The good, the bad,
and the better
Infrastructure July 2015

Nicklas Garemo
Stefan Matzinger
Robert Palter

Infrastructure megaprojects are crucial to the future classes. urbanization. most All these megaprojects have worked as intended. and their demand is overestimated by 51. it’s important to remember that building and maintaining infrastructure is a critical and The risks associated with megaprojects—those that sometimes even lifesaving undertaking. it is almost impossible to think of these York a decade just to begin the $3. and metro projects still will not have materialized. Construction began in 2000. Brazil. and the promised road. Megaprojects: The good.1 Rail Protection Works.9 billion project places without them. Much of the Netherlands project management at Oxford’s business school. states. infrastructure. keep diseases one influential study. of 44. the bad. the MTR. Also not surprisingly. The increased economic growth. are being expanded. McKinsey has estimated that Big infrastructure projects can also be economically bridges and tunnels incur an average 35 percent cost transformative. which has enabled the densely Consider the metro system in Salvador. the premise that projects need to work on two levels—in the short term for recovering financial McKinsey estimates that the world needs to spend outlays and the longer term for creating social about $57 trillion on infrastructure by 2030 to impact—often becomes a barrier to taking action. for roads. go over budget by an average country’s landscape. Of that. the cost of maintaining the worn-out and airport projects are both running late and well bridge keeps increasing. Bent Flyvbjerg. about two-thirds will be required in especially in places where revenues from a project developing markets. of the work is still unfinished. bridge. Consider the Panama Canal. the canal meanwhile. and individual livelihoods. and the better Building big infrastructure projects is always risky. It overrun. 27 percent of its GDP. all of them to rebuild the 59-year-old Tappan Zee Bridge. for example. Finally. And Hong Kong would surely grind to a halt without its clean and speedy subway Time overruns. However. 2 . this leaves governments to accounting for 21 percent of Dubai’s employment and pick up the tab for the rest. and of cities. Given that many accounts for a significant share of the country’s GDP. it’s 20 percent. for example. Sewage and cost $1 billion or more—are well documented. These countries need problem is that these projects often go off the rails. but there are ways to improve the odds of a smooth landing. too. In water-supply systems. Even projects that are needed do not get executed. projects are approved with a 20 percent return on Dubai’s international airport is the world’s busiest. above the initial budget.4 percent. It has taken New indeed. system. population growth. an expert in such as cholera at bay. investment expected. even the hyperefficient MTR has run into delays with some of its projects. but all too often many years will pass either with regard to budget or time—or both. packed city to build beyond the downtown districts. enable the anticipated levels of GDP growth globally. but it took more than a dozen years for the first passengers to ride it. where there are rising middle are unlikely to cover its cost. would be under water without the North Sea estimated that nine out of ten go over budget. Not surprisingly.7 percent. which guards that low-lying projects. are a perennial problem.

productivity will not compensate for these shortfalls and dozens of private contractors. we consider the most important look at other cities that have built similar lines to reasons that megaprojects falter. For example. the delay can stall the entire project. For example. In part. In order to and construction firms that have been contracted to justify a project. Having delivered an unrealistically Why projects go bad low project budget. for example. that poor execution was responsible for cost and such projects are on a fast track to failure. time overruns in 73 percent of the cases.2 Up-front preparation 3 . issues are not deeply considered or priced to the construction productivity has remained flat or even fullest before launching a project. Just one issue can stall the process project system. it took two countries a decade to work out the diplomatic Another challenge is low productivity. and because such delays tend to ripple through the entire end users. and then we understand the true cost and time dynamics. it should that question. declined. All too often. numerous local governments. indefinitely.How can companies and governments do better that don’t necessarily justify the preferred course to build megaprojects that deliver their societal of action. Poor execution. In one case. A deemed affordable. suggest principles for improvement. and even mathematical the data until they come under the limit of what is errors in scheduling and risk assessment. McKinsey study of 48 troubled megaprojects showed would make a project unpalatable. a new railway could involve three national example. Flyvbjerg argues that problems such as incomplete design. resulting in One useful reality check is to compare the project higher costs for the same results. for the rest.” this process addresses confirmation infrastructure more efficiently can reduce its bias by forcing decision makers to consider cases whole cost by 15 percent. execution is poor because many projects cross state or national borders and involve projects are so complex that what might seem like a mix of private and government spending. the temptation is to cut corners There are three main reasons for failure. Wages. if steel does not arrive at the job site on governments. he writes. have continued to rise faster than inflation in many markets. under consideration to similar projects that have already been completed. sometimes costs and timelines deliver the project. time. one of the specialty trades has a problem. if a city wants to build a ten- benefits and do so on time and on budget? To answer kilometer metro line with four stations. Known as “reference-class McKinsey studies have shown that delivering forecasting. from design are systematically underestimated and benefits and planning through construction. stating the real cost. is riddled with systematically overestimated. these were due to politics. ill-advised shortcuts. however. Higher varied degrees of skills and wage expectations. suppliers. While considerations that allowed them to build a the manufacturing sector has approximately hydroelectric dam. Ditto if different environmental and health standards. such as new governments A common example of this comes when big or laws. lack of clear project managers competing for funding massage scope. these complicating doubled its productivity over the past two decades. to maintain cost assumptions and protect the (typically slim) profit margins for the engineering Overoptimism and overcomplexity. For routine issues can become major headaches. Project execution. From the outset.

excess power supplies. it established the Public and Private megaprojects in a lifetime. the transit accounting for 75 percent of journeys. the first three layers are looking for this ideal than others. are another issue. misalignment ƒƒ Layer 3: Construction manager to owner’s among stakeholders. has more work and more money. and are compounded by the speed at which projects get construction can generate savings of as much as started. selects transport projects. The following but regularly going over time and over budget structure is common: implies that there are systematic errors at work. All told. Costs are commonly underestimated and benefits ƒƒ Layer 4: Owner’s representative to project overestimated. down from the top leadership. All these problems gains in approval. and representative inability to find or use the appropriate capabilities. Because each one is Infrastructure Investment Management Center unique. the learning curve is steep every time. megaprojects may 25 percent on new projects. projects for a living may execute only three or four In 2005. with public are looking to deliver on time and budget. while the later ones a national goal for dense urban living. This authority to make final decisions is often remote aspiration guides how the Land Transport Authority from the action. Large projects are typically either sponsored by the government or by an entrepreneur with bold The process for selecting projects needs to be fact aspirations. And that means these problems can be identified ƒƒ Layer 1: Subcontractor to contractor and addressed. insufficient planning. The key is to first establish social and economic priorities ƒƒ Layer 5: Project sponsor to business executive and only then to consider what projects are best suited to deliver them. people in 12 to 18 months—a significant operational and managerial challenge equivalent to creating a Weakness in organizational design and new start-up company. Starting from scratch. Even individuals who build large infrastructure South Korea shows one promising approach. capabilities. That requires developing This is a problem because each layer will have a independent and robust analyses on the true cost view on how time and costs can be compressed. without compromising have to create organizations of several thousand the quality of outcomes. and benefits. Without such rigor and oversight. As a result. for instance. Also. you can imagine bridges to nowhere. Capabilities. and empty roads. engineering. and (PIMAC) to get accurate data on costs and 4 . hindsight shows or managing contractor that the problems began at the outset. Some countries are closer to achieving For example. Singapore. due to poor justification and need for the project. Many entities involved in building megaprojects have an organizational setup in How to deliver on the promise of megaprojects which the project director sits four or five levels Any big project carries a big opportunity for failure. the baseline for assessing sponsor overall project performance is wrong. procurement. based and transparent to ensure accountability. they can take 10 to 15 years to finish. or lack thereof. pays for itself many times over. efficiency the skills needed are scarce. ƒƒ Layer 2: Contractors to construction manager Typically. when projects go wrong.

First. however. The state of Virginia starting construction. 5 . Edward Merrow. payments to the time to approve power-industry infrastructure contractors may not correlate to actual construction from 12 months to 9. Finally. A more sophisticated approach is to use real-time data that measures activity in the field. including on public review. Measuring progress on the basis of cash limits all along the way. the project timelines and costs by roughly 20 percent. It is problematic when there are or project sponsors spend 3 to 5 percent of the multiple estimates of the cost and time performance capital cost of the project on early-stage engineering of the project relative to the baseline. though doing so clearly improves project megaprojects. This differs a project than to build it. That means the applying these approaches.benefits. which is to track progress permits involve prioritizing projects. compared with an average of progress. saving time and money. baselines get adjusted time and again. not unusual that it takes longer to get approvals for relative to work plans and budgets. they do of money on early-stage engineering and design not have robust risk-analysis or risk-management for three reasons. It’s cubic meters of concrete poured or earth moved. performance. Improving project performance requires four years in the rest of Europe. better planning and preparation in three areas. Projects can also be designed to reduce time- Do the engineering and risk analysis before consuming land battles. on-time and on-budget delivery. moreover. By it takes more than 30 days to pay. because usually Providing “one stop shop” permitting can help. Second. This is because the This limits the partners’ ability to figure out how to design process will often raise challenges that need accelerate project delivery and control cost overruns. outdated (as they generally rely on payments to they worry that the design will be modified once contractors rather than on actual work performed) construction is under way and thus make the and not aligned with the true progress of the project. such as Streamline permitting and land acquisition. and establishing time budget. Most project-development organizations or project Distressed projects have another thing in common: sponsors are reluctant to spend a significant amount they lack adequate controls. is less than ideal. Before PIMAC. England and Wales cut related data are out of date. This is often stated but rarely completed a plan to widen Interstate 495 in 2012. Best practices in issuing from the usual approach. In addition. defining clear by measuring funds paid to contractors relative to roles and responsibilities. to be resolved before construction starts. flow. they are eager used to report on project progress are typically to break ground and start construction. expenditure on up-front design pointless. conducts value-for-money tests. rejection figure was 3 percent. and evaluates projects when they of a book on megaprojects. and contractors and owners use different metrics to Our experience indicates that if project developers measure progress. Specifically.3 has shown that the best are done. which means and design. they often lack the funds protocols and do not provide timely reporting on in the early stage of a project to spend significantly progress relative to budgets and timelines. PIMAC does feasibility studies on public followed. PIMAC has rejected almost half examples of project-definition work reduce both the projects that it has reviewed. the founder of the sponsors comprehensive research on improving Independent Project Analysis consultancy and author public investment. So far. The data on design and engineering. it tends to produce far better results in there is no common understanding of performance.

1 Bent Flyvbjerg. they will fail. It is not enough 2 For more. 2011. An experienced project manager is Nicklas Garemo is a director in McKinsey’s Abu Dhabi not enough. go well beyond a specific bridge. on mckinsey. 6 .” Project Management Journal. see “Infrastructure productivity: How to save for them to have a vague theoretical overview of $1 trillion a year. and Robert Palter is a director in the Toronto office. escalating contractor’s costs. a detailed. regulatory approval. contract management. January 2013. and government and community relations. Merrow. for everyone. putting together the project team. and Practices for 6–19. At worst. Stefan Matzinger is a director in the Hong Kong all the requisite skills. Getting it right. expertise. including legal and technical office. projects will not deliver the best possible return on investment. first edition. players must assemble a team that has office. “What you should know about megaprojects and why: An overview. Hoboken. or replacing a high- tech supplier. Industrial Megaprojects: Concepts. or sewage Without a well-resourced and qualified network system. lack and to create the economic growth that will pay for them. 2014. They need to create 3 Edward W. tunnel. stakeholder management. likely eventualities as managing quality risks. how the project should work. is good of project managers. or at least better. Investors and owners need to take an active role in Volume 45. Number 2. project reporting.” McKinsey Global Institute. and controllers. after a private design company came up with a plan The world needs megaprojects to deliver the that cut costs markedly and eliminated the need to economic and social goods that billions of people remove hundreds of homes. advisers. NJ: John Wiley & Sons. practical approach to deal with such Strategies. pp. But a bad project has consequences that Build a project team with the right mix of abilities.

July 2015 Designed by Global Editorial Services Copyright © McKinsey & Company .