Moving Beyond Optimism Bias and Strategic Misrepresentation: An Explanation for Social Infrastructure Project Cost Overruns

560 IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT, VOL. 59, NO. 4, NOVEMBER 2012

Moving Beyond Optimism Bias and StrategicMisrepresentation: An Explanation for Social

Infrastructure Project Cost OverrunsPeter E.D. Love, David J. Edwards, and Zahir Irani

Abstract—Infrastructure projects regularly experience cost andschedule overruns. Research led by Flyvbjerg has suggested thatmisrepresentation and optimism bias are primary causes for over-runs. While Flyvbjerg’s research has made a significant contribu-tion to ameliorating understanding as to why economic infrastruc-ture projects experience overruns, it does not adequately explainwhy this is the case for such social infrastructure. In addressing thisshortcoming, case studies are used to determine the intermediaryevents and actions that contributed to project cost overruns. Thepathogens, events, and actions that contributed to overruns areidentified and analyzed. The analysis of the cases’ findings led tothe propagation of a nomological framework for social infrastruc-ture project overruns. Acknowledgment of the systemic pathogenicinfluences has enabled the establishment of an orthodoxy, whichprovides an impetus for addressing the issues needed to improvethe performance of social infrastructure projects.

Index Terms—Design errors, overruns, pathogens, socialinfrastructure.

I. INTRODUCTION

H IGH-PROFILE infrastructure projects that experiencecost and time problems, or contractual disputes, attract

media attention as the community contributes to funding theirdelivery. Several notable projects in recent times include Den-ver’s US$ 5 billion airport that experienced a cost overrun of200%, the DKK 800 million Oresund Bridge that experienceda 68% cost overrun, and the Scottish Parliament Building thatwas over three years late and experienced more than a 900%cost overrun.

In Australia, several large-scale social infrastructure projects(i.e., hospitals, law and order, museums, schools, and recre-ational facilities) have experienced considerable delays due topoor project governance and design errors [36]. More contem-porary examples of this phenomenon include the Western Aus-tralian Perth Arena that had an original contract value of A$168 million but is forecast to cost more than three times thisamount and be delivered at least three years later than expected.

Manuscript received January 26, 2011; revised April 2, 2011 and May 28,2011; accepted July 16, 2011. Date of publication October 3, 2011; date ofcurrent version October 16, 2012. Review of this manuscript was arranged byDepartment Editor J. K. Pinto.

P. E. D. Love is with the School of Built Environment, Curtin University,Perth, W.A. 6845, Australia (e-mail: [email protected]).

D. J. Edwards is with the Business School, Birmingham City University,Birmingham, B42 2SU, U.K. (e-mail: [email protected]).

Z. Irani is with the Brunel Business School, Brunel University, Uxbridge,Middlesex, UB8 3PH, U.K. (e-mail: [email protected]).

Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TEM.2011.2163628

Flyvbjerg et al. [9] suggested that average cost overruns for in-frastructure projects can range between 20.4% for roads, 33.8%for bridges and tunnels, and 44.7% for rail. In contrast, Loveet al. [22] report significantly lower levels of average cost over-run, with 13% for roads and 5.5% for bridges. This observeddifferential can be explained by the monetary value of projects.Research report in Flyvbjerg et al. [9] focused on magprojectswith contract values in excess of US$ 1 billion whereas thecontract values reported in Love et al. [22] were considerablysmaller with an average value of A$ 33 million.

According to Flyvbjerg et al. [12], there are two rudimentaryreasons why projects experience cost overruns. First, strategicmisrepresentation that is an Orwellian euphemism for describ-ing deceptive actions used by politicians and planners to ensurethat projects proceed. Second, optimism bias that encapsulatesthe systematic tendency for decision makers to be overoptimisticabout the outcome of planned actions [11]. This includes over-estimating the likelihood of positive events and underestimatingrisk and loss.

Despite comprehensive research that explains “why” and“how” infrastructure projects experience cost and schedulegrowth, this issue remains an unresolved problem for govern-ments internationally. Irrespective of malignant practice or un-substantiated optimism, infrastructure projects continue to besubjected to significant levels of rework, which predominatelyarises through design changes and errors [22]. While Flyvb-jerg’s research has made a significant contribution to ameliorat-ing understanding as to why economic infrastructure projectsexperience overruns, it does not necessarily explain why this isthe case for such social infrastructure. In addressing this short-coming, case studies are used to determine the intermediaryevents and actions that contributed to cost overruns. A series ofcausal chains of design errors are presented through a processof causal judgment; the analysis of causation in this instancepredicts certain judgments about certain cases.

II. FROM TRADITIONAL TO PUBLIC–PRIVATE PARTNERSHIPS

Governments utilize an assortment of procurement methodsto deliver their infrastructure projects [29]. Traditionally, how-ever, derivatives’ of build-own-operate transfer (BOOT) anddesign-build-own-transfer have been used to fund high-profileeconomic infrastructure such as roads, bridges, and tunnels [17].Private-sector funding of critical infrastructure has enabled gov-ernments to service community needs and free up scarce funds toprop-up public services, such as education and health care. No-table examples of high-profile BOOT-type projects undertaken

0018-9391/$26.00 © 2011 IEEE

LOVE et al.: MOVING BEYOND OPTIMISM BIAS AND STRATEGIC MISREPRESENTATION 561

in Australia include the Melbourne Trans Urban City Link andthe Sydney Harbour Tunnel.

Significant social infrastructure projects include theA$ 155 million National Museum of Australia, delivered us-ing an alliance and the A$ 1.8 billion Fiona Stanley Hospital inWestern Australia (WA), delivered using management contract-ing. In stark contrast to the aforementioned projects, traditionallump sum (TLS) methods, or hybrids thereof, have tended topredominate, particularly within W.A. [20]. Successive govern-ments contend that this form of procurement provides cost andtime certainty, lower risk, and greater transparency and account-ability. Yet traditional “procurement” methods are unsuitablefor the management and delivery of infrastructure projects andservices [5].

Studies that have employed comparative analytical techniquessuggest that traditional procurement methods fail to meet valuefor money (VfM) assessment criteria and are prone to cost andtime overruns as well as disputes [29]. In addition, these methodsdo not adequately cater for whole-life costs, future maintenance,and standards for sustainability and energy usage because de-sign, construction operation, and maintenance of facilities areoften undertaken independently. Surprisingly, it has been sta-tistically demonstrated that the extent and influence of designerrors does not vary with the procurement method adopted andthe project type [22]. This finding, prima facie, contradicts thecomparative studies that have been undertaken considering theadverse impact of design errors on project performance. An ex-planation for this incongruity can be found at the heart of thedesign process: the design and contract documentation producedby design professionals consistently contains errors [1].

The quest for increased efficiency in public service delivery,government budget difficulties, and the growing receptivity ofpublic opinion to the discourses on rehabilitating market regu-lations have led to a growing number of forms of associationbetween the public and private sectors. Consequently, there hasbeen a subtle shift away from the use of conventional methods(such as TLS, design and construct management contracting,and construction management) to public–private partnerships(PPPs). Despite widespread acknowledgement that PPPs canprovide VfM (see, e.g., [29]), many academics and practitionershave questioned their underlying proposition as they increaseproject complexity and are prone to time and cost overruns.Thus, PPPs should be used only if they provide better VfMthan conventional methods [7]. In Australia and the U.K., forexample, VfM is explicitly required for proceeding with a con-cession PPP project. VfM is typically determined (a priori) by apublic-sector comparator who must determine whether PPP candeliver on this criterion for the benefit of the taxpayer. This issueremains a prevailing problem irrespective of the procurementmethod used to deliver infrastructure because eliminating designerrors and rework pose a major challenge when seeking VfM.

III. GENESIS OF DESIGN ERRORS

Design firms, such as architects, position themselves (andthe portfolio of services offered) using a differentiation strategy[27]. Design firms have also streamlined their activities to meet

the challenge of an increasingly volatile and competitive marketenvironment. Against this backdrop, most design organizationshave adopted a market-driven strategy based on their fees vis-a-vis service quality [32]. Winch and Schneider [37, p.930]suggest that only a few practices within the industry develop thedistinctive competence to sustain premium fees based upon theirstrategic orientation. Most other firms, however, earn averagefees and are unable to differentiate themselves [37].

Management within architectural firms has been conceivedas the administration of contracts and offices rather than themanagement of projects and practices [38]. The insularity andaversion of the architectural profession to “management” hasresulted in poor service quality and its marginalization withinthe industry [37]. Architects have been historically inflexibleand thus unable to respond to customer needs and demands.Therefore, a significant gap exists between consumers’ a prioriexpectations of an architectural service and their perceptions ofthe actual service delivered. Powell [27, p. 84], for example,states “architects still cling to the notion that their future liesin building original works of art to last forever.” Symes et al.[34, p.55] concluded that “without more training in how tomanage a firm and adapt to the needs of their customers, mostarchitects may well be doomed to work in small practices andthereby be further marginalized in the building process.” Thisproblem is not unique to architects as engineering design con-sultants have been confronting a similar dilemma.

Design quality is a function of the level of consultancy ser-vices required, the method of selecting these, and how feesare negotiated. When design consultants’ services are procuredthrough competitive tendering, fees levels are reduced but thedegree of services offered is restricted [37], [38]. Lowry [24]examined fee levels and concluded that fees were reduced toprimarily ensure organizational survival. Juxtaposed with re-duced fee levels are the excessive and unrealistic demands thatare often imposed upon design consultants by the public sectorto document projects [1]. When the period to design and docu-ment is constrained by strict time and cost parameters during theinitial planning and feasibility stages, design consultants haveto tradeoff the future risk posed by design shortcuts against feemaximization [23].

Individuals readily violate established conventions when un-der pressure to produce documentation, despite knowing thatsuch a practice is professionally inappropriate [23]. If concur-rent omissions occur in a single task, a common occurrenceduring the design process, the effects are additive and create arecurrent error trap. However, the decomposition of individualtasks into a series of steps enables errors traps to be identifiedin advance [30]. Variations in design procedures and protocolsmay lead to bespoke problems, and accordingly, distinctive ap-proaches are adopted to resolve issues as they arise [25]. Yet,this understanding of circumstance does not explain why de-signers continue to omit tasks and commit errors, particularly,in public-sector projects.

When a project commences, design consultants have everyintention of producing error-free documentation. Yet revisionsare frequently undertaken on the design documents that theyproduce [25]. Drawing on the “1:10:100” rule [6], which is


somewhat akin to the traditional medical axiom “an ounce ofprevention is worth a pound of cure,” an error identified duringthe predesign phase would have a proportional impact of $1 tothe project. If an error remains unidentified until completionof the design phase, the cost could increase to $10, and $100if unidentified when commencing construction on site. Con-tractors are frequently provided with erroneous contract docu-mentation requiring a number of design clarifications and revi-sions to be obtained from design consultants. Often, informationcontained within contract documentation is contradictory and,subsequently, requires changes to be made [13].

A. Emergence of Pathogens

The actions of strategic misrepresentation and optimism biasmay stimulate pathogens to arise within the environment thatinfrastructure projects are procured. Pathogens are latent condi-tions and lay dormant within a system until an error comes tolight. Before discovery, designers remain unaware of the impactthat particular decisions, practices, or procedures can have uponproject performance [4]. Pathogens can arise because of strate-gic decisions taken by top management or key decision makers.Latent conditions can reside within a system for a considerableperiod of time and become an integral part of everyday workpractices. Once they combine with “active failures” (which aresimilar to Deming’s common causes), the errors can arise andthe consequences may be significant [23].

Active failures are unsafe acts committed by people who areembedded within a system. Such acts include slips, lapses, mis-takes, and procedural violations [30]. Active failures are oftendifficult to foresee and cannot be eliminated by simply react-ing to the event that has occurred. Latent conditions, however,can be identified and remedied before an adverse event occurs,while pathogens can be defined by a number of qualities as theyare relatively stable phenomena that exist for a substantial timebefore the error occurs, not seen as obvious stages in an identi-fiable sequence of failure before the error occurs, and stronglyconnected to the error and are identifiable as principal causes ofthe error once discovered [4]. The pathogens that contribute toerror occurrence can be categorized as follows [4]:

1) Practice: arising from people’s deliberate practices.2) Task: arising from the nature of the task being performed.3) Circumstance: arising from the situation or environment

the project was operating in.4) Organization: arising from organizational structure or op-

eration.5) System: arising from an organizational system.6) Industry: arising from the structural property of the indus-

try.7) Tool: arising from the technical characteristic of the tool

(e.g., software).Many pathogen-based errors in engineering firms are based on

practices (i.e., those pathogens from people’s deliberate prac-tices) that attempted to solve a particular problem [4], [23],for example, recycling design details, specifications, and othercontract documentation to reduce time and save money with-out giving due consideration to the project’s bespoke nature.

Commencing work using tentative information is often a con-sequence of working with nontraditional procurement methods(where design and construction activities are overlapped) andtherefore, short lead times are often needed to meet a project’sschedule.

Individuals may repeat inappropriate practices, such as tak-ing shortcuts and not following due processes. When a practiceprovides an individual with a satisfactory outcome, this practiceis used again on future projects irrespective of its suitability [4].For example, the designer’s decision to eschew audits, checks,verifications, and reviews prior to releasing documentation forpricing or construction. Due to financial pressures being im-posed upon design firms by their clients, this practice has be-come the norm [19]. Lack of attention to quality managementduring the design process has resulted in the notion of reworkbecoming entrenched in work practices, and consequently, lessprofit is experienced by design firms [32]. It has been suggestedthat design-error-induced rework may have settled within theindustry at an insidiously comfortable level [21]. If design con-sultants repeatedly produce design-error-induced rework due tocomplacency (i.e., the cost of doing business), then a chronicmalaise becomes “normal.” Whatever the percentage increasethat is taken up by rework, that percentage will be added to thefirm’s costs. So, if rework accounts for 10% of design firm’sregular work, this would lead to everything being increased by10%: supervision, cycle time for administrative procedures, an-swering requests for information, and so on. The time elementtranslates into costs, which are then buried within “normal”operating costs.

B. Causation and Causal Chains

The relationship between strategic misrepresentation and op-timism bias with cost overruns implicitly assumes causality interms of counterfactual dependence of the effect on the cause:the cause is rendered counterfactually necessary for the ef-fect [18], [14], [2]. For instance, to say that strategic misrep-resentation caused a cost overrun is to say that if the misrepre-sentation had not occurred, then the cost overrun would not haveensued. To be more precise, causality can be defined by referenceto a causal chain of counterfactually dependent events, wherea sequence of events (C, E, F, ...) is a chain of counterfactualdependence if E counterfactually depends on C, E counterfac-tually depends on F, and so on. Basically, Lewis [18, p. 167]asserted that “one event is a cause of another if and only if thereexists a causal chain leading from the first to the second.”

Various forms of counterfactual dependence have beenadopted through the application of structural equations (see,e.g., [14]). Within the extant literature, very limited studieshave applied structural equations to examine the causal fac-tors that contribute to the poor performance of infrastructureprojects (see, e.g., [23] and [26]). While such studies have pro-vided a valuable contribution to understanding causal inferencesthrough generalization, they have not provided a nomologi-cally possible context. Such a context provides detail about howevents unfold according to an underlying “event theory,” a setof background laws that define the outcome of events [2], [3].


Fig. 1. Nomologically different contexts each represent a different history.

Fig. 1 illustrates three nomologically different contexts, wherestrategic misrepresentation A and/or optimism bias C could giverise to a cost and/or time overrun E. Each node represents anevent. In this instance, the occurrence of event A or C or both(at some implicit point in time) is the cause of the occurrenceof event E (at a later point in time).

In the context of Fig. 1(a), C and D are proximate (as are Aand D) and C and E are remote. The occurrence of C stimulatesD, in this case a design error, but inhibits B. The occurrenceof D then results in E (i.e., a cost overrun). Besides, strategicmisrepresentation and optimism bias, pathogenic influences cangive rise to C and A, which can then trigger a series of eventsthat result in E, a time or cost overrun (or both). Determiningthe conditions that enabled events to occur would provide richerand more meaningful understanding of events that lead to designerrors and time and cost overruns.

IV. RESEARCH APPROACH

A case study approach based upon analytic induction is usedto further examine the underlying dynamics that may contributeto cost and time overruns in social infrastructure projects. Acase study is exploratory in nature, based on interviews andrelies heavily on verbal reports and unobtrusive observation asdata sources. This methodology should be used to investigatethe technical aspects of a contemporary phenomenon within areal life context [39], particularly, in critical and unique cir-cumstances [10]. It is, particularly, useful when the boundariesbetween phenomenon and context are difficult to ascertain andwhen multiple sources of evidence are used [39]. A case studycan provide analytical rather than pure statistical generalizationsand can capture the complexity and dynamism of organizationalsettings in projects [10].

Analytic induction refers to a systematic examination of sim-ilarities between various social phenomena to develop genericconcepts or ideas. It facilitates modification of social conceptsand their relationships throughout the research process, withthe goal of most accurately representing the reality of the sit-uation [28]. No analysis, however, can be considered final asreality is inexhaustible and dynamic [40]. The determinationof causal events that lead to design errors can explain the in-teraction that can exist between variables. The conjunction ofevents can result in orthodoxies being established and throughthe process of observation, generalizations can be made. If such

generalizations can be repeatedly tested and confirmed, they canlead to the discovery of a lawful relationship.

A. Data Collection

An Australian public-sector client was invited (and subse-quently accepted) to participate in the research during 2009.The client identified two projects where cost overruns wereforecasted to occur due to significant design problems that werebeing experienced. Specifically, the problems pertained to de-sign errors within contract documents that were produced bythe design consultants. With the client’s unequivocal supportand permission, the researchers approached both design con-sultants and contractors involved with the identified projectsand explained the nature of the research. All agreed to activelyparticipate in the research.

A series of in-depth interviews were conducted over a four-month period with participants who had been involved with thedelivery of a hospital and school. Data collected during and aftereach project’s completion enabled the researchers to develop anin-depth understanding of the key issues and errors that emergedduring each project’s procurement. A variety of methods wereused to obtain data, which included interviews and documentarysources (e.g., variation lists, site instructions, architect’s instruc-tions, drawings and specifications, and nonconformances). Forthe purposes of this research, the following operational defini-tion of an error was adopted: “Where a deviation of some kind,whether a departure from an intended course of action, depar-ture from a path of planned action toward a desired goal or adeviation from the appropriate behaviour at work” ([31, p. 39]).

Primary data emanating from interviews were used to deter-mine the conditions and events that led to design errors and costoverruns. Interviews were kept open using phrases such as “tellme about it” or “can you give me an example.” This approachenabled the researcher to ask respondents about pertinent factsas well as their opinions about design-error causation. The opennature of the questions stimulated avenues of interest to be pur-sued as they arose without introducing bias in the response.Additional notes were taken during the interview to support thedigital transcription process to maintain validity and safeguardagainst the digital recorder’s failure. A total of 24 in-depth inter-views were undertaken, which ranged between 60 to 90 min induration with architects (5), engineers (7), project managers (2),contract administrators (2), contractor’s project managers (2),and clients (2). The interviews were digitally recorded and, then,transcribed and distributed to each interviewee for comment andapproval.

B. Data Analysis

Content analysis was used as the primary analysis technique.In its simplest form, this technique can be used to extract andcategorize information from documents [15]. The text derivedfrom interviews and other documentary sources was analyzedusing NVivo 9. This software application enabled data sourcesto be analyzed [33]. One advantage of such software is thatit enables additional data sources and journal notes to be in-corporated into the analysis. The recurrent development and


TABLE IPROJECT DETAILS

reassessment of themes as the analysis progressed accords withcalls for avoiding confining data to predetermined sets of cate-gories [33]. Kvale [16, p. 204] suggests that ad hoc methods forgenerating meaning enable the researcher to access “a variety ofcommon-sense approaches to interview text using an interplayof techniques such as noting patterns, seeing plausibility, mak-ing comparisons, etc.” NVivo facilitated an organic approach tocoding to be developed because it enabled triggers or categoriesof interest within the text to be coded and used to keep trackof emerging and developing ideas [16]. These codings can bemodified, integrated, or migrated as the analysis progresses andthe generation of reports, using Boolean search, facilitates therecognition of conflicts and contradictions.

V. PREFACE: SOCIAL INFRASTRUCTURE CASES

Details pertaining to the social infrastructure, projects exam-ined can be found in Table I. With regards to the hospital, theState Government sought to politicize the hospital upgrade bydemonstrating that they were attending to the shortfall in “bedspaces” and “long patient waiting” lists. The project’s scope wasin development and meetings with key hospital stakeholderswere being undertaken when the decision to deliver the projectahead of the original intended commencement date was made.This resulted in the project’s scope only being partly definedwhen the design and documentation process commenced. To ad-dress the need for timely delivery of the project, the client useda construction management form of the procurement methodand eschewed the use of a TLS form. In doing so, it was en-visaged that design and construction would proceed in parallelthus ensuring a far more expedient process than other methodsavailable.

At the project’s outset, design consultants and contractormembers held a degree of skepticism in regard to the ambi-tious program of works, the inadequate scope and the restrictedaccess to areas during the hospital’s daily functioning: 96 weekswere indeed considered to be an optimistic schedule. A differentscenario was encountered with the school project, and notably,the client preferred to use a TLS method. Contracts for designservices were procured using competitive tendering and the low-est tenders were awarded contracts. Likewise, the contractor wasselected based on a competitive basis and thus the lowest ten-der. The State Government has a policy in place not to simplyaccept the lowest bid, but rather offer contracts on a VfM basis.

Therefore, it appeared that VfM was overlooked as a budget hadbeen established for the school’s development program. Theinitial project estimates developed for the client overlooked theextent to which inflationary pressures had influenced materialand labor costs within the local market. A dialogue between theresearcher and client suggested that the client may have beenslightly overoptimistic or perhaps overconfident with the initialestimate.

Evidently, however, issues surrounding accountability andprobity appeared to be drivers for securing lowest price; therewas little room to adjust the project budget. A culture of un-certainty avoidance within the client’s organization and criti-cal decision-making processes appeared to reside. As will bedemonstrated herein after, a combination of the lowest bids anda lack of project governance provided the stimulus for firms toadopt self-serving practices in an attempt to maximize their feesand profits.

In both case study projects, design errors contained within thecontract documentation that was produced provided a dysfunc-tional working environment where self-preservation prevailed.A number of different types of pathogen emerged through theanalysis and interpretations of findings. The most prominentpathogens identified, as will be discussed hereinafter, relatedto organization, circumstance practice, and task. Figs. 2 and 3demonstrate how the pathogens worked together to contributeto a series of errors for the hospital and school, respectively.There was no evidence that any form of strategic misrepresenta-tion arose during the interviews with the client, but a degree ofoptimism bias regarding each project’s design and constructionschedule prevailed in all cases. While the hospital and schoolwere to be upgraded, political expediency to satisfy immediatecommunity needs and stymied criticisms from the main oppo-sition party provided the inertia for their earlier completion.Table II presents examples of quotes that were obtained fromthe interviews and their pathogenic classification; interestingly,many of the pathogenic influences identified were deemed to beinterdependent.

VI. HOSPITAL CASE

The main challenges for the contractor were gaining accessto different areas of the hospital and integrating new and exist-ing services (i.e., electrical, mechanical, hydraulic, etc.). Bothdesign and construction were being undertaken in parallel, andit became evident to the client and contractor during the ini-tial stages that the design consultants were unable to providecomplete design documentation when required. The project’sschedule and client demands were far too demanding for thedesign consultants. Design consultants were reluctant to meetthe contractor on site to discuss issues due to their prevailingworkload, and they mentioned that they were often required tomultitask due to their involvement with several projects at once.Such pressure subjected them to workplace stress and thus pro-vided the ripe conditions in which mishaps, mistakes, lapses,and omissions arise.


Fig. 2. Hospital pathogens: Circumstance, task, and practice.

A. Public-Sector Demands

The use of a construction management procurement methodrequired design consultants to document as the building was be-ing constructed and in accordance with subcontract trade pack-ages. As a result, design firms were required to work togetherin a collaborative and integrated manner. Serendipitously, theclient, design consultant and contractor had limited experiencewith this form of the procurement method. The client’s de-fault choice for previous projects had been predominately aTLS form. Even though the project scope was incomplete at thecommencement of construction, the client suggested that design

changes occurred due to a “mismatch of the project team’s un-derstanding” of their requirements. The contractor always triedto accommodate the client’s changes and demands within theconfines of the project’s program. To ensure a timely projectcompletion, the contractor was forced to employ a specialist con-sultant scheduler to monitor and revise the program due to therate of design errors and changes that were being incurred. Thecontractor tolerated the substandard documentation producedas they were simultaneously negotiating contractual terms withthe client to secure a multibillion dollar social infrastructureproject.


Fig. 3. School pathogens: Circumstance, task, practice, and organization.

B. Professional Fees

While the client furnished additional costs for any changesthey made to the project scope, they regularly disputed manyof the contractor’s claims that resulted from design errors. Theclient acknowledged that the contract documentation may havecontained errors, but they obstinately believed that the contrac-tor had the capacity and capability to attend to errors beforethey resulted in rework or further design changes. The client’sstance indirectly rewarded the design consultants for their “poorquality” work. One particular design consultant even adopted abelligerent stance toward the contractor who had complained to

the client on several occasions about the standard of design doc-uments produced stating: “If the contractor tries to make a claimagainst our work, then we will counter claim. The contractor hasno idea how difficult it is for us; we’re being paid a pittance.”The contractor had become a “gate keeper” and every drawingissued had to be checked to ensure it married with the con-struction works undertaken. It was not within the contractor’scontractual requirements to adopt this role, but they deemed itnecessary if the project was to be completed successfully. Ad-ditional time and resources were needed to identify and solvethe errors generated within the documentation produced by the


TABLE IIPATHOGEN CLASSIFICATION AND EXAMPLES

design consultants. This point was consistently made during in-terviews and it was evident that the contractor was frustrated astheir “profits” were being unnecessarily hemorrhaged.

C. Managerial Procedures and Practices

The design consultants did not conduct audits, reviews, orverifications as it was perceived that they did not have the time.They were not an explicit requirement of the client and, there-fore, excluded from their professional fees. Moreover, limitedcoordination and integration between engineering and archi-tectural drawings took place. Surprisingly, design errors wereconsidered to be an integral part of professional practice withan engineer stating: “There’s always going to be figures thatdon’t match up. It’s a fact of life, and we normally sort outany problems that arise with the contractor.” As a result of de-sign consultants continually omitting fundamental proceduraltasks (e.g., design reviews) problems emerged on site for thecontractor on a daily basis.

During the initial stages of the project, drawings were pro-duced by the service engineer indicating that services needed tobe laid in ground. There was no consultation with the structural

engineer. As a result the foundations failed to take into accountthe positioning of in-ground services. Accordingly, the founda-tions had to be modified on site, which resulted in ground beamsbeing modified at considerable expense. The hydraulic and elec-trical services laid in-ground clashed with a 2-m-deep lift pit de-signed by the consultant structural engineer; the services had tobe diverted round the pit. A plethora of examples exist where thearchitectural and engineering drawings conflicted with one an-other. For example, the architectural drawing indicated a brickwall to a particular staircase, whereas the structural drawingsindicated that the same wall was reinforced precast concrete.Furthermore, the dimensions of the staircase were different. Inanother example, the structural drawings indicated that steelcolumns were required on the first floor to support the newGastroenterology department’s roof. No columns, however, hadbeen specified on the ground floor to support the floor above.In this case, the circumstance, task, and practice pathogens op-erated in congruence to create a climate for design errors toensue.

Within the same set of architectural and structural drawings,a plant room had been designed but the ducting and equipmentrequired would not fit within the specified dimensions as theplant was too small. Due to the design consultants having limited


time to revise their documentation, the contractor’s team workedwith the mechanical subcontractor to redesign the configurationof the plant room to ensure the ducting and equipment could beaccommodated. The revised design was then approved by thedesign consultants. It was fortuitous that all five stages of theproject were completed and delivered to the project client ontime considering the amount of design errors and changes thatoccurred. The contractor’s proactive approach to identifying andsolving design errors resulted in a three-month time saving forthe design consultants.

VII. SCHOOL CASE

In stark contrast to the hospital case, the engineering designfor the school was subjected to multiple reviews, checks, andsignoffs. While such policies and procedures are a vital part ofthe design process, they can also be detrimental and a waste oftime if design consultants do not coordinate and integrate theirwork. In this case, a clear misunderstanding of the expectationsand requirements of the client and project team members oc-curred. It was revealed that the structural engineer and architectworked independently with limited design coordination and so,consequently, design omissions, errors, and changes occurred.Despite the occurrence of omissions and errors, the structuralengineer was adamant that they had acted professionally anddiligently. A focus on simply “doing things right, rather thandoing the right things” appeared to subconsciously reverberatefrom the structural engineer during the interview. The undertonefrom the structural engineer was that the engineering designwas correct and that they were not responsible for errors thatoccurred. The engineer was, however, cognizant not to blameanyone, but implied that the architect failed to provide completeinformation when required.

Recognizing the potential for errors in calculations and load-ing assumptions that could be made, the structural engineer or-ganized for their work to be checked by a third party, specificallyto ensure compliance to Australian standards. The structural en-gineer adopted a risk mitigation strategy to prevent errors fromoccurring. Risks were identified and prioritized and more timewas then allocated to checking the critical designed componentsand structure. This process was deemed to be an essential partof the verification process; if it was not undertaken, the poten-tial for an engineering failure and accident was considered tosignificantly increase.

A. Design Coordination

Audits, reviews, and verifications are only useful if designdocumentation is coordinated systematically between designconsultants. Despite the structural engineers’ rigorous reviews,they omitted to design the storm water drainage. In fact, thearchitect, who was acting as the project’s superintendent, over-looked this issue as well. As a result, the contractor did notinclude the storm water drainage in their tender submission andan independent hydraulics engineer was employed to attend tothis oversight. Rectifying this omission error cost A$ 100 000.In another related example, the structural engineer omitted to“design ferrule connections” which were required to be cast into

the 60 precast reinforced concrete columns. Such connectionswould serve to accommodate the tensile forces imposed uponcolumns from the floor above. Again, this omission occurreddue to inadequate information flow between the architect andthe engineer.

B. Project Governance

Akin to the hospital upgrade case, the contractor in this projectobserved that the contract documentation produced by the de-sign consultants was contradictory. According to the contrac-tor, the design consultants simply needed to “communicate witheach other and this issue would have been resolved.” It appearedthat the procurement method and contractual obligations of theparties hindered “effective” communication from taking place.Various consultants had been appointed at different stages of thedesign process and were, therefore, constrained by what had al-ready been designed, particularly, by the architectural drawings.Both the architect and the contractor believed that many of theerrors that did arise could have been prevented if the contractorhad been included in the design process.

The client preferred to use TLS contracts in conjunction withcompetitive tendering. Issues surrounding accountability, pro-bity, and familiarity were deemed to be the drivers underpinningtheir use. It was perceived that the client paid limited attentionto project governance, i.e., they did not actively participate inthe project, which resulted in key decisions being taken by thesuperintendent, who also assumed the role of architect. Ac-ceptance of the lowest bids, from design consultants and thecontractor, provided the impetus for firms to adopt self-servingpractices to maximize their fees. The verification of the projectteam’s performance was also left to the superintendent. As tax-payers fund such projects, it is expected that “VfM” rather thanthe lowest cost is adhered to, so as to ensure they are deliveredon time and within budget.

C. Silos: Creation of Walls

The mechanical and electrical engineer suggested that a “silomentality” had become ingrained within the way in whichprojects were being procured. The engineers proceeded to statethat there was no motivation or incentive in place to encouragedesign consultants to work together in a harmonious and collab-orative manner to ensure that documents had been coordinatedand were complete. Fees were considered to be tight and withan unrealistic schedule in place to document the design, consul-tants implemented “time boxing” to ensure that the key areasof the design were complete (i.e., a fixed period of time maybe allocated to complete each task, irrespective of whether thedocumentation of each individual task is complete or not). Inadopting this approach, however, they reduced their “scope ofwork.” This resulted in aspects of their designs being incom-plete, which in turn led to inconsistencies occurring betweenarchitectural and engineering drawings as well as contractualvariations being raised and issued during construction.


Fig. 4. Nomological framework of design error and project overruns.

VIII. NOMOLOGICAL FRAMEWORK FOR SOCIAL

INFRASTRUCTURE OVERRUNS

The association between strategic misrepresentation and op-timism bias with project overruns, which has been promulgated,does not adequately explain why social infrastructure projectsconsistently under perform in terms of time and cost. The delimi-tation in Flyvbjerg and his colleagues’ theorem is that intermedi-ary conditions and events that lead to project overruns occurringare not examined or explained. A chronological lacuna exists be-tween the initial event and final outcome. Only by understandingand knowing the causal of events occurring within a system, canprogress be made to ameliorating its performance. To assumeoverruns occur due to strategic misrepresentation or optimismbias, or a combination of both, disregards the complexities andunderlying dynamics associated with the delivery of social in-frastructure projects. Even when the aforementioned issues donot arise, the likelihood that a project would incur overruns isstatistically high when factors such as structural and technicalcomplexity, procurement arrangement, contractual conditions,managerial effectiveness, location, size, and type of project aretaken into account (see, e.g., [35] and [23]). An issue that com-pounds this quandary is the differing goals and objectives ofthe “ephemeral coalition,” which can often provide a forum for“project dissonance” [5]. Vehicles that can be used to reducesuch dissonance and align goals and objectives are alliance-based project structures. Such structures can provide movementtoward consensus and reduce negative tensions that can prevailwithin projects.

In addressing the shortcomings of Flyvbjerg’s research, theevents and actions that led to overruns in social infrastructureprojects are identified and presented in a nomological frame-work in Fig. 4. The organizational- and project-related aspectsthat lead to errors and rework are made explicit. Retrospec-tive analysis of events and actions that occurred has been use-ful for determining the pathogenic influences. In particular, thepathogens of organization, circumstance, practice, and task wereleitmotivs within the cases, which are denoted by the shadedboxes. Acknowledgment of these features translates into thefollowing orthodoxy being produced.

“The use of competitive tendering for selecting design consultantsfor social infrastructure projects establishes an environment wheretheir services are reduced or omitted to maximize profit. The omis-sion of critical tasks and practices such as design audits, reviewsand verifications invariably leads to contract documentation beingerroneously produced and therefore increases the propensity for costand schedule overruns to be experienced.”

The framework is intuitive and implies that competitive ten-dering stimulates design firms to adopt opportunistic and un-ethical behaviors. More often than not, when design firms areselected for a public-sector project, overly competitive feescharged are deemed to be so low that there is minimal scopefor profit. When these are combined with an overtly unrealisticexpectation to design and prepare documentation, a breedingground for practice- and task-related pathogens may begin tofester.


Design firms need to differentiate themselves from their“mass of rival competitors by articulating a distinctive com-petence” (see [37, p. 928]). At present, however, many firmsappear to subconsciously assimilate themselves with Porter’s[42] ‘cost-leadership strategy’, whereby public-sector commis-sions are obtained by appealing to their price-sensitive nature.A strategy achieved by offering a “no-frills service” wherebystandardized details and specifications are issued, time boxingis implemented, design audits and reviews are eschewed, and er-roneous documents are produced and distributed to project teammembers. Furthermore, staff members are often required to mul-titask, which limits their ability to address timely responses toqueries that may be raised by other project team members. Issuesof cognitive dissonance can readily emerge among individualswho are subjected to high work demands. Overheads may becompressed by paying reduced salaries, locating to premises inlow-rent areas, and establishing a cost-conscious culture. Main-taining such a strategy requires a continuous search for costreductions in all aspects of the business.

In a quest to improve project performance, the public sec-tor has entered a period of introspection. For example, therehas been a proclivity to switch from conventional project de-livery methods to PPPs, though the improvements gained havebeen marginal with many reported prominent failures; promi-nent Australian examples include the Port Macquarie Hospital inN.S.W., the Sydney Airport Railway Link, and the Sydney CrossCity Tunnel. While significant efforts are being undertaken todevelop innovative ways to reduce overruns, it would appear thatan understanding as to why they occur has been, and continuesto be, overlooked. A reason contributing to such an oversightpertains to a shortfall in knowledge causation and the interde-pendence that exists between events and actions. Strategies andtechniques embedded within concepts of lean production andsupply chain management are being applied to project environ-ments by public-sector clients, in an attempt to reduce overrunswithout eliminating existing pathogenic influences. There is adanger that a whole series of “new” problems may materializeand become intertwined with those that already exist [41]. Thisis somewhat nihilistic and may further blur the epistemologicaland ontological lenses that are being used to examine the natureof project performance within the public sector. The developedframework and orthodoxy portray the fundamental issues thatneed to be addressed by public-sector clients and managerswithin design firms to ensure that projects are delivered moreeffectively than they currently are.

IX. LIMITATIONS OF THE STUDY AND AVENUES

FOR FUTURE RESEARCH

Using case studies, the research sought to determine thecausal nature of project overruns that were experienced in socialinfrastructure projects. This represented a challenging task dueto the political and commercial sensitivity attached to this issue.In this paper, this extensive and complex phenomenon has beenexamined from a rather narrow empirical perspective, that ofdesign errors and their causation. The selection and use of twocases naturally brings forth several limitations as far as the reli-ability and generalizations of the research results are concerned.

When case studies of this nature reveal some of the complex-ities associated with project delivery, particularly, in organiza-tional settings, there is often a problem with representation. It isdifficult to present accessible and realistic views of events thatlead to design errors occurring. For example, interviews wereused to identify design errors and obtain information about howand why they occurred. Thus, there was a potential for inter-viewees to intentionally or unintentionally conceal information.Moreover, there are often several different ways to present thesame set of issues, each one of which is subtly different in itsapproach and emphasis. This situation can make the findings de-rived from interviews difficult to summarize and thus contributeto the occurrence of situational biases.

Another limitation of this research is the adopted perspec-tive. Instead of trying to understand how and why strategicdecisions are generally made within projects, this research hasbeen first and foremost limited to an operational focus. Al-though the research has also taken into account an array ofviews along the development of its theoretical analysis, the mainperspective from which conclusions have been drawn are fromthose who were generally operating at the “coalface” within theprojects sampled. Understanding the rationale for key strate-gic decisions adopted by Government and organizations thatwere involved within the selected projects would have no doubtprovided a more fruitful and comprehensive understanding ofthe pathogenic influences contributing to project time and costoverruns.

Building upon the research presented in this paper and thenormative literature, it is suggested that additional descriptivecase-study-based research should be undertaken to further de-rive the conditions for quantitative studies to assess the proba-bility of project overruns. The use of Bayesian networks (BN)to develop risk management models throughout the formativestages of a project’s life cycle would reduce the likelihood ofdesign errors and subsequent rework from occurring. A BN isa graphical inference technique used to express the causal re-lationships among variables. They can be used to predict theprobability of unknown variables or to update known variablesgiven the certain state of other variables (i.e., evidence suchas provided in this research) through the process of probabilitypropagation or reasoning.

X. CONCLUSION

Cost and schedule overruns have become an innate featureof infrastructure projects. Flyvbjerg and his colleagues’ sem-inal work has suggested that strategic misrepresentation andoptimism bias are the primary contributors to economic infras-tructure overruns. However, a significant amount of projectsexperience overruns without such actions being present. To sim-ply assume that strategic misrepresentation and optimism biasare overarching actions that lead to the unsuccessful deliveryof social infrastructure projects is misleading considering thecomplex array of conditions and variables that interact with oneanother during the project’s procurement. More importantly,solely focusing on addressing such actions may mask the un-derlying conditions that continually contribute to the adoptionof opportunistic project and managerial practices. In this pa-


per, it has been revealed that the underlying pathogens that areresident in the environment within which social infrastructureprojects are procured act in congruence to destabilize the man-agement and functioning of projects. Unnecessary products ofthis destabilization process are design errors.

Understanding the conditions that result in design errors oc-curring is necessary to reduce their incidence within projects.An examination of two social infrastructure cases presented inthis paper revealed that errors contributed to a great deal ofdissonance between project team members as well as increasedproject costs. An investigation as to why and how design errorsoccurred was undertaken. This resulted in the construction ofa nomological framework and orthodoxy of design error andproject overruns. Unless there is a fundamental shift away fromthe use of competitive tendering to select design consultantsand a mandate for design reviews, checks, and verifications tobe undertaken throughout each phase of the design process, er-rors will remain and continue to contribute to overruns in socialinfrastructure projects.

ACKNOWLEDGMENT

The authors would like to thank Associate Editor, Prof. J.Pinto, and the three anonymous reviewers for their valuable andconstructive comments, which helped improve this manuscript.

REFERENCES

[1] S. Andi and T. Minato, “Design document quality in the Japanese con-struction industry: Factors influencing and impacts on the constructionprocess,” Int. J. Project Manage., vol. 21, pp. 537–546, 2003.

[2] J. Bell, “Causation and causal conditions,” in Proceedings of the 9thInternational Conference on the Principles of Knowledge Representationand Reasoning, D. Dubios, C. A. Welty, and M.-A. Williams, Eds. MenloPark, CA: AAAI Press, 2004, pp. 2–11, Eds.

[3] J. Bell, “Natural events,” J. Artif. Intell. Res., vol. 30, pp. 361–412, 2007.[4] J. S. Busby and E. J. Hughes, “Projects, pathogens, and incubation peri-

ods,” Int. J. Project Manage., vol. 22, pp. 425–434, 2004.[5] S. O. Cheung and T. W. Yiu, “Are construction disputes inevitable?,”

IEEE Trans. Eng. Manage., vol. 53, no. 3, pp. 456–470, Aug. 2006.[6] P. B. Crosby, Quality is Free: The Art of Making Quality Certain. New

York: New Amer. Library, 1979.[7] European Commission. (2003). Guidelines for Successful Public-

Private Partnerships. European Commission. [Online]. Available:http://ec.europa.eu/regional_policy (last accessed May 15, 2010).

[8] B. Flyvbjerg, M. K. Skamris, and S. L. Buhl, “What causes cost overrun intransport infrastructure projects?,” Transp. Rev., vol. 24, no. 1, pp. 3–18,2004.

[9] B. Flyvbjerg, M. K. Skamris, and S. L. Buhl, “How (in)accurate aredemand forecasts in public works projects,” J. Am. Planning Assoc.,vol. 71, no. 2, pp. 131–146, 2005.

[10] B. Flyvbjerg, “Five misunderstandings about case-study research,” Qual-itative Inquiry, vol. 12, no. 2, pp. 219–245, 2006.

[11] B. Flyvbjerg, “Curbing optimism and strategic misrepresentation in plan-ning: Reference class forecasting in practice,” Eur. Planning Stud., vol. 16,no. 1, pp. 1–21, 2008.

[12] B. Flyvbjerg, M. Garbuio, and D. Lovallo, “Delusion and deception inlarge infrastructure projects: Two models for explaining and preventingexecutive disaster,” California Manage. Rev., vol. 51, no. 2, pp. 170–193,2009.

[13] S. Fox, L. Marsh, and G. Cockerham, “Design for manufacture: A strat-egy for successful application to buildings,” Construct. Manage. Econ.,vol. 19, no. 5, pp. 493–502, 2001.

[14] C. Hitchcock, “The intransitivity of causation revealed in equations andgraphs,” J. Philosophy, vol. 98, pp. 273–299, 2001.

[15] K. Krippendorf, Content Analysis: An Introduction to its Methodology.Newbury Park, London, U.K.: Sage, 1980.

[16] S. Kvale, Interviews: An Introduction to Qualitative Research Interview-ing. Thousand Oaks, CA: Sage, 1996.

[17] Y. H. Kwak, Y. Y. Chih, and C. W. Ibbs, “Towards a comprehensive un-derstanding of public private partnerships for infrastructure development,”California Manage. Rev., vol. 51, no. 2, pp. 51–78, 2009.

[18] D. Lewis, “Causation,” J. Philosophy, vol. 60, pp. 17–25, 1973.[19] P. E. D. Love, D. Edwards, and J. Smith, “Contract documentation quality

and rework in Australian projects,” J. Architectural Eng. Design Manage.,vol. 1, pp. 247–259, 2006.

[20] P. E. D. Love, P. Davis, D. Baccarini, and D. Edwards, “Uncertaintyavoidance: Public sector clients and procurement selection,” Int. J. PublicSector Manage., vol. 21, no. 7, pp. 753–776, 2008.

[21] P. E. D. Love, D. J. Edwards, and Z. Irani, “Forensic project management:An exploratory examination of the causal behaviour of design-inducederror,” IEEE Trans. Eng. Manage., vol. 55, no. 2, pp. 234–248, May2008.

[22] P. E. D. Love, D. J. Edwards, J. Smith, and D. H. T. Walker, “Congruenceor divergence? A path model of rework in building and civil engineeringprojects,” ASCE J. Perform. Construct. Facilities, vol. 23, no. 6, pp. 480–488, 2009.

[23] P. E. D. Love, D. J. Edwards, Z. Irani, and D. H. T. Walker, “Projectpathogens: The anatomy of omission errors in construction and resourceengineering projects,” IEEE Trans. Eng. Manage., vol. 56, no. 3, pp. 425–435, Aug. 2009.

[24] N. Lowry, “Government procurement of capital works and competitionpolicy: Their effect on quantity surveying profession and ramifications forthe community,” Australian Inst. Quantity Surveyors, Brisbane, Queens-land, Australia, Res. Rep., 1996.

[25] M. R. Manavazhi and Z. Xunzhi, “Productivity oriented analysis of designrevisions,” Construct. Manage. Econ., vol. 19, no. 4, pp. 379–391, 2001.

[26] K. Petroutsatou and S. Lambropoulos, “Road tunnels construction costestimation: A structural equation model development and comparison,”Oper. Res., vol. 10, no. 2, pp. 163–173, 2009.

[27] C. Powell, “Responding to marginalisation,” Architectural Res. Quart.,vol. 2, pp. 84–89, 1997.

[28] C. C. Ragin, Constructing Social Research: The Unity and Diversity ofMethod.. Thousand Oaks, CA: Pine Forge Press, 1994.

[29] P. Raisbeck, C. Duffied, and M. Xu, “Comparative performance of PPPsand traditional procurement in Australia,” Construct. Manage. Econ.,vol. 28, pp. 345–359, 2010.

[30] J. T. Reason, “Human error: Models and management,” Brit. Med. J.,vol. 320, no. 7237, pp. 768–770, 2000.

[31] J. T. Reason and A. Hobbs, Managing Maintenance Error: A PracticalGuide. Burlington, VT: Ashgate Publishing Co., 2003.

[32] G. Rounce, “Quality, waste, and cost consideration in architectural build-ing design management,” Int. J. Project Manage., vol. 16, no. 2, pp. 123–127, 1998.

[33] D. Silverman, Interpreting Qualitative Data. London, U.K.: Sage, 2001.[34] M. Symes, J. Eley, and A. D. Seidal, Architects and Their Practices: A

Changing Profession. London, U.K: Butterworths, 1995.[35] D. H. T. Walker, “An investigation into construction time performance,”

Construct. Manage. Econ., vol. 13, no. 3, pp. 265–274, 1995.[36] Western Australian Auditor General’s Report. (2010, May 15). The

Planning and Management of Perth Arena. [Online]. Available:www.audit.wa.gov.au.

[37] G. M. Winch and E. Scheider, “Managing the knowledge based organiza-tion: The case of architectural practice,” J. Manage. Stud., vol. 30, no. 6,pp. 923–937, 1993.

[38] G. M. Winch and E. Scheider, “The strategic management of architecturalpractice,” Construct. Manage. Econ., vol. 11, pp. 467–473, 1993.

[39] R. Yin, Case Study Research: Design and Methods (Applied SocialScience Research Methods Series), vol. 5. Beverley Hills, CA: Sage,1984.

[40] F. Znaniecki, The Method of Sociology. New York: Farrar and Rinehart,1934.

[41] P. E. D. Love, D. Edwards, and E. Wood, “Loosening the Gordian Knot:The role of emotional intelligence in construction engineering, construc-tion and architectural management, vol. 18, no. 1, pp. 50–65, 2011.

[42] M. E. Porter, Competitive Strategy. New York: Free Press, 1980.

Authors’ photographs and biographies not available at the time of publication.

Documents

Moving Beyond Optimism Bias and Strategic Misrepresentation: An Explanation for Social Infrastructure Project Cost Overruns