The Water Industry Commission for Scotland Scottish Water. Scottish Water. Reporter s Report SR10 2 nd DBP Appendix G CAPEX Estimating Systems

Transcription

1 The Water Industry Commission for Scotland Scottish Water Scottish Water Reporter s Report SR10 2 nd DBP Appendix G CAPEX Estimating Systems March 2009

2 SCOTTISH WATER REPORTER S REPORT SR10 2 ND DBP APPENDIX G - CAPEX ESTIMATING SYSTEMS CONTENTS G1 INTRODUCTION 1 G1.1 General...1 G1.2 Approach to Audit of the Engineering Estimating System...2 G1.3 Summary of material changes from the 1 st DBP...2 G1.4 Key Points...3 G2 THE ENGINEERING ESTIMATING SYSTEM 6 G3 DIRECT COST FUNCTIONS 9 G3.1 Sources of Cost Data...9 G3.2 Non-infrastructure Direct Cost Functions...9 G3.3 Infrastructure Cost Functions...12 G4 PROJECT ON-COSTS 16 G4.1 Overview of Project On-costs...16 G4.2 Assessment of On-Costs for the 2 nd DBP...18 G4.3 Application of On-costs for the Cost Base...20 G4.4 Site Specific on-cost...21 G4.5 Contractors on-costs...23 G4.6 Delivery Partner on-cost...24 G4.7 SW Client on-costs...26 G4.8 Risk (tender to out-turn adjustment)...27 G4.9 Estimate to Outturn Calibration...28 (i)

3 GLOSSARY AND ABBREVIATIONS 1DBP SR10 1 st DBP (May 2008) 2DBP SR10 2 nd DBP (April 2009) CAPEX Capital Expenditure EES Engineering Estimating System OPEX Operational Expenditure Q&S2 Quality and Standards Period 2 ( ) Q&S3 Quality and Standards Period 3 ( ) SR06 Strategic Review 2006 SR10 Strategic Review 2010 WLC Whole life cost Document issue details: B&V project no Client s reference no. SR10 2nd DBP Version no. Issue date Issue status Distribution 01 Various Draft Scottish Water 02 27/03/09 Final WICS and Scottish Water Notice: This report was prepared by Black & Veatch Limited (BVL) solely for use by Scottish Water and the Water Industry Commission for Scotland (WICS). This report is not addressed to and may not be relied upon by any person or entity other than Scottish Water and WICS for any purpose without the prior written permission of BVL. BVL, its directors, employees and affiliated companies accept no responsibility or liability for reliance upon or use of this report (whether or not permitted) other than by Scottish Water and WICS for the purposes for which it was originally commissioned and prepared. In producing this report, BVL has relied upon information provided by others. The completeness or accuracy of this information is not guaranteed by BVL. (ii)

4 SCOTTISH WATER REPORTER S REPORT SR10 2 ND DBP APPENDIX G CAPEX ESTIMATING SYSTEMS G1 INTRODUCTION G1.1 General This report on Scottish Water s SR10 Costing Systems has been prepared for the Water Industry Commission for Scotland (WICS) by the Reporter as part of the Reporter s report on Scottish Water s 2 nd DBP for SR10. This report covers the development and application of Scottish Water s CAPEX estimating systems which have been used to cost: Elements of the non-infrastructure and infrastructure enhancement programmes. The MEAV. Elements of the Cost Base. In total, these costing systems have been used to cost 62% of the 2 nd DBP by value. The detail findings of our report are set out in the following sections: Section G2 Section G3 Provides an outline of the Engineering Estimating System (EES) which is the proprietary software system which provides the platform for SW s cost estimating systems. Describes the development of direct cost functions for infrastructure and noninfrastructure assets. Section G4 Describes the development of project on-costs applied to complete the estimates. This introduction concludes with: An outline of our approach to the audit (Section G1.2) A summary of material changes from the 1 st DBP A summary of the key points arising from our audit (Section G1.4) (G-1)

5 G1.2 Approach to Audit of the Engineering Estimating System During our audit of the costing system we: Met with SW staff to understand the methodologies used to develop and populate the Engineering Estimating System. Reviewed the detailed methodology statements and user guidance prepared by SW. Reviewed the direct cost functions prepared by SW and selected an audit sample from the projects used to populate the system. Audited sample projects for allocation of costs and yardsticks which form the data points included in the resulting cost models. Audited the calculation of project-on costs. During subsequent audits of the Investment Plan, MEAV and Cost Base, we used knowledge developed from the audit of the costing system to challenge the application of cost functions, in particular, assumptions made regarding the scope of works covered by individual cost functions. The outcome of these audits is addressed in the Investment Plan sections of our main report on the 2 nd DBP. G1.3 Summary of material changes from the 1 st DBP SW has continued to develop and improve its internal costing systems from the 1 st DBP. We note the following material changes to the Capex estimating system from the 1 st DBP: 1. The proportion of the business plan costed using EES cost functions and oncosts through the PEF analysis sheet has increased to 62%. 2. SW has updated its cost data so the majority of data used to prepare EES cost functions are now derived from target costs for Q&S2 and Q&S3 projects. This has allowed SW to retire a large proportion of the AR04 legacy data which formed the basis of much of the non-infrastructure cost functions used to price the 1 st DBP. 3. SW has updated its on-cost percentages. The revised on-cost percentages have been derived from Q&S3 projects. As such, they represent current procurement practice and are consistent with the Q&S3 direct cost data points used to populate the cost models. Updating this data has clarified audit trails to the oncosts applied and demonstrated consistency with the direct cost estimates. 4. Because of the use of Q&S3 data, SW has been able to develop a wider range of matching Investment Plan and Cost Base data points which have formed the basis of separate Investment Plan and Cost Base cost functions. 5. Further work has been done on the quality assurance of the cost data and development of the direct cost functions. (G-2)

6 G1.4 Key Points 1. SW s CAPEX estimates for the SR10 2 nd DBP have been prepared using a range of systems including econometric analysis, bottom up estimates and unit rates for outputs derived from the Q&S3A programme delivery. Approximately 62% of the costs in the Business Plan have been derived from a structured estimating system called EES (the Engineering Estimating System). This report addresses the development and application of the EES. Other cost estimating systems used in the 2 nd DBP are described in the relevant sections of the main report. 2. Cost functions derived from the EES have been used to cost the following elements of the SR10 2 nd DBP: a. The majority of the quality enhancements for water and wastewater treatment and UIDs as well as elements of the capital maintenance plan. b. The modern equivalent asset valuation. c. The infrastructure Cost Base and elements of the non-infrastructure Cost Base. 3. SW has made significant improvements to its estimating systems from those used in the Q&S3 process and SR06 Business Plan submissions. In particular: a. In-house estimating systems have been used. b. The estimating system is largely based on SW s own data including historic project data and current tendered rates. c. SW has updated the source data used so that the costing system is largely based on CAPEX 3 data for Q&S2 and Q&S3A. As a result, SW has been able to retire much of the legacy AR04 data used in previous submissions. This brings the cost estimating system up to date and has improved confidence in the consistency and analysis of the data. d. SW has developed on-cost factors from the analysis of historic projects. This analysis draws on the same recent projects used to populate the direct cost functions. As a result, SW has developed a consistent costing system which reflects recent experience and methods of procurement. e. A proprietary estimating system (the Engineering Estimating System) has been adopted which provides a structured and controlled environment for capturing and applying cost data. f. A consistent estimating system has been applied demonstrating clear links between the SR10 CAPEX estimates, the asset valuation and the Cost Base submission. g. The system is well documented. h. Good audit trails have been maintained of the review and assessment of data. i. SW has further reinforced the need for consistency between the way direct cost yardsticks are derived and the way they are applied. (G-3)

7 4. From our audits of development and application of the EES estimating system we note the following concerns: a. Our detailed audits exposed a range of issues regarding the quantification of yardsticks and the consistent allocation of costs between cost functions. SW has worked to address the concerns we raised; however, we believe that there is further work to do to resolve these issues and put in place systems which reduce the scope for error in the future. b. While these issues affect confidence in individual cost functions, we recognise that SW aims to capture all costs of new work on a project within the project data points used in the EES cost functions. In addition, we did not conclude that yardstick errors were systematic. c. We had particular concerns regarding the allocation of costs and the derivation of yardsticks for the Cost Base cost functions. SW has addresses these issues for the specific cost functions used in the Cost Base which we have reviewed. We note that this targeted approach does not address the risk or errors in other cost functions. d. The estimating system is based on CAPEX3 estimates. SW has undertaken some work to calibrate the estimating system against historic out-turn costs. However, this work is not yet complete and SW has not applied any calibration factor to the Business Plan. Given that SW is not anticipating any significant pain/gain share we accept that CAPEX3 estimates are likely to be close to outturn costs. We also recognise that any estimate to out-turn adjustment would also be included in Cost Base and would affect SW s efficiency position. Despite this, the completion of a estimate to out-turn calibration would have provided additional comfort that the costing system was robust. e. We believe that the level of percentage on-costs applied to the estimates is relatively high. However, SW has been able to demonstrate that it is based on recent project estimates and reflects current experience and procurement practice. In addition, where appropriate, SW has applied the same level of oncost to the Cost Base which will impact on its efficiency position. f. When considering the level of on-cost applied to the estimates, we considered whether a high level of on-costs reflects an intensive effort in project development to minimise the scope of works delivered within a challenging SR06 determination. A combination of high resource cost and low direct asset cost would result in a relatively high on-cost percentage. We are concerned that for the Business Plan, these high on-cost percentages are often applied to a scope of works based on standard solutions where the scope has not been reduced to reflect the challenge inherent in the level on-cost applied. SW has responded to our concern noting that additional resource costs to reduce scope may only be part of the reason and on specific projects only. SW has rightly expressed concern about drawing any judgement from these observations. SW is of the view that the high on-costs now being experienced can be explained by: o Earlier contractor and designer involvement in the project definition stage. o Less opportunity for contractual disputes. (G-4)

8 o Assistance in re-building the SR06 enhancement programme and recover from Year 1 delays. o Provision of specialist design/ commercial expertise and construction capacity in a buoyant marketplace. o Transfer of 3rd party risks (land, planning, utilities etc.). o Cost and process design risk transfer, backed by parent company guarantees. o Obliged to make investment decisions on lowest whole life cost. o CDM client responsibility transferred; incentivised to improve H&S (KPC). o Market testing of 20% of allocated programme. o Specific challenges associated with remote areas. g. SW s estimating systems includes a site-specific on-cost function which relates to general site works such as roads, fencing, ducting, drainage and general site wide cabling. SW has applied these percentage on-costs across the Business Plan including capital maintenance items for the replacement of plant. We were concerned that on-costs derived for major plant construction or replacement work are applied to maintenance costs for direct plant replacement. We recognised that there is a need to maintain consistency at a programme level between the way on-costs are derived and the way they are applied. In this respect, we were worried that the projects used to derive the on-costs contained a material element of direct plant replacement work suggested that SW should consider whether it is appropriate to apply the full on-costs to all types of work in light of the balance of work included in their derivation. SW responded that it had tried to separately analyse on-costs for CM projects but found that it was not possible to produce CM only on-costs across the 5 service types and 3 value bands. SW also noted that such projects were rarely straightforward particularly working on live, aged assets which need to be left in a safe working condition. It therefore considered that the site specific on-costs used for SR10 CM projects were reasonable. We accept that working with live plant on operating sites can be complex. h. It is noted that the full available data set was used in the analysis and that there was no sampling. In a limited number of examples we found that the cost functions had not been applied in the same way as the data had been provided. For example, where yardstick data for net process volumes had been applied to cost functions based on a gross structure volume thus underestimating costs. We believe that there is a need for continuing checks and controls spanning the development and use of the cost functions to ensure that they are applied correctly. (G-5)

9 G2 THE ENGINEERING ESTIMATING SYSTEM SW has introduced a proprietary estimating system known as the Engineering Estimating System (EES) which provides a structured and controlled environment for capturing cost data. The EES is widely used in the water industry supported by an active user group of which SW is part. This allows sharing of knowledge and the sharing of user experience across the industry. It has allowed the functionality of EES to develop to fit the regulatory requirements of the water industry including the development of high level estimating facilities and development of Cost Base estimates. The engineering estimating system has a wide range of functionality which allows it to capture costs and prepare estimates at a range of levels from detailed schedules of plant labour and materials through to all in costs for a treatment works. Typically, data capture and estimating capability is available at the following levels of detail: Plant, labour and material estimates prepared from detailed resource costs and productivity rates entered in the system. Item rates, which can be built up from specific plant, labour, material estimates or entered directly in the system. An item rate might be the all-in rate for reinforced concrete placed or laying a pipeline to a standard depth in standard conditions. Element rates built up from a combination of item rates. For example, the sum of standard specification items or the extra-over items required to allow a complete pipeline process rate to be built up. The element rate is created from a combination of item rates. Process rates, where a process is a section of the works, such as a pipeline or a settlement tank, which can be described by a limited number of simple yardstick quantities. The process rates can be built up from one or more element rates or can be entered directly in the system. These process level rates are used to develop the direct cost functions which were used to prepare estimates for the 2 nd DBP. It is possible to introduce source data at any level of estimating system. By proper structuring of the data it is possible to roll up information from one level to another, for example, allowing data from detailed plant labour and material estimates to be rolled up automatically to generate item rates, element rates or process rates. The cost data used for the 2nd DBP was generally entered into EES at one of two ways: As detailed plant, labour and material rates and productivity rates which are used to prepare Q&S3A CAPEX2 and CAPEX3 estimates. The detailed costs and productivity rates used are tendered rates from the Q&S3A Associate Delivery Partner framework tenders. The CAPEX2 and CAPEX3 estimates are generated within a structured coding system which allows costs to be rolled up to a process level rate. Individual process level rates from other sources including SW s AR04 data and CAPEX3 estimates from Q&S2. These are prepared and entered as a process level (G-6)

10 cost and yardstick and do not carry detailed underlying information on rates or quantities. SW employed cost consultants to undertake checks and reviews on the cost data which were used to develop the direct cost functions used in the 2 nd DBP. In part, this provided a quality check on the cost data. Of equal importance, it provided a sense check that the data points were valid in terms of the scope of works included. It acts as the control point to data points used in the direct cost functions. The check and review process was well structured and detailed records are maintained to support the information. The process level cost functions used for the 2 nd DBP were developed within EES. The cost functions are a statistical best fit analysis to the cost data. The best fit analysis routine considered linear and power cost functions. Two possible power functions were considered, one constrained to pass through the origin and the other unconstrained. In general, the best fit cost curve was selected on the basis of lowest standard error. While the preparation of the cost functions in EES is an automated process, SW has also engaged the services of a statistician to review the data points and cost functions to identify outliers and test the impact of data points which have a particular influence on the cost functions. The EES cost functions can be updated as new data is entered on the system. For the 2 nd DBP SW has fixed the cost data at a point in time to develop a defined series of Business Plan cost functions. These cost functions have been captured in a Project Estimating Form (PEF) spreadsheet. The PEF spreadsheet made output of the EES costing system available to the Business Plan teams in a controlled fashion. In addition, the PEF spreadsheet provided an output of structured costing information linked to asset codes and drivers which mapped through a database to the individual Business Plan table lines. We were impressed by the structure, quality and visibility of data from the PEF estimating sheet and associated summary sheets. This provided access to blocks of data in a structured format. It improved our ability to construct and follow audit trails from project estimates to the 2 nd DBP. The process level cost functions are cost functions aimed at capturing the plant, labour and material costs which are the direct construction costs of specific items of work. To provide a complete cost it is necessary to add the indirect costs which include: General site costs which cannot be allocated to a specific process. Costs of the construction management and facilities on site. Professional services including design costs and programme management costs. Client on costs. Indirect costs can also be added at different levels of detail in EES. For the purpose of the 2 nd DBP, SW has developed a matrix of standard indirect on-cost percentages for five asset groups and three levels of investment which are applied to the estimated direct cost to provide a complete project estimate. (G-7)

11 The development of the direct cost functions is described in Section G3. The development of the on-costs applied to the direct cost functions so as to provide a complete estimate is described in Section G4. (G-8)

12 G3 DIRECT COST FUNCTIONS G3.1 Sources of Cost Data The cost functions used to cost the 1 st DBP were derived from four main sources of data: AR04 legacy data. Q&S2 CAPEX3 target cost estimates. Q&S3A CAPEX2 and CAPEX3 estimates based on detailed plant labour and material rates and productivity rates tendered as part of the Q&S3A Associate Delivery Partner framework tenders. GAP model data. The GAP model data are bottom up estimates which have been developed to populate a limited number of non-infrastructure cost functions. For the 2 nd DBP SW has continued to rely on the same sources of data but has refreshed the data used so that the majority of data used to prepare the EES cost functions are now derived from target costs for Q&S2 and Q&S3 projects. This has allowed SW to retire much of the AR04 legacy data which formed the basis of the cost functions used to price the 1 st DBP. G3.2 Non-infrastructure Direct Cost Functions Non-infrastructure direct cost functions are based on either: G3.2.1 Q&S3A cost data supplemented where necessary by limited quantities of AR04 legacy data. GAP models Non-infrastructure Cost Functions based on SW data The majority of SW data used to populate the EES non-infrastructure cost functions for the 2 nd DBP have been derived from Q&S3A CAPEX3 project estimates. Q&S3A CAPEX3 estimates are prepared using tendered rates for the Associated Delivery Partner (ADP) framework contracts for Q&S3A. The ADP tenders are based on priced schedules of work for items which mirror the standard cost estimates of the SR06 Cost Base. The detailed pricing schedules in the tender documents included an estimate of quantities for each standard cost item and an estimate of the plant labour and materials required to construct them. These schedules, prepared by SW, were developed to match the structure of the EES estimating system. Tenderers were asked to tender costs for the plant labour and materials identified by SW and rates of production which would then generate unit rates for the work. The tendered information was used as the basis of the commercial assessment of the tenders. The tendered information for the selected framework contracts was fed back into the EES estimating (G-9)

13 system and is used to prepare estimates including CAPEX 2 estimates, CAPEX 3 estimates and target costs. Cost estimates are generally prepared at a Cost Plan level setting out quantities for individual elements of the work. The estimates include detailed labour plant and material schedules. However, these are generally created from templates based on the tendered costs subject to minor adjustments of labour, plant and material rates where appropriate. For example, individual tendered rates for materials may be replaced by Scottish Water wide framework rates. The data points used in the analysis are derived from a range of projects across Scotland and include estimates based on the range of Associated Delivery Partner tenders received. Therefore, the costing system allows for the regional variation of costs across Scotland. Individual CAPEX3 cost estimates are inflated forward from the tendered date to mid-point of construction date based on future projections of inflation provided by Scottish Water s professional advisers. The data points for the estimating cost functions are converted to a common COPI index. The final Business Plan figures have been adjusted to COPI = In some key process areas SW does not have sufficient Q&S3A data to confidently develop a cost function. In these circumstances it has continued to rely on AR04 legacy data. In our report on the 1 st DBP we expressed concerns regarding the audit trails for AR04 data and the difficulty in confirming that the current EES allocation rules had been applied consistently in to the AR04 data points. These concerns are reduced by the limited use of AR04 data for the 2 nd DBP but some remain. For example, it is not clear that the legacy data used to prepare the water process membrane plant cost function excludes the disinfection plant and ph adjustment which is priced as separate items in the Investment Plan. Cost data points are related to a yardstick such as process volume or flow rate. Cost functions are prepared relating costs to yardsticks. The quality of the cost functions is reliant on three key activities: The correct collation and transfer of cost information from the estimate to the costing system data point. The correct allocation of process yardstick. The quality control undertaken by SW to ensure that only appropriate data has been selected for use in the analysis of the cost function. During our audit we audited a sample of Q&S3A CAPEX3 estimates to confirm the transfer of cost data into the EES data points. Form these audits we concluded that: The estimates are well structured and the transfer of information into the Process Unit rates, which is automated, appears to be carried out accurately, subject to the correct coding of costs in the estimates. There are errors in the allocation of costs between process units and the site specific on-costs which are described in Section G4. In particular, we noted one example where cabling costs which should have been allocated to individual process units had (G-10)

14 been allocated to general site specific costs because of the way the cost estimate had been constructed and coded. During our audit we audited a small sample of projects to confirm the yardstick data such as volume or flow rate. From this we identified a number of issues regarding the use of the correct yardsticks and the quality control of the yardstick information. In response to this, SW undertook a further review of its systems. Following this review we carried out a more systematic audit and concluded that the basic yardstick data still included a wide range of errors. SW has processes in place to screen the data used in cost functions to remove outliers. We take some comfort from the fact that this will have identified and eliminated some erroneous yardstick data from the analysis. However, this will also exclude valid data points which could be used if the yardstick value had been calculated properly. We did not determine a systematic bias in the yardstick estimates which would result in a systematic bias to the overall programme estimate. We raised particular concerns about the allocation of costs and yardsticks for cost functions which were used in the Cost Base which could bias the Cost Base estimates. SW has reviewed and revised the project data for these cost functions to correct for the errors identified and reviewed other data points in the same cost functions. We reviewed the quality control and quality assurance work undertaken by SW and noted the work undertaken to check and review the robustness of the cost functions including eliminating outlier data points. We noted the steps undertaken to check and quality assure the coding of project information which is the basis for allocating particular cost data to individual cost curves. However, we remain concerned by the level of error identified in the project yardsticks data and note that this had not been identified by SW s own review processes. We suggest that there is a need for end to end quality checks on data before a project is accepted into the costing system which would check individual project data and would also identify particular systematic issues which could be corrected before they create legacy data issues. SW has begun to review the costing system data in light of our concerns and has taken steps to improve the data which has a material impact on the Investment Plan and the Cost Base. SW has advised us that the changes arising from this review were not material. However, we remain concerned that there is a wider legacy issue of yardstick data quality which remains to be addressed. SW has accepted this and has put in place a process to review the data. G3.2.2 GAP Model Cost Functions SW identified a number of key non-infrastructure cost functions required to cost the 2 nd DBP where AR04 data is not available and SW has not been able to source Q&S2 or Q&S3 data. These cover chemical dosing (for ferric sulphate, sodium hydroxide, orthophosphoric acid) and wastewater tertiary sand filters. For these cost functions, SW employed consultants to prepare generic scopes of work for at least five different sizes of plant which were costed using a combination of SW s cost data and information supplied by the consultant. The data points were then used to prepare a cost function for each GAP models. (G-11)

15 During our audit we reviewed the process for developing these cost functions and audited the estimates for two of the cost models. From these audits we noted the following: These models form a material element of the overall enhancement capital programme in the 2 nd DBP The GAP models have been used for types of process where SW has experience of similar work in Q&S2. It should have been possible to take scope and cost information from Q&S2 to either inform or verify the GAP cost functions. Notwithstanding these concerns, based on our experience of similar work in the past, we believe that the models used are reasonable for planning purposes. Since the 1 st DBP, SW has undertaken further work to develop the yardstick values for these plant to ensure that they are applied in a way which is consistent to the way the cost curves were developed. G3.3 Infrastructure Cost Functions G3.3.1 Introduction In this section of the report we address the direct cost functions developed to cost infrastructure. General sub-sections cover sewerage and water mains. We also include commentary on cost functions for water mains rehabilitation and communication pipes. G3.3.2 Sewerage Cost Functions The sewerage cost functions have been prepared from data abstracted from Q&S3A CAPEX2 or CAPEX3 estimates. The sewerage cost functions have been used to prepare the MEAV for the 2 nd DBP. The sewerage cost functions have also been used to cost the UID programme of work in the Business Plan, although SW has expressed general concerns regarding the definition, scoping and costing of this work and has assumed that the work would be developed through Q&S3B and regulated under an extension of the 7-Stage Process. Other sewerage infrastructure in the capital programme has been costed through a mix of EES and historic unit rates from SR06 or Q&S2 for maintenance and enhancements. Q&S3A CAPEX2 and CAPEX3 estimates are prepared using tendered rates for the Associated Delivery Partner (ADP) framework contracts for Q&S3A. The basis and structure of this tendered data is described in the section on non-infrastructure cost data above. Separate cost functions have been prepared for manholes and sewer laying. Nine separate sewerage cost functions were developed for sewerage covering: Three surface types: of grassland; rural/suburban road; and urban road. Three depth bands of: shallow = 0-2m; medium 2-4m; and deep more than 4 m. (G-12)

16 Three cost functions were developed for manholes differentiated by depth band each covering a mix of surface types. Within each depth band the yardstick used is number and the cost function does not distinguish manholes of different diameters. The cost data used to develop the cost functions has been refreshed since the 1 st DBP and is now dominated by Q&S3A CAPEX3 data for shallow and medium sewers. However, the deep sewer cost functions continue to rely on data taken from CAPEX2 estimates prepared as part of the assessment of options for the UID Strategic Studies. We are not convinced that this data is fully representative of SW s general sewerage stock. In particular, it includes: G3.3.3 a disproportionate quantity of very deep sewers (which will impact on the deep sewer cost functions only); a disproportionate quantity of sewers constructed in difficult conditions such as in contaminated land or adjacent to water courses; Water Mains Cost Function The water mains cost functions have been prepared from data abstracted from Q&S3A CAPEX3 estimates. The water mains cost functions have been used to prepare the MEAV for the 2 nd DBP and the capital programme. SW has used the sewer cost function for water aqueducts. SW has used the water mains cost function for all sizes of sewage rising mains. Q&S3A CAPEX3 estimates are prepared using tendered rates from the Associated Delivery Partner (ADP) framework contracts for Q&S3A. The ADP tenders are based on priced schedules of work for items which mirror the standard cost estimates of the SR06 Cost Base. The detailed pricing schedules in the tender documents included an estimate of quantities for each standard cost item and an estimate of the plant labour and materials required to construct them. Theses schedules, prepared by SW, were developed to match the structure of the EES estimating system. Tenderers were asked to tender costs for the plant, labour and materials identified by SW and rates of production which would then generate unit rates for the work. The tendered information was first used as the basis of the commercial assessment of the tenders. The tendered information for the selected framework contracts was fed back into the EES estimating system and is used to prepare estimates including CAPEX 2 estimates, CAPEX 3 estimates and target costs. Three separate cost functions were developed: WAMLG water mains in grassland WAMLR water mains in rural and suburban highway WAMLU water mains in urban highway. SW has noted that the majority of water mains are laid at shallow depth and that it is not practical or necessary to prepare different cost functions by depth bands. The cost functions (G-13)

17 reflect the average depth of mains laid in the sample data and is assumed to be representative of Scotland. Each cost curve is well populated with more than 40 data points for each surface type with diameters up to or just greater than 600 mm. As with previous estimates, the data is concentrated on smaller diameter pipes, reflecting the work undertaken by SW, which then dominate the best fit analysis. However, due to the increased quantity of data available, the cost function appears to be less influenced by individual large diameter data points than in previous submissions. The CAPEX3 data used to populate the models reflect the regional and site specific issues affecting construction. The data points used are spread across Scotland, reflecting regional variation. In addition to including pipes at varying depth and under different surface types, the data points reflect the impact of services, traffic management and ground conditions and the type and frequency of fittings found identified in the sample of data used. G3.3.4 Water Mains Rehabilitation Cost Function SW has developed four separate cost functions for mains rehabilitation covering: WAMR-A WAMR-B WAMR-L WAMR-SWAB Water mains rehabilitation by surface applied internal coating. Water mains replacement by pipebursting Water mains rehabilitation by slip lining Swabbing water mains. The water mains rehabilitation cost function is based on data abstracted from Q&S3 CAPEX3 estimates. Q&S3A CAPEX3 estimates are prepared using tendered rates from the Associated Delivery Partner (ADP) framework contracts for Q&S3A as described above. Much of the water mains rehab work is delivered through a single contractor experienced in this type of work and the cost functions depend on their tendered rates. The cost functions for pipe-bursting and sliplining are well populated with each having almost 60 data points. The data generally goes up to 225 mm diameter pipes with a very limited number of data points for slip-lining up to 700 mm diameter. There is a limited data set for spray applied internal coating with eight data points drawn from three project estimates for the same area covering pipe diameters up to 315 mm. The swabbing cost function is based on four unit rates taken from the framework contract and does not rely on specific project data. The CAPEX3 data used to populate the models reflects the regional and site specific issues affecting construction. The data points used are spread across Scotland, reflecting regional variation. The data points also reflect the impact of services, traffic management and ground conditions and the type and frequency of fittings found identified in the sample of data used. (G-14)

18 G3.3.5 Communications pipe cost functions The communications cost function has been used in the 2nd DBP to price: The replacement of lead communication pipes in the investment plan. The communication pipes in Table B2.6 and B2.7. The communication pipe standard costs in Table A.2 of the Cost Base submission. The communication pipe cost function data was taken from the analysis of Q&S3A target cost estimates at CAPEX3 prepared as part of the planned water mains rehabilitation programme. Each data point is a mix of long side and short side communication pipes and therefore reflects the mix of communication pipes found on actual schemes. During our audit we noted only limited additions to the cost estimates over and above the standard cost base functions. (G-15)

19 G4 PROJECT ON-COSTS G4.1 Overview of Project On-costs SW s detailed EES estimating system generates a series of direct cost functions which are used to estimate future project costs. In addition to these direct costs SW has identified indirect costs required to deliver a project which cover: Site specifics and site wide items including roads, drainage, fencing, landscaping, intersite pipework, ducts and site wide cabling between processes (see Section G4.4). Contractors on-costs including site facilities, site supervision, contractor design and management fees (see Section G4.5) Delivery partner on-cost including project development, design, project management, delivery partner accommodation and facilities, commercial management and management fee (see Section G4.6). SW Client on-costs including specific project costs such as land purchase and professional fees incurred directly by Scottish Water and the allocation of Scottish Water capitalised costs of staff and facilities required to manage the capital programme (see Section G4.7). Risk allowance based on the risk allocation in CAPEX3 estimates and intended to estimate the project estimate to outturn variation (see Section G4.8). For the 1 st DBP, SW prepared two separate on-costs for infrastructure and non-infrastructure assets. The on-costs percentages were developed from a mix of AR04 data and analysis of historic projects. They were applied cumulatively to give a total cost estimate as follows: First DBP on-costs Asset Type Infrastructure Non-infrastructure Cost Element On-cost (%) Cum Factor On-cost (%) Cum Factor Base construction costs Site specifics 2.00% % 115 Contractor on-costs 29.22% % 146 Delivery partner on-costs 11.91% % 171 SW Client on-cost 14.84% % 196 Estimate to out-turn 5.05% % 206 Cumulative factors are for a Base Construction Cost = 100. For the second DBP, SW has completely revised its on-cost estimate based on the analysis of 377 recent CAPEX3 estimates to provide on-costs from a single source which are consistent with current project estimates and procurement experience. SW has further sub-divided its analysis of on-costs into five asset categories and three direct cost investment bands. (G-16)

20 The specific on-costs for each 5x3 matrix are set out in the individual sections below. The overall mark up on direct costs in each category is summarised in the table below (as an outturn based on direct costs of 100: Direct Direct Cost Band k k 250k to 500k > 500k Sewerage Infrastructure Sewerage non-infrastructure UID schemes Water Infrastructure Water Non-infrastructure Note the use of a single average multiplier applied to all UID schemes in light of the lack of sufficiency data to provide robust factors for each cost band. As a result of the revised analysis: The on-cost multiplier on water infrastructure projects has declined while the multiplier on sewerage infrastructure projects has increased. The lower level of water infrastructure on-costs may reflect the delivery of water infrastructure through large planned programmes of main rehabilitation work while the sewerage infrastructure programme is delivered as a series of individual schemes. There has been an overall increase in non-infrastructure on-cost multipliers with the increase on smaller schemes being proportionately larger. In the absence of scheme information for mid range UID schemes and a limited number of projects in the upper range, SW has adopted an average on-cost multiplier of 2.84 on direct costs. Changes in individual on-cost percentages are set out in the subsequent sections. The overall movement in on-costs contributing to the schemes above are: The site specific on-costs percentages for sewerage have increased in the order of 5-10% and site specific costs for non-infrastructure work have increased by the order of 10%. The contractor on-costs percentage for UID schemes have increased by the order of 15%. The Delivery Partner on-costs have increased by the order of 20% across all categories with the exception of water infrastructure where they have remained broadly constant. The Scottish Water on-cost percentage has reduced by the order of 10% across all categories of expenditure. The risk allocation percentage (which replaces the tender to outturn allowance of the 1 st DBP) has increased by the order of 5-10%. (G-17)

21 No specific allowance has been made for estimates for pain/gain mechanism included in the contracts to incentivise contractors. SW provided information on Q&S3A CAPEX3 to CAPEX5 out-turn to date which indicates that projects are delivered at their CAPEX 3 budget including risk allowance. G4.2 Assessment of On-Costs for the 2 nd DBP SW s approach to the development of on-costs for the 2 nd DBP is outlined in the introduction above and described below. SW has based its assessment on the analysis of 377 recent CAPEX3 estimates from the SR06 investment programme to provide on-costs from a single source which are consistent with current project estimates and procurement experience. The projects analysed had a total value of 301 million. SW has further sub-divided its analysis of on-costs into five asset categories and three direct cost investment bands. The 5x3 matrix of categories and the number of projects analysed in each categories is set out in the table below. Number of projects analysed to develop on-costs for the 2 nd DBP Direct Cost Band k Asset category k 250k to 500k > 500k All Sewerage Infrastructure Sewerage non-infrastructure UID schemes Water Infrastructure Water Non-infrastructure All We noted the wide ranging project analysis undertaken and concluded that it was sufficient to provide a reasonably representation of project on-costs. We believe that it would be useful to extend the analysis to include at least one additional direct cost band at the upper end of the range to provide greater confidence in the estimates for larger projects in the Investment Plan. However, given the number of data points available, introducing an additional cost band at this stage would reduce the number of projects available in the upper bands to a level would reduce confidence in the factors for the individual bands. SW has indicated that this will be considered for future exercises but will be dependent on the data available. It would be useful to further sub-divide the asset categories to provide on-cost estimates for individual process units. This information would address any concerns relating to the application of on-costs to a future programme of work which may have a different mix of process types. However, a further allocation of projects between process areas would reduce the number of projects in any one process area and size band and would reduce confidence in (G-18)

22 the on-costs estimated. SW has indicated that this will be considered for future exercises but will be dependent on the data available. We were concerned about the level of average on-costs being applied to plant replacement projects and in future we would also like to see a category of base maintenance analysed separately to enable the detailed on-costs to be more reflective of this type of work. SW has indicated that this will be considered for future exercises but will be dependent on the data available. SW also noted that it had tried to derive a base maintenance on-cost separately but found it could not do so across the 5 service areas and 3 price bands. It also stated that such projects were rarely straightforward particularly working on live, aged assets which need to be left in a safe working condition. It therefore considered that the site specific on-costs used for SR10 CM projects were reasonable. We accept that working with live plant on operating sites can be complex. The structure of the CAPEX3 estimates prepared in EES provides a sub-division of on-costs which has been reflected in the on-cost analysis as follows: EES Ref Description Commentary Direct costs Costs captured in the EES cost functions used to estimate the direct costs of schemes. 099 Site specifics Selected items are captured in site specific cost functions (for example MCCs) which are applied in the same way as direct cost functions. Remaining site wide items applied as a general site specific percentage. 100 Contractors on-cost Includes site facilities and site supervision as well as contractors profit and overheads. 101 Design Separate design costs have been added to the Delivery Partner on-cost and this item is not used in the estimating system. 103 Delivery Partner. Delivery partner on-cost include project development, design, project management, delivery partner accommodation and facilities and management fee. Separate design costs have been rolled up into the Delivery Partner oncost for the purpose of the estimating system. 104 SW as Client Includes project costs such as land purchase and professional fees incurred directly by Scottish Water as well as the allocation of Scottish Water capitalised costs of staff and facilities required to manage the capital programme 102 Risk This covers the probabilised risk allowance included in the CAPEX3 estimates as a best estimate of out-turn costs. It includes residual risks managed by the contractor, the delivery partner and the client which, if they materialise would be included in the actual costs charged to the client. CAPEX3 estimate costs were identified in each category and summed by the 5x3 matrix of categories of asset type. The on-costs percentages were calculated as the sum of the on-cost divided by the cumulative total of direct costs and on-costs in the order set out above. (G-19)

23 During the audit we were able to confirm the costs abstracted from a sample of project estimates into the on-cost analysis and noted the use of use of projects contributing to the direct cost functions in the on-cost analysis. We noted a number of minor errors in the cost transfer but concluded that the on-cost percentages represented the on-costs allowed in the broad sample of recent CAPEX3 estimates considered in the analysis. During our audit we were able to check and confirm that the calculation of on-costs followed the same order as they are applied and the percentage are based on the cumulative cost included preceding on-costs. G4.3 Application of On-costs for the Cost Base SW calculated an amended set of on-cost functions for the Cost Base, based on the same project data set. The revised Cost Base on-costs aim to reflect the general and specific requirements of the Cost Base specification. In particular: SW has excluded items such as roads and pavements from the site specific on-costs in line with the Cost Base specification. SW has only applied the site specific on-cost to Cost Base estimates where a number of process elements have been used and an allowance is required to cover items such as inter-process pipework and site wide electrics. In practice, this is limited to Table C2.8 Line 11 for additional nutrient removal at a 2000 population equivalent sewage treatment works. SW has applied the full level of contractors on-costs used in the Investment Plan for the Cost Base. SW has applied the full level of Delivery Partner on-costs used in the Investment Plan for the Cost Base. SW has a reduced level of SW Client on-cost to the Cost Base to reflect the Cost Base exclusions for planning and environmental work and land purchase. SW has limited the level of risk applied to the Cost Base to the historic SW Client risk. It has excluded the risk items managed by the contractor and delivery partner on the basis that these items reflect delivery risks such as weather and ground conditions and project definition risks which are excluded by the general requirements of the Cost Base or the specific definition of Cost Base items. In the subsequent sections on individual on-costs we have set-out the Cost Base on-cost percentages, which can be compared with the equivalent Investment Plan on-cost percentages, and commented on the Cost Base exclusions. (G-20)

24 G4.4 Site Specific on-cost Site specific on-costs cover items of work which are not included in the direct cost functions and are not covered by on-cost additions. This includes items such as inter-process pipework, provision of roads and fences, provision of site-wide cabling between processes and other items which cannot be attributed to particular element of the works. For the 1 st DBP SW included a nominal 2% for infrastructure projects and 14.75% for noninfrastructure projects based on an analysis undertaken for the AR04 Cost Base. We expressed concerns regarding this approach including the lack of an audit trail to support the assessment. For the 2 nd DBP SW has updated its assessment of site specific on-costs as described in Sections G4.2 and G4.3 above. The revised site specific on-costs are set out below. 1DBP Direct Cost Range Asset type IP < 250 k k >500 k on-cost IP CB IP CB IP CB Water Infrastructure 2.00% 2.17% 0.00% 2.42% 0.00% 3.09% 0.00% Water Non-infrastructure 14.75% 21.66% 17.55% 27.20% 19.87% 20.72% 17.00% Sewerage Infrastructure 2.00% 9.03% 0.21% 16.34% 0.64% 9.39% 1,87% Sewerage Non-infrastructure 14.75% 32.89% 22.56% 26.61% 19.54% 25.58% 20.00% UID Schemes 2.00% 16.27% 4.69% 16.27% 4.69% 16.27% 4.69% During our detailed project audits we noted examples where all cabling costs had been allocated to site specifics whereas the EES estimating rules require cabling within a process unit to be allocated to that process unit and cabling between units to be allocated to site specifics. The issue arose from the common practice of procuring all cabling under one subcontract. It was not necessary to allocate the sub-contract costs across process units to prepare a robust project estimate. The impact of this approach on the EES estimating system is to increase the site specific oncost percentage and to reduce the costs allocated to individual process units. This is of particular importance where the process units are used in the Cost Base where the site specific on-cost is not applied. In these limited cases, SW has reviewed the individual projects contributing to the relevant cost functions and corrected any misallocation of cabling costs. However, SW has not adjusted the on-cost percentages for the Cost Base or the wider Investment Plan. We accept that there will be a portfolio effect in the Investment Plan estimates where an overestimated site specific on-costs and underestimated process unit costs will cancel. However, the extent to which the costs cancel will depend on how the balance of the programme reflects the balance of projects used to develop the estimating system. For example, the site specific on-cost percentage for water non-infrastructure, which has been over estimated to some extent due to the inclusion of process level cabling has been applied to membrane plant estimates (G-21)