Flood prevention schemes: guidance for local authorities

SECTION 5: COMPARING COSTS AND BENEFITS

Estimating the costs

Whole-life costs

5.1 For any economic appraisal, the anticipated benefits of a scheme must be compared with its expected whole-life costs. All relevant capital, maintenance and running costs must be included. Maintenance estimates should allow for storm damage repairs and, where significant, decommissioning costs.

5.2 Only benefits and costs resulting from implementing each option should be considered. The appraisal should exclude 'sunk' costs, which have already been incurred, such as previous investments in defences and expenditure on feasibility studies.

Discontinuities in costs

5.3 In the same way as benefits may vary in a stepped fashion, option costs are also likely to increase in specific increments. For example, this could occur where the form of construction needs to change to accommodate a higher water level for an increased standard of service. The points at which these steps occur should be examined in detail, to assess the range of option standards where benefits may have increased without increased costs.

Residual values

5.4 Some assets may have a lifetime beyond that used in the analysis. The residual values should be taken into account in the estimation of costs and benefits only where this is required to ensure equality of assessment between different options. A straight-line depreciation over the asset life, which presupposes a decision to continue using the asset, will usually be appropriate.

Rates and prices

5.5 Estimates should reflect the nature and scope of the work to be constructed. Aspects to be considered should include site conditions, location, size, complexity, risks, programming and timing constraints, availability of resources, construction methodology, specification and conditions of contract. Wherever possible the estimate should be based on appropriate cost data from historical tenders, published articles and estimating price books, estimates and quotations from companies for specialist work, and the estimator's own experience.

5.6 Where less common items constitute a significant part of the overall cost, it is often necessary to make a careful assessment of quotations and estimates obtained from operators with commercial experience in that sector.

Price indices

5.7 The base year used for pricing should always be stated. When data are not available for that particular year, it may be necessary to use appropriate indices to convert historical prices to the same base. At feasibility level, costs and benefits can generally be indexed using the all-items index of retail prices. For construction costs, the public works non-road index is appropriate for most uses.

5.8 Particular components may constitute a large proportion of a project cost, or the cost of those components may be expected to vary in real terms over time. In such cases, sensitivity analysis should be used to explore the implications for option choice.

Other cost elements

Contributions from others

5.9 When considering project costs, it may be incorrect to deduct payments from developers, highway authorities or other contributors. Generally, such windfalls only affect the distribution of costs and not the total resources required for the project. If the associated benefits were excluded from the appraisal, it would be reasonable also to exclude the contributions. However, particularly when those contributions come from other budgets of public money, it is preferable to include all benefits and costs. In this way the benefit-cost analysis will demonstrate whether the project as a whole is justified.

5.10 When the contributor is a commercial organisation, it may be prepared to pay on the basis of the financial benefits to that organisation, which may be greater than the economic benefits. For example, the owner of a supermarket might be prepared to contribute an amount which reflects not just the direct losses that would be experienced from flooding, but also the loss of trade where in economic terms the loss is simply a transfer payment. The economic implications of such contributions therefore need careful consideration, having regard to the particular circumstances.

Negative costs

5.11 Sales that offset the costs of construction are to be treated as benefits of the scheme. For example, the sale of sand or gravel excavated as part of a channel widening scheme, or charges raised for the incorporation into the scheme of arisings from others, should be treated as a benefit (paragraphs 3.22, 3.23).

Negative benefits

5.12 Disbenefits such as noise and disruption caused by project works, and obstructions to views, should be treated as costs (paragraph 3.23). In general, however, they are likely to be better handled as part of the consultation process and environmental appraisal, through which ameliorating actions are likely to be identified and included in scheme costs. If not, the costing of the disbenefits is likely to be disproportionate to their magnitude. Their value will only require consideration when there are significant differences of impact between different groups or between different options.

Treatment of project risk

5.13 There is a widely recognised, general tendency for appraisers to be overly optimistic in their early assessment of project costs, time-scales and benefits, when these are compared with final outturn values. This is termed "Optimism Bias". Prior to the revision in 2003 of the Treasury Green Book, this bias was taken into account in a generalised way through a percentage premium included within the test discount rate. HM Treasury have unbundled the issue from the discount rate, which has been reduced from a flat 6% to a variable 3.5% (reference 2). Consequently, an explicit consideration of Optimism Bias is required through (i) the application of suitable uplift factors to early best estimates of project costs, and (ii) sensitivity analysis of predicted benefits and project time-scales.

5.14 The best estimate for any project should be the appraiser's assessment of the most likely outturn costs of the project. These should include, for example, all labour, materials, supervision, land purchase, compensation, access costs and contractors' overheads associated with both temporary and permanent works, and all long-term costs associated with operation and maintenance. Where the estimator judges that additional sums are likely to be required for particular areas of work, for example, for dealing with poor ground conditions, these should be included, but general contingencies should be estimated as part of the process of deriving the optimism bias adjustment. All elements of the estimate should be based on experience of projects of similar character and should recognise the likely difficulties involved in carrying out works in particular circumstances, for example, the high cost of working in confined spaces or within, or adjacent to, private properties in urban areas.

5.15 Sensitivity analysis of benefits and project time-scales is an important element of best practice. However, the new approach to Optimism Bias in cost estimates requires a strengthened procedure. Best practice guidance in relation to strategy and scheme costs is set out below. This develops the interim guidance issued by the Executive to local authorities in March 2003 (reference 11).

• Strategy costs (initial feasibility stage)

At this stage it is assumed that no detailed design has been carried out and that, therefore, cost estimates are based on broad assumptions about the scope and nature of the work.

Step 1: For each option, identify best estimates of all capital, operating and maintenance costs given current information.

Step 2: Take a starting value for Optimism Bias of 60% of total Present Value costs (including capital, operating and maintenance costs over the whole life of the option). This percentage reflects the current view of the average cost uplift from strategy/pre-feasibility stage to the final account stage.

Step 3: Study Annex C, which sets out the current view of the key components of risk that make up the overall 60% factor. Assess whether the contributions of these components should be higher or lower for the particular situation under consideration. Where demonstrable action has been taken to minimise individual risks, a case can be made to reduce the relevant component(s). Conversely, where the project is considered to be more risky than average in certain areas, perhaps because of innovation, say, the relevant risk component contributions should be increased. In the absence of evidence either way, the default risk component percentages (Annex C values) should be left unchanged.

Step 4: Rework the overall Optimism Bias factor based on any revisions to risk components. Apply the revised Optimism Bias factor as a percentage uplift to total Present Value costs (in place of any contingency estimate).As an alternative, where a full "Monte Carlo" type risk evaluation has been undertaken, then the 95% confidence level estimate should be used to derive the optimism bias factor. Where any present value costs are not included in the risk approach these should be adjusted using the above approach and added to the 95% confidence risk based result.

• Scheme costs (detailed design stage)

To reach this stage it is assumed that appropriate site investigations and detailed design of the main works have been carried out, so that major cost items are based on detailed assessments of works required from substantially complete working drawings and specifications.

Proceed as for strategy costs, but use a starting Optimism Bias factor of 30% and the relevant risk component guidelines in Annex C.

Discounting and economic efficiency

5.16 To test the economic efficiency of different options on a comparable basis, it is necessary to discount all of the costs and benefits of the scheme, from the time they arise in the future, to their present value. The test discount rates specified in the Treasury Green Book are 3.5% for years 0-30, 3% for years 31-75, and 2.5% thereafter (reference 2).

5.17 The convention that should be adopted is to take all costs and benefits in any given year as accruing at the midpoint of that year, and to discount all these streams back to their present value at mid-year 0. This is the time at which capital expenditure is also to be taken to start to accrue.

5.18 In terms of good practice:

• the test discount rates specified by the Treasury are to be used for all streams of benefits and costs;
• each and every benefit and cost should be taken to accrue in the middle of the year when it occurs, and
• present values should be calculated as at the mid-year of year 0.

5.19 The economic efficiency of options can be assessed by comparing their benefit-cost ratios (ie the present value of benefits divided by the present value of costs), and their net present values (ie the present value of benefits less the present value of costs).

5.20 Where different options offering different standards of protection are compared, the incremental benefit-cost ratio of the higher standard of protection equals the increase in benefits resulting from that higher standard divided by the additional cost.

5.21 Because of the limitations inherent in comparing schemes by use of a single indicator, it is good practice to plot the changes in the different streams of benefits and costs over time. Discontinuities are either a sign of a change in conditions or an arithmetic error.

Economic sustainability

5.22 The economic sustainability of different options can be examined by plotting the distribution of net annual benefits over time. For example, Figure 5.1 shows 3 options with very different distributions of net annual benefits (ie the difference between expected annual benefits and costs for each year of the scheme life) but very similar benefit-cost ratios and net present values ( NPVs). Option 1 has (marginally) both the highest benefit-cost ratio and NPV. However, unlike the other 2 options, the net annual benefits of option 1 are negative in the long run. With option 2 there are some significant initial costs and the benefits are not immediately realised in full, but in the long- term stable benefits are achieved. Option 3, shows increasing benefits over time but also high recurrent costs.

5.23 In such a case it would not be appropriate to attach significant weight to the relatively minor differences in benefit-cost ratio or NPV but rather to examine the wider area of economic sustainability. This might include at least an outline appraisal over a longer time period so that longer-term gains and losses can be taken into account.

Figure 5.1 Comparison of options with different expenditure profiles