"What You Should Know About Megaprojects and Why: An Overview", Flyvbjerg 2014 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2424835

The McKinsey Global Institute (2013) estimates global infrastructure spending will be US$3.4 trillion per year between 2013 and 2030, or approximately 4% of the total global gross domestic product, mainly delivered as large-scale projects. The Economist (2008) similarly estimated infrastructure spending in emerging economies at US$2.2 trillion annually for the period between 2009 and 2018. To illustrate the accelerated pace at which spending is taking place, consider that in the five years between 2004 and 2008, China spent more on infrastructure in real terms than during the entire 20th century, which is an increase in spending rate of a factor of 20. Similarly, between 2005 and 2008, China built as many kilometers of high-speed rail as Europe did in two decades; Europe was extraordinarily busy building this type of infrastructure during this period as well. Not at any time in the history of mankind has infrastructure spending been this high, measured as a share of world GDP, according to The Economist, (2008), who calls it “the biggest investment boom in history.” And that’s just for infrastructure.
If we include the many other fields in which megaprojects are a main delivery model—oil and gas, mining, aerospace, defense, ICT, supply chains, mega events, and so forth—then a conservative estimate for the global megaproject market is between US$6 and US$9 trillion per year, or approximately 8% of the total global gross domestic product. To put this into perspective, consider this is the equivalent of spending five to eight times the accumulated U.S. debt to China, every year. That’s big business by any definition of the term. Moreover, megaprojects have proved remarkably recession proof. In fact, the downturn from 2008 has helped the megaprojects business grow further by showering stimulus spending on everything from transportation infrastructure to ICT. From being a fringe activity— albeit a spectacular one—mainly reserved for rich, developed nations, megaprojects have recently transformed into a global multi-trillion-dollar business that affects all aspects of our lives, from our electricity bill to how we shop, what we do on the Internet to how we commute.

The following characteristics of megaprojects are typically overlooked or glossed over when the four sublimes are at play and the megaproject format is chosen for the delivery of large-scale ventures:
1. Megaprojects are inherently risky due to long planning horizons and complex interfaces (Flyvbjerg, 2006).
2. Often, projects are led by planners and managers without deep domain experience who keep changing throughout the long project cycles that apply to megaprojects, leaving leadership weak.
3. Decision making, planning, and management are typically multiactor processes involving multiple stakeholders, both public and private, with conflicting interests (Aaltonen & Kujala, 2010).
4. Technology and designs are often non-standard, leading to “uniqueness bias” among planners and managers, who tend to see their projects as singular, which impedes learning from other projects.3 Uniqueness bias is here defined as the tendency of planners and managers to see their projects as singular. This particular bias stems from the fact that new projects often use non-standard technologies and designs, leading managers to think their project is more different from other projects than it actually is. Uniqueness bias impedes managers’ learning, because they think they have nothing to learn from other projects because their own project is unique. This lack of learning may explain why managers who see their projects as unique perform significantly worse than other managers (Budzier & Flyvbjerg, 2013). Project managers who think their project is unique are therefore a liability for their project and organization. For megaprojects this would be a mega-liability.
5. Frequently there is overcommitment to a certain project concept at an early stage, resulting in “lock-in” or “capture,” leaving analyses of alternatives weak or absent, and leading to escalated commitment in later stages. “Fail fast” does not apply; “fail slow” does (Cantarelli, Flyvbjerg, & Rothengatter, 2010; Ross & Staw, 1993; Drummond, 1998).
6. Due to the large sums of money involved, principal-agent problems and rent-seeking behavior are common, as is optimism bias (Eisenhardt, 1989; Stiglitz, 1989; Flyvbjerg, Garbuio, & Lovallo, 2009).
7. The project scope or ambition level will typically change significantly over time.
8. Delivery is a high-risk, stochastic activity, with overexposure to socalled “black swans”; i.e., extreme events with massively negative outcomes (Taleb, 2010). Managers tend to ignore this, treating projects as if they exist largely in a deterministic Newtonian world of cause, effect, and control.
9. Statistical evidence shows that such complexity and unplanned events are often unaccounted for, leaving budget and time contingencies inadequate.
10. As a consequence, misinformation about costs, schedules, benefits, and risks is the norm throughout project development and the decisionmaking process. The result is cost overruns, delays, and benefit shortfalls that undermine project viability during project implementation and operations.
In the next section, we will see just how big and frequent such cost overruns, delays, and benefit shortfalls are. The Iron Law of Megaprojects Performance data for megaprojects speak their own language. Nine out of ten such projects have cost overruns; overruns of up to 50% in real terms are common, over 50% are not uncommon. The cost overrun for the Channel Tunnel, the longest underwater rail tunnel in Europe, connecting the United Kingdom and France, was 80% in real terms. The cost overruns for the Denver International Airport were 200%; for Boston’s Big Dig, 220%; and for the Sydney Opera House, 1,400% (see more examples in Table 2).
Overrun is a problem in private as well as public sector projects, and things are not improving; overruns have stayed high and constant for the 70-year period for which comparable data exist. Geography doesn’t seem to matter either; all countries and continents for which data are available suffer from overruns.
Similarly, benefit shortfalls of up to 50% are also common and above 50% not uncommon, again with no signs of improvements over time and geography (Flyvbjerg, Holm, & Buhl, 2002, 2005). Combine the large cost overruns and benefit shortfalls with the fact that business cases, cost–benefit analyses, and social and environmental impact assessments are typically at the core of planning and decision making for megaprojects and we see that such analyses can generally not be trusted. For example, for rail projects, an average cost overrun of 44.7% combines with an average demand shortfall of 51.4%, and for roads, an average cost overrun of 20.4% combines with a 50-50 risk that demand is also incorrect by more than 20%. With errors and biases of such magnitude in the forecasts that form the basis for business cases, cost–benefit analyses, and social and environmental impact assessments, such analyses will also, with a high degree of certainty, be strongly misleading. (Flyvbjerg, 2009) “Garbage in, garbage out,” as the saying goes.
As a case in point, let’s consider the Channel Tunnel in more detail. This project was originally promoted as highly beneficial both economically and financially. At the initial public offering, Eurotunnel, the private owner of the tunnel, tempted investors by telling them that 10% “would be a reasonable allowance for the possible impact of unforeseen circumstances on construction costs” (The Economist, 7 October, 1989, 37–38). In fact, costs went 80% over budget for construction, as mentioned above, and 140% over budget for financing. Revenues have been one half of those forecasted. As a consequence, the project has proved non-viable, with an internal rate of return on the investment that is negative, at minus 14.5% with a total loss to the British economy of US$17.8 billion; thus, the Channel Tunnel detracts from the economy instead of adding to it. This is difficult to believe when you use the service, which is fast, convenient, and competitive with alternative modes of travel. But, in fact, each passenger is heavily subsidized—not by  the taxpayer this time, but by the many private investors who lost their money when Eurotunnel went insolvent and was financially restructured. This drives home an important point: A megaproject may well be a technological success, but a financial failure, and many are. An economic and financial ex post evaluation of the Channel Tunnel, which systematically compared actual with forecasted costs and benefits, concluded that “the British Economy would have been better off had the Tunnel never been constructed” (Anguera, 2006, p. 291).
Other examples of non-viable megaprojects are Sydney’s Lane Cove Tunnel, the high-speed rail connections at the Stockholm and Oslo Airports, the Copenhagen Metro, and Denmark’s Great Belt Tunnel, the second-longest underwater rail tunnel in Europe, after the Channel Tunnel.

Delays are a separate problem for megaprojects and they cause both cost overruns and benefit shortfalls. For example, preliminary results from a study undertaken at Oxford University, based on the largest database of its kind, suggest that delays on dams are 45% on average. Thus, if a dam was planned to take 10 years to execute, from the decision to build until the dam became operational, then it actually took 14.5 years on average. Flyvbjerg, Holm, and Buhl (2004) modeled the relationship between cost overrun and length of implementation phase based on a large data set for major construction projects; they found that, on average, a one-year delay or other extension of the implementation phase correlates with an increase in percentage cost overrun of 4.64 percentage points.
To illustrate, for a project the size of London’s US$26 billion Crossrail project, a one-year delay would cost an extra US$1.2 billion, or US$3.3 million per day. The key lesson here is that in order to keep costs down, implementation phases should be kept short and delays small. This should not be seen as an excuse for fast-tracking projects, in other words, rushing them through decision making for early construction start. Front-end planning needs to be thorough before deciding whether to give the green light to a project or stopping it before it starts (Williams & Samset, 2010). But often the situation is the exact opposite. Front-end planning is scant, bad projects are not stopped; implementation phases and delays are long; costs soar, and benefits and revenue realization recedes into the future. For debt-financed projects this is a recipe for disaster, because project debt grows, whereas there is no revenue stream to service interest payments, which are then added to the debt, which increases interest payments, and so on in a vicious cycle. As a result, many projects end up in the so-called “debt trap,” where a combination of escalating construction costs, delays, and increasing interest payments makes it impossible for income from a project to cover costs, rendering the project non-viable.
That is what happened to the Channel Tunnel and Sydney’s Lane Cove Tunnel, among other projects.

It is far easier, however, to produce long lists of projects that have failed in terms of cost overruns and benefit shortfalls than it is to produce lists of projects that have succeeded. To illustrate this, as part of ongoing research on success in megaproject management, this author and his associates are trying to establish a sample of successful projects large enough to allow statistically valid answers; but, thus far have failed. Why? Because success is so rare in megaproject management that, at present, it can only be studied as small-sample research; whereas, failure may be studied with large samples of projects.

On one side of the paradox, megaprojects as a delivery model for public and private ventures have never been more in demand, and the size and frequency of megaprojects have never been larger. On the other side, performance in megaproject management is strikingly poor and has not improved for the 70-year period for which comparable data are available, at least not when measured in terms of cost overruns, schedule delays, and benefit shortfalls.

One may argue, of course, as famously done by Hirschman (1967a, pp 12–13) that if people knew in advance the real costs and challenges involved in delivering a large project, “they probably would never have touched it” and nothing would ever get built; so, it is better not to know, because ignorance helps get projects started, according to this argument. The following excerpt is a recent and particularly candid articulation of the nothing-would-ever-getbuilt argument, by former California State Assembly Speaker and Mayor of San Francisco, Willie Brown, discussing a large cost overrun on the San Francisco Transbay Terminal megaproject in his San Francisco Chronicle column (27 July 2013, with emphasis added):
“News that the Transbay Terminal is something like $300 million over budget should not come as a shock to anyone. We always knew the initial estimate was way under the real cost. Just like we never had a real cost for the [San Francisco] Central Subway or the [San Francisco–Oakland] Bay Bridge project. So get off it. In the world of civic projects, the first budget is really just a down payment. If people knew the real cost from the start, nothing would ever be approved. The idea is to get going. Start digging a hole and make it so big, there’s no alternative to coming up with the money to fill it in.”
Rarely has the tactical use by project advocates of cost underestimation, sunk costs, and lock-in to get projects started been expressed by an insider more plainly, if somewhat cynically.

The head of the World Bank’s Economics Department told Hirschman:
“You’ve helped in part to remove the unease that I have had in reflecting on the fact that if our modern project techniques had been used, much of the existing development in the world would never have been undertaken” (Adelman, 2013).
Hirschman’s thinking also eventually penetrated academia. Teitz and Skaburskis (2003) follow the Hiding Hand logic when they ask of the huge cost overrun on the Sydney Opera House: “Did people really think that the Sydney Opera House would come in on budget? Or did we all agree to accept the deception and engage in wishful thinking in order to make something that we really wanted happen? ... [D]o Australians really regret those dramatic sails in the harbour? Or would they have regretted more the decision [not to build] that would most reasonably have been based on a fair prediction of costs?”
The logic is seductive, yet precarious. In retrospect, of course Australians do not regret the Sydney Opera House, given what it has done for Australia though, at first, the building was not called “dramatic sails in the harbour,” but “copulating white turtles” and “something that is crawling out of the ocean with nothing good in mind” designed by an architect with “lousy taste” (Reichold & Graf, 2004, p. 168).
Non-Australians may feel regret, however; for example, the architect of the Opera House: What’s his name? Does anybody know? Only few do, which seems surprising given we are talking about the architect of arguably the most iconic building of the 20th century. And, if anybody knows the architect is the Dane Jørn Utzon, how come they can hardly ever mention another building designed by him? Because the overrun on the Opera House, and the controversy that followed, destroyed Utzon’s career and kept him from building more masterpieces. He became that most tragic figure in architecture: the one-building-architect. This is the real regret—and real cost—of the Sydney Opera House, not premier Joe Cahill’s deliberate deception about the cost— to get approval in Parliament—and the consequential huge cost overrun (Flyvbjerg, 2005). In a meeting held in support of Utzon at Sydney Town Hall in March 1966—six weeks before the controversy made Utzon leave Australia and the Opera House, in the middle of construction and never to return—the Viennaborn Australian architect Harry Seidler said, “If Mr. Utzon leaves, a crime will have been committed against future generations of Australians” (Murray, 2004, p. 105).
Seidler was more right than he could have imagined, except the crime would not be limited to Australians—it became a crime against lovers of great architecture everywhere. After winning the Pritzker Prize,  the Nobel for architecture, in 2003, Utzon again became widely acclaimed, even in Australia, where the Sydney Opera tour guides for years had been forbidden to even mention his name. But it was too late. Utzon was now 85 years old and had not built anything major for decades. So instead of having a whole oeuvre to enjoy, as we have for other architects of his caliber, we have just the one main building. Utzon was 38 when he won the competition for the Opera House. How would other works by the mature master have enriched our lives? We will never know.
As a thought experiment, consider the collected works of architect Frank Gehry, who is in the same league as Utzon; then consider which building you would choose, if you could choose only one, and the rest would have to go. So if you chose, say, the Guggenheim Museum Bilbao, then Los Angeles’ Disney Concert Hall, Chicago’s Jay Pritzker Pavilion, Prague’s Dancing House, and Seattle’s Experience Music Project Museum would be eliminated. This illustrates the high price the government of New South Wales has imposed on the world by mismanaging the planning of the Sydney Opera House

Hirschman’s and Sawyer’s theories are also flawed on a more basic level, that of validity. A close look reveals the theories to be based on small samples and biased data. Hirschman studied only 11 projects or a few more if we take into account the subprojects, and Sawyer studied 10 to 15. This important fact is typically ignored when the Hiding Hand principle is discussed. Hirschman (1967a, pp. 7, 14) seemed aware of the weak foundations and limited applicability of the principle when he called it “speculative” and useful only “[u]p to a point.” To a colleague he admitted at the time of publication that his book was “an exploration, an experiment”; to another he said he had deliberately biased his analysis “to emphasize unexpected successes” (Adelman, 2013, pp. 404–405). Even so, Hirschman went on to call the Hiding Hand a “general principle of action” and brazenly used a name for it with clear connotations to Adam Smith’s famous Invisible (Hidden) Hand. Evidently, the temptation to formulate an “economic law” was too strong, despite the weak and biased data. Sawyer (1952, p. 204) warned the reader up front that his study must be considered a “marginal and distinctly limited note.” He admitted the study considers only a “quite special kind of case” and neglects projects that were “failures” in order to focus on projects that were “successful” in the sense that “an original gross miscalculation as to costs ... was happily offset by at least a corresponding underestimation of demand.” Sawyer’s results, thus, do not describe a general characteristic of large projects, but a characteristic of his biased sample that includes only projects lucky enough to have had large underestimates of costs compensated by similarly large or larger underestimates of demand. Some would call this dubious data fishing, and the only redeeming factor is that Sawyer was disarmingly honest and tongue-in-cheek humoristic about it. He appears to not have expected to be taken wholly seriously, which he unfortunately was by some, including Hirschman.

In particular, their odd asymmetrical assumption that optimism would apply to cost estimates, yet pessimism to estimates of benefits, has been solidly disproved by Kahneman and Tversky (1979a, 1979b) and by behavioral economists building on their work. They found that optimism bias applies to estimates of both costs and benefits. An optimistic cost estimate is low and leads to cost overrun, whereas an optimistic benefit estimate is high and results in benefit shortfalls. Thus, errors of estimation do not cancel each other out, as Hirschman would have it; the exact opposite happens—errors generally reinforce each other.

The problem is not that projects worth undertaking do not exist or cannot be built on time and on budget. The problem is that the dubious and widespread practices of underestimating costs and overestimating benefits used by many megaproject promoters, planners, and managers to promote their pet project create a distorted hallof-mirrors in which it is extremely difficult to decide which projects deserve undertaking and which do not.
In fact, the situation is even worse than that. The common practice of depending on the Hiding Hand or creative error in estimating costs and benefits, thus “showing the project at its best” as an interviewee put it in a previous study, results in an inverted Darwinism, i.e., the “survival of the unfittest” (Flyvbjerg, 2009). It is not the best projects that get implemented in this manner, but the projects that look best on paper, and the projects that look best on paper are the projects with the largest cost underestimates and benefit overestimates, other things being equal. But the larger the cost underestimate on paper, the greater the cost overrun in practice; and the larger the overestimate of benefits, the greater the benefit shortfall. Therefore, the projects that have been made to look best on paper become the worst, or unfittest, projects in reality, in the sense that they are the very projects that will encounter the most problems during construction and operations in terms of the largest cost overruns, benefit shortfalls, and risks of non-viability. They have been designed like that—as disasters waiting to happen.