## The capital cost estimate

This section illustrates the use of quantitative risk analysis for the evaluation of the capital cost estimate for the timber development. The base estimate at the preliminary design stage is summarized in Table 21.4.

The first stage of the analysis estimated a proportional variation about the base value for each major item. This was estimated as a probability distribution of percentage changes from the base, as shown in Figure 21.2 and Figure 21.3. Detailed documentation extended the form in Figure 21.3 to include a statement of data sources, major assumptions and the reasons for the particular numerical values.

The variation distributions for items in the estimate were converted to cost distributions by multiplying by the base cost values. The cost distributions were then added, to form a distribution for the total budget cost (Figure 21.4). The additions were performed using the @Risk simulation package embedded in a standard Excel spreadsheet. The analysis indicates a very high likelihood (better than 99%) of achieving the budget with contingency, and a 95% likelihood of achieving a cost less than 105.

In the addition process, substantial positive dependence was assumed between most items: if the cost of one item is higher than expected, the costs of other items are likely to

 Item Base cost 1. Log handling 17 2. Saw mill 15 3. Drying s 4. Dry mill and planer mill 7 5. Pressure treatment 3 6. Boiler 11 7. Turbo-alternator 6 s. Fresh water system 4 9. Offices 1 10. Workshop 2 11. Mobile equipment 4 12. Mill site 3 13. Port facilities 5 Total civil and equipment s6 14. Construction overhead 3 15. Project engineering 6 16. Construction management 1 17. Start-up costs 1 18. Training 3 Base estimate 100 19. Contingencies 10 Total capital budget 110

be high also. Relationships of this kind arise due to common dependencies on such factors as imported equipment, local labour rates and productivity, or an overall conservatism in the generation of the estimate. Failure to account properly for dependence can lead to serious underestimation of risk: Figure 21.5 shows the effect of interdependence on the cost variation associated with the four most expensive items, Items 1, 2, 3 and 6. (The 'kinks' in the curves in Figure 21.5 are associated with asymmetries in the underlying distributions, as seen in Figure 21.2.)

Figure 21.4 showed the build-up of cost components to form the total construction cost. 'Sliding' the curves in Figure 21.4 to a common zero-probability point allows the variations attributable to groups of cost items to be identified and compared (Figure 21.6). Here, the four most expensive items — the log handling plant, the sawmill, the dryer and the boiler — appear to contribute the major part of the overall variability. Analysis of this kind facilitated assessment of those areas to which management effort might usefully be directed to provide better estimates, to obtain more precise forecasts or to monitor more closely. In our experience, risk analysis and cost control processes during preliminary design, final design and project implementation contribute significantly towards capital cost reduction and cash flow optimization.

Financing charges and interest during construction were not included in the capital cost breakdown. However, for optimization purposes, they were taken into consideration when

Item 3, drying

Item 3, drying

Item 3, drying

Figure 21.2—Variation distribution in histogram and cumulative forms

Figure 21.2—Variation distribution in histogram and cumulative forms

 3. Drying Description ... Description of the item, including references to preliminary designs, equipment lists and estimates, other data sources, major assumptions... Base cost 8 Sources ... Sources of variation in the base cost... Variation (%) -20 -15 -10 -5 0 +5 +10 Probability 0.15 0.18 0.28 0.30 0.06 0.02 0.01 Discussion ... Notes on risk treatment and management actions...

Figure 21.3—Uncertainty about the base cost for Item 3, drying, illustrated in Figure 21.2

Figure 21.3—Uncertainty about the base cost for Item 3, drying, illustrated in Figure 21.2

Cost

Figure 21.4—Preliminary estimates for costs excluding contingency

Figure 21.4—Preliminary estimates for costs excluding contingency

Cost

-Independent ---Dependent

Figure 21.5—Effect of dependence on the combined cost of the four largest items preparing the project time schedule. In this respect, the influence of phasing, completion time reduction and delayed or reduced financial exposure was further analysed for improved robustness.

For some purposes, the individual lines of the cost estimate might be sub-divided into component costs, or into quantities and unit prices, or a range of separate sources of variation might be estimated. Such refinements are useful at later stages of a project, when they provide the project owners and management with detailed guidance for contingency planning and control, for setting contract terms and conditions when tenders are let, and for allocating risks, costs and rewards between owners, contractors, suppliers and insurers. A detailed analysis of this kind was not necessary for evaluating the preliminary design estimate for this project.

Cost variation

Figure 21.6—Cost variation contributed by Items 1, 2, 3 and 6, civil items, and all other items excluding contingency

The risk analysis yielded a distribution of estimated capital cost in Figure 21.4, rather than the single value in Table 21.4. This enabled realistic figures to be given for the capital budget and contingency allowances, and it allowed the probabilities that the actual construction cost would be within these targets to be assessed. This is particularly important for projects that are unique in some way, are in new areas, or have special risks associated with them, where the use of a rule-of-thumb percentage contingency, based on experience in a stable environment, may not be appropriate. As Figure 21.4 indicates, the capital cost targets for this project were conservative, with a high probability of achievement, which is unusual.