Professional Documents
Culture Documents
Article information:
Access to this document was granted through an Emerald subscription provided by emeraldsrm:462515 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.
Paper submitted 13 July 1994; accepted for publication 28 July 1995; discussion open until
September 1996
Abstract The aim of the paper is to examine attitudes of general contractors operating in
the Sydney region to the potential use of probability estimating and databases in cost
estimating. A sample of 10 large general contractors with a turnover over $100m was
selected for the study, which took place in 1993. Responses of the contractors to a
standard questionnaire were obtained using face to face interviews. The research
described in this paper confirmed the popularity of traditional single value estimating and
highlighted the lack of use of probability cost estimating by the general contractors surveyed. The limited availability of client-prepared bills of quantities for tendering has neither diminished their popularity among bidding contractors nor increased the use of
elemental cost planning. Although databases are generally available, subjective judgements of estimators are of greater value in cost estimating. The research has concluded
that a change in the estimating paradigm towards probability cost estimating, and the use
of databases, are unlikely to occur in the near future.
Keywords bill of quantities, construction, database, general contractors, probability
estimating
INTRODUCTION
Tendering is the commercial function through which a contractor compiles a bid
for a project on the basis of the net cost estimate with provision for general
overhead and profit. The objective of every contractor is to submit a bid that gives
an appropriate balance between its competitiveness and the desired profitability.
The bidding process emphasizes the need for a sound knowledge of bidding
strategies and the ability to prepare a stable and accurate estimate of project costs.
Estimating is the technical process carried out by a contractor in order to calculate the likely net cost of a proposed project from available information. It can
be approached in two ways:
1 Using the traditional single value estimate with the addition of a contingency
allowance
2 By probability estimating where estimates of variable cost items are expressed in
83
84
Uher, T.E.
the form of ranges (probability distributions) of possible values and the combined effect ofrisks on the final cost determined mathematically or stochastically.
Traditionally, contractors compile bid prices in the form of single-value estimates. A bid price comprises explicitly in addition to the value of materials,
labour, plant and preliminary cost items - a risk contingency. A risk contingency is
an assessment of the assessed impact of risk on a project and is commonly
expressed as a percentage of a project cost. It is the goal of every estimator to be
able to perceive the presence of risks, and accurately predict their magnitude and
likely impact on the project. However, there is little evidence (Perry & Hayes
1985) that contractors undertake a systematic and detailed assessment of individual risks and the expression of a contingency is often at best an educated guess of
a stab in the dark.
Risk and uncertainties are inherent in construction. While they are almost
always present, their presence may not necessarily be a problem, particularly
where their impact is low. However, even high impact risks may serve a useful
purpose as an estimator may want to utilize a greater level of risk to generate a
potentially higher level of income. The effective management of risk during cost
estimating requires the application of a systematic and disciplined approach to the
identification and analysis of, and response to, risks. Probability estimating provides a suitable framework for such an effective management of project risks.
Because it focuses the attention of estimators on ranges (probability distributions)
of possible values of various risk variables, it helps to model project variables
vigorously and systematically and provides estimators with a more robust and
reliable decision making tool (Raftery 1990).
Diekmann (1983) outlined four probability estimating methods suitable for
cost estimating:
1 Direct analysis technique
2 Methods based on the central limit theorem
3 Approximate methods for general estimating models
4 Simulation techniques such as the Monte Carlo method.
These and other methods of cost estimating, such as the successive principle
approach, have been reviewed by Toakley (1989). Probability estimating using the
Monte Carlo method has been the main focus of this research.
A successful application of probability estimating is dependent on both the
structural requirements, such as the development of appropriate modelling tools
and the availability of data, and human (cultural) factors such as knowledge, skills
and attitudes of decision makers. Bills of quantities have traditionally provided a
satisfactory structure for the generation of contractors' single value cost estimates.
However, this structure is inappropriate for probability estimating due to the very
large number of cost items in a typical bill of quantities. Jaafari (1990) and Uher &
Levido (1992) proposed the use of an elemental cost planning structure for
probability estimating while Brotherton (1993) preferred a trade summary
structure. These issues will need to be carefully addressed in developing the most
appropriate approach to probability estimating.
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 I 1, 2, 83-95
85
86.
Uher, T.E.
Norman (1984), Perry and Hayes (1985), Cooper & Chapman (1987), Jaafari
(1990), Raftery (1990), Uher (1990, 1991), Morgan (1991), Moselhi et al. (1991),
Rowland & Curran (1991), Shafer (1991) & Yeo (1991), whose work promotes
the use of probability estimating in preference to the traditional single value
(deterministic) estimating method.
Rowland & Curran (1991) & Uher (1991) believe that because deterministic
estimates are largely based on past experiences of estimators and/or historical data
sets, they are often expressed as averages or estimates close to average of such past
data sets. In probability terms, such deterministic estimates represent 50% or
higher probability or 50% or lower probability that they would actually be
achieved, which represents a risky decision. Probability cost estimating enhances
the decision making process by systematically assessing the impact of the key
project variables on an estimate and by generating a range of possible decision
outcomes.
While unit cost rates and monetary allowances, such as prime cost items and
contingencies, are the most common examples of risk variables in a probability
estimate, risks associated with quantities can also be expressed as risk variables.
An enhanced probability cost estimate can further include modelling of correlations among different risk variables.
Considering the risky nature of the construction industry and the degree of
exposure of contractors to risk under a typical construction contract, contractors
would be expected to embrace estimating methods which reduce their exposure to
risk and improve their decision making capabilities. However, probability estimating is largely ignored by contractors and only a small number of specific
applications have been reported so far in the US construction industry (Golden
1991; Moselhi et al. 1991; Rowland & Curran 1991; Shafer 1991; Townley 1991).
Risk variables in a probability cost estimating model need to be expressed as
probability distributions. In the absence of data bases, rigorous and systematic
elicitation and analysis of subjective data is required. Techniques such as Delphi
analysis and a fractile method (Pratt 1964; Morrison 1967; Raiffa 1968; Winkler
1967; Huber 1974; Herz & Thomas 1983), and the use of simple uniform and
triangular probability distributions (Raftery 1990) meet the requirements. Some
researchers, notably Betts (1987, 1991), Ivkovic (1991) and Townley (1991)
believe that there is no apparent reason why past project data could not be analysed and filed in a database for use in risk analysis. Such a database would be
more flexible, permitting data to be retrieved in the user defined format.
RESEARCH METHODOLOGY
The research concentrated on the cost estimating activities of the 10 large general
contractors who operated in the Sydney region. These contractors average in
excess of A$100m annually. All the contractors sampled operated nationally in
both the public and private sectors of the construction industry and were
experienced in different forms of project delivery. Eight contractors were engaged
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 I 1, 2, 83-95
across the whole spectrum of construction activities while the remaining two
contractors specialized in building construction only. For this reason the research
sought views and opinions on building cost estimating practices only. Interviews
with selected general contractors took place in 1993.
Although there are close to 10 000 registered construction firms in New South
Wales, the selected sample of 10 contractors represented the top end of construction firms in terms of turnover. Although the sample may seem small, it
actually represents all but two building contractors working in the Sydney region
with a turnover over $100m annually. It was assumed that because of their size
and diversity the contractors surveyed would be near or at the forefront of
applying modern cost estimating technologies.
Responses of the selected contractors were collected through face to face
interviews with chief estimators. The initial aim was to examine estimating
practices of each contract for up to 10 specific building projects to identify variations in approach. However, after preliminary interviews this strategy was
abandoned as no significant differences in approach were detected.
87
88
Uher, T.E.
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1, 2, 83-95
Total scored
values
43
37
30
30
10
%
values
29
25
20
20
6
Note: % values column expresses the scored values for each benefit as a percentage of the total scored value for all the
benefits.
) 1995 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1, 2, 83-95
89
90
Uher, T.E.
Table 2 Ranked order of perceived limitations of bills of quantities as a tender document, response scored
on a 5,4, 3, 2, 1 scale in decreasing order of importance
Ranked order of benefits of bills of quantities
as a tender document
Presence of errors
Incompleteness
Cannot be used to control project costs
Cannot facilitate the use of probability cost estimating
Poorly structured and too 'bulky'
Total scored
values
39
36
30
25
20
%
values
26
24
20
17
13
Note: % values column expresses the scored values for each benefit as a percentage of the total scored value for all the
benefits.
Client-produced bills of quantities are rapidly disappearing as a tender document from major Australian commercial and industrial projects. Principally, clients argue that the production of a document which is almost always incomplete,
inaccurate and often misleading and which for reasons of incompleteness and
inaccuracies gives rise to claims against clients, is not worth having and paying for.
It is not the aim of this paper to argue for or against the use of a bill of quantities.
Of more interest is the examination of attitudes of general contractors towards a
bill of quantities and the manner in which general contractors compile tenders
when a client-prepared bill of quantities is unavailable.
The limited availability of client-prepared bills of quantities has not diminished
their popularity as a tender document among contractors. Seven of the contractors surveyed preferred to use bills of quantities in tendering and all agreed
that the structure of a typical bill, where various bill items are broken up into trade
packages, facilitated a more effective elicitation and formulation of subcontract
bid prices.
If bills of quantities were not provided by the client, the above seven contractors
either prepared their own bills or purchased them for quantity surveyors. The remaining three contractors in the sample preferred to develop an elemental cost plan
first, in order to establish a better control over tendering in terms of cost and time.
An interesting observation here is that these contractors were experienced in the
design and construct (turnkey) method of project delivery. This mode of delivery
requires the use of cost planning in the absence of the traditional tender documentation. The three contractors benefited from an inherent ability of the method
to facilitate a more accurate and faster comparison of alternative construction
strategies, with an added benefit of combining cost estimating with time scheduling. This approach also forms a basis of a project cost control system (Table 3).
The survey also showed that in addition to elemental cost planning, the above
contractors also prepared an electronically compiled builder's bill, a simplified
version of a traditional bill of quantities, to ensure the most effective form of
elicitation and analysis of subcontract and supply bid prices. Combining elemental cost planning with billing provides an added benefit of a vital check of the
cost estimate accuracy.
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1, 2, 83-95
Table 3 Ranked order of perceived benefits of elemental cost planning as a tender document, response
scored on a 4, 3, 2, 1 scale in decreasing order of importance
Ranked order of benefits of using elemental cost
planning as a tender document
Total scored
values
%
values
19
32
15
14
12
25
23
20
Because elemental cost models have significantly smaller numbers of cost items
than traditional cost estimating models based on a bill of quantities, the process of
modelling the behaviour of random risk variables using elemental cost models is
generally less arduous. A successful application of probability cost estimating is
dependent not only on the knowledge of the fundamental principles of risk
management but also the ability to express cost items in an elemental form. A
number of probability cost estimating models express cost items as elements, e.g.
those developed by Campbell (1971), Picardi (1973), Doyle (1977), Flanagan &
Norman (1984), Perry & Hayes (1985), Cooper & Chapman (1987), Jaafari
(1990), Raftery (1990), Uher (1990, 1991) and Yeo (1991). Similarly a trade
summary approach proposed by Brotherton (1993) would significantly reduce the
number of cost items and would enhance modelling of random risk variables.
91
92
Uher, T.E.
Table 4 Ranked order of perceived reasons for not having databases, response scored on a 6, 5, 4, 3, 2, 1
scale in decreasing order of importance
Total scored
values
33
27
25
24
19
19
%
values
23
18
17
16
13
13
rather than intuitive estimating proess and that estimators would be distracted
from examining projects in detail.
The results given in Fig. 4 show a 'weak' support for computerized databases
with only three contractors surveyed using them. However, considering that an
additional five contractors intended to develop them in the future, their use in cost
estimating will become more prominent.
CONCLUSION
The research has shown that the use of a single value cost estimating method is
popular among the ten large Sydney based general contractors and that the limited
availability of the client-prepared bills of quantities for tendering has had little
impact on their cost estimating practices. When the client-operated bills are
unavailable, these contractors either prepare their own quantities or commission a
1996 Blackwell Science Ltd, Engineering. Construction and Architectural Management 3 | 1, 2, 83-95
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1, 2, 83-95
93
94
Uher, T.E.
References
Ashley, D.B. (1989) Project risk identification using inference subjective expert assessment and historical data. Proceedings of the Internet International Expert Seminar in Connection with the PMI/Internet
Joint Symposium, Atlanta, October 12-13, 9-28.
Betts, M. (1987) Co-ordinated system of information retrieval for building contractor's tendering.
Transactions of Building Cost Research Conference, University of Salford, January, 339-355.
Betts, M. (1991) Achieving and measuring flexibility in project information retrieval. Construction
Management and Economics, 9(3), 231-245.
Brotherton, R. (1993) Risk analysis of construction cost estimates. B. Build. Thesis, School of Building,
University of New South Wales.
Campbell, D.W. (1971) Risk analysis. American Association of Cost Engineers Bulletin, 13, 4 - 5 .
Cooper, D . & Chapman, C. (1987) Risk Analysis for Large Projects. John Wiley & Sons, New York.
Cooper, D., Macdonald, D.H. & Chapman, C. (1985) Risk analysis of a construction cost estimate.
International Journal of Project Management, 3(3), 141-149.
Diekmann, M. (1983) Probabilistic estimating: mathematics and applications. Journal of Construction
Engineering and Management, ASCE, 109(3), 297-308.
Doyle, R.C. (1977) How good is your estimate. American Association of Cost Engineers Bulletin, 19(3),
93-97.
Flanagan, R. & Norman, G. (1984) Risk analysis - an extension of price prediction techniques for
building work. Construction Papers, 1(3), 27-34.
Golden, L. (1991) Learning to live with probable outcomes. The 1991 Transactions of the American
Association of Cost Engineer, 5th Annual Meeting, Seattle, Washington, pp. G.1.1-G.1.6.
Hertz, D.B. & Thomas, H. (1983) Risk Analysis and Its Applications. John Wiley and Sons Inc., New
York.
Huber, G.P. (1974) Method of quantifying probabilities and multivariate utilities. Decision Sciences, 5,
12-17.
Ivkovic, B. (1991) Project modelling and uncertainty. Proceedings of International Conference on Construction Project Modelling and Productivity, Dubrovnik, 15-32. Faculty of Civil Engineering, University of Zagreb.
Jaafari, A. (1990). Probabilistic unit cost estimation for project configuration optimisation. Transactions
of the 1990 National Engineering Project Management Conference and Forum, Sydney, 221-246.
Lewellen, W.G. & Long, M.S. (1972) Simulation versus single value estimates in capital expenditure
analysis. Decision Sciences, 3(4), 19-23.
Morgan, P.R. (1991) Probabilistic cost estimating - use and interpretation. The 1991 Transactions of the
Institution of Engineers Australia, 36-44.
Morrison, N . (1967) Critique of ranking procedure and subjective probability distributions. Management Science, 14, B253-254.
Moselhi, O., Hegazy, T. & Fazio, P. (1991) Practical cost estimation in a competitive environment. The
1991 Transactions of the American Association of Cost Engineers, 5th Annual Meeting, Seattle,
Washington, pp. H.5.1-H.5.9.
Newton, S. (1992) Methods of analysing risk exposure in the cost estimates of high quality offices.
Construction Management and Economics, 10(5), 431-449.
Perry, J.G. & Hayes, R.W. (1985) Risk and its management in construction projects. Proceedings
Institution of Civil Engineers, Part I, 78, 499-521.
Picardi, A. (1973) Closing the credibility gap in construction cost estimating. APEC Journal, February,
p.4.
Pratt, J.W. (1964) Risk aversion in the small and in the large. Econometrica, 32, 122-136.
Raftery, J. (1990) Risk analysis for construction forecasting. Proceedings of Risk Analysis for Construction
Forecasting Workshop, Department of Surveying, Thames Polytechnic.
Raiffa, H. (1968) Decision Analysis. Addison-Wesley, Reading, Massachusetts.
Rowland, W.P. & Curran, K.M. (1991) Range estimating in value engineering. The 1991 Transactions
of the American Association of Cost Engineers, 35th Annual Meeting, Seattle, Washington, pp. G.3.1G.3.5.
Samid, G. (1994) Contingency revisited. Cost Engineer, 36(12), 23-26.
Shafer, S.L. (1991) Estimate and project risk analysis approaches. The 1991 Transactions ofthe American
Association of Cost Engineers, 35th Annual Meeting, Seattle, Washington, pp. K.5.1-K.5.5.
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1, 2, 83-95
Toakley, A.R. (1989) Risk, uncertainty and subjectivity in the building procurement process - a critical review.
School of Building, University of New South Wales.
Townley, T.W. (1991) Historical data-driven range estimating. The 1991 Transactions of the American
Association of Cost Engineers, 35th Annual Meeting, Seattle, Washington, pp. G.2.1-G.2.5.
Uher, T.E. (1990) The effect of risks and uncertainties on subcontract bid prices. PhD Thesis, University of New South Wales.
Uher, T.E. (1991) Practical approach to risk management in construction. Proceedings ofthe Continuing
Education Seminar. School of Building, University of New South Wales.
Uher, T.E. & Levido, G.E. (1992) Construction Planning and Control, 2nd edn. School of Building,
University of New South Wales.
Winkler, R.L. (1967) The assessment of prior distributions in Baysian analysis. Journal of American
Statistical Society, 62, 776-800.
Yeo, K.T. (1991) Project cost sensitivity and variability analysis. International Journal of Project
Management, 9(2), 111-116.
1996 Blackwell Science Ltd, Engineering, Construction and Architectural Management 3 | 1,2,83-95
95
1. Peter R. Davis, Peter E.D. Love, David Baccarini. 2009. Bills of Quantities: nemesis or nirvana?.
Structural Survey 27:2, 99-108. [Abstract] [Full Text] [PDF]
2. Aminah Fayek, David M. Young, Colin F. Duffield. 1998. A Survey of Tendering Practices in the
Australian Construction Industry. Engineering Management Journal 10, 29-34. [CrossRef]