There is a very old story about an accounting student who, before graduation, must sit an examination in front of the board of accountants. The student has performed brilliantly in their exam so far and is asked a final question by the chief examiner, “what is two plus two?” The accounting student immediately responds with the correct answer, “what do you want it to be?”.
The architects of business cases and cost benefit analyses (or benefit cost analyses) often apply the same philosophy. That is, they will provide an answer to their customer, sponsor, or supervisor that is a fantasy, simply so that they can give their customer what they want. This phenomenon is rife in the public sector where businesses cases regularly overstate benefits, ignore key risks and dis-benefits, disregard credible competing options, cherry pick data, and adopt creative discounting methods to get the desirable answer as opposed the correct answer. In this blog, we explore techniques to manipulate business cases and economic analyses to get the answer we want.
When presenting options to customers, management consultants around the world are familiar with the strategy of providing only three options. The first option is the cheap and nasty alternative that provides negligible value (sometimes labelled the ‘do nothing’ scenario). This option is immediately rejected. The second option is hideously expensive and has an implementation timeframe of decades. This option will also face immediate rejection. The third option, of course, is the option that the customer or sponsors want selected. This is why framing the business case is so important. For example, if the public sector outcome is to improve traffic congestion, cater for urban sprawl, or improve access to amenities for citizens then there is a plethora of potential solutions, not just one megaproject. Nonetheless, the solution space is often constrained to a single popularist option, or more cynically, the option that creates jobs in marginal electorates. If the sponsor or responsible minister has a specific solution in mind for political or ideological grounds, then we need to accept that the answer has already been selected. This answer may be very far away from the best answer, nonetheless we can craft business cases to justify this answer by simply framing the business case to a subset of solutions that exclude all other credible options. This is especially important if those other options have higher benefit cost ratios or higher net present values. This is relatively easy to achieve by simply not mentioning the other competing options, claim that these other options are inconsistent with government policy, or any other contrived excuse. This will allow you to justify the selection of almost any course of action you want.
In some circumstances you may be forced to explore competing options that could jeopardise the selection of your preferred solution. This is where cherry picking provides opportunities to manipulate the process. Every business case incorporates assumptions about future states, risks, issues, and opportunities. The key is to select information from datasets that increase the benefit cost ratio for your desired option and decrease it for all competing options. For example, if your preferred option creates substantial future benefits with reduced carbon pollution then it is imperative that you select a social cost of carbon at the upper extreme using the Intergovernmental Panel on Climate Change (IPCC) price of $1,472 per tonne of Carbon as illustrated in Table 1 below. Conversely, if your preferred option is carbon emissions intensive, then you should select the lowest figure, such as the current market price of carbon.
Table 1 Carbon Values under IPCC Calculations (2010 prices).
‘When manipulating business cases, it is vital to cherry pick data to bolster your preferred option.’
Opportunities exist to cherry pick data in many other areas of your economic analysis such as travel time savings, cost of lives saved, construction disruption costs etc. This provides almost limitless opportunities to get the answer you want.
Discount Rate Manipulation
If you are fortunate enough to be in a jurisdiction where treasury does not specify a discount rate to be used in your economic analysis, then opportunities abound to exploit the numbers to get the answer you want. To illustrate, consider a project that will generate economic benefits of $100 million per annum over a thirty year period. The following two discount rates yield very different present values:
Option 1. 3% discount rate – Present Value = $1.96 Billion
Option 2. 7% discount rate – Present Value = $1.24 Billion
If your desired project has benefits that span decades, then you should aim to use the lowest discount rate that is plausible. Conversely, if you have costs or disbenefits that eventuate into the far future, then you should select the highest discount rate you can get away with. If you are forced to use a discount rate provided by treasury, then there is always an option to use nominal prices and ‘accidently’ use a real discount rate. This will overstate the benefits of your economic analysis considerably. With luck, such errors will not be detected until after your project is approved and considerable costs have been expended.
Large, complex projects involve high risks and significant uncertainty. Business cases for such projects will be fuzzy to say the least. Despite these limitations, we should aim to create the impression that the business case is highly credible, robust, and precise. Augustine’s, 35th law embodies this approach:
“The weaker the data available upon which to base one’s conclusion, the greater the precision which should be quoted in order to give the data authenticity.”
Consequently, we should present data to as many significant figures as possible and avoid any reference to confidence levels. Large complex projects typically have cost and schedule probability distributions that resemble beta distributions with very long tails (large positive skewness). This means that the 90th percentile cost and schedule estimates will be substantially greater than the expected 50th percentile values. Such figures will terrify those accountable for funding such projects and therefore the worst case estimates should be omitted from our business case.
Wider Economic Effects
Wider economic effects are very difficult to estimate and should not normally be included in benefit cost ratios. Nonetheless, some jurisdictions do allow the inclusion of wider economic effects as part of an ‘adjusted benefit cost ratio.’ This allows us to manipulate our business cases further. Make sure your wider economic impact assessment overstates the benefits and ensure that no wider economic dis-benefits are included. Broadening the scope for the economic analysis also provides opportunities to bolster your business case, especially when it comes to jobs creation opportunities. Where an economy is operating at full employment, job creation claims in business cases are often bogus. A project that engages thousands of employees means that those resources cannot be used elsewhere.  In other words, your project is merely moving jobs from one area to another with no net benefit. Do not be dissuaded from such logic; however, as politicians salivate at the prospect of spruiking their economic credentials and their perceived ability to create jobs. Make sure your business case has job creation front and centre even though such claims are spurious.
If your best efforts to ‘cook the books’ and manipulate the numbers to get a benefit cost ratio greater than one does not eventuate, then all is not lost. Simply add the costs and benefits of your project to a related project which has a high benefit cost ratio. By blending multiple project costs and benefits within a portfolio, you can manipulate the numbers so that in aggregate a programme or portfolio of several projects has a combined benefit cost ratio greater than one. Whilst this appears to be an outrageous approach that no sensible person would pursue, this is precisely what the Australian Capital Territory Government did in their Stage 2A light rail business case. Despite this individual project having a benefit cost ratio of only 0.6, the project was blended with the previously completed stage to create the illusion of a positive net present value. Audaciously, the government presented the benefit cost ratio (BCR) of this single project as 1.2. As observed in a subsequent audit report for this business case, such an approach is misleading:
‘blended BCR is novel; no example of blending the result of a past investment with a future investment is known’ ‘[this approach] has no relevance to the [Light Rail Stage 2a] investment decision because the Stage 1 costs are ‘sunk’, i.e. cannot be recovered.’
With luck, your business case may not be scrutinised by auditors and blending your benefit cost ratio with another project may go unnoticed.
Transparency, Scrutiny and How to Avoid It
Flyvbjerg and Bester advocate for independent auditing of cost benefit forecasts. Some jurisdictions recommend business cases be made publicly available. Having your business case critiqued could undo all the hard work you put into fudging the numbers and you should therefore do everything imaginable to avoid such scrutiny. This can be achieved by contaminating your business case with as much commercially sensitive material as possible (ideally in every figure and paragraph). This will mean that any redacted version of your business case is meaningless and cannot be effectively verified if it is made publicly available. This may not thwart independent auditors but will hopefully slow down the review process until it is too late.
The techniques available to manipulate business cases are many and varied. We can cherry pick data, include serious errors of omission, and make many assumptions that allow us to justify almost any course of action. Those readers with a moral compass would not stoop to such skulduggery, rather you will use the strategies identified in this blog to recognise where business cases are being manipulated and hopefully put an end to such activities. Bogus business cases result in billions of taxpayers’ dollars wasted and lost opportunities. This situation should not be tolerated
Though we have focussed on public sector business cases, there is also ample opportunity to cook the books in the private sector, especially with taxation treatment. This brings us back to Hitchhikers Guide to the Galaxy canon, wherein Disaster Area’s chief research accountant’s thesis, General and Special Theories of Tax Returns, it is shown that, “the whole fabric of the space-time continuum is not merely curved, it is in fact totally bent.”
 Adapted from IPCC (2018) figure 2.26 – Dominique Bureau, Alain Quinet, and Katheline Schubert “Benefit-Cost Analysis for Climate Action” Journal of Benefit-Cost Analysis Vol 12 issue 3, 26 November 2021.
 N. Augustine Augustine’s Laws (1986).
 See e.g UK Department of Transport ‘TAG UNIT A2.1 Wider Economic Impacts Appraisal’ (2019)
 Queensland Government “Cost Benefit Analysis Guide – Business Case Development Framework” (2021) p 17.
 ACT Government- Major Projects Canberra ‘City to Woden Light Rail: Stage 2a City to Commonwealth Park Business Case’ (2019) p 123 https://www.transport.act.gov.au/__data/assets/pdf_file/0003/1412634/Stage-2A-Light-Rail-Business-Case-redacted.pdf
 Auditor General’s Report CANBERRA LIGHT RAIL STAGE 2A: ECONOMIC ANALYSIS REPORT NO. 8 /2021, p37 https://www.audit.act.gov.au/__data/assets/pdf_file/0005/1859630/Report-No.8-of-2021-Canberra-Light-Rail-Stage-2A-Economic-Analysis.pdf
 Flyvbjerg, Bent and Dirk W. Bester, “The Cost-Benefit Fallacy: Why Cost-Benefit Analysis Is Broken and How to Fix It,” Journal of Benefit-Cost Analysis, October 2021, pp. 1-25.
 Infrastructure Australia ‘Delivering Outcomes – A roadmap to improve infrastructure industry productivity and innovation’ (May 2022) [Recommendation 4.1.1.] p14 https://www.infrastructureaustralia.gov.au/sites/default/files/2022-03/IA%20Delivering%20Outcomes%20March%202022%20Final_1.pdf
 Douglas Adams, The Restaurant at the End of the Universe (1980).