The Five Most Common Gaps in a Green Book Business Case

Back to all posts

The Green Book is 130 pages. The supplementary guidance adds another 400. Every business case author references about 20 pages of it. The gaps that follow are predictable.

This isn't a knowledge problem. The people writing business cases for major programmes are experienced professionals. They know the Green Book exists. They've read the relevant sections. But knowing what the guidance says and producing evidence that satisfies an assessor are different things. The gaps below appear in programme after programme, not because teams are incompetent, but because the Green Book demands a level of documented rigour that most programme timelines don't accommodate.

These five gaps account for the majority of AMBER and RED findings in Green Book business case assessments. If you're preparing a Full Business Case for review, check these first.

Gap 1: Optimism bias applied without documented methodology

"We applied 24% optimism bias uplift to the capital cost estimate."

That's a number. It's not a methodology. And it's exactly the kind of statement that triggers a finding.

The Green Book supplementary guidance on optimism bias is specific about what's required. You need a reference class — which category of project are you benchmarking against? You need to document the empirical basis for the percentage — where did 24% come from? You need to show what adjustments you made for programme-specific factors. If your programme has characteristics that differ materially from the reference class, the uplift should reflect that, and the reasoning should be documented.

What assessors actually check is the audit trail. Can they trace from the final number back through documented reasoning to an empirical source? If the answer is "24% is what we've always used" or "our cost consultant recommended it," that's not a methodology. That's a habit.

The most common optimism bias finding isn't that the percentage is wrong. It's that there's no documented methodology behind it. The number might be perfectly reasonable — but without the workings, an assessor can't confirm that.

The fix is straightforward but time-consuming: document the reference class, cite the empirical source, record the programme-specific adjustments, and keep the working file as part of the business case evidence pack. Treat optimism bias like a calculation, not a judgement call.

Gap 2: Options appraisal that isn't genuine

The Green Book requires a long list of options assessed against spending objectives, narrowed to a shortlist through documented criteria, with each shortlisted option appraised on its merits. In practice, what assessors often find is a long list of six options, conveniently narrowed to two, with the preferred option clearly pre-selected before the appraisal began.

"Option 3 was discounted due to cost" is not an assessment against spending objectives. It's a sentence that tells the assessor someone decided the answer before they did the work.

A genuine options appraisal has several characteristics that are easy to check. Each long-list option is assessed against the same criteria. The criteria are derived from the spending objectives, not invented to favour a particular answer. The shortlisting rationale explains why each discounted option was removed, with reference to the assessment. And the shortlisted options include at least one that is genuinely different from the preferred option — not a straw man designed to lose.

The do-minimum option is particularly important. The Green Book requires it as a baseline. "Do nothing" and "do minimum" are different things. Do-minimum means the least intervention that addresses the problem — not zero intervention. Getting this wrong undermines the entire options comparison because the baseline is flawed.

Assessors spend time on the options appraisal because it's where confirmation bias is most visible. If your preferred option looks inevitable from page one, the appraisal hasn't done its job.

Gap 3: Management Case written as intentions, not evidence

The Management Case is the consistent weak spot in Green Book business cases. It's written last, it's written under time pressure, and it reads like a plan rather than a body of evidence.

"A robust monitoring and evaluation framework will be established during the delivery phase." That sentence appears in some form in nearly every business case that receives an AMBER or RED finding on the Management Case. It tells the assessor that the framework doesn't exist yet. At Full Business Case stage, that's a problem.

What assessors expect to see by Gate 3:

• A benefits register with baselines already measured, not "to be established"
• A monitoring and evaluation design that specifies what will be measured, how, and by whom
• Governance that is evidenced by minutes and decisions, not just described as a structure
• Named responsible owners for each benefit, each risk, each workstream
• A change control process that has been used, not just documented

The pattern is consistent: the Management Case describes what governance and assurance will look like, rather than providing evidence that it already works. Structure diagrams without board minutes. Risk frameworks without evidence of risk discussions. Assurance plans without evidence of assurance activity. The Management Case needs to demonstrate that the programme can be delivered, which means showing that the management arrangements are already functioning — not that they will be set up later.

Gap 4: Do-nothing counterfactual not properly defined

"Without the programme, the status quo continues."

That's not a counterfactual. That's an assumption that nothing changes if you don't intervene. The Green Book is explicit: the counterfactual must quantify what happens without the programme, including costs that will be incurred, risks that will materialise, and benefits that will be foregone.

A proper counterfactual is a scenario, not a sentence. It includes the cost of maintaining existing assets or services without the programme. It quantifies the risks of inaction — regulatory penalties, safety incidents, service degradation, missed policy objectives. It estimates the economic cost of foregone benefits. And it applies the same analytical rigour to the counterfactual that it applies to the preferred option.

Get the counterfactual wrong and everything downstream is compromised. The net present value calculation compares the preferred option against the counterfactual. If the counterfactual understates the cost of inaction, the programme's NPV is understated. If it ignores risks that would materialise without intervention, the benefit-cost ratio is wrong. The entire economic case is built on the difference between "with programme" and "without programme." If "without programme" is a placeholder, the difference is meaningless.

A weak counterfactual doesn't just affect one section. It undermines the entire economic case because every NPV and BCR calculation is measured against it. Assessors check the counterfactual early because if it's flawed, nothing that follows can be relied upon.

Gap 5: Benefits quantification that doesn't survive scrutiny

The economic case is where assessors spend the most time. The numbers either work or they don't, and the workings either exist or they don't.

The most common finding is monetised benefits without transparent assumptions. A benefit of £450 million over 30 years sounds significant. But how was it calculated? What unit values were used? Where did those values come from? What growth rates were applied, and what's the basis for those rates? If the assessor can't trace from the headline number to documented assumptions, the benefit is unverifiable.

Discount rates are another frequent issue. The Green Book specifies the social time preference rate — 3.5% for years 0-30, declining thereafter. But programmes regularly apply discount rates without documenting which rate they used or why. Health and environmental benefits have different discount rate guidance. Using the wrong rate, or not documenting which rate was applied, is a basic error that appears more often than it should.

Sensitivity analysis is the third common gap. The Green Book requires testing how results change when key assumptions vary. What assessors find is either no sensitivity analysis at all, or a perfunctory version that tests one variable by an arbitrary percentage. A credible sensitivity analysis identifies the assumptions that most affect the result, tests them across a realistic range, and presents the findings so that decision-makers understand where the uncertainty lies.

Distributional analysis — who benefits and who bears the costs — is increasingly scrutinised, particularly for programmes with significant regional or socioeconomic impacts. The Green Book's supplementary guidance on distributional effects is clear about when it's required. Programmes that skip it when it's relevant are inviting a finding.

The common thread

These five gaps account for the majority of AMBER and RED findings in Green Book assessments. They're not knowledge gaps — teams know the Green Book. They're evidence gaps. The analysis has been done, the thinking has happened, but the documented evidence trail is incomplete.

The fix is not better analysts or more guidance. It's systematic assessment against each Green Book requirement, with evidence cited to specific documents, pages, and paragraphs. The same way an assessor would check it — but before the assessor arrives.

Programmes that conduct this kind of structured self-assessment before submission consistently perform better at review. Not because their business cases are fundamentally stronger, but because they found and fixed the evidence gaps that would otherwise become findings.