We live in an era of expensive, overrun, and overbudgeted ‘megaprojects’ says Daniel Baker. But what are the solutions to bringing the curtain down on these?
A rational person would be forgiven for thinking the word ‘megaproject’ cursed. That is, if the Oxford Global Projects Database (OGPD) – which holds information on nearly 12,000 different megaprojects – is to be believed.
According to the OGPD’s own data, nearly every category of megaprojects failed to meet planned budgets, and to be completed on time. This includes nearly 40%of rail projects and 61% of projects in the aerospace sector. Even more distressingly, only a tiny 0.5% of megaprojects provided all of the ‘extended benefits’ they were initially planned to deliver.
Infamous megaprojects and risk
Two notorious megaprojects currently ongoing in the UK are the rail projects Crossrail and HS2. The former was due to open in 2018, but will not open until 2021 at the earliest. As a result, Crossrail’s contingency had to be more than doubled — from £3bn to more than £6.2bn. This is sobering reading for anyone already looking at HS2’s first-phase budget of £57bn. If HS2 ends up requiring a contingency budget anything like Crossrail’s, then that could amount to an additional £20bn — or in other words, more money than was initially budgeted for the Crossrail project altogether. Latest estimates are that HS2 could be as high as double the original budget and double the original timescales.
But even these rail projects pale in comparison to some of the other megaprojects of history. The Suez Canal, when it was eventually opened, cost nearly 2,000 times the initial estimates. The Holyrood parliament building in Scotland cost £400m, after an initial budget of £40m, and the Sydney Opera House overran its budget by 1,400%.
Why do megaprojects fail?
The problems that tend to afflict megaprojects are difficult to anticipate because the megaprojects themselves are by nature staggeringly complicated. At least, that is according to OGPD’s Bent Flyvbjerg. It can be bewildering for a project manager to even know where to begin.
Traditionally, megaprojects are broken down into lots of smaller, more manageable projects and then each one is run through a risk management process known as Quantitative Schedule Risk Assessment (QSRA) that uses Monte Carlo analysis. The Monte Carlo simulations provide project managers with the ability to apply what is known as ‘three-point estimates’ to each activity in their plans. It all sounds good in theory, but according to Flyvbjerg innate human biases, such as ‘uniqueness bias’, undermine the Monte Carlo simulations.
Biases, black swans, and a lack of communication
Uniqueness bias is the assumption that, because a megaproject is so big and so complex, it must be unique. Project managers then tend to assume that they cannot learn anything from projects before them. But, Crossrail, for example, a large railway programme in London, is not all that different from previous railway work that has been done in major European cities – such as the Metro extensions in Paris.
Conversely, three point estimating in Monte Carlo analysis assumes that the middle or ‘most likely’ estimate is accurate when in fact it will be using the durations in the plan that have already been ‘squeezed’ to meet the required milestones.
In addition, mega-projects are more prone to Black Swan events. These are events that are extremely low probability but could have devastating impacts. Because megaprojects take so long to complete, they are more likely to invite chance to dish up a Black Swan event.
At the most fundamental level, there are the inevitable risks that are introduced by lack of effective communication. The sheer scale of megaprojects and their team sizes means that this is going to be more prevalent than ever.
Taking the risk out of megaprojects
The first stage of getting-real with mega-project estimates is getting away from the slavish belief in QSRA estimating. Techniques such as strategic delivery assurance and strategic target analysis in particular, are proven ways of getting away from the squeezed-plan syndrome and providing meaningful estimates.
In addition, innate human uniqueness biases will need to be challenged. This is exactly what psychologists Daniel Kahneman and Amo Tversky are trying to do with their ‘reference class forecasting’ database, which works by collecting and analysing past megaprojects, along with the experiences of the people working on them. Using complex AI-algorithms, data analysts working on future megaprojects will be able to directly learn from the real-world case studies of the past.
But in order for any database like this one to be thorough enough, the construction industry will have to work closely together. The runners of every project would be required to record every minute detail: including mistakes made, inspection data, any compensation events, and then insert the information into the database. This would require a kind of multilateral culture of fostering data ‘trust’ which won’t be easy.
In fact, many organisations have already cited their misgivings for such a database, as some of the information required might be commercially sensitive.
The debate goes on, but one possible solution might be a sort of federated ownership and access, which only provides access via third-parties and not rival construction companies. Building a culture of data trust will not be easy, but if it works, could finally bring down the curtain on an era of expensive, overrun, and over-budgeted megaprojects.
Daniel Baker is Risk Analyst at De-Risk, a strategic programme risk management company based in Surrey, UK.
