Mike Baird can make real reforms in response to the light rail saga
Published by The Sydney Morning Herald, Friday 9 December
Premier Mike Baird’s renewed commitment to deliver infrastructure projects “on time and on budget” was still ringing in the air last week when, just hours later, the NSW Auditor General released a report blaming poor planning for another leap in the cost of the Sydney Light Rail project.
But closer monitoring of project costs – the department’s solution – is unlikely to make much difference to a problem that is entrenched in all Australian states.
Instead, the Premier should reinterpret the Sydney Light Rail saga as a prompt for genuine reform.
The first lesson: Only promise infrastructure after you’ve done your due diligence
Kristina Keneally was the first politician to commit to the Sydney Light Rail project, in 2010, at a cost of $500 million – before the route had been determined. In 2012, Barry O’Farrell revised the project’s cost up to $1.6 billion – without a properly assessed business case. Given how little homework had been done on the project, it’s unsurprising that the cost shot up to $2.1 billion by the time Mike Baird signed contracts, in late 2014.
The Sydney Light Rail is a textbook case of premature announcement. Grattan Institute analysis has found that 75 per cent of the value of cost blowouts come from projects where a politician made the first cost promise too early – before a formal budget commitment and generally before a proper business case. This is because these prematurely announced projects need larger cost upgrades not just early on, but throughout their lives.
The current guess-and-check approach to planning multi-billion dollar investments is negligent. If Premier Baird wants to deliver projects “on time and on budget”, the first lesson to take from the Sydney Light Rail would be to improve the governance of project announcements. Proper business cases should be reviewed by Infrastructure NSW or Infrastructure Australia and tabled in Parliament before public funds are committed.
The second lesson: Learn from history
The fate of the Sydney Light Rail project demonstrates that that we need to get scientific about estimating projects’ cost risks. Cost estimates for Australian transport infrastructure projects are systematically optimistic – over the past 15 years, these projects have cost 24 per cent more than first estimated.
The NSW government’s estimates of cost risk on their current megaprojects seem implausible. For example, the WestConnex “worst case” cost estimate is only 6 per cent above the “likely” cost – less than a quarter of what it should be if the experience of the past 15 years is any guide. If the kinds of problems that have pushed past Australian transport into “worst case” scenarios materialise, WestConnex’s costs could keep rising, by as much as a further $3.4 billion more than the “worst case” the NSW government has envisaged.
And that’s precisely the problem. In the absence of detailed historical data on projects’ cost outcomes, the cost estimators are expected to arrive at realistic estimates of projects’ risks with good information on the costs of concrete, wages and equipment, but in the dark about the “unknown unknowns” that, in practice, can make up a substantial component of project costs.
The saga of the Sydney Light Rail is a perfect illustration of this problem. For a large project that is announced prematurely, expect the cost to be substantially higher – 38 per cent on average – by the time a contract is signed. Nor are the overruns over and done with at that point. Projects with troubled beginnings like the Sydney Light Rail typically incur an additional 10 per cent cost overrun during the construction period. If this were to occur on the Sydney Light Rail, the project’s benefit cost ratio would fall further, to below 1.3.
Of course, insiders may explain the cost overruns on infrastructure projects by pointing to scope changes. And it’s true that scope changes do explain a modest share of cost overruns – 9 per cent, in this latest cost hike. But even where the scope changes are so substantial as to alter the whole character of a project, cost changes matter because it’s unlikely that the project will be cancelled even if it is no longer worth building.
If Premier Baird wants his legacy projects to finish on budget, he should demand much more than closer monitoring of costs as projects are built. The actions that would really make a difference would be to curb his fellow politicians from promising to spend public money before a business case and project evaluation have been tabled in the Parliament, and to publish the post-completion report on projects afterwards. Then perhaps Mike Baird could be the Australian premier that finally nails “on time and on budget”.