Color profile: Disabled Composite Default screen 3 Cost and time modelling Summary This chapter seeks to establish that a need exists for both the simplification and the rationalisation of the bill of quantities (BoQ) cost model. The present day uses of the BoQ are compared to those for which it was originally intended. This comparison demonstrates that, in practice, the BoQ and its current uses are at variance with, or have exceeded, original intentions. Having established the need for new improved cost models in Chapter 2, this chapter looks at some of the methods used by academic researchers and practitioners to, firstly, overcome the difficulties faced by the BoQ and, secondly, to integrate cost with time. The methods are compared and critically evaluated. The evidence supports the proposition that simplification using “significance theory” alone simplifies the BoQ, but does not improve or enhance its role in the systematic quantification of delay and disruption claims. In short, simplification using significance theory does not make the BoQ capable of fulfilling the various tasks required of today’s cost model. Because most problems associated with the conventional BoQ exist because it fails to model work items and cost realistically, it is proposed that improvement and increasing the realism within the BoQ can best be achieved by simplifying the method of measurement or by a change from work items towards “work packages” or “operational groups”. 3.1. Introduction The BoQ has been in use for a number of centuries. Originating in the UK, evolved significantly in the nineteenth century, mainly as a result of the growth in transport such as canals and railways. The evolution of the BoQ was accompanied by amendments to conditions of contract, which were in turn revised in line with changes in case law and statute. Case law at this time centred on liability for Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:37 PM Color profile: Disabled Composite Default screen 28 Quantifying and managing disruption claims inaccurate BoQs, measurement rules, errors in composite descriptions, and alleged omissions of items, e.g. Moon v. Witney Union.1 “Measurement problems” probably led to standardisation and development of the first standard methods of measurement. According to a Royal Institute of Chartered Surveyors (RICS) report in 1993, the BoQ is used in approximately 50 % of all construction contracts by value. This figure is nearly a decade old, and the use of traditional contracting has been reduced by the increased use of the design and build, BOOT and BOO procurement routes, PFI/ PPP arrangements, and partnering and alliancing relationships. However, the 50 % figure is probably low, since it does not appear to take into account the wide use of BoQs in the above procurement options such as design and build. The BoQ is still the cost model used by the parties in estimating cost in the iterative process of refining the design and specification and for appointing subcontractors. This reliance, or continued use of the BoQ even in these untraditional procurement routes, lies in the fact that estimators, quantity surveyors and cost accountants all appear to use the standard methods of measurement. Reliance upon detailed standard methods of measurement means that estimators, quantity surveyors and the cost accountants, naturally, develop or compile detailed and lengthy BoQs. Pricing or estimating such BoQs is then carried out using “norms” or data from either long-established commercial databases or using the contractor’s/estimator’s own, equally longestablished “book of rates”. The practical problem is that although the popularity of BoQs may not have changed, the uses to which they are now put have, in contrast, changed dramatically. 3.2. The original purpose of the BoQ The BoQ was originally designed to: • • • • Provide an equal basis for tendering. Convey the client’s requirements. Save contractor’s time and cost in preparation. Act as a payment document during the construction phase. Professor John Uff QC, in his recent paper Are we all in the wrong job? Reflections on construction dispute resolution,2 when discussing the rise of the quantity surveyor, says: 1 Moon v. Witney Union (1837) 3 Bing NC 814. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:38 PM Color profile: Disabled Composite Default screen Cost and time modelling 29 Bills of quantities then took on more significant roles both as a vehicle for calculation of interim payments and doubling as the specification, under the Joint Contracts Tribunal (JCT) with quantities forms of building contracts. These are still lump sum contracts, but for the Institution of Civil Engineers (ICE) and other types of contract, the bill of quantities later acquired full contract status, the work becoming subject to “remeasurement”. In the past few decades both natural evolution in forms of procurement and conscious policy have led to a move against bills of quantities, towards milestone interim payments and lump sum contracts of an increasingly turnkey nature. No one today wants to incur the cost of interim measurements, nor to run the risk of interim payment disputes. Even though the use of interim measurements may be decreasing, the BoQ still remains a very useful document, so much so that the BoQ and its present day diverse practical uses are clearly at variance with its original theoretical objectives, see Table 3.1. Table 3.1. Present day uses of BoQs Client Is an aid to defining employer’s requirements/material specification. Allows comparison of tenders. Allows costing of alternative designs. (In traditional contracting) construction supervision, (remeasurement). Contractor Estimating and tendering. Obtaining quotations for materials. Planning. Cost control. Cashflow prediction. Construction site management. Contractual interface Interim payment valuation. Valuation of variations. Work remeasurement. As early as 1980,3 surveys had concluded that, in practice, most use was made of the BoQ during the construction and final account 2 3 Are we all in the wrong job? Reflections on construction dispute resolution. Uff J. A Society of Construction Law Paper based on the talk given to a joint meeting of the Society of Construction Law and the Society of Construction Arbitrators in London, 3 July 2001. The Contractor’s Use of Bills. Skinner DWH and Jepson B. Chartered Quantity Surveyor, pp. 63–68, September 1980. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:38 PM Color profile: Disabled Composite Default screen 30 Quantifying and managing disruption claims phases, and not at the tendering stage as was initially intended. This is not surprising, since once the BoQ is used in the tendering stage, it adopts its contractual purpose during the construction phase. The contractual purpose of the BoQ appears to be to produce a cost model for payment purposes during the construction contract. However, because the BoQ is the only contract document concerned with detailed financial issues, in practice it is often used in all matters involving money and naturally becomes the central vehicle for the financial control of civil engineering and building works. In short, experience and academic surveys appear to indicate that the BoQ is now used as a cost estimating model, tendering model, planning model and a contractor’s site cost control model, lies at the centre of most payments, and is used in the valuation of variations and claims (including those for delay and disruption). The important question involves an analysis of whether the BoQ adequately fulfils the varied requirements of the industry. In other words, let us investigate if the BoQ warrants improvement, total replacement or commendation. As a starting point, the positive and negative aspects of the BoQ are evaluated. 3.3. Positive aspects of the BoQ • Allows contractors to tender on a common basis. • Because the BoQ is produced only once, costs to the employer may be ultimately minimised. • It is a convenient method to price provisional sums and other parts of the work that are not designed before tenders are sought. • The BoQ unit rates generally provide a “pre-agreed” basis for the preparation of interim valuations, pricing (certain, simple) variations and for preparation of the final account. 3.4. Negative aspects of the BoQ 3.4.1. Estimating 3.4.1.1. Unit rate or operational estimating? BoQs usually describe the type and net quantity of material to be used in the construction process. Some estimators believe that the material price is relatively simple and will have little effect on the overall price for the site work operation. For example, “a brick is a brick and will cost the same wherever it may be incorporated into the structure”. Equally true is the fact that probably all the contrac- Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:38 PM Color profile: Disabled Composite Default screen Cost and time modelling 31 tors will be able to purchase the bricks at similar prices. However, a contractor could seek advantage over its competitors by minimising material wastage. Because the material cost can, in some cases, represent up to 50 % of the total project cost profile, the estimator needs to be careful when predicting the wastage allowance. Experience shows that typically, most contractors simply use handbooks to calculate/predict wastage. Interestingly, a report on material wastage found that although the mean actual wastage on construction sites in 1993 was 12.3 %, the mean wastage allowance used by estimators was only 6.6 %.4 In contrast, the estimation of labour and plant costs is more complex, subjective and will depend upon the location and, in practice, each contractor’s chosen method of site construction. Harris and McCaffer,5 Moyles,6 Merriefield7 and Harrison8 all suggested that labour and plant usage/costs are best estimated using an “operational” method. Operational estimating involves calculating the total direct cost of resources based on the total time the resources are required, whereas unit rate estimating involves calculating costs based on the outputs of each individual resource. The considered view of the above writers was that, certainly for civil engineering works, operational estimating links very well with site planning and is effective in allowing for idle time, which is common in plant-intensive civil engineering work. For this reason, such writers concluded that the operational estimating method was more reliable and, in practice, more realistic than the unit rate methods. Indeed, over 30 years ago in 1971, operational estimating was beginning to be reported as gaining in popularity: Barnes9 wrote that operational estimating’s “superiority for describing the cost of interrelated mechanised operations stems from the more realistic representation of the variables affecting their cost”. 4 5 6 7 8 9 Organizational Productivity – A Case Study of Materials Management in UK Construction Firms. Proverbs D, Olomolaiye P and Harris FC. Construction Papers, No. 44, edited by Peter Harlow, the Chartered Institute of Building, Ascot, 1995. Modern Construction Management. Third Edition. Harris F and McCaffer R. Blackwell Science, 1989. An Analysis of the Contractor’s Estimating Process. Moyles BF. M.Sc. Thesis, Loughborough University of Technology, 1973. Construction Estimating. Merriefield CM. Management of Contracts and Projects, Project Management Group, Department of Civil and Structural Engineering, UMIST, 13–17 May 1991. Operational Estimating. Harrison RS. Construction Papers, No. 33, edited by Peter Harlow, the Chartered Institute of Building, Ascot, 1994. Civil Engineering Bills of Quantities. Barnes NML. CIRIA Research Report No. 98. Department of Civil and Structural Engineering, UMIST, 1971. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:38 PM Color profile: Disabled Composite Default screen 32 Quantifying and managing disruption claims Today, there appears to be a distinct shortage of empirical data/ records on the use of operational estimating. However, it is clear, or rather assumed, that because increases in mechanisation promote operational estimating, it may be reasonable to suggest that since the former has increased, then operational estimating would also have found greater use, certainly with civil engineering estimators. However, the problem for such estimators is that they need to subjectively manipulate the operational estimate to become commensurate with the detail in the typical BoQ. In practice, the estimator is required to group together various BoQ items to form what he or she believes represents an actual site work operation, e.g. laying a section of sewer. This group of BoQ items may be represented in the construction programme as one or more activities. Typically, the estimator assumes a total duration and then selects the labour/resources to be used for that site activity. This total site activity duration is then used to calculate the time-related costs of that activity. Paradoxically, because the cost model, the BoQ, is so detailed, the operationally estimated site activity cost is then “assigned” back to each detailed BoQ item. This means that whilst a group/section of BoQ unit rate items may be estimated using a realistic approach, the BoQ itemisation encourages/requires a divergence between actual site costs and BoQ unit rates. Harris and McCaffer10 suggested that the “method of assignment” is open to each estimator and that the simplest method is to apportion the cost pro rata with the item quantity. But, in practice, experience shows that contractors and estimators usually produce another cost model, based on operations, solely for their estimating and planning purposes. They typically call this manipulation of the BoQ “reworking and adaptation”. In practice, this means that from a construction site’s point of view, the employer’s BoQ is totally unsuitable for the contractor’s site planning and construction of the works. Barnes’s11 empirical research reinforces the above. He stated, “The conventional BoQ appears to be imposing an unrealistic model of the factors influencing construction costs upon financial control due to its adherence to the quantity-proportional hypothesis”. He went on to report that the estimator man-hours involved in manipulation and “assignment” procedures could be significant, and stated, “This conversion process would have been unnecessary if the bill items had themselves been operationally grouped”. Further, it is better to provide operational estimates to site management so that each site operation is controlled and resourced 10 11 See Footnote 5. See Footnote 9. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:39 PM Color profile: Disabled Composite Default screen Cost and time modelling 33 effectively and efficiently and is at least commensurate with the estimator’s assumptions. In practice, site contract managers do not want information on BoQ work items for their purposes – they want to control, pay for and resource actual site operations. In short, their interest is very different from that of the estimator, and BoQ items discourage any useful sharing of information. In practice, experience tells us that whilst many contractors may (in private) admit that the current BoQ system is not suitable for accurate/realistic estimating and planning, they quite willingly manipulate the BoQ unit items in order to develop their own cost models based on proposed site operations. So although, in practice, the structure and associated complexity of the BoQ may not be a “perceived” problem, when situations arise where delays and variations need to be evaluated, conflict arises simply because the BoQ does not reflect the true manner in which the fixed, time-related and other resource costs are actually planned to be incurred on site. 3.4.1.2. Level of detail in the BoQ Too many low-value items It is widely accepted that BoQs contain many low-value items. The academic starting point for this proposition was Lichtenberg,12 who highlighted the “problem of overwhelming detail” in the BoQ. The BoQ is so detailed because it acts as a specification, a long list of all the materials to be incorporated and, most importantly, because it is developed by applying the rules from, typically, a standard method of measurement. As a consequence, employers seek to provide as much detail as possible, confusing hundreds of pages in a BoQ with being precise. The standard methods of measurement simply categorise or specify at such detail that reliance on them, in practice, generates very lengthy BoQs. The level of detail chosen for itemisation in the method of measurement has a direct effect on the level and on the number of BoQ items. For example, the methods of measurement “itemise” each time there is a new material description, e.g. a change from 16 mm diameter reinforcement to 20 mm diameter reinforcement. Barnes stated, “If the method of measurement were to be revised, the opportunity should be taken of structuring the rules for itemisation and classifying them more precisely, 12 The Successive Principle, Procedures for a Minimum Degree of Detailing. Lichtenberg S. Sixth Annual Proceedings of the Project Management Institute, Washington, pp. 570–578, 1974. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:39 PM Color profile: Disabled Composite Default screen 34 Quantifying and managing disruption claims so it can be simplified”. However, in practice, some employers justify the large amount of detail in the BoQ, claiming that the BoQ helps to “build up a picture of the complex nature of the project”. Others in the industry suggest that, in practice, drawings and the actual specification provide a complete description of the works; hence, the BoQ needs to concentrate only on the features that significantly affect cost. Not surprisingly, the Royal Institute of Chartered Surveyors in its Principle of Measurement (International) for Works of Construction (POMI) adopted this less detailed approach. In the POMI, the items are more composite in nature than in a traditional BoQ and much of the work, which would normally be described as separate items, is deemed to be included in the major items. Similarly, in practice (and in private), some estimators admit, especially in relation to tendering to DB, BOOT and PFI/PPP projects, that they amend the common methods of measurement, either by direct omission or by amalgamation, with the result that their BoQs contain at least 50 % less items. In the 1970s, it was perceived that 2.5 % of the total project cost of civil engineering projects was consumed by the compiling and administration of BoQs (excluding the negotiation of claims). Not surprisingly, subsequent academics and certain estimators in contracting entities sought to reduce the costs of compiling and pricing BoQs by reducing their sheer complexity by concentrating on the most important BoQ items or the “significant items” in the BoQ. Estimating accuracy There is a correlation between the estimating accuracy and the value of each BoQ item. More specifically, high-value BoQ items are, in practice, estimated to a higher degree of accuracy than low-value items, and as the relative value of the BoQ item falls so does the accuAccuracy level + Design process – Fig. 3.1. The relationship between estimating accuracy and the design process Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:40 PM Color profile: Disabled Composite Default screen Cost and time modelling 35 racy with which it is estimated. This may be because, perhaps, lowvalue items simply receive low priority or because low-value items generally relate to abstract work. Accuracy level Cost of estimating as percentage of cost of job Fig. 3.2. Estimating accuracy against the cost of estimating It is generally well known that the accuracy of a cost estimate will improve as the level of design detail increases (see Fig. 3.1). Accuracies quoted for the feasibility estimate for civil work range from ±20 % to ±40 %.13 Interestingly, and perhaps not so well known is that if an estimator has priced a certain number of items, there is little increase in accuracy to be gained by pricing any further items (see Fig. 3.2).14 As Lichtenberg commented, “It is of major importance not to waste effort in detailing to a greater degree than can be justified by the accuracy of the total”.15 3.4.2. Valuation of change, delay and disruption Barnes16 has commented, “One of the principal shortcomings of the traditional BoQ was that it was only a BoQ of the permanent work 13 14 15 16 Design Phase Estimating: State of the Art. Ogunlana O and Thorpe T. International Journal of Construction Management and Technology, Volume 2, No. 4, pp. 34–47, 1987. Probabilistic Approach to Estimating and Cost Control. Vergera AJ and Boyer LT. ASCE Journal of Construction Division, Volume 100, pp. 543–552, 1974. The Successive Principle, Procedures for a Minimum Degree of Detailing. Lichtenberg, S. Sixth Annual Proceedings of the Project Management Institute, Washington, pp. 570–578, 1974. Civil Engineering Standard Method of Measurement. Third Edition. Barnes NML. The Institution of Civil Engineers. Handbook, Thomas Telford, London 1992. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:40 PM Color profile: Disabled Composite Default screen 36 Quantifying and managing disruption claims left behind when the contractor’s men and machines had moved on”. The BoQ was originally introduced in a labour-intensive building industry and did not take account of plant and temporary works. As the industry developed, the use of plant increased but the BoQ rates still compounded all the resource costs and relied on the principle that quantity is proportional to cost. Barnes suggested that this situation could be tolerated, so long as the “work was never varied after tenders were accepted”. However, most practitioners would agree that variations and delay and disruption claims are simply inevitable. It is in the evaluation of changes, delay and disruption that unit rates (compounding all costs) primarily fall down; those BoQ items that are not quantity-proportional are often evaluated on a totally unsystematic, inaccurate or unrealistic basis. In 1971, Barnes found that 70 % of all claims required consideration of the distinction between costs. Five years later in 1976, Barnes introduced the concept of method-related charges (MRC) to the Civil Engineering Standard Method of Measurement (CESMM). It was now possible for contractors to price certain work items that they considered not to be simply quantity-proportional. Three charges were available to estimators: (1) the quantity-proportional charge, (2) the time-related method-related charge and (3) the fixed method-related charge. It was intended that the three types of charge would be more closely related to the manner in which costs were intended to be actually incurred on site. Accordingly, it was hoped that the MRCs would allow the automatic evaluation of variations and increase the realism of quantifying claims for delay and disruption. It was decided that the MRCs were a function of each contractor’s own method of working and therefore should not be set against predetermined items and the MRCs were not mandatory. As a consequence, in practice, MRCs were not commonly used in the UK, and in 1992 Barnes conceded that the “three price bill does not model value perfectly because it does not model all the many variables which can affect cost”. The problems that MRCs were initially designed to overcome still remain today. 3.4.3. BoQ unit rate abuse Because the unit rate in the BoQ does not reveal the way in which a contractor’s price builds up, in practice the pricing of BoQ items is open to manipulation. “Tricks of the trade” such as “front-end loading”, designed to improve the contractor’s cash flow, demonstrate the “smokescreen” between the realistic costs of construction and BoQ item rates. This smokescreen surrounding the contractor’s resource assumptions also allows “claims spotting” to go undetected. This Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:40 PM Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:41 PM Contractual interface:. Contractor: Client: Partial Work remeasurement. Fail Site management. Fail Fail Cashflow prediction. Valuation of variations and claims. Partial Cost control. Partial Partial Planning. Interim payment valuations. Good Partial Comparison of tenders. Obtaining quotations. Partial Construction supervision. Partial Fail Costing of alternative designs. Estimating and tendering. Partial Cost and time modelling Time-related and other charge are not acknowledged in payment. BoQ rates promote “cost is proportional to quantity” idea. No link to time model, unit rate compounds all types of cost. Measurement is based on material components installed. Control model is different from payment model. Too many work items in work operation hence no effective control. BoQ not directly linked to time model. Even the “S” curve approach is approximate. BoQ items and rates are converted into site operations. Control is “crude” and feedback non-existent. BoQ based on materials to be incorporated, for planning function of site activities BoQ requires subjective conversion. Sub-contractors like the materials detail. Main contractors simply photocopy BoQ and send out for enquiry. Unit rate restricts use of operational estimating. Too many low-value items. Unit rate manipulation makes this difficult. Need to increase tender realism. The BoQ requires “conversion” into site activities. No integration of estimating and planning is promoted. Unit rate does not link the design and costing exercises. Large variances between work items rates and tender, therefore use by clients to estimate is questionable. Drawings and specification may be more suitable. Success Remarks Aid to define client’s requirements. Use Table 3.2. Summary of effective uses of BoQ Color profile: Disabled Composite Default screen 37 Color profile: Disabled Composite Default screen 38 Quantifying and managing disruption claims may be described as a process by which contractors decrease their rates on items they believe to have been overmeasured, and increase the rates for items they consider to be undermeasured. In practical terms, employers very often fail to detect such devices because the actual variances between each contractor are simply much greater than can be attributed to the effect of loading. 3.4.4. Summary of BoQ uses Table 3.2 provides a summary of the uses of the BoQ in each of the roles for which it is now typically employed. The proposition in this book is that the BoQ now fails to fulfil its various desired uses and that improved cost models should aim to satisfy the original use for which the BoQ was developed (tendering on an equal basis) but must also overcome the deficiencies in the BoQ caused by its varied and extended use. Table 3.3 provides a summary of the requirements of an “ideal cost model”. The next section recounts previous attempts made to simplify and rationalise the way costs are modelled in the construction process in order to develop a direct link with time as typically modelled in the programme. Table 3.3. Ideal cost model requirements Cost model requirement Client Allow tendering on a common basis. Allow contractor flexibility in method and pricing. Provide transparency in pricing in order to aid tender evaluation/comparison. Contractor Improve and simplify estimating; encourage realistic planning. Facilitate integration between the cost model and time model. Allow and promote effective site control. Contractual interface Allow realistic measurement and valuation. Facilitate the quick and automatic evaluation of variations and claims involving delay and disruption. 3.5. Previous attempts to rationalise the BoQ 3.5.1. Significant items approach In order to reduce the number of items in the BoQ and also increase estimating/planning attention on the higher-value items (thus Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:41 PM Color profile: Disabled Composite Default screen Cost and time modelling 39 increasing the overall estimating accuracy), various academics and practitioners have used two practices: the “successive principle” and the “80/20 principle”. The pattern underlying the 80/20 principle was discovered in 1897 by Italian economist Vilfredo Pareto (1848– 1923). Pareto discovered that 80 % of the wealth of a nation is in the hands of only 20 % of the population. The 80/20 rule has been found to apply to many situations, and it has long been known that 80 % of the value of a BoQ is contained in approximately 20 % of the items. 3.5.1.1. Successive principle Empirical research by Lichtenberg17 concluded that maximum estimating accuracy was obtained by concentrating estimating effort on the major items with greatest variance. He used this principle to develop the method of “successive estimating”. In this procedure, a project is divided into a small number of main work items or work operations. Each work item/operation is quantified and three estimates (maximum, minimum, most likely) made of the costs. Similar to the PERT method, the mean and the variance can be “approximated” for each item. Items with a large variance are subdivided and the variance recalculated. This procedure is repeated until the estimator is satisfied that the “variance derived from items cannot be estimated with greater accuracy than the variance achieved”. The underlying philosophy in Lichtenberg’s work is that there is little sense wasting time and effort in estimating items that do not improve overall accuracy. However, the process of selecting the “major items” is highly subjective and, in practice, will vary between contractors and estimators. Further, Lichtenberg made no attempt to model or compensate for low-value sections of the work. Lichtenberg’s study did not relate directly to BoQ unit items, instead his “work items” reflected site tasks. Lichtenberg’s analysis was aimed at reducing the BoQ complexity, purely to optimise the accuracy, cost and effort associated with estimating. The need for site control, cost and time integration, and systematic evaluation of the effects of delay and disruption claims were not investigated. 17 See Footnote 12. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:41 PM Color profile: Disabled Composite Default screen 40 Quantifying and managing disruption claims 3.5.1.2. Significant items estimating Short (1970)18 At the Institute of Municipal Engineers’ conference, Short discussed the concept of cost-significant items. He highlighted earlier research conducted at the University of Manchester that had established that 90 % of total project costs were consumed by 10 % of the BoQ items. Short concluded that it might be beneficial to tender only the 10 % of significant items and negotiate a lump sum cost for the remaining 10 % of the project cost. He believed that this method would remove the cost associated with estimating the abstract and low-value items such as “fillets in manholes vertically downward”. Barnes (1971) 19 Barnes used “Pareto’s law” and concluded that 20 % of the total cost in a BoQ was occupied by 75 % to 95 % of the BoQ items. Nevertheless, Barnes suggested that the low-value items were necessary to convey the detail of the proposed work. However, because his empirical research had concluded that the low-value items had little influence on the final account, Barnes concluded that they could be eliminated from the financial control of the contract. Barnes noted the inadequacy of the standard form methods of measurement used to compile BoQs, stating, “It is one of the paradoxes of the contractual system that the BoQ contains prices for fixing components of the work which are so small that their fixing cost cannot economically be defined”. As discussed above, Barnes espoused the many advantages of operational estimating and the need to adapt standard form methods of measurement to encourage operational estimating. Moyles (1973) 20 Moyles’s investigation into items in the BoQ “proved the existence of a Pareto distribution”. Accordingly, he suggested that this 80/20 relationship be used to simplify BoQs. The method used by Moyles to select the high-value items in the BoQ or the cost-significant items 18 19 20 Summing Up. Short GS. Second Conference on Contracting in Civil Engineering since Banwell. EDC for Civil Engineers and Municipal Engineers, National Economic Development Office, 1970. See Footnote 9. An Analysis of the Contractor’s Estimating Process. Moyles BF. M.Sc. Thesis, Loughborough University of Technology, 1973. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:41 PM Color profile: Disabled Composite Default screen Cost and time modelling 41 was still laborious and involved pricing all the BoQ items and ranking them in descending order. When 80 % of the total value was reached, the corresponding items were selected. Harmer (1983) 21 Harmer, with the Property Services Agency (PSA), also decided to make use of Pareto’s law. The PSA developed an estimating system using significance theory, basing it on the rationale that “by restricting the measurement and pricing to significant items and using data with reasonable sample sizes it was considered that the unreliability of rates from the BoQ would be minimised”. Harmer reported that, on average, 20 % of the BoQ items in a work classification (e.g. excavation, brickwork, concrete) represented 69–89 % of the work section total value. Saket (1986) 22 Saket used an earlier finding that the cost-significant items were those whose value was greater than the average value, and developed what he termed “iterative estimating”. This method of estimating did not require all the BoQ items to be priced; instead, attention was focused on the cost-significant items. Allman (1988) 23 Allman, like Harmer, researched the simplification of the BoQ within the PSA. Allman’s review of the significant items estimating system highlighted that results analysed by the PSA and checked independently by the Building Research Station (BRS) “indicated significant items estimating as being the most accurate method in use within the PSA”. The PSA published a “significant items booklet” that set out the estimating procedure and contained details relating to the percentage cost of insignificant items in each trade. The PSA also produced a computer programme that incorporated Monte Carlo simulation to increase estimating accuracy. 21 22 23 Identifying Significant BoQ Items. Harmer S. Chartered Quantity Surveyor, pp. 95–96, October 1983. Cost Significance Applied to Estimating and Control of Construction Projects. Saket MM. Ph.D. Thesis, The University of Dundee, 1986. Significant Items Estimating. Allman I. Chartered Quantity Surveyor, pp. 24–25, September 1988. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:42 PM Color profile: Disabled Composite Default screen 42 Quantifying and managing disruption claims Betts and Gunner (1993) 24 Betts and Gunner reported undertaking similar research in Singapore, which “revealed that the 80/20 relationship remains constant over a wide variety of building functions”. They highlighted the growing popularity of cost-significant item estimating worldwide and reported that in Singapore, the Construction Industry Development had begun issuing tender price indices based upon significant items. McGowan (1994) 25 McGowan further refined the cost significance technique in order to develop models based on the principle of “resource significance”. A resource-significant item was identified as any item in a BoQ that had a labour, plant or material resource value (cost and hours) greater than the mean labour, plant or material resource value. McGowan developed a model for reinforced concrete bridges and a model for roads (flexible construction), which accounted for a consistent and high (typically 80–90 %) proportion of the resources and costs. The primary failure of such models was that they were project-specific and extremely sensitive to change. 3.5.1.3. Positive aspects of using significant items approach Low-value items, which are often associated with large variances in accuracy but which still require time and effort to price, are no longer required to be estimated. The significant-items approach appears to provide a cheaper method of estimating whilst preserving sufficient accuracy. 3.5.1.4. Negative aspects of using significant items approach 1. Still leads to producing a BoQ – and still encourages unit rate estimating. 2. Still relies on detailed standard methods of measurement. 3. Still leads to BoQ unit rates based on materials. 4. BoQ items are not related to site operations or work packages. 5. Cost and time are not directly linked, and realistic planning and site construction process control are not encouraged. 24 25 Financial Management of Construction Projects: Cases and Theory in the Pacific Rim. Betts M and Gunner J. Longman Inc., White Plains, New York, USA, 1993. Integrated Cost and Time Models for Measuring, Valuing and Controlling Construction Projects. McGowan PH. Ph.D. Thesis, The University of Dundee, 1994. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:42 PM Color profile: Disabled Composite Default screen Cost and time modelling 43 6. Low-value BoQ items are still required to be purchased and hence need to be unambiguously itemised in the contract documents. The principle of significance does, of course, lead to simpler, less detailed cost models. Such models are still as accurate in predicting total cost as is a BoQ containing 100 % of the items. However, costsignificant items (CSIs) retain the same BoQ item definition, based upon detailed methods of measurement. CSIs are not work packages, i.e. they certainly do not represent work operations or activities on a programme; neither did any of the proponents of significant items estimating purport that CSIs were capable of being such work packages or operations. In short, CSIs are all about reducing the number of items that an estimator needs to price. The number of items in a BoQ is clearly dependent upon: 1. The degree of detail required in the particular method of measurement. 2. Format or structure of the BoQ, e.g. elemental, sectionalised bills usually have greater duplication of items. 3. The complexity and repeatability of the design. 4. The size of the project (obviously!) The proposition in this book is simply that because all BoQs are required to be compiled using a method of measurement, simplification of the cost model can only come about by reducing the level of detail in the methods of measurement. Further, in order to allow site construction process control, the cost model must contain work packages that are directly linked to site activities. Further still, to define these work packages, additional criteria are needed: significance theory alone appears to be insufficient. In short, it appears that in order to simply reduce the complexity of BoQs for estimating purposes, the significant items approach is sufficient, but that the integration of time with cost, thus facilitating realistic planning and site control, can, in practice, only be realistically achieved by reducing the detail of measurement in the method of measurement. In short, simplification and rationalisation of BoQs must begin with changing the way standard form methods of measurement generate the vast number of BoQ items. 3.5.2. Work packaging or “operational grouping” This method involves grouping similar BoQ items and considering them as one. Like items may be grouped into work operations such as fixing steel reinforcement or fixing formwork. Prior to a comparison of the relative advantages of the two “simplification” methods, Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:42 PM Color profile: Disabled Composite Default screen 44 Quantifying and managing disruption claims let us take a look at the historical development of the school of thought behind the “operational” or work packaging methodology. 26 Skoyles (1968) – operational bill Skoyles stressed that the BoQ system “is based on labour cost being related to the physical measurement of materials” and suggested that, in contrast, the estimation of labour/plant costs should be related to site operations. He argued that the BoQ was simply a description of the completed project rather than the work to be done, and therefore did not allow any form of construction process control. Skoyles believed that cost models should instead reflect the manner in which costs are actually incurred on site, and hence relate to site activities or operations. He suggested these site operations or work packages would naturally encourage operational estimating, increase overall estimating realism and also facilitate the integration of cost with time. Cost models alone cannot allow systematic evaluation of change, and hence Skoyles encouraged project programmes to be produced based on operations (pre-defined in the cost model). Skoyles also concluded that changes to the standard method of measurement were necessary to drive the changes in the BoQ. However, he was cautious, and cited the inertia of the industry to change in his decision to “retain familiar methods of measurement and rely on pricing unit rates”. Because the items in the cost model were still differentiated by material function, abstract and low-value items still remained in the operational bill. Indeed, the operational bill format created more items, because it separated items further by individual elements and sections of the project. The bill format failed to encourage operational estimating and hence the original estimating problems relating to complexity, time and cost of estimating were still prominent. In fact, it was found that some quantity surveyors reported 80 % increases in the time required to produce such bills and that the pre-defined precedence diagram was greeted with contempt by the industry. Contractors simply thought it was “too restrictive”.27 The Building Research Station (BRS) developed the operational bill (OB), and the government agency encouraged its use on some projects to demonstrate its intended benefits. In the early termination of a contract, Skoyles and Lear28 demonstrated the benefits of using 26 27 Introducing Bills of Quantities (Operational Format). Skoyles ER. Current Paper 62, Building Research Station, 1968. Operational Bills: Users Comments. Abbot B, et al. The Architect’s Journal, pp. 617–620, March 1970. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:43 PM Color profile: Disabled Composite Default screen Cost and time modelling 45 work operations in the measurement and valuation. They suggested that operations allowed easier measurement since it was easy to relate the cost (or payment) model to site operations. They maintained that if a BoQ (as previously prepared) had been used, the process of termination would have “been much more complex, involving a good deal of remeasuring and greater liability to dispute”. 29 Barnes (1971) – operational grouping Barnes sought to develop a cost model that was site-realistic and encouraged operational estimating. His approach was to group similar BoQ items to form site activities. Barnes reported that a 40 % reduction in the number of BoQ items to be priced could be achieved by “changing descriptions and amalgamating items”. Barnes’s operational grouping did not group work performed by one gang working with one productivity, but was simply related to site activities. For example, Barnes did not group concrete placing, steel reinforcement fixing and formwork erection separately, but together to form the activity of “constructing a concrete foundation”. 3.5.2.1. Positive aspects of using work packages 1. Work packages are directly related to site operations and encourage, even allow, operational estimating. If operational estimating is more realistic, better estimating accuracy may be expected. Operational estimating reflects realistic resource usage commensurate with developing cost-sensitive charges. 2. The work package definition is identical in the cost and time models, so integrated planning and control is feasible. 3. Work packages relating to site activities may allow effective site control, which in turn generates feedback, which can contribute to more accurate or realistic estimates. 4. Work packages and a work breakdown structure (WBS) are essential for cost and time integration, thereby generally allowing objective, systematic evaluation of a greater number of variations and delay and disruption claims. 5. Practical experience indicates that grouping of work items into work packages is currently common practice within larger 28 29 Practical Application of Operational Bills Two. Skoyles ER and Lear RF. Current Paper Series, Building Research Station, 1968. See Footnote 9. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:43 PM Color profile: Disabled Composite Default screen 46 Quantifying and managing disruption claims contracting organisations, and hence, in practice, should face less inertia to change. 3.6. Time in construction contracts Pickavance30 defines programming as merely deciding when and where work will be performed and how it is to be sequenced in relation to other activities. By necessity, programmes involve decisions concerning the duration of the work, who will perform the work, the labour/plant resources to be applied, how the work will be monitored, and how the programme will be updated. A number of authors are dismissive of the role of the programme in evaluating the effects of changes and delays and disruption (Fenwick-Elliot,31 Duncan-Wallace).32 The reasons given for this are: 1. Because the programme is not stated in the forms of contract to be a contract document, it could be no more than a statement of best endeavours. 2. Because, historically, construction programmes have tended to be prepared as simple bar charts without apparent construction logic, the effect of one activity upon another cannot be predicted. Ideally, programmes should be realistic and reflect the contractor’s optimum working method, and should be the most economically beneficial for the contractor. Pickavance has suggested that there is a need for contractors to have access to adequate databases in order to plan realistically. Currently, “bills of quantities are not generally produced in a form which will assist them”. Similarly, in practice, contractors have stated that a “realistic estimation of the productivity of gangs” is a main factor in a successful programme. The importance of the role of a realistic programme in the assessment of extension-of-time claims caused by delays was highlighted in Barker Construction Ltd v. London Portman Hotel Ltd [1996] 12 Const LJ 277. This case held that in order to calculate a fair and reasonable extension of time, the engineer/architect must: 1. Apply the rules of the contract. 2. Recognise the effects of the constructive change. 30 31 32 Delay and Disruption in Construction Contracts. Pickavance, K. First Edition. Construction Practice Series, LLP Reference Publishing, 1997. Building Contract Litigation. Fourth Edition. Fenwick-Elliot R. Longmans, 1993. Construction Contracts: Principles and Policies in Tort and Contract. Duncan-Wallace I. Sweet and Maxwell, London, 1986. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:43 PM Color profile: Disabled Composite Default screen Cost and time modelling 47 3. Make a logical analysis, in a methodical way, of the effect of delay event on the contractor’s programme. 4. Calculate rather than make an impressionistic assessment of the time taken up by the delaying event. This case demonstrates the importance of realistic and resourced programmes for the contractor, employer and the contract administrator. However, there is no consistency of approach to the requirements for programming in the standard forms of contract. For example, the Agreement for Minor Building Works, MW 80, does not require any programme from the contractor. The JCT 80, whilst requiring a programme, gives no definition of its content or how it is to be prepared and monitored. A realistic programme that is directly linked to the cost model and allows simple, sufficiently accurate, monitoring of site progress is vital in order to allow systematic evaluation of the effects of delay. For example, consider delays caused by failure to provide drawings, instructions and details (breach of JCT 80 Clause 25.4.6). If these delays are to constitute a compensational delay event, then by definition the information must be requested at a time not too distant from nor too close to the time at which the contractor actually needs it, and not the time he might reasonably be expected to need it if everything went to plan (London Borough of Merton v. Stanley Hugh Leach [1985] 32 BLR 68). The contractor’s actual need must be considered in the light of actual progress, not the theoretical progress that might have been made had the contractor been proceeding efficiently. Realistic programmes that allow resource monitoring would be especially important in cases where the contractor alleges that additional costs have been suffered as a result of disruption, or that the progress of the work has had to be accelerated (constructive acceleration). The contractor must be able to demonstrate that in the event of the delaying factor, the resources, or lack of them, were not a cause or principal cause of the delay and, secondly, where acceleration has been alleged, the contractor must prove that the reasonable resources have been increased beyond the resources planned at the time of tender. This “ideal” of the role of the programme is reflected in the Delay and Disruption Protocol issued by the Society of Construction Law in October 2002. Core Principle 1 of the Protocol states that the: “Contractor should prepare and the Contract Administrator (CA) should accept a properly prepared programme showing the manner and sequence in which the Contractor plans to carry out the works. The programme should be updated to record actual progress and any extensions of time (EOT) granted.” Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:44 PM Color profile: Disabled Composite Default screen 48 Quantifying and managing disruption claims The Society goes on to suggest that, if this is done, “the programme can be used as a tool for managing change, determining EOTs and periods of time for which compensation may be due”. At the time of writing, it is simply too soon to comment on what “properly prepared” means. However, the guidance and the model clause in the Protocol make reference to the Programme being constructed and accepted in accordance with a Method Statement “describing in detail how the Contractor intends to construct the works, and the resources …it intends to use to do so”. As we shall see in Chapter 4, the Protocol’s approach is similar to the role of the programme envisaged in this book. However, the method in this book goes much further in specifying how the work operations on a method statement must relate to work activities (work packages) on the contractor’s programme. 3.6.1. Damages and delay Where a contractor loses money by virtue of the construction project being delayed, it is trite law that losses can only be recovered from the employer/guilty party if and to the extent that party has been the cause of the delay. Put simply, the law of contract applies a “fivefold” test: 1. Was the loss caused by the default/delay event? (The causation issue.) 2. Was the loss foreseeable as the consequence of the breach? (The foreseeability issue.) 3. Is the loss too remote from the delay event? (The remoteness issue.) 4. Was the loss contributed to by the default of the contractor? (The contribution issue.) 5. Did the contractor take reasonable steps in an attempt to avoid the loss? (The mitigation issue.) 3.7. Practical conclusions 1. BoQs were initially developed to provide an equal basis for tendering and for interim payment purposes, typically on a remeasurement basis. The present day uses of the BoQ have exceeded its original objectives. 2. Today, the BoQ has a role in estimating, tendering, planning and control and also has a vital role in the evaluation of the monetary consequences of variations, delays and disruption. The majority of the negative comments about the BoQ are related to the level of detail. Too many low-value items exist Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:44 PM Color profile: Disabled Composite Default screen Cost and time modelling 49 which have little influence on the estimating accuracy of the overall BoQ, yet which still consume cost and effort in their pricing. 3. The BoQ does not encourage operational estimating, which is considered to be more realistic to the manner in which costs are actually incurred in the construction process, and therefore capable of generating estimates of greater accuracy. 4. Because the work items are based on the materials to be incorporated, they do not require the contractor to reflect the factors to which the costs are actually sensitive, and the systematic evaluation of variations is, in practice, limited to changes in quantity. The automatic evaluation of the cost effects of disruption is beyond the scope of the current BoQ. 5. Previous improvements in the BoQ as a cost model have come in two forms. Both have sought to reduce the complexity of the BoQ, thereby increasing the estimating accuracy and reducing the costs and time involved with its production (Table 3.4). Table 3.4. Summary of research attempts to improve the BoQ Significant items Operational modelling Short (1970) Skoyles (1968) Barnes (1971) Barnes (1971) Moyles (1973) Harmer (1983) Saket (1986) Allman (1988) Betts and Gunner (1993) McGowan (1994) 6. The significant items approach to estimating is effective. It reduces the number of items to be estimated and model factors can be used to account for the insignificant items. However, significant items retain the description based on materials to be incorporated and are still based on a detailed method of measurement. They still advocate a unit rate type of estimating and do not readily link cost with time. Significant work items alone cannot be isolated and controlled in practice. The real practical problem is that the structure of the work described in a BoQ cost model needs to be improved. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:44 PM Color profile: Disabled Composite Default screen 50 Quantifying and managing disruption claims 7. Improvements need to take account of the fact that the lowvalue items, whilst not strictly necessary for estimating, must be itemised somewhere in the contract. There is a need to increase the realism of estimating and to create a direct link between cost and time. The approach in this book seeks to allow the direct effects of delay and disruption to be evaluated both in terms of cost and time. It also facilitates sufficiently accurate construction site control. Further, the approach does not restrict the contractor’s flexibility in deciding methods of construction and pricing, and allows tendering on an equal basis. 8. Resource/labour productivity is at the centre of estimating, the planning process, site control and the evaluation of delays and disruption. The proposition is that resource productivity should be used to develop a new approach to modelling cost and time. In short, a cost model composed of work packages related to site operations, linked to a time model composed of the identical work packages, would form a direct tangible link between cost and time. A single labour and/or plant productivity would characterise the work package. Chapter 4 demonstrates the development of such “characteristic productivity” work package models. The process represents a synthesis of the operational modelling and significant items approaches with the distinct advantage that the models are not project-specific and do not involve highly sensitive model factors. Thomas Telford\Quantifying and Managing Disruption Claims Friday, November 15, 2002 3:07:44 PM