'Over budget, over time, over and over again' - the curse of thinking too big
As Dublin looks to expand its tram system, the plan risks running into the Iron Law of Megaprojects, warns Colm McCarthy
Last Thursday saw the release by the National Transport Authority (NTA) of the preferred routing for the next addition to the Dublin tram system, formerly called Metro North and re-christened MetroLink. This will be an expansion of capacity on the existing Green line from Sandyford in the south of the city and a new line, much of it underground, through the city centre and on northwards to the airport and Swords. The trams will be bigger than those originally ordered for the Red and Green lines but will operate on the same gauge of track, narrower than mainline or suburban rail.
They will squeeze in 500 passengers, mainly standing, when fully loaded, versus around 300 for the shorter first-generation Luas trams. Suburban trains are bigger and can take 1,000 passengers or even more. Total cost for MetroLink is estimated at €3,000m, which would make it easily the most expensive single project ever undertaken in Ireland.
In addition to the routing, on which there will be a public consultation, the NTA released a 28-page summary cost-benefit study prepared by the engineering consultants Systra, a French company whose principal shareholders are RATP, which operates the metro and the regional railways around Paris, and SNCF, the French national rail company. A full cost-benefit study is promised once the consultation phase has been completed. The ratio of capital cost (three thousand million euro) to the length of the cost-benefit study as published works out at more than €100m per page and it is surprising that the board members at the NTA felt able to sign off on this huge project on the basis of such a succinct evaluation.
MetroLink qualifies, in the language popularised by Oxford economics professor Bent Flyvbjerg, as a 'megaproject', dwarfing the four recent schemes completed in Dublin put together. These were the Red (Tallaght) and Green (Sandyford) Luas lines, the Port Tunnel and the new Cross City Luas line which cost, in aggregate, just over €1900m. Each of the four over-shot the initial estimate on which the political go-ahead was based and all experienced delays in commissioning. This Dublin experience would not have surprised Professor Flyvbjerg, the most prominent critic on failures in the economic evaluation of megaprojects. He has studied cost over-runs on hundreds of similar ventures from dozens of countries and has enunciated what he calls the Iron Law of Megaprojects: they end up "over budget, over time, over and over again".
Reviewing the experience with hundreds of transport megaprojects around the world, Flyvbjerg and his collaborators have documented a depressing pattern. Their out-turn costs tend to exceed the initial estimates by large margins and their benefits tend to be exaggerated by the project promoters. With cost estimates too low and benefit estimates too high, the excess of benefits over costs used to justify these projects are rarely realised in practice. There have been numerous cases where costs were so high, and benefits so low, that the projects would never have secured political support had better cost-benefit estimates been prepared, according to Flyvbjerg.
Moreover, he has found that there has been no tendency for the accuracy of megaproject evaluations to improve over time. They have always, with the occasional exception, been too rosy, and this continues to be the case. This rules out some innocent explanations for the continuing errors in cost estimation and the over-estimation of project benefits, namely that it takes time to learn from mistakes, or that the technology is novel. In urban transport, the technology is not new and the record of unrealistic economic evaluations goes back over a century. Flyvbjerg and many other economists around the world have regretfully settled on an alternative explanation. Megaproject evaluations suffer from more than optimism bias and are plain deceitful: project promoters have turned cost-benefit analysis into a branch of the public relations industry.
From the brief document prepared by the engineers at Systra, it is difficult to judge whether the MetroLink evaluation conforms to Bent Flyvbjerg's Iron Law of Megaprojects. However, there are reasons to be fearful that even the enormous €3,000m cost budget may not be enough and that the excess of benefits over costs is illusory.
The first baby step has been taken in preparing for a possible cost over-run. Calculations are now shown for capital costs up to €4,000m - when the project was unveiled as part of the National Development Plan on February 16, it mentioned only the €3,000m figure. But the benefits are claimed to have been quantified at up to €6,000m, so the Systra engineers felt able to give an enthusiastic thumbs-up for MetroLink.
They appear to have made no provision for uncompensated disruption costs imposed during construction. This was also a feature of the Luas project evaluations but the breakdown of cost components has been omitted by Systra, so it difficult to tell. The issue is important. Urban transport projects often involve a sizeable bill for incurred compensation costs, particularly property acquisition, and no doubt a provision has been made. But uncompensated costs can be even bigger. These arise, for example, when retail business is damaged by construction work or when amenities cannot be accessed for three or four years. Retailers in Dublin city centre claim to have gone bust during the Luas construction.
Two famous northside sports clubs, Na Fianna GAA in Glasnevin and the illustrious Home Farm football club, the nursery of a long succession of Irish international stars, claim that their pitches will be unusable for years due to construction works on MetroLink. The whole point of cost-benefit analysis is to go beyond the narrow financial impact and to allow fully for factors such as uncompensated disruption.
There will be more ructions in the leafy suburbs along the Sandyford line south of the city centre, since extensive re-engineering along the Green line is also required. If the French engineers are unfamiliar with uncompensated costs, the ecumenical coalition of Na Fianna, Home Farm and the burghers of Ranelagh will bring them up to speed in short order.
There is another risk on the cost side. Archaeologists have uncovered yet more Viking artefacts recently while exploring a hotel site in the Liberties and Finn McCool must be buried somewhere. MetroLink requires extensive tunnelling right through the centre of the old city of Dublin and there will be more happy discoveries.
The most puzzling aspect of the summary cost-benefit report is the apparent assumption of high passenger volumes. The engineers appear to have assumed intensive future development along the route. The record around Dublin is not encouraging - there are unopened stations on both the Green line and the Kildare suburban line, for example.
Moreover, the study appears to compare the MetroLink project to a 'do-nothing' scenario - economists prefer to compare projects also to the best available alternative. In the absence of published detail, it is always possible that benefits have been counted twice: for example, fare revenues should not be credited in addition to passenger time savings.
This project falls under the public spending code and will have to pass muster with the economists at the Department of Public Expenditure. Systra has given them plenty to work on.