lable at ScienceDirect

Ocean & Coastal Management 54 (2011) 164e172

Contents lists avai

Ocean & Coastal Management

journal homepage: www.elsevier .com/locate/ocecoaman

Coastal risk management: How to motivate individual economic decisions tolower flood risk?

Tatiana Filatova a,b,*, Jan P.M. Mulder b,c,d, Anne van der Veen c,e

aCentre for Studies in Technology and Sustainable Development, Faculty of Management and Governance, University of Twente, P.O. Box 217, 7500AE Enschede, The NetherlandsbDeltares, P.O. Box 177, 2600 MH Delft, The NetherlandscDepartment of Water Engineering and Management, University of Twente, P.O. Box 217, 7500AE Enschede, The NetherlandsdNetherlands Centre for Coastal Research/NCK, The Netherland1e International Institute for Geo-Information Science and Earth Observation, Hengelosestraat 99, P.O. Box 6, 7500AA Enschede, The Netherlands

a r t i c l e i n f o

Article history:Available online 5 November 2010

* Corresponding author. Centre for Studies inDevelopment, Faculty of Management and GovernanBox 217, 7500AE Enschede, The Netherlands. Tel.: þ 3489 4850.

E-mail addresses: T.Filatova@utwente.nl (T. Filat(J.P.M. Mulder), Veen@itc.nl (A. van der Veen).

1 www.nck-web.org.2 Flooding, which is caused either by a break of a d

two physical effects: inundation and erosion. Here whwe assume both effects.

0964-5691/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.ocecoaman.2010.10.028

a b s t r a c t

Coastal flood risk is defined as a product of probability of event and its effect, measured in terms ofdamage. The paper is focused on coastal management strategies aimed to decrease risk by decreasingpotential damage. We review socio-economic literature to show that total flood damage depends onindividual location choices in the housing market and on individual flood risk awareness. Low flood riskawareness leads to inefficient spatial developments and increased flood risk. We show that personalexperience, risk communication, financial instruments like insurance from flooding and technicalinstruments like building on high elevations, are factors that increase individual risk awareness. Evidencethat these factors indeed affect housing prices and land use patterns is provided. We discuss proactiveinstruments that can be used in coastal zone management in the Netherlands to increase individual riskawareness. We argue that policy-makers may create incentives giving individuals a possibility to makelocation choices that lead to less total flood risk in the coastal zone area.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Worldwide the amount of capital in coastal zones susceptible toflooding2 is increasing.According to the IPCCthedamage fromnaturaldisasters in Europe has rapidly increased over the past decades,mainly because of the growth of capital accumulated in flood-proneareas (Nicholls etal., 2007). In theNetherlandswhereabout70%of theGross National Product is earned in the areas below sea level (Veraartet al., 2007), the issue of decreasing flood risk attracts a lot of atten-tion.Riskofflooding inEuropeanwatermanagement (alsocommonlyknown as expected loss) is defined as a function of the probability ofa flood event and its potential effect (in terms of monetary damageand human causalities) (Rijkswaterstaat, 2005a). This implies that

Technology and Sustainablece, University of Twente, P.O.1 53 489 3530; fax: þ 31 53

ova), jan.mulder@deltares.nl

ike or high water levels, hasen we say flooding (or flood)

All rights reserved.

lower flood risk can be achieved either by decreasing probability offlooding, or by decreasing potential damage from flooding or bycombining the two. In Dutch water management traditionally thefocus has been on reducting the probability of flooding by means ofengineering defense constructions (i.e. strengthening dikes anddunes) (Rijkswaterstaat, 2002; Smits et al., 2006; Bucx et al., 2008).However, thedecrease in totalflood riskdue to loweredprobabilityofflood defense failure is vanished if the economic value of the areacontinues togrow inzonesvulnerable toflood. Flood riskcanbereallyreduced only if engineered coastal defense measures are com-plemented with an economic use of a flood zone that ensures lesspotential damage.

The extent of flood damage depends on spatial patterns ofresidential and commercial areas and their values (Rijkswaterstaat,2005c). Both patterns and prices are the outcomes of many indi-vidual interactions in the urban landmarket. Due to several reasonseconomic developments occur in proximity to old economiccenters, which originated close to water ways or harbors. First,production and business companies benefit from clustering (Fujitaand Thisse, 2002). Second, households are attracted to economi-cally developed areas because of employment opportunities. Third,in addition to these economic factors coastal zones provideimportant environmental amenities, which are highly valued by

Fig. 1. Self-reinforcing loop that fuels the growth of flood risk in coastal zones (CZ).

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172 165

households (Bin et al., 2008). Thus, all economic forces work topromote growth of capital in flood-prone areas.

In coastal zones another vital factor in thedecision tobuypropertyor to invest is subjective probability of flooding (MacDonald et al.,1987). Low-risk awareness leads people to buy properties in thezones vulnerable to flooding at higher prices and in higher amountsthan would be beneficial for a society as a whole. It was shown thatpeople who underestimate flood probability drive urban develop-ments to expand to economically inefficient3 zones (Tatano et al.,2004; Filatova, 2009). This paper explores stimuli to motivate indi-viduals to make microeconomic decisions at the housing market inthe Netherlands that lead to less flood risk at the aggregated level. Atthe same time we highlight the aspects of shared responsibilitiesbetween government and individuals with regard to flood risk.

We focus on two research questions: 1) How do individualsmake economic decisions at the housing market if there is a prob-ability of flooding? 2) What instruments may policy-makers use topromote outcomes of a housing market that would lead to lessflood risk? First, we outline the challenges associated with thecurrent water management in the Netherlands. Next, from thereview of economic studies we provide both theoretical andempirical evidence of the influence of flood probability on landprices and spatial patterns and explore the role of individual riskawareness. Third, the results of surveys aimed to elucidate indi-vidual coastal flood risk awareness in the Netherlands arereviewed. We then discuss four factors that increase risk awarenessin coastal zones, and have a direct measurable effect on land usepatterns and housing prices and, thus, on potential flood damage.Finally we draw conclusions on the effectiveness of these factorsand possibilities to use them in water management.

2. Challenges for flood risk reduction in the Netherlands

In the Netherlands, at the macro-level the government took onthe responsibility to minimize flood risk. Thus, society as a whole iscollectively responsible for flood protection financed via the taxsystem irrespectively of the risk levels in the location of taxpayersand actual tax payment. Today the country is protected by a systemof dike rings with different safety levels, whose standards areenforced by law (Wet op de Waterkering, 1995). The dikes anddunes in the provinces of North and South Holland are supposed towithstand a storm surge with a probability of occurrence of once in10 000 years. Although the probability that a disaster may happenis low, the consequences will be dramatic (€ 300 billion at risk ofloss for the “Centraal-Holland” dike-ring alone in prices of 2000(Kok et al., 2002)). The present safety standards, related to proba-bilities of flooding, are based on the guidelines developed by thefirst Deltacommissie (1960). In the Netherlands the spatiallydifferentiated probabilities (controlled by dike/dune strength) aredetermined by policy-makers taking into account the number ofpeople living in a certain area and the economic value of the area.The second Deltacommissie (2008) has proposed to increase thecurrent safety levels by decreasing the probability level by a factor10.

However, by only decreasing probability levels one shadowssome important hidden feedbacks between micro (individual) andmacro (policy) decisions in the process of risk reduction. The higherthe density of population and economic value of a area (boxes I. andII. in Fig. 1) the more the government is under pressure tominimizethe probability of flooding (box III. in Fig. 1). However, there is

3 Economic outcome is considered to be economically efficient if no single personcan be made better off, without making somebody else worse off. Inefficiency is theopposite.

a clear danger of a positive feedback here. The safer it becomes tolive in the potentially vulnerable areas, the more people andbusinesses are attracted to settle and to invest there (box IV. inFig. 1). Moreover, economic forces work to attract more businessand households seeking jobs to the existing economic clusters incoastal zones. The economic value of the area increases even more(boxes V. and I. in Fig. 1) and again may push the government toincrease the safety in this area (box III. in Fig. 1).

It is a self-reinforcing cycle that has a negative effect on flood risk:the safer it becomes to live somewhere, the more economic agentswould like to live and work there, the higher the total flood riskbecomes and the more the state should invest to increase overallsafety standards. Note, that individuals investing in coastal zone donot bear any risk personally: the government has the collectiveresponsibility to keep total flood risk from growing. There areneither motivation nor means for individuals who choose to locatein coastal zone to share this responsibility for safety. Thus, thegovernment needs to invest public money to increase safety levelsof dikes and dunes (i.e. flood probability). Eventually, a cost-benefitanalysis (CBA) e calculating the cost of technical protectionmeasures and the benefits of avoided risk e will indicate whetherthe cycle may continue. The key question is what will happen at thecritical point in time when the CBA will appear to be negative. Inorder to postpone or even prevent this point in time, it seemsworthwhile to investigate ways to interrupt the self-reinforcingcycle and turn it into a positive direction of maintaining or evendecreasing flood risk.

With respect to damage reduction the following aspects deservespecial attention:

1. Damage from a flood event is calculated as a sum of direct andindirect economic damage and damage from business inter-ruption (Rijkswaterstaat, 2005a, 2005c). Both spatial patternsand prices of properties play an essential role in the potentialdirect damage. In the Netherlands spatial patterns of devel-opment are strictly controlled by the government via spatialplanning (Rijkswaterstaat, 2002). Housing prices, however, arethe outcomes of market allocation of land between competitiveuses where individual choice plays the main role. Thus, the riskof flooding is related to land/housing market outcomes, i.e. toindividual demands for particular locations.

2. Safety from flooding is a public good. Formally in theNetherlands the government has an overall responsibility todecrease flood risk along the coast, while individuals bear noneof the risk in case of a disaster. However, individual locationdecisions create capital at stake making citizens partlycontributing to the increased risk of flooding in the coastalzone. The present system of water management does not havemechanisms to account for these shared responsibilities. There areno incentives at the individual level to make housing decisionsthat would lower potential flood damage.

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172166

Characteristics of the traditional approach to flood safety in theNetherlands, as described above, are the dominant role of thegovernment, no active mechanisms for individuals to make landmarket decisions leading to less flood damage and the emphasis ontechnical measures including technical expressions of safety levels.Even though, technically speaking, the probability of a dike failureis not zero, this water management practice has created a feeling ofabsolute safety amongst the population (see Section 4) and ofabsolute trust in the government taking care of the safety. Cana change in this attitude contribute to a reduction in flood risk?

We state that if individuals who buy properties in the land marketare aware of risks that they are taking when settling in a vulnerablezone (box IV in Fig. 1), then flood risk in coastal zone decreases. Weargue that this measure can reduce the growth of flood risk in coastalzones even if the safety standards (probabilities of disaster occur-rence) (Wet op de Waterkering, 1995) remain constant. To supportour argument we first provide evidence that individual riskawareness influences land prices and spatial patterns and, conse-quently, flood risk. Second, we discuss possibilities to increaseindividual risk awareness, which affects individual location deci-sions, so that flood risk in coastal zones can be decreased.

3. Land use and housing values in flood-prone areas: howdoes probability of flood enter into economic decisions athousing market?

3.1. Theory: urban economics and economic decisions under risk

The functioning of markets is well studied in economics (Varian,1992). Economic agents in a land market have preferences forproperties, fromwhich the demand for land can be determined. Thesupply of housing depends on geographical conditions, spatialplanning, and the structural density at which developers supplyresidential or commercial buildings. Conceptually, the price ofa spatial good (i.e. a house or land) is the intersection point of theaggregated demand and supply curves. This is a very simplifiedview on a housing market; however, it gives us some powerfulinsights into how everything is interconnected. If demand forproperties goes up then prices also increase (Buurman et al., 2001).For a coastal city this also implies that potential direct damage fromflooding will increase. For example, if households have strongpreferences for a seaside view, then demand for coastal propertiesincreases and so do prices. Alternatively, if economic agents areaware of the risk of flooding in the locations close to the seaside,then the proximity to the coast might serve as a repulsive factor.The aggregated demand for land is likely to decrease pushing landprices down. Consequently, the lower the property price the loweris the direct economic damage from flooding. Obviously, these twofactors are spatially correlated and both affect property pricesdriving them in opposite directions. As a matter of fact, in theNetherlands an average price of properties at the coast is higherthan the average for coastal provinces. Specifically, in the provinceof Zuid-Holland the difference between average property pricesalong the coast and those more landward, was € 99 400 in 2005(Vliz, 2005). Considering the fact that flood probability is the sameeverywhere in the dike-ring covering Zuid-Holland, this mayindicate that coastal amenities and economic attractiveness of theDutch coast exhibit strong influence.

Urban economics studies location decisions of individualhouseholds and firms in a city and aggregated urban features suchas land prices and spatial structure. The majority of urban modelsare based on the monocentric urban model (Alonso, 1964).Households search for a location that maximizes their utility underbudget constraint. Firms search for a location that maximizes theirprofits. This standard urban model was extended to account for

several employment centers (Fujita and Thisse, 2002), as well as forthe influence of environmental amenities (such as a seaside view)(Wu and Plantinga, 2003).

Under conditions of uncertainty, such as when buying proper-ties in an area potentially vulnerable to flood or erosion, peoplemaximize expected utility (expected profit). Thus, a choice betweenlocating in a flood-prone coastal area and a safe area, might be seenas maximization of expected utility function (Varian, 1992). Thespecific characteristic of the Dutch coastal zone is that the proba-bility of flooding is very low. At the same time, benefits of locatingclose to the economically most developed area and coastalamenities are very appealing, which causes urban developments inthe coastal zone to expand.

Moreover, economic agentsmay hold subjective beliefs about theprobabilities, causes, and effects of flood events in the future. Thesebeliefs depend on their experience, income, education and otherpersonal factors. As discussed in literature (Varian,1992; Slovic,1987;Sjoeberg, 2000) economic agents have subjective perceptions ofa flooding probability and associated risk. Experimental economicshas proven that subjective risk perceptions very much depend onhow the risky situation is framed or presented to an individual deci-sion-maker (Kahneman and Tversky, 1979). Subjective probabilitiesof an event are updatedwhenever a hazard event occurs and becomelower if it does not occur for a long time. Empirical studies also revealthat the perceptionof risk increases after anevent (Kaiser et al., 2004)and gradually reduces in the long-term, if an event does not happenperiodically (Chivers and Flores, 2002).

Tatano et al. (2004) proposed an urban model, in which theprobability of a disaster was integrated as a qualitative character-istic of land. This model borrows much from the traditional mon-ocentric model, except for the fact that a city is divided into twozones: a safe zone and a zone vulnerable to a natural disaster.Households are assumed to have some “perception bias” for thechance of a disaster. Equilibrium urban spatial pattern and landprices depend on individual risk perceptions of a natural disaster.The authors concluded that if risk perception biases exist, then anefficient allocation of land in the city is not possible.

3.2. Empirical evidence: hedonic analysis of housing prices in flood-prone areas

The influence of risk of flooding on the land prices is also empir-ically justified by hedonic studies (MacDonald et al.,1987; IWR,1998;Daniel et al., 2007). In hedonic price studies a spatial good is consid-ered as a differentiated market good with specific quantitative andqualitative attributes (Rosen, 1974). The probability of flooding is animportant characteristic of the quality of a spatial good and itcontributes to the overall property price. All else equal, buyerschoosing to locate in an area vulnerable to natural disaster andaccepting potential future losses, have lower bid prices. This capital-izes in housing prices, meaning that property prices normallydecrease in thevulnerableareas.Thepropertypricediscountdependson the perceived loss and probability of a disaster (MacDonald et al.,1987). It develops in the competitive markets with perfectlyinformed buyers and sellers. However, information asymmetry, riskperception bias, and expectations that the government fully providespublic safety, add to the complexity of market dynamics (IWR,1998).

Using property transaction data, hedonic price researchersestimate the monetary value of every significant characteristic ofa spatial good. In this case, the probability of flooding will havea negative value, which might reduce the total value of a spatialgood in the market. Numerous studies trying to identify thediscount of the housing prices caused by flood risks, were done inthe USA. The common finding is that a probability of flooddecreases property values (Table 1).

Table 1Hedonic estimations of the housing price discount in the flood hazard areas.

Authors, year of publication Case-study area Type of natural disaster Housing price reduction in the flood-pronearea compared to the safe area

MacDonald et al., 1987 Monroe, Louisiana Spring river flooding 2.8%e2.9%Speyrer & Ragas, 1991 New Orleans, Louisiana, USA Flooding 4.2%e6.3%Shultz and Fridgen, 2001 Fargo, North Dakota and Moorhead,

Minnesota, USA;Spring river flooding 8.8%

Daniel et al., 2007 North Carolina, USA River flooding 7%e13%Bin et al., 2008 North Carolina, USA Coastal flooding 6.2%e7.8%

4 http://comrisk.hosted-by-kfki.baw.de/index.html

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172 167

In the Netherlands, the hedonic price method rarely has beenapplied to evaluate flood effects on housing values. To our knowl-edge there is only one study applying it to river flooding (Danielet al., 2007). The results of this hedonic price analysis for thedwellings in the flood plain of Meuse River, showed that localhousing markets in the Netherlands are sensitive to flood risk. Thehousing sales during the period from 1990e2004 were analyzedand the result was that the prices of houses located in an areainundated in 1993, were ex-ante 7% lower than a similar houselocated in a safe zone, and went down by 14% after the flood.

River flooding is a more frequent event than coastal flooding inthe Netherlands; people are more aware of river flood risks andreact accordingly in the housing market. This makes the situationdifferent from coastal flooding, which is rarer.

To summarize the economic literature, we outline threeimportant facts:

1. Individual decisions matter: Land prices and land use patterns(on which flood risk depends) are determined by the totaldemand and supply of land/housing in the region. The lattertwo are formed by aggregating individual supplies anddemands (i.e. individual preferences for locations and financialpossibilities);

2. Probability of flood matters: There are attractive and repulsiveforces, which influence individual decisions on where to locateor to invest, i.e. the individual demand for location. These forcesare formed by preferences of economic agents over character-istics of spatial environment. Probability of flooding, if inte-grated in location decisions, affects housing prices and land usepatterns;

3. Risk awareness matters: Individuals have subjectively perceivedprobability of the hazardous event, which cause economicallyinefficient biased aggregated outcomes such (i.e. land usepatterns and land prices).

4. Perception of risk of coastal flooding in the Netherlands

Water managers operate with probabilities, but people verypoorly understand probabilistic risks (Slovic, 1987). Especiallywhen an event has a low-probability, such as coastal flood in theNetherlands, people hardly integrate risk into their decisions (Binand Kruse, 2006; Camerer and Kunreuther, 1989). Risk perceptionis shaped by socio-economic, demographic and cultural factors andby previous individual experience of the hazardous event, and maydiffer from region to region (Kaiser et al., 2004; Williams et al.,1999; Raaijmakers et al., 2008).

Several surveys, which investigate the level of coastal riskawareness, were carried out in the Netherlands. Recently there werethree surveys studying flood risk perception in the Netherlandscarried out (Krywkow et al., 2008; Terpstra and Gutteling, 2008;Bo�ckarjova et al., 2009). The studies observed low level of riskperception together with a rather high indicator of trust in govern-ment. Specifically, the mean value of the variable ‘worry’ character-izing perceived risk was 2.5 on the 11-point scale (0 meaning

‘absolutely do not feel any risk’; number of respondents ¼ 1052)(Bo�ckarjova et al., 2009). Respondents in another study (n¼ 658) alsorevealed low-risk perceptions measured in terms of personal risk,perceived likelihoodand feelings of fear choosing the average score of1.77 (measured on a five-point scalewhere 1means no personal risk,and 5 a very high personal risk). At the same time the majority of therespondents (73%)believe thatgovernment isprimarilyorcompletelyresponsible for damage mitigation (Terpstra and Gutteling, 2008). Inaddition, various characteristics of risk perception (Worry, Vulnera-bility and Severity) decrease as the level of Trust (also measured ona likert-scale) in authorities rises (Bo�ckarjova et al., 2009).

To our knowledge there are two other surveys that have beencompleted earlier. In 2000 there was a survey conducted in sixDutch coastal towns (van den Berg et al., 2002). Themainpurpose ofthat surveywas to understand thewillingness of coastal residents toaccept flood risks. The sample was very small and included only 12households (and 48 for the set of questions concerning alternativepolicies). The main conclusion of the survey was that citizens of thecoastal towns felt themselves safe and had trust in the defensemeasures designed by the government. The most common answerto the question “Do you think that flooding could happen in the areawhere you live?”was negative (see (van den Berg et al., 2002) p. 20)although each of the towns is certainly prone to flooding.

Another survey was conducted under the umbrella of the COM-RISK4 project (Kaiser et al., 2004). It covered several Europeancountries in the North Sea region, including the Netherlands. One ofthe objectives was the analysis of the present state of publicperceptions of flood and flood defense. About 22.3% of the 400questionnaires distributedwere returned (making the sample groupof about 89 respondents). Although 34% of respondents had expe-rienced coastal flooding, 74% of them considered the probability ofcoastal flooding as “low” and “very low”. Also, 95% of respondentsdid not take any personalmeasures to be prepared for a storm flood.The survey suggested that knowledge about the probability ofdisaster did not imply awareness about consequences. Although theprobability of coastal flooding in the Netherlands is low, flood risk(i.e. probability times effect) is enormous.

It is surprising to conclude that in the Netherlands where half ofthe area is below sea level, citizens have very low coastal flood riskawareness and hardly carry out any preparedness measures. Onepossible explanation for this may go back to the history of floodsand water management in the Netherlands. The experience ofa disaster and the time since it occurred strongly influencesperceptions of risk (Kaiser and et al., 2004). The last coastal floodhappened in the Netherlands in 1953. That event took away over1800 human lives, destroyed 3000 houses and 300 farms, over40 000 houses and 3000 farms were damaged. The inundation ofalmost 200 000 ha of land caused an enormous material damage of3.1 billion discounted Dutch guilders (Marchand et al., 2006). Afterthis tragedy the Dutch government promised to improve flooddefense, so that such a disaster would not happen again

Fig. 2. Conceptual links between flood risk, housing market and individual riskawareness.

6 The chain of safety is an approach to flooding disaster prevention aimed toprovide proaction, prevention, preparation, response and recovery strategies tominimize the casualties and damage caused by coastal flooding. See details athttp://www.chainofsafety.com/

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172168

(Deltacommissie, 1960). The enormous Delta works resulted inconstruction of a strong wall against the sea in the south-west ofthe country. Coastal engineering constructions, which made theNetherlands a country with the highest flood safety standards inthe world, have been considered as ‘the’ solution to the floodproblem and have created a feeling of absolute safety. Perhaps,feeling safe behind strong flood defenses, the Dutch society hasstarted to experience an effect of dissonant perception, i.e.a perception of the past hazardous event as a coincidence, anda conviction that this is unlikely to happen again (Kaiser and et al.,2004). Dissonant risk perception is attributed to people who havemuch material wealth and have to live their day-to-day lives witha chance of a major disaster (Smith, 2001).

Another explanation for low awareness of coastal flooding in theNetherlands might be the fact that the threats along the Dutchcoast are actually not visible for ordinary citizens. Indeed with thedecision to maintain the coastline in the 1990 position(Rijkswaterstaat, 1990), coastal nourishments5 fix the erosion thathappens during winter storms. The civil engineering solution tomaintain the coastlinewas successful in preventing dune erosion tospread landward. However, it may exacerbate a false feeling ofsafety and further decrease individual risk awareness.

5. Increasing individual flood risk awareness asa complementary measure to reduce flood risk

Following the logic outlined in the previous sections (see Fig. 2)we argue that in order to decrease risk of flooding in coastal zones, animportant option is to increase individual coastal flood risk aware-ness. In this sectionwe discuss how risk awareness can be positivelyinfluenced by personal experience, risk communication, financialmechanisms and technical engineering solutions. We focus on themeasurable effect of an increased risk awareness by one of thesefactors, on individual location choices and consequently on landprices and on overall damage from flooding in the vulnerable areas.

5.1. Personal experience of disaster and risk awareness

Perceived risk and experience of disaster are closely related.People who have ever experienced flooding themselves, are likelyto have higher flood risk awareness than those who never did(Kaiser et al., 2004; Lave and Lave, 1991). There is clear evidence

5 Coastal nourishments (also known as beach or sand nourishments) is a processof supplying sand to the beach to replace its loss during coastal storms and erosion.In the Netherlands sand nourishments are done annually.

that risk awareness, increased by a natural disaster, can changeindividual location decisions.

Example (effect of a hurricane on a housing market): Hallstromand Smith (Hallstrom and Smith, 2005) evaluated whether theprices of properties known to be in risky locations, have adjusted inresponse to the information provided by the hurricane Andrew inFlorida, USA. In 1992 this area experienced the strongest storm overthe last 20 years. This hurricane “nearly-missed” Lee County andpassed the area 75 miles to the south without any damage to LeeCounty. However, the hurricane provided information about theeffect of the natural disaster. The housing prices in the flood hazardareas went down by 19% after the “nearly-missed” hurricane. Thisshows how people perceive risk implying that home buyers andsellers appear to have incorporated the risk in their decisions in thehousing market, not just because the actual disaster happened tothem, but even because the disaster ‘nearly’ happened. It was notdamage from flooding that made housing prices go down. People’sflood risk awareness was affected and they responded at thehousing market with lower housing prices in the risk-prone zone.

Economically speaking, if subjective probability of flooding isupdated to the actual probability level, then the individuallyexpected utility from a location that is vulnerable to flood,decreases. Thus, the demand for urban land in vulnerable zonedecreases followed by decrease in property prices and in potentialdirect flood damage.

Conclusion 1: If individual flood risk awareness increases, thenspatial patterns and land prices in vulnerable areas may change.These changes decrease potential economic damage from floodingand thus total flood risk.

Examples above show that personal experience is indeed animportant factor in increasing flood risk awareness. However,stimulating floods in order to promote personal flooding experi-ences, of course is not a reasonable policy instrument. Neverthe-less, broadcasting and visualizing these previous experiences maybe an important component of a risk communication strategy.

5.2. Risk communication and risk awareness

By applying a chain-of safety approach6, European watermanagement, in the first place, aims at proactive measures.Increased risk awareness is essential to achieve this. Not only doesincreased risk awareness lead to changes in location behavior anddecreaseof totalflood risk in thearea, asdiscussedabove; it is also anexpression of the willingness of people to respond or acceptproactive actions of government. For example, citizens will be morelikely toaccept “roomfor the river7” as afloodmanagement strategy,if they are aware that riverfloods can happen. Thus, increasing floodrisk awareness is notonlyaproactivemeasure in itself; it also createsground for acceptance of other proactive flood protection initiativesof government. Consequently, the government needs to finda proper flood risk communication tool or another way to increaseindividual risk awareness (Rijkswaterstaat, 2005b).

Decision-makers inwater management operate with a technicalnotion of risk. This, however, is quite different from the notion of thegeneral public about risk (Raaijmakers et al., 2008; Plough and

7 ‘Room for the River’ is a package of measures concerning the spatial planningkey decisions in the Netherlands. In particular, it aims to create more space forrivers to flood in the zones where the least damage is expected to occur (e.g. part ofagricultural land along the rivers is being transferred into the ‘planned’ floodplains). For details see http://www.ruimtevoorderivier.nl/

Table 2Factors relevant to the technical and cultural rationality of risk (After (Plough andKrimsky, 1987)).

Technical Rationality Cultural Rationality

Trust in scientific methods,explanations, evidence

Trust in political culture anddemocratic process

Appeal to authority and expertise Appeal to folk wisdom, peergroups, and traditions

Risks are depersonalized Risks are personalizedEmphasis on statistical

variation and probabilityEmphasis on the impacts of riskon the family and community

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172 169

Krimsky, 1987) (see Table 2 for the differences between technicaland cultural rationality). Following a need to communicate differenttypesof risks (health risk, risk fromnatural disasters, industrial risks,risk from genetically modified food and so on) to people, a newscientific field has emerged (Plough and Krimsky, 1987). The mainchallenge in risk communication is that the public perception ofhazards is inconsistentwith the objective information. The questionis how to present information to the public in order to gain itsunderstanding and acceptance for policy decisions. Thus, increasingrisk awareness is all about presenting information in a way accept-able for the public, specifically, addressing it’s feelings (Baker,1990).However, informing about potential flood risks and not providingreal instruments for individuals to act upon this risk, does notprovide a ground for a shared responsibility for risk eliminationbetween government and individuals.

Risk communication strategies can be based on different theo-ries; for example organizational theory (Chess, 2001) or mentalmodel approaches (Kolkman et al., 2007; Wagner, 2007). Specialattention is paid to the communication of the risk of natural hazards(Lave and Lave, 1991; Baker, 1990). In particular, Baker (1990)provides a detailed 8-step guide on how to transfer informationabout environmental risks from experts to a non-expert public.

Risk communication and increase of risk awareness is a verycomplex and long-term process since it has to deal with promotionand establishment of new social or public norms or opinions ina society. Bier (2001) reviewed several empirical studies andinferred positive and negative lessons learned. There are manyobstacles to effective risk communication, including some externaland internal barriers inside a governmental agency or ministry(such as inadequate resources, lack of time, special interests orinternal politics) (Johnson and Chess, 2006). Moreover, one needsa delicate balance in communicating flood risk to avoid panic andrunaway strategies.

Example (effect of media coverage of risk on a housing market):Risk is communicated by providing information about a hazardousevent. With new information, individual beliefs about risk areupdated. McCluskey and Rausser (2001) tried to measure theevolution of risk perception due to information people obtain fromthe media. Particularly, they have analyzed the effect of news abouta hazardous waste site published in a local newspaper, on changesin risk perception. The hypothesis was that “if the media affects thepublic perception of risk, then media coverage of environmentaldamage should be a significant factor in determining propertyvalues”. They concluded that media coverage increases perceivedrisk, which in turn decreases property prices. Gayer with colleagues(2000) analyzed changes in perceived risk of cancer fromhazardous waste sites, due to newspaper coverage, and effect ofthese changes on the housing market. The analysis showed thatnewspaper publicity about the local hazardous waste sites hada negative effect on housing prices.

Conclusion 2: If a flood risk communication strategy is designedin an appropriate way, it may increase individual risk awareness ofenvironmental hazards. With increased risk awareness economic

agents are likely to integrate flood risk in their location decisions,which might be reflected in decreased land prices and decreasedflood risk for the area.

5.3. Insurance against flooding as a measure to increase coastalflood risk awareness

Dikes and insurance are two totally different measures of floodprotection. The former reduces the probability of a disaster, and thelatter reduces the individual damage from flooding. Insuranceagainst flooding has been implemented for flood-prone areas inseveral countries, such as France, Germany, the UK and the USA, butnot in the Netherlands. Compulsory flood insurance is supposed toimprove the economic efficiency of the use of the flood-prone areas(Chivers and Flores, 2002). One of flood insurance’s primarypurposes is to guarantee premiums proportional to risk for indi-viduals in vulnerable areas (Krutilla, 1966). Botzen and van denBergh (2008) discuss the feasibility of flood insurance as a risk-sharing mechanism in the Netherlands.

Yet, in addition to risk-sharing, flood insurance may serveanother objective, which seems to be ignored in the discussionsover its implementation in the Netherlands: in areas where peopleignore risks, which is often the case for low-probability high-impact events such as coastal floods in the Netherlands, insurancemay serve as a measure to communicate this risk and to persuadepeople to integrate it in their decision making. As noted by Chiversand Flores (2002) “even if people do not appreciate the risk theyface by locating in the flood plain for whatever reason, compulsoryinsurance,.will force them to face the social cost of locating in theflood plain”. If compulsory flood insurance is present, then risk offlooding becomes explicitly known to people making a choice tobuy a house or to invest. Since the late 1960s the federal govern-ment in the USA has emphasized flood insurance as a primary toolfor improving location behavior in flood-prone areas (Lave andLave, 1991). To make insurance work as both risk-sharing and riskcommunication device, information asymmetry should be avoided(Bin and Polasky, 2004) and spatial discrimination of risk and itsaccurate prices should be emphasized.

In practice insurance implies that individuals and governmentshare the responsibility in the case of disaster. As for any public good,safety from flooding may cause a discussion about who has to payfor its provision. In the Netherlands, traditionally the governmentand Water Boards are responsible for building and maintainingflood protection structures and guaranteeing a certain safety level.All citizens, independently of the fact whether they live in flood-prone or in safe areas, have to pay their share of taxes to financeflood protection measures for vulnerable zones. As such there isa collective responsibility for flood protection.

There is an important difference between collective and sharedresponsibility. Collective responsibility means that society asa whole is responsible, but nobody in particular. In this case flooddefense measures are financed via the tax system, which does notdifferentiate between risk levels in the location of taxpayers andactual tax payment. Shared responsibility means that society asa whole (represented by a government decision-maker) and indi-viduals, both play active roles in risk reduction. Insurance is anexample of a shared responsibility arrangement. Insurance paymentdiscriminates on a risk of a natural hazard and individuals canactually choose whether to take it. However, since the introductionof an insurance against flooding implies a partially individualresponsibility for flood risk, the Dutch are very reluctant to it.

Nevertheless, nowadays insurance against flooding is beingdiscussed by policy-makers and scientists in the Netherlands. Theissue of redistribution of responsibilities for damage betweengovernment and individuals entered the decision making process

9 See Communication strategy Nederland leeft met Water (Netherlands Livingwith water) of the Netherlands Ministery of Transport, Public Works and Watermanagement (http://www.nederlandleeftmetwater.nl/).10 Rijkswaterstaat, an official web site of the Directorate-General for Public Worksand Water Management, http://www.rijkswaterstaat.nl/themas/bescherming_

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172170

of the Dutch Ministry of Public Works and Water Management,when risk management strategies for coastal towns under floodrisk were discussed (Rijkswaterstaat, 2005b). Each of the strategiesincluded a set of options with respect to coastal defense measures,spatial planning, insurance options and risk communication. Threeof the four proposed policy options implied an establishment ofa pool of insurers. For some coastal towns under risk, insurance ifsupported by government seems feasible (Kok, 2005).

An option to extend insurance coverage for coastal flood risks atthe scale of the whole Netherlands is argued to be unattractive dueto the potential high damage from flooding and impossibility tocover all insurance claims (Kok et al., 2002). However, it might beattractive to use in coastal cities under risk, such as discussed byPoelmann commission (2005). For large-scale flooding in theNetherlands the re-insurance schemes should obviously be offered,probably involving the global private insurancemarket discussed inliterature (Bouwer and Vellinga, 2002); this offers new products toeffectively transfer risk. Botzen and van den Bergh (2008) proposeda set of insurance arrangements for the Dutch situation. They arguethat varying risk premiums across risk classes, introducingdeductibles, co- and re-insurance and upper limits on coverage,will not only make flood insurance feasible in the Netherlands, butwill also increase economic efficiency.

Example (flood insurance and decreased housing values): Housingmarkets in the countries where flood insurance is present, reveala decrease of prices due to insurance pressure on individualbudgets. Flood insurance largely explains housing price discountingin flood plains where mandatory insurance has been introduced(Speyrer and Ragas, 1991). Shultz and Fridgen (2001) showed thatflood insurance premiums account for about 80% of a house pricedecrease in a flood plain. They also associated the decrease inhousing prices with increased flood risk awareness. Bin et al. (2008)reported that the value of flood insurance premiums capitalizedinto housing prices. They concluded that flood insurance conveyedrisk information to participants in the coastal housing market.

Conclusion 3: If mandatory flood insurance is introduced, itserves as a flood risk communication device that helps to integrateflood risk in the urban land market. Capitalization of flood insur-ance leads to a house price discounting, leading to less directdamage from flooding to an area as a whole. Prices in vulnerablezones go down, as it should be from the viewpoint of market effi-ciency. Where possible, Dutch water management needs to applymarket mechanisms to reduce flood risk in coastal zones.

5.4. Building on higher elevation levels as a measure to increase riskawareness

Coastal engineering solutions have significantly decreased floodprobability but have made people unaware of coastal risk (Smitset al., 2006; Wesselink, 2007). There are also engineering solu-tions that can decrease risk of flooding without creating a falsefeeling of safety. Currently in the Netherlands, an option of raisingground level for new spatial developments is discussed as one ofthe adaptation strategies to climate change (Bucx et al., 2008).Under the framework of EU project ComCoast there is a pilotproject “Perkpolder”, that aims to develop a new area with housingand a hotel to be built on elevated grounds in the Dutch province ofZeeland8.

Building new developments at higher ground levels does notonly decrease potential damage from flooding. We claim that it alsomakes people living in the neighborhood aware of flooding since itcreates a visible sign that houses are in the vulnerable zone. The

8 http://www.comcoast.org/

traditional methodology to increase individual risk awareness is toeducate people about risk as discussed in Section 5.2; which meanstalking about a problem and making people think about a problem.In combination with traditional hydraulic engineering measures,a traditional way of increasing risk awareness is emphasizing“water as a threat”: basically a negative message that is givena positive twist by the appeal “to livewith water”9. Contrary to that,an alternative engineering measure like building on higher groundlevels, offers the opportunity to communicate a positive message: itis about building, instead of being liable for erosion; it is aboutbeing protected instead of being threatened. Moreover, the secondDeltacommissie (2008, p 52) refers to the idea of building on highelevations as to one of the options to decrease damage due to highwater levels.

Example (lessons from Dutch “Terpen”): Artificial mounds orterpen were used in the Netherlands back in 500 B.C.10 and evenearlier11. Back in the middle ages terpen, on which villages werebuilt, could reach up to 5.5 m above NAP (mean sea level)(Groenendijk, 1997). The fact that in the Netherlands in the pastthere were developments on artificially elevated grounds impliesthat people were aware of flood risk. It illustrates that “higher issafer” is easy to understand, which makes developments on terpena perfect flood risk communication instrument. From this point ofview, terpen not only provide means to reduce actual flood risk, butalso increases the awareness of people about this risk. Individualssearching for houses to buy and being aware that floods occur in thearea, will more likely prefer to choose a location at higher elevation.By choosing between a house in the elevated and safe area or in thelow and vulnerable place, individuals share a responsibility withtheir government with respect to flood risk reduction. Due to theirimage of a cultural and historical heritage, terpen “do not scorebadly in the Netherlands” (Groenendijk, 1997). It is likely that theaverage Dutch person will have a positive attitude towards buyinga house on a terp. Themore developments in the zone vulnerable toflood occur on man-made mounds, the less the potential risk is.

Example (lessons from Hurricane Katrina): The analysis of theconsequences of Hurricane Katrina led to many recommendations.In addition to calls for better governmental reaction, forstrengthening the levees, and for individual preparedness andevacuation plans, there are also recommendations for future urbandevelopment. Waugh and Smith (2006) mention that plans forNew Orleans’ redevelopment assume that new developments arebuilt at higher elevation. Specifically, redevelopment plansincluded elevated residences and an elevated rail system betweenthe city and the airport. According to Rodiek (2007) urban plan-ners should consider landscape reconfiguration by creating topo-graphic terraces of varying height with commercial and residentialdevelopments at the highest level. Naturally, the areas of NewOrleans that were on higher ground had less physical damage(Cigler, 2007).

Conclusion 4: Constructing new developments on highergrounds not only decreases direct damage in the case of flooding. Italso serves as a persistent flood communication signal that remindsthat flooding may happen, thus reviving individual risk awareness.

tegen_het_water/hoe_beschermt_rws/terpen/index.aspx11 An official web site of Friesland province in the Netherlands http://www.friesland.org/red/n_terpen.htm

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172 171

6. Discussions and overall conclusions

Risk of coastal flooding is an important issue in the Netherlands,where the majority of economically developed areas is below sealevel. Risk, definedasaproductofprobabilityanddamage, canbebestreducedbydecreasingbothmultipliers.Whilemostof theeffort so farhas been on reducing probability, in this paper we have been dis-cussing policy options to decrease potential damage.We are trying toachieve that, and consequently to define options to reduce flood riskin the coastal zone in the Netherlands, by affecting microeconomicdecisions in the housingmarket. Changes in flood risk happen due tochanges in individual demand for land in flood-prone areas, whichaffects property prices and consequently potential direct damage.With theDutch government being totally responsible for safety, therearenorealmechanisms for individuals tomakechoices in thehousingmarket that would lead to less potential damage. Thus, there are noinstruments to share the responsibility for flood risk reductionbetween government and individuals, although inpractice individualchoices do affect potential damage. As a result the existing watermanagement practice is a distortion of a free market by a kind ofa subsidy that provides equal costs of living on flood-prone and safeareas.Consequently, theDutchgovernment creates all themotives forurban developments to be skewed toward costal development.

Economic theoryandempirical evidence showthat individual riskawareness affects individual location choices. Property prices andspatial patterns change due to increasing risk awareness, leading toa decrease indirect damage from coastalflooding. However, results ofrecent surveys on coastal flood risk perception show that riskawareness is low in the Netherlands. It may lead to inefficienteconomic landuse outcomes. Thus, forwatermanagement policy it isimportant to findways of increasing individual flood risk awareness.

By reviewing literature and experience in different countries wefound four factors that influence risk perception and havemeasurable effects on property prices: 1) personal experience, 2)risk communication, 3) financial and 4) technical instruments asrisk communication devices. We showed evidence of the effects ofthese four factors on individual risk perception and locationbehavior. Ourmain and general conclusion is that if individual floodrisk awareness is increased, then it is likely that coastal flood risksare integrated in the individual economic decisions at the housingmarket. In this case the housing market reacts by changing spatialpatterns and housing prices leading to less total flood risk in thearea. Thus, we argue that coastal flood risk can be reduced byincreasing individual coastal flood risk awareness.

Personal experience of a disaster increases risk awareness andhas a strong effect on individual location choices and, consequently,can change flood risk in the area. However, being a factor thatincreases awareness, it cannot be used directly by policy-makers.Nevertheless, personal experiences may play an important role incommunicating risk through outreach and education.

Risk communication,financial and technicalmechanisms, can beused by policy-makers as complementary instruments to increaseindividual risk awareness and decrease coastal flood risk. Riskcommunication is an important but a long-termprocess, potentiallyrequiring a change in social norms for a Dutch society which forcenturies believed in collective responsibility for flood safety. Bothinsurance and building on higher elevation have direct measurableeffects on individual location choices and, consequently, on floodrisk reduction. Moreover, these two instruments give individuals anopportunity to act, in contrast to risk communication where anindividual remains passive. They also create a ground for sharedresponsibility between government and individuals for flood riskreduction. There have to be possibilities for individuals to act uponcommunicated risk. Insurance schemes allow individuals to chooseto pay different premiums depending on different risks in different

locations. Similarly, developments built at higher elevationwill giveindividuals a choice to live either at, or below, or above sea level,even in the coastal provinces in the Netherlands.

Mandatory flood insurance is an effective risk communicationtool. It increases economic efficiency of land use in coastal zones.However, it is a forced action and due to the cultural and historicbackground, Dutch citizens might be reluctant to accept it. Anadvantage and at the same time a drawback of this measure is thatit may drive investments to more safe areas. On the one hand itimplies less flood risk in the areas prone to flooding. On the otherhand, however, it may mean a loss of investment for some coastalmunicipalities and lower housing prices for house owners.

Building new developments on higher elevation levels is aneffective engineering instrument which also increases risk aware-ness. In addition, it decreases potential flood damage in theelevated areas. Nevertheless, attention has to be paid to potentialnegative effects on flood risk in the surrounding, not elevated area.Another problem is the financial feasibility of new spatial devel-opments at high elevation. Both concerns can be addressed bydesigning a long-term gradual introduction of surface accretion,which will slowly (in, say, 100 years time) decrease risk in elevatedareas. Regarding to the financial aspects, preliminary estimatesshow that the cost of raising surface levels of planned new devel-opments will constitute only about 2% of total current Dutchbuilding investments (Meulen, 2007). The major advantage of thistechnical instrument to increase risk awareness is that it does notdeter investments from coastal zones. Instead it creates extra space,safe from flooding, in the coastal zone. Moreover, the governmentdoes not need to force people to buy houses that are at higherlevels, it just needs to provide opportunities for them to make thischoice. The more aware people become about coastal flood risk, themore they may choose to prefer locations at higher elevations.Increasing demand for land uphill will provoke land market forcesto come into play driving up housing prices on high ground andmaking it appealing to developers to invest more in the elevatedhouses. The more developments occur on higher ground, the moreindividuals become aware of flood risk and the less potentialdamage is going to occur. This is potentially a self-reinforcing cyclewith a positive effect of decreasing flood risk. In this case not onlythe government achieves its goal of decreasing flood risk, but alsoindividuals and private investors benefit. It has the potential tomake microeconomic behavior working in line with policy-goalsand thus to exploit the capacity of shared responsibilities for floodrisk reduction between the government and the citizens.

Ethical statement

Hereby we would like to approve that the material presented inthe current manuscript has not been published or submitted else-where. The paper is an entirely original work. All the work and/orwords of others, which is mentioned in the paper, has beenappropriately cited or quoted.

References

Alonso, W., 1964. Location and Land Use. Harvard University Press, Cambridge, MA.Baker, F., 1990. Risk communication about environmental hazards. Journal of Public

Health Policy 11 (3), 341e359.Bier, V.M., 2001. On the state of the art: risk communication to the public. Reliability

Engineering and System Safety 71, 139e150.Bin, O., Kruse, J.B., 2006. Real estate market response to coastal flood hazards.

Natural Hazards Review 7 (4), 137e144.Bin, O., Polasky, S., 2004. Effects of flood hazards on property values: evidence

before and after hurricane Floyd. Land Economics 80 (4), 490e500.Bin, O., Kruse, J.B., Landry, C.E., 2008. Flood hazards, insurance rates, and amenities:

evidence from the coastal housing market. The Journal of Risk and Insurance 75(1), 63e82.

T. Filatova et al. / Ocean & Coastal Management 54 (2011) 164e172172

Bo�ckarjova, M., van der Veen, A., Geurts, P.A.T.M., 2009. Flood disaster in theNetherlands: a trade-off between paying for protection and undertakingaction?. ITC Working paper Series, ISBN 978-90-6164-280-0.

Botzen, W.J.W., van den Bergh, J.C.J.M., 2008. Insurance against climate change andflooding in the Netherlands: present, future, and comparison with othercountries. Risk Analysis 28 (2), 413e426.

Bouwer, L.M., Vellinga, P., 2002. In: Beniston, M. (Ed.), Changing climate andincreasing costs - Implications for liability and insurance in Climatic Change:Implications for the Hydrological Cycle and for Water Management. KluwerAcademic Publishers, The Netherlands, pp. 429e444.

Bucx, T., et al., 2008. Staat en Toekomst van de Delta. Achtergronddocument ThemaVeiligheid. Deltares: Delft, The Netherlands.

Buurman, J., Rietveld, P., Scholten, H., 2001. The Land Market: A Spatial EconomicPerspective. In: Stillwell, J., Scholten, H. (Eds.), Land Use Simulation for Europe.Kluwer Academic Publishers, pp. 65e83.

Camerer, C.F., Kunreuther, H., 1989. Decision processes for low probability events:policy implications. Journal of Policy Analysis and Management 8 (4), 565e592.

Chess, C., 2001. Organizational theory and stages of risk communication. RiskAnalysis 21 (1), 179e188.

Chivers, J., Flores, N.E., 2002. Market failure in information: the National FloodInsurance Program. Land Economics 78 (4), 515e521.

Cigler, B.A., 2007. The “Big questions” of Katrina and the 2005 great flood of NewOrleans. Public Administration Review 67, 64e76.

Daniel, V.E., Florax, R.J.G.M., Rietveld, P., 2007. Long term divergence between ex-ante and ex-post hedonic prices of the Meuse River flooding. In: TheNetherlands. in 47th Congress of the European Regional Science Association"Local Governance and Sustainable Development" Paris.

Deltacommissie, 1960. Rapport Deltacommissie. Staatsdrukkerij- en Uitgeverijbe-drijf: ’s-Gravenhage.

Deltacommissie, 2008. Samen werken met water: Een land dat leeft, bouwt aan zijntoekomst. Bevindingen van de Deltacommissie, 2008. Hollandia Printing.

Filatova, T., van der Veen, A., Parker, D., 2009. Land market interactions betweenheterogeneous agents in a heterogeneous landscapedtracing the macro-scaleeffects of individual trade-offs between environmental amenities and dis-amenities. Canadian Journal of Agricultural Economics 57 (4), 431e459.

Fujita, M., Thisse, J.-F., 2002. Economics of Agglomeration. Cities, Industrial locationand Regional growth. Cambridge University Press.

Gayer, T., Hamilton, J.T., Viscusi, W.K., 2000. Private values of risk tradeoffs atsuperfund sites: housing market evidence on learning about risk. The Review ofEconomics and Statistics 82 (3), 439e451.

Groenendijk, H.A., 1997. Terpen: a shared responsibility for a shared interest. In:Willems, W.J.H., Kars, H., Hallewas, D.P. (Eds.), Archaeological HeritageManagement in the Netherlands. Uitgeverij Van Gorcum, pp. 239e255.

Hallstrom, D.G., Smith, V.K., 2005. Market responses to hurricanes. Journal ofEnvironmental Economics and Management 50, 541e561.

IWR, 1998. Empirical Studies of the Effect of Flood Risk on Housing Prices, in IWRREPORT 98-PS-2. Institute for Water Resources, U.S. Army Corps of Engineers,Alexandria, USA. 1e72.

Johnson, B.B., Chess, C., 2006. From the inside out: environmental agency viewsabout communications with the public. Risk Analysis 26 (5), 1395e1407.

Kahneman, D., Tversky, A., 1979. Prospect theory:an analysis of decisions under risk.Econometrica 47 (2), 263e292.

Kaiser, G., et al. 2004. COMRISK Subproject 3: Public Perception of Coastal FloodDefence and Participation in Coastal Flood Defence Planning. Final Report. Kiel,Germany. p. 1e212.

Kok, M., et al., 2002. Risk of flooding and insurance in The Netherlands. In: Wu, B., etal. (Eds.), Flood Defence. Science Press, New York, pp. 146e154.

Kok, M., 2005. Een waterverzekering in Nederland: mogelijk en wenselijk?, Water,O.A., Editor. 2005, HKV: LIJN IN WATER. p. 1e33.

Kolkman, M.J., van der Veen, A., Geurts, P.A.T.M., 2007. Controversies in watermanagement: frames and mental models. Environmental Impact AssessmentReview 27 (7), 685e706.

Krutilla, J.V., 1966. An economic approach to coping with flood damage. WaterResources Research 2 (2), 183e190.

Krywkow, J., Filatova, T., van der Veen, A., 2008. Flood risk perceptions in the Dutchprovince of Zeeland: does the public still support current policies? in FLOODrisk2008. The European Conference on Flood Risk Management, Oxford, UK.

Lave, T.R., Lave, L.B., 1991. Public perception of the risks of floods: implications forcommunication. Risk Analysis 11 (2), 255e267.

MacDonald, D.N., Murdoch, J.C., White, H.L., 1987. Uncertain hazards, insurance, andconsumer choice: evidence from housing markets. Land Economics 63 (4),361e371.

Marchand, M., Crosato, A., . Klijn, F., 2006. Flood Risk Analysis for the River ScheldtEstuary in FLOODsite Report T25-06-01. 1e25.. Delft Hydraulics: Delft, TheNetherlands.

McCluskey, J.J., Rausser, G.C., 2001. Estimation of perceived risk and its effect onproperty values. Land Economics 77 (1), 42e55.

van der Meulen, M.J., et al., 2007. Regional sediment deficits in the Dutch lowlands:implications for long-term land-use options. Journal of Soils and Sediments7 (1), 9e16.

Nicholls, R.J., et al., 2007. In: Parry, M.L., et al. (Eds.), Coastal systems and low-lyingareas. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribu-tion of Working Group II to the Fourth Assessment Report of the Intergovern-mental Panel on Climate Change. Cambridge University Press, Cambridge, UK,pp. 315e356.

Plough, A., Krimsky, S., 1987. The emergence of risk communication studies:social and political context. Science, Technology, and Human Values 12 (3/4),4e10.

Poelmann Commissie, 2005. Advies van de Commissie Bescherming en Ontwikk-eling van Buitendijks gebied in Kustplaatsen Haarlem.

Raaijmakers, R., Krywkow, J., van der Veen, A., 2008. Flood risk perceptions andspatial multi-criteria analysis: an exploratory research for hazard mitigation.Natural Hazards. doi:10.1007/s11069-007-9189-z.

Rijkswaterstaat, 1990. Kustverdediging na 1990 (1e Kustnota); Beleidskeuze voor deKustlijnzorg. Ministerie van Verkeer en Waterstaat: Den Haag.

Rijkswaterstaat, 2002. Towards an Integrated Coastal Zone Policy e Policy agendafor the coast. The Hague: Ministerie van Verkeer en Waterstaat Rijkswaterstaat,Direct Dutch Publications BV.

Rijkswaterstaat, 2005a. Flood Risks and Safety in the Netherlands (Floris). Florisstudy - interim report. Ministry of Transport, Public Works and WaterManagement, Rijkswaterstaat: Delft.

Rijkswaterstaat, 2005b. Hoofdrichtingen voor risicobeheersing in kustplaatsen.Ministerie van Verkeer en Waterstaat, Den Haag.

Rijkswaterstaat, 2005c. Standard Method 2004. Damage and Casualties Caused byFlooding. Ministerie van Verkeer en Waterstaat Rijkswaterstaat.

Rodiek, J., 2007. Landscape planning in hazardous zones, lessons from HurricaneKatrina, August 2005. Landscape and Urban Planning 79 (1), 1e4.

Rosen, S., 1974. Hedonic prices and implicit markets: product differentiation in purecompetition. The Journal of Political Economy 82 (1), 34e55.

Shultz, S.D., Fridgen, P.M., 2001. Floodplains and housing values: implications forflood mitigation projects. Journal of the American Water Resources Association37 (3), 595e603.

Sjoeberg, L., 2000. Factors in risk perception. Risk Analysis 20 (1), 1e11.Slovic, P., 1987. Perception of risk. Science 236, 280e285.Smith, K., 2001. Environmental Hazards: Assessing Risk and Reducing Disaster.

Routledge Physical Environment Series. London.Smits, A.J.M., Nienhuis, P.H., Saeijs, H.L.F., 2006. Changing estuaries, changing views.

Hydrobiologia 565 (1), 339e355.Speyrer, J.F., Ragas, W.R., 1991. Housing prices and flood risk: an examination

using spline regression. Journal of Real Estate Finance and Economics 4 (4),395e407.

Tatano, H., Yamaguchi, K., Okada, N., 2004. In: Okuyama, Y., Chang, S. (Eds.), Riskperception, location choice and land-use patterns under disaster risk: long-term consequences of information provision in a spatial economy, in Modellingspatial and economic impacts of disasters. Springer, pp. 163e181.

Terpstra, T., Gutteling, J.M., 2008. Households’ perceived responsibilities in floodrisk management in TheNetherlands. International Journal of Water ResourcesDevelopment 24 (4), 551e561.

Van den Berg, A., Jacobs, M., Langers, F., 2002. Beleving kustveiligheid, in Alterra-Rapport 583. Alterra, Research Instituut voor de Groene Ruimte, Wageningen.p. 1e98.

Varian, H.R., 1992. Microeconomic Analysis. W. W. Norton. 563.Veraart, J.A., et al., eds. Climate Changes Spatial Planning. Introduction to the Dutch

National Research Programme. CcSP-Brochure. 2007: Wageningen, TheNetherlands.

Vliz, 2005. Europese duurzaamheidsindicatoren voor kustgebieden in Neder-land:een eerste inventarisatie. Vlaams Instituut voor de Zee Ostende, Belgium.

Wagner, K., 2007. Mental models of flash floods and landslides. Risk Analysis 27 (3),671e682.

Waugh, W.L., Smith, R.B., 2006. Economic development and reconstruction on theGulf after Katrina. Economic Development Quarterly 20 (3), 211e218.

Wesselink, A.J., 2007. Flood safety in the Netherlands: the Dutch response tohurricane Katrina. Technology in Society 29, 239e247.

Wet op de Waterkering. 1995. Den Haag, Editor. Ministerie van Verkeer en Water-staat, Ministerie van Justitie.

Williams, B.L., et al., 1999. Risk perception in context: The Savannah river sitestakeholder study. Risk Analysis 19 (6), 1019e1035.

Wu, J., Plantinga, A.J., 2003. The influence of public open space on urban spatialstructure. Journal of Environmental Economics and Management 46 (2),288e309.