When Cost-Benefit Analysis Becomes Advocacy Instead of Analysis

The danger is not that cost-benefit analysis produces numbers. It is that numbers can make a predetermined choice look more objective than it really is.

Leaders need analytical tools because consequential decisions involve competing interests, uncertain consequences and scarce resources. Cost-benefit analysis can bring useful discipline: define alternatives, identify consequences, value what can reasonably be valued and make trade-offs more explicit.

But the same analytical authority can be used differently.

I Putu Sudiana's 2010 Australian National University discussion paper examined cost-benefit analysis in two Australian Government departments through a public-value lens. His study found a paradox. Public managers regarded CBA as conceptually useful, yet the observed practice could be pragmatic, with analysis used to support decisions rather than operate as a purely rational choice mechanism. The paper also raised concerns about transparency, capability, methodological difficulty and the risk of analysis being used to justify a predetermined decision.

This is historical, limited case-study evidence, not a universal description of public administration. Yet the governance problem extends well beyond government.

Any organisation can turn analysis into advocacy.

The Strategic Context

Executives rarely make decisions in neutral environments. A proposal usually has a history before the formal business case arrives.

A sponsor has invested reputation in it. A team has spent months developing it. A vendor may have shaped the solution. A public commitment may have been made. A budget window may be closing. A competitor's move may have created urgency.

These forces do not make the decision illegitimate. They make analytical governance necessary.

The 2006 Australian Government CBA Handbook itself recognises several limitations that matter here: false accuracy, self-serving analysis, quantification problems, complexity and equity concerns. It also emphasises that assumptions should be transparent and that analysis assists rather than replaces judgement.

The executive question is therefore not simply, “Was a CBA completed?”

It is, “Was the decision process designed so that the analysis could genuinely change the decision?”

What Leaders Commonly Misread

One mistake is equating analytical sophistication with analytical independence. A complex model can still be biased by which alternatives were included, which assumptions were chosen and which consequences were excluded.

Another is treating transparency as publication of the final number. Transparency means decision-makers can inspect the logic, assumptions, ranges, exclusions and treatment of conflicting evidence.

A third mistake is assuming bias must involve dishonesty. Most advocacy enters more subtly. Teams investigate their preferred option in detail while leaving alternatives generic. Benefits receive quantified models while disbenefits remain qualitative. Upside assumptions are described as “expected”; downside assumptions are described as “conservative”.

A fourth mistake is outsourcing judgement along with analysis. Sudiana's interviews raised the issue of limited internal capability and dependence on consultants. External expertise can be essential, but an organisation that cannot interrogate the model remains vulnerable. It may receive analysis without developing decision capacity.

Finally, leaders may assume that following the model is objective. It can be just as irresponsible to follow a result mechanically when the method does not capture all relevant value.

Reframing the Issue

CBA should be governed as a decision process, not merely a calculation.

That requires attention to four layers.

Problem framing: Who defined the problem and the objectives?

Alternative design: Were credible alternatives considered, including different scale, timing and non-investment responses?

Analytical construction: Which costs, benefits, time horizons and assumptions were included?

Decision accountability: Who made the final judgement, and how were departures from the analysis explained?

The quality of the final number is bounded by the quality of these choices.

Strategic Analysis: How Advocacy Enters the Model

Control the alternatives and you influence the answer

A preferred option can be made to look compelling when it is compared with weak alternatives.

For example, an organisation proposing a major new system might compare “full replacement” with “do nothing”, while ignoring process redesign, targeted modernisation, staged replacement or a managed-service option.

The CBA may be internally consistent and still answer a badly framed question.

This is why alternative generation should occur before the preferred solution becomes too entrenched.

Assumptions carry strategic judgement

Forecast growth, utilisation, labour savings, residual values, implementation time, benefit duration and discount rates can all materially influence an appraisal.

The 2006 Handbook's discussion of optimism bias is important because it treats favourable assumptions as a recurring analytical problem, not a rare ethical failure. Clear assumptions and, where appropriate, independent assessment make it harder for optimism to hide inside the model.

Governance should therefore focus on assumption ownership. Who proposed the estimate? What evidence supports it? How wide is the credible range? Who can challenge it?

Quantification can create asymmetry

What is easy to monetise may dominate what is strategically important.

Direct labour savings might be modelled precisely while safety, resilience, public trust, workforce capability or environmental impacts remain outside the main calculation. The quantified portion then acquires apparent authority simply because it is measurable.

The correct response is not to abandon quantification. It is to prevent the model boundary from becoming the decision boundary.

Capability determines whether challenge is real

Sudiana's case study linked the effectiveness of CBA to operational capacity. Interviewees reported limitations in internal capability, and the paper noted methodological difficulties such as valuation, discount rates and data availability.

An executive committee does not need every member to become an economist. It does need enough internal capability to ask whether the analysis is fit for purpose.

If no one inside the organisation can challenge the model, assurance becomes procedural rather than substantive.

Transparency changes incentives

Sudiana argued that limited transparency could weaken legitimacy and public trust. In enterprise settings, transparency has a similar internal function.

When assumptions, alternatives and decision rationales are visible to peers, boards, auditors or future reviewers, teams have stronger incentives to build defensible analysis. When only the final recommendation is visible, favourable framing is easier.

Transparency is therefore not merely disclosure. It is a control mechanism.

Decision Framework

A high-integrity investment process should pass seven governance tests.

One further test is particularly powerful:

Could this analysis produce a “no”?

If the process is structured so that every path leads to approval, it is not functioning as decision analysis.

From Strategy to Execution

Immediate action

For material decisions, separate the sponsor from the person who leads analytical challenge.

Require comparable treatment of credible alternatives.

Publish an assumption register with evidence quality, uncertainty range and named owner.

Ask the investment committee to record what evidence would cause it to reject or defer the preferred option.

Where the final decision departs from the model, record why. Political, strategic, ethical or operational judgement may legitimately outweigh a narrow economic result, but the departure should be explicit rather than hidden inside the numbers.

Medium-term capability building

Develop an internal appraisal capability even if modelling is outsourced. The organisation must be able to specify the question, review the method and challenge results.

Create peer review for high-value or high-irreversibility decisions.

Benchmark forecast assumptions against actual outcomes. Repeated bias should influence future assurance intensity.

Use standard disclosure requirements for alternatives, counterfactuals, distributional effects, assumptions and sensitivity.

Protect dissent. Analysts should be able to surface unfavourable results without being treated as obstacles to delivery.

Long-term strategic positioning

High-integrity decision systems create institutional memory.

Over time, the organisation can learn which benefit classes are systematically overstated, which costs are repeatedly omitted, which sponsors forecast well, which suppliers underperform and which uncertainties deserve early testing.

This converts appraisal from a one-off approval activity into an enterprise learning system.

It also strengthens trust. Boards and executive teams can move faster when they trust that bad news will surface before capital is irreversibly committed.

Signals to Monitor

Leaders should investigate when:

• one preferred option is modelled in detail and alternatives are superficial;
• analysis begins after the solution has been selected;
• all sensitivities preserve the original recommendation;
• consultants present the model but internal leaders cannot explain the assumptions;
• difficult-to-monetise effects are relegated to a footnote;
• a project's business case improves materially each time it approaches an approval threshold;
• the organisation rarely rejects projects after formal appraisal;
• benefits are revised upwards more often than downwards;
• negative findings disappear between analytical drafts and executive papers;
• no one is responsible for comparing forecast and actual outcomes.

These are governance signals, not proof of misconduct. They justify stronger challenge.

Questions for the Leadership Team

1. Could this analysis genuinely lead us to reject the proposal?
2. Who selected the alternatives, and which credible alternatives were excluded?
3. Which assumptions are matters of evidence and which are matters of judgement?
4. What important consequences sit outside the monetised model?
5. Do we have enough internal capability to challenge the analysts or consultants?
6. If we disagree with the CBA result, will we record the reason openly?
7. What have previous investment outcomes taught us about our forecasting bias?

Closing Perspective

Cost-benefit analysis earns its authority when it improves the quality of choice. It loses that authority when it becomes a sophisticated way to legitimise a decision that could no longer be changed.

The answer is not to distrust analysis. It is to govern it.

Protect alternative generation. Expose assumptions. Maintain internal challenge capability. Make non-monetised consequences visible. Document judgement rather than disguising it as mathematics.

A strong executive decision system does not ask numbers to remove responsibility from leaders. It uses numbers to make that responsibility harder to avoid.

Related article: Portfolio Governance Is a Decision-Rights System