Land Use Policy 47 (2015) 6677

Contents lists available at ScienceDirect

Land Use Policy

jo ur nal ho me pag e: www.elsev ier .com/ locate / landusepol

Negoti arbrown he

B. Gluma nb,c

a Eindhoven Un 3, 560b Radboud Univ HK Nc University of

a r t i c l

Article history:Received 24 February 2014Received in revised form 8 March 2015Accepted 27 March 2015

Keywords:Browneld redPublicprivateGame theoryFuzzy Delphi m

The redevelopment of a browneld can provide a range of societal, environmental but also economicbenets for a number of entities. In the Netherlands (and elsewhere), publicprivate partnerships arecommon practice for such projects, because of two main reasons. First, limitations to public funding haveled governments to invite the private sector into various long-term arrangements for capital-intensiveprojects. Second, a comprehensive approach for the whole browneld area may be more efcient and

Introductio

Several d(CABERNETbrowneld used or devbe partiallyor contamiimmediate authors (CaGanser and

CorresponE-mail ad

w.f.schaefer@t

http://dx.doi.o0264-8377/ evelopment (BR) partnership (PPP)

ethod (FDM)

protable, compared to piecemeal development via interventions by individual owners. This article inves-tigates, with respect to browneld redevelopment, the interaction behavior of two key parties in formingpartnerships: the municipality and a private developer. It is assumed that, apart from their mutual inter-est to redevelop the browneld area, they will have different interests as well. In order to indicate theirspecic interest and the negotiation outcome regarding the forming of a public private partnership, thispaper makes use of an experimental game theory approach. Three specic negotiation issues were ana-lyzed in our research: a building claim, future land use and reparcelling of the land. In addition, this papersuggests an eight-step procedure to conduct a game theoretical experiment. A survey was conducted inorder to gather the required data for the experiment. The data have been used to estimate the payoffsvariations between the two key parties in the mentioned negotiation games. Finally, by comparing subgame perfect Nash equilibrium generated game outcomes and direct expected outcomes of respondents,this paper experimentally proves that the game theoretical analysis provides a valid representation ofa real world browneld redevelopment negotiation within the Dutch institutional-economic context.The outcome of the experiment conrms the Dutch tradition of public private partnerships in urbandevelopment practice, with public and private bodies willing to share nancial risks and returns in theseprojects.

2015 Elsevier Ltd. All rights reserved.

n

enitions for a browneld can be found in the literature, 2002; Yount, 2003). This paper uses the following: Asite is any land or premises which has previously beeneloped and is not currently fully in use, although it may

occupied or utilized. It may also be vacant, derelictnated. Therefore, a browneld site is not available foruse without intervention (Alker et al., 2000). Numerousrroll and Eger Iii, 2006; Chen et al., 2009; De Sousa, 2002;

Williams, 2007; Lange and McNeil, 2004a,b; Wang et al.,

ding author. Tel.: +31 402472373.dresses: b.glumac@tue.nl (B. Glumac), q.han@tue.nl (Q. Han),ue.nl (W. Schaefer), e.vanderkrabben@fm.ru.nl (E. van der Krabben).

2011) have argued that the redevelopment of a browneld canprovide a range of economic, social, and environmental benets.Leaving brownelds unmanaged brings a potential loss of economicopportunities to the community in which they are located.

In most cases, a browneld redevelopment (BR) seeks a formof partnership. A public private partnership (PPP) is a concept fre-quently used in development practice (Koppenjan and Enserink,2009) although a uniform denition is still lacking (Weihe, 2005).PPPs are particularly useful when circumstances are not favor-able for a piecemeal development via interventions by individualowners (Grimsey and Lewis, 2002). In such cases a comprehensiveintegrated approach, with private owners/developers collaborat-ing with the responsible public authorities, may be more efcientand protable. Another important reason for the establishmentof a PPP can be limitations to public funding available, mak-ing a public sector-led redevelopment impossible. This has led

rg/10.1016/j.landusepol.2015.03.0182015 Elsevier Ltd. All rights reserved.ation issues in forming publicprivate peld redevelopment: Applying a game t

ca,, Q. Hana, W. Schaefera, Erwin van der Krabbeiversity of Technology, Department of the Built Environment, Den Dolech 2, P.O. Box 51ersity, Geography, Spatial Planning and Environment Department, P.O. Box 9108, 6500

Ulster, School of the Built Environment, United Kingdom

e i n f o a b s t r a c ttnerships fororetical experiment

0 MB Eindhoven, The Netherlandsijmegen, The Netherlands

B. Glumac et al. / Land Use Policy 47 (2015) 6677 67

local governments to invite the private sector into various long-term arrangements for capital-intensive real estate developmentprojects.

Forming a PPP can be problematic as a consequence of differ-ences in goaddresses tviding variHieminga, 2in BR mayinterest of aby the otheProviding mwhen formintroduces standing offorming PPPoper.

Game thonly few apMost gameopment focwith regard2013) and Sgame theorcapturing iing strategihelp of gamMartnez an2010; Wu enegotiationFor exampllease contrenvironmenvated.

Regardindecision-museful to dredevelopm2011), to cdevelopmeet al., 2008)lic opinion Tam et al., ing decision(Blokhuis e2008; Wanthe latter ries interesand generaoptimally bAlthough grelated to dbeen put onproject. Thnegotiationland; see Sthe game) game-theorsuggesting experiment

Classicalnotion of amaker (e.g. a classical gings based omodeling th

experimentally tested. Usually, an experiment consists of severalphases: description of the game environment, the assumptionsunderlying the game, and estimation of players preferences. Thisexperiment introduces an eight-step procedure. First, the game is

propivideativeectinolutt of tnducithin

of ga for soth gerts.arty. d to pond

expeationendscolleused

expdenttical tionninges, inlicies

suppng thtion s pap

for tg PPg PPgotitep p

urbadentand theorarize

gamportae thementent

theo

n, ins doeby citablg situ

assuheors in wRasm

ecisito thals amongst potential partners. The existing literaturehe general diversication of goals and interests by pro-ous typologies of potential parties (Coiacetto, 2001;006). The preferences of the potential parties involved

vary substantially. It is also possible that the self-n individual decision-maker can be heavily inuencedr parties that are present in a certain decision moment.ore insight into these interactions may be of help

ing new publicprivate coalitions for BR. This paperconcepts of game theory in order to improve the under-

the interactions among two key decision-makers in for a BR project: the municipality and a private devel-

eory has been applied in many elds of research, butplications can be found in urban development practice.-theoretical applications with respect to urban devel-us on negotiations, applying game-theoretical concepts

to the interaction of players. Samsura et al. (2010,amsura and van der Krabben (2011, 2012) have usedy to model negotiation processes with respect to valuen land and property development. In addition, pric-es with respect to land use have been modeled withe-theoretical concepts (Forester, 1987; Ma et al., 2007;d Henrquez, 2007; Mu and Ma, 2007; Sibdari and Pyke,t al., 2014; Zellner et al., 2009). Modeling this kind ofs has proved to be able to generate practical advice.e, Pfrang and Witting (2008) have demonstrated howact negotiations can be smoothened and how a socialt between the tenant and the landlord can be culti-

g the application of game theory with respect toaking processes for BR, analyzing negotiations may beecide how to allocate cost and benets in browneldent negotiations (Liang et al., 2008; Wang et al., 2007,ompare the costs and benets of BR and greeneldnt, in order to support BR with effective policies (Liang, and to evaluate the potential conict in engaging pub-in redevelopment processes (Tam and Thomas, 2011;2009). Most applications, however, refer to improv--making processes in establishing various partnershipst al., 2012; Sounderpandian et al., 2005; Walker et al.,g et al., 2008; Youse et al., 2007, 2010). Ultimately,esearch helps to develop decision-support tools, clar-ts, identies tradeoffs, recognizes party satisfaction,tes optimal solutions, preparing a decision maker toenet from the negotiation (e.g. Youse et al., 2010)ame theory can help to negotiate favorable conditionsifferent partnerships types, little attention so far has

isolated negotiable issues in forming a PPP for a BRis paper elaborates on three specic issues in theses (building claim, future land use and reparcelling theection Dening the institutional-economic context ofand aims to contribute to the further development ofetical approaches to urban development practice bya formal procedure for applying a game-theoretical.

game theory has been largely criticized due to the homo economicus, a completely rational decision-Camerer, 2003; Raiffa, 2002). Therefore, instead of usingame-theoretic approach, this paper provides the nd-n experimental game theory results. Rather than onlye outcome of the negotiations, the games have been

set in abeen dcooper(2) selgame scontexwas cotions wtypes game)tion, bBR expvate preferrethe resreticalaggregiment

To tool is

Therespontheoreapplicaconceranalysous poaim atreducimenta

Thiarguesforminforminover neeight-sfor anresponnique game summmatedthe ima gamexperivelopm

Game

Ofteprocesenced is a sumakintheorygame tuation2004; 2008).

A d has er institutional-economic environment. This phase hasd into ve separate steps: (1) selecting a game class

vs. non-cooperative and conict vs. common interest;g a game form strategic vs. strategic; (3) selecting aion concept; (4) describing the institutional-economiche game, here it is important to mention that the studyted in the Netherlands; (5) designing the game condi-

the game set environment. Further (6), two differentmes have been assumed (ultimatum and bargainingpecic negotiation issues in forming PPP for BR. In addi-ames are experimentally validated by a survey among

The players in both games are a public party and a pri-In the remaining of the paper, these two parties areas a municipality (M) and a developer (D). To estimateents preferences (7), a standard phase in game theo-riments, the fuzzy Delphi method (FDM) with similarity

method (SAM) has been applied. (8) Finally, the exper- with the analysis of the outcomes.ct the data for validating the results an on-line survey

(Berg Enqute System 2007).eriment explores whether the self-prediction of thes about the game outcome corresponds to the game-predictions. This provides insight in the suitability of the

of game theory in predicting real-world actor behavior BR projects. In addition, based on the outcomes of theterventions can be designed and through them, vari-

may be considered. The eventual new policies wouldorting the cooperation between relevant parties, thuse number of conicts and stimulating the actual imple-of BR projects.er rst explains the basic elements of a game tree andhe implementation of game theoretical experiments inP for BR projects (Section Game theory applications inP for a BR project). Further, Section Designing gamesation issues in forming PPP for a BR project explains therocedure of conducting a game theoretical experimentn development project. Section Data collection ands characteristics reports on the data collection tech-the background of respondents that are used for theetical experiment. Section Game experiment resultss the empirical results of validated game trees and esti-e outcomes. Finally, Section Conclusions concludes onnce of using rigorous procedural steps of conductingoretical experiments and the contribution that suchs can provide to represent real world browneld rede-negotiation.

ry applications in forming PPP for a BR project

urban development the outcome of a decision-makings not only depend on individual choice but is also inu-hoices made by other decision-makers. Game theorye theory to test the behavior of interactive decision-ations (e.g. Neumann et al., 1944). Even more, gamemes that decision-making is always interdependent;y mainly aims to provide a better understanding of sit-hich decision-makers interact (Colman, 1995; Osborne,usen, 2007; Shoham and Leyton-Brown, 2009; Stengel,

on-maker a player in game-theoretical terminologyink ahead and is assumed to devise a strategy based

68 B. Glumac et al. / Land Use Policy 47 (2015) 6677

on the expthe main ogies the pltheir own so. Game ttions in whinterests anand relatedplayers areMu and MOther reseaprovide insrelevant forSounderpan2010).

The basiations for ais based onplayers. Thing PPPs foplayers marepresentinkey playersindividualstext, it hasFinally, gambased on cgame. Contnegotiationtions a useproject.

A game tree

In genersive form. Tplayers actor sequentigames in exno conventexample (Fon top. Thecessor nodeconnect thepossible plamaker (1, 2Alternativethe node is predened named termthat is a rea

Game elements: players, strategies, payoffs, equilibriumsBasic assumptions in game theory include that decision-makers

pursue well-dened, exogenous objectives (they are rational), theyn innite good memory (perfect recall), and they take intot theor (t

absto bet intets: p

playlitar

stralaye

for uctingoice f the,. . .,

eforeally, an btcomthe ache.

cond as etw

cisione payor ths ovencesrdercus trate

playh of ,. . .,size tyersts arncepriumns.

ent

eraltie

stratnformheorly onFig. 1. An example of game tree

ected countermoves of the other player(s). Therefore,bjective of game theory is to determine what strate-ayers, rationally, ought to choose in order to pursueinterests and what the outcomes will be if they doheory is well-suited to describe and analyze situa-ich players have both conicting and supplementaryd interdependency in behavior in urban development

decision-making situations in which two or more involved (e.g. Liang et al., 2008; Mayer et al., 2005;a, 2007; Samsura et al., 2010; Wang et al., 2007).rch has already demonstrated that game theory canight in the negotiation strategies over specic issues

BR (Arentze and Timmermans, 2003; Forester, 1987;dian et al., 2005; Wang et al., 2011; Youse et al.,

c arguments for applying game theory in PPP negoti- BR can be summarized as follow. First, game theory

the premise of relational interdependency betweenis responds to the multi-actor environment in form-r BR projects. Secondly, game theory assumes thatke rational decisions based on their utility functiong their needs and interests. Having in mind that two

in BR (municipality and developer) are a group of responding in a certain institutional-economic con-

to be veried if their decisions are indeed rational.e theory analysis reveals the possible course of actionsertain payoffs, structure and present players in anyrolled changes of these terms might inuence different

outcomes, thus making game-theoretical applica-ful policy or strategy tool in forming a PPP for a BR

and its elements

al, games can be represented in strategic or exten-he main difference between them underlines how the. They can either act simultaneously (strategic form)ally (extensive form). It is important to underline thattensive form are represented with a game tree. There ision, however, how to design a game tree. The following

have aaccounbehavihighlythem tpredicelemen

Thebe a so

Thewhat paimingby selethe cheach oSN = {snis ther

FinThis cble ouby all are attsystem

Thedenemade bthe dedenitcome fsystemprefere

In olysts fomore sby thefor eacs* = (s*1maximthe placonception coequilibfunctio

Experim

Genbe idenples ofsuch igame ting onig. 1) draws the tree downwards starting with the root nodes represent the state of the game while two suc-s represent the move of the player. Lines called edgesm. The line that connects two successors nodes are ayers move or an action (X, Y, a, b, c, d). If a decision-) performs an action, this node is called decision node.ly, an action can be determined by nature: in that casecalled chance node. Then the move or action is random,with the probability. The nodes without a successor areinal nodes. At such a node, every player gets a payoffl number.

often leadsgame. Althothe controlmental tecthe relatioment (Crawstrongly coto gather eexperimentin predictinbeen reliabir knowledge or expectations of other decision-makershey reason strategically). Game theoretical models areract representations of real-life situations, which allow

used to study a wide range of phenomena. In order toraction outcomes a game consists of at least three basiclayers, strategies, and payoffs.ers are the decision-makers; a player n is assumed toy actor who makes decisions as a single decision body.tegy sn is a plan of all possible actions An = {an}, deningr n might do in any given situation during the game,tility maximization. All players make their own choices

a strategy, but the result for each player depends onof the other player. The resulting set of strategies for

N players in the game is denoted as a strategy prolesN}. If the game has only two players, a strategy prole

a pair of strategies with one strategy per player.the payoff for player n is denoted as Un (s1,. . .,sn).e dened as a number associated with each possi-e resulting from a complete set of strategic selectionsplayers in a game. Generally, higher payoff numbersd to outcomes that are better in the players rating

junction of chosen strategies and related payoffs isthe outcome of the game. A clear distinction has to beeen the concepts of outcome and payoff; an outcome is, if any, arrived at by the players collectively, while theoff of an outcome for a player is the value of that out-e player. Because players will have different valuationr the set of possible outcomes, and hence have different

over the outcomes, this is where conicts can arise. to predict the outcome of a game, game theoretical ana-on possible strategy proles and on selecting one orgy proles as a reection of the most rational behaviorers. A strategy prole that consists of the best strategythe n players in the game is dened as an equilibrium*n). Players choose equilibrium strategies in trying to

heir individual payoffs. In order to nd equilibriums, most preferred strategies should be dened. Solutione suitable for dening such preferred strategies; a solu-t F: {S1,. . .,Sn, U1,. . .,Un} s* is a rule that denes an

based on the possible strategy proles and the payoff

al game theory

ly speaking, the emergence of game experiments cand by a need for empirical information about the princi-egic behavior and the ability of experiments to provideation (Crawford, 2002). The predictions of classical

y are very sensitive to the structure of the game. Rely- the existing (data) input of a certain research context

to an unobserved or uncontrolled structure of theugh experiments often share some of these problems,

and observation given by applying modern experi-hniques provides a notable advantage in identifyingnships between strategic behavior and the environ-ford, 2002). Still, theory and experiment may have

mplementary roles. While theory provides a frameworkmpirical data and interpret the respondents behavior,s indicate which parts of the theory are most usefulg and identifying behavioral parameters that have notly determined by theory (Crawford, 2002).

B. Glumac et al. / Land Use Policy 47 (2015) 6677 69

Designing games over negotiation issues in forming PPP fora BR project

Besides the argued applicability, detailed insight is requiredwhen an anment in a cattention toiment. The when applyment problthe game inParticularlyconsists of detailed ins

First, themajor taskssible game that playersmany solutever, only othat ts bebe to descridesigning ttext. Withoexperimentsuggests thtally validatpredictionsways by appbeen used, ia relatively in this sectiand direct gproof that inthat the expBR negotiat

Identifying a

Games cgames. BotLeyton-Brotions in whiplayers aimjoint protsindividual ption of playan interactiprots. Howbinding agrvate playerpotential beas non-coobehave townation is to

Moreovemon interesthe interactoutcome. Thily, in a conopposed orchoose an obe the interinuence onother. As anoutcomes fo

Given the explanation of four possible classes of games, thispaper selects a non-cooperative, conict game as a credible mimicof the negotiation issues in forming PPP for BR project betweena municipality (player M) and developer (player D). In particu-

ee nlande weesene bas repd rep

ing a

ent repretweulta

Givepresxtenpreventtial atedlayer

conf usua2007lmoserefotimeur statinger knistor, futummbben

formd re

assu

a so

lutio playo-calced

extenes Nriums. Theis pa

why thn noat th

in t hav

2 chtwee

2 chalyst wants to implement a game-theoretical experi-ontext of BR. Therefore, this section provides special

the procedure for conducting a game theoretical exper-purpose of setting these dened methodological stepsing a game theoretical framework for an urban develop-em should be regarded as a constraint to pre-inuence

order to reduce the biased assumptions of an analyst. in this study, the construction of a game experimentthe mentioned eight steps. This section provides aight for each of these steps.

game class and form needs to be identied. Here the for an analyst are identifying one out of the four pos-class combinations and identifying if it is more suitable

act simultaneously or sequentially. Further, there areion concepts that might be used as a reference. How-ne proper game solution concept has to be adopted

st to the problem itself. Further necessary step wouldbe the institutional-economic context of the game andhe statistically valid game conditions within the con-ut dening and presenting them to the respondents,al results would not be meaningful. This article alsoat the assumed game structures should be experimen-ed as well. Estimating the respondents preferences and

of the game outcomes can be performed in numerouslying various methods. In this article FDM with SAM hast enabled experimentally valid results although havingsmall sample of the respondents. A nal step describedon is interpreting the respondents feedback with SPNEame outcome predictions that provided on one hand the

this experiment player act rationally and on the othererimental game theory can be successfully applied inion.

game class: non-cooperative conict game

an be classied as cooperative and non-cooperativeh are used to study players interaction (Shoham andwn, 2009). Cooperative game theory deals with situa-ch groups of players already agreed to cooperate. These

for coordinating their actions, eventually resulting in. Because these joint prots often exceed the sum of therots, cooperative game theory deals with the interac-ers within binding agreements such as PPP. In this case,on could address the division of the PPPs expenses andever, often some of the interactions are not part of a

eement. This might be the case when public and pri-s negotiate about the division of risks or developmentfore creating a PPP. Such negotiations are characterizedperative games, which model how actors strategicallyard each other when the cost of bargaining and coordi-o high.r, game theory distinguishes between conict and com-t games (e.g. Bowles, 2004). In a common interest game,ions have a pattern of a trafc jam that is an overall poorerefore, it is benecial for everyone to avoid it. Contrar-ict game, the interests of several decision-makers are

only partly coincide. Each decision-maker will usuallyption contributing to his own interest, which has not toest of others. For example, negotiating about the players

a future land-use means more for one and less for the aside, these individual decisions could result in worser all players compared to a coordinated decision.

lar, thrfuture that arto reprand thissue iuse angame.

Identify

As mcan beence bact simform). form rein the ein the ously msequenaggregboth pbeforecesseset al., ment aIt is thsingle

In onegotia playplete hclaimsered coder Kralack inland antheory

Finding(SPNE)

A sowill beis the sintroduform.

In producEquilibplayeroff. Thactionsson whdecisioaction chosenactionsplayerand beplayeregotiation issues have been observed: building claim, use and future reparcelling of the land. Two gamesll known in the game theoretic literature are selectedt these negotiation issues, respectively, the ultimatumrgaining game, where the building claim negotiationresented by a ultimatum game and the future landarcelling of the land negotiation issues by a bargaining

game form: extensive form

ioned in Section A game tree and its elements, gamesesented in strategic or extensive form. The main differ-en them underlines how the players act. They can eitherneously (strategic form) or sequentially (extensiven the time perspective of the negotiation, in strategicentation the players act in a single time sequence whilesive form players are being aware of all moves of playersious time sequences. The negotiation between previ-ioned player M and D ts better with the concept of thegames, because of two reasons. First, both players are

entities. This implies a multitude of opinions betweens thus resulting in a structured decision-making processronting each other. Secondly, urban redevelopment pro-lly consist of several phases (e.g. Hieminga, 2006; Miles; Peiser and Frej, 2003). This makes urban redevelop-t by denition a time consuming multi-phase process.re unlikely that the negotiation will be resolved in a

sequence negotiation.udy we assume perfect information for M and D when

over forming a PPP. Perfect information implies thatows the status of the game and therefore the com-y of the game up to then. Negotiating about buildingre land use and reparcelling of the land can be consid-on practice for both M and D (Samsura et al., 2010; Van

and Jacobs, 2013). On the other hand, the players mayation with respect to for instance market conditions andal estate prices. Nevertheless, this application of gamemes perfect information for players.

lution concept: sub game perfect Nash equilibrium

n concept is a formal rule for predicting how the gameed. The central concept of non-cooperative game theoryled Nash equilibrium that is named after John Nash who

this solution concept in the 1950s for games in strategic

sive games, a backward induction process alwaysash equilibrium, also called Sub game Perfect Nash

(SPNE), since it represents strategy prole for both optimal play of any player should maximize his pay-

yoff can be decided irrespectively of the others playersen observing the players last action. That is the rea-e backward induction process always starts with thede closest to the leaves. A player naturally chooses theis node giving him the maximum payoff. An action ishis way for every decision node when all subsequente been decided. In the example given above (Fig. 1),ooses between the payoffs 3 and 0 on the left side noden payoffs 2 and 0 on the right side node. Therefore,ooses actions a and c (indicated by the two arrows).

70 B. Glumac et al. / Land Use Policy 47 (2015) 6677

Going backward in time, player 1 chooses action Y (also indicated byan arrow) assuming the previously described behavior of player 2.The action selected by the backward induction is not always unique,since it is possible that more than one action provides the maximalpayoff. Thisin a compleThe result oextensive fostrategy pro

Dening the

To set upcontext. Forerty develoet al., 2010)(3) PPP moextensive ato Van der characterizthe ownersservice it anthat acquire

This stumodel as ain the initiaowns the lacompany (Jand the devlocation. Thprivate devright of rsting claim; innewly formrate for fur

The gamissues that JVC or not. Tsingle BR prthe municipformed, thebuild-on buproduct tha

The playthe JVC or n(actual playinition (Alkonly to thenot). Third,hectares. Fu(affecting this assumed Therefore, t

Setting a ga

The desset of the retion settingproject) tharigorous degame experexogenous future landbeen used:

Table 1Condition state: attributes allocation and the treatment combinations.

Treatment combination Attributes allocation

tion: is

ssibilpublieddetegrainist

of corgy wding

se as et rger,013)

thre is anation

an en fac

vali thrent deign r

1.) snatio

assuther dentation. Fors of

Tabl

ssum

enton oject.

to sere. They both investigate negotiation or more specic how

ning occurs between two rational actors. First, it is impor- underline for both games that they are non-cooperative,e players estimate the costs of bargaining and cooperationo the binding agreement. This leads to a players decisionblish a PPP or not. Second, both games are conict gamese the players larger inuence on a future land-use, for exam-eans more for one and less for the other player. As an aside,ould assume that the partnership had already been agreedn we would investigate the interaction under binding agree-nd therefore classify these games as cooperative, conict procedure denes every action in all decision nodeste game tree and describes a strategy for each player.f backward induction is therefore a strategy prole. Inrm games, it is then also possible to state that SPNE andle obtained by backward inductions are synonyms.

institutional-economic context of the game

the game, we rst dene the institutional-economical this purpose, we refer to the common land and prop-pment models applied in the Netherlands (Samsura: (1) public land development, (2) building claim model,del and (4) private land development model. For annalysis of the characteristics of these models, we referKrabben & Jacobs (2013). Each development model ised by the initial situation on the land market related tohip structure, the dened parties that acquire the land,d reparcel the land into building plots, and the parties

the building plots.dy addresses only the building claim development

specic type of PPP. In this PPP development model,l situation the original owner or a private developer

nd. Subsequently, the land is acquired by a joint ventureVC) a type of PPP that is formed by the municipalityeloper that already owned (part of the) land on the BRe land is serviced and reparcelled by the same JVC. Theeloper participating in the JVC has usually claimed the

buyer of the building plots, against a xed price (build- return for his willingness to take part in the JVC) of theed building plots (see Section Sample size and responsether explanation).es presented in this paper address certain negotiationnally inuence the decision of both players: to form thehe municipality invites a developer to form a JVC for aoject. In order to simplify the game, it is assumed thatality also owns part of the land on the BR site. When

JVC will service the land and reparcel it into ready-to-ilding plots. The building plots can be seen as the nalt is produced by the JVC.ers involved in the game base their decision to formot on several general conditions. First, the respondentsers) should be familiar with the proper browneld def-er et al., 2000). Second, the decision problem is limited

initiative phase of a BR (to produce building plots or the size of a browneld is in the range of one to tenrther, future land use is assumed to be a mix land usee price of the building plots that can be sold). Finally, it

that the decisions that will be made are country-specic.his study focuses only on the Netherlands.

me condition state

criptive part of the game is composed from a limitedlevant attributes that represent changing BR negotia-s. An attribute is the characteristic of a product (e.g. BRt consists of various levels (Louviere et al., 2000). Thislineation is made for the purpose of the statistically validiment named game condition. In order to describe theconditions of the BR negotiation upon building claim,-use, and reparcelling of the land, four attributes have

1 2 3 4 5 6 7 8 9

Locaimityacceand

Embbe in

Admtions

Syneroun

The(Adamand Beet al., 2sists ofA levelnegoti

Forfractiotisticalhavingdiffererial des(Tablecombidesigneach oresponnegoti(Fig. 5)consist0), see

Game a

As msentatiBR prochoicestructubargaitant tobecausprior tto estabecausple, mif we won, thament aL E AS S

0 0 0 00 1 1 20 2 2 11 0 1 11 1 2 01 2 0 22 0 2 22 1 0 12 2 1 0

refers to the proximity and accessibility of a site. Prox-a distance to the key city locations (e.g. CBD); andity refers to how good the access is to the site by carc transport.dness: the extent to which the redevelopment area canted into the existing urban fabric.

rative support: refers to the transparency and percep-ntinuity in governance, politics and the administration.ith surrounding users: the extent to which the sur-

area inhabitants/users support the redevelopment.

ttributes emerged as signicant in related studiesal., 2001; Glumac et al., 2011; Lange et al., 2013; Page

2006; Peiser, 2007; Syms, 1999; Thomas, 2003; Wang. Each attribute that is used in the game condition con-e levels (e.g. location excellent, moderate, and poor).

ordinal value of an attribute that enables changing BR settings.xperimental design, this study generated orthogonaltorial design (Hahn and Shapiro, 1966) to secure the sta-dity of this experiment. Given these four attributes alle levels, the full factorial design would suggest 34 = 81cision moments. Instead, an orthogonal fractional facto-educed the number of treatments. The following designhows the attributes allocation over the nine treatmentns (Hahn and Shapiro, 1966). Predened orthogonalres that the attributes levels are unconfounded withthus providing a statistical validity. As an aside, each

responds to only one treatment combination or one setting when rating and selecting the game outcomes

example, the treatment combination in the rst row (1)the lowest levels for each of the four attributes (0, 0, 0,e 1.

ption

ioned above, two games are assumed to be a valid repre-f three different negotiation issues in forming PPP for aTherefore, this subsection provides argumentation for alect these two non-cooperative conict games and their

B. Glumac et al. / Land Use Policy 47 (2015) 6677 71

games suchLeyton-Bro

The ultimatuWe have

negotiationDutch conta situation on a (re)devits plot of ltion that thbuilding plobuilding plooriginally otime the ordevelopmeerty develoon two assuservicing it and therefopensation fhouses (or be protabplots from tit attractiveThis buildinmunicipalitefcient anthe locationused. Two oin the struceach optionation issue a building cbasic ultimcounteroffeteroffer wowould ask however, wissue. The nwhere NBC is describedfer extensiothe case of are possiblecounteroffeone out of fThis implieopment mothe construdevelopmecreate a PPPthe previou

Player description, information, and strategy. The ultimatum gameis regarded as a 2 2 game. That stands for a game where there areonly two players each having only two strategies. This ts with theproposed building claim ultimatum game since there are only two

als aM) a

M iscontrmat eas knoers kies ofs; (4fs.

2 illto plC) oryer Dn nod.s proions

diff fromn of yer Dey d

adab payd (FD

expzzy

rgains gam

respellingocesject.

the n is rding garddecis

devent

llingndevearcemedi

Theshedce, DationtancFig. 2. Ultimatum game: building claim

as cake-cutting or fair division games (e.g. Shoham andwn, 2009).

m game: building claim assumed that the ultimatum game can represent the

on one issue: the building claim negotiation. In theext of land development, the building claim refers toin which a private developer that owns a piece of landelopment location agrees with the municipality to selland to the municipality (or the JVC) under the condi-e municipality will service and reparcel the land intots and offers the private developer the rst right to buyts (the building claim; equal in size to the plot of land itwned) against a price that has been agreed upon at theiginal plot of land was sold (Samsura et al., 2010). Thisnt model has been common practice in land and prop-pment in the Netherlands since the 1990s. It is basedmptions: (1) acquiring (previously built-on) land andand selling it again as building land is not without risksre, private developers require a building claim as a com-or sharing this risk with the municipality; (2) buildingother properties) on the building plots is expected tole and therefore, in general, the demand for buildinghe private sector exceeds the supply. The latter makes

for private developers to have the rst right to buy.g claim model is assumed to be favorable both to the

y and the private developer, because it allows for a mored protable comprehensive development strategy for. To select or design a game, the following postulate isptions of a negotiation issue need to be accommodatedture of the game. This postulate is addressed by setting

as an action in the game. Therefore, the rst negoti-regarding the building claim has two options: to offerlaim (BC) or not (NBC) (Fig. 2). As an alternative to theatum game it is also possible to assume an altering orr ultimatum game. In this case, that altering or a coun-uld appear at the decision node of player D where hefor BC after player M offers NBC. Such an assumption,ould be redundant due to the nature of this negotiationegotiation issue is regarded as dummy ordinal variablehas less value. Having in mind that this negotiation issue

by the ordinal dummy variable, altering or counterof-n to the game structure would only be appropriate ininterval or ratio variable where quantitative divisions

propospality (playerDutch

Infoperfecplayerall playstrategpay-ofpay-of

Fig.offers able (BM, pladecisioreache

Thible actactionschoosethe plathe plaNBC. Ththe re

Themethosectionwith fugame.

The baThi

and D:reparction prBR proaffectslocatioof builsion republic privateimplemthe wiand rethe rep(high, ment).establiinuennegotiFor ins. Also, it would not be realistic to present solely thisr as a new negotiation round since a building claim isour urban development models (Samsura et al., 2010).s that player D would also include three other devel-dels at this decision node. Further, this would lead toction of the game where players choose the optimalnt model (Samsura et al., 2010) instead of deciding to

or not. As an aside, the building claim game is set insly described game condition.

smaller shaSimilar t

gaining gamsimilar to tues of an astructure otied for thof the playpotential tovailable: BC and NBC. The two players are the munici-nd developer (D) that would potentially form a JVC. The

the initiator of the game, which is consistent with theext of public land development (Samsura et al., 2010).tion available to the players is dened as follows: (1)ch player knows his position in the game tree and allw the previous moves of the other players; (2) certain now the payoff of playing a particular strategy given thef other players; (3) asymmetric players have different) incomplete a player does not know the other players

ustrates the game. At the rst decision node, player Mayer D a deal in which a building claim is either avail-

not available (NBC). For both possible actions of player can accept (a) or reject (r) the deal on the succeedingde in the game. The game stops when the end nodes are

cedure practically explains the complete plan of possi-(strategy by denition) of the players M and D. Theirer and a branch represents each action. Player M can

two possible actions: BC, NBC. These two actions denepossible actions Am = {BC, NBC}. Similar, the actions of

are: a, r as a reaction on the BC and a, r as a reaction onene the Ad = {aBC, rBC, aNBC, rNBC}. Note that, because ofility of Fig. 2 the actions aBC and aNBC are marked as a.offs are estimated empirically with the fuzzy DelphiM). Rather than providing a game solution, this sub-lains the game design. In Section Estimated resultsDelphi method we further analyze the outcome of the

ing game: future land use and reparcelling of the lande addresses two other negotiation issues between Mectively inuence on the future land use and on the

of the land. Both issues are relevant to the negotia-s, because they may inuence the protability of the

The future land use allowed on the location directlyprice per m2 of the building plots, while the way theeparcelled in building plots may inuence the total sizeland that can be sold (e.g. Zellner et al., 2009). The deci-ing future land use and the reparcelling of the land is aion (as part of a decision on the land use plan), but theeloper may be able to inuence it, because for the actualation of the land use plan the municipality depends oness of the private developer to buy the building plots

lop the site. The inuence over the future land-use andlling of the land has been expressed in the ordinal scaleum and low inuence of developer in future develop-

outcome of the negotiation indicates if JVC would be or not. For instance, if M would allow D only minor

might reject to join JVC. As an additional relevancy, this can be linked to the division of the future JVC shares.e, if M would allow D high inuence, D might accept are in the returns from the JVC.o the ultimatum game, this game can be seen as a bar-e that has one additional proposal in the table. Also

he previous game, the postulate is that the ordinal val-ttribute (or variable) need to be accommodated in thef the game. Therefore, three inuence values are iden-is issue: high (H), medium (M), and low (L). The levelers (e.g. developer) inuence (H, M, L) expresses the

adjust the land use ratio within the mixed-use zoning.

72 B. Glumac et al. / Land Use Policy 47 (2015) 6677

F

The (H) inregulated bthe size andveloped. Minuence oFig. 3 descrpresented uassume thedelays. Thisthe four methe institut

Player descrgaining gaminformation

At the the deals Huse and repof player Mnodes. The that the higwas the initplayer M ofdeal on thethe highest M, then D cor reject (r)decision nois similar fonode. In anreached.

Similar tthe compleplayers M aaction. Playgame: H, Mactions Amreaction onreaction onfor the play

Fig. 3 alsfuzzy Delph

Game tree v

As a secothe validatstructured q

For both games, every decision node is textually described. Forexample, the rst question in the experiment (Fig. 4.) correspondsto the initial decision node in the building claim game (Fig. 2.). In

ample, the description is: A municipality (as initiator) nego-ith ationsiblele-chif all

not. be thng th

a resn nor oth

a mihe idd me v

re.

feren

nadentodet, thutco

condn Ids des, eacpturle, Fips th

to th D ple in

towe w

pref it is putcorobas areed wreforg. 5)ax, men

ollecig. 3. Bargaining game: future land use and parcellation

uence means that a player can carry out any land usey a mixed-use zoning plan and completely determines

the shape of any parcel in the land that will be rede-edium inuence (M) grants a developer less and lownly minimal possibilities to adjust the land use ratio.ibes the bargaining game. Contrary to the previouslyltimatum game, in the bargaining game it is possible to

existence of the counteroffers that might cause costly is appropriate since this negotiation can occur in any ofntioned urban development models (Section Deningional-economic context of the game).

iption, information, and strategy. The players in the bar-e are again municipality (M) and developer (D). Their

is set to be the same as in the previous game.rst decision node, player M offers to player D one of, M, or L linked to different inuence on the future landarcelling of the land. For each of the possible actions, player D can react differently on each of the decisionstructure of every sub-tree has been designed in a wayhest level of inuence can be reached no matter whatial offer from the player M (H, M, L). For example, whenfers H, then player D can only accept (a) or reject (r) the

succeeding decision node in the game. This is becauselevel of inuence is already offered. However, if M offersan ask for the highest inuence (h) or either accept (a)

the offer. If D ask for (h) there is one more succeedingde where M can accept (A) or reject (R) that offer. Thisr the branch when the player M offers L at the initialy case, the game stops when the end nodes have been

o the previous game, this procedure practically explainste plan of possible actions related to the strategies ofnd D. Their actions differ and a branch represents eacher M may choose from nine possible actions in this

this extiates wnegotiathe posmultipverify rect orwouldmeaniondly,decisioanswetion ofcase, tselectethe gamstructu

The pre

Therespon end n

Firsgame oa game(Sectioand it iSecondtion caexampand stospondsplayeroutcomerenceoutcomtree. Apayoffgame omost pplayerexplain

Thevey (Fi(min, mrequire

Data c, L, A, R, A, R, A, R. These actions dene the plan of all. Player D may also choose from nine actions: a, r as a

the H. Then h, a, r, as a reaction on M, and h, m, a, r, as a L. Together they dene the plan of all possible actionser D, Ad.o presents the payoffs. In Section Estimated results withi method we further analyze the outcomes of this game.

alidation

nd part of the game experiment, this study introducesion of the game trees (Fig. 4.), with help of a semi-uestionnaire with multiple-choice answers.

To collecsemi-structbeen used method (FDdictions of traditional DcontributedIshikawa etFDM mainlDelphi quetions and thproblem tosion makin developer over the building claim. What are the possible options? Given this description, every respondent states

actions (branches) at that decision node by lling in theoice answers. In each question, a respondent can rstassumed actions at the specic decision node are cor-Therefore, the ideal validation of the game tree structureat all respondents selected all of the multiple answers at they exist in real market negotiation situations. Sec-pondent checks if there is a missing action(s) at certainde. If an action is missing, a respondent chooses theer. Additionally, a respondent can provide the descrip-ssing action in the text line below every question. In thiseal validation would be that none of the answers wereeaning that there are no missing actions. The results of

alidation are presented in Section Validated game tree

ces and predictions of the game outcomes

l part of this experiment concerns the estimation of thes preferences and predictions over the game outcomess in both games (Figs. 2 and 3).e respondents are asked to give their estimation on ame in a specic decision moment. This moment is set asition within the given institutional-economical context

entifying a game class: non-cooperative conict game)cribed with specic set of attributes and levels (Table 1).h game outcome is textually described. The text descrip-es all lines of action leading to the game outcome. Forg. 5 describes outcome 1 as: A developer rejects the offere negotiation when building claim is available. This corre-e following line of actions: player M plays BC, and then

ays r (Fig. 2). Same logic is used to describe every game a game tree. Finally, every respondent provides its pref-ard a certain game outcome and its prediction that thisill occur. This is repeated for every outcome in a gameerence is regarded here as a payoff and by knowing theossible to apply backward induction and estimate SPNEme. By comparing the SPNE solution with the estimatedble outcome, it is possible to investigate whether the

rational and if market behavior in a BR project can beith the application of the game theory.e, there were two estimations per outcome in the sur-. Further on, each estimate requires three input valuesand optimal) from a respondent in order to meet thets of FDM.

tion and respondents characteristics

t the data two approaches have been employed. First, aured questionnaire with multiple-choice answers hasto validate a game tree. Second is the fuzzy DelphiM) and it is used to estimate preferences and pre-

the game outcomes. FDM derived from combining theelphi method and fuzzy set theory. Various researchers

to the origin of this approach (Hsu and Chen, 1996; al., 1993; Murray et al., 1985; Noorderhaven, 1995).y aims to improve the characteristics of the traditionalstionnaire that has the tendency that both the ques-e answers are indistinct. Additionally, there is a notable

solve the fuzziness in expert consensus in group deci-g. These two key issues resulted in the proposed (Hsu

B. Glumac et al. / Land Use Policy 47 (2015) 6677 73

rimen

and Chen, 1Hsu and Chthe group cber becauseand Yuan, 1with fuzzy

The dataBerg EnqutApril to Sepsize, responistics.

Sample size

The exprespondentexperimentof respondein classical

The groupendent dedevelopmerespondent

perimse raond

y theferedenteted ponsteris

f res

foll2). Ths a r

ince numbzy DFig. 4. Game validation expe

996) similarity aggregation method (SAM). Although,en (1996) used a trapezoidal fuzzy number to estimateonsensus, this study uses the triangular fuzzy num-

it is the least demanding for the respondents (e.g. Klir995). Results are discussed in Section Estimated resultsDelphi method.

in this experiment is collected with the survey tool e System 2007; the period the survey took place wastember 2011. This section reports briey on the samplese rate, and the distribution of respondents character-

and response rate

eriment relies on FDM to collect and asses thes observations. The minimum requirements of this

are determined by the rule of thumb that each groupnts should have 1015 people each, as recommended

The exrespon86 respinitiallthe preresponcompling rescharac

Types o

The(Table D showever, sof the for fuzDelphi literature (Delbecq et al., 1975).ps of respondents that were investigated are: (1) inde-velopers; (2) contractors; (3) asset developers; (4)

nt agencies; (5) municipalities. Type (1), (2), and (3)s refer to player D; type (4) and (5) refer to player M.

looked. Thethe respondmeet with tFDM-SAM ctrees in thi

Fig. 5. Game rating and game choice expet

ent consists of two data collection parts in which thete is different. For the validation of the game trees,ents participated in the survey out of 563 that visited

on-line survey; response rate: 15.28%. For estimatingnces and most expected game outcomes, the number ofs dropped during the survey. Precisely, 43 respondentsthis part of the experiment thus making the correspond-e rate of 7.64%. The distribution of the respondents

tic is similar in the two parts of the experiment.

pondents, years of experience and BR experience

owing table gives an insight of the respondent typese division of respondents between player M and playeratio of 3:7. Preferably, this should be improved. How-the main concern is to have the minimum requirementer of experts per group (1215 respondents per groupelphi method), the unfavorable ratio has been over- overall experience in years and the BR experience ofents are presented in Table 3. This does not completelyhe descriptive purpose, but it is nevertheless used in thealculation. Overall, it is possible to validate the game

s experiment because the response rate is regarded as

riment

74 B. Glumac et al. / Land Use Policy 47 (2015) 6677

Table 2Type of respondents.

Type of respondent Frequency Percent

(1) Independent developers 21 24.4(2) Contractors 26 30.2(3) Asset developers 13 15.1(4) Development agencies 12 14.0(5) Municipalities 14 16.3Total 86 100.0

Table 3Characteristics: years and BR experience.

Characteristic Levels Frequency Percent

Years of experience

B. Glumac et al. / Land Use Policy 47 (2015) 6677 75

number. All calculations are based on Hsu and Chen (1996). Still adistinction must be made between these two tables.

Table 5 represents the estimation of players M and D generalpreferences over the building claim game outcomes. In technicalterms, the binations dexperimenttion of the pAfter generinduction reexplained pmarked witdents aggreX in this taboutcome. ThM and playebe reportedthe same pAppendix 2

When cAppendix theoretical outcomes adents. Seve

In the stact rationaltheir own mand strategory about iavoided in tbe traced togated opinicollect aggrtorial experexperimenta proper inavoiding biacomes. Finaempirical rethe studiedthe precond

Second, outcome imthe behavioultimatum seen that thexperimentfor examplepayoffs valbut rather ainuencinginclude hidwillingnesscritic is thepayoffs incremains threpresentedrepresentedtheir levelswork, both institutionaconditions rthat are crualso reectreported thmethod ena

preference of previously identied attributes and their levels. Theperfect t shows that the players behave exactly as we wouldexpect them to, meaning that the mentioned common mistakeshas been avoided and that experimental game theory can be

sfullyides tifyi

by a y. Thith

pub prepestab

thered games.ned tnmenroce

a soation

withhis sh pled as

studthin hin tde ons. Firh of. 2).o joincesper. F

D repim ageivedal. Ond bes thble founicill prwe muallyprefenot layerdditieachork

ng thfs, (3

sion

isione mopporhis p

on ltipleen inecises, general preference covers all nine treatment com-iscussed in the setting up of the game-theoretical. In addition, the preferences (S) are regarded as indica-ayoffs. This can also be traced in the game tree (Fig. 2).ating the payoffs, the game can be solved by backwardferring to SPNE within the perfect information games,reviously. In the far right column, the indicated SPNE ish X. On the other hand, Table 6 reports on the respon-gated opinion about the most probable outcome. Thele refers to the highest score (S) for the most probablee score X in this table is assigned separately for playerr D. This implies that the most probable outcome could

differently for each player. For the bargaining game,rocedure is conducted. The results of this game are in.omparing the last columns of Tables 5 and 6 and2, there is an evident match between the game-solution provided by SPNE based on the preferrednd the most probable outcome estimated by the respon-ral conclusions can be drawn based on this perfect t.udied context, the rst conclusion is that the playersly. This perfect t indicates that players can perceiveoves and strategies, but also the other players moves

ies. Therefore, the most common critic of game the-ll-assumed rationality of a decision maker has beenhe studied context. This rational behavior probably can

the fact that the players are represented by the aggre-on. Besides the implemented fuzzy Delphi method toegated opinions, this study introduces the fractional fac-imental design (Table 1) to provide a statistically validal set-up. Such design equips respondents (player) withformation background about a decision problem thussed reports on both preferred and expected game out-lly, this methodological approach evidently led to thesults that discover rational behavior of the players in

context and the proof of players rational behavior isition to solve a SPNE with the backward induction.the perfect t between estimated SPNE and probableplies that the bargaining games are suitable to interpretr of the real-life negotiations in BR projects. When thegame is representing the real world problem, it is oftene bargaining pie has even or near-even splits, like in this

as well. This implies that the game outcome estimated with SPNE is based on very small differences in the

ue. However, this is not an imperfection of game theory common mistake of an analyst to dene all attributes

each players payoff. Commonly, an analyst misses toden attributes such as fairness that would indicate a

of a player to reject one-sided offers. Another common relevance of the estimated SPNE in the case when therease in value but the ratio over all game outcomese same. This is only a problem when the payoffs are

by the attributes levels that have certain value and not by the players utility that captures all attributes and

. By applying an experimental game theoretical frame-obstacles have been overcome. Together the describedl-economic context of the game and designed gameepresent the most important attributes and their levelscial for the game. In addition, any hidden attributes areed in the payoff because the respondents themselvese payoffs by fuzzy Delphi method (Fig. 5). The samebles to represent payoffs as utility based on the players

succesBes

of quanenced countra PPP wDutch:monlya well-2010),repeattiple tiundeenvirothese plead tonegotigamesfrom tfor botregardto thisbut wi

Witconclupayofffor eacand Figvated tpreferedeveloplayering claD percable desame eindicatfavorais in mthat wtively, that usThese issues both p

In aity of rframewchangipay-of

Conclu

Decbecomcan suest in tdependon muhas being of doutcom used in elaborated context.the included set of attributes, their levels, and a wayng the payoffs, the projected SPNE might also be inu-negotiation culture specic for a certain region or aerefore, the game outcome in which the players enterbuilding claim agreement can be regarded as typicallylic and private bodies in urban development are com-ared to share nancial risks and returns. This outcome islished Dutch urban development model (Samsura et al.,efore the reported SPNE can be regarded as a part of ame. In this type of games, players play the game mul-

In addition, this repeated game is a nite game withime span. It is nite because the institutional-economict is changeable, and it is not dened in time becausesses are long and unpredictable. Usually, these gamescially optimum strategy. This can be interpreted as a

culture and one essential part of nitely repeated undened time span is punishing players who deviatetrategy. The punishment may reect in lower payoffsayers. Therefore, the estimated SPNE should again be

typical for the Dutch context. An interesting additiony would be a comparison with SPNE of a similar gamea different institutional-economic context.his institutional-economic context, it is also possible to

both players preferences by only observing the gamest, it can be noticed that player D has higher payoffs

the games outcomes compared to player M (Table 5 This can be interpreted as that player D is more moti-n a PPP. Such an outcome should contribute to higher

to develop a site because this is the key activity of aurther, if the deal would be rejected (BC r; NBC r) thenorts higher payoff when rejecting a deal without build-reement (BC r < NBC r). Therefore, it is clear that player

the potential consequence of accepting a less favor-n the other hand, the higher payoff of player M for theranch is counterintuitive at rst. However, such a payoffat actually having a building claim agreement is morer this player too. This can be mainly explained that itipal interest to have also a rmly attached developerovide secure development of building plots. Alterna-ight say that player M also prefers to share the BR risks

exceed greeneld development risks (De Sousa, 2002).rence ndings could be used when negotiating othercovered by this study, such as nancial obligations ofs.on, these ndings provide a base to check the possibil-ing a different game outcome. In the game theoretical, this can be achieved with interventions such as: (1)e information of the involved players, (2) changing the) changing the playing rules (Jost and Weitzel, 2008).

s

processes in urban (re)development projects havere complex. Therefore, it is useful to nd theories thatt the governance of such processes. Of special inter-aper are urban (re)development processes that do notthe individual choice by one stakeholder, but depend

stakeholder decisions. A game-theoretical frameworktroduced in this paper to provide a better understand-ion-makers interactive behavior and expected decisionalong with the recommendations concerning the

76 B. Glumac et al. / Land Use Policy 47 (2015) 6677

application of intervention strategies in conict situations. Gametheory also provides a relevant framework to gather empirical datain order to interpret the respondents interactive behavior. There-fore, an experiment has been carried out to indicate which partsof the basicbehavioral pthat purposprovided byinitiative ph

An expeand it consclass; (2) seconcept; (4(5) design ment; (6) vaof the respocomes; (8) direct gameantees thatexperiment

In genercal framewin applyingrelying onlythe perspecnovel perspis not so muwith the prainstitutionainuence oence the ousuch analysprovide an to show its

Furthermnegotiationfor the builof the land. tion of ratiogame-theorble outcomthe players backward ithe perfect fully descriexactly as wshould be rwell-establ2010), theroutcome ofAs such, oudition of ppartnershiprisks and reach game,ing ordinal values leadpotentials fan alternatiattractive oferent negostrategies uthe two maity and dev

Finally, presents a

interaction processes in BR projects have been covered, anddeliberately so. The aim is to represent the interaction structureas a tool to understand the behavior of the involved parties, not tocompletely mimic the real-world to every detail. As mentioned,

gotigametruct

insto be

dix A

plem in epol

nces

D., Disnel

, Joy, V. Man, T., Tieen merical, E., Snelry to S., 200. PresET, 20ic Reg, C.F., .A., E

Publi, Hipeem foro, E., 2n Pol

A., 19nces. Rd, V., 2.

a, C.A.neld280.

, A.L., Vning: view,, J., 19

in locR., Wi? Evid

B., Has: app, D., Lnfrast., Shapof Orteral Ela, G.,chte H

M., Ching. Fua, A., Amin

Syst. ., Weducti

, YuanHall Pan, J.Ftures:in. Re., McN. Dev.., McN

deve theory are most useful in predicting and identifyingarameters that theory does not reliably determine. Fore, an illustration of urban (re)development process is

elaborating the three specic negotiation issues in thease of a BR project.rimental game theory procedure has been introducedists of eight consecutive steps: (1) selection of a gamelection of a game form; (3) selection of a game solution) description of the institutional-economic environmentof the game conditions within the game set environ-lidation of the assumed game structure; (7) estimationndents preferences and predictions of the game out-interpretation respondents feedback with SPNE and

outcome predictions. The suggested procedure guar- the decision moment is properly isolated and that the

is statistically valid.al, the benets of using an experiment game theoreti-ork to planning and development practice can be found

a prescriptive decision-making approach, instead of on the normative nature of game theory. The idea oftive approach is to provide usable outcomes such asectives through decision aids. Therefore, this approachch concerned with a theoretical contribution but rathergmatic value provided to the end user. In the describedl-economic, an analyst would be able to investigate thef different external effects (Table 1) that would inu-tcome of negotiations. However, to be able to performis additional data is required since it is not possible toempirical evidence of the use of a certain method andperformance on the same data set.ore, this paper provides an empirical example of the

s taking place in a BR project with respect to the choiceding claim model, the future land use and reparcellingIn the studied context, it is rst shown that the assump-nal players is correct. The evident match between theetical solution provided by SPNE and the most proba-e estimated directly by the respondents suggests thatdo act rationally. This test is required to assure that thenduction can be used to estimated SPNE. In addition,match implies that negotiations in BR can be success-bed by the bargaining games because players behavee would expect them to. However, the estimated SPNE

egarded as typical for the Dutch context. The game is aished Dutch urban development model (Samsura et al.,efore the reported SPNE can be regarded partly as the

a repeated game with nite and undened time span.r analysis provides empirical evidence of the Dutch tra-ublic and private bodies entering into public privates in urban (re)development, willing to share nancialeturns, even in very complex situations. Anyhow, for

all negotiable attributes are described with the accord-values. As the ndings reveal, the increase in attributess to the higher-level acceptance of a deal. This reectsor interventions in order to reach a certain outcome. Asve, these ndings can be used to set a borderline for anffer. Therefore, this article has demonstrated how dif-tiation issues in BR can be quantied and what are thender dened institutional-economic environment forjor players in urban development process, a municipal-eloper.the reader should realize that game theory alwaysn abstraction. Not all engagements of real-life

the nesingle base (ssimilarneed t

Appen

Supfound,landus

Refere

Adams, brow

Alker, S.Plan

Arentzebetwnum

Blokhuisbrowtheo

Bowles, Univ

CABERNnom

CamererCarroll, D

Rev.Chen, Y.

systCoiacett

UrbaColman,

ScieCrawfor

115De Sous

gree251

DelbecqPlanGlen

Forestergies

Ganser, gets

Glumac,ture

Grimseyfor i

Hahn, Gysis Gen

Hieminggeri116.

Hsu, H.-mak

IshikawmaxSets

Jost, P.-JIntro

Klir, G.J.tice

KoppenjtrucAdm

Lange, DPlan

Lange, Deldation attributes have been represented as separates. This implies that the proposed games provide a rmure and procedure) for the similar attributes (with theitutional-economic context and type of variable) thatstudied in the negotiation context.

. Supplementary data

entary data associated with this article can bethe online version, at http://dx.doi.org/10.1016/j.

.2015.03.018.

berry, A., Hutchison, N., Munjoma, T., 2001. Ownership constraints tod redevelopment. Environ. Plan. A 33, 453477.., Roberts, P., Smith, N., 2000. The denition of browneld. J. Environ.age. 43, 4969.mmermans, H., 2003. A multiagent model of negotiation processesultiple actors in urban developments: a framework for and results of

experiments. Environ. Plan. B 30, 391.nijders, C., Han, Q., Schaefer, W., 2012. Conicts and cooperation ind redevelopment projects: application of conjoint analysis and gamemodel strategic decision making. J. Urban Plan. Dev. 138, 195205.4. Microeconomics: Behavior, Institutions, and Evolution. Princeton

s, Princeton, NJ.02. Browneld Denition. Concerted Action on Browneld and Eco-eneration Network.2003. Behavioral Game Theory: Experiments in Strategic Interaction.ger Iii, R.J., 2006. Brownelds, crime, and tax increment nancing. Am.c Admin. 36, 455477.l, K.W., Kilgour, D.M., Zhu, Y., 2009. A strategic classication support

browneld redevelopment. Environ. Model. Softw. 24, 647654.001. Diversity in real estate developer behaviour: a case for research.icy Res. 19, 4359.95. Game Theory and Its Applications in the Social and Biologicaloutledge, London.002. Introduction to experimental game theory. J. Econ. Theory 104,

, 2002. Measuring the public costs and benets of browneld versus development in the Greater Toronto area. Environ. Plan. B 29,

an de Ven, A.H., Gustafson, D.H., 1975. Group Techniques for ProgramA Guide to Nominal Group and Delphi Processes. Scott. Foresman,

IL.87. Planning in the face of conict: negotiation and mediation strate-al land use regulation. J. Am. Plan. Assoc. 53, 303314.lliams, K., 2007. Browneld development: are we using the right tar-ence from England and Germany. Eur. Plan. Stud. 15, 603622.n, Q., Smeets, J., Schaefer, W.F., 2011. Browneld redevelopment fea-lying fuzzy Delphi. J. Eur. Real Estate Res. 4, 145159.ewis, M.K., 2002. Evaluating the risks of public private partnershipsructure projects. Int. J. Project Manage. 20, 107118.iro, S., 1966. A Catalog and Computer Program for the Design and Anal-hogonal Symmetric and Asymmetric Fractional Factorial Experiments.ectric, Research and Development Center, New York, pp. 73.

2006. Projectontwikkeling Marketperspectief & Integrale Gebieds-erstructurering. ING Bank, Economisch Bureau ING, Amsterdam, pp.

en, C.-T., 1996. Aggregation of fuzzy opinions under group decisionzzy Sets Syst. 79, 279285.magasa, M., Shiga, T., Tomizawa, G., Tatsuta, R., Mieno, H., 1993. TheDelphi method and fuzzy Delphi method via fuzzy integration. Fuzzy55, 241253.itzel, U., 2008. Strategic Conict Management: A Game-Theoreticalon. Elgar, Cheltenham, UK., B., 1995. Fuzzy Sets and Fuzzy Logic: Theory and Applications. Pren-TR, USA..M., Enserink, B., 2009. Publicprivate partnerships in urban infras-

reconciling private sector participation and sustainability. Publicv. 69, 284296.eil, S., 2004a. Browneld development: tools for stewardship. J. Urban

130, 109116.eil, S., 2004b. Clean it and they will come? Dening successful brown-lopment. J. Urban Plan. Dev. 130, 101108.

B. Glumac et al. / Land Use Policy 47 (2015) 6677 77

Lange, D., Wang, D., Zhuang, Z., Fontana, W., 2013. Browneld development selectionusing multi-attribute decision-making. J. Urban Plan. Dev. 140, 04013009.

Liang, Y.H., Guo, P., Hu, J.F., 2008. Analysis of browneld redevelopment by evolu-tionary game. In: IEEE International Conference on Industrial Engineering andEngineering Management, Singapore, pp. 481485.

Louviere, J.J., Hensher, D., Swait, J.D., 2000. Stated Choice Methods: Analysis andApplications. Cambridge University Press, Cambridge.

Ma, L., Arentze, T., Borgers, A., Timmermans, H., 2007. Modelling land-use deci-sions under conditions of uncertainty. Comput. Environ. Urban Syst. 31,461476.

Martnez, F.J., Henrquez, R., 2007. A random bidding and supply land use equilib-rium model. Transport. Res. B 41, 632651.

Mayer, I.S., van Bueren, E.M., Bots, P.W.G., van der Voort, H., Seijdel, R., 2005. Collab-orative decisionmaking for sustainable urban renewal projects: a simulation gaming approach. Environ. Plan. B 32, 403423.

Miles, M.E., Berens, G., Weiss, M.A., 2007. Real Estate Development: Principles andProcess. Urban Land Institute.

Mu, L., Ma, J., 2007. Game theory analysis of price decision in real estate industry.Int. J. Nonlin. Sci. 3, 155160.

Murray, T.J., Pipino, L.L., van Gigch, J.P., 1985. A pilot study of fuzzy set modicationof Delphi. Hum. Syst. Manage. 5, 7680.

Neumann, J., Morgenstern, V.O., 1944. Theory of Games and Economic Behavior.Wiley, New York.

Noorderhaven, N., 1995. Strategic Decision Making. Addison-Wesley, UK.Osborne, M.J., 2004. An Introduction to Game Theory. Oxford University Press, New

York.Page, G.W., Berger, R.S., 2006. Characteristics and land use of contaminated brown-

eld properties in voluntary cleanup agreement programs. Land Use Policy 23,551559.

Peiser, R.B., 2007. Regenerating Older Suburbs. Urban Land Institute, Washington,USA.

Peiser, R.B., Frej, A.B., 2003. Professional Real Estate Development. Urban Land Insti-tute, Washington.

Pfrang, D.C., Witting, S., 2008. Negotiating ofce lease contracts: from agame-theoretical towards a behavioural view. J. Eur. Real Estate Res. 1,88105.

Raiffa, H., 2002. Negotiation Analysis: The Science and Art of Collaborative DecisionMaking. Harvard University Press, Cambridge.

Rasmusen, E., 2well Publi

Samsura, A., vment throHuijbregtsSpringer, B

Samsura, A.A.,ment: a ga4865.

Samsura, D.A.A2013. Bargmental ap

Samsura, D.A.approach tPolicy 27,

Shoham, Y., Leyton-Brown, K., 2009. Multiagent Systems: Algorithmic, Game-Theoretic and Logical Foundations. Cambridge University Press.

Sibdari, S., Pyke, D.F., 2010. A competitive dynamic pricing model when demand isinterdependent over time. Eur. J. Oper. Res. 207, 330338.

Sounderpandian, J., Frank, N., Chalasani, S., 2005. A support system for medi-ating brownelds redevelopment negotiations. Ind. Manage. Data Syst. 105,237254.

Stengel, B.V., 2008. Game Theory Basics. Department of Mathematics, London Schoolof Economics, London, UK.

Syms, P., 1999. Redeveloping browneld land: the decision making process. J. Prop.Invest. Finance 17, 481500.

Tam, C.M., Thomas, K.L.T., 2011. Conict analysis study for public engagementprogramme in infrastructure planning. Built Environ. Project Asset Manage. 1,4560.

Tam, C.M., Zeng, S.X., Tong, T.K.L., 2009. Conict analysis in public engagementprogram of urban planning in Hong Kong. J. Urban Plan. Dev. 135, 5155.

Thomas, M.R., 2003. Browneld redevelopment: information issues and the affectedpublic. Environ. Pract. 5, 6268.

Van der Krabben, E., Jacobs, H.M., 2013. Public land development as a strategictool for redevelopment: reection on the Dutch experience. Land Use Policy30, 774783.

Walker, S., Hipel, K.W., Inohara, T., 2008. Attitudes and coaliions within browneldredevelopment projects. In: IEEE International Conference on Systems, Man andCybernetics, Singapore, pp. 29012906.

Wang, L., Fang, L., Hipel, K., 2011. Negotiation over costs and benets in browneldredevelopment. Group Decis. Negot. 20, 509524.

Wang, L., Fang, L., Hipel, K.W., 2007. A game-theoretic approach to browneld rede-velopment: negotiation on cost and benet allocation. In: IEEE InternationalConference on Systems, Man, and Cybernetics, Montreal, QC, pp. 18001805.

Wang, Q., Hipel, K.W., Kilgour, D.M., 2008. Conict analysis in browneld redevelop-ment: the erase program in Hamilton, Ontario. In: IEEE International Conferenceon Systems, Man and Cybernetics, Singapore, pp. 29132918.

Wang, W.-M., Lee, A.H.I., Peng, L.-P., Wu, Z.-L., 2013. An integrated decision mak-ing model for district revitalization and regeneration project selection. Decis.Support Syst. 54, 10921103.

Weihe, G., 2005. Publicprivate partnerships: addressing a nebulous concept. In:Kaspersen, L.B. (Ed.), 10th International Research Symposium on Public Man-agement (IRSPM X). International Center for Business and Politics Copenhagen

ness Shang,pact

.R., 20t. 5, 25

S., Hi over247.

S., Hiristicserencics EnM.L., P. The ctive 007. Games and Information: An Introduction to Game Theory. Black-shing, Malden, UK.an der Krabben, E., 2011. Funding transport infrastructure develop-ugh value capturing: a game theoretical analysis. In: Nunen, J.A.E.E.,, P., Rietveld, P. (Eds.), Transitions Towards Sustainable Mobility.erlin, Heidelberg, pp. 5980.

van der Krabben, E., 2012. Negotiating land and property develop-me theoretical approach to value capturing. J. Eur. Real Estate Res. 5,

., van Deemen, A.M.A., van der Krabben, E., van der Heijden, R.E.C.M.,aining for value capturing: a game-theoretical analysis and experi-proach. Environ. Plan. B 40, 234253.A., van der Krabben, E., van Deemen, A.M.A., 2010. A game theoryo the analysis of land and property development processes. Land Use564578.

BusiWu, Y., Z

its imYount, K

PracYouse,

ation240

Youse,acteConftron

Zellner, 2009collechool, Copenhagen. X., Skitmore, M., Song, Y., Hui, E.C.M., 2014. Industrial land price and

on urban growth: a Chinese Case Study. Land Use Policy 36, 199209.03. What are brownelds? Finding a conceptual denition. Environ.33.pel, K., Hegazy, T., 2010. Considering attitudes in strategic negoti-

browneld disputes. J. Legal Aff. Dispute Resol. Eng. Construct. 2,

pel, K.W., Hegazy, T., Witmer, J.A., Gray, P., 2007. Negotiation char- in browneld redevelopment projects. In: 2007 IEEE Internationale on Systems, Man and Cybernetics, Institute of Electrical and Elec-gineers (IEEE), New York City, NY, USA, pp. 18661871.age, S.E., Rand, W., Brown, D.G., Robinson, D.T., Nassauer, J., Low, B.,emergence of zoning policy games in exurban jurisdictions: informingaction theory. Land Use Policy 26, 356367.

Negotiation issues in forming publicprivate partnerships for brownfield redevelopment: Applying a game theoretical experi...IntroductionGame theory applications in forming PPP for a BR projectA game tree and its elementsGame elements: players, strategies, payoffs, equilibriums

Experimental game theory

Designing games over negotiation issues in forming PPP for a BR projectIdentifying a game class: non-cooperative conflict gameIdentifying a game form: extensive formFinding a solution concept: sub game perfect Nash equilibrium (SPNE)Defining the institutional-economic context of the gameSetting a game condition stateGame assumptionThe ultimatum game: building claimPlayer description, information, and strategy

The bargaining game: future land use and reparcelling of the landPlayer description, information, and strategy

Game tree validationThe preferences and predictions of the game outcomes

Data collection and respondents characteristicsSample size and response rateTypes of respondents, years of experience and BR experience

Game experiment resultsValidated game tree structureEstimated results with fuzzy Delphi method

ConclusionsAppendix A Supplementary dataReferences