Page - 305 - in Loss and Damage from Climate Change - Concepts, Methods and Policy Options

12 IntegratedDisasterRiskManagementandAdaptation 305 It is important to realise that thefinancial compensation arrangement should be designed so that it is complementary to, and facilitates, the undertaking of cost- effective adaptationmeasures, and not acts as a substitute or financial disincentive for implementing suchmeasures.Amoral hazard effect canarisewhen individuals prepare less for a risk after theyhaveobtained insurance coverage against the risk. Thiscanoccurwhenpolicyholdersexpecttoreceivecompensationfromtheirinsurer irrespectiveof risk reductionefforts and ifpolicyholders receivenofinancial incen- tives, like lowerpremiums, from their insurer to limit risk.This canposeproblems for the insurerwhendue to information asymmetries, the insurer does not observe theheightened risks takenbyaparticular policyholder.This implies thehigher risk isnot adequately reflected inahigher risk-basedpremium.Moreover, suchamoral hazardeffect isevidentlyundesirablewhenclimatechangeincreasesnaturaldisaster riskssince ithampers the implementationofadaptationmeasuresbypeoplecovered by insurance. Hudsonetal. (2017)examinedtheexistenceofthismoralhazardeffectusingdata fromsamplesofhouseholds livingalong the riverElbe inGermany.This isdoneby estimating relations between flood insurance coverage and the implementation of flood damagemitigationmeasures, and by estimatingwhether flood damage out- comesdifferbetween the insuredanduninsured,whilecontrolling foradiversityof other relevant explanatory variables. The results show that amoral hazard effect is absent (Hudsonet al. 2017). Inparticular,flooddamagesof insuredhouseholds are notsignificantlyhigherthanthoseofuninsuredhouseholdswhendifferencesinflood hazardcharacteristicsareaccountedfor.Moreover, individualswithfloodinsurance coverage aremore likely to have taken specific flood damagemitigationmeasures thanpeoplewithoutfloodinsurance.Theseinsuredindividualsdidnotreceiveapre- miumdiscountfortakingflooddamagemitigationmeasures,whichimpliesthatother reasonsexplainwhythe insuredwerebetterpreparedforfloodrisk.Theresultssug- gest that behavioural characteristics, like high risk aversion, imply that individuals havepreferences forboth insurancecoverageand riskmitigation. Although the relationsbetween risk reductionandnatural disaster insurancehas receivedlittleempirical research(seealsochapterbyLinnerooth-Bayeretal.2018), the findings by Hudson et al. (2017) do not stand by themselves. Thieken et al. (2006)alsoobserved that individualswithflood insurancecoverage inGermanyare betterpreparedforfloodingthanpeoplewithoutfloodinsurance.Botzenetal.(2017) findpositive relations between havingflood insurance coverage and implementing flood-proofingmeasuresamonghomeownersinflood-proneareasinNewYorkCity. Hudsonetal. (2017)showthatsimilarpositiverelationsbetweeninsurancecoverage andrisk reductioncanbe foundforwindstormrisks in several areas in theU.S. that were impactedbyhurricanes Irene, Isaac, andSandy.Thesefindings are consistent with positive relations betweenwindstormcoverage andwindstorm risk reduction activities reported inCarsonet al. (2013)andPetrolia et al. (2015). Additional calls have beenmade to design natural disaster insurance arrange- ments in away that they incentivise risk reduction by policyholders. For instance, insurance could reward investments in damagemitigationmeasureswith premium discounts (Kunreuther 2015). There are few examples of flood insurance arrange-