that run out of control.

Professor Richard Ericson, All Souls College, University of Oxford, Oxford OX1 4AL, UK.E-mail: [email protected]

01 RESO330201.fm Page 135 Wednesday, May 12, 2004 9:39 AM

Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

136

Economy and Society

His focus is not on natural hazards, such as earthquakes, that have alwaysplagued human settlements, but on ‘

un

natural, human-made, manufactureduncertainties and hazards beyond boundaries we are confronted with’ (Beck2002: 41). Indeed, he says that these new uncertainties – such as those that resultfrom nuclear, biological, and chemical production – are not simply ‘beyondboundaries’, but are ‘de-bounding’ because they transcend existing boundariesand eventually transform them. They do so spatially (e.g. across nation-stateboundaries), temporally (e.g. they are long-tailed and, in some cases, seeminglyinfinite), and socially (e.g. they create social trouble regarding attributions ofliability, accountability, responsibility, and response-ability).

Seen through the lens of first modernity, these new uncertainties are butnegative side-effects of accountable and rational activity: residual risks that canbe subject to cost-benefit analysis and future control through advances inscience and technology. Seen through the lens of second modernity, these newuncertainties delegitimate institutional authority, erode scientific rationality,create distrust in technologies, and insinuate radical doubt about the future.They ‘abolish the four pillars of risk calculus’: compensation, limitation,security, and classification (Beck 1992c: 102). They contribute to a strongeremphasis on the responsibility of individuals, organizations, and communities toembrace more risk (Baker and Simon 2002). They also engender a precautionaryapproach to a wide range of risks that have catastrophic potential (Ewald 2002).

The private insurance industry is treated as a pivotal institution in the trans-formation from first modernity to second modernity. It is in the business ofapplying advances in science and technology to practices of risk spreading andharm minimization. It is also at the vanguard of the new emphasis on embracingrisk and exercising precaution.

Beck uses lack of insurability as

the

indicator that we now live in risk society(second modernity) which, given his views, he should refer to as uncertainsociety. For example, in

World Risk Society

we count fifteen separate passagesabout private insurability and the transition to risk society. In a succinct state-ment, Beck declares, ‘It is the insurance companies which operate or mark thefrontier barrier of risk society’ (1999: 77). A full expression of his views iscontained in the following passage.

[In risk society] there may be a failure of the norms and institutions developedwithin modern society: risk calculation, insurance principle, the concept of anaccident, disaster prevention, prophylactic aftercare (Ewald 1991; Bonß1995). Is there a ready indicator of this? Yes, there is. Controversial industriesand technologies are often those which not only do not have private insurancebut are completely cut off from it. This is true of atomic energy, geneticengineering (including research), and even high-risk sectors of chemicalproduction. What goes without saying for motorists – not to use their carwithout insurance cover – seems to have been quietly dropped from wholeindustrial branches and sunrise technologies, where the dangers present toomany problems. In other words, there are highly reliable ‘technological pessi-mists’ who do not agree with the judgment of technicians and relevant


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

Richard V. Ericson and Aaron Doyle: Catastrophe risk, insurance and terrorism

137

authorities about the harmlessness of their product or technology. Thesepessimists are the insurance actuaries and insurance companies, whoseeconomic realism prevents them from having anything to do with thesupposed ‘nil risk’. World risk society, then, balances its way along

beyond thelimits of insurability

.(Beck 1999: 31–2)

Bolstered by his world view, Beck repeatedly declares what he thinks is unin-surable: ‘nuclear power plants have suspended the principle of insurance’ (ibid.:55); ‘industrial-technical projects are not insurable’ (ibid.: 77); ‘nuclear, chemi-cal, ecological and genetic engineering risks…cannot be compensated for orinsured against (Beck 1994: 2)’ (ibid.: 77).

Beck hinges his views on the assumption that insurers will insure only whatthey can calculate through scientific expertise applied from the field of riskconcerned. He imagines that, in first modernity, scientists, including actuarialscientists, could routinely calculate risks and thereby manage them. ‘[D]ecisionswere previously undertaken with fixed norms of calculability, connecting meansand ends or causes and effects’ (ibid.: 4). He imagines that, in second modernity,calculation and control evaporate. ‘These norms are precisely what “world risksociety” has rendered invalid. All this becomes very evident with private insur-ance, perhaps the greatest symbol of calculation and alternative security’ (ibid.:4).

There is undoubtedly a trend towards more large-scale catastrophic lossesthat create problems for the insurance industry (Froot 1999; Bougen 2003).However, Beck’s contentions about a massive exit of insurers from catastrophicloss coverage are overstated. Three key points need to be kept in mind.

First, insurers have always been selective about the risks they assume. Forexample, in the United States, which has about 35 per cent of all insurancecapacity in the world, there is a long history of variously excluding particularrisks, partnering with government on other risks, and/or creating special insur-ance arrangements (captives, facilities, alternative risk transfer, etc.) beyond theregular private insurance market (Moss 1999, 2002). Some risks that haveproven too difficult to insure in the private market – for example, the risk ofunemployment or of flooding in specific regions – are addressed by governmentinsurance schemes. Other such risks are not insured at all.

Second, many of the fields imagined by Beck to be uninsurable are insured.For example, nuclear energy facilities in Canada, the United States, and othercountries are routinely insured through special facility pools involving theinsurance industry in collaboration with state regulators (Ericson

et al

. 2003).While in some times and places insurers are indeed what Beck terms ‘techno-logical pessimists’ and ‘economic pessimists’, they also take risks for profit. Inother words, they speculate and gamble. They thrive on uncertainty, to thepoint where in some respects their calculation of risk is not so much a matter offrequency and severity, but rather threat and opportunity.

Beck contends that insurers can no longer insure broad areas of catastrophic riskbecause of problems in scientific knowledge and in technical control of risks. His


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

138

Economy and Society

specific focus in this regard is the incalculability of these risks within the conven-tional formula of frequency and severity. However, as Dean (1999: 183–4) hasnoted, insurers routinely handle such incalculabilities by converting them intotheir own capital risk logic, and by using non-scientific forms of knowledge (seealso O’Malley 2003; Ericson and Doyle 2004). Indeed the insurance industry hasalways acted in this way. For example, the life insurance industry developed longbefore the invention of actuarial tables (Clark 1999).

The capital risk logic of insurance is based on loss ratios: premium revenueminus administrative expenses minus loss prevention expenses minus reinsur-ance premiums minus claims paid plus reinsurance claims plus investmentreturns (Ericson

et al

. 2003: 101). Where the insurer feels that a risk can behandled through an acceptable loss ratio, it may be insured regardless of scien-tific and technical uncertainty. This decision depends on the financial conditionof each insurance company. Indeed, each company will have a different defini-tion of catastrophe depending on its loss ratio arrangements and financial condi-tion. A loss of several hundred million dollars may be a catastrophe for onecompany and indeed spell insolvency, while another company can absorb such alarge loss in the normal course of business and not experience it as a catastrophe.

Insurers make decisions in conditions of uncertainty as a daily routine (Ericsonand Doyle 2004). Scientific data on risk are variously absent, inadequate, contro-versial, contradictory, and ignored. Insurers impose meaning on uncertaintythrough non-scientific forms of knowledge that are intuitive, emotional, aesthetic,moral, and speculative. Of course the nature of the uncertainty and the responseto it varies substantially across various types of threat. Insurers confront realempirical variation in these types that is not simply reducible to their own scienceof actuarialism, technologies, and organizational practices. Insurance is a highlydiverse and adaptable industry, with multiple logics and capacities for addressingthreats and the limits of their knowledge about them.

In his article on world risk society and terrorism, Beck (2002: 41) recognizesthat insurers sometimes make non-scientific decisions to insure in conditions ofextreme uncertainty (non-calculability). However, he refers to these efforts as adesperate attempt to ‘feign control over the uncontrollable’ (ibid.). In our view,insurers’ decisions in conditions of extreme uncertainty are not cynical exercisesin feigning control. Rather, they are necessary expressions of authoritativecertainty in face of limited knowledge, allowing definitive courses of action to bepursued. There is a real need to finance damages even if considerable risk-takingis involved, and even if new configurations of preventive security and compen-sation must be imaginatively created.

Beck (1999: 142) feels that ‘damages can no longer be compensated financially– it makes no sense to insure against the worst-case ramifications of the globalspiral of threat’. Here he is pointing to catastrophes such as Bhopal andChernobyl, which have caused human and environmental devastation not onlyto those immediately affected, but also to future generations. Beck’s message isan important one: insurance cannot replace the loss of loved ones nor one’streasured environments and personal effects. However, it can protect against


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


139

loss of capital by offering financial compensation for loss. While there is alwaysan issue of fair compensation, many victims of earthquakes, hurricanes,terrorism, asbestosis, and nuclear, biological, and chemical accidents arethankful for what insurance can offer.

Third, Beck, as well as other participants in debates about risk society, makeshis assertions about insurability without empirical evidence regarding how theinsurance industry actually operates in conditions of uncertainty. As Bougenobserves, ‘Just how the catastrophic underwriter arrives at key decisions hasproved a source of fascination for many years, yet literature on the topic isparticularly scarce’ (2003: 258). Similar comments have been made in otherfields, such as political science: ‘How and why the insurers and risk managersexercise such power over outcomes and with what consequences for the worldmarket economy and for the allocation of values among social groups, nationaleconomies and business enterprises is a fundamental question for contemporaryinternational political economy. For fifteen years I have waited, in vain, forsomeone to write a definitive account’ (Strange 1996: 123). In debates on risksociety, this lacuna is especially remarkable because much of the debate hingeson how insurers make decisions as gatekeepers of risk and uncertainty at thefrontiers of risk society.

The empirical project is to investigate how insurers make new threats govern-able through preventive security initiatives and innovative forms of capital riskdistribution and compensation. In pursuing this project, researchers must besensitive to the variability of types of threat and how they are responded todifferentially by various private insurance and state officials. Threats, and therationalities of risk in response to them, are multiple and heterogeneous. Theinsurance-based governance of threats involves diverse elements assembled indifferent ways (Ericson

et al

. 2003; Ericson and Doyle 2004).Insurers play key but often hidden roles in establishing preventive security

and loss prevention infrastructures, whether based on environmental design,electronic surveillance technologies, or private security operatives (ibid.). Theirinitiatives in this regard are increasingly within the precautionary principle,which emphasizes that low frequency but high severity risks must be addressedthrough extraordinary control measures that reflect ‘zero tolerance’ and aspireto ‘zero risk’ (Ewald 2002).

O’Malley criticizes Beck for claiming that the insurance industry influencesthe control of risks, and asserts that insurance is only a system for compensatingloss.

But in what sense does insurance control threats? Insurance is a means ofdistributing risks in order to provide compensation after the event. It does not‘control’ the source of the risk any more than do such venerable methods as‘state intervention’ in the form of disaster relief funding, which distribute theimpact of risks through taxation and related ‘spreading devices’.

(O’Malley 2003: 276)

O’Malley’s contention is peculiar, especially since his own research shows how


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

140

Economy and Society

insurers are crucial mobilizers of loss prevention (O’Malley 1991, 1992) andbecause he cites Higgins (2001) regarding efforts to redefine drought in Aus-tralia from natural disaster to manageable risk and thereby responsibilize farm-ers into being agents of loss prevention.

Higgins’ research exemplifies how catastrophe threats are subject to diverseformulations with direct implications for insurance systems of loss prevention aswell as compensation. Another illustration is the Canadian approach to flashhailstorms, one of the largest sources of catastrophic loss in Canada. After heavylosses in the 1990s, the insurance industry and government collaborated in a lossprevention programme that involved having airplanes drop silver iodide instorm clouds to make the resulting hail smaller and therefore less damaging.Aware of their own risk-taking in this risk reduction effort, the organizationsinvolved formed a separate corporation called the Alberta Severe WeatherManagement Society. This move provided corporate limited liability as a way ofprotecting against liability claims that might ensue if the experiment went badlywrong and was blamed for further damage.

Although Bougen (2003) focuses on forms of financing catastrophe risk thatare still relatively insignificant, he at least shows that there are innovativetechnologies at work to help absorb catastrophic loss and at the same timedistribute risks to insurance industry capital. He also demonstrates some of theways in which scientific uncertainty about threats can be converted into thecapital risk logic of insurance, and thereby indicates that uncertainty and risk arenot mutually exclusive. Moreover, he argues that some of the new technologiesof capital risk distribution articulate with the emphasis on embracing risk char-acteristic of neo-liberalism. Hence, the growing precautionary logic with respectto preventing catastrophe risk does not displace neo-liberal risk-taking logic.Rather, the two logics coalesce. Catastrophe risk and neo-liberalism meet where‘a seemingly insatiable appetite for financial protection from danger [intersects]with a seemingly equally insatiable appetite for the investment potential of risk’(ibid.: 257).

Bougen’s contributions are important because they help to modify anotherclaim by Beck, namely, that new catastrophic losses spell the end of neo-liberalregimes. For example, in his rumination on the world after the terrorist activityof September 11, 2001 (hereafter ‘9/11’), Beck proclaims, ‘the terrorist attackson America were the Chernobyl of globalization. Just as the Russian disasterundermined our faith in nuclear energy, so September 11th exposed the falsepromise of neo-liberalism … in times of crisis neo-liberalism has no solution tooffer’ (2002: 47–8).

In what follows we demonstrate empirically the solutions neo-liberalism hadto offer in the wake of 9/11. Our particular focus is how neo-liberal responses tothis catastrophe were negotiated between the insurance industry and UnitedStates government.

1

Our analysis addresses ‘the as-yet-unanswerable questionof whether the insurance and reinsurance techniques that are being assembledinto “catastrophe insurance” will be effective when the crunch comes’(O’Malley 2003: 276). 9/11 was by far the largest event-specific catastrophe


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


141

‘crunch’ in history. Yet, as we shall demonstrate, the insurance system workedreasonably well in compensating losses in this context, and in subsequentlydevising new means of preventive security and capital risk distribution withrespect to continuing terrorism insurance coverage.

We follow O’Malley’s (2003: 276) suggestion that the best way forward in risksociety debates is ‘to chart the variety of ways in which catastrophe risks are

already

being governed in this new environment’. Our case study of how theinsurance industry and state addressed the catastrophe of 9/11 spells out in onecontext what world risk society actually looks like to decision-makers within itand the central role of insurance in that society.

Terrorism as intentional catastrophe

For insurers, terrorism is intentional catastrophe. Like insurers, terrorists areexplicitly in the business of uncertainty. They play on randomness to keepwhole populations in fear, anticipation, and disestablishment. They precipitatethe urge for more certainty, expressed through escalating security measures. Atthe same time, they are adept at grasping the rationality of each new securitysystem in order to subvert it and induce more uncertainty.

The terrorist power of uncertainty is especially strong because we do live in arisk society. This society is characterized by a cultural desire to tame chance andeffect security, and by institutions increasingly organized around risk manage-ment (Garland 2003). Terrorism strikes at the foundation of risk society becauseit is a stark reminder of the limits to risk assessment and management. Humanbeings can always subvert rational systems by manipulating their rationality.The terrorists involved in the 9/11 events organized their attack without policeintervention, passed through airport security on the day, and used the techno-logical power of turning airplanes into bombs to create catastrophic loss. Theyhit home as never before the potential ‘ungovernability of modern societies’(Stehr and Ericson 2000): how those with little power can work cheaply andefficiently against powerful institutions to destroy.

As of August 2002, the estimated insured loss for 9/11 was $55b. This was thefirst insured loss from terrorist activity that exceeded $1b (Swiss Re

Sigma

No.1 2002: 17). The previous highs for any type of event-specific catastrophe ininsurance history were $21b for Hurricane Andrew in Florida and $15b for theNorthridge earthquake in California. According to the estimate of a reinsurancecompany in March 2002, there was also $50b uninsured economic damagearising from 9/11, making the total loss more than $100b.

Insured losses were suffered across various lines of insurance. According toestimates as of 31 January 2002 (Swiss Re

Sigma

No. 1 2002, citing TillinghastTowers Perrin), the losses in various lines were as follows: property $10–12b,business interruption $3.6–7b, workers’ compensation $3–5b, aviation $3–6b,liability $5–20b, other non-life $1–2b, life and health $4.5–6b.

Many insurance company groups suffered heavy insured losses arising from


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

142

Economy and Society

9/11 claims. Very rough estimates at an early stage of claims processing (A. M.Best, 31 December 2001; pre-tax, net of reinsurance) give some indication of thesignificant impact on large groups. For example, Lloyd’s experienced an esti-mated $9b in claims, although $6–7b was recoverable from reinsurers theythemselves went out to. Munich Re reportedly lost $2.37b and Swiss Re $1.07b.Swiss Re’s loss is now known to be greater than this figure, and for the first timein their history they reported an overall loss for the fiscal year.

The tragic loss of life from the 9/11 terrorist attacks also far exceeded anyprevious terrorist event. While 3,212 lost their lives from the terrorist attacks on9/11, the next nine most deadly terrorist attacks each involved a few hundredvictims (Swiss Re

Sigma

No. 1 2002: 17).

Uncertainty and risk prior to 9/11

In some respects there was neither uncertainty nor risk regarding terroristactivity of the magnitude that occurred on 9/11. A catastrophic loss such astotal destruction of the World Trade Center (WTC) was simply not on theradar screen of most insurers. Indeed, as one interviewee observed, if someonehad imagined a total loss of the WTC, he would have been deemed crazy anddismissed. Furthermore, it would have been impossible to sell insurance toclients based on a total loss scenario because they also did not perceive such arisk. The interviewee observed that insurance is ‘very much an after-the-factindustry, because it’s hard to sell something to clients if it hasn’t happened, ifthey don’t perceive it as a risk’.

In other respects there was a known risk. Insurers had suffered significantlosses in terrorist attacks over the previous decade. For example, an IRA bombin London’s financial district in 1992 resulted in an insured loss of $671m, andanother one in the same district the following year cost $907m. An IRA bombingin Manchester in 1996 cost $744m. The 1993 bombing of the WTC by al-Qaedaresulted in an insured loss of $725m.

Following these previous losses, insurers took measures to limit theirexposure to terrorism-related losses. After the bombings in the London finan-cial district, insurers and reinsurers moved quickly to exclude terrorismcoverage. However, they were brought back into the market through the estab-lishment of a pool, known as Pool Re, among over 200 companies, and theinvolvement of the British government above pool limits (Walker andMcGuiness 1997). After the 1993 WTC bombing, one of the major reinsurancecompanies revised some underwriting criteria and the insured limits on itsWTC contracts. These decisions saved this company from even greater losses inthe 9/11 event. Another leading reinsurance company produced a documentafter the 1993 WTC event stating the need to restructure the underwriting ofinsurance and reinsurance coverage for terrorism fundamentally. However, aswe shall learn, such advice was not taken up. It took the 9/11 insured loss of$55b to precipitate significant change.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


143

Some insurance executives were aware of the literature on new forms ofterrorism prior to 9/11. For example, a reinsurance executive said he was awareof a Rand Corporation publication in which there is considerable detail onreligiously motivated terrorism, including reference to the continuing goal oftotal destruction of the WTC twin towers (Lesser

et al

. 1999).This publication and others (e.g. Stern 1999) mapped the rise of a new

‘religious terrorism’ over the 1980s and 1990s. These publications emphasizedthe peculiar dimensions of this terrorism, which make it especially difficult topredict and control. Its support base is spread geographically rather thanconcentrated in a particular jurisdiction. It has amorphous constituenciesbeyond specific state governments, which free it to use weapons of massdestruction without alienating a state-based constituency. Its strategy is irreg-ular attacks, playing on randomness and the uncertainty and fear it fosters. AsBeck observed at the same time:

the danger of regional or global self-destruction through NBC weapons hasby no means been exorcized – on the contrary, it has broken out of thecontrol structure of the ‘atomic pact’ between the superpowers. To thethreat of military conflict between states is now added the (looming) threatof fundamentalist or private terrorism. It can less and less be ruled out thatthe private possession of weapons of mass destruction, and the potentialthey provide for political terror, will become a new source of dangers in theworld risk society.

(Beck 1999: 36)

Journalists also traced the evolution of al-Qaeda terrorism in regular reports,including interviews with Osama bin Laden and other major figures, in whichterrorist activity was threatened (for a retrospective account, see Miller

et al

.(2002)). For example, they identified Ramzi Yousef as the mastermind behindthe 1993 WTC attack and his subsequent unexecuted plot to explode twelvejumbo jets more or less simultaneously. Yosef was connected with bin Ladenand others in the US embassies in Africa plot, the USS Cole bombing, and 9/11 (ibid.). These patterns were also under active investigation by the FBI(Public Broadcasting System,

Frontline

, 2002).While this prior knowledge was not readily expressible in conventional risk

assessment terms of frequency and severity, it was nonetheless evident thatterrorism was a catastrophe risk. In this context, an obvious question arises: whydid insurers and reinsurers not severely limit or exclude terrorist activity fromtheir insurance contracts? We offer four reasons.

First, as mentioned at the beginning of this section, many insurers did notvisualize a total loss scenario for major commercial properties such as the WTC.The data available from past terrorist events gave no indication of the severity ofthe 9/11 insured loss. Indeed, previous loss magnitudes suggested that themaximum terrorist activity loss would be in the millions rather than billions.There were no developed probable maximum loss (PML) models on terroristactivity comparable to those in use regarding other types of catastrophe, such as


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

144

Economy and Society

earthquakes and storms. Instead, insurers simply assumed partial loss in majorcommercial properties. For example, PML scenarios for the WTC imagined thedestruction of two floors. An interviewee observed:

The magnitude of the loss in the 1993 [WTC] bombing was not that great.…It caused injury and it caused physical damage to the property, but it didn’tbring the building down. And so again they said, well, “What was our loss?How many millions of dollars? That’s not so bad. Keep on going”.

Second, some insurers said they were abreast of patterns in the new religiousterrorism but did not include it in their portfolio of catastrophe risks. Theyoffered a version of Mary Douglas’s theory (Douglas 1966, 1992; Douglas andWildavsky 1982; see also Hacking 2003): among all catastrophe risks, only a fewcan be selected for attention on the basis of social, political, economic, and cul-tural considerations. A reinsurance executive said:

An international reinsurer is supposed to have a certain experience withcatastrophes. But the question is . . . given the diversity of liability scenarios,how can we meaningfully process such experience? This is scarcely possibleusing actuarial methods alone. It calls for professional methods that areprobably more akin to those of a social historian than those of an actuarialscientist or legal expert.

He went on to say that his company had heavy losses in the 9/11 attacks, and inprevious terrorist attacks in different parts of the world. He also said he wasaware of the literature on the new terrorism, including that which indicated col-lapse of the WTC twin towers had been a goal.

So the problem is not know or not to know, the problem is . . . what youshould consider. . . . This is a sophisticated game, obviously, but this sophis-ticated game is based on the selection of four, five, six catastrophes out oftwenty, forty, fifty. And this selection is absolutely based on, I don’t knowwhat,

casual

developments. And so the whole basis of this sophisticated andrational game is not so sophisticated.

Third, there were insurance conventions and legal restrictions on excludingterrorism in different lines of insurance. For example, most employees who losttheir lives in the WTC had group life policies through their employers. Thesepolicies generally have few, if any, exclusions, and not ones that would apply todeath arising from terrorist activity. Employees who were injured were coveredby compulsory workers’ compensation insurance policies. State regulations forsuch policies require coverage for anything that happens at work, including ter-rorist activity. Again terrorism coverage was simply embedded in policies with-out a thought. In the commercial property line, terrorism was part of ‘all risk’coverage. Insurers were legally obligated to offer a limited, but significant,amount of terrorism coverage for property losses in any United States jurisdic-tion with a statutory fire insurance policy. Indeed, it is likely that, even if agiven state’s Department of Insurance regulatory body approved a terrorism


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


145

exclusion, any loss from terrorist activity involving fire would still be covered(see Baker 2003: ch. 4, and in particular his discussion of Watson V. USAA, 566N.W.2d 683 [Minn. 1997]). That is, a fire loss precipitated by terrorist activitywould be covered in the same way as fire loss precipitated by careless smoking,faulty wiring, arson by a third party, and so on.

Fourth, in most lines of insurance, including especially commercial propertyand casualty lines, there was a fiercely competitive soft market for insurancethroughout the 1990s. Insurance companies were chasing premium dollars andtherefore were not keen to make stringent contracts for fear of losing business.Although some insurers and especially reinsurers would have preferred ‘namedperils’ contracts – naming terrorist activity as a peril and either charging for it orexcluding it – the soft market limited the scope for such contracts.

Uncertainty and risk after 9/11

The insurance curse

Industry insiders describe 9/11 as an exemplification of the ‘insurance curse’.Risks that were unimaginable before a catastrophic loss have two impacts afterthe loss. The first is the immediate cost of unexpected indemnity payments.The second is the effort to learn about the previously unknown risks, but, asknowledge accumulates, uncertainty magnifies. These two impacts combine tomake underwriters precautionary. The precautionary approach is compoundedby the organizational culture in which underwriting decisions are made.Underwriters often get into trouble and sometimes lose their jobs for acceptingan insurance risk that proves undesirable and costly

ex post

. In contrast, they areless likely to be in jeopardy for rejecting what proves, in retrospect, to be a prof-itable risk because that decision is much less visible.

Following the catastrophic losses of 9/11, the insurance curse was especiallyacute regarding four types of risk exposure.

Natural catastrophes such as earthquakes and storms can be reasonably‘walled off’ by geographic region and allocation of dedicated capital up tospecific limits. This ‘geographic risk’ spreading is especially difficult withterrorist activity because it can occur anywhere. Moreover, terrorists intention-ally induce the uncertainty of randomness not only in the lives of their actualvictims, but also in those who fear becoming victims. The psychological effectsof terrorism as a business of uncertainty are seen as having additional economiceffects beyond the direct costs of recovery from a terrorist attack. Thus, aninsurance company executive said that if terrorist strategy turned to randombombings in shopping malls in different regions of the United States, ‘it wouldhave a tremendous psychological impact on the U.S. public [and henceeconomic effect]. And so there are just so many variations . . . without predicta-bility, we may become very concerned relative to application [insurability]’.

The 9/11 attack also created new uncertainties for insurers regarding


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

146

Economy and Society

‘aggregation risk’: the over-concentration of insured loss exposure in a particularlocation. Prior to 9/11, high-rise commercial buildings with established tenantswere perceived as ideal risks. On the property side, a building with contem-porary construction standards and a sprinkler system was viewed as an easilyassessable and lucrative insurance risk. For group life, health, and workers’compensation insurers, well-paid professionals working in such buildings werealso among the best risks. As one insurance executive expressed it, 9/11 ‘turnedthe nature of risk on its head’. The prime commercial property in a downtownlocation was now looked upon as a potentially bad risk: what is the total lossestimate? Is it a symbolic ‘trophy’ target for terrorists? What exclusions, lowerinsured limits, and increased prices should be imposed by any single insurer orreinsurer in the name of prudence? Aggregation risk is a particular problem forreinsurers because they are many steps removed from first-hand knowledge ofthe risks they are assuming and how they are concentrated. A reinsurancespecialist for an insurance group addressed the topic in a speech to colleagues inthe spring of 2002.

Think about the difficulty that [each primary insurance company] has indetermining the aggregation of risk in various locations around the country.And then picture yourself as a reinsurer who picks up a part or all of thesecompanies’ exposures as well, and so the aggregation of risk is compoundedand the knowledge level is at least once, if not twice, removed. As a result,until the middle of September [2001], I don’t know that the reinsurancecommunity had ever evaluated the kind of hazard that was demonstrated in amultiple explosion event in the heart of the financial district in New York.That became very clear to the reinsurance industry in the months thatfollowed. . . . The reinsurance community globally is unregulated in mostways and is free to move quickly in and out of markets and in and out ofcoverages. . . . Retrocessionaires [reinsurers of reinsurers] are so far removedfrom [direct knowledge of aggregation] risk that there is no capacity forterrorism reinsurance in the retro market today. As a result, reinsurers . . . ifthey want any kind of exposure to terrorism reinsurance are taking this netfrom their own account and therefore have little capacity or appetite for it.

The 9/11 losses also made insurers more sensitive to ‘correlation risks’ acrosslines of insurance. As indicated previously, there were multibillion-dollar lossesin each of property, business interruption, workers’ compensation, aviation, lia-bility, and life and health insurance. The problem for large insurance groups,and especially reinsurers, was exposure across several of these lines arising fromthe same incident in ways they had not appreciated previously. An executive ofa major reinsurance group stressed in interview that 9/11 ‘highlighted for theindustry the tremendous correlation risk that exists within our business, fromdifferent lines of business, that we never thought would correlate to a singleevent. With that awareness now, we are looking at all our business with that as aprimary focal point.’ A colleague added that a future catastrophic event withsimilar correlation risk ‘could truly, severely impair even our company with all


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


147

of the risk they have accumulated worldwide. So the whole concept of riskmanagement has, I think, been taken to a whole different level. . . . Maybe thisis a wake-up call to the industry.’

New forms of ‘enterprise risk’ were also appreciated in the aftermath of 9/11.The 9/11 insured losses interacted with the broader capital market downturnand low interest rate environment to provide new understanding of how under-writing, investment, and credit risks can correlate in an actual case. A reinsur-ance executive said in interview that this new understanding amplifieduncertainties about the magnitude of future losses.

Suppose there is a dirty bomb in New York City, in the financial area, what doyou think the impact would be on stock market performance? Ultimately, thattrickles through into the balance sheet of insurance and reinsurancecompanies. . . . So there’s a correlation not only in terms of insurance productsthat we provide, but also relative to the financial strength and the foundation forproviding these products . . . enterprise risk, a whole risk evaluation.

The precautionary principle

Insurers struggled to impose meaning on these new uncertainties. Their initialreactions exemplified the precautionary principle: when faced with a threat thatis very severe even if infrequent, adopt an equally severe approach to preventivesecurity (Ewald 2002). Baker (2002) observes that 9/11 crystallized a societaltrend towards precaution regarding various types of risk – for example, regard-ing health (AIDS, BSE, GMF), public safety (especially child safety), and theenvironment (nuclear energy production, global warming) – that have cata-strophic potential (see also Haggerty 2003). There is retreat from the belief thatharm can be subject to precise risk analyses and thereby predicted and control-led in advance. This retreat threatens the insurance system to the extent that itis based on the belief that one can predict both the efficient level of preventionthat will minimize harm and the future cost of harm.

To be responsible within the dictates of the precautionary principle, one mustengage in prevention at all costs and appreciate that the future cost of harm isimmeasurable. The only response is extreme vigilance, a kind of pre-caution.One should even exercise caution about how one is being cautious.

As we document in subsequent sections, insurers exercised extreme precau-tion in both underwriting terrorism risk and ensuring that policyholders whoretained terrorism coverage had intensified preventive security. One of the firstprecautionary steps was to change PML estimates on major commercial proper-ties from partial loss to total loss. The pre-9/11 catastrophic imagination envis-aged a major incident that would cause significant but containable damage. Thepost-9/11 catastrophic imagination refers to the WTC as ‘ground zero’,language drawn from nuclear war, and assumes the worst. A senior industryofficial said, ‘now the PMLs include complete destruction of the building, anybuilding . . . that is being [under]written. The insurer’s PML is that another


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

148

Economy and Society

plane is going to go into it and completely destroy the building. Or a nuclearattack which is completely going to render the building useless for the next fiftyyears.’

The dread of nuclear, biological and/or chemical terrorist attacks moved tothe forefront of arguments about the inability to estimate severity of loss. Thisdread was fuelled by public discourse. For example, an OECD report on‘Economic Consequences of Terrorism’ included a feature box, ‘How to preparefor the risk of mega-terrorism’ (OECD 2002: 120). This feature provides detailon how ‘security specialists’ imagine ‘terrorists could at some stage attempt toexplode a nuclear device or release contagious viruses in a populous metropol-itan area’. One scenario even draws on familiar fear about Russian sources ofexternal threat to envisage all of Manhattan as ground zero. With catastrophicimaginations running rampant, the New York State Superintendent of Insur-ance testified to the US House of Representatives Financial ServicesCommittee, ‘In the insurance world, perceptions of a threat could be just asimportant as a real threat’ (Serio 2002: 8).

Models of uncertainty

While precaution pervaded the insurance industry in the aftermath of 9/11, itdid not lead to unremitting efforts to remove exposure to terrorism risks. Asstated previously, regulatory requirements and market considerations com-pelled terrorism coverage before 9/11, and these forces remained in play after9/11. More fundamentally, insurers can thrive on conditions of extreme uncer-tainty. They market themselves as experts on uncertainty who take risks onbehalf of the insured. The uncertainties of experts about the dynamics of risk,and of the insured in search of protection from the risk, are reconfigured intocapital against whose loss the insurer offers a guarantee. Insurers speculateaggressively within prevailing insurance and capital market conditions, andunderwrite uncertainties that have catastrophic potential for their own viability.The reality is that many perils have been insured when the source of cata-strophic loss was not even known at the time of underwriting (e.g. asbestosis),the science is shaky (e.g. earthquakes), or the insured losses are largely shapedby the inter-institutional culture of claims (e.g. disability and medico-legalclaims inflation). Insurance is a business of uncertainties: the key question iswhich party becomes more averse and decides either to withdraw from the con-tractual relationship or to seek more enabling conditions of participation.

Addressing an industry conference on insurance markets after 9/11, areinsurance executive declared:

We love ambiguity. We know how to handle uncertainty. And I think that weknow it much better than the average consumer or the buyer of our products.Risk has always two dimensions. It is a threat, a peril, but it also contains theaspect of opportunity. And the art is to balance these two, threat and oppor-tunity, in a smart and conscious way.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


149

Many interviewees informed us that they were still participating in terrorisminsurance markets in a smart and conscious way. For example, an insurancecompany executive observed:

The challenge is really not to just walk away from the experience but . . . tounderwrite it. . . . Even though we are not offering a specific terrorism product

per se

, we know that we have embedded in our business, in our portfolio, a lot ofterrorism exposure. And that’s the exposure that we are trying to manage, andcontain, and budget for, such that we can still participate in what is now – allodds equal – a very attractive marketplace with rising prices.

Hoping to capitalize on market opportunities, insurers sought models of ter-rorism exposure that promised to make the incalculable seem more calculable,to turn their uncertainties into risks. Immediately following 9/11, three com-panies with expertise in natural catastrophe modelling – AIR Worldwide Cor-poration, Risk Management Solutions, and Equecat Incorporated –developed terrorism risk models for insurers (Green 2002; Treaster 2002;

Business Wire

2002). A

Business Wire

(2002) announcement of the AIR Ter-rorism Loss Estimation Model (TLEM) quotes the president and CEO ofAIR as saying: ‘“The problem is the current uncertainty about coverage,exclusions and price.”’ The solution is TLEM, which ‘analyzes variousthreats posed by domestic extremists, formal international and state spon-sored terrorist organizations, and loosely affiliated networks. . . . It considersfrequency and severity of attacks, likely targets and impacts on both theintended targets and on surrounding structures – all within a fully probabilis-tic framework’. Perhaps anticipating scepticism that the uncertainties of ter-rorism can be converted into ‘a fully probabilistic framework’ of risk, thenews release invokes the authoritative certainty of Buck Revell, a formerassociate deputy director of the FBI responsible for criminal investigations,counterterrorism, and counterintelligence, who served on an expert panel tohelp develop TLEM: ‘The evaluation process that AIR has undertakenregarding terrorism is extremely important. . . . The first and most importantelement of being prepared is to understand the true nature of the risk and theconsequences of not being prepared for a terrorist attack.’

The terrorism models draw upon conventions developed in modelling naturalhazard catastrophic loss. Indeed, one model even claims to take weather condi-tions into account because winds would affect the dispersal of chemical andbiological agents. A primary focus is the diversification of geographic andconcentration risk. Knowledge for diversification is provided through data onurban aggregates and zonal distributions of insurance written. Risk Manage-ment Solutions catalogued 1000 buildings across the United States as primetargets of terrorism and made them the subject of various terrorist activitysimulations involving different types of weapons (e.g. airplanes, bomb blasts,nuclear, biological, chemical). AIR did the same for 330,000 properties, rangingfrom the Empire State Building in New York to the Capital Building in CarsonCity, Nevada. AIR also analysed properties for concentration risk exposure in


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

150

Economy and Society

property and workers’ compensation insurance lines, and for multiple lines ofinsurance exposure.

Severity estimates are based on various types and capacities of weapons andtheir projected impact. Risk Management Solutions used about two dozendifferent weapons in its scenarios, for example, airplanes of different design andcapacities, trucks loaded with explosives, cake boxes loaded with explosives, andso on. The AIR TLEM modellers ‘figuratively triggered different sizes ofexplosives, sometimes at more than one location on the site. Then they calcu-lated the damage to the landmark and the surroundings. The program can alsodetonate blasts at each property in an insurance company’s portfolio.’ AIR is adivision of Insurance Services Organization, the mega-provider of informationto the industry. As such, AIR has access to that organization’s Specific PropertyInformation database on two million properties in the United States to helpassess its terrorism target buildings regarding size, construction, occupancy, andpreventive security protection.

In spite of claims about ‘a fully probabilistic framework’ and an under-standing of ‘the true nature of the risk’, terrorism loss estimate models areheavily dependent upon the subjective opinions of selected experts. Both AIRand Risk Management Solutions hired former FBI and CIA counterterrorismoperatives to make educated guesses about landmark buildings that might be thetargets of terrorism. AIR also used its experts to estimate frequency. As theTLEM product manager declared, ‘“There’s no physical science behind[frequency estimates]. You can’t extrapolate from past history, so we use expertopinion and codify that”’ (Green 2002). The AIR experts used the DelphiMethod of estimating frequency and severity as developed by the RAND corpo-ration during the cold war (

Business Wire

2002). This approach ‘turns educatedguesses into numerical rankings that are cranked into a computer to determinerisk’ (Treaster 2002).

Of course the developers of TLEM, including the experts themselves,realized the limits to this form of knowledge. According to the president andCEO of AIR, at one juncture ‘her team considered leaving it up to clients tocome up with a frequency estimate. But she said she was convinced that clientswanted her to set down a marker. “They want to have some kind of default” shesaid’ (ibid.). She conceded that TLEM is only ‘a first step in imposing rationalmeasurement on what has been totally unquantifiable, but argued it is “betterthan anything else our clients have to use”’. Her counterpart at Risk Manage-ment Solutions was more reticent about selling uncertainty as if it were risk. Hesaid, ‘it might be six months before his team figures out an estimate of thefrequency of attacks. He may just give up. “This is very speculative,” he said.“It requires a level of insight into what they [terrorists] are thinking that rightnow is not there”’ (ibid.). Meanwhile the president of Equecat simply declared,‘“We don’t know where and we don’t know when, but we know it’s going tohappen”’ (Green 2002).

Appreciating the limits of such models, insurance companies struggled withalternative ways of gaining insight into what terrorists are thinking. One


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


151

reinsurance company we studied formed its own panel of experts consisting ofsenior company executives. Their main source of knowledge was news reportsof ‘the war on terrorism’, which were used to inform their underwriting guide-lines and decisions. One executive observed in interview:

I don’t think we have much of anything right now. . . . We file all the tradejournals, we get a lot of information on what’s found in the Caves of Afghan-istan and so we know all that. And we actually build underwriting guidelinessomewhat around those types of information. . . . We kind of got in a roomand just tried to theorize, tried to get in the head of a terrorist. . . . We’ve beentrying … unsuccessfully so far, to get the FBI in. . . .[We want them] to shareknowledge, to either validate our underwriting approach, or add to it, ormaybe invalidate it, based on what they found through interrogation of peoplethat have been captured, based on some information that comes out publicly.

The pass-the-exposure express

In the aftermath of 9/11, all sectors in the insurance relationship sought newterms of participation. There was a particular focus on new terms for terrorisminsurance, but many other fields of insurance were also subject to reconfigura-tion. Testifying before the US House of Representatives Committee on Finan-cial Services, the New York State Superintendent of Insurance referred to theseefforts at reconfiguration as ‘the pass the exposure express’ (Serio 2001a).

In a precautionary mode, albeit with an eye on new opportunity for profits,reinsurers cancelled contracts, failed to renew contracts, or underwrote newcontracts on much more stringent terms. As we analyse in more detail shortly,reinsurers are relatively free to exclude terrorism coverage from their agree-ments with primary insurers. Primary insurers have some latitude to excludeterrorism coverage also, although their degrees of freedom vary substantiallyacross lines of insurance and the regulatory regime of each state jurisdiction.Without terrorism reinsurance, primary insurers are vulnerable and try to passthe exposure to the insured through some combination of exclusions, morestringent contract terms, and steep price increases. If governments fail to stepin, for example, through a backstop programme in which they serve as rein-surers of last resort, the insured, insurance companies, and political economy ingeneral are left to embrace more risk.

This is a story of how, in conditions of extreme uncertainty, insurers havedifficulty forming a market, and seek the help of governments as the insurers oflast resort. Governments, meanwhile, seek both the capital and preventivesecurity capabilities of the insurance industry to spread at least some of the risk.Insured organizations and individuals are especially vulnerable, as they embracethe risks of being uninsured, underinsured, and/or paying substantial premiumincreases to insurers and governments scrambling for more capital.

As we proceed to document in separate consideration of the position ofreinsurers, insurers, insureds, and governments, the boundaries of private


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

152

Economy and Society

insurance do not ‘dissolve’ in face of extreme catastrophe risk (Beck 1999), butthey do undergo major reconfiguration. Indeed, in his article on terrorism andworld risk society, Beck explicitly pulls back from his stronger statementsabout lack of insurability in the face of new catastrophe risks. He observesthat, after 9/11, ‘the principle of

private

insurance is partly being replaced bythe principle of

state

insurance . . . [resulting in] new questions and potentialconflicts, namely, how to negotiate and distribute the

costs

of terrorist threatsand catastrophes between businesses, insurance companies and states’ (Beck2002: 44).

The new questions include, for example, who is willing to embrace what riskat what price? Who will share risk distribution at what levels of exposure? Willhigher premiums exclude the most vulnerable? At what point does the mutualityrequired for insurance begin to unravel? How much can neo-liberal govern-ments leave to the market before dissimulation occurs?

The pass-the-exposure express involved a flurry of efforts to cancel or fail torenew insurance contracts deemed to be unattractive in retrospect. This is acommon response by insurers following catastrophic loss. For example, after theMount St Helens’ volcanic eruption in 1980, Hurricane Andrew in 1992, andthe Northridge earthquake in 1993, insurance coverage that was previously‘given away’ as part of standard all perils contracts was suddenly pre-emptedthrough exclusions. As an insurance official remarked about reported losses inthe months following 9/11:

It was like playing, ‘Can you top this?’ because every day you read anotherfigure. It started out at $25b, 30, 40, 50, 70, 80 – nobody really knew how bigthe loss was going to be. And it was in that climate the insurance industrysaid, wait a minute, no, we cannot take a chance on that happening again. . . .I mean the wholesale, we are not going to cover anybody for terrorismanywhere under any circumstances, was kind of a knee-jerk reaction, it was anemotional reaction.

Of course, terrorism coverage was never entirely removed from the insurancemarket. First, after 9/11 some multi-year contracts with terrorism includedremained in force. Second, some terrorism coverage remained compulsory.Workers’ compensation schemes in every state require coverage for injury atwork however caused, including acts of terrorism. The US National Associa-tion of Insurance Commissioners voted to ensure that personal lines propertyand casualty insurance continued to include terrorism coverage in all states.Regulators in five states – New York, Florida, California, Texas, and Georgia –ruled that terrorism coverage must be made available in commercial lines prop-erty and casualty insurance. These five states constituted about 38 per cent ofthe US commercial market at the time. Third, some insurers decided to takeadvantage of the opportunity for selling terrorism insurance on the view thatterrorist activity is a low probability, high profit risk.

Where legal requirements and/or the thirst for profits compelled terrorismcoverage, insurers proceeded to reconfigure the terms of underwriting. First,


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


153

there was selective exclusion of terrorism coverage in specific contexts. As aninsurance company executive expressed it:

Exclusions are being applied to make sure that we are not just picking up anytype of terrorism. . . . It’s the uncertainty factor. . . . We don’t excludeterrorism on everything. . . . We do exclude terrorism on the liability side ona target risk basis. That means a directors’ and officers’ liability on an airportor on a large office building that contains, for example, a stock exchange; wewould exclude terrorism there.

Second, the key to a given insurer’s exclusion and inclusion strategy was nego-tiating the meaning of terrorism on a contract-by-contract basis. As a reinsur-ance executive stated in interview, ‘We need to define it in order to exclude it’.There was a marketplace of ‘contracted’ definitions open to negotiationdepending on how the reinsurer or primary insurer wished to participate in aspecific terrorism risk. A reinsurance executive involved in contract negotia-tions with primary insurers said that his company worked from a very broaddefinition of terrorism in order to negotiate terrorist activity exclusions in rein-surance contracts with ceding companies (primary insurers) where desirable.‘We might throw out these words, someone else might throw out another set ofwords to define terrorism, and we work through it. . . . Everybody usually has adifferent point of view on things, which are guided by maybe corporate riskappetites and such.’ A colleague added, ‘There are as many terrorism exclu-sions in the marketplace as there are ceding companies. So I think at last countwe were up above 250 [companies]. So there’s probably 250 some-odd word-ings for terrorism exclusion in the current marketplace.’ He stated that the finaldefinition of terrorism settled on for a given insurance contract depends onother terms of the contract.

There’s a lot of negotiation that goes back and forth with the companiesdepending upon our participation in the programme, limits offered, attach-ment point, whatever, where we may come off some of those [definitions thatexclude terrorism] objectives . . . sub-limits are one of the ways to obviouslyreduce your exposures, to reduce your limits, then maybe we can give a littlebit on the wording.

This involvement of insurers in defining terrorism belies the view that it is onlypowerful governments and states that define the new transnational terrorist(Beck 2002: 44), and the related views that ‘the power of definition of expertshas been replaced by that of states and intelligence agencies; and the pluraliza-tion of expert rationalities has turned into the simplification of enemy images’(ibid.: 45).

Third, the contracted conditions of terrorism coverage included sharppremium increases in an already hardening insurance market. With risk andprice left to the imagination, and many insurers not willing to take risks at anyprice, insurers still in the terrorism coverage business could seek whatever pricethe market would bear.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

154

Economy and Society

Fourth, the hardening market meant that insurers could also harden theircontractual requirements of preventive security. Insureds in need of terrorismcoverage could be compelled to ‘target harden’ their persons and property byhiring more private security personnel, buying more electronic surveillancetechnologies, and opening up to audits of their employees, operations, securityinfrastructure, and so on.

We now consider how the pass-the-exposure express shaped, and was shapedby, reinsurers, primary insurers, insureds, and governments.

Reinsurers

Reinsurance is an essential component of a catastrophe insurance programme.Its most important feature is a global capacity to spread risk efficiently overtime and place. In this capacity, it provides protection to primary insurancecompanies with a local domicile, which in turn protect the insured and relievegovernment of full responsibility for disaster relief.

From the perspective of reinsurers, the initial pull-back from participation incontracts with terrorism exposure was a responsible act of precaution whileinsurance markets were reconfigured. Compared to the primary insuranceindustry, the reinsurance industry had far less surplus capital to cover anothercatastrophic loss. Its losses from 9/11 were estimated by some interviewees at 60per cent or more of the $55b total. Too much exposure to the uncertainties ofterrorism could threaten the solvency of some reinsurance companies, which inturn would affect the recoveries of primary insurance companies and perhapstheir solvency as well. Insolvent primary insurers would mean that the insuredmight not recover through an insurance claim.

Reinsurers, along with primary insurers, were already facing peculiar marketdynamics prior to the record catastrophic loss of 9/11. Throughout the 1990s,the industry was operating in a soft (highly competitive) market in which lowpricing and lax underwriting were used to attract business. It was the stronginvestment markets of the 1990s that buoyed their overall profitability, andindeed encouraged further competitive pricing and lack of vigilance on theunderwriting side. With the sharp decline in interest rates and investmentmarkets, there was a problem of capital adequacy affecting many parts of theindustry.

The soft market condition was one reason why broad terrorism coverage wasincluded in reinsurance contracts rather than excluded or treated as a specialcoverage with higher premiums. The market was so soft that low-priced, multi-year reinsurance contracts were written, although the preferred practice isannual contracts. This meant that some contracts still included broad terrorismcoverage well beyond the normal renewal time of January 2002. This continuingexposure provided another reason for reinsurers to be precautionary aboutincluding terrorism coverage where contractual agreements allowed themchoice.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


155

An additional charge on capital adequacy came from a series of long-tailedcatastrophic losses before and after 9/11. Medical cost inflation, medicalmalpractice claims, the continuing costs of asbestosis liability, Enron andWorldcom corporate crime liabilities, and toxic mould liabilities each consti-tuted multibillion-dollar insurance catastrophes. Asbestosis still stands as thelargest insured loss ever, with approximately $65b in claims paid in the UnitedStates alone. Senior executives of reinsurance companies also singled outmedical claims inflation as a particular problem. For example, one executiveobserved that in the 1990s soft market there was ‘systematic underpricing ofproduct, and maybe a lack of realization of how substantial medical inflation hadbecome. And medical inflation disproportionately affects reinsurance entities.This was way before the World Trade Center, way before corporate scandals,and cost the industry billions, and maybe billions to come’. Many of the medicalclaims problems were concentrated in the area of personal injury claims to autoinsurers and workers’ compensation insurers (for case examples, see Ericson

etal

. 2003; Ericson and Doyle 2004: ch. 3).Continuing problems at Lloyds were another source of uncertainty for the

reinsurance industry. The 1990s soft market and severe losses in 9/11compounded an already weak environment in London. Lloyd’s syndicates(private investors with unlimited liability) experienced enormous losses in theearly 1990s, primarily arising from their exposure to the Piper Alpha catas-trophe (Luessenhop and Mayer 1995). They did not recover, and have becomedisplaced by the admission of limited corporate capital into the London market.

In recent years, the Lloyd’s market-place has been highly unprofitable:between 1998 and 2001 the accumulated loss totalled more than GBP 6bn!This is equivalent to about 60 per cent of the average underwriting capacity.At the solvency margin of around 50 per cent, this means that over 100 percent of the equity committed has been lost within the past four years!’

(Swiss Re

Sigma

No. 3 2002: 4)

The weakness of the London market was significant because it constitutedabout 15 per cent of reinsurance capacity worldwide and was

the

internationaltrading centre for ‘reinsurance covers that are large, unusual and complex. . . .[I]f difficult risks are going to find capacity at all, this is where they will find it’(ibid.: 4). The limited corporate capital that has come into Lloyd’s is organized,in effect, through insurance companies that have formed within Lloyd’s struc-tures, half of which are actually from the United States and Bermuda. Bothcapital and insurance operations are increasingly offshore, making the Londonmarket resemble an e-business virtual out-station (ibid.: 25).

The global and virtual features of insurance capacity were also evident in the$30b of new capital that flowed into the business in the months following 9/11(OECD 2002: 3; Swiss Re

Sigma

No. 3 2002: 15). Lloyd’s had at most a 7 percent share of this equity, again limiting its ability to build capacity in a hardeningmarket. Almost half of the new capacity was created in Bermuda, mostly in newcompanies. The new capital was opportunistic ‘hot’ money rather than


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

156

Economy and Society

‘committed’ money, venture capital in search of short-term gains from hard-ening insurance market conditions. The new companies that formed into thesemarket conditions did not carry the liabilities of 9/11, nor did they bear any ofthe previous investment decline and catastrophic liabilities we have outlined.They could simply gamble that there would not be another major catastrophicloss from terrorism in the short term, and make a lot of money if their guessproved correct.

If the new companies used their competitive advantage of no past liabilities toundercut reinsurance prices too much, then they could negatively affect existingreinsurers who are willing to manage catastrophic losses in the long term.Existing reinsurers do have liabilities from the past and need the hard market tocontinue for awhile to bolster their capacity. A senior executive of a reinsurancecompany observed, in interview:

New capital has gushed into the business. . . . And that in itself dampens theeffects of the replenishment of [reinsurers] . . . that have been in businesshistorically, that were in business on September 11th, and suffered thoselosses, and suffered underpricing in the soft market prior to that, directors’and officers’ scandals, and corporate scandals and so on.

Given all of the above considerations about their market conditions and posi-tioning, reinsurers argued that, while they had a social responsibility to providesome terrorism coverage, they also had a social responsibility to exclude andlimit coverage where it seemed to threaten their viability as a profitable busi-ness. A reinsurance executive said in interview, ‘There is an economical limit toeverything. . . . At one point, social responsibility [to provide reasonable rein-surance] can become social irresponsibility because, if we go bankrupt as a com-pany, it would affect a larger scope of people.’

At the same time, although they have the greatest regulatory freedom towithdraw from participation in insurance markets, reinsurers are obviouslybound by relationships with major partners in primary insurance companies andgovernments. For example, with 35 per cent of all insurance capacity in theworld based in the United States, major reinsurers are unlikely to withdrawunilaterally from an area of coverage important to the US government andprimary insurers. Partnerships with the US government and primary insurerswere especially compelling in the wake of 9/11. Reinsurers had much to gaininstrumentally from business relations that would help form new markets forterrorism coverage as a low probability, high profit enterprise. They also recog-nized the symbolic value – moral and political – of being seen as integralparticipants in the ‘war on terrorism’.

Instrumentally, hard markets gave reinsurers an opportunity to do hardbargaining with ceding primary insurance companies. They could redefine themeaning of risks covered, such as terrorist activity. They could command muchhigher prices for reinsurance, a cost that primary insurers would in turn pass onto the insured. They could require more stringent loss prevention enforcementby primary insurers. A reinsurance company executive enthused in interview,


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


157

‘The questions that we’re now asking are as important in a soft market as theyare in a hard market. Only now it is a hard market, we’re in the driver’s seat. Sowe’re taking advantage of the situation. You can perform better underwriting ina hard market than you can in a soft one.’

There was also hard bargaining with governments. In particular, the reinsur-ance industry wanted the US federal government to serve as the terrorismreinsurer of last resort. Such government reinsurance backstop programmeswere negotiated in other countries, for example, France (Standard and Poor’s2002a: 3–4). According to senior executives in the reinsurance industry weinterviewed, the initial pull-back of reinsurers from terrorism coverage was inpart an exercise in prudent responsibility about the capacity of their industry,and in part a political strategy to force government participation as reinsurers oflast resort. As one interviewee explained:

Some of the restriction on the reinsurance side is actually meant to initiate thecreation of governmental facilities. . . . [That intention is] specificallyexpressed at least in some of the external memos. So both bodies, the govern-mental bodies and the reinsurers and insurers . . . their representatives meetand there is no ivory tower with the reinsurers.

They

decide what to do, and

that

is the reality. Every step is very controlled I would say. . . . Part of thatpullout [of terrorism coverage after 9/11] was to

force

some action and someresponse on the part of government.

Primary insurers

The pre-emptive actions of reinsurers to exclude, limit, and redefine terrorismcoverage posed serious consequences for primary insurers. For example, thegroup life insurance industry was not generally in favour of excluding terror-ism, but could reasonably maintain it only with the participation of reinsurers.Without reinsurance, primary insurers would have to embrace the risk of ter-rorism and face possible insolvency in the event of another catastrophe. Thisthreat could be relieved only if government regulators allowed primary insurersto exclude terrorism, but in several state jurisdictions, including New York,regulators did not allow exclusion. If exclusion was permitted, the burden ofterrorism would be shifted to the level of the individual, and both the insuranceindustry and the regulators who are supposed to protect the consumer would ineffect be admitting the failure of their respective institutions.

A life insurance executive emphasized that:

reinsurance . . . is a critical risk-spreading mechanism for the group business.We work in the law of large numbers, we have very minimal underwriting,and disasters would tend to have a huge impact on us because of the clusteringof employees. So reinsurance enables a life carrier to spread that risk tomaintain some stability and favourable pricing.

He said that, as of March 2002, approximately one-half of his company’s


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

158

Economy and Society

reinsurance partners were not offering terrorism coverage upon renewal of con-tracts. Moreover, those who were offering coverage introduced new exclusions,for example, nuclear, biological, or chemical acts of terrorism. He also said that,upon renewal of contracts, reinsurance rates were rising exponentially. Withterrorism excluded, the increases ranged from 700 per cent to 1300 per cent.With terrorism included, the increase was another 250 per cent.

Property insurers also felt the impact of precautionary measures byreinsurers. Most lost their terrorism reinsurance coverage into 2002, and therewere dramatic price increases for the little coverage still available (Standard andPoor’s 2002a: 1).

As mentioned previously, regulators in five states, including New York, didnot allow primary insurers blanket exclusion of terrorism coverage on commer-cial properties. At the same time, these regulators appreciated that, withoutreinsurance, primary insurers were left to embrace the catastrophic conse-quences of another terrorist attack. Unable to force reinsurers to participate, theregulators’ choice was to make either primary insurers or consumers bear theburden of terrorism. Forty-five states chose consumers, five states choseprimary insurers. A regulator in one of the later states said in interview that thedecision was taken not to let the consumer

be the last stop in the pass the exposure express. So the consumer now has thecoverage, the insurance companies are providing that coverage, but now thereinsurance companies are not providing them with reinsurance for theirexposure. And that is why we support some sort of federal [government]intervention to eliminate the gap that currently exists because of the actionsof reinsurers. . . . Even forcing insurers to cover this risk does not make senseto some extent because they do not have the ability to rate, and we know that. . . [without federal government intervention] the alternative is to completelyeliminate coverage: let the reinsurers eliminate coverage because they can’trate it, let the primary insurers eliminate coverage because they can’t rate it,eventually we will all bear this risk ourselves. And that in our opinion is not agood outcome.

This decision by regulators intersected with market conditions. Primary insurersalso suffered from the downturn in capital markets at the time, which resulted insignificantly lower returns on the investment side of operations. This situationled debt-rating agencies to pay more attention to the underwriting side of opera-tions. If the insurance company’s underwriting was seen as problematic, forexample, because it had terrorism exposure without reinsurance, it faced the pos-sibility of a rating agency downgrade, with implications for its equity price andfinancing capacities. In the autumn of 2001, Standard and Poor’s (2002a: 2–3) puttwenty-two insurers and reinsurers on ‘credit watch’, and lowered the ratings offourteen companies, commenting that, ‘The continuing uncertainty accounts forsome of this’. The managing director of insurance company ratings for Standardand Poor’s declared, ‘If regulators force the industry to accept risk without a [fed-eral government backstop], then there will be ratings consequences.’


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


159

Like reinsurers, primary insurers were also operating in soft market condi-tions throughout the 1990s. Indeed, the US property insurance industry experi-enced losses on the underwriting side (premium revenue minus claims andadministrative expenses) in every year since 1979, culminating in the biggestloss ever in 2001, estimated at $50b (Insurance Information Institute website,<www.iii.org>, 2002). Nevertheless, in the 1990s the industry was able tounderprice insurance, buy market share, and sustain profitability because theywere deriving excellent investment returns.

In jurisdictions we studied, property and casualty insurance companies arelegally required to take into account investment income in determining insur-ance rates. According to a regulator we interviewed, in the late 1990s theinvestment return assumptions were as high as 15 per cent, whereas by 2002,they were at 5 per cent. The downturn in investment returns, especially inequity markets, contributed to a hardening of the insurance market andpremium increases.

A capital markets specialist informed us that, while insurance regulators maybe concerned about heavy investment-side subsidization of premiums during asoft market, they more or less go along with it because it is unpopular to suggestthat consumers should be paying more. ‘It’s hard to deal with a really long-termproblem in the short term if things are going well.’ He added that, even at theheight of the equity market bubble in 2000, regulators’ ‘stress testing’ of acompany’s loss ratios would use a worst case scenario of a 20 per cent correctionof the NASDAQ index, not an 80 per cent downturn as was experienced by2002. During a soft market, everyone is compelled to go with competitiveunderpricing, reduced vigilance in loss prevention, and aggressive investment.

The problem is that if the companies and the people would assume scenarioswhich with hindsight would happen, then nobody would be in business.Because, if you would have known with certainty, or forced companies toestimate that the stock markets would do what they did, that would havedriven an awful lot of companies out of business! The permanentoptimism. . . . Especially if your peers are doing much better, you have to, toa certain extent, go with the flow because the investors would abandon you ifyou don’t perform. As well as your competitors, even if they perform for thewrong reasons.

In company with reinsurers, primary insurers also faced long-tailed liabilitiessuch as asbestosis, medical inflation, and toxic mould. These liabilities, alongwith lower investment returns and some retraction in reinsurance marketsbefore 9/11, meant that the primary insurance market was already hardening atthe beginning of the decade. Significant capital taken out of the industry and anew sense of precaution both led to demand exceeding supply, driving up pre-miums. In the words of the Superintendent of the New York State InsuranceDepartment, ‘Unfortunately, September 11th has changed what might havebeen a gradual hard-market “curve” into a disaster-driven “hair-pin” bend.The combination of lack of terrorism coverage and rising insurance rates is


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

160

Economy and Society

likely to create availability and affordability problems for consumers in theshort run’ (Serio 2001b: 9). One analysis concluded that 9/11 accounted forabout one-half of the aggregate equity decline in the global insurance industry.

The terrorist attack of September 11 triggered the largest loss ever for theinsurance industry, with the cost to direct insurance companies andreinsurers being estimated at between USD 35 and 55bn. Together withother large insurance claims and losses on investments in the order of aboutUSD 50bn due to the downturn on the capital markets, this caused theaggregate equity of the global insurance industry to decline by about USD100bn. The relative shortage of insurance capacity, especially for large risksand in reinsurance, in conjunction with the need to re-assess many risks, hasled to the current hardening of the market. But September 11 has acceleratedand drastically intensified a trend that had already been underway since theyear 2000.

(Swiss Re

Sigma

No. 3 2002: 11)

The reference to assessment of risks is also significant in accounting for thehardening market. As addressed previously, 9/11 led insurers to re-examinePML estimates on commercial properties, and to project total loss rather than aminor partial loss. There was also a reassessment of the amount of risk capitalneeded to support the insurance industry, and of the liquidity needs in theevent of another major catastrophe. Considerations such as these led to newunderwriting standards and the establishment of lower limits on a given risk,forcing many insureds to seek policies from multiple carriers and at muchhigher prices.

Even in the fourth quarter of 2001, there were very steep increases in variouscommercial insurance lines (Table 1). The OECD (2002: 125) said that the rateincreases had the largest influence on the aviation industry, but there was also asignificant impact on other transportation, construction, tourism, and energygeneration industries. It estimated a 30 per cent rate increase across commercialand liability insurance as a whole, with exceptionally high increases for terrorismtarget structures such as iconic buildings and chemical and power plants.

While there were myriad contributors to the hardening market, opportunismafter 9/11 must be considered. By law, insurers are not permitted to recoup pastlosses directly through higher premiums. The increased premium rate issupposed to be prospective. To rate retrospectively, for example with respect to9/11 losses, would obviously skew premiums significantly. Nevertheless,various interviewees observed that 9/11 was used by company agents andbrokers to justify higher rates. For example, a senior executive of an insurancecompany took the view that, while he personally ‘would not peg’ the sharppremium increases on 9/11, ‘sometimes I view it as a broker’s easy way of sellinga rate increase to a client . . . pegging it on 9/11 as opposed to understanding theeconomics behind it, to explain it to a client better’. He proceeded to explain the‘economics behind it’ with reference to small commercial firms that were expe-riencing 25 to 30 per cent increases upon renewal of their insurance contracts,


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


161

which is well in line with what was received on the hard markets of the past aswell. The ones that enjoyed the benefit of, let’s say, naïve underwriting overa number of years are probably seeing multiples [i.e. increases of severalhundred per cent]. Losses that were not taken into consideration in pricing inthe past because of the soft market are being corrected for that in thefuture…. I think it’s more of a correction which 9/11 just accelerated a lotquicker than we anticipated because it took a lot of capital out of the market.

Insureds

The efforts of reinsurers and primary insurers to reconfigure risk following 9/11had the greatest impact on insured commercial organizations, and ouranalysis therefore focuses on them. A Prudential Securities survey in the firstquarter of 2002 reported that one-half of US commercial property insurancebuyers had no terrorism coverage, 26 per cent had limited coverage, 6 per centhad stand-alone coverage, and only 18 per cent had full coverage (Hartwig2002: 4). In effect, trillions of dollars of exposure were pre-emptively trans-ferred to business enterprise. Moreover, as documented in the last section,whether or not terrorism coverage was available, there were enormous priceincreases when contracts became due for renewal, often with much lowerinsured limits on a given contract.

In the months following 9/11, the New York State Insurance Departmentheld public hearings across the state to learn more about the experiences of theinsured. Reports of commercial insurance premium increases between 100 and300 per cent were commonplace. For example, one group of twenty-threehospitals in New York City found that their insurance premiums almost tripled,their coverage limits were cut in half, and they had to obtain this new coveragefrom more than twenty different insurance companies rather than the onecompany they had used previously. The New York Metropolitan Transporta-tion Authority’s (NYMTA) policy up to 9/11 provided $1.5b coverage,including terrorism, for an annual premium of $6m. Upon renewal, coverage

Table 1

Rate survey, fourth quarter 2001

Rate increases by line of business

No change Up 1–10% Up 10–30% Up 30–50% Up >50%

Commerical auto 1% 5% 51% 36% 5%Workers comp. 1% 20% 46% 24% 3%General liability 0% 7% 64% 22% 5%Commerical umbrella 1% 2% 21% 34% 41%Commerical property 0% 2% 28% 38% 31%Business interruption 1% 8% 53% 25% 10%

Source

: Council of Insurance Agents and Brokers


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

162

Economy and Society

was cut in half, terrorism was excluded from the basic policy, and the price wasincreased 300 per cent. Separate terrorism insurance was underwritten for just$70m coverage, with a $30m deductible and $7.5m premium. In other words,this modest terrorism coverage alone was costing the NYMTA 25 per cent morethan its entire coverage, including terrorism, before 9/11 (Standard and Poor’s2002c: 3).

Jewish organizations in New York – synagogues, schools, and charitablefoundations – complained that they were being singled out for termination ofinsurance or steep premium increases. Termination or higher pricing on theirproperties may have been based on location (e.g. Lower Manhattan) or the factthat these properties were deemed to be especially vulnerable to terroristactivity. Termination of an insurance agreement on the basis of the religiousaffiliation of the insured was prohibited by state regulators. However, the abilityto detect actual discrimination by insurers was difficult because of the range offactors that are taken into account in underwriting a particular property and thehard market conditions prevailing at the time.

While the problems were especially acute in New York City, they extendedacross the state. For example, a steel merchant and fabricator in up-state NewYork complained that his premiums increased from $45K in 2001 to $165K in2002: general liability premiums went from $8K to $75K, automobile coveragefrom $31k to $56k, and commercial umbrella policy premiums from $6K to$34K. Moreover, early in 2001 the company selected their insurance providerafter receiving quotations from more than a dozen insurers, while in 2002 onlytwo insurers agreed to provide quotations. The company estimated that it wouldhave to raise prices to its steel fabrication customers by 10 per cent to cover theseincreased insurance premiums (Serio 2002: 17–18).

In a study of the economic consequences of 9/11 on New York City, CityComptroller Thompson (2002: 6) argues that it is wrong to think of insuranceclaims payments as offsetting the economic impact. He observes that ‘a majordisaster affects future premiums so that it [insurance] is best seen, in economicterms, as a loan to the affected economy, to be paid out of an ongoing differentialin premiums that reflects the loss that insurers and reinsurers have sustained’(ibid.). Furthermore, the increased spending on both public and private securityfollowing 9/11 contributed to ongoing insurance operations to the extent that ithelped reduce future claims payments.

The insurance availability and affordability problems had the potential toaffect investment, for example, in new construction projects, and thereforegrowth (OECD 2002: 125). In the year following 9/11, organizations coped withinsurance availability and affordability problems in different ways. Some simplywent without insurance coverage they held previously, although they did notnecessarily reveal their exposure to parties that would be concerned about it,such as creditors. In some contexts, banks amended loan agreements to allowcommercial clients to operate without terrorism insurance coverage. There wasalso some flexibility in the energy, utility, and power sectors, as lenders madedecisions about requiring terrorism insurance coverage on a case-by-case basis.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


163

In some instances, lenders allowed such facilities to go without terrorismcoverage on a temporary basis (Standard and Poor’s 2002a: 6).

Another adaptation was for companies in particular sectors to form their own‘captives’ or self-insurance pools, usually in collaboration with insurers andgovernments. The practice was already growing, but it took on new dimensionsfollowing 9/11. For example, firms in the energy, utility, and power sectors findmarket-based insurance very expensive and it is often advantageous to form self-insurance pools. Power companies for example can pay between 1 and 10 percent of the insurance limits they cover. This means that $50m of coverage couldcost up to $5m annual premium (ibid.: 5). Following 9/11, the airlines faced aninsurance capacity crisis and tried to form their own industry insurancecompany with the help of a US federal government backstop.

The precautionary approach of insurers after 9/11 also shifted responsibilityfor more preventive security onto commercial entities. The Superintendent ofthe New York State Insurance Department said, ‘We know of individuals andbusinesses vacating the upper floors of high rise buildings not only because theyare unable to obtain appropriate insurance coverage but out of additionalconcern for personal safety’ (Serio 2002: 7). A survey of chief executive officersof major commercial entities, conducted in November and December 2001,asked them to identify security issues that were of greater concern after 9/11(Insurance Information Institute, <www.iii.org>, 25 April 2002). Among thematters that were of greater concern were mail processing (86 per cent), travel(85 per cent), employee protection (79 per cent), infrastructure protection (75per cent), risk assessment capabilities (71 per cent), office/plant protection (69per cent), employee morale (69 per cent), supply-chain distribution (51 percent), customer security (50 per cent), and productivity (47 per cent). The samesurvey asked respondents to identify ‘post 9/11 precautions’ actually taken, andthe following were mentioned most frequently: review disaster plan (90 percent), background checks on contractors (51 per cent), background checks onemployees (39 per cent), limiting number of staff on a single flight (36 per cent),and considering alternative office space (35 per cent).

With visions of the Titanic as

the

infallible technology that failed, it wasstressed that the WTC had been touted as ‘collapse proof’. Those contemplatinghigh rise construction were implored to take note of total loss risks, while thoseresponsible for existing buildings were asked to address remedial measures thatmight reduce such risks. On a mundane level, advice flowed on the need fortarget hardening. ‘For instance, installing a film across windows can reduce therisk of injuries from flying glass. By putting a barrier around a building, acompany can keep potentially dangerous cars and trucks away’ (Green 2002).Caught in the marketing of (in)security following 9/11, organizations boughtinto an array of surveillance technologies aimed at monitoring people and lethalweapons. According to the publisher of the new

Homeland Defense Journal

, ‘TheSept. 11 attacks will be the biggest catalyst for U.S. technological innovationsince the Soviets launched their Sputnik Satellite in 1957, spurring thecompeting U.S. space program’ (Krane 2002: D16).


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

164

Economy and Society

Private security was a fast growing and substantial part of the U.S economyprior to 9/11. One estimate of private security spending in the United Statesduring the 1990s is $40b annually, with almost half devoted to private policestaff and the remainder to preventive security technologies (Anderson 1999).The OECD projected that private security would expand considerablyfollowing 9/11, with possible effects on economic growth and productivity: ‘Adoubling of private security might reduce the level of potential output by 0.6 percent after five years and the level of private sector productivity by 0.8 per cent’(2002: 136). The New York City Comptroller reported that, in 2001, over 1 percent of all workers in New York City were security guards. In the four yearsfollowing the 1993 WTC bombing, there was a 22 per cent increase in spendingon private security guards. Assuming the same level of growth over four yearsafter 9/11, the additional cost would be about $1b. ‘Security-related spending,however, is a cost that may be seen as an investment because it helps to narrowthe future property/casualty [insurance] premium between NYC and othercities’ (Thompson 2002: 22).

The insured were highly vulnerable to insurance companies riding the pass-the-exposure express. Insurers are eager to gamble on uncertainties as long asthey keep reaping profitable returns. However, an extraordinary catastrophicloss leads them to hop on the pass-the-exposure express and seek refuge untilthey can restructure the market and gain confidence that they will again be onthe winning side. The refuge is provided by governments, who help underwritefaltering insurance markets. This socialism for business enterprise can super-sede any socializing effects on behalf of the insured.

Governments

Government by the state can be understood with respect to risk managementfunctions and activities. As Moss (2002) has shown, government is ‘the ultimaterisk manager’ because it has unique capacities to intervene in imperfect marketsin order to shift, spread, and reduce the direct impact of risks. Even apparentlyanti-statist governments such as the United States must invest heavily in eachof these areas of risk management because markets in risk are imperfect. ‘Per-haps one of the strongest criticisms of a system of freely competitive markets isthat the inherent difficulty in establishing certain markets for insurance bringsabout a sub-optimal allocation of resources’ (Arrow and Lind 1970: 374, citedby Moss 2002: 3). This difficulty must be addressed by government, whoseresponsibility is to ‘undertake insurance in these cases where [a private marketfor insurance], for whatever reason, has failed to emerge’ (Arrow 1963: 961,cited by Moss 2002: 3).

Government has a number of advantages in risk management compared to theprivate sector. First, it has legal power, for example, to compel private insurancecoverage, shift the risk of corporate limited liability from shareholders onto credi-tors, or shift the risk of product liability from consumers onto manufacturers.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


165

Second, it has the power to compel revenue through taxation and printing money,which in turn give it the deepest pockets and greatest creditworthiness. Third, ithas power over temporal risk: both its legal power to compel insurance and itsfiscal power to compel revenue give it the capacity to force future generations toparticipate in risk spreading over time. Fourth, it has the power of information,both through its own databases and by accessing those in private-sector institu-tions. Fifth, it has the power of preventive security, especially in the form of policeand military.

In what follows we examine some of the most salient dimensions of how theUS federal government served as the ultimate risk manager after 9/11. We focusin particular on efforts to insure and otherwise bail out the airline industry, todevelop a backstop programme for the insurance industry, and provide preven-tive security.

The US federal government took extraordinary steps to protect the airlineindustry, which was already in difficulty before 9/11. In addition to the fourairplanes destroyed and lives lost in the 9/11 attack, the industry suffered animmediate decrease in passenger traffic.

The Air Transportation Safety andSystem Stabilization Act

(ATSSSA) provided a number of temporary measuresto bail out and stabilize the industry: $5b direct compensation; $10b loan guar-antees for both 9/11 costs and any subsequent terrorist attack; a $100M liabilitycap on future terrorist attack claims; discretionary reimbursement of airlineinsurance premium increases; discretionary reinsurance protection; and aninterim insurance programme (Campbell 2002; Standard and Poor’s 2002a: 4).

The most extraordinary provision of the ATSSSA was the Victims Compen-sation Fund (VCF). This provision was designed to compensate victims of 9/11in a direct and timely manner, as an alternative to the tort liability system.Anyone who suffered personal injury in the terrorist attacks of 9/11, or benefi-ciaries of those who experienced wrongful death, could apply to the VCF. Sincethe claim was payable by the US Treasury, the VCF was not really a ‘fund’ orinsurance but simply a contribution to disaster relief. The VCF payment wasreduced by any collateral sources available to the claimant, such as life andhealth insurance, pension benefits, and specified charitable donations the victimreceived from the Liberty Fund or similar sources.

If the victim or beneficiary opted for the VCF, they were required to relin-quish any right to sue through the tort liability system. The events of 9/11precipitated

the

mass tort of all time, and the VCF provided an alternative thatwould short-circuit the extraordinary legal, time, and psychic costs of litigation.In the legal opinion of some, the fact that the VCF option pre-empted the rightto sue made it unconstitutional. As such, the VCF arrangement was consistentwith the many other legal enactments following 9/11 that short-circuited legalprocess or avoided it altogether.

The VCF was the ‘purest’ no-fault statute ever adopted in the United States.The statute was drafted by leading representatives of the plaintiffs’ bar, a groupthat would benefit enormously from litigation. The plaintiffs’ bar co-operated inthe VCF process because it felt it had no choice but to do so. Indeed, it not only


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

166

Economy and Society

helped to develop the VCF alternative to the tort system, but also offered freelegal services to victims and their families as a gesture of solidarity.

On one view the VCF was ‘an unprecedented government social welfare reliefprogram’ (Campbell 2002: 211). It was the first direct obligation federal billsince medicare thirty-six years earlier, social security being the only other directobligation provision in US history. The official in charge of the programmecalled it an ‘“Unprecedented expression of compassion on the part of theAmerican people to the victims and their families . . . designed to bring somemeasure of financial relief to those most devastated by the events of September11 . . . [and] an example of how Americans rally around the less fortunate”’(ibid.: 15–16).

The VCF was also an example of how Americans rally around their lessfortunate industries. The VCF was a crucial ingredient in efforts to bail out theairline industry. The ATSSSA capped airline liability for 9/11 to their insur-ance coverage at the time. In effect, the VCF provided a federal immunity tocommercial entities (airlines) for state-based tort claims, dispossessing the stateof legal jurisdiction. Speaking at an insurance industry conference, a legal expertwho had been close to the ATSSSA legislative process said:

When this bill originally came up in Congress it started out as a pure bailoutfor the airline industry. . . . [T]heir insurance was about to be cancelled acrossthe board by all insurance carriers for all airline companies. They could notrenew their lines of credit. . . . [The argument was made] you cannot bail outthe industry and leave the victims behind. That caught on like wildfire. . . .What the government really has done here is spread the risk over all of thetaxpayers of the United States.

In the year following 9/11, there were protracted debates and different modelsfor involving the federal government in a terrorism insurance backstop pro-gramme. The government finally passed the

Terrorism Risk Insurance Act

on 26November 2002 (Bumiller 2002). This bill covers terrorist activity by foreigngroups or governments, not Americans. It includes a provision that allows vic-tims of terrorist activity to sue corporations for punitive damages, thereby pro-viding an incentive for corporations to intensify preventive security of theirproperties and operations. It requires all insurers of commercial properties tooffer terrorism coverage. The most significant backstop layer begins at aninsured loss of $10b and is capped at $100b for a given terrorist event. In thisrange the government programme pays 90 per cent of insured loss. Below $10bthe government’s programme pays a lesser proportion (for models in othercountries, see OECD 2002: 127).

The arguments in favour of government participation in a backstop insuranceprogramme encapsulate our analysis of the pass-the-exposure express. First,from a primary insurance industry perspective, terrorism exposure cannot beadequately measured, monitored, or predicted, and therefore it was difficult toallocate specific insurance capacity to terrorism coverage or to obtain reinsur-ance. Primary insurers were left vulnerable, compounded by the fact that


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


167

regulators in some states, including New York, continued to require them tomaintain terrorism coverage in various lines of insurance. The uncertainties ofterrorism seemed infinite, while industry capacities were finite.

Second, it was clearly unfair for the insured to be forced to embrace the riskof terrorism. Placing the onus on innocent victims to deal with the consequencesof terrorism was politically unacceptable, yet the pass-the-exposure express washeading swiftly in that direction. Where there was no coverage, businesses wereleft vulnerable to not being able to maintain lines of credit, their expansion plansand other forms of risk-taking were inhibited, their employee benefitprogrammes were threatened, and they were exposed to total loss in the event ofanother terrorist attack. Where there was some terrorism coverage available insome contexts, it was often at a price that many could not afford, leaving themwith the ‘choice’ of going without coverage or reorganizing the financialmanagement and other aspects of their enterprise.

Third, the government was in the best position to insure against terrorismcompared to either insurers or insureds. As introduced previously, governmentalone can create and enforce legal rules to overcome insurance capacity prob-lems, and it has pockets whose depth can be tailored through the power to tax

expost

and finance indemnity over time. Government participation as a reinsurerof last resort could provide premium revenue to build a substantial reserve and/or fund anti-terrorist security measures that might further reduce the risk.Government also has the greatest information resources about terrorism, andtherefore its participation in an insurance programme is vital to both under-writing and preventive security measures in an overall protection package forboth insurers and insured.

The federal government backstop insurance programme was designed tolimit the terrorism exposure of insurers to a known degree, thereby addressingthe severity side of the risk equation. Government military and police securitymeasures were designed to prevent terrorist attacks, thereby addressing thefrequency side of the equation. Speaking at an insurance industry conference, aconsultant to the industry emphasized that ‘the US government is actuallyalready a risk manager in this event. The amount of money that is being spent. . . on the war against terrorism is akin to just loss control that an insurancecompany would undertake in evaluating what it is going to charge.’ An insur-ance company interviewee underscored the importance of governmentalsecurity measures in restoring public and industry confidence that they couldreturn to business as usual.

[T]he more success that our government has in fighting and uncoveringterrorist cells and arresting people and foiling plots, the longer that happens,the more comfortable people get, the more comfortable the industry willget. . . . Perception is a very big part of the insurance industry. They arelemming-like . . . they hear it on the news and everybody is doing it. . . . I seethe government involvement as almost a psychological support more thanfinancial.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

168

Economy and Society

Conclusion

Insurance industry capacity to address catastrophe risk has been central todebates on whether we are entering a risk society of uncontrollable dangers withradical implications for social change. In particular, Ulrich Beck uses lack ofinsurance capacity to manage catastrophe risks as a cornerstone of his theorythat we now live in a ‘world risk society’ of global threats that are beyond thenormal bounds of institutional, scientific, and technological controls.

The

question in these debates is how insurers make decisions about catas-trophe risks. However, relevant data are sparse, and many of the debates are in avacuum, if not vacuous. Beck and most other participants in the debates fail toconduct their own empirical enquiries, and, when they turn to secondarysources, they find a paucity of relevant research.

In order to better inform these debates, we undertook an empirical investiga-tion of how various participants in the insurance relationship responded to theterrorist activity of 9/11, the largest single-event insured loss in history. Thisstudy is instructive on a number of points regarding risk society and insurability.

Beck assumes that insurers will insure only what they can calculate throughscientific expertise applied to the field of risk concerned. According to Beck, infirst modernity there were ‘fixed norms of calculability’ and the insuranceindustry was therefore assured of actuarial precision and predictable loss ratios. Insecond modernity – post-risk-calculation society – these norms are ‘renderedinvalid’ and, by implication, so are institutions such as insurance that depend onthem. To the contrary, our case study shows that insurers will insure that which isnot calculable through scientific expertise or controllable through certain technol-ogies. They convert scientific uncertainty and lack of technological control in thefield of risk concerned into their own science and technology of capital riskdistribution that limits severity of losses. They enforce preventive securityrequirements on the insured – involving environmental design, electronic surveil-lance, and private police – thereby fostering precaution and vigilance that limitsfrequency of loss. They collaborate with governments in both capital risk distribu-tion and preventive security. While their tireless efforts to contain severity andfrequency are always fallible, they nevertheless continue to thrive on uncertainty,always eager to turn threat into opportunity. They impose meaning on uncertaintythrough non-scientific forms of knowledge that are intuitive, emotional, aesthetic,moral, and speculative. They take risks and gamble for profit. As a moderninstitution, insurance has always operated in this way.

Beck concedes that insurers sometimes continue to insure catastrophe risks inconditions of extreme uncertainty, but he regards these efforts as desperateattempts to feign control over what they know is uncontrollable. Our case studyshows that, while newly discovered catastrophe risks pose unique problems ofknowledge and control, insurers do not back off from this ‘insurance curse’ butrather turn threat into opportunity. They respond to the real need to financedamages even if considerable risk-taking is involved, and even if new configura-tions of compensation and preventive security must be imaginatively created.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4


169

They also recognize the real need to take authoritative decisions in face of limitedknowledge, allowing definitive courses of risk management to be pursued.

9/11 was a ‘big crunch’ and the insurance system worked reasonably well inpaying compensation. In this case at least, a modernization risk was governablethrough the insurance system. Of course, such an enormous insured loss hasserious consequences for the insurance system and its continuing capacity toaddress catastrophe risk. In the aftermath of 9/11, there were many tribulations:the pass-the-exposure express, individual cases of unfair compensation, andnew inequalities in insurance prices, terms, and exclusions. But far from‘dissolving’, as Beck would have it, the boundaries of private insurance werereconfigured in collaboration with the state in the ongoing, neo-liberal negotia-tion of political economy. Governments helped insurance markets by usingregulatory mechanisms to forge manageable risk pools, participating in backstopreinsurance and other financing programmes, and expanding preventivesecurity arrangements. Insurance markets helped governments by generatingcapital to address future catastrophe risk and by enforcing preventive securityrequirements on their clients. In sum, the state and insurance industrycombined to address the uncertainties of a new catastrophe risk through both aneo-liberal emphasis on embracing risk and a precautionary approach to lossprevention.

While insurance remains a central institution in world risk society, it doeshave limits. Beyond his overstatements and lack of attention to empirical detail,Beck does offer some pointers in this regard. First, there are growing problemsof both long-tailed and event-specific catastrophe risk that threaten global insur-ance capacity. This was one of the significant fears that permeated the post-9/11 precautionary environment: the insurance system worked this time, but howmany such losses can it take? Second, insurance is a limited form of compensa-tion. It protects against loss of capital, but it cannot replace loss of loved ones,treasured environments, or personal effects that are priceless. Third, insurance-driven preventive security can only go so far, and much depends on govern-ments to provide peace and security. Fourth, in the end insurers are primarilydriven by profits, and capital risk-taking in this regard. All too often theirperceptions and decisions about uncertainty – with potential for windfall profitsas well as catastrophic losses – create crises in insurance availability and provokenew forms of inequality and exclusion. Hence, while the insurance industry is acentral bulwark against uncertainty, insurers can also play a key role in fosteringit. More case studies and systematic research are needed on insurance as anuncertain business.

Note

1 Field research for this case study on terrorism insurance was conducted in 2002 inCanada and the United States. It included attendance at an industry conference oninsurance markets following the terrorist activity of 9/11. The entire conference pro-ceedings were tape-recorded and selectively transcribed. There were sixteen presenters


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

170

Economy and Society

at the conference, including insurance and reinsurance company executives, insuranceindustry regulators, experts in legal liability, and experts in insurance risk and decisionprocesses. In separate field trips, we also conducted open-focused interviews with thir-teen senior executives in insurance companies, reinsurance companies, and regulatoryagencies. We obtained documents from conference participants and interviewees, as wellas from the websites of insurance companies, insurance industry information servicecompanies, and insurance regulatory bodies.

Our analysis is also informed by our wider research programme on the insuranceindustry (Ericson

et al

. 2003; Ericson and Doyle 2004). We conducted 224 open-focusedinterviews in Canada and the United States. Interviewees included people with a widerange of responsibilities in the insurance industry, professionals in expert systems thatserve the industry, representatives of consumer associations, members of the generalpublic who were consumers, industry regulators, senior civil servants, and members ofparliament. Observational research included attending industry conferences, andobserving insurance sales, loss prevention and claims processes. The research wasfunded by the Social Sciences and Humanities Research Council of Canada, the CanadaCouncil Killam Research Fellowship Programme, and the Visiting Fellows Programmeof All Souls College, Oxford.

The terrorist activity of 9/11 continued to reverberate on insurance markets andgovernment policy well beyond our data collection period. Therefore, the reader mustappreciate the limitations of our analysis. We tried to capture how terrorism wasaddressed in insurance markets and government policy in the first year after the catas-trophe.

References

Anderson

,

D.

(1999) ‘The aggregate burden of crime’,

Journal of Law and Economics

42: 611–42.Arrow, K. (1963) ‘Uncertainty and the welfare economics of medical care’, American Economic Review 53: 941–73.Arrow, K. and Lind, R. (1970) ‘Uncertainty and the evaluation of public investment decisions’, American Economic Review 60: 364–78.Baker, T. (2002) ‘Liability and insurance after September 11th: embracing risk meets the precautionary principle’, The Geneva Papers on Risk and Insurance 27: 349–57.—— (2003) Insurance Law and Policy: Cases, Materials and Problems, New York: Aspen Publishers.Baker, T. and Simon, J. (eds) (2002) Embracing Risk: The Changing Culture of Insurance and Responsibility, Chicago, IL: University of Chicago Press.Beck, U. (1992a) Risk society: Towards a New Modernity, London: Sage.—— (1992b) ‘Modern society as a risk society’, in N. Stehr and R. Ericson (eds)

The Culture and Power of Knowledge: Inquiries into Contemporary Societies, Berlin and New York: de Gruyter, pp. 199–214.—— (1992c) ‘From industrial society to risk society’, Theory, Culture and Society 9: 97–123.—— (1994) Ecological Enlightenment: Essays on the Politics of Risk Society, Atlantic Highlands, NJ: Humanities Press.—— (1999) World Risk Society, Cambridge: Polity Press.—— (2002) ‘The terrorist threat: world risk society revisited’, Theory, Culture and Society 19: 39–55.Bougen, P. (2003) ‘Catastrophe risk’, Economy and Society 32: 253–74.Bumiller, E. (2002) ‘Government to cover most costs of insurance losses in terrorism: Bush signs bill seen as aid to building industry’, New York Times 27 November.Business Wire (2002) ‘AIR Worldwide releases terrorism loss estimation model to analyze financial impact of terrorist attacks on insurers, reinsurers and corporations’, 3 September.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

Richard V. Ericson and Aaron Doyle: Catastrophe risk, insurance and terrorism 171

Campbell, R. (2002) ‘America acts: the swift legislative response to the September 11th attack on America’, paper to the conference, Liability and Insurance after September 11th, Insurance Law Center, University of Connecticut School of Law, 21–2 March.Clark, G. (1999) Betting on Lives: The Culture of Life Insurance in England, 1695–1775, Manchester: Manchester University Press.Dean, M. (1999) Governmentality: Power and Rule in Modern Society, London: Sage.Douglas, M. (1966) Purity and Danger: An Analysis of the Concepts of Pollution and Taboo, London: Routledge & Kegan Paul.—— (1992) Risk and Blame: Essays in Cultural Theory, London: Routledge & Kegan Paul.Douglas, M. and Wildavsky, A. (1982) Risk and Culture: An Essay on the Selection of Technical and Environmental Dangers, Berkeley, CA: University of California Press.Ericson, R. and Doyle, A. (2004) Uncertain Business: Risk, Insurance, and the Limites of Knowledge, Toronto: University of Toronto Press.Ericson, R., Barry, D. and Doyle, A. (2000) ‘The moral hazards of neoliberalism: lessons from the private insurance industry’, Economy and Society 29: 532–58.Ericson, R., Doyle, A. and Barry, D. (2003) Insurance as Governance, Toronto: University of Toronto Press.Ewald, F. (2002) ‘The return of Descartes’s malicious demon: an outline of a philosophy of precaution’, in T. Baker and J. Simon (eds) Embracing Risk: The Changing Culture of Insurance and Responsibility, Chicago, IL: University of Chicago Press, pp. 273–301.Froot, K. (ed.) (1999) The Financing of Catastrophe Risk, Chicago, IL: University of Chicago Press.Garland, D. (2003) ‘The rise of risk’, in R. Ericson and A. Doyle (eds) Risk and Morality, Toronto: University of Toronto Press, pp. 48–86.Giddens, A. (1990) The Consequences of Modernity, Cambridge: Polity Press.—— (1994) Beyond Left and Right, Stanford, CA: Stanford University Press.

—— (1998) The Third Way: The Renewal of Social Democracy, Cambridge: Polity Press.Green, M. (2002) ‘Cat models look to predict losses from future attacks’, Best Wire 29 August.Hacking, I. (2003) ‘Risk and dirt’, in R. Ericson and A. Doyle (eds) Risk and Morality, Toronto: University of Toronto Press, pp. 22–47.Haggerty, K. (2003) ‘From risk to precaution: the rationalities of personal crime prevention’, in R. Ericson and A. Doyle (eds) Risk and Morality, Toronto: Univerity of Toronto Press, pp. 193–214.Hartwig, R. (2002) ‘Industry financials and outlook’, Insurance Information Institute Media 26 June.Higgins, V. (2001) ‘Calculating climate, “advanced liberalism” and the governing of risk in Australian drought policy’, Journal of Sociology 37: 299–316.Krane, J. (2002) ‘Terrorism worries bring big business: pitch the product the right way and “you’d be amazed what you can sell”’, The Toronto Star 29 August: D16.Lesser, I., Hoffman, B., Arguilla, J., Ponfellt, D. and Zanini, M. (1999) Countering the New Terrorism, Washington, DC: RAND.Luessenhop, E. and Mayer, M. (1995) Risky Business: An Insider’s Account of the Disaster at Lloyd’s of London, New York: Scribner.Miller, J., Stone, M. and Mitchell, C. (2002) The Cell: inside the 9/11 Plot and Why the FBI and CIA Failed to Stop It, New York: Hyperion.Moss, D. (2002) When All Else Fails: Government as the Ultimate Risk Manager, Cambridge, MA: Harvard University Press.O’Malley, P. (1991) ‘Legal networks and domestic security’, Studies in Law, Politics and Society 11: 171–90.—— (1992) ‘Risk, power and crime prevention’, Economy and Society 21: 252–75.—— (2003) ‘Governable catastrophes: a comment on Bougen’, Economy and Society 32: 275–79.Organization of Economic Cooperation and Development (2002)


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

172 Economy and Society

‘Economic consequences of terrorism’, OECD Economic Outlook 71: 117–40.Public Broadcasting System, Frontline (2002) The Man Who Knew, broadcast video, October.Serio, G. (2001a) Testimony of New York State Insurance Department Before U.S. House Committee on Financial Services, 26 September.—— (2001b) Testimony of New State Insurance Department Before Assembly Speaker Sheldon Silver and Assembly Standing Committee on Insurance, 12 December.—— (2002) Testimony of New York State Insurance Department Before U.S. House of Representatives Committee of Financial Services, Subcommittee on Oversight and Investigations, 27 February.Standard and Poor’s (2002a) ‘Terrorism coverage remains in doubt’, Standard and Poor’s Insurance, 15 April.—— (2002b) ‘U.S. property/casualty insurance midyear outlook 2002: negative fundamentals outweight higher pricing’, Standard and Poor’s Insurance, 28 May.Stehr, N. and Ericson, R. (2000) ‘The ungovernability of modern societies: states, democracies, markets, participation, and citizens’, in R. Ericson

and N. Stehr (eds) Governing Modern Societies, Toronto: University of Toronto Press, pp. 3–25.Stern, J. (1999) The Ultimate Terrorists, Cambridge, MA: Harvard University Press.Strange, S. (1996) The Retreat of the State: The Diffusion of Power in the World Economy, Cambridge: Cambridge University Press.Swiss Re Sigma No. 1 (2002) Natural Catastrophes and Man-Made Disasters in 2001: Man-Made Losses Take on a New Dimension, Zurich: Swiss Reinsurance Company.Swiss Re Sigma No. 3 (2002) The London Market in the Throes of Change, Zurich: Swiss Reinsurance Company.Thompson, Jr., W. (2002) One Year Later: The Fiscal Impact of 9/11 on New York City, New York: Comptroller of the City of New York.Treaster, J. (2002) ‘The race to predict terror’s costs’, New York Times 1 September.Walker, C. and McGuiness, M. (1997) ‘Political violence and commercial victims: high treason against the political economy’, unpublished paper, Faculty of Law, University of Leeds.

Richard Ericson holds the Chair of Criminology and is Director of theCentre for Criminological Research, University of Oxford. He is also a Fel-low of All Souls College, Oxford. From 1993 to 2003 he was Principal ofGreen College and Professor of Law and Sociology, University of BritishColumbia. Prior to that he served as Director of the Centre of Criminologyand Professor of Criminology and Sociology, University of Toronto. He iscurrently working on studies of decision-making and governance in condi-tions of uncertainty. His recent books include Risk and Morality (editedwith Aaron Doyle, 2003), Insurance as Governance (with Aaron Doyle andDean Barry, 2003), and Uncertain Business: Risk, Insurance, and the Limits ofKnowledge (with Aaron Doyle, 2004), all published by University ofToronto Press.

Aaron Doyle is Assistant Professor Sociology, Carleton University. He is cur-rently working with Richard Ericson on studies of insurance industry practices.A former journalist, he is also working on studies of crime, policing, and themedia, and on the relationship between journalistic and social scientific ways ofknowing. His most recent books include Arresting Images: Crime and Policing in


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

Richard V. Ericson and Aaron Doyle: Catastrophe risk, insurance and terrorism 173

Front of the Television Camera (2003), Risk and Morality (edited with RichardEricson, 2003), Insurance as Governance (with Richard Ericson and Dean Barry,2003), and Uncertain Business: Risk, Insurance, and the Limits of Knowledge (withRichard Ericson, 2004), all published by University of Toronto Press.


Dow

nloa

ded

by [

The

UC

Irv

ine

Lib

rari

es]

at 2

0:37

21

Oct

ober

201

4

Documents

Catastrophe risk, insurance and terrorism