Flood prevention schemes: guidance for local authorities

SECTION 3: STAGES IN A BENEFIT-COST ANALYSIS

Main stages

3.1 Figure 3.1 shows the main stages in a benefit-cost analysis. It also shows the iterative loops by which the identification of the 'do something' options, and the estimation of the benefits and costs of those options, can be progressively refined.

Figure 3.1 Main stages in a benefit - cost analysis

Identifying the 'do nothing' option

3.2 Benefits and costs for all 'do something' options should be compared with those of the 'do nothing' case. The latter provides a convenient common baseline against which the other options can be assessed. Scheme benefits are calculated from the losses avoided by carrying out the proposed works rather than doing nothing.

3.3 Identifying the 'do nothing' option correctly is therefore important to the analysis and needs careful consideration. Where there is no existing scheme, the 'do nothing' option is obvious; there is no intervention in natural processes.

3.4 Where there is an existing scheme, the 'do nothing' option will be to walk away and abandon all associated maintenance and repair, allowing nature to take its course. Simply continuing with maintenance and repair of the existing structure then becomes one of the 'do something' options.

3.5 While it might appear impossible for political or other reasons to abandon the area, this should mean that the advantages of preserving what is there are overwhelming. It should therefore be easy to demonstrate that continuing present practice is better than the 'do nothing' option. If this is not the case, the non-economic reasons for continuing must be carefully considered. For health and safety reasons, it may be necessary to take minimal steps to make any abandoned works safe and these costs should be taken into account.

3.6 In some cases, due to the statutory duty to maintain certain watercourses (Flood Prevention (Scotland) Act 1961, section 4B(1)), it may be necessary to use the 'do minimum' option as the baseline. Otherwise, this option will only be appropriate when there is interdependence between action on the site and other areas, such that the cost of analysing the 'do nothing' case is disproportionate compared to the size of the project.

3.7 The 'do nothing' options should be:

where there is an existing scheme, walk away: cease all maintenance, repairs and similar activities immediately;

where there is an existing scheme, and walking away is not permissible for legal or statutory reasons, adopt the 'do minimum' option;

where there is no existing scheme, do nothing; do not intervene in natural processes.

Identifying the 'do something' options

3.8 In the early stages of analysis, the range of options should be as wide as possible; the process of analysis may itself suggest new options. In terms of design standards, the guideline for grant purposes is protection against the 1 in 100 year flood event. However, this is not intended to constrain the design process. It is simply a practical benchmark to assist with the administration of the grant scheme. Different design standards should always be considered, including the guideline standard. Such an approach should help to justify departures from that standard.

3.9 The options selected should always ¹ include the best solution in environmental terms, provided such a solution is technically feasible. ( ¹ The Water Framework Directive, Article 4(7), requires demonstration of disproportionate cost where the beneficial objectives served by the physical modification of a water body could be delivered by a significantly better environmental option that would minimise the risk of a deterioration in status.) The options should also include any reasonable proposals suggested by consultees.

3.10 Although it is good practice to start with a wide range of options for several different standards of defence, these can usually be reduced to a smaller range of standards and options for detailed analysis. For example, in economic terms, the solution with the lowest present-value cost will be the most cost-effective solution for any particular standard.

3.11 Alternative lines of defence should be considered, wherever feasible. Local protection works, or works which only protect particular areas of high value, should also be considered. These may include floor-raising and flood-proofing individual properties. In addition, remote protection should be examined; for example, by providing off-line storage/attenuation in the upper catchment.

3.12 An appropriate range of options should be considered. These should normally include:

• different standards of protection;
• alternative alignments;
• different approaches to solution of the problem.

Identifying the probabilities of failure

3.13 Changes in probability are central to the benefit-cost analysis. For flood prevention schemes, the 'do something' options reduce the probability of flooding. Commonly, the probability changes over time; for example, the likelihood that an existing flood embankment will fail may well increase with time. Climate change, development of catchments, and alterations in ground level may also change the probabilities of events of particular magnitudes. Chapter 6 provides guidance on the use of risk assessment in project appraisal, and the derivation of probability functions.

The nature of benefits and costs

3.14 Benefit-cost analysis is only concerned with changes in the total value of benefits and the total cost of the resources used. People will often adjust to a flood loss, and do so in a way that minimises their losses. If flooding closes a factory, production may be increased elsewhere. In such a case, the total national value remains the same. If the alteration simply varies the distribution of benefits and costs across the country, then no economic change occurs. Changes only in the distribution of consumption and resources are termed 'transfer payments' and should be excluded from the benefit-cost analysis (Annex A).

Boundaries of the benefit-cost analysis

3.15 Boundaries in space and time have to be drawn for any benefit-cost analysis. They should be set taking into account the consequences of the different options. It may not be possible to assess all such consequences, and a reasoned decision must therefore be made as to how far the process should be pursued. Options may have an influence outside the immediate area of the scheme but, in addition, their consequences may depend on factors outside the project boundary. For example, urbanisation of the upstream catchment could affect flood risk. The assumptions made about external conditions should be realistic, not simply convenient. Any significant impacts beyond the project boundary should be included.

The project domain

3.16 In some cases, the form of appraisal will depend on whether the solution is in fact a single scheme, or a series of independent projects. If the latter, a separate benefit-cost analysis should be carried out for each independent element. For example, if separate flood embankments are proposed for several different villages, the protection of each should be justified on its own. Where, however, it is possible to protect all the villages with a single scheme, such as a barrage, or flood storage reservoir, it is still necessary to consider the option of protecting each one individually. In this case, the aggregate costs and benefits of the best worthwhile individual protection schemes should be compared with those of the single scheme.

Phasing of a project

3.17 When a single scheme is phased, and there are no independent elements, the benefit-cost analysis should be carried out for the project as a single entity. However, a review of the justification of each stage of the phased works should also be undertaken.

3.18 As a scheme is progressed, the probability of a failure of one part of the scheme will change. It is this change in probability that should be used in the review of each stage of the works. Further, there are likely to be differences in the consequences of a failure, depending upon the particular works already completed.

Time span of the analysis

3.19 The appraisal period should reflect the physical life (with maintenance) of the longest-lived asset under consideration for a scheme. The presumption is that for most conventional schemes involving major earthworks, concrete or masonry structures, a 100-year timeframe will be appropriate. If, exceptionally, a shorter time horizon is used, terminal values may be applied to ensure that different options are evaluated on a comparable basis. These values should equate to the residual values of any assets, normally calculated using straight-line depreciation. Clearly all options should be evaluated over the same period.

Price basis

3.20 Benefit-cost analysis should be undertaken using real prices; that is, inflation is ignored where 'inflation' has the everyday meaning of the price of a resource increasing without its relative value also increasing. Commonly, the relative prices of the different streams of costs and benefits are assumed to be constant over time; this is generally a conservative practice. In reality, they may change over time.

3.21 Growth factors may be adopted to reflect predicted changes in relative prices or demand. However, if such factors are used for one stream of benefits or costs, they should logically be used for all streams. Since prices are relative, it follows that, over time, some will fall relative to others, in the same way that the real price of many electrical goods has fallen over the last 30 years. Therefore, any use of selective growth factors should be considered carefully.

• Inflation should be ignored in undertaking the analysis.
• Real prices should be used for all streams of benefits and costs.

What is a benefit and what is a cost?

Treatment of negative costs and negative benefits

3.22 In deriving a benefit-cost ratio, there is no universally agreed basis for classifying a particular item as either a positive cost, or a negative benefit (disbenefit), or vice versa. While the particular approach adopted will have no effect on the net present value of an option, it can have a significant effect on the benefit-cost ratio. To ensure a consistent approach between options and schemes, it is important to have a common rule. The following conventions should therefore be adopted.

• Any 'negative costs' should be regarded as benefits.
• Any 'negative benefits' should be regarded as costs.

3.23 The benefit arising from a flood prevention scheme is the net difference between total present value damages with and without the scheme - that is, the damage avoided in comparison with the 'do nothing' option. Any negative benefit, or disbenefit, arising from the project represents a loss to society, and should therefore be treated as a scheme cost. Conversely, resources which become available to society as a result of project implementation should be regarded as a benefit. For example, if an option restricts the views of residents, the loss of amenity should be counted as a cost; if widening or dredging a river yields gravel which can be sold, this is a benefit.

Stocks and flows

3.24 When identifying and valuing the different streams of benefits and costs, it is helpful to think in terms of 'stocks' and 'flows'. It is easy to make the mistake of including the same benefit or cost twice because of a failure to distinguish sufficiently between the stock of some asset and the flow of resources, or consumption, which that stock generates. Typically, stocks give rise to some flow of consumption or resource so that the current capital value of the stock is determined by the discounted value of the future benefits, which flow from it. In some cases the flow diminishes the value of the stock value (eg mining coal necessarily diminishes the stock of coal) whereas in other cases it does not (eg catching fish at below the rate of replacement).

3.25 An appraisal can include either the stock value of a resource or the sum of all the flows that it yields but not both. However, the market value of a stock might not always fully reflect the value of the flow of services which it provides since these can include unpriced public goods, such as the provision of opportunities for enjoyment of the countryside. In such cases allowance can be made for these additional flow values, suitably discounted.

3.26 In some cases it may be easier to estimate the change in the annual flow of benefits from the stock than the change in the capital value of the stock. For instance, there is no obvious market price for a riverside park that might be destroyed through the construction of a flood prevention scheme. An estimate of the value placed by users on visits made to that park might represent a better approach to valuing its loss (section 4.2).

Level of detail

3.27 The level of detail in a benefit-cost analysis should be proportional to the scale of the project. For projects involving a small amount of expenditure, a detailed benefit-cost analysis may not be economically justified. Decisions will therefore be required on the appropriate level of detail, the streams of benefits and costs to be included, and the amount to be spent on the analysis.

3.28 For example, where appropriate, strategic and prefeasibility studies should be undertaken using readily available data, since the most expensive part of a benefit-cost analysis is data collection and collation. A simplified analysis should not, however, be interpreted as one that lacks economic rigour. This may include the simple technique of benefit transfer, which involves taking values derived in one context and applying them elsewhere; for example, taking the values of enjoyment for a visit to one river, and using them to estimate the value of visits to another.

3.29 For the more detailed design studies, the benefit-cost analysis should be sufficiently extensive to show with reasonable confidence whether or not it is worth adopting any of the 'do something' options. Ideally, the first streams of benefits and costs to be included should be those contributing the greatest proportion of the total. While it is not always possible to determine in advance which will be the largest, there are 2 areas which should generally be considered first. These are:

• benefits and costs which accrue earliest in the lifetime of the scheme; and
• those which have the highest probability of occurrence.

3.30 The cost of evaluating a stream of benefits and costs depends on the precision required. It also varies markedly according to the stream category (Table 3.1). For recreational benefits or non-use values, it will cost as much to evaluate these for a small scheme (or area) as a large one, since the costs are largely fixed. The assessment cost also depends on the extent to which standard data can be used, such as the depth-damage data for residential properties. Further, if the flood surface and topography are complex, evaluation costs will increase.

Table 3.1 Relative costs of evaluation

Benefit or cost stream	Relative cost of assessment
Flood prevention scheme: protecting residential plus some small commercial or industrial properties	?
Flood prevention scheme: protecting some large commercial or industrial properties	? ?
Traffic disruption	? ?
Recreation benefits	? ? ? ?
Environmental assets: replacement cost method	? ?
Environmental assets: evaluation of non-use value	? ? ? ? ?

Key: The greater the number of ?, the greater the relative cost.

Post project evaluation

3.31 Post project evaluations demonstrate whether past investment has been worthwhile and has achieved its objectives. The exercise will be considerably eased if the fully documented original appraisal is readily available to evaluators. For example, it will be possible to compare actual performance against the sensitivity analysis in the benefit-cost appraisal. This will enable a check on the areas of greatest uncertainty, to see whether variations in practice are within predicted limits

3.32 In undertaking benefit-cost analysis as part of post project evaluation of flood prevention schemes it can be difficult to identify realised benefits, which are based on the avoidance of losses which would have occurred had the scheme not been implemented. Hence, it is difficult to determine whether the actual benefits are equal to those predicted. It will usually, therefore, be necessary to judge success on the accuracy of related predictions such as the costs of construction and maintenance, rates of environmental enhancement or measures of residual damage.