Goodwin College of Professional Studies

Goodwin College of Professional Studies

Drexel E-Repository and Archive (iDEA)

http://idea.library.drexel.edu/

Drexel University Libraries www.library.drexel.edu

The following item is made available as a courtesy to scholars by the author(s) and Drexel University Library and may contain materials and content, including computer code and tags, artwork, text, graphics, images, and illustrations (Material) which may be protected by copyright law. Unless otherwise noted, the Material is made available for non profit and educational purposes, such as research, teaching and private study. For these limited purposes, you may reproduce (print, download or make copies) the Material without prior permission. All copies must include any copyright notice originally included with the Material. You must seek permission from the authors or copyright owners for all uses that are not allowed by fair use and other provisions of the U.S. Copyright Law. The responsibility for making an independent legal assessment and securing any necessary permission rests with persons desiring to reproduce or use the Material.

Please direct questions to [email protected]

http://www.drexel.edu

http://idea.library.drexel.edu/

www.library.drexel.edu

mailto:[email protected]

http://www.drexel.edu/goodwin/

I n terms of the broad approaches bywhich risks are analyzed, constructionproject risks can be broadly classified

as either objective or subjective.Risks that are purportedly analyzed by

the actual observation or calculation oftheir occurrence and impact on a projectare often described as objective risks.Analyzes of objective risks are quantitativein nature and often structured inprobabilities. They involve experimentalevidence, long-term experience, orcomplicated analytical calculations thatdescribe actual or potential risks. Risks thatare assessed based on beliefs about the risksrather than objective recorded risk data areoften referred to as subjective risks.

Analyzes of subjective risks are oftenqualitative and based on the analyst’sknowledge and experience of the risks andthe process by which the analyst selects andorganizes such knowledge and experiencesinto meaningful patterns.

The majority of construction contractrisks are subjective; there are ofteninsufficient historical data to enable theirobjective analysis. Therefore, their analysiswill, at best, be based on the subjectivepredictions of the analyst(s).

The use of rigorous analyticaltechniques for objective risks inconstruction is well reported in the

literature. However, reports of empiricalwork on the application of similar rigor tocontract risks such as payment delays,adverse ground conditions, etc. is scant.Evidence suggests that such risks are oftenanalyzed in a somewhat arbitrary manner.

Contractors tend to resort to theaddition of a single arbitrary percentagecost contingency to give their overallimpression of the total risk rather thanassessing the risks that they are asked tocarry. However, applications of analyticalrigor to similar risks in other industriespoint to the potential benefits that suchmethods present to the constructionindustry. For example, subjectiveprobability forecasting has been usedreliably, in combination with classicalmethods involving objective probabilities,in short-range weather forecasting.

Elicitation of expert beliefs has beenused in the development of probabilisticexpert systems for the diagnosis ofcongenital heart disease, and in populationprojection [12, 18]. Elicitation of priorbeliefs has also been successfully used inestimating the future maintenance costs ofwater treatment plants, in determining thehydraulic conductivity of rocks for nuclearwaste repository development, and inuncertainty analysis for radiologicalprotection [13, 14].

Subjective predictions made on theweather are based on the forecaster’sknowledge and experience of weatherpatterns in much the same way as aconstruction expert’s knowledge andexperience of soil characteristics that leadto risks such as adverse ground conditionsdrive his predictions about such risks.Furthermore, unlike recurring events thatare subject to the laws of large numbers,most contract risks represent rare eventsamples for which analysis of long-runfrequency and traditional statisticalapproaches are often inapplicable.

A. Tversky argues that predictions andanalysis of such subjective risks are oftenbased on the intuitive judgements of thedecision-maker [19]. These intuitivejudgements and estimates are influencedby individual perceptions and biases thatare often not addressed during the analysis.Appropriate analytical methods forconstruction contract risks are thus neededto enhance risk management efficiency.Bayesian methods, for example, enablesubjective opinions about uncertainties tobe combined with sample evidence aboutthe risk to form posterior probabilitydistributions of the risk. Such distributionscan then be used as input variables in thesystematic analysis of the uncertainty. Theadjustment of the individual expert’sestimate by the available sampleinformation will lead to a more accuraterepresentation of the probability of the risk.

J. Ansell (1992) argues that one of themajor problems in evaluating the risk of asystem is the identification of the full rangeof risks to which the system could besubject [3]. The identification process ismade difficult especially since what isconsidered “risk” is significantly influencedby perceptions that are very dynamic innature. A thorough identification of bothinternal and external project-related risksrequires that the risks analyst is not onlysystematic and experienced, but alsocreative.

J. Raftery argues that the best way togain access to such a range of qualities is touse a team of experienced constructionprofessionals or experts [15]. Thus, aconstruction contract risk managementapproach that uses appropriate methods ofeliciting and aggregating opinions frommultiple experts will be better at reducingthe impact of individual perceptions andbiases on the estimates and enhancing riskmanagement effectiveness [1].

22 Cost Engineering Vol. 50/No. 1 JANUARY 2008

TECHNICAL ARTICLE

Construction Contract RiskManagement: A Study of Practicesin the United Kingdom

ABSTRACT: Although formal analytical processes are applied to the management of economicrisks in construction projects, very little is reported on the application of similar processes to themanagement of construction contract risks. Evidence from other industries however, point tothe benefits that such rigorous approaches will offer the construction industry. This article pres-ents the findings of a study conducted to assess the extent to which available techniques foridentifying and analyzing risks are applied to construction contract risks by construction pro-fessionals in the United Kingdom. The study also sought to evaluate the appropriateness of thepredominant techniques used by determining whether the techniques adequately address thenature of contract risks and the significant impact that personal perceptions and biases have ontheir analysis. The study employed a structured questionnaire survey to collect primary data col-lection for analysis in order to provide an empirical basis for the major conclusions of theresearch study. The study found that the predominant approaches to construction contract risksidentification and analysis are inappropriate as they rely heavily on single expert assessmentsand as such do not address the effect of individual perception and biases on the subjective esti-mates used in their analysis.

KEY WORDS: Contract risks, analysis, identification and management

Dr. Francis K. Adams

Aims and Objectives of the StudyThis article presents the findings of a

study conducted to assess the extent andfrequency with which available techniquesfor identifying and analyzing risks areapplied to construction contract risks bythe various professions within the UKconstruction industry.

The study also sought to evaluate theappropriateness of the predominantpractices of contract risk identification andanalysis in the UK. Appropriateness in thiscontext was determined by whether theprocesses and techniques used adequatelyaddress the nature of contract risks and thesignificant personal biases and perceptionsto which their analysis are subjected.

In essence, the study sought to test thefollowing operational assertions:

• There is very little application, if any,of systematic and rigorous probabilisticmethods to contractual risk inconstruction. And,

• Analytical methods currently used tomanage contractual risks inconstruction do not adequately dealwith the effect of perception on thesubjective estimates used in theseanalytical techniques.

These assertions are theoreticalconstructs that needed testing with real liferesponses. The research therefore adopteda positivist approach, using structuredstandardized questionnaires as the mainmethod of primary data collection. Majorconclusions of the research are thereforeempirically based; induced from theanalysis of the data collected.

The Risk Management ProcessAlthough opinions vary in the

literature as to what constitutes the stagesin the systematic process of project riskmanagement, experts generally agree onthe intended objectives, content, andoutcomes of the total process.

Within the context of constructioncontract risks, the objectives of the riskmanagement process can be summed upin the following.

• To gain a greater awareness andunderstanding of the types and natureof risks inherent in the project, and thelikelihood of their occurrence.

• To assess the potential impact of therisks on the viability of the project andcontract and to determine how best toeliminate or control the risks. And,

• To inform the development of themost appropriate contract strategies byhelping to identify which of theproject parties are best placed to takeresponsibility for the risks and theoptimum premiums or contingenciesto allow for the risks.

For the purposes of the study presented inthis article, the following three-stage riskmanagement system was adopted [2].

• Risk identification: ‘the process ofsystematically and continuouslyidentifying, categorizing, and assessingthe initial significance of risksassociated with a construction project’.

• Risk analysis and evaluation: ‘theprocess which incorporatesuncertainties in a quantitativemanner, using probability theory, toevaluate the potential impact of risk.’

• Risk response management: ‘strategiesaimed at removing as much as possiblethe potential impact of risk orincreasing control of risk.’

Methods for Identifying Project RisksAn effective risk identification

technique would aim at answering thequestion: “What are the discrete features ofthe project or its general environment (risksources) which might cause failure?”

The technique would thereforeinvolve an investigation into all possiblepotential sources of project risks, theinterrelationships among them, and theextent to which they can be controlled.

Methods for identifying risks inconstruction projects are also very welldescribed in literature. J.E. Diekman andothers, C. Chapman and S. Ward, forexample, discuss the main range oftechniques available for identifying risks [8,10]. Those surveyed in the present studyare summarized as follows.

Types of Assessments• In-house individual expert assessment• In-house multi-disciplinary/panel

group assessment• In-house synectic team assessment

Assessment Techniques• pondering or “what can go wrong”

analysis;• free and structured brainstorming;• synectics;• checklists;• risk records;• prompt lists; and• structured/expert interviews.

Techniques for Analyzing ContractualRisks in Construction

The risk identification process wouldhave highlighted risks that may beconsidered by project management to bemore significant and selected for furtheranalysis.

The objective of a risk analysistechnique is therefore to determineprobability numbers that quantity beliefsabout uncertainty and thereby quantify theeffects on the project of the identifiedmajor risks. This analysis progresses fromthe collection of data relevant to themodeling of the uncertainty through theanalyzes of the frequency, severity, andprobability of the risk event, to theassessment of the consequences andimpact of the risks on the risk targets.

The range of techniques available foranalyzing and evaluating risks are alsosummarized in the following.

Techniques for Assessing Risk Likelihood

• quantitative probability assessmentsbased on historical data.

• subjective probability assessmentsbased on expert judgment. And,

• scaled assessments (e.g.,high/medium/low) based onexperience

Techniques for Assessing Risk Severity

• quantitative probability assessmentsbased on historical data.

• subjective probability assessmentsbased on expert judgment.

• scaled assessments (e.g.,high/medium/low) based onexperience. And,

• Individualized estimation of actualseverity should risk occur. Thisapproach differs from the three abovein that detailed calculations of the costto be incurred to time to be lost aredone for each key risk, assuming therisk occurred.

Cost Engineering Vol. 50/No. 1 JANUARY 2008 23


Techniques for Assessing Risk Impact

• decision trees analysis;• fault trees analysis;• event trees analysis;• probability analysis;• sensitivity analysis;• scenario analysis;• simulation analysis; and• ranking options.

The risks analysis techniquessummarized above are not exhaustive, butreflect the key models that are commonlyused on construction projects. Forexample, although techniques such ascatastrophe theory, fussy set analysis, gametheory and multi-criterion decision makingmodels were known by most practitionersin the UK.

S.J. Simister (1994) reported that thesetechniques were either not used at all orused by only a handful of practitioners [16].

Research DesignAccording the findings of D. Simonton

in 1996, and the analysis of S. Vick, in2002, on the development of engineeringexpertise, it would appear that engineeringexperts reach the height of their expertisebetween the career ages of 10 and 33 whichcorresponds to chronological ages of 35and 53 [17, 20].

Against this background, and in viewof the specialist nature of constructioncontract risks and the relative infrequencyof such risks, respondents were selectedfrom construction experts based on aminimum of 10 years of industrialexperience.

All the experts were sampled randomlyfrom the United Kingdom (UK) usingdirectories such as the ‘Chartered BuildingCompany’ Directory, the ‘Contractors’File’ and the ‘Consultants’ File.’

Telephone and fax surveys were thenused to identify qualifying and interestedparticipants. Quantity surveyors,construction managers, contracts managersand project managers are the mainconstruction professionals usuallyresponsible for project risk management.

To achieve optimum efficiency in thestudy, the mail survey focused on suchprofessionals from the civil engineeringand building construction industrialsectors. These two sectors were chosen asthe study was concerned with risks arisingfrom contracts in these industries. 160

experts comprising 40 each from the fourprofessions above were selected.

The research instrument comprisedmainly of a set of questions each of whichhad a domain or a set of answers fromwhich respondents were to choose. Theoption was also provided in each domain ofanswers for respondents to indicate theirresponses if their answer was not includedin the list provided.

In addition to the benefit of being ableto test the research constructs with real liferesponses, the questionnaire approach wasconsidered beneficial because of its highefficiency in terms of research cost, time,and effort. The lower cost of using self-completing questionnaires was consideredparticularly important since the samplewas widely spread geographically.

Also, the questions and responsedomains in the research used language thatis part of the training and the regularprofessional practice of both the researcherand respondents. This would enhance thebenefit of reduction in biasing errors thatself-completion questionnaires provide [9].One other advantage of the questionnaireapproach is the reliability of measurementsand its amenability to making statisticalinferences and generalizations from datacollected.

Restricted domain of responsesensures that consistent responses areobtained over all respondents. In addition,since responses have to lie in givendomains, applying formal statisticaltechniques in the analysis is a relativelystraightforward process. The potentialadverse impact on the research by theconstraints imposed on responders by thepredetermination of the appropriatequestions to ask and their response domainswould also be minimized by the significantexperience and knowledge of the fieldunder investigation, as well as extensiveliterature review by the researcher.Consultations were also made with otherresearchers and construction experts toascertain the appropriateness of thequestions asked and the completeness ofthe response domain for each question.

Results and AnalysisBefore proceeding to use the survey

results to test the hypotheses, it is necessaryto explain some basic computationsapplied to the data collected. An inspectionof the questionnaire (Appendix 1) showsthat apart from questions seeking factual

information about respondents, thequestions asked respondents to indicatehow often (on a 5-point scale of: never,occasionally, frequently, very frequently,always) they use any of the key techniquesand methods identified through theliterature for identifying and analyzingcontractual risks.

Answers to factual questions are in theform that can be readily applied in analysisof the respondents and thus the variousprofessions within the constructionindustry. Frequency tables were alsoconstructed and inferences based on theresults. For the analysis of responses to the5-point scale questions, first a carefullyconsidered weight reflecting the degreeusage of a particular technique beingmeasured is assigned to each possibleanswer. Thus, responses were weightedaccording to the following scale:

Response Scale ValueAlways 4Very Frequently 3Frequently 2Occasionally 1Never 0

Using these scale values, the responseswere then converted into rating values thatare later plotted to provide a summary viewof the risk identification and analysispractices within the industry. The totalrating value (R) for a risk identification oranalysis technique, T, was computed as:

equation 1

where ni is the total number of respondentsassigning scale value i to technique T. As anillustration, consider the response to thequestion on how often respondents usedpondering in seeking to identify things thatcould go wrong on a project. Details of theresponses and the corresponding ratingscale values provided by the respondentsare provided in table 1. From equation 1,the total rating value for pondering iscomputed as 86 with a mean rating value of2.97:

equation 2


The approach and rating scale issimilar to that used by J.F. Burchett andothers in the analysis of the extent of riskidentification within electrical supplyprojects, and by A.Y. Antwi for the analysisof urban land markets in Ghana [4, 7].

It should be pointed out that since therating scale is applied in the same manneracross all responses to the differenttechniques, direct comparisons of theresulting rating values can be madebetween different techniques and differentsets of questions. Mean values calculatedusing this scale thus also represent themeasures of central tendency of the usagesof the different techniques [7].

Characteristics of the SurveyRespondents

The breakdown of the responses fromthe 160 questionnaires sent to the selectedprofessionals is given in table 2. The“Other” category represents respondentswho are now working in differentprofessional capacities from the ones forwhich they were selected. They include abuilding manager, an estates director, twomanaging directors and a facilitiesmanager. Seven other participantsresponded to indicate that they wereunable to complete the questionnaire for anumber of reasons (e.g. lack of relevantexperience, busy personal schedule, etc).These were not included in the analysis.

As table 2 shows, the majority ofrespondents were either quantity surveyorsor project managers. This may be reflectiveof the fact that as one of the primaryrepositories of construction cost and riskinformation, quantity surveyors are bestplaced to provide the required information.

Although the effective response rate of18 percent appears small, it is consistentwith survey responses within theconstruction industry generally and in theUK specifically [6, 7, 19].

The low response rate fromconstruction managers was mainlyattributed to either inexperience or lack oftime. These were also the main reasonsgiven for non-response by the other 131survey subjects who were unable toparticipate in the study.

Two facts explain this reason. The firstis that in view of the relative infrequency ofcontractual risks, experts would need tohave been in the field that provides themwith experience of the risks for significantnumber of years for them to be able to

make informed judgements about the risks.Secondly, for most non-responses, thequestionnaire was completed by a juniorstaff member in the original expert’scompany because of a lack of time on thepart of the selected expert. The problemwas that the average experience amongsuch junior staff members (as revealed bythe average years of experience of the non-respondents) was between 2-5 years.

Respondents also came from a varietyof industrial sectors: 32.6 percent fromquantity surveying consulting, 21.7 percentfrom building construction, 13.0 percentfrom project management consulting, 6.5percent from civil engineering and 8.7percent from each of legal and contractsconsulting and property development.

With the exception of two respondentswho were construction managers, all therespondents had more than 10 years ofindustrial experience (94 percent had morethan 15 years of construction experience),and over 80 percent had more than 10years of experience in their currentprofession within construction. This highpercentage of highly experienced

respondents is very significant in the lightof the works of D. Simonton in 1996, andS. Vick in 2002, and lends further supportto the credibility of the informationobtained [20, 23].

The annual turnovers of almost 70percent of the respondents were less thanfive million pounds (41.4 percent hadturnovers less than £1 million, 27.6percent between £1 million and £5million, 13.8 percent between £16 millionand £25 million, 6.9 percent between £26million and £50 million, and 10.3 percentover £50 million). This contradicts thebelief that formal risk managementapproaches could only be afforded bylarger companies because of perceivedextra cost of implementing such systems.

Extent of Use of Risk AssessmentApproaches

Among the four professional categoriesstudied, the task of risk assessment appearsto be undertaken predominantly by oneindividual within the organization. Thisapproach was used by about 76 percent ofthe respondents.

Table 1 — Respondents’ Ranking of Their Use of Pondering for Risk Identification

Table 2 — Profile of Survey Sample

This is particularly so among quantitysurveyors, about 83 percent of whom usethis approach (see figure 1 (a)). As figure 1(b) shows, this usage is consistent withlevels of usage within the quantitysurveying sector of the industry where over93 percent of respondents use thisapproach compared to the 13 percent whouse the in-house multidisciplinary groupapproach.

Whereas a case justifying this practicecan be made in view of the high level ofexperience of those involved in the study, ithas been generally argued that the complexand dynamic nature of constructionprojects requires more experience inidentifying project risks than one expertcan provide [5].

Modern construction spans severalindustries. Within the constructionindustry itself, technological advancementshave created such myriad specializations inproduct and component technologies thatno one professional can any longer claim tobe a sole repository of construction riskknowledge.

The in-house Individual approach isthe most susceptible to personal biases andperceptions and would be considered theleast suitable among three main assessmentapproaches for contractual risk analysis,unless the projects being analyzed werevery familiar, highly identical, or repetitiveof previous projects for which rigorous riskanalysis had previously been conducted.This hardly ever happens in construction.

About 24 percent of respondents usedthe in-house multidisciplinary groupapproach which also appears to be thestandard practice within the civilengineering sector (100 percent ofrespondents as shown in figure 1 (b)). Thismust however, be interpreted against thebackground that the civil engineeringsector makes up only 6.5 percent of therespondent population. Only about 14percent of respondents used the in-housesynectic team approach and 7 percent ofrespondents used other approaches

Extent of Use of Risk IdentificationTechniques

Figure 2 summarizes the results of thesurvey concerning the use of riskidentification techniques in the UKconstruction industry.

Pondering appears to be the key riskidentification technique employed in theindustry. All the respondents use this

technique to varying degrees (rating values:mode = 4.00; median = 3.00; mean = 2.97;inter-quartile range = 2.00 - 3.00) and about85 percent of respondents use it at leastfrequently.

This is followed by the use ofchecklists. Like pondering, all but one ofthe respondents uses this method, about 72percent of respondents using it at leastfrequently.

It is worth noting that the riskidentification techniques listed in thesurvey instrument appear to be the onlyones with which that the greater majority(97 percent) of the respondents werefamiliar.

Among the risk identificationtechniques, synectics and expert interviewsare the least used among almost all theprofessions. These results are consistentwith results from the analysis by industrialsector.

The higher usage of pondering andchecklists (compared to synectics andexpert interviews) can be attributed to theease with which one person can use suchtechniques. In terms of corporateeconomics, they are cheaper techniquestoo!

While the practice of pondering andchecklists are good approaches to riskidentification, it is doubtful if they canadequately highlight all the risks in a


Figure 1 — Usage of Risk Assessment Approaches

complicated construction project,especially if these techniques are being usedby just one individual on the project. Asstated earlier, the multidisciplinary, cross-industrial and technologically specializednature of modern construction makes itinappropriate for one professional toassume sole responsibility for theidentification of construction project risks.

Extent of Use of Risk Likelihood andImpact Assessment Techniques

Tables 3 and 4 summarize the results ofthis area. The predominant practices here(including practices among the professionsand in different industrial sectors) involvethe use of scaling methods and subjectiveprobability assessments.

This is not surprising, as contractualrisks by their nature do not lend themselveseasily to the use of quantitative probabilityassessments. However, the fact that thesetechniques are not ‘always’ or ‘veryfrequently’ used would seem to confirm thework of R.W. Hayes and others in 1986[11]. It reported that contractors hardlyassess the separate risks that they are askedto carry, but resort to the addition of a singlepercentage cost contingency. This gives anoverall impression of their perception of thetotal risks that they are asked to carry.

These findings are also consistent withthe findings of S.J. Simister in 1994, in hissurvey of construction project risk analysistechniques used in the UK, and those of J.F.Burchett and others in the 1999 worldwidesurvey of risk management practices amongelectrical supply companies [7, 19].

The industry as a whole does not seemto have made significant progress inemploying available analytical techniquessince the 80s! Although scaling methodsand subjective probability assessments areused in risk likelihood and impactassessments, they are on average not used“frequently.’ The assessments are generallyconducted using the in-house individualapproach.

This combination fails to maximize thepotential benefits of the subjectiveprobability approach. The assessmentsbecome heavily subject to effects of thepersonal perceptions and biases of theindividual. After over a decade oftechnological, research, and managementadvances, the construction industry doesnot appear to have shifted very significantlyfrom old practices.


Figure 2 — Usage of Risk Identification Techniques


Risk Analysis TechniquesIt is evident from table 5 that risk

analysis in the form that is applied ineconomic risk analysis is very much anunexplored area when it comes to contractrisk analysis.

Generally, none of the techniquessurveyed by the study is used to anysignificant degree. On average, probabilityanalysis, sensitivity analysis, scenarioanalysis, and ranking options are generallyeither never used or used only occasionally.

It is also significant to note that most ofthe experts who use the probabilistictechniques were from comparativelysmaller companies with annual turnovers ofless than £5 million, and that no othermethods of risk analysis are used by any ofthe respondents.

At first sight, the results appear tocontradict the findings of S.J. Simister in1994, in whose study about 72 percent and60 percent of respondents indicated thatthey currently used Monte Carlosimulation and sensitivity analysisrespectively [19]. However, Simister’s 1994study merely reported how many of thesurveyed participants currently use thetechniques but did not investigate howfrequently they use them [19].

The majority of the 72 percent thatused Monte Carlo simulation according toS.J. Simister’s 1994 study could well beoccasional users of the probabilisticapproach [19].

The present study does not onlymeasure the numbers of participants whouse the various approaches, but also howfrequently they actually use the differentapproaches.

Even though scaling methods andsubjective probability assessments are thetechniques used most frequently forevaluation risk likelihood and impact, theyare only used occasionally when makingfinal analytical decisions about the risks.

This suggests that the estimates that arederived for risk likelihood and impact areoften single point estimates and not of thetype that can be used in a rigorousprobabilistic analysis such as Monte Carlosimulation. Although management systemsare becoming more reliable and efficientand construction project environmentsmore complex, it appears that not mucheffort is being made by the industry toincorporate available and adaptable systemsinto its risk management practices.

Testing the HypothesesHypothesis (i) is considered as

supported if the majority of the participantsuse it “less-than-frequently.” Thus, on therating scale of 0 to 4 explained earlier, amean rating value of less than “2” (whichrepresents “frequent” use) would indicate asupport for the hypothesis.

The various risk identification andanalytical techniques and their capacity todeal with the effect of perception on thesubjective estimates used in risk analysiswere discussed earlier.

Hypothesis (ii) is considered assupported if the majority of the participantsfrequently use those techniques that do notadequately deal with the effect ofperception on the subjective estimates, or ifthey use techniques that adequately deal

with the effect of perception on thesubjective estimates “less-than-frequently.”

Thus, on the rating scale of 0 to 4, amean rating value of more than “2” (whichrepresents “frequent” use) for the first set oftechniques, less than “2” for the techniquesthat deal with perception would indicate asupport for the hypothesis (ii). Table 6presents a listing of all the risk managementtechniques evaluated in this study ranked inthe order of their overall mean ratingvalues.

Testing of Hypothesis (i)For risk Identification, the only

techniques with mean ranking values ofmore than “2” are pondering andchecklists. Synectics and expert interviewsthat have the highest potential for

Table 3 — Usage of Risk Likelihood Assessment Techniques

Table 4 — Usage of Risk Impact Assessment Techniques


generating the broadest listing of risks havemean rating values of less than one,indicating that at best, they are only usedoccasionally. The mean rating values forrisk likelihood, risk impact and risk analysistechniques are all less than “2.” It is worthalso noting that these techniques (includingthe risk identification techniques) aregenerally used within an in-houseindividual assessment setting. The surveyresults support hypothesis (i).

Testing of Hypothesis (ii)The only contractual risk management

techniques that can adequately deal withindividual perceptions and biases aresubjective probability estimates derivedwithin a group setting (synectics or in-housemultidisciplinary group) and probabilityanalysis including simulation analysis. Themean rating values for these techniques areall less than 2. Furthermore, they are alsogenerally used within an in-houseindividual assessment setting. These resultssupport the hypothesis (ii).

It is evident from these results thatthere is a significant gap between thetechniques available to the constructionindustry and what are actually used in themanagement of contractual risks.Contractual risks by their nature make ithighly unlikely that one individual will

Table 5 — Usage of Risk Analysis Techniques

Table 6 — Ranking of Mean Rating Values of Risk Management Techniques


have sufficient first hand experience of eachrisk to enable him/her to conduct accurateidentification and analysis. That is withoutmaking the whole risk managementexercise heavily subject to the errors causedby personal biases and perceptions.

This is supported by the fact thatalthough 160 professionals were selected toparticipate in the study, based on at least 10years of industrial experience, only 29believed they had sufficient experience toenable them respond to the surveyquestions. Yet the predominant practices inthe industry seems to center on oneindividual dealing with risk management.These practices do not adequately takeaccount of the nature of contractual risksnor help to make risks explicit.

The economic practicality of bringingtogether a team of experts for risk analysison small projects is undoubtedlyquestionable, but it is possible to employavailable computing and informationtechnology or consultative approaches (e.g.expert interviews) in overcoming thedisadvantages of the individual expertassessment approach.

This appears not to be the case in theindustry generally. Very few professionalsuse risk analysis techniques such asprobability analysis, sensitivity analysis, andscenario analysis. Those who use themappear to do so only occasionally. There is ause of scaled and subjective judgmentalprobability assessments techniques tovarious degrees, but very little application ismade of these assessments. Suchassessments would include input variablesfor a formal, systematic probabilisticanalysis and quantification of the risks.

Further interviews with some of theexperts suggest the perception that such aformalized process is perhaps only withinthe resource capabilities of the very largeconstruction firms. This perception is notsupported by the findings of the research.On the contrary, the majority (about 68percent) of companies using all the variouskinds of risk management techniques havea turnover of less than £5 million. In fact,companies with turnover of under £1million, who use the various techniques,account for over 40 percent of therespondents.

The author’s view is that the use offormalized processes for risk management isnot, and need not, be only within theresource capabilities of very largeconstruction firms. This is based upon

advances in modern technology andevidence from this research and from otherindustries.

Most of the techniques surveyed areused on a regular basis by other industriesthat are prone to similar risk as those facedby the construction industry, and areconsidered generally beneficial to the riskmanagement effort.

I t is evident from the study that theextent of application of systematic andrigorous probabilistic methods to

contract risks in construction is very scant.Also, the analytical methods currently usedto manage such risks do not adequately dealwith the effect of perception on thesubjective estimates used in these analyticaltechniques.

The reliability of these findings isreinforced by the results obtained throughthe analysis of the mean rating values(MRV) of the various techniques. The onlyrisk management techniques used that hada MRV of more than “2” (“2” signifying thatthe technique is used frequently) arepondering and checklists.

Assessments are also largely done bysingle individual in an organization.Construction contract risks are rare bynature and largely undocumented. It istherefore highly unlikely that oneindividual will have sufficient first handexperience of each risk to enable him/herconduct accurate and thoroughidentification and analysis of such risks inany major construction project.

This fact was evident in the survey bythe fact that although 160 professionalswere selected in the UK, based on at least10 years of industrial experience, toparticipate in the study, only 29 believedthey had sufficient experience to enablethem respond to the survey questions.

The study highlights the fact that thereis a significant gap between the techniquesavailable to the construction industry andwhat are actually used in the managementof contract risks. Further research is neededin investigating the use of availabletechniques that fully address the subjectivenature of construction contract risks. Suchtechniques should also be able to obtainsubjective risk estimates for analysis in amanner that minimizes the impact ofindividual perceptions and biases on theestimates used for the analysis. ◆

REFERENCES1. Adams, F.K. The Management of Risks

in International Infrastructural Projects.PhD Thesis. The University ofEdinburgh. United Kingdom, 2004.

2. Al-Bahar J.F., and K.C. Crandall.Systematic Risk Management Approachfor Construction Projects, Journal ofConstruction Engineering andManagement; 116, 3 (1990):533-546.

3. Ansell J. Reliability: Industrial RiskAssessment. J. Ansell and F. Wharton,editors. Risk: Analysis, Assessment andManagement. UK: Wiley, 1992

4. Antwi A.Y. Urban Land Markets inSub-Saharan Africa: A QuantitativeStudy of Accra Ghana. Ph.D. Thesis,Napier University. United Kingdom,2000.

5. Ashley DB, S.L. Stokes, and Y.H.Perng. Combining Multiple ExpertAssessments for Construction RiskIdentification. Proceedings of the 7thInternational Conference onOffshore Mechanics and ArcticEngineering, Houston, Texas, 1988. p.183-192

6. Baker S, D. Ponniah D, and S. Smith.Risk Response Techniques CurrentlyEmployed for Major Projects.Construction Management andEconomics, 17 (1999): 205-213.

7. Burchett, J.F., V.M. Rao Tummala andH.M. Leung. A Worldwide Survey ofCurrent Practices in the Managementof Risks Within Electrical SupplyProjects. Construction Managementand Economics 17 (1999): 77-90.

8. Chapman C, and S. Ward. ProjectRisk Management: Processes,Techniques and Insights. UK: Wiley,1997.

9. Cicourel A.V. Methods andMeasurement in Sociology. NewYork: Free Press, 1964.

10. Diekmann J.E., E.E. Sewester and K.Taher. Risk Management in CapitalProjects. Texas: CII, University ofTexas at Austin, 1988.

11. Hayes R.W., J.G. Perry,P.A. Thompsonand G. Willmer. Risk Management inEngineering Construction. London:Telford, 1986.

12. Kadane J.B., and L.J. Wolfson.Experiences in Elicitation. TheStatistician. 46, 4 1997): 1-17.

13. O’Hagan, A. Eliciting Expert Beliefs inSubstantial Practical Applications; TheStatistician. 46, 4 (1997).


Appendix 1— Survey Questionnaire


Appendix 1— Survey Questionnaire Continued


14. O’Hagan A, and R.G. Haylock.Bayesian Uncertainty Analysis andRadiological Protection. V. Barnett andK.F. Turkman, editors. Statistics for theEnvironment 3, Pollution Assessmentand Control, Chichester, UK, 1997. p.109-128

15. Raftery J. Risk Analysis in ProjectManagement. London: E & FN Spon,1994.

16. Simister S.J. Usage and Benefits ofProject Risk Analysis and Management.International Journal of ProjectManagement, 12, 1 (1994):5-8

17. Simonton D. Creative Expertise: ALifespan Developmental Perspective. K.Ericsson, editor. The Road toExcellence: The Acquisition of ExpertPerformance in the Arts and Sciences,Sports and Games, Erlbaum,Mahwah, N.J. p. 227-253.

18. Spiegelhalter D.J., N.L. Harris, K. Bulland R.C.G. FranklinG. EmpiricalEvaluation of Prior Beliefs AboutFrequencies: Methodology and a CaseStudy in Congenital Heart Disease.Journal of the American StatisticalAssociation, 79, 387 (1994):489-500

19. Tversky A. Assessing Uncertainty.Journal of the Royal StatisticalSociety, Series B, 36, 2 (1974):148-159.

20. Vick S. Degrees of Belief - SubjectiveProbability and EngineeringJudgement. Virginia: ASCE Press,2002.

RECOMMENDED READING1. Dey P.K. Quantitative Risk

Management Aids RefineryConstruction. HydrocarbonProcessing (International Edition) 81,3 (2002).

2. Flanagan R, and G. Norman. RiskManagement and Construction.Oxford: Blackwell ScientificPublications, 1993.

3. Godfrey P.S. Control of Risk: A Guideto the Systematic Management ofRisk from Construction. London:CIRIA, 1996.

ABOUT THE AUTHORDr. Francis K. Adams is an assistant

professor of construction managementwith Goodwin College of ProfessionalStudies in Philadelphia, PA. He can becontacted by sending e-mail [email protected].

Documents

Goodwin College of Professional Studies