ARTICLE IN PRESS

1747-7891/$ - se

doi:10.1016/j.en

�Tel.: +1 213

E-mail addr

Environmental Hazards 7 (2007) 383–398

www.elsevier.com/locate/hazards

Economic resilience to natural and man-made disasters:Multidisciplinary origins and contextual dimensions

Adam Rose�

School of Policy, Planning, and Development (SPPD) and Center for Risk and Economic Analysis of Terrorism Events (CREATE),

University of Southern California, 230 RGL 650 Childs Way, Los Angeles, CA 90089, USA

Abstract

Economic resilience is a major way to reduce losses from disasters. Its effectiveness would be further enhanced if it could be precisely

defined and measured. This paper distinguishes static economic resilience—efficient allocation of existing resources—from dynamic

economic resilience—speeding recovery through repair and reconstruction of the capital stock. Operational definitions are put forth that

incorporate this important distinction. The consistency of the definitions is examined in relation to antecedents from several disciplines.

The effectiveness of economic resilience is evaluated on the basis of recent empirical studies. In addition, its potential to be enhanced and

eroded is analyzed in various contexts.

r 2007 Published by Elsevier Ltd.

Keywords: Resilience; Disasters; Business interruption; Adaptation; Natural hazards; Terrorism

1. Introduction

During the past several years, the world has witnessedsome unprecedented disasters such as the World TradeCenter attacks, Asian Tsunami, and Hurricane Katrina.Policy-makers rushed to assure us of remedial actions toreduce the risk of future potential catastrophes. Wherepossible, they have emphasized preventative measures. Butthe reality is that all future disasters cannot be prevented,in part because of the likelihood that these events willinvolve unexpected forms, magnitudes, or locations.

What is often overlooked is the fact that individuals,institutions, and communities have the ability to deflect,withstand, and rebound from serious shocks in terms of thecourse of their ordinary activities or through ingenuity andperseverance in the face of a crisis. Moreover, this‘‘resilience’’ is often implemented in a relatively costlessmanner, such as conserving resources in short supply,recouping lost production at a later date, or reallocatingresources in response to market signals.

e front matter r 2007 Published by Elsevier Ltd.

vhaz.2007.10.001

740 1716; fax: +1 213 821 3926.

ess: adam.rose@usc.edu

The general concept of resilience was first put forth byecologists more than 30 years ago (see, e.g., Holling, 1973).It has been adapted or re-invented for the case of short-term disasters (see, e.g., Tierney, 1997; Bruneau et al., 2003;Rose, 2004b) and long-term phenomena, such as climatechange (see, e.g., Timmerman, 1981; Dovers and Handmer,1992). Klein et al. (2003) have noted, however, and thisauthor concurs, that some definitions of resilience are sobroad as to render the term meaningless. At the same time,few analysts have delved deeply into its economic aspects.The focus of this paper is the economic dimensions of

resilience. One key dimension relates to time. Anotherrelates to the context in which resilience takes place. Theconcept of static economic resilience is essentially makingthe best of the resources available at a given point in time,as distinct from the dynamic implications of repair andreconstruction, which affect the time-path of the economy.This general definition and the distinction is very much inkeeping with the essence of the economic problem ingeneral—the efficient allocation of resources at a pointin time and in regard to the future. Disasters make thisproblem all the more dire. Static resilience has greatpotential to reduce disaster losses in a straightforward andinexpensive manner, but this aspect has usually been

ARTICLE IN PRESS

(footnote continued)

considerations, the parameters of Computable General Equilibrium

models are typically based on normal operating experience and long-term

adjustment, and would therefore overstate the flexibility, and hence the

A. Rose / Environmental Hazards 7 (2007) 383–398384

overlooked in favor of its dynamic resilience counterpart,which is focused on the speed of recovery and which hasdominated most of the engineering-based literature on thesubject. Dynamic resilience is thus more complex from aneconomic standpoint and more expensive. It is, of course,no less important, but is given less attention in this paper infavor of compensating for the neglect of static resilience inthe literature to date. Likewise, contextual effects ofresilience analyzed in this paper will primarily be relatedto the static economic version. Context refers to conditions(e.g., economic structure, severity of damage) that affectthe enhancement, erosion, and implementation of thisimportant way of coping with disasters.

The purpose of this paper is to establish the basis for aconsistent and comprehensive formulation of economic

resilience. This is accomplished by first identifying aspectsunique or primary to the economic realm and making animportant distinction between static and dynamic versionsof this concept. Second, I examine and reconcile definitionsof resilience from several disciplines. Third, I put forthoperational definitions. Fourth, I discuss an estimate of thestrength and cost-effectiveness of this important feature ofdisaster response. Fifth, I identify tangible actions that canenhance economic resilience and discuss how resilience iseroded by extreme conditions.

The presentation of a precise definition is importantbecause resilience is in danger of becoming a vacuousbuzzword from overuse and ambiguity. For example, severalmajor academic works on the subject published recently donot even bother to define it (see, e.g., Chernick, 2005; Valeand Campanella, 2005).1 Some progress has been made inseveral major public policy initiatives (CBO, 2004; DHS,2006; TISP, 2006). However, definitional ambiguities, incon-sistencies within some guidelines, and incompatibilities acrossthem will further hinder policy design and implementation.

The presentation here is also important because eco-nomic resilience does not appear to be adequatelyappreciated as a cost-effective tool to manage catastrophicrisk. One example is the view that price increases in theaftermath of a disaster represent ‘‘gouging’’ rather than, inmany cases, a useful signal of increased scarcity. Alsounderestimated is the effectiveness of common senseresponses by individuals, as well as the professionalizationof this strategy through the formation of the new businesscontinuity service industry.

Consideration of resilience is critical in assessingpotential losses from disasters and evaluating the benefitsof their mitigation. Disaster loss estimation is still less thanfully developed, and many models in current use are limitedby rigidities and lack of behavioral content, or by overlyflexible structures (see the review by Rose, 2004a).2

1This is not to detract from the many excellent papers in both of these

volumes. However, only one of the 20 papers between them offers a

precise definition of resilience.2Input–Output models are linear and lack behavioral content. Econo-

metric models are typically based on time series, which means they are an

extrapolation of the past. Although non-linear and including behavioral

Accurate estimates of disaster losses at the level of theindividual firm, the market, and the macroeconomy arecritical to the evaluation of risk-management strategies.Under-estimation of losses will result in too few resourcesapplied to the problem, while over-estimation of losses willlead to excess resources being applied. Therefore, theanalysis below is intended to be helpful to academicresearchers in several disciplines to understand the natureand role of economic resilience. It should be useful to thoseinvolved in economic loss estimation in accurately asses-sing the net economic impacts of disasters. Finally, itshould be useful to public and private decision-makersforming judgments about the promotion, safeguarding,and implementation of resilience.

2. Defining economic resilience

I define static economic resilience as the ability of anentity or system to maintain function (e.g., continueproducing) when shocked (see also Rose, 2004b). It is thusaligned with the fundamental economic problem—efficientallocation of resources, which is exacerbated in the contextof disasters.3 This aspect is also interpreted as staticbecause it can be attained without repair and reconstruc-tion activities, which affect not only the current level ofeconomic activity but also its future time-path. Anotherkey feature of static economic resilience is that it isprimarily a demand-side phenomenon involving users ofinputs (customers) rather than producers (suppliers). Itpertains to ways to use resources available as effectively aspossible. This is in contrast to supply-side considerations,which definitely require the repair or reconstruction ofcritical inputs.A more general definition that incorporates dynamic

considerations is the speed at which an entity or systemrecovers from a severe shock to achieve a desired state.This also subsumes the concept of mathematical or systemstability because it implies the system is able to bounceback. This version of resilience is relatively more complexbecause it involves a long-term investment problemassociated with repair and reconstruction. It is also morecomplicated because it involves serious tradeoffs, as, forexample, when haste in reconstruction might leavebusinesses or the economy in general more vulnerable tofuture disasters. Of course, resilience can be enhancedbefore a disaster, as in the design of more flexible

resilience of a system. All these models can be refined, though to different

degrees, to explicitly incorporate resilience (see, e.g., Rose and Lim, 2002;

Rose and Liao, 2005; Rose et al., 2007a).3Other considerations relating to resource allocation besides efficiency

are appreciated by the author, as, for example, the importance of equity,

or fairness (see, e.g., Rose et al., 1988; Rose and Kverndokk, 1999).

However, they are beyond the scope of this paper.

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398 385

production processes, stocking of inventories, and byholding emergency planning drills, but these investmentsare rather modest compared to both the mitigation anddynamic resilience reconstruction decisions.

Ability implies a level of attainment will be achieved.Hence, the definition is contextual—the level of functionhas to be compared to the level that would have existed hadthe ability been absent. This means a reference point ortype of worst case outcome must be established first.Further discussion of this oft-neglected point is deferred toSection 4 below.

I also distinguish two types of resilience in each context.Inherent resilience refers to the ordinary ability to deal withcrises (e.g., the ability of individual firms to substitute otherinputs for those curtailed by an external shock, or theability of markets to reallocate resources in response toprice signals). This is itself a type of resource already inplace that can be enhanced prior to a disaster and so thatcapabilities that are not damaged or eroded can beimplemented in the disaster aftermath.

Adaptive resilience in contrast refers to the ability incrisis situations to maintain function on the basis ofingenuity or extra effort (e.g., increasing input substitutionpossibilities in individual business operations or strength-ening the market by providing information to matchsuppliers with customers). This corresponds to pushing theefficiency frontier outward, though not necessarily with thehelp of any investment.

Resilience emanates both from internal motivation andthe stimulus of private or public policy decisions (Mileti,1999). Also, resilience, as defined in this paper, refers topost-disaster conditions and response (Comfort, 1994),which are distinguished from pre-disaster activities toreduce potential losses through mitigation (cf., Bruneauet al., 2003). In disaster research, resilience has beenemphasized most by Tierney (1997) in terms of businesscoping behavior and community response, by Comfort(1999) in terms of non-linear adaptive response oforganizations (broadly defined to include both the publicand private sectors), and by Petak (2002) in terms of systemperformance. These concepts have been extended topractice. Disaster recovery and business continuity indus-tries have sprung up that offer-specialized services to helpfirms during various aspects of disasters, especially poweroutages (see, e.g., Business Continuity Institute, 2002;Salerno, 2003). Key services include the opportunity tooutsource communication and information aspects of thebusiness at an alternative site. There is also a growingrealization of the broader context of the economic impacts,especially with the new emphasis on supply chain manage-ment (Hill and Paton, 2005; Sheffi, 2005). One enlightenedcompany executive recently summarized the situation quitepoignantly and in modern business terms: ‘‘In short,companies have started to realize that they participate ina greater ecosystem—and that their IT systems are only asresilient as the firms that they rely on to stay in business’’(Corcoran, 2003, p. 28). Experience with Y2K, 9/11,

natural disasters, and technological/regulatory failures, aswell as simulated drills, have sharpened utility industry andbusiness resilience (Eckles, 2003). Similar activities ofpublic sector agencies have improved community disasterresilience (Godschalk, 2003).Resilience can take place at three levels:

�

Microeconomic—individual behavior of firms, house-holds, or organizations. � Mesoeconomic—economic sector, individual market, or

cooperative group.

� Macroeconomic—all individual units and markets com-

bined, including interactive effects.

Examples of individual resilience are well documented inthe literature, as are examples relating to the operation ofbusinesses and organizations (Tierney, 1997; Rose andLim, 2002). What is often less appreciated by disasterresearchers outside economics and closely related disci-plines is the inherent resilience of markets. Prices act as the‘‘invisible hand’’ that can guide resources to their bestallocation even in the aftermath of a disaster. Some pricingmechanisms have been established expressly to deal withsuch a situation, as in the case of non-interruptible servicepremia that enable customers to estimate the value of acontinuous supply of electricity and to pay in advance forreceiving priority service during an outage (Chao andWilson, 1987). The price mechanism is a relatively costlessway of redirecting goods and services. Those priceincreases, to the extent that they do not reflect ‘‘gouging,’’serve a useful purpose of reflecting highest value use, evenin the broader social setting (see also Schuler, 2005).Moreover, if the allocation does violate principles of equity(fairness), the market allocations can be adjusted byincome or material transfers to the needy. Of course,markets are likely to be shocked by a major disaster, in ananalogous manner to buildings and humans. In this case,we have two alternatives for some or all of the economy:(1) substitute centralized decree or planning, though at asignificantly higher cost of administration; (2) bolster themarket, such as by improving information flows (e.g., thecreation of an information clearinghouse to matchcustomers without suppliers to suppliers without custo-mers). Both approaches are forms of resilience.At the macroeconomic level, there are a large number of

interdependencies through both price and quantity inter-actions that influence resilience. That means resilience inone sector can be greatly affected by activities related to orunrelated to resilience in another. This makes resilience allthe more difficult to measure and to influence in the desiredmanner. This includes situations in which the whole is notsimply the sum of the parts. An example is offered by Roseand Benavides (1999), where a system of individuallystructured non-interruptible service premia may not besocially optimal, because individual businesses makedecisions on whether to pay the premium on the basis oftheir own benefits, but ignore benefits to their direct or

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398386

indirect suppliers and customers. In this context, resilienceis not only a function of individual business or householdactions but also all the entities that depend on them or thatthey depend on directly or indirectly.

3. Comparison with related concepts of resilience

The purpose here is not demonstrate that the definitionsof this author are correct and those of others are not. Infact, the intent is to focus on points of agreement and toincorporate the work of others into the formulation ofeconomic resilience. Criteria for conceptual and opera-tional definitions of resilience, inclusive of some of itsvarious dimensions and at the exclusion of others, shouldbe based on consistency with fundamental economicprinciples, the needs of potential users, and the practicalmatters of data availability and computational manage-ability. My formulation of resilience is dependent onprecedents in the established literature in ecology, econom-ics, and related fields over the past 30 years.

3.1. Ecological origins

As in many other fields, some researchers on the subjectof resilience have re-invented the wheel narrowly in theirown discipline, rather than looking carefully for precedentsor at the big picture. To begin, ecologists have pioneered auseful, broad definition of resilience relating to the survivalof complex systems. Holling (1973, p. 17) is typically citedas the first to have defined resilience, his definition being‘‘the ability of systems to absorb changes y and stillpersist.’’ He sometimes refers to it as ‘‘buffer capacity.’’

Adger (2000) suggests there is no single definition ofecological resilience, and offers two definitions analogousto my static and dynamic economic definitions above. Animportant contrast in the static definitions exists, however.The ecological definition emphasizes the amount ofdisturbance the system can absorb without incurring achange in its state. In economics, only the most severehazard (a catastrophe) results in such a change, and thussuch a definition would be of very limited usefulness.Instead I use the term resilience here more in line with thebuffer concept, as the ability to mute the influence of theexternal shock. It is not just the decrease in economicactivity, but rather the actual decrease relative to thepotential decrease (see also the mathematical definitions inSection 5 below). Perrings (2001, p. 322) also definesresilience in a relative manner: ‘‘As a first approximation,this may be measured by an index of the level of pollutionor depletion relative to the assimilative or carrying capacityof the ecological system concerned.’’ Subsequently, Per-rings (p. 323) defines it in terms of the ‘‘gap betweencurrent and critical loads’’ to the ecosystem and even theecological economic system (though this would seem tohave application to engineered systems as well).

Here and below it is important to distinguish the conceptof resilience and related terms. For example, Holling (1973,

p. 17) defines stability as ‘‘the ability of a system to returnto equilibrium after a temporary disturbance.’’ Thisdefinition is often put forth as the essence of resilience orat least a special dimension. However, it is clear thatresilience and stability are distinct. As Handmer andDovers (1996) point out, a stable system may not fluctuatesignificantly, but a resilient system may undergo significantfluctuation and return to a new (and, implicitly, animproved) equilibrium rather than the old one.Several ecologists and ecological economists have linked

resilience to the concept of sustainability, which refers tolong-term survival and at a non-decreasing quality of life(see, e.g., Common, 1995; Perrings, 2001). Common (1995)suggests that resilience is the key to this concept. A majorfeature of sustainability is that it is highly dependent onnatural resources, including the environment. Destroying,damaging, or depleting resources undercuts our longer-term economic viability, a lesson also applicable to hazardimpacts where most analysts have omitted ecologicalconsiderations. Klein et al. (2003) note that, from aneconomic perspective, sustainability is a function of thedegree to which key hazard impacts are anticipated.However, I agree with the position that it is also a functionof a society’s ability to react effectively to a crisis, and withminimal reliance on outside aid (see Mileti, 1999).In the context of longer-term disasters, such as climate

change, Timmerman (1981) defined resilience as themeasure of a system’s capacity to absorb and recover fromthe occurrence of a hazardous event. Dovers and Handmer(1992) note an important feature that distinguishes manfrom the rest of nature in this context—human capacity foranticipating and learning. They then bifurcate resilienceinto reactive and proactive, where the latter is uniquelyhuman. I maintain that proactive efforts can enhanceresilience by increasing its capacity prior to a disaster, butthat resilience is operative only in the response/recovery/reconstruction (often referred to as ‘‘post-disaster’’) stages.Adger (2000) was one of the first to extend the ecological

definition of resilience to human communities as a whole.He measured social resilience as related to social capital andin terms of economic factors (e.g., resource dependence),institutions (e.g., property rights), and demographics (e.g.,migration). Mileti and collaborators (Mileti, 1999) analyzedmany aspects of resilience to hazards in the attainment ofsustainable communities. However, Mileti (1999, p. 5) wenttoo far in defining a resilient community as not only onethat ‘‘can withstand an extreme event with a tolerable levelof losses’’ but also one that ‘‘takes mitigation actionsconsistent with achieving that level of protection.’’ Mycomment is no way a criticism of mitigation, which has beenfound to be cost-effective in countless applications and isstill underutilized (see, e.g., Rose et al., 2007c), but ratherthat mitigation is distinct from resilience for several reasonsdiscussed in the following sub-section.Timmerman and others also relate resilience to vulner-

ability. Specifically, Pelling (2003) decomposes vulnerabil-ity to natural hazards into three parts: exposure, resistance,

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398 387

and resilience. As does Blaikie et al. (1994), Pelling definesresilience to natural hazards as the ability of an individualto cope with or adapt to hazard stress. My view is thatvulnerability is primarily a pre-disaster condition (see alsoCutter, 1996), but that resilience is the outcome of a post-disaster response. Resilience is one of several ways toreduce vulnerability, the others being adaptation and theentirely separate strategy of mitigation.

3.2. Engineering-based definitions

Bruneau et al. (2003) provide a comprehensive andsophisticated analysis of the many aspects of earthquakeloss reduction all under the heading of resilience. Theauthors apply the concept at four levels: technical,organizational, social, and economic. They contend thatresilience has four dimensions, which are listed below alongwith a definition applied to the economic level:

(1)

4B

perf

syst

in d

resp

case

and

obst

Robustness—avoidance of direct and indirect economiclosses;

(2)

Redundancy—untapped or excess economic capacity(e.g., inventories, suppliers);

(3)

Resourcefulness—stabilizing measures (e.g., capacityenhancement and demand modification, external assis-tance, optimizing recovery strategies);

(4)

Rapidity—optimizing time to return to pre-eventfunctional levels.

Bruneau et al. also stipulate that the resilience of asystem has three aspects:

(a)

e

y

reduced probability of failures;

(b) reduced consequences from failures; (c) reduced time to recovery.

The relationship between the dimensions and aspects ofa resilient system differs from the definition of economicresilience presented here in the following ways:

�

My definition excludes the dimension of reducedprobabilities of failure because this is more pertinentto measures taken before an event, primarily for thepurpose of mitigation. � Reduced consequences from failure comes the closest to

my definition of static resilience.

5

� In a similar vein, Chang and Shinozuka (2004, p. 741) state that: ‘‘It is

useful to view robustness and rapidity as the desired ends of resilience-

enhancing measures. Redundancy and resourcefulness are some of the

Reduced time to recovery is the same as my definition ofdynamic resilience, though the state of restoration ismore general in my formulation.4

runeau et al. include ‘‘restoration of the system to its ‘normal’ level of

ormance’’ to their definition. This definition subsumes whether a

m can ‘‘snap back’’ at all, i.e., the concept of stability as typically used

namics. My use of the term desired state is a generalization of possible

onses, which would include return to pre-disaster status as a special

, but would at the same time allow for growth and change over time

implementation of mitigation practices, as well as considering

acles to achieving the desired state.

�

me

sta

diff

the

sub

eco

its

haz

ind

Robustness is also similar to my definition of staticresilience and is a commonly used term in engineering toconvey this more narrow definition of resilience.

� Redundancy is primarily a supply-side mitigation

strategy; however, some examples are a subset ofresilience responses in my formulation (e.g., diesel-powered electricity generators).

� Resourcefulness is a major feature affecting adaptive

resilience as I have defined it.

� Rapidity is consistent with my definition of dynamic

resilience, though the Bruneau et al. formulation ismore restrictive in that it requires the condition ofoptimization.

This discussion is not intended as a criticism of theexcellent analytical framework developed by Bruneau et al.(2003) per se. Rather, it is a criticism of their choice ofterminology, which includes all aspects of hazard lossreduction under the banner of resilience. The exposition byKlein et al. (2003) is consistent with my argument to keepthe definition of resilience from becoming too broad. Theypropose the concept of ‘‘adaptive capacity’’ as the umbrellaconcept that covers many of the features identified byBruneau et al. This is also more consistent with definingresilience as an outcome or system attribute rather than asa tactic like mitigation.5 Adaptation is also the complementto mitigation. When negative forces (e.g., conventionalhazards, climate change) cannot be or are not mitigated, wetypically resort to adaptation.It would appear that some analysts, such as Mileti and

Bruneau et al., have envisioned a goal of a community thatis able to take many steps to minimize its vulnerability tohazards. Resilience has become a convenient term tocharacterize all of these possibilities. However, this broadusage is inconsistent with the etomology of the term ingeneral (resilio, meaning rebounding), and its use inecology, economics and other areas of hazards research.Ideally, another term can be found to modify this idealcommunity, so that the term ‘‘resilience’’ can be applied tothe sub-set of characteristics to which it is well suited.

3.3. Organizational behavior

Organizational (also institutional) behavior focuses onresilience as a process (Hill and Paton, 2005). As such, it isa strategy in risk management under the sub-heading of

ans to these ends.’’ Again, robustness and rapidity correspond to my

tic and dynamic definitions of resilience, respectively. The major

erence between ends and means is an important reason not to extend

definition of resilience beyond the ends theme. Note also that

sequently in their paper, Chang and Shinozuka define robustness in

nomic terms as the reduction in Gross Regional Product, rather than

deviation from a maximum possible level given the characteristics of the

ard stimulus, as in this paper. Rapidity is defined by them,

ependently, in the same manner as in this paper, however.

ARTICLE IN PRESS

(footnote continued)

however, would not be prudent in the case of resilience and various related

A. Rose / Environmental Hazards 7 (2007) 383–398388

crisis and continuity management. Paton and Johnston(2001) define resilience in this dimension as ‘‘a capacity ofpeople and systems that facilitate organizational perfor-mance to maintain functional relationships in the presenceof significant disturbances as a result of a capability todraw upon their resources and competencies to manage thedemands, challenges and changes encountered.’’ Thisviewpoint extends even more fundamentally to naturalecosystems, whereby The Resilience Alliance (2005) in-cludes as one of its three dimensions of resilience ‘‘thedegree to which the system is capable of reorganization.’’Adger et al. (2005) extend this to the social-ecologicalnexus.

Comfort (1994) was one of the first researchers toventure into this area. Her definition is more narrow thanthe generic one that was the focus of the previous sub-section, because she confines resilience to actions andprocesses after the event occurs, or, as noted in my critiqueof Bruneau et al. (2003), appropriately limits the definitionto reducing the consequences of failure. This also relates toprocess-oriented counterparts of the concept of dynamicresilience, where the focus is not on attaining a target levelof output but rather a target level of ‘‘functioning.’’However, the trajectory of this functioning is clear from themajor themes of non-linear and adaptive dynamics(Comfort, 1999). It also leaves no doubt that the dynamicversion of resilience, the ability to bounce back (or therapidity to do so) is uniquely applicable to the post-disasterstages. Moreover, the recovery process this characterizes isanother way of reducing the consequences of the hazardensuing from structural or system damage (‘‘failure’’ in theBruneau et al. terminology).

Klein et al. (2003) have taken this even further to suggestthat resilience goes beyond the Holling definitions, and byimplication those I propose, to include the functioning andinteraction of interlinked systems (see also UN/ISDR,2002). Again, I note that this does not go as far assuggesting resilience includes all aspects of adaptation ormitigation.

In contrast to resilience activities I have previouslymodeled and discuss in this paper (e.g., conservation,import substitution, market strengthening), the focus oforganizational theory is on ‘‘competencies and systems’’(Hill and Paton, 2005). The relationship between the twoapproaches can be viewed as follows: most standardtreatments of resilience in economics identify a set ofoptions and assume that managers can optimize amongtheir choices (see, e.g., Rose and Liao, 2005). Organiza-tional analysis identifies vulnerabilities and limitations inmanagerial abilities and how they can be overcomethrough resilience.6 The economics approach to reconciling

6Shaw and Harrald (2004) stress that for this to be successful

organizational relationships and authorities must first be defined. In a

situation analogous to issues emphasized in this paper, they point to the

need to reconcile basic disagreements over the definitions of key concepts

such as ‘‘crisis management’ and ‘‘business continuity management.’’

Their solution of combining the two terms into a single umbrella concept,

these two views would be to assume some form of‘‘bounded rationality’’ (see, e.g., Gigerenzer and Selten,2002) and to view managerial resilience as an improvementover the basic outcome. Hill and Paton (2005) analyzeseveral aspects of the theory and practice of businesscontinuity management and how it relates to resilience.They emphasize that a major prerequisite of success in thisarea is the willingness of an organization to adapt to itsnew environment.7

3.4. Planning

Sustainable communities and smart growth emanatefrom the collaborative visions of ecologists, economists,and planners. Thus far, the planners have been mostprominent at practical approaches to the broader design,while the two former disciplines have been more niche-oriented, including the nexus of ecological economics inreorienting individual business operations to principles ofindustrial metabolism (see, e.g., Ayres and Simonis, 1989;Daly and Farley, 2004).The planning profession has as a goal the creation

of hazard-resilient communities (Burby et al., 2000;Godschalk, 2004), primarily through the area of land use.This holistic approach is superior to the piecemeal way thatordinary hazard mitigation is usually promulgated, whichhas actually enticed development in hazardous areas. Forexample, the presence of dikes and levees in New Orleansgave residents a feeling of false security. Many similarexamples have led to the general trend of fewer disasterevents, but the ones now taking place have relatively muchlarger damages. Smart growth has tended to avoid suchoutcomes. Mileti (1999) has started that ‘‘no singleapproach to bringing sustainable hazard mitigation intoexistence shows more promise at this time than increaseduse of sound and equitable land-use management.’’Burby et al. (2000) identify four major themes of how to

integrate mitigation in land-use planning can promotecommunity resilience, but only one of them, and only inpart, pertains to the post-disaster period. This points to thetension in the planning field about terminology, similar tothe discussion in other fields. Godschalk (2003, p. 137)concludes that ‘‘Traditional hazard mitigation programshave focused on making physical systems resistant todisaster forces’’ [my emphasis added]. He goes on to state,however, that ‘‘future mitigation programs must also focuson teaching the city’s social communities and institutionsto reduce hazard risk and respond effectively to disasters,

terms such as sustainability, adaptation, vulnerability, or mitigation.7Broader dimensions of resilience in terms of the social fabric or

community are not discussed here because they are beyond the scope of

this paper (see, e.g., Tobin, 1999; Paton and Johnston, 2001). These

dimensions focus on aspects of resilience, such as psychology, sociology,

and community planning, that are important to a holistic view of the topic

of resilience and are discussed elsewhere in this volume.

ARTICLE IN PRESS

8Note that the static definition presented here (based on Rose, 2004b) is

couched in deterministic terms. Though their definition of resilience (an

off-shoot of the definition by Bruneau et al., 2003) differs from the one

presented here, Chang and Shinozuka (2004) make a major contribution

by providing a framework and illustrative example for evaluating

economic resilience in probabilistic terms and in relation to performance

objectives.

A. Rose / Environmental Hazards 7 (2007) 383–398 389

because they will be the ones most responsible for buildingultimate urban resilience.’’ In fact, Geis (2000) hasexplicitly stated a preference for the term ‘‘disaster-resistance’’ with respect to planning themes and practicesin this area, concluding it is more ‘‘fitting and moremarketable than disaster resilient.’’ At the same time, otherplanners have come to apply the term ‘‘resilient’’ to theinteraction of physical and social systems (Olshansky andKartez, 1998).

Godschalk makes the point, however, that ‘‘Resilientcities are constructed to be strong and flexible, rather thanbrittle and fragile.’’ It is this flexibility (adaptability) that isthe key to resilience as interpreted by others (e.g., Comfort,1999; Rose, 2004b). Foster (1997) interprets this in terms ofcoping with contingencies. He has put forth 31 principlesfor achieving resilience, among them in the general systemsrealm, such characteristics as being diverse, renewable,functionally redundant, with reserve capacity achievedthrough duplication, interchangeability, and interconnec-tions.’’ Godschalk summarizes the work of severalresearchers to identify eight categories of resilienceresponses, seven of which have been emphasized by Rose(2004b, 2006) and in this paper: redundant, diverse,efficient, autonomous, strong, adaptable, and collabora-tive. Finally, Godschalk proposes a more enlightened set ofmitigation measures for social and institutional resiliencethrough the reduction of business interruption impacts,though the specific policy instruments he mentioned arelimited to loans and general government assistance, ratherthan the self-motivated coping behavior emphasized in thispaper.

4. Quantifying resilience

In this section, I provide admittedly crude mathematicaldefinitions of resilience in both static and dynamiccontexts. Direct static economic resilience (DSER) refersto the level of the individual firm or industry (microand meso levels) and corresponds to what economistsrefer to as ‘‘partial equilibrium’’ analysis, or the operationof a business or household entity itself. Total staticeconomic resilience (TSER) refers to the economy as awhole (macro level) and would ideally correspond to whatis referred to as ‘‘general equilibrium’’ analysis, whichincludes all of the price and quantity interactions in theeconomy (Rose, 2004b). In terms of actual measurement ofthe ‘‘indirect’’ portion of resilience, input–output (I–O)models of disaster impacts capture only quantity inter-dependence, often referred to as multiplier effects.Computable general equilibrium (CGE) models andmacroeconometric models capture both price and quantityinteraction through the explicit inclusion of market forces(see Rose, 2005).

An operational measure of DSER is the extent to whichthe estimated direct output reduction deviates from thelikely maximum potential reduction given an externalshock, such as the curtailment of some or all of a

critical input:

DSER ¼%DDY m �%DDY

%DDY m , (1)

where %DDYm is the maximum percent change in directoutput and %DDY the estimated percent change in directoutput.In essence DSER is the percentage avoidance of the

maximum economic disruption that a given shock couldbring about. A major measurement issue is what should beused as the maximum potential disruption. For ordinarydisasters, a good starting point is a linear, or proportional,relationship between an input supply shortage and thedirect disruption to the firm or industry. This would beconsistent with the context of an I–O model, which isinherently linear. The application of a simple version ofthis type of model implicitly omits the possibility ofresilience. Note that while a linear reference point mayappear to be arbitrary or a default choice, it does have anunderlying rationale. A linear relationship connotesrigidity, the opposite of the ‘‘flexibility’’ connotation ofstatic resilience defined in this paper. Aspects of non-linearities in the context of an extreme disaster, or acatastrophe, are discussed below.8

Analogously, the measure of TSER to input supplydisruptions is the difference between a linear set of generalequilibrium effects, which implicitly omits resilience and anon-linear outcome, which incorporates the possibility ofresilience. From an operational modeling standpoint this isthe difference between linear I–O multiplier and CGE, orother comprehensive, non-linear (e.g., econometric) model,impacts as follows:

TSER ¼%DTY m �%DTY

%DTY m ¼M%DDY m �%DTY

M%DDY m ,

(2)

where M is the economy-wide input–output multiplier,%DTYm the maximum percent change in total output,%DTY the estimated percent change in total output.My definitions of economic resilience have been stated in

flow terms in relation to economic output for a givenperiod in time. Is resilience applicable to stocks, i.e.,property damage, as well? While property is important, theflow of goods and services it contributes to economic well-being is paramount. In relation to ecosystems, Holling(1973) defines resilience in terms of flow (productivitymeasures) as opposed to stocks. Resilience of moreconventional capital assets (buildings, infrastructure)would pertain to the ability of the stock to absorb shocks(e.g., a building to withstand ground motion or the blast

ARTICLE IN PRESS

Initial drop in customer output due to electricity outage Productivity improvements by customers Upper limit of customer resilience Erosion of customer resilience Repair & reconstruction of electricity system

YDR

YDU

t9 t10 t11 t12 t4 t1 t2 t5 t7 t8

A B

t0 t3 t6

Regional Economic

Output

Y0

YN

Time

YD

Fig. 1. Static and dynamic resilience in the context of business interruption.

9Haimes et al. (2005a) combined engineering and economic considera-

tions in a useful disaster impact and policy framework—the Inoperability

input–output model. ‘‘Inoperability’’ notes the system’s dysfunction,

‘‘expressed as a percentage of the system’s ‘as planned’ level of operation’’

(p. 68). They define resilience in dynamic terms in two contexts: (1) for

sector damage by a terrorist attack, it is the recovery rate; and (2) for a

sector affected by an ensuing (interdependent) demand reduction, it is the

production adjustment rate. An important contribution is their dynamic

A. Rose / Environmental Hazards 7 (2007) 383–398390

from a terrorist bomb). This would best be consideredunder the purview of engineering resilience. A morecomplex system, however, raises other issues. For example,an electricity system might be said to be less likely to fail ifit has incorporated redundancy of power lines, or bettercommunication between operators to avoid cascadingfailures (see, e.g., Lave et al., 2007). Again, this might beconsidered engineering resilience or perhaps economicresilience on the supply side (as opposed to the demand-side resilience that is the focus here).

Also, while the entire time-path of resilience is key to theconcept for many analysts, it is important to rememberthat this time-path is composed of a sequence of individual

steps. Even if ‘‘dynamics’’ are the focal point, it isimportant to understand the underlying process at eachstage: why an activity level is achieved and why that leveldiffers from one time period to another. As presented here,static resilience helps explain the first aspect, and changesin static resilience, along with repair and reconstruction ofthe capital stock, help explain the second.

Several considerations discussed thus far can be illu-strated in Fig. 1, which represents resilience in the wake ofa total power outage caused by a natural or man-madehazard. In it, the vertical axis represents the level ofeconomic activity, Y, and the horizontal axis representstime, t. The normal level of output (abstracting fromconsiderations of economic growth for ease of expositionand without loss of generality) proceeds at YN until someexternal shock takes place. The result of this disruption inthe presence of static resilience is a reduction in output toYD, as opposed to a total shutdown of the economy to Y0.That is, static resilience is the ratio of the avoided drop inoutput and the maximum potential drop to Y0, or(YD�Y0)/(YN�Y0), or the ratio of line segments B and A

(B/A). In the initial period, adaptive behavior (ingenuity) islikely to be minimal, and the measure is likely to bedominated by inherent resilience.Note also that Fig. 1 provides some important insights

into dynamics of the issue. In the literature on resilience,dynamics often refers to the issue of stability or to thespeed of recovery. The real interesting question here is thepattern of recovery—how much recovery takes place ineach time period and why. The case of individual businessrecovery, as distinct from the repair and rebuilding of thecapital stock provides us with considerable insight. In thisregard, suppose the disruption of the electricity network isdue to the destruction of a major transformer that requiresseveral time periods (ti) to replace. The upward movementin output following the initial decline due to the disaster,YD, would represent basic improvements in resiliencethrough adaptive behavior in t1 and t2. A temporaryequilibrium is reached and persists until t5, when deteriora-tion in static resilience might start to take place (e.g.,inability to sustain Draconian conservation, permanentloss of customers that reduces the possibility of productionrescheduling, and even dissipation of inherent resiliencesuch as substitution possibilities). The next upswing in YD

does not take place until t9, and then as a combination ofrepair/replacement of the transformer (and its phasing in ofoperation) and of remaining static resilience capabilities.9

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398 391

Dynamic resilience would then be defined as the loss-reducing effect of hastening repair and reconstruction ofthe capital stock over and above business as usualpractices. It is best defined in terms of its total effect:

TDER ¼Xn

t¼0

Y DR �Xm

t¼0

Y DU , (3)

where m4n.This would have two interpretations. First, it would be

the overall reduction in the ‘‘loss triangle’’ (this losstriangle is the area between YN and the path of YD). Thisreduction would be the difference between the resilientresponse path, YDR, and the ‘‘normal’’ course of therecovery, YDU (note that YDR and YDU overlap until timeperiod 9). This definition would include static resilience andthe loss-reducing effects of hastening repair and recon-struction. A purist, however, would probably argue that itshould only include the latter feature:

TDER0 ¼Xn

t¼0

Y DR �Xm

t¼0

Y DU � TSER. (30)

For the sake of consistency one might want to exclude therepair and reconstruction aspects from the static definitionas well, thereby limiting it to the YDU time-path.

The definitions of resilience are sufficiently general toallow for an important extreme outcome. Note that in Fig.1 the level of economic output increases beginning at t9 buttapers off beginning at t11 to a level below the pre-disasterlevel (YN). This reflects the possibility that a lengthyrecovery will cause customers of disrupted businesses tolook to other suppliers, possibly on a permanent basis (seealso the discussion in Section 6 below).

Another dimension of economic resilience is the ex-istence of both demand-side and supple-side considera-tions. The discussion in this paper focuses on the demand-or customer-side for a good reason. From Fig. 1, we seethat customer resilience alone is responsible for the increasein output from t0 to t2. Recovery of the capital stock, orsupply-side efforts, are the domain of the electricityprovider and are a completely separate matter. At a givenpoint in time, meaning with a given fixed capital stock,resilience is completely a demand-side issue in the contextof an electricity, or any critical input, supply disruption. Ifthe physical plant of the business is damaged as well, aswould more likely be the case in an earthquake as opposedto a targeted terrorist attack on an electric utility, the

(footnote continued)

model, which explicitly includes capital stock variables critical to the

ultimate recovery process. Although they acknowledge the pace of

recovery is variable, they offer no additional insights into variations in

resilience, beyond equipment related considerations pertaining to the

electric power sector or electrical equipment (see Haimes et al., 2005b). In

fact, they emphasize how ‘‘hardening’’ and other risk mitigation efforts

increase resilience during recovery (Haimes et al., 2005a, p. 74) and

otherwise neglect post-disaster considerations. Ironically, they overlook

the obvious definition of static resilience as the complement of their

inoperability definition.

situation becomes more complicated but can be addressedby evaluating the components of static resilience anddynamic resilience of the customer separately, as well astaking into account any interactive effects.

5. Measuring resilience

To date, most of the efforts to formally measureeconomic resilience in the face of disasters pertain tobusiness interruption associated with utility lifeline disrup-tions. These analyses use as the measure of DSER thedeviation from the linear proportional relation between thepercentage utility disruption and the percentage reductionin customer output. One of the most obvious resilienceoptions for input supply interruptions in general is relianceon inventories. This has long made electricity outagesespecially problematic, since this product cannot typicallybe stored. However, the increasing severity of the problemhas inspired ingenuity, such as the use of non-interruptiblepower supplies (capacitors) in computers. Other resiliencemeasures include backup generation, conservation, inputsubstitution, and rescheduling of lost production. In manybusiness enterprises, these measures are adequate tosubstantially cushion the firm against some losses of arather short or moderate duration.In the case of TSER, both individual business and

market-related adjustments suggest some muting of generalequilibrium effects. As a starting point, it is appropriate tomeasure market, or net general equilibrium, resilience asthe deviation from the linear outcome, e.g., the multipliereffect that would be generated from a simple I–O analysisof the outage. Adjustments for lost output of goods andservices other than electricity include inventories, conser-vation, input substitution, import substitution and produc-tion rescheduling at the level of the individual firm, and therationing feature of pricing and the re-contracting amongsuppliers and customers at the level of the market.The number of empirical studies is rather sparse, because

I have limited inclusion to those studies that used customerlost output as the unit of measure and that have alsoexplicitly or implicitly included indirect (either ordinarymultiplier or general equilibrium) effects.10 Admittedly theexamples refer only to an isolated type of shock to aneconomy, but they provide some important insights intothe effectiveness of resilience.The first major attempt to measure resilience is that of

Tierney (1997), who collected responses to a surveyquestionnaire from more than a 1000 firms following theNorthridge Earthquake. Note that maximum electricityservice disruption following this event was 8.3%, and thatnearly all electricity service was restored within 24 h.Tierney’s survey results indicated that direct output losses

10Nearly all studies of power outages exclude resilience (see, e.g., Caves

et al., 1992), except for those that use a resilience response as a proxy value

of service continuity, as in the case of back-up generators (see, e.g., Bental

and Ravid, 1986).

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398392

attributable to the electricity outage amounted to only1.9% of a single day’s output in Los Angeles County asinterpreted by Rose and Lim (2002) from the Tierney data,meaning DER is 77.1%.

A study by Rose and Lim (2002) of the aftermath of theNorthridge Earthquake used a simple simulation model ofthree resilience options to estimate DSER of 95% and usedan I–O model to estimate market resilience (the differencebetween TSER and DSER) at 79.3%. Although this studydid not include the full range of resilience tactics as wasinherent in the Tierney study, it is also likely that in theTierney study the effects of production rescheduling wouldbe under-reported because not all businesses connectactivities undertaken long after the event with the affectsof the disaster. This helps explain the relatively higher levelof resilience in the analysis by Rose and Lim.

A study by Rose and Liao (2005) for a hypotheticalearthquake in Portland, Oregon, and for water rather thanelectricity utilities, incorporated engineering simulationestimates of direct output losses into a CGE model. Thefirst simulation, which represented a business-as-usualscenario, indicated a DER of 88.7%. A second simulation,representing the case of a $200 million capital expenditureinitiative to replace cast-iron pipes with modern materials,estimated DER at 88.6%. Direct resilience declinedfollowing mitigation (direct output losses as a proportionof utility outage levels increased), because mitigationreduces initial loss of service and hence ironicallynarrows the range of resilience options that can be broughtinto play.

More recently, Rose et al. (2007a, b) performed simula-tions for hypothetical terrorist attacks on the power andwater systems of Los Angeles. They simulated total supplydisruptions for the entirety of Los Angeles County for a2-week period. Their analysis incorporated an extensive setof resilience options and estimated DER at 90.6% for thecase of the power outage and 89.8% for the water outage.Market resilience was found to be almost as high.Resilience to these targeted attacks is likely to be relativelyhigher than that for natural hazards. The former arefocused on a key aspect of a community’s infrastructure inthe absence of any other devastation. On the other hand,for natural disasters and more widespread terrorist attacks(e.g., a ‘‘dirty bomb’’), other aspects of a regional economyare affected. This will reduce the ability to substituteinputs, bring in additional imports, rely on an effectivelyworking market, etc.11

Rose et al. (2007a, b), also evaluated the relativeeffectiveness of various resilience responses to water and

11It should be noted that the various studies summarized are not entirely

independent. For example, Rose and Liao used some of the Tierney survey

findings on resilience to recalibrate their production function parameters.

In addition the same production rescheduling (recapture) factors used in

the Rose and Lim study were applied to all of the other study results by

Rose and associates. It should be kept in mind, however, that these are

only a few of several considerations that influence the numerical value of

the results.

power outages. Production rescheduling was estimated tobe by far the most effective option, and this resultgeneralizes to other contexts. On the other hand, therelatively high effectiveness of ‘‘alternative sources’’ ofelectricity (e.g., back-up generators and solar panels) aremore site-specific. Many businesses and even a goodnumber of households purchased portable electricitygenerators in the aftermath of the Northridge Earthquake,and solar power is a cheap alternative source of supply in aconducive area like Los Angeles. The relatively loweffectiveness results for water storage and alternativesources of water are region specific—there is little storagein Los Angeles, no major rivers to tap, and groundwaterextraction is severely limited by law.12

Note that the above analysis is generalizable andoperational beyond the case of a utility service disruption.It can be applied to business interruption from propertydamage in general through the use of capital–output ratios(or related ‘‘functionality’’ factors in engineering).Two limitations of my definition of resilience and its

measurement should be noted. First, I have used acommon denominator of economic output to defineresilience. Although I have not done so in my ownresearch, and neither has anyone else, I have indicatedhow the standard measures of gross output, gross regionalproduct, and value added can be extended to include thevalue of un-priced or under-priced goods and services aswell. Still, the single measure, while being additive,reasonably comprehensive, and readily measurable, tendsto obscure specific elements of an economy, such as itsrelative competitiveness or equity.Second, most of the simulation studies performed on this

subject come closer to measuring potential resilience ratherthan actual. For one thing, they fail to take into accountfactors beyond the disruption of utility services. A terroristattack targeted at the electricity system will likely leavefactories and shops unscathed, but an earthquake will not,thereby making it less than automatic to rescheduleproduction (see the more extensive discussion below).Also, the existence of coping measures does not mean theywill be optimally used given the likelihood of the situationof bounded rationality and market failures. At the sametime, all analysts on the subject may have underestimatedhuman ingenuity. Overall, however, the estimates ofresilience presented above are likely biased toward thehigh side.

6. Contextual insights into resilience

6.1. Enhancing resilience

While much resilience is inherent in the economic systemand in the human spirit, and therefore a ‘‘natural’’occurrence, resilience can be enhanced by deliberate action

12Sizeable amounts of resilience have also been found in a recent survey

of Japanese firms (Kajitani and Tatano, 2007).

ARTICLE IN PRESS

Table 1

Evaluation of individual business resilience actions

Action Example Ordinary

effectiveness

Effectiveness time trend Potential

effectiveness

Effectiveness in Catastrophe

Inherent resource

substitution

Bottled water for piped water Minor Constanta Limited by cost Lowered because substitutes less

available

Adaptive resource

substitution

Drilling new water wells Minor to

moderatebIncreases w/learning Increases w/

planning

Lowered by limited substitution

options

Inherent import

substitution

Importing bottled water Minor Constanta Limited by cost Lowered if transport network

damaged

Adaptive import

substitution

Importing trucked water Moderate Increases w/re-

contracting

Increases w/

planning

Lowered if transport network

damaged

Adaptive conservation Using less water by recycling Minor to

moderatebIncreases w/learningc Increases w/

technology

Weakened by property damage

Resource inventories Using stored water Minor Decreasing Limited by

capacity

Weakened by property damage

Resource importance Portion of operation not

requiring water

Moderate to

largebConstant Increases w/

technology

Unlikely to be affected

Production

rescheduling

Making up lost production

afterward

Moderate to

immensebDecreases w/length of

disruption

Improvements

unlikely

Weakened by property damaged

aIncreases are associated with the adaptive version of this action.bDepends significantly on sector.cDraconian measures are likely to be sustainable for only short periods, however.dAlso weakened by decreased availability of other inputs and cancellation of customer orders.

A. Rose / Environmental Hazards 7 (2007) 383–398 393

(see also Bockarjova, 2007). The best but not the only timeto do so would be before the disaster strikes. Note that Iam not contradicting my earlier position that resilienceapplies to the post-disaster context, because this is the timeperiod when it is actually implemented. Examples ofresilience enhancement include the obvious increase ininventories, improving substitution possibilities for keyinputs, and broadening the supply chain. Most of theseapply to inherent resilience. Allenby and Fink (2005) havepointed out the many changes in business practices andbroader systems changes that improve disaster resiliencesecondarily.

More subtle forms of enhancement would affectadaptive resilience, such as making production processesmore flexible in general and holding emergency planningdrills that focus on business management and logisticdecisions.13 At other levels, these would include theestablishment of information clearinghouses that canimprove this decision-making, as well as compensating

13In relation to some concepts mentioned earlier, we point out another

important feature of the time dimension of disasters. Dovers and

Handmer (1992) emphasize a major distinction between natural ecosys-

tems and society—the latter’s greater ability to anticipate and learn. These

features are key to adaptive capacity. They are operable not only during

the course of a single event, but also over multiple and disparate events.

For example, the rush of companies in Los Angeles to buy back-up

electricity generators after the Northridge earthquake in 1994, and after

the rolling blackouts (caused by poorly designed deregulation) in

2000–2021, makes them less vulnerable to the possibility of a terrorist

attack on the electric power system.

for information that the market might be unable to provideduring a crisis. The continued development and experienceof the business continuity industry would overlap withboth inherent and adaptive enhancement. Of course,government at various level plays a key role in economicrecovery, so resilience can be key here as well. Improve-ments in the quality and quantity of emergency services canbe thought of as resilience enhancement. Increases infinancial or in-kind disaster assistance and the effectivenessof their distribution to the affected parties promoterecovery as well. However, the provision of aid can havedisincentive effects on resilience, just as it does formitigation in the ‘‘bail-out’’ sense.In the discussion above, there is a tradeoff between

resilience enhancement during pre-disaster and post-disaster time periods. Post-disaster initiatives have costadvantage because they involve targeting of resources whenthey are actually needed rather than probabilisticallyanticipated. One needs to include the dual use or co-benefit of both mitigation and resilience that apply toactivities unrelated to disasters in this calculation (seeAllenby and Fink, 2005). Two other tradeoffs are evenmore important. The first is one between static anddynamic aspects of resilience. Improved resource allocationin a static sense may increase economic production butmay get in the way of repair and reconstruction (and visaversa). This proper balance could be best determined by adynamic optimization model.This optimization process dovetails with the tradeoff

between resilience and mitigation. The funds for resilience

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398394

enhancement could also go toward mitigating the disasterin the first place. Again, post-disaster resilience enhance-ment has the edge. In general, resilience (pre- or post-disaster) includes many low-cost and even cost-savingoptions. At the same time, mitigation has a relativeadvantage if society requires an initial target level ofsafety, e.g., if saving lives is the priority or if there is amaximum level of economic disruption that can betolerated, even if the economy is capable of bouncing backfrom breeching it.

The final set of tradeoffs pertains to the extent to whichresilience and mitigation may undercut each other. In acase study of hypothetical major earthquake, Rose andLiao (2005) have found that economic resilience decreasedslightly when mitigation was increased. This was explainedas mitigation narrowing the set of resilience options, butthis phenomenon can stand much closer scrutiny. Manyexamples of a resilience undercutting mitigation apply topost-disaster decisions and tradeoffs. In a rush to improvestatic and dynamic resource allocation, opportunities forimproving mitigation of future disasters may be compro-mised. Thus, the ideal economic model would be one ofdynamic optimization over an extended time period thatincluded a time sequence of potential disasters.

Finally, initiatives to enhance resilience are only one sideof the coin. In the following section, we raise the issue thatresilience can be eroded by ordinary and extreme condi-tions. Thus, preventing this deterioration is of a similarnature to resilience enhancement.

(footnote continued)

period during which the business, market, or economy as a whole has not

recovered). Second, we offer no specific definition of the threshold at

6.2. Eroding resilience

Additional insight into resilience can be gained by moreclosely examining the context in which it operates and howchanges in this context affect the concept. By context, werefer to internal and external conditions affecting aphenomenon. The former includes characteristics ofbusinesses, such as size, age, inherent flexibility of pro-duction process, skills of management and workers, andlocation. Pertinent characteristics at other levels would be abusiness’s connection to the supply chain, competitivenessof its market, etc. The external context refers to thefrequency, magnitude, and duration of the external shock,interdependence of the market system, and inflow ofexternal funds (both insurance and aid).

Here I examine how resilience changes in relation totwo of the external factors: duration and severity of thedisaster. More specifically, we examine the time trend andthe effectiveness of different resilience responses and howeffectiveness at a given point in time and over a period oftime differs between an ordinary disaster and a cata-strophe.14

14Note two considerations. First, duration and magnitude are not

independent. Larger magnitude events are likely to have longer durations

(duration here is defined from an economic standpoint as not simply being

the period of ground shaking or flood waters, but rather the subsequent

Table 1 summarizes a set of individual business resilienceactions in relation to a water service disruption for the sakeof illustration. Column 1 lists the type of action, whilecolumn 2 provides a concrete example. Column 3 lists thecurrent effectiveness based on a study by Rose et al.(2007b).15 The conclusion from this study is that mosttypes of resilience reduce potential losses by only a fewpercentage points each. The major exception is productionrescheduling, which ranges from 30% to 99% in terms ofpotential loss reduction capability depending on the sector(see FEMA, 2004; Rose and Lim, 2002). ‘‘Resourceimportance’’ refers to the proportion of business operationthat can continue without water. ATC (1991) estimatesthat this ranges from 0% to 85% depending on sector.The effectiveness of the various options over time is

presented in column 4. By definition, inherent substitutionis constant, since any improvement in it is assigned to theadaptive version, which increases with learning, as well aswith availability of substitutes. The situation for importsubstitution is analogous. Adaptive responses, on the otherhand, are likely to increase with learning and managerialand market efforts, such as re-contracting. Inventories(e.g., stored water in small containers or large tanks) is themost limited option for most businesses because it is a fixedamount that is not readily continued (replenished) overtime; in fact it is characterized by depletion. Resourceimportance is likely to be rather constant except iftechnological change takes place. Ironically, the mostpotent resilience option, production rescheduling, de-creases over time, as firms reach their productive capacitylimits or lose market share permanently.Column 5 provides a summary of potential effects in the

context of ordinary disasters. Inherent capabilities arelimited by definition, though it is possible to enhance thembefore (‘‘capacity building’’). This is also the case forinventories by increasing storage capacity. Conservationand resource importance can be increased after the shockthrough improvements in technology. Production resche-duling is likely to defy improvement, e.g., it is notworthwhile to increase productive capacity to make uplost production if this additional capacity is needed onlysporadically.Catastrophes can have major effects on resilience. Their

sheer magnitude and associated duration are likely tochallenge not only individual businesses but the economyas a whole, e.g., multiple failures in the provision ofinfrastructure. They may also reduce decision-makingcapability by reducing information flows or creating stressand trauma.

which a disaster becomes a catastrophe. We simply point to clear-cut

examples that we have in mind, such as Hurricane Katrina, Indian Ocean

Tsunami, and World Trade Center attacks.15See also Rose et al. (2007a) for a counterpart assessment of electricity

service disruptions.

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398 395

Several of these factors directly or indirectly affectresilience options. In the case of inherent substitution, acatastrophe, because it is relatively more widespread, islikely to reduce the availability of substitutes. This is alsolikely to be the case for adaptive substitution. Bothinherent and adaptive import substitution are highlyvulnerable to damage to the transportation system.Adaptive conservation is weakened by property damage.Resource inventories are also likely to be weakenedby damage to structures and containers. Resource im-portance is unlikely to be affected in any other than arandom way. Production rescheduling is also weakened byproperty damage, as well as by decreased availability ofneeded inputs and cancellation of customer orders (loss ofmarket share).

Overall, the brief analysis here indicates that cata-strophes are likely to lower resilience significantly. Thiswill stem from a combination of damage to physicalaspects of the business enterprise, as well as damage to theremainder of the economy on which it is dependent.Catastrophes will also weaken decision-making ability.

One other consideration that is critical in the context ofcatastrophes is the baseline from which we measureresilience. Earlier, I used a linear damage function as thisreference point, but it is likely there are complexities andinteractions that make damages exponential in the contextof catastrophes (i.e., an X% loss of a critical input will yielda loss of output larger than X%). At the extreme there areirreversibilities or ‘‘flips’’ that can lead to a state of decay inan eco-system (see, e.g., Perrings, 2001), which areapplicable to human catastrophes as well. These variousfactors make resilience all the more important, while at thesame time posing an even greater challenge to itseffectiveness. Note, however, that a total absence of staticresilience would result in only a linear reduction ineconomic activity. The damage states exceeding the linearoutcome would appear to be related to aspects of dynamicresilience in reverse—decay versus rebuilding.

7. Policy design and implementation

Individual businesses have an incentive to implementresilience, since they capture its direct benefits. However,they are not likely to capture most of its indirect benefits.The reward structure is even more challenging in the caseof infrastructure service providers, since they produce apublic good. Moreover, given the rather novel aspects ofmany resilience practices, and the fact that that they mustbe implemented in crisis situations, resilience is even morelikely to be under-provided. Thus, there is a role forgovernment in promoting resilience for the sake of thebroader public interest.

Below I analyze three recent strategic initiatives intendedto lay the groundwork for promoting resilience to disastersin the United States, each inspired by the September 11terrorist attacks. This discussion is not intended to evaluatethe documents in their entirety but to focus on how they

define, measure, and intend to implement the concept.Actual implementation of resilience policy is still in itsinfancy, so the attention of the documents is focused onplanning rather than on any actual attempt to measureperformance. My intent is to shed some light on theseefforts from the perspective presented in this paper.The first document is Regional Disaster Resilience: A

Guide for Developing an Action Plan by the InfrastructureSecurity Partnership (TISP, 2006), a consortium of morethan 100 infrastructure-related groups, governments at alllevels, the academic community, and various organizationsinterested in disasters. Although the term ‘‘resilience’’ isemployed in the title, this report focuses primarily onmitigation and preparedness. Resilience is thus used as anall-encompassing term to include mitigation, as well asimprovements in recovery and reconstruction activities, asin Bruneau et al. (2003), perhaps not surprisingly since theAmerican Society of Civil Engineers (ASCE) was a primemover in TISP.The discussion of resilience focuses on infrastructure and

its service providers. Examples of resilience include ‘‘systemhardening, building in redundancy, implementing back-upsystems, and other mitigation measures (p. 3; emphasisadded by the author). Only a handful of considerationsaddress resilience on the customer side, the most notableone being the supply chain. Discussion of response,restoration, and reconstruction is also much less developed.The section on them emphasizes the need for a regionallevel resource management plan and the importance ofawareness to the threats. Only a few resilience strategies, ofthe many discussed in this paper, are noted, such asstockpiling of critical inputs and emergency planningexercises. No mention is made of the resilience providedby the marketplace, and the resilience of the economy as awhole is only tangentially noted. The discussion is slightlymore advanced in the area of business continuity, withmention of emergency planning exercises, supply chainmanagement, redundant systems, and off-site storage. Thereport also points to the fact that many of these optionsrepresent low hanging fruit.The second report is Homeland Security and the Private

Sector by the Congressional Budget Office (CBO, 2004). Itpoints out the potential market failures of the privatesector responses to terrorism and the potential role ofgovernment in correcting them. Three major strategies areconsidered: (1) for the private sector to internalize more ofthe social costs, (2) to assign more responsibility to thegovernment sector, and (3) to improve information tofacilitate private sector action. In the case of the latter, anunderstanding of the concept of resilience and the variousways to implement it are key.The report confines most of its attention to four sectors

deemed to be most critical and vulnerable: civilian nuclearpower, chemicals and hazardous materials, electricity, andfood and agriculture. The term resilience is rarely used, butexamples are emphasized on both the supply and demandsides. For example, in the case of electric power outages,

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398396

the report notes: ‘‘The potential losses from such disrup-tions would be limited and of relatively short durationbecause the industry and electricity customers are generallywell prepared for such failures’’ (p. xi). It also acknowl-edges the role of electricity market restructuring in aidingthe cause. Finally, it cites the ‘‘flexibility’’ of the economyin general to adapt to power losses. Still, the focus is onensuring reliable supplies and not how customers canmaintain more function without them.

The third report is the National Infrastructure Protection

Plan by the US Department of Homeland Security (DHS,2006). It presents a unifying framework for the potential ofboth critical infrastructure and key resources (CI/KR) in arisk-management context. It is intended to promoteresilience by protecting CI/KR and by strengtheningpreparedness, response and recovery from a range ofthreats. The initial use of the term in the report would seemto be the broad one in the TISP report, and its use in someother junctures is broad as well. However, the glossary andthe most prevalent uses are similar to the more focuseddefinitions of this paper: ‘‘the capacity of an asset, system,or network to maintain its function during or to cover froma terrorist attack or other incident’’ (p. 104). How thiscapability is to be measured and in relation to whatbenchmark is never explained, however.

Protection is defined broadly to include actions tomitigate risk, such as hardening facilities, building resi-liency and redundancy, and incorporating hazard resis-tance into design. The report indicates it is complementaryto response strategies, which it refers to as actions taken inthe aftermath of the event. Protection is used in the NIPPto cover both pre- and post-incident time periods. This isbecause risk is expressed in standard assessment terms of:Risk ¼ f (threat, vulnerability, consequence). Hence busi-ness continuity and resilience initiatives are represented asexamples of protection-related plans. The wording hereand in other parts of the report indicates that resilience isprimarily used to refer to infrastructure providers (supply-side) restoration abilities, as opposed to customer con-tinuity (demand-side).

Finally, the report promotes the use of three types ofquantitative indicators to measure program performance.Resources are then to be allocated to the most effectiveactivities. The metrics are in the process of beingdeveloped. Thus, it would seem that resilience will actuallyhave to be measured to perform the evaluation, but noexplicit insights are offered on how this might beaccomplished.

8. Conclusion

The objectives of this paper have been several fold. Ihave clarified the major features of economic resilience andhow it compares with closely related dimensions in ecology,engineering, organizational theory, and planning. I havealso extended an operational definition of the concept andsummarized measures of its effectiveness. Finally, I offered

insights into how resilience differs according to temporaland contextual dimensions.Several major conclusions can be drawn from the paper.

First, the definition of economic resilience can learn from,inform, and nicely complement definitions from otherdisciplines. Second, in both static and dynamic terms,resilience is best conceived as reducing the consequences ofdisaster, so as to contrast it from mitigation, which reducesthe probability that a disaster will occur. This bifurcationhelps more clearly delineate the tradeoffs between pre- andpost-disaster strategies. Third, operational definitions ofvarious aspects of the concept can be formulated. Fourth,resilience has been found to be a powerful way of reducinglosses from disasters. Fifth, resilience is likely to beseriously challenged by major catastrophes that render itless effective in these contexts. Klein et al. (2003, p. 41)concluded that decades of research have not been able ‘‘totransform the concept [of resilience] into an operationaltool for policy and management purposes.’’ This paper isintended to provide a solid foundation toward this goal inthe economic realm.

Acknowledgments

The research in this paper is supported by funding fromthe DHS Center for Risk and Economic Analysis ofTerrorist Events (CREATE) and by a grant from the NSF-sponsored Multidisciplinary Center for Earthquake En-gineering Research (MCEER). The author would like tothank Kathleen Tierney and other colleagues at MCEER,and Detlof von Winterfeldt and other colleagues atCREATE for their helpful comments on various stagesof this research. The views expressed in this paper,however, are solely those of the author and not necessarilythose of the institutions with which he is affiliated nor ofhis funding sources. Also, the author is solely responsiblefor any errors or omissions.

References

Adger, W.N., 2000. Social and ecological resilience: are they related?

Progress in Human Geography 24 (3), 247–364.

Adger, W.N., Hughes, T.P., Folke, C., Carpenter, S.R., Rockstrom, J.,

2005. Social-ecological resilience to coastal disasters. Science 309,

1036–1039.

Allenby, B., Fink, J., 2005. Toward inherently secure and resilient

societies. Science 309, 1034–1036.

Applied Technology Council (ATC), 1991. Sesimic Vulnerability and

Impacts of Disruptions of Utility Lifelines in the Cotermininous

United States, Report ATC-25. Applied Technology Council, Red-

wood, CA.

Ayres, R., Simonis, U., 1989. Industrial Metabolism. United Nations

Press, Tokyo.

Bental, B., Ravid, S.A., 1986. Simple method for evaluating the marginal

cost of unsupplied electricity. Bell Journal of Economics 13, 249–253.

Blaikie, P., Cannon, T., Davis, I., Wisner, B., 1994. At Risk: Natural

Hazards, People’s Vulnerability and Disasters. Routledge, London,

UK.

Bockarjova, M., 2007. Mafor Disasters in Modern Economics. University

of Twente, Enschede, The Netherlands.

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398 397

Bruneau, M., Chang, S., Eguchi, R., Lee, G., O’Rourke, T., Reinhorn, A.,

Shinozuka, M., Tierney, K., Wallace, W., von Winterfeldt, D., 2003.

A framework to quantitatively assess and enhance seismic resilience of

communities. Earthquake Spectra 19, 733–752.

Burby, R., Deyle, R., Godschalk, D., Olshansky, R., 2000. Creating

hazard resistant communities through land-use planning. Natural

Hazards Review 1 (2), 99–106.

Business Continuity Institute, 2002. Good practice in business continuity

management. Continuity 6 (4), 2.

Caves, D., Harriges, J., Windle, R., 1992. The cost of electric power

interruptions in the industrial sector: estimates derived from inter-

ruptible service programs. Land Economics 68, 49–61.

Chang, S., Shinozuka, M., 2004. Measuring and improving the disaster

resilience of communities. Earthquake Spectra 20, 739–755.

Chao, H.P., Wilson, R., 1987. Priority service: pricing, investment and

market organization. American Economic Review 77, 899–916.

Chernick, H. (Ed.), 2005. Resilient City. Russell Sage Foundation,

New York.

Comfort, L., 1994. Risk and resilience: inter-organizational learning

following the Northridge Earthquake of 17 January 1994. Journal of

Contingencies and Crisis Management 2 (3), 157–170.

Comfort, L., 1999. Shared Risk: Complex Seismic Response. Pergamon,

New York.

Common, M., 1995. Sustainability and Policy Limits to Economics.

Cambridge University Press, Cambridge, UK.

Congressional Budget Office (CBO), 2004. Homeland Security and

the Private Sector /http://www.cbo.gov/ftpdocs/60xx/doc6042/12-20-

HomelandSecurity.pdfS.

Corcoran, P., 2003. IBM business continuity services. Disaster Recovery

16 (4), 26–28.

Cutter, S.L., 1996. Societal responses to environmental hazards. Interna-

tional Social Science Journal 47 (4), 525–536.

Daly, H., Farley, J., 2004. Ecological Economics. Island Press, Washing-

ton, DC.

Department of Homeland Security (DHS), 2006. National Infrastructure

Protection Program /http://www.dhs.gov/xprevprot/programs/editor-

ial_0827.shtm#0S.

Dovers, R., Handmer, J., 1992. Uncertainty, sustainability and change.

Global Environmental Change 2 (4), 262–276.

Eckles, J., 2003. SunGard availabiltiy services. Disaster Recovery 16 (4),

28.

Federal Emergency Management Agency (FEMA), 2004. Earthquake

Loss Estimation Methodology (HAZUS). National Institute of

Building Sciences, Washington, DC.

Foster, H., 1997. The Ozymandias Principles: Thirty-One Strategies for

Surviving Change. UBC Press, Victoria, Canada.

Geis, D., 2000. By design: the disaster-resistant and quality-of-life

community. Natural Hazards Review 1 (3), 151–160.

Gigerenzer, G., Selten, R. (Eds.), 2002. Bounded Rationality: The

Adaptive Toolbox. MIT Press, Cambridge.

Godschalk, D., 2003. Urban hazard mitigation: creating resilient cities.

Natural Hazards Review 4 (3), 136–143.

Godschalk, D., 2004. Land use planning challenges: coping with conflicts

in visions of sustainable development and livable communities. Journal

of the American Planning Association 70 (1), 1–9.

Haimes, Y., Horowitz, B., Lambert, J., Santos, J., Lian, C., Crowther, K.,

2005a. Inoperability input–output model for interdependent infra-

structure sectors, I: theory and methodology. Journal of Infrastructure

Systems 11 (2), 67–79.

Haimes, Y., Horowitz, B., Lambert, J., Santos, J., Crowther, K., Lian, C.,

2005b. Inoperability input–output model for interdependent infra-

structure sectors, II: case studies. Journal of Infrastructure Systems 11

(2), 80–92.

Handmer, J., Dovers, S., 1996. A typology of resilience: rethinking

institutions for sustainable development. Industrial and Environmen-

tal Crisis Quarterly 9 (4), 482–511.

Hill, R., Paton, D., 2005. Managing Company Risk and Resilience

Through Business Continuity Management.

Holling, C., 1973. Resilience and stability of ecological systems. Annual

Review of Ecology and Systematics 4, 1–23.

Kajatani, Y., Tatano, H., 2007. Estimation of lifetime resilience factors

based on empirical surveys Japanese industries. Earthquake spectra,

forthcoming.

Klein, R., Nicholls, R., Thomalla, F., 2003. Resilience to natural hazards:

how useful is this concept? Environmental Hazards 5, 35–45.

Lave, L., Apt, J., Morgan, G., 2007. A worst case electricity scenario: the

benefits and costs of prevention. In: Richardson, H., Gordon, P.,

Moore, J. (Eds.), The Economic Cost and Consequences of a Terrorist

Attack. Edward Elgar Publishing Company, Cheltenham, UK.

Mileti, D., 1999. Disasters by Design: A Reassessment of Natural Hazards

in the United States. Joseph Henry Press, Washington, DC.

Olshansky, R., Kartez, J., 1998. Managing land use to build resilience. In:

Burby, R. (Ed.), Cooperating with Nature: Confronting Natural

Hazards with Land-Use Planning for Sustainable Communities.

Joseph Henry Press, Washington, DC.

Paton, D., Johnston, D., 2001. Disasters and communities: vulnerability,

resilience and preparedness. Disaster Prevention and Management 10,

270–277.

Pelling, M., 2003. The Vulnerability of Cities: Natural Disasters and

Social Resilience. Earthscan, London, UK.

Perrings, C., 2001. Resilience and sustainability. In: Folmer, H., Gabel,

H.L., Gerking, S., Rose, A. (Eds.), Frontiers of Environmental

Economics. Edward Elgar, Cheltenham, UK, pp. 319–341.

Petak, W., 2002. Earthquake resilience through mitigation: a system

approach. Paper Presented at the International Institute for Applied

Systems Analysis, Laxenburg, Austria, July, 2002.

Resilience Alliance, 2005. Research on social–ecological systems: A basis

for sustainability /http://www.resilience.orgS.

Rose, A., 2004a. Economic principles, issues, and research priorities of

natural hazard loss estimation. In: Okuyama, Y., Chang, S. (Eds.),

Modeling of Spatial Economic Impacts of Natural Hazards. Springer,

Heidelberg, pp. 13–36.

Rose, A., 2004b. Defining and measuring economic resilience to disasters.

Disaster Prevention and Management 13, 307–314.

Rose, A., 2005. Analyzing terrorist threats to the economy: a computable

general equilibrium approach. In: Richardson, H., Gordon, P., Moore,

J. (Eds.), Economic Impacts of Terrorist Attacks. Edward Elgar,

Cheltenham, UK, pp. 196–217.

Rose, A., 2006. Macroeconomic impacts of catastrophic events: the

influence of resilience. In: Quigley, J., Rosenthal, L. (Eds.), Real

Estate, Catastrophic Risk, and Public Policy. University of California

Press, Berkeley, CA.

Rose, A., Benavides, J., 1999. Optimal allocation of electricity after major

earthquakes: market mechanisma versus rationing. In: Lawrence, K.,

et al. (Eds.), Advances in Mathematical Programming and Financial

Planning. JAI Press, Greenwich, CT, pp. 147–168.

Rose, A., Kverndokk, S., 1999. Equity and environmental policy: with

special application to global warming. In: van den Bergh, J. (Ed.),

Handbook of Environmental Economics. Edward Elgar Publishing

Co., Cheltenham, UK.

Rose, A., Liao, S., 2005. Modeling resilience to disasters: computable

general equilibrium analysis of a water service disruption. Journal of

Regional Science 45 (1), 75–112.

Rose, A., Lim, D., 2002. Business interruption losses from natural

hazards: conceptual and methodology issues in the case of the

Northridge Earthquake. Environmental Hazards: Human and Social

Dimensions 4, 1–14.

Rose, A., Stevens, B., Davis, C., 1988. Natural Resource Policy and

Income Distribution. Johns Hopkins University Press, Baltimore.

Rose, A., Oladosu, G., Liao, S., 2007a. Business interruption impacts of a

terrorist attack on the electric power system of Los Angeles: customer

resilience to a total blackout. Risk Analysis 27 (3), 513–531.

Rose, A., Oladosu, G., Liao, S., 2007b. Regional economic impacts of a

terrorist attack on the water system of Los Angeles: a computable

general disequilibrium analysis. In: Richardson, H., Gordon, P.,

Moore, J. (Eds.), The Economic Costs and Consequences of a

ARTICLE IN PRESSA. Rose / Environmental Hazards 7 (2007) 383–398398

Terrorist Attack. Edward Elgar Publishing Company, Cheltenham,

UK.

Rose, A., Porter, K., Dash, N., et al., 2007c. Benefit-cost analysis of

FEMA hazard mitigation grants. Natural Hazards Review 8 (4),

97–111.

Salerno, C., 2003. Powered up when the lights go out, continuity

insights: strategies to assure integrity. Availability and Security 1 (6),

23–28.

Schuler, R.E., 2005. Two-sided electricity markets: self-healing

systems. Paper Presented at the Second Annual CREATE Sym-

posium on the Economic of Terrorism. USC, Los Angeles, CA,

August, 2005.

Shaw, G., Harrald, J., 2004. Identification of the core competencies

required of executive level business crisis and continuity management.

Journal of Homeland Security and Emergency Management 1 (1),

1–13.

Sheffi, Y., 2005. The Resilient Enterprise. MIT Press, Cambridge, MA.

The Infrastructure Security Partnership (TISP), 2006. Regional Disaster

Resilience: A Guide for Developing an Action Plan. American Society

of Civil Engineers. /http://www.tisp.org/rdr_guideS.

Tierney, K., 1997. Impacts of recent disasters on businesses: the 1993

midwest floods and the 1994 Northridge Earthquake. In: Jones, B.

(Ed.), Economic Consequences of Earthquakes: Preparing for the

Unexpected. National Center for Earthquake Engineering Research,

Buffalo, NY, pp. 189–222.

Timmerman, P., 1981. Vulnerability, Resilience and the Collapse of

Society: A Review of Models and Possible Climatic Applications.

Institute for Environmental Studies, University of Toronto, Canada.

Tobin, G., 1999. Sustainability and community resilience: the holy grail of

hazards planning? Environmental Hazards 1, 13–25.

UN/ISDR, 2002. Living with Risk: A Global Review of Disaster

Reduction Initiatives by the ISDR Geneva, Switzerland.

Vale, L., Campanella, T., 2005. The Resilient City: How Modern Cities

Recover from Disaster. Oxford, New York.