Website of the Department of Health

4. In addition to the new Green Book itself, the website contains:

Transitional Arrangements

5. The new Treasury guidance came into effect from 1 April 2003. Treasury's transitional arrangements are that business cases for PFI schemes that reached Preliminary Invitation to Negotiate stage (or Invitation to Negotiate stage for those undertaking the accelerated route) by 31 March 2003 may continue under the previous guidance. Conventionally funded schemes that reached Invitation to Tender stage by 31 March 2003 may continue under the previous guidance. Projects that had not reached any of these stages must use the new guidance.

Key Changes

6. The key changes in the new Green Book are:

7. The following points should be noted in relation to each of the key changes announced in the Green Book.

Monetary Valuation of Benefits

8. The new Green Book places greater emphasis on the identification, quantification and monetary valuation of benefits in order to help assess more explicitly whether investment proposals are value-for-money. By encouraging investment appraisals to quantify most, if not all, benefits in monetary terms, the Green Book is placing more emphasis on undertaking Cost Benefit Analysis (CBA). Under CBA, investments should only be approved if the present value of benefits is higher than the present value of costs.

9. The Green Book puts greater emphasis on quantifying and valuing benefits with the caveat of where it is feasible and practical to do so (and the costs of valuing benefits are not incommensurate with the size of the project).

10. The Department of Health is undertaking further work and research on the feasibility and practicality of undertaking Cost Benefit Analysis for NHS investment. Until further guidance is issued by DH, business cases (including those that did not reach ITN stage by 1 April 2003) may continue to follow the guidance on assessing benefits in the current Capital Investment Manual, which focuses on weighted benefit scores (OBC) and cost-effectiveness analysis (FBC).

11. This should not preclude Trusts or PCTs undertaking work on valuing benefits on individual schemes - though it would be sensible for this to be shared with the relevant SHA and the Capital Investment Branch and Economic and Operational Research Branch of the Department of Health before formal incorporation into a business case.

Distributional Effects

12. It is important that the distributional implications of options, i.e. their differential impact on various groups in society, are considered during appraisal. The new Green Book requires any distributional effects to be explicitly stated and quantified as much as possible.

13. The new Green Book focuses on differing impacts by income group, but differing impacts according to age, gender, ethnic group, health or location need also to be stated and quantified wherever feasible.

14. The Green Book recognises that one pound of benefit is worth more to a poor person than a rich person. To account for this, it suggests multiplying the value of benefits by a scaling factor, with the largest scaling or weighting factor applied to benefits accruing to the least well off in society (measured by income groups or the closely related socio-economic groups). Guidance on values for the scaling factors are provided in the Green Book. Proposals that deliver greater net benefits to households or individuals in lower income groups are rated more favourably than those that benefit higher income groups.

15. The quantification of distributional effects and the application of scaling factors only has full force when benefits can be quantified, especially when they can be quantified in monetary terms.

16. As far as NHS business cases are concerned, as most NHS services are provided to either the very old or the very young, this will increase the attractiveness of NHS schemes. However, in many cases adjusting for distributional effects may make little substantive difference to the ranking of options presented in individual cases - options usually present alternative ways to provide healthcare for similar catchment populations. This implies the distributional effects for each of the options will be similar, and benefits will be scaled by the same factor.

17. Examples of where distributional effects may be more important include where there are different options for the location of services, which differ in accessibility by, for example, public transport. If options improve the quality of different services differentially, there may be a distributional effect as certain services are used more intensively by some groups than others (e.g. by different age groups). Where distributional effects are likely to be important, and it is not feasible to scale the benefits by income group, they should be included as one of the benefit criteria for the proposal in the weighted benefit scores.

18. The distributional adjustment will however be helpful to SHAs and CPAG when determining which schemes to prioritise.

19. In summary, for NHS and DH business cases, on distributional effects, the Green Book requires that:

Changes to the Discount Rate

20. A central feature of the new Green Book is the unbundling of the discount rate so that it relates to only one factor - the social time preference rate (which is society's preference for receiving benefits now rather than later, and equivalently, society's preference for bearing costs later rather than now). The discount rate is now set at 3.5% real. Other factors which have been implicitly bundled up in the old 6% figure are dealt with explicitly and separately. The most important of these is optimism bias. The new capital charge rate is also 3.5% real.

21. Treasury have introduced a differentiated discount rate for the long term (beyond 30 years). The discount rate to be used for years 0 to 30 in a project appraisal is 3.5% real, while for years 31 to 75 years the discount rate is 3.0% real (the discount rate declines further after 75 years, as set out in the Green Book).

22. The new discount rates must be applied in business cases, both OBCs and FBCs, for schemes for which the new Green Book guidance applies (see transitional arrangements section above).

23. The new capital charge rate of 3.5% applies from 1 April 2003. It is possible to have a Full Business Case applying the new capital charge rate but the old discount rate as the two are mutually exclusive in the appraisal - the former applies to affordability, the latter to value for money. This would be the case for schemes that reached ITN by 31 March 2003.

Optimism Bias

24. Appraisers have a demonstrated tendency to be overly optimistic when assessing projects. Costs and time-scales are underestimated and benefits are overestimated. This observation is more descriptive than judgemental as there are many causes of optimism bias.

25. To redress this tendency, the new Green Book requires that explicit adjustments are made to allow for optimism bias in all appraisals; that is estimates of costs and works duration are adjusted upwards to counteract the known tendency of underestimation. The new Green Book says the adjustments should be empirically based, using data from past projects or similar projects elsewhere (which may need to be adjusted for the unique characteristics of the project in hand). This places greater importance on Post Project Evaluation to provide a better empirical base for adjustments for optimism bias.

26. The new Green Book says the size of the optimism bias adjustments should be based on the best empirical evidence relevant to the stage of the appraisal. Adjusting for optimism bias should provide a better estimate earlier on of key project parameters, and the size of the adjustments for optimism bias should be subsequently reduced in later stages of the appraisal according to the increasing degree of confidence in the estimates, the extent of management of generic risk, and the extent of the work undertaken to identify and mitigate project specific risks. The new Green Book says, wherever possible, the cost estimates and the adjustments for optimism bias should be reviewed externally.

27. The new Green Book requires explicit adjustments for capital cost and works duration. On the basis of studies of actual cost increases between OBC and FBC for NHS schemes, the annex to this note sets out indicative optimism bias uplifts for NHS large build schemes at OBC stage. Cases need to consider also to what extent the optimism bias has been mitigated (see annex) so that a lower percentage uplift is instead applicable. The reasons for the extent of mitigation of optimism bias need to be clearly stated and justified in business cases.

28. The Green Book notes optimism bias can arise also in relation to operating costs and under delivery of benefits, but states that if there is no evidence available to support adjustments to operating costs or benefits' shortfalls, appraisers should use sensitivity analysis.

29. The annex provides more information on the application of optimism bias to NHS business cases, which is now required for business cases to which the new Green Book guidance applies (see transitional arrangements above).

Taxation

30. The new Green Book places more emphasis on tax issues, though noting that, generally, it is relatively rare that adjustments for taxation between options are required, because similar tax regimes usually apply to different options. The Green Book, however, says also that where the tax regimes applying to different options vary substantially, this should not be allowed to distort option choice. In such cases, it is important to adjust for any major differences between options in the incidence of tax arising from different contractual arrangements. In particular, where publicly financed options are compared to PFI options, the new Green Book expects the tax liabilities to differ and requires that an adjustment for this difference in taxation is made in the business case.

31. Treasury commissioned KPMG to quantify the size of any tax liabilities which might result from PFI compared with traditional public procurement. KPMG reviewed various taxes, including stamp duty, VAT, employment taxes, business rates, and concluded that only a SPV's additional corporation tax liabilities are relevant when considering differential tax regimes between PSC/PFI options. Other taxes apply equally to PSC/PFI options.

32. The Treasury's guidance on estimating the size of this differential tax liability adjustment for PFI versus PSC comparisons is based on advice provided by KPMG and is contained in 'Supplementary guidance on adjusting for taxation in PFI vs PSC comparisons.' This guidance from Treasury needs to be consulted to calculate the size of the adjustment to be applied for each scheme, which will vary between schemes.

33. In the Treasury's guidance the size of the tax adjustment has a starting value of 2% which is the bottom of the likely range and represents a tax efficient project with relatively low capital expenditure and low risk. The size of the adjustment is increased over and above the starting value of 2% by set amounts according to how the scheme scores against each of four criteria:

34. Treasury expect the size of the adjustment to be in the range 2%-10% for the majority of projects across the whole of government, and possibly more for a minority.

35. The tax adjustment is included in the value for money analysis in the business case. The size of the tax adjustment in £s is added to the risk adjusted PSC. The size of the tax adjustment in £s is calculated by taking the relevant percentage (from using the Treasury's guidance) of the appropriate base, where the base is the risk adjusted present value of the costs of the public sector comparator that would otherwise transfer to the PFI option. Other PSC costs that would remain with the public sector under a PFI option (and therefore not subject to corporation tax) need to be deducted to reach the correct base number (e.g the cost of clinical services and retained risk need to be deducted).

36. The size of the tax adjustment factor is expected to be accurate to +/- 3 percentage points in most cases, which should be used as the upper and lower bounds for sensitivity analysis.

Publicly funded schemes and leases (non-PFI)

37. The new Green Book methodology applies also to publicly funded schemes - e.g. where it has always been accepted from their outselt that PFI would not be appropriate. Taxation adjustments are not relevant in the appraisal of publicly funded schemes.

38. The new arrangment s apply also to buisness cases where leases are envisaged (e.g. for equipment, vehicles or property). In some of these there may be less scope for optimism bias or taxation adjustment. However, risk transfer and benefits (whether quantified or simply scored and weighted) may well be material factors, which also need to be taken into account in assessing leasing options.

Further Guidance

39. DH will be updateing the Generic Economic Model (GEM). The guidance in this note is likely to be reviewed regularly and comments to assist in this are welcome. Questions or comments on this guidance should be sent to:

Michael Chaplin
Economics and Operational Research
Department of Health
Room 1N33
Leeds
LS2 7UE
Telephone: 0113 254 5571

• Email Michael Chaplin: (opens new window)

ANNEX: OPTIMISM BIAS

1. The new Green Book requires that explicit adjustments are made to allow for optimism bias in all appraisals. Optimism bias provides, in particular, a measure of the relative increase in costs from those estimated in the Business Case and the actual outturn costs.

2. Treasury argue that adjusting for optimism bias should provide better estimates earlier on in the development of a business case of key project parameters. Treasury require that the adjustments should be empirically based, using data from past projects or similar projects elsewhere (which may need to be adjusted for the unique characteristics of the project in hand). The size of the adjustments for optimism bias should be reduced in later stages of the appraisal according to the extent of confidence in the estimates, the extent of management of generic risk, and the extent of work undertaken to identify and mitigate project specific risks.

3. The Treasury guidance says also that wherever possible, the cost estimates and the adjustments for optimism bias should be reviewed externally and cites using Gateway reviews for large projects and internal audit reviews of smaller projects for this purpose.

4. In the absence of Department specific data on which to base the adjustments, Treasury recommend using the percentages for optimism bias in the Mott MacDonald report set out in the table below. The table covers only capital expenditure and works duration - it does not cover operating costs or benefits as the Mott MacDonald report had insufficient data on these. Treasury recommends that if there is no evidence to support adjustments to operating costs or benefits' shortfalls, appraisers should use sensitivity analysis.

Table 1: Optimism bias adjustments from the Mott MacDonald report

+Mott MacDonald's definitions of their project types are: Standard buildings are those which involve the construction of buildings not requiring special design considerations, including most general hospitals. Non-standard buildings are those which involve the construction of buildings requiring special design considerations due to space constraints, complicated site characteristics, specialist innovative buildings or unusual output specifications. Equipment/development covers projects concerned with the development of equipment and/or development of software and systems (i.e. manufactured equipment, Information and Communication Technology development projects) or leading edge projects. Outsourcing projects are concerned with the provision of hard and soft facilities management services (i.e. ICT services, facilities management or maintenance projects).
*The optimism bias for outsourcing projects is measured for operating expenditure.

5. The upper bound is said to be applicable at OBC stage and the lower bound, through improvements in the accuracy of the costing and management of risks, is applicable at FBC stage. The upper and lower bounds are indicative, and they do not represent the highest and lowest possible values for optimism bias.

6. For NHS projects, explicit adjustments for optimism bias for initial capital costs and works duration are required in business cases.

7. For NHS projects no data are available on adjustments to operating costs for optimism bias. Following the Treasury recommendation, sensitivity analysis should be used to show the range of outcomes for operating costs under different assumptions. Treasury's supplementary guidance on optimism bias recommends in particular calculating switching values. For NHS projects, the most relevant switching values at OBC stage may be how much operating costs can increase if the proposal is to remain affordable, and how much operating costs need to change to switch the ranking of options. At FBC stage for PFI projects, the switching value likely to be of most interest is how much PSC operating costs can change for the PFI proposal to remain value for money.

8. The remainder of this annex focuses on initial capital costs for construction cases only. Further guidance will be issued on IM&T projects.

9. The key issues in applying optimism bias for initial capital costs in NHS construction business cases are:

The 'upper bound'

10. An examination of a sample of past business cases and of the returns to the Health Select Committee suggest the following percentage increases in capital costs between OBC and FBC, after stripping out increases due to inflation as measured by the MIPS index:

11. These figures should be used for optimism bias and used to uplift capital costs at the OBC stage, unless there is good cause to vary them because of the specific characteristics of the project and/or mitigation (see section on mitigation below). A particularly difficult site is likely to require a higher level of optimism bias while a very straight forward site is likely to require a lower level of optimism bias. The Department of Health intends to issue further guidance on the related issues of varying the size of the upper bound and mitigation.

12. The percentage uplift for optimism bias is over and above the sum allowed for contingencies on line 9 of the OB/FB forms. The addition for optimism bias should be calculated by applying the percentage uplift for optimism bias to initial capital costs that exclude the allowance for contingencies.

13. Little information is available on cost increases between SOC and FBC stage. The above figures for optimism bias at OBC stage of 30%- 40% should also be used at SOC stage unless there is reason to increase them because of, for example, particular uncertainty about the scope of the project.

14. In terms of works duration, NHS Estates data suggest outturn works durations are some 15%-20% longer than anticipated. These figures should be used at OBC stage, unless there is good cause to vary them because of the specific characteristics of the project and/or mitigation (see section on mitigation below). Works duration here literally means how much longer it takes to complete the works. Longer durations often raise costs, and any such higher costs should already be incorporated into the optimism bias for costs.

15. The levels of optimism bias and the reasons for the chosen levels must be recorded in business cases for both capital expenditure and works duration.

16. If the costs of provision under PFI are estimated at OBC stage, the above figures for 30% - 40% for the initial level of optimism bias should also be applied to the initial capital cost component of the estimated PFI costs. The 30% - 40% range reflects what has historically happened since OBC stage, due to changes in scope and costs omitted or underestimated at OBC stage. Such increases are likely also to affect the PFI costs. One reason for applying a different level of optimism bias at OBC stage for PFI provision may be if the estimated PFI cost includes risk transfer, which is a form of mitigation (see section on mitigation below).

The 'lower bound'

17. At the time of the FBC, through improved accuracy of costing, management of risk (including the valuation of project specific risks) and mitigation, Treasury expect the adjustments for optimism bias to be around the lower bounds shown in table 1.

18. The data available from NHS Estates' Time and Cost Performance Indicators suggest that for publicly delivered projects:

19. The data available do not identify the causes of these overruns. The reasons will be a combination of optimism bias that persists at FBC stage and project specific risks that actually materialise.

20. In light of these data, it is expected for the PSC in NHS business cases, that at FBC stage, the risk adjustments and remaining optimism bias together sum to similar figures to those given in paragraph 18. The figures in paragraph 18 are indicative averages and actual levels may differ between projects according to their specific characteristics. The level of optimism bias and risk adjustments need to be clearly set out in the FBC.

21. As an example, one project may apply 40% optimism bias at OBC stage to the capital costs. By FBC stage, increases in the scope of the project and emerging difficulties with the site have increased estimated costs by 20%. Project specific risks have been identified and quantified by FBC stage summing to 15% of capital costs. 2% remaining optimism bias is included at FBC stage. Of the original 40% optimism bias at OBC stage: 20% points has been incurred with the increase in costs to FBC stage, 18% points is represented by the project specific risks (=15% of the higher costs at FBC), and 2% points is the remaining optimism bias.

22. At the FBC stage, the remaining optimism bias to be applied to the PFI costs should be very low and encapsulated in the estimates of the risk retained in the public sector. The evidence suggests there is not a need to adjust the PFI unitary payments for optimism bias as these do not vary after financial close.

23. The evidence from NHS Estates' Time and Cost Performance Indicators suggests percentage time overruns for publicly delivered projects are around 15% to 20%. As for capital costs, this is the combined effect of remaining optimism bias and project specific risks. It is expected that at FBC stage, the risk adjustments and remaining optimism bias together sum to around these figures for overruns on works duration. Data more specific to a particular project may be used where available. The remaining optimism bias for works duration should be applied to the PSC only. Any costs associated with longer durations should already be incorporated into the risk/optimism bias for costs.

Mitigation of optimism bias

24. The upper bound for optimism bias is an initial assessment of optimism bias. Treasury's Supplementary Guidance on the treatment of optimism bias sets out how this level of optimism bias is reduced according to the management processes in place and characteristics of the project. Mitigation is a way of expressing how firm are the costings for a project.

25. Treasury's Supplementary Guidance on optimism bias sets out a list of 'contributory factors' that have been shown across the public sector as a whole to lead to cost and duration overruns. It sets out also the percentage contribution of each factor to the upper bound for optimism bias. The list of contributory factors and the percentage contribution of each factor to the upper bound for optimism bias may be changed in individual NHS business cases, where appropriate, to give a more relevant list of contributory factors and their relative importance for a particular scheme. The reasons for any such changes must be set out and justified in business cases.

26. Treasury's guidance requires an assessment of the extent to which each contributory factor has been mitigated through good project management, including provision of reliable costings, management of risk, and identification and quantification of project specific risks.

27. The extent of mitigation of each contributory factor is a matter of judgement. This judgement leads to an estimated scale of mitigation for each contributory factor, with mitigation values lying between 0.0 and 1.0. A mitigation value of 0.0 for a contributory factor means the factor has not been mitigated at all, and a value of 1.0 means the contributory factor has been fully mitigated. Values of between 0.0 and 1.0 represent partial mitigation. After applying the mitigation values to each contributory factor, the new overall level of optimism bias applicable is calculated. An example is given in Treasury's supplementary guidance on optimism bias. The overall level optimism bias applicable is likely to change at various points in the development of a business case, including between OBC and FBC.

28. Treasury's guidance sets out different sets of contributory factors for standard buildings and non-standard buildings, and also different sets for works duration and capital expenditure.

29. For NHS projects, the initial levels of optimism at OBC normally expected are set out above, which may need to be varied due to projects' specific characteristics. Lower levels may instead apply at OBC stage as a consequence of the contributory factors having been partially mitigated. By FBC stage, it would be expected that any remaining optimism bias is very low, partly due to its replacement by the quantification and valuation of specific risks. It needs to be shown at FBC stage that nearly all the factors contributing to optimism bias have been fully mitigated, or accounted for in separately quantified risks or in risks transferred to the private sector.

30. Factors relevant in assessing the level of mitigation to be applied may include the following example, which are not exhaustive or comprehensive:

31. Business cases must clearly set out the initial 'upper bound' level of optimism bias used, and set out and justify any differences from the indicative levels given above for the 'upper bounds', any changes made to the list of contributory factors and their percentage contributions, and the rate of mitigation applied to each contributory factor.

Further work

32. This guidance on the levels of optimism bias is preliminary. Further work and research is needed to identify the contributory factors to optimism bias that are specific to the NHS and appropriate levels of mitigation.