HE CONOMICS OF ATURAL DISASTERS S H · 2020-04-17 · ments of the total disaster cost to the population, i.e. an estimate of welfare losses. Moreover, disaster impacts can have positive

CESifo Forum 2/2010 14

Focus

THE ECONOMICS OF

NATURAL DISASTERS

STÉPHANE HALLEGATTE* AND

VALENTIN PRZYLUSKI**

Introduction

Large-scale disasters regularly affect societies overthe globe, causing huge destructions and damages.The 2010 earthquake in Port-au-Prince and the hur-ricane Katrina in 2005 have shown that poor as wellas rich countries are vulnerable to these events,which have long lasting consequences on welfare,and on human and economic development.

After each of these events, media, insurance com-panies and international institutions publishnumerous assessments of the ‘cost of the disaster’.However these various assessments are based ondifferent methodologies and approaches, and theyoften reach quite different results. Beside technicalproblems, these discrepancies are due to the multi-dimensionality in disaster impacts and their largeredistributive effects, which make it unclear what isincluded or not in disaster cost assessments. Butmost importantly, the purpose of these assessmentsis rarely specified, even though different purposescorrespond to different perimeter of analysis anddifferent definitions of the ‘cost’.

This confusion translates into the multiplicity ofwords to characterize the cost of a disaster in pub-lished assessments: direct losses, asset losses, indirectlosses, output losses, intangible losses, market andnon-market losses, welfare losses, or any combinationof those. It also makes it almost impossible to com-pare or aggregate published estimates that are basedon so many different assumptions and methods.

To clarify the situation, this article proposes a defin-

ition of the cost of a disaster, and emphasizes the

most important mechanisms that explain this cost. It

does so by first explaining why the direct economic

cost, i.e. the value of what has been damaged or

destroyed by the disaster, is not a good indicator of

disaster seriousness and why estimating indirect loss-

es is crucial. Then, it describes the main indirect con-

sequences of a disaster and of the following recon-

struction phase, and discusses the methodologies to

measure them. Finally, it proposes a review of pub-

lished assessments of indirect economic conse-

quences, which confirm their importance and the

need to take them into account.

What is a disaster? How to define its cost? For which purpose?

There is no single definition of a disaster. From an

economic perspective, however, a natural disaster

can be defined as a natural event that causes a per-

turbation to the functioning of the economic sys-

tem, with a significant negative impact on assets,

production factors, output, employment, or con-

sumption. Examples of such natural event are

earthquakes, storms, hurricanes, intense precipita-

tions, droughts, heat waves, cold spells, and thun-

derstorms and lightning.

Disasters affect the economic system in multiple

ways and defining the cost of a disaster is tricky.

Pelling et al. (2002), Lindell and Prater (2003),

Cochrane (2004), Rose (2004), among others, dis-

cuss typologies of disaster impacts. These typolo-

gies usually distinguish between direct and indirect

losses.

Direct losses are the immediate consequences of the

disaster physical phenomenon: the consequence of

high winds, of water inundation, or of ground shak-

ing. Direct losses are often classified into direct mar-

ket losses and direct non-market losses (also some-

times referred to as intangible losses, even though

non-market losses are not necessarily intangible).

Market losses are losses to goods and services that

* Centre International de Recherche sur l’Environnement et leDéveloppement, Paris and Ecole Nationale de la Météorologie,Météo-France, Toulouse.** Ecole Nationale de la Météorologie, Météo-France, Toulouse.This research is supported by the European Community's SeventhFramework Programme (FP7/2007-2013), ConHaz Project, ContractNo. 244159.

CESifo Forum 2/201015

Focus

are traded on markets, and for which a price can eas-ily be observed. Even though droughts or heat wavesaffect directly the economic output (especially in theagriculture sector), direct market losses from mostdisasters (earthquakes, floods, etc.) are losses ofassets, i.e. damages to the built environment andmanufactured goods. These losses can be estimatedas the repairing or replacement cost of the destroyedor damaged assets. Since building and manufacturedgoods can be bought on existing markets, their priceis known. Direct market losses can thus be estimatedusing observed prices and inventories of physicallosses that can be observed (as recorded, e.g. in theEM-DAT database or insurance-industry databases)or modelled (using, e.g. catastrophe models of theinsurance industry).

Non-market direct losses include all damages thatcannot be repaired or replaced through purchases ona market. For them, there is no easily observed pricethat can be used to estimate losses. This is the case,among others, for health impacts, loss of lives, natur-al asset damages and ecosystem losses, and damagesto historical and cultural assets. Sometimes, a pricefor non-market impacts can be built using indirectmethods, but these estimates are rarely consensual(e.g. the statistical value of human life always leadsto heated controversies).

Indirect losses (also labelled ‘higher-order losses’ inRose (2004)) include all losses that are not pro-voked by the disaster itself, but by its consequences.Often, the term ‘indirect losses’ is used as a proxyfor ‘output losses’, i.e. the reduction in economicproduction provoked by the disaster. Output lossesinclude the cost of business interruption caused bydisruptions of water or electricity supplies, andlonger-term consequences of infrastructure and cap-ital damages. Indirect losses can also have ‘negative-cost’ components, i.e. gains from additional activitycreated by the reconstruction. Sometimes, non-mon-etary indirect consequences of disasters are alsoincluded, like the impact on poverty or inequalities,the reduction in collected taxes, or the increase innational debt.

Another difficulty in disaster cost assessment lies inthe definition of the baseline scenario. The cost ofthe disaster has indeed to be calculated by compar-ing the actual trajectory (with disaster impacts) witha counterfactual baseline trajectory (i.e. a scenario ofwhat would have occurred in absence of disasters).This baseline is not easy to define and several base-

lines are often possible. Moreover, in cases whererecovery and reconstruction does not lead to areturn to the baseline scenario, there are permanent(positive or negative) disaster effects that are diffi-cult to compare with a non-disaster scenario.

For instance, a disaster can lead to a permanentextinction of vulnerable economic activities in aregion, because these activities are already threat-ened and cannot recover, or because they can moveto less risky locations. In that case, the disaster is nota temporary event, but a permanent negative shockfor a region and it is more difficult to define the dis-aster cost. Also, reconstruction can be used to devel-op new economic sectors, with larger productivity,and lead to a final situation that can be consideredmore desirable than the baseline scenario. Thisimprovement can be seen as a benefit of the disaster.It is however difficult to attribute unambiguouslythis benefit to the disaster, because the same eco-nomic shift would have been possible in absence ofdisaster, making it possible to get the benefits with-out suffering from the disaster-related human andwelfare losses.

Most importantly, defining the cost of a disaster can-not be done independently of the purpose of theassessment. Different economic agents, indeed, areinterested in different types of cost. Insurers, forinstance, are mainly interested in consequences thatcan be insured. Practically, this encompasses mainlythe cost of damages to insurable assets (e.g. damagedhouses and factories), and short-term business inter-ruption caused by the disaster (e.g. the impossibilityto produce until electricity is restored).

For affected households, insurable assets are also amajor component, but other cost categories are asimportant. Primarily, loss of lives, health impacts andperturbation to their daily life are crucial. But inaddition, households are concerned about theirassets but also about their income, which can bereduced by business interruption or by loss of jobs,and about their ability to consume, i.e. the availabil-ity of goods and services.

At the society scale, all these aspects are important,but local authorities, governments and internationalinstitutions are also interested in other points. First,to manage the recovery and reconstruction periodand to scale the necessary amount of internationalaid, they need information on the aggregated impacton economic production, on unemployment and


Focus

jobs, on the impact of inequality and poverty, onlocal-businesses market-shares, on commercial bal-ance, on collected taxes, etc. Second, to assesswhether investment in prevention measures aredesirable, they need the broadest possible assess-ments of the total disaster cost to the population, i.e.an estimate of welfare losses.

Moreover, disaster impacts can have positive or neg-ative ripple-effects at the global scale, as shown byhurricane Katrina, which led to a significant rise inworld oil prices. Depending on the purpose and ofthe decision-making spatial scale, the perimeter ofthe cost analysis will be different. For instance, acountry may want to assess the losses in the affectedregion, disregarding all out-of-the-region impacts, tocalibrate the financial support it wants to provide tothe victims. But it may also want to assess total loss-es on its territory, including gains and losses outsidethe affected region, for example to assess the impacton its public finance.

Clearly, depending on the purpose of the assessment,some of the cost components have to be included ornot in the analysis. In the following, we focus on theeconomic cost for the affected region, with the aimof informing decision-makers on post-disaster finan-cial aid and prevention measure desirability. To doso, it is obvious that the direct cost is an insufficientmeasure, and that the loss of welfare is much morerelevant. Assessing a loss of welfare is complicated,as it includes many economic and non-economiccomponents. Here, we focus on the economic com-ponent of welfare losses, and we define the econom-ic cost of disaster as the lost consumption, consid-ered as an important component and a good proxyof economic-related welfare losses.1 Of course, thisarticle does not try to be comprehensive, and majorcost components are left out of the analysis, like lossof lives, health consequences and loss of jobs. Theseadditional component are important for the popula-tion welfare and therefore for prevention measureassessments.

Consumption losses and output losses

This section explains how to assess consumptionlosses from asset and output losses. More precisely, itexplains why the sum of asset losses and output loss-es is a good proxy for the loss of consumption. To do

so, Figure 1 (a), (b) and (c) show simplified repre-sentations of a post-disaster situation. Figure 1(a)depicts the situation in which only output losses areestimated: in this situation the disaster leads to atemporary reduction in output during the recon-struction phase. We assume here that reconstructionis a return to the baseline scenario (i.e. a no-disastercounterfactual scenario). As already stated, this isnot always the case, but making an assumption onthe final state is necessary to define the cost of thedisaster, and the assumption of a return to the non-disaster baseline scenario is likely to be the mostneutral one for this type of assessment.

The sum of instantaneous output loss is what is oftenreferred to as the indirect loss. But reconstructionneeds in the disaster aftermath means that a signifi-cant share of the remaining production will have tobe devoted to reconstruction, as shown inFigure 1(b). In other terms, the resources used torebuild damaged houses cannot be used to build newhouses, or to maintain existing ones. This reconstruc-tion output is included in total output, and is not aloss of output. But it is a ‘forced’ investment, in addi-tion to the normal-time investment – consumptiontrade-off. It causes, therefore, a loss of welfare. Thevalue of this forced investment is the replacementvalue of damaged asset, which is referred to as thedirect losses. This is what is represented inFigure 1(c): the sum of the output loss and of thereconstruction output is the amount that cannot beused for consumption and non-reconstructioninvestment. This is referred to here as ‘total losses’.

In this framework, total costs are the sum of the indi-rect cost (i.e. the reduction of the total value addedby the economy due to the disaster) and the directcost (i.e. the portion of the remaining value addedthat has to be dedicated to reconstruction instead ofnormal consumption). Capital and output losses cantherefore be simply added to estimate consumptionlosses.

Of course, Figure 1 shows a simplified situation inwhich production has no flexibility. In this case,reconstruction needs cannot be satisfied throughincreased production and it has to crowd out otherconsumptions and investments. Figure 2 depicts adifferent case, in which there is a limited flexibility inthe production process: capital destruction leads to areduction in output; but unaffected capital canincrease its own production to compensate thisreduction, for instance through an increase in work

1 In an utilitarist framework, what matters is not output and pro-duction, but consumption.


Focus

hours by workers at unaffectedfactories and businesses. In prac-tice, there are gross indirect loss-es, and gross indirect gains (dueto the stimulus effect of thereconstruction). But there is stilla fraction of the remaining pro-duction that is used for recon-struction instead of normal con-sumption, even though this shareis smaller than in absence of pro-duction flexibility.

In this situation also, the con-sumption loss is still the sum ofdirect (asset) and indirect (out-put) losses (see Figure 2(c)),making it necessary to estimateoutput losses. But output lossesare not only the lost productionfrom the affected capital, butalso the output gains and lossesfrom unaffected capital in therest of the economy. It makes theassessment of output lossesmore complicated, since it de-pends on complex economicmechanisms and trade-offs.

In practice, moreover, the reduc-tion in consumption can be miti-gated or amplified by (i) changesin prices; (ii) flexibility in theproduction process; (iii) changesin the saving-consumption trade-off for the remaining production;and (iv) the fact that the rebuiltcapital will be more recent thanbefore the disaster, with poten-tial benefits (Hallegatte andDumas 2009). The following sec-tion will describe methodologiesto assess output losses and high-light the most important process-es to take into account.

How to assess output losses?

If output losses represent animportant component of totallosses, it becomes essential todevelop methodologies to assessthem. To do so, we propose to

Figure 1DIRECT LOSSES, INDIRECT LOSSES AND TOTAL LOSSES WHEN THERE IS

NO FLEXIBILITY IN THE PRODUCTION PROCESS

a)

b)

c)

Source: Authors’ conception.


Focus

start by assessing the lost outputfrom the directly affected capi-tal. In the second subsection, weinvestigate the systemic impactsof disasters, including the effecton the capital that is not directlyaffected by the disaster.

From asset losses to output losses

The first step in an assessmentof output losses is to estimatehow much output is lost becauseof direct asset losses. Economictheory states that, at the eco-nomic equilibrium and undercertain conditions, the value ofan asset is its expected futureproduction, and this equality hasbeen widely used to assess disas-ter output losses. Assuming thisequality is always verified, theoutput loss caused by capitalloss is simply equal to the valueof the damages, capital lossesand output losses are simplyequal, and the sum of asset andoutput losses is the double ofasset losses.

The assumption that output loss-es are equal to capital losses ishowever based on strong as-sumptions, which are not alwaysverified. In estimates of disasterconsequences, ‘asset loss’ is thereplacement value of the capital.To have the equality of asset lossand output loss, a double equali-ty needs to be verified: replace-ment value has to be equal to market value; and marketvalue has to be equal to the net present value of expectedoutput. In an optimal economy atequilibrium, these two equalitiesare valid: first, the market valueof an asset is by definition the net value of its output; second,if market value were higher(lower) than replacement value,then investors would increase(decrease) the amount of capitalto restore the equilibrium.

Figure 2DIRECT LOSSES, INDIRECT LOSSES AND TOTAL LOSSES WHEN THERE IS

A LIMITED FLEXIBILITY IN THE PRODUCTION PROCESS

a)

b)

c)



Focus

Therefore, in theory, there is no difference betweencapital losses and the reduction in output from thiscapital. But the assumption of the economics beingoptimal and at equilibrium is questionable. First, forthe replacement value and the market value to beequal, the economy needs to be at its optimum, i.e.the amount of capital is such that its return is equalto the (unique) interest rate. This is unlikely for thecapital that is affected by natural disasters, especial-ly as infrastructure and public assets are heavilyaffected. Since these assets are not exchanged onmarkets, they have no market prices. Moreover, theyare not financed by private investors, but decidedabout through a political process under the con-sideration of multiple criteria (e.g. land-use planningobjectives), and there is no reason for their purely-financial return to be equal to the (private) interestrate. Practically, some assets may have an outputvalue lower than their replacement value (e.g. a sec-ondary road that is redundant and does not providea significant gain of time or distance), while somemay have an output value much larger than theirreplacement value (e.g. a bridge that cannot beclosed without large consequences for users).

Second, for market values to be equal to net presentvalue of expected output, expectations have to beunbiased and markets need to be perfect. This is notalways the case especially in sectors affected by dis-asters, where expectations can be heavily biased (e.g.in housing market).

Also, output losses are most of the time estimatedfrom a social point of view. The equality betweenmarket value (for the owner) and expected output(for the society) is valid only in absence of external-ities. Some assets that are destroyed by disasters mayexhibit positive externality. It means that their valueto the society is larger than the value of the owner’sexpected output. Public goods have this characteris-tic, among which most infrastructures.

An example is provided by the San FranciscoOakland Bay Bridge, which is essential to the eco-nomic activity in San Francisco and had to be closedfor one month after the Loma Prieta earthquake in1989. Its replacement value has no reason to beequal to the loss in activity caused by the bridge clo-sure, because the bridge production is not sold on amarket, the bridge has no market value, and thesocial return on capital of the bridge is unlikely toexhibit decreasing returns and is likely to be muchhigher than the interest rate.

Another example is the health care system in NewOrleans. Beyond the immediate economic value ofthe service it provided, a functioning health care isnecessary for a region to attract workers. AfterKatrina landfall on the city, the absence of healthcare services made it more difficult to reconstruct,and the cost for the region was much larger than theeconomic direct value of this service.

The systemic impact of natural disasters

The equality between output losses and asset lossesis questionable for any economic shock, small orlarge. The most important issues appear when con-sidering very large shocks, or systemic events, whichare the events that perturb the functioning of theentire economic system and affect relative prices. Inthis case, output losses may be damped or amplifiedby several mechanisms.

(a) Changes in prices

Figures 1 and 2 show output in real terms, i.e. withno monetary effects. But output losses can be esti-mated assuming unchanged (pre-disaster) prices orcon-sidering the impact of the disaster on prices.Both assumptions lead to the same result if the dis-aster has only a marginal impact on the economy,with little impact of prices, but can be very differentin the opposite situation. In other terms, one canassume that if a house is destroyed, the family whoowns the house will just have to rent another houseat the pre-disaster price. But this assumption isunrealistic if the disaster causes more than a mar-ginal shock. In post-disaster situations, indeed, a sig-nificant fraction of houses may be destroyed, lead-ing to changes in the relative price structure. In thiscase, the price of alternative housing can be muchhigher than the pre-di-saster price, as a consequenceof the disaster-related scarcity in the housing mar-ket. Estimating the value of lost housing serviceshould then be done using this higher cost instead ofthe pre-disaster one, which can lead to a significantincrease in the assessed disaster cost. Unfortunately,it is difficult to predict ex ante the change in pricesthat would be caused by a disaster, making lossassessment more complicated.

The same reasoning is possible in all other sectors,including transportation, energy, water, health, etc.In extreme cases, reconstruction may even be impos-sible, at all prices. This is because markets are not atequilibrium in disaster aftermath (i.e. price is not


Focus

such that demand equals production). The ‘if I canpay it, I can get it’ assumption is not valid in post-dis-aster situations. In these situations, therefore, thevalue of lost production cannot always be estimatedas the product of lost produced quantity and pre-dis-aster prices. Providing an unbiased estimate requiresan assessment of the disaster impact on prices.

Often considered as resulting from unethical behav-ior from businesses, which are thought to benefitfrom the disaster, post-disaster price inflation canalso have positive consequences. This inflation,indeed, helps attract qualified workers where theyare most needed and creates an incentive for allworkers to work longer hours, therefore compensat-ing for damaged assets and accelerating reconstruc-tion. It is likely, for instance, that higher prices afterhurricane landfalls are useful to make roofers fromneighboring unaffected regions move to the landfallregion, therefore increasing the local productioncapacity and reducing the reconstruction duration.Demand surge, as a consequence, may also reducethe total economic cost of a disaster, even though itincreases its burden on house owners.

(b) Length of the reconstruction phase

Importantly, there is a large difference between los-ing a home for one day (in this case the total loss isthe reconstruction value, i.e. the direct loss) and los-ing a home for one year (in this case the total loss isthe reconstruction value, i.e. the direct loss, plus thevalue of one year of housing services, i.e. the outputloss). Of course, the longer the reconstruction peri-od, the larger the total cost of the disaster.

The reconstruction phase, and the economic recov-ery pace, will ultimately determine the final cost ofthe natural disasters. The reconstruction pace islinked to the constraints to the reconstructionphase, which are of two types. First, they can befinancial. This concerns situations in which house-holds and businesses can simply not finance thereconstruction. This is of particular importance incountries with limited resources (Freeman et al.

2002; Mechler et al. 2006).

Constraints are also technical. Technical limits to theability to increase production are obvious in the con-struction sector, which experience a dramatic in-crease in demand after the disaster. In spite of thisdemand, production does not follow, because thereare strong constraints on reconstruction. Many

households are able to pay for reconstruction, butcannot find workers and contractors to carry out thework. The same is true for businesses and factories.This explains why reconstruction often takes severalyears, even for limited damages (e.g. the 2004 hurri-cane season in Florida – see McCarty and Smith2005). Examples of constraint include the availabili-ty of equipment and qualified workers. For instance,the limited availability of glaziers increased the costof reconstruction and lowered the reconstructionpace after the 2001 chemical explosion in Toulouse(France), despite glaziers coming from all the coun-try to carry out the work.

(c) Output gains and losses from the non-affectedcapital

Damages in crucial intermediate sectors may lead tonegative ‘network effects’ in the economy, leading toproduction losses even for businesses that are notdirectly affected by the disaster. Water, electricity,gas and transportation are the most critical sectors,and their production interruption has immediateconsequences on the entire economic system. In pastcases, it has been shown that the indirect conse-quences of utility services had larger consequencesthan direct asset losses, both on households(McCarty and Smith 2005) and on business (Tierney1997). Of course, when capital cannot producebecause of a lack of input (e.g. electricity, water),input substitution, production rescheduling, andlonger work hours can compensate for a significantfraction of the losses (see Rose et al. 2007). Thesemechanisms can damp the output losses, and canespecially reduce the crowding-out effects of recon-struction on normal consumption and investment(see Figure 2). But their ability to do so is limited,especially when losses are large.

There are many sources of flexibility in the econom-ic system. First, production capacity is not fully usedin normal times and idle production capacity can bemobilized in disaster aftermath to compensate forlost production from lost assets. Second, behaviorscan change in disaster aftermath and workers canincrease their work hours in unaffected businesses tohelp society cope with disaster consequences (andsometimes benefit from increased prices). As a con-sequence, unaffected capital can often increase pro-duction to compensate for output loss from affectedcapital. After mild disasters, net output gains caneven be observed, explained by the non-zero priceelasticity of production, and by the under-optimality


Focus

of the pre-disaster situation that leaves some roomfor increased production. In an economy that fullyuses all resources and cannot increase its productionover the short-term (whatever the price level), sucha gain would be impossible. In a more realistic econ-omy that does not use efficiently all resources (withunderemployment, and imperfect allocation of capi-tal), additional demand does not lead only to infla-tion, but also to increased output.

The ‘adaptability’ and ‘flexibility’ of the productionsystem and its ability to compensate for unavailableinputs is largely unknown and largely depend on theconsidered timescale. Over the very short term, theproduction system is largely fixed and the lack ofone input can make it impossible to produce.Moreover, over short timescales, local productioncapacity is likely to be highly constrained by existingcapacities, equipments and infrastructure. Onlyimports from outside the affected region and post-ponement of some non-urgent tasks (e.g. mainte-nance) can create a limited flexibility over the short-term. This is what is represented in economic Input-Output model (e.g. Okuyama 2004), in which pro-ducing one unit of output requires a fixed amount ofall input categories.

Over the longer term and the entire reconstructionperiod which can stretch over years for large-scaleevents, the flexibility is much higher: relative priceschange, incentivizing production in scarce sectors;equipments and qualified workers move into theaffected region, accelerating reconstruction andreplacing lost capacities; and different technologiesand production strategies can be implemented tocope with long-lasting scarcities. The production sys-tem organization can also be adjusted to the new sit-uation: one supplier that cannot produce or cannotdeliver its production (because of transportationissues, for instance) can be replaced by another sup-pliers; new clients can be found to replace bankruptones; slightly different processes can be introducedto reduce the need for scarce inputs (e.g. oil-runningbackup generator can be installed if electricity avail-ability remains problematic). These types of substi-tution are represented in Calculable GeneralEquilibrium models (e.g. Rose et al. 2007), in whichthe scarcity of one input translates into higher price,and reduced consumption of this input, compensatedby larger consumption of other inputs.

IO models are often considered too pessimistic,since they assume that prices are fixed and that no

substitution can take place in the production sys-tem. On the opposite CGE models are consideredas too optimistic, since they assume that marketsfunction perfectly (even in post-disaster situa-tions), and that optimal prices balance productionand demand and, act as signals to incentivize pro-duction of the most needed goods and services. Thereality probably lies somewhere in between thesetwo extremes, prompting the work on intermediatemodels. These intermediate models are either IOmodels with flexibility like those in Hallegatte(2008) or CGE models with reduced substitutionelasticity like those in Rose et al. (2007).

(d) The stimulus effect of disasters

Disasters lead to a reduction of production capaci-ty, but also to an increase in the demand for thereconstruction sector and goods. Thus, the recon-struction acts in theory as a stimulus. However, asany stimulus, its consequences depend on the pre-existing economic situation, or the phase of thebusiness cycles. If the economy is in a phase of highgrowth, in which all resources are fully used, thenet effect of a stimulus on the economy will be neg-ative, for instance through diverted resources, pro-duction capacity scarcity and accelerated inflation.If the pre-disaster economy is depressed, on theother hand, the stimulus effect can yield benefits tothe economy by mobilizing idle capacities. Thiscomplex interplay between business cycles andnatural disasters economics is analyzed in detail inHallegatte and Ghil (2008), who support thecounter-intuitive result that economies in recessionare more resilient to the effect of natural disasters.This result appears to be consistent with empiricalevidence. For instance, the 1999 earthquake inTurkey caused destructions amounting to 1.5 to 3 percent of Turkey’s GDP, but consequences ongrowth remained limited, probably because theeconomy had significant unused resources at thattime (the Turkish GDP contracted by 7 percent inthe year preceding the earthquake). In this case,therefore, the earthquake may have acted as astimulus, and have increased economic activities inspite of its terrible human consequences. In 1992also, when hurricane Andrew made landfall onsouthern Florida, the economy was depressed andonly 50 percent of the construction workers wereemployed (West and Lenze 1994). The reconstruc-tion needs had stimulus effects on the constructionsectors, which would have been impossible in a bet-ter economic situation.


Focus

(e) The productivity effect

When a disaster occurs, it has been suggested thatdestructions can foster a more rapid turn-over of cap-ital, which could yield positive outcomes through themore rapid embodiment of new technologies. Thiseffect, hereafter referred to as the ‘productivityeffect’, has been mentioned, for instance, by Albala-Bertrand (1993), Stewart and Fitzgerald (2001),Okuyama (2004), and Benson and Clay (2004).Indeed, when a natural disaster damages productivecapital (e.g. production plants, houses, bridges), thedestroyed capital can be replaced using the mostrecent technologies, which have higher productivities.Examples of such upgrading of capital are: (i) forhouseholds, the reconstruction of houses with betterinsulation technologies and better heating systems,allowing for energy conservation and savings; (ii) forcompanies, the replacement of old production tech-nologies by new ones, like the replacement of paper-based management files by computer-based systems;and (iii) for government and public agencies, theadaptation of public infrastructure to new needs, likethe reconstruction of larger or smaller schools whendemographic evolutions justify it. Capital losses can,therefore, be compensated by a higher productivityof the economy in the event aftermath, with associat-ed welfare benefits that could compensate for the dis-aster direct consequences. This process, if present,could increase the pace of technical change andaccelerate economic growth, and could therefore rep-resent a positive consequence of disasters.

As an empirical support for this idea, Albala-Bertrand (1993) examined the consequences of 28natural disasters on 26 countries between 1960 and1979 and found that, in most cases, GDP growthincreases after a disaster and he attributed thisobservation, at least partly, to the replacement of thedestroyed capital by more efficient one.

However, the productivity effect is probably notfully effective, for several reasons. First, when a dis-aster occurs, producers have to restore their produc-tion as soon as possible. This is especially true forsmall businesses, which cannot afford long produc-tion interruptions (see Kroll et al. 1991; Tierney1997), and in poor countries where people have nomean of subsistence while production is interrupted.Replacing the destroyed capital by the most recenttype of capital implies in most cases to adapt compa-ny organization and worker training which takestime. Producers have thus a strong incentive to

replace the destroyed capital by the same capital, inorder to restore production as quickly as possible,even at the price of a lower productivity. In extremecase, one may even imagine that reconstruction iscarried out with lower productivity, to make recon-struction as fast as possible, with a negative impacton total productivity. Second, even when destruc-tions are quite extensive, they are never complete.Some part of the capital can, in most cases, still beused, or repaired at lower costs than replacementcost. In such a situation, it is possible to save a partof the capital if, and only if, the production system isreconstructed identical to what it was before the dis-aster. This technological ‘inheritance’ acts as a majorconstraint to prevent a reconstruction based on themost recent technologies and needs, especially in theinfrastructure sector.

This effect is investigated in Hallegatte and Dumas(2008) using a model with embodied technicalchange. In this framework, disasters are found toinfluence the production level but cannot influencethe economic growth rate, in the same way as thesaving ratio in a Solow-like model. Depending onhow reconstruction is carried out (with more or lessimprovement in technologies and capital), indeed,accounting for the productivity effect can eitherdecrease or increase disaster costs, but is never ableto turn disasters into positive events.

(f) Poverty traps

It is crucial to also take into account the possibilitythat natural disasters increase poverty. In particular,because they destroy assets and wipe out savings, theycan throw households into ‘poverty traps’, i.e. situa-tion in which their productivity is reduced, making itimpossible for them to rebuild their savings andassets.These micro-level poverty traps can also be cre-ated by health and social impacts of natural disasters:it has been shown that disasters can have long-lastingconsequences on psychological health (Norris 2005),and on children development (from reducing inschooling and diminished cognitive abilities – see, forinstance, Santos (2007); Alderman et al. (2006)).

These poverty traps at the micro-level (i.e. thehousehold level) could even lead to macro-levelpoverty traps, in which entire regions could be stuck.Such poverty traps could be explained by the ampli-fying effect reproduced in Figure 3. Poor regionshave a limited capacity to rebuild after disasters: ifthey are regularly affected by disasters, they do not


Focus

have enough time to rebuild between two events,and they end up into a state of permanent recon-struction, with all resources devoted to repairsinstead of addition of new infrastructure and equip-ments. These obstacles to capital accumulation andinfrastructure development lead to a permanent dis-aster-related under-development. This effect hasbeen analyzed by Hallegatte et al. (2007) with areduced-form model showing that the average GDPimpact of natural disasters can be either close to zeroif reconstruction capacity is large enough, or verylarge if reconstruction capacity is too limited (whichmay be the case in less developed countries).

This type of feedback can be amplified by otherlong-term mechanisms, like changes in risk percep-tion that reduces investments in the affected regionsor reduced services that makequalified workers leave theregions. Because of these mech-anisms, the consequences of adisaster can last much longerthan what is normally consid-ered to be the recovery and re-construction period.

An example of assessment on

Katrina in New Orleans

For the landfall of Katrina onNew Orleans, the availability ofa large amount of data allowedmany modelling analyses.Hallegatte (2008), for instance,estimated using a regional input-

output model that indirect eco-nomic losses in Louisiana afterKatrina amounted to 42 billionUS dollars compared to 107 bil-lion US dollars of direct eco-nomic losses. More generally,this analysis concludes thatregional indirect losses increasenonlinearly with direct losses,suggesting the existence ofthreshold in the coping capacityof economic systems. In thisanalysis of Louisiana, indirectlosses remain negligible (or evennegative) for direct losses below50 billion US dollars, and thenincrease nonlinearly to reach 200billion US dollars for direct loss-

es of the same amount (see Figure 4). Also, indirectlosses decrease rapidly if it is possible to ‘import’reconstruction means (workers, equipment, finance)from outside the affected region. This result high-lights the importance of considering the interregion-al interactions.

Conclusions

This article highlights the main difficulties in defin-ing, measuring and predicting the total cost of disas-ters. It focuses on indirect (or output) losses, consid-ered as a major component of the total loss of popu-lation welfare. There are several methodologies toassess these indirect losses, but they are all based onquestionable assumptions and modelling choices,

AMPLIFYING EFFECT ILLUSTRATING HOW NATURAL DISASTERS COULD

BECOME RESPONSIBLE FOR MACRO-LEVEL POVERTY TRAPS


Limited reconstruction capacity

Reduced economic development

Reduced accumulation of capital and infrastructure

Large economic cost of natural disasters

Long reconstruction period after disasters Amplifying

effect

Figure 3

- 50

0

50

100

150

200

250

300

0 50 100 150 200 250 300

Source: Hallegatte (2008).

INDIRECT (OUTPUT) LOSSES AS A FUNCTION OF DIRECT (ASSET) LOSSES

IN LOUISIANA FOR KATRINA-LIKE DISASTERS in billion US dollars

Indirect losses

Direct losses

Figure 4


Focus

and they can lead to very different results. The mainconclusion is of this article twofold.

First, it is impossible to define ‘the cost’ of a disaster,as the relevant cost depends largely on the purposeof the assessment. The best definition and methodobviously depend on whether the assessment is sup-posed to inform insurers, prevention measure cost-benefit analyses, or international aid providers. Afirst lesson from this article is that any disaster costassessment should start by stating clearly the pur-pose of the assessment and the cost definition that isused. Following this recommendation would avoidmisleading use of assessments, and improper com-parison and aggregation of results.

Second, there are large uncertainties on indirect dis-aster costs, and these uncertainties arise both frominsufficient data and inadequate methodologies.Considering the importance of unbiased estimates ofdisaster cost, for instance to assess the desirability ofprevention measures, progress in this domain wouldbe welcome and useful. To do so, much moreresearch should be devoted to this underworkedproblem.

References

Albala-Bertrand, J. M. (1993), The Political Economy of LargeNatural Disasters with Special Reference to Developing Countries,Oxford: Clarendon Press.

Alderman, H., J. Hodditnott and B. Kinsey (2006), “Long-termConsequences of Early Childhood Malnutrition”, Oxford EconomicPapers 58, 450–474.

Cochrane, H. (2004), “Economic Loss: Myth and Measurement”,Disaster Prevention and Management 13, 290–296.

Freeman, P. K., L. A. Martin, J. Linnerooth-Bayer, R. Mechler, S.Saldana, K. Warner and G. Pflug (2002), Financing Reconstruction:Phase II Background Study for the Inter-American DevelopmentBank Regional Policy Dialogue on National Systems forComprehensive Disaster Management, Washington DC: Inter-American Development Bank.

Lindell, M. K. and C. S. Prater (2003), “Assessing CommunityImpacts of Natural Disasters”, Natural Hazards Review 4, 176–185.

Hallegatte, S. (2008), “An Adaptive Regional Input-Output Modeland Its Application to the Assessment of the Economic Cost ofKatrina”, Risk Analysis 28, 779–799.

Hallegatte S., J. C. Hourcade and P. Dumas (2007), Why EconomicDynamics Matter in Assessing Climate Change Damages: Illustrationon Extreme Events”, Ecological Economics 62, 330–340.

Hallegatte, S. and P. Dumas (2008), “Can Natural Disasters HavePositive Consequences? Investigating the Role of EmbodiedTechnical Change”, Ecological Economics 68, 777–786.

Hallegatte, S. and M. Ghil (2008), “Natural Disasters Impacting aMacroeconomic Model with Endogenous Dynamics”, EcologicalEconomics 68, 582–592.

Kroll, C. A., J. D. Landis, Q. Shen and S. Stryker (1991), EconomicImpacts of the Loma Prieta Earthquake: A Focus on Small Business,Studies on the Loma Prieta Earthquake, University of California,Transportation Center.

McCarty, C and S. K. Smith (2005), Florida’s 2004 Hurricane Season:Local Effects, Florida Focus, University of Florida,http://www.bebr.ufl.edu/system/files/FloridaFocus1_3_2005_0.pdf.

Mechler, R., J. Linnerooth-Bayer and D. Peppiatt (2006),Microinsurance for Natural Disasters in Developing Countries:Benefits, Limitations and Viability, ProVention Consortium,Geneva,http://www.proventionconsortium.org/themes/default/pdfs/Microinsurance_study_July06.pdf.

Norris, F. H. (2005), Range, Magnitude, and Duration of the Effectsof Disasters on Mental Health: Review Update 2005, DartmouthMedical School and National Center for PTSD, Hanover andBoston.

Okuyama, Y. (2004), “Modeling Spatial Economic Impacts of anEarthquake: Input-Output Approaches”, Disaster Prevention andManagement 13, 297–306.

Pelling, M., A. Özerdem and S. Barakat (2002), “The Macro-eco-nomic Impact of Disasters”, Progress in Development Studies 2,283–305.

Rose, A. (2004), “Economic Principles, Issues, and ResearchPriorities in Hazard Loss Estimation”, in: Okuyama,Y. and S. Chang(eds.), Modeling Spatial and Economic Impacts of Disasters, Berlin:Springer, 14–36.

Rose, A., G. Oladosu and S. Y. Liao (2007), “Business InterruptionImpacts of a Terrorist Attack on the Electric Power System of Los Angeles: Customer Resilience to a Total Blackout”, RiskAnalysis 27, 513–531.

Santos, I. (2007) Disentangling the Effects of Natural Disasters onChildren: 2001 Earthquakes in El Salvador, Doctoral Dissertation,Kennedy School of Government, Harvard University.

Tierney, K., (1997),“Business Impacts of the Northridge Earthquake”,Journal of Continencies and Crisis Management 5, 87–97.

West, C. T. and D. G. Lenze (1994), “Modeling the Regional Impactof Natural Disasters and Recovery: A General Framework and anApplication to Hurricane Andrew”, International Regional ScienceReview 17, 121–150.

Documents

HE CONOMICS OF ATURAL DISASTERS S H · 2020-04-17 · ments of the total disaster cost to the population, i.e. an estimate of welfare losses. Moreover, disaster impacts can have positive