A new report from Infrastructure Australia, Improving Public Transport: Customer Focused Franchising, and its associated technical report from Pricewaterhouse Coopers (PwC), will have state and federal treasurers salivating. The accompanying press release is clearly intended to set their fiscal juices flowing. It suggests that:
… subjecting the operation of Australia’s government-operated bus and rail services to competitive tender processes could save Australian taxpayers up to $15.5 billion by 2040…
This A$15.5 billion is the estimated high operational cost savings scenario, in present value terms (2016 prices; 7% real discount rate). Infrastructure Australia/PwC’s conservative savings estimate is still a very substantial A$11.6 billion.
The magic pudding is competitive tendering of a range of government-operated rail and bus public transport services, as well as Melbourne’s privately run bus services, which are not currently awarded by competitive tendering.
Just how realistic are these savings?
Ownership and contracting of public transport services has been a major transport policy focus for 30 years. This issue has had its own two-yearly conference, the Thredbo International Conference Series, for most of that time.
The Thredbo series (named after the location of the first conference) provides a vast array of international material that sheds light on ways to improve the efficiency and effectiveness of public transport under different operating regimes. Australian federal and state government officials rarely attend.
The PwC report presents Australian and international examples of competitive tendering of rail and bus services. It concludes that first-round rail tenders will deliver operational cost savings of 15-20% (conservative and high estimates respectively) by the end of the first tender and 25-32% by the end of the second. For bus services, the report estimates savings of 15-20% from the first tender and 30-35% by the end of the third round.
International experience on rail privatisation is sparse, with the sources that PwC cites among the more well known. We are concerned, however, that contrary evidence is given no attention. This raises questions about the objectivity of the analysis.
Andrew Smith and colleagues, in a paper to the Thredbo 6 Conference, reviewed the UK rail experience. They found costs rose after privatisation.
Total passenger train operating costs (excluding infrastructure) per train-kilometre were 14% higher in 2006 than at privatisation ten years earlier. Between privatisation and 2008-09, real average salary costs rose much faster than in the economy as a whole.
Pedro Cantos and colleagues reviewed the European experience across 16 countries. Their study concluded that tendering regional passenger services did not have a significant effect on efficiency and productivity.
As Public Transport Users Association president Tony Morton wrote in May 2017 to The Age:
Victoria’s train/tram privatisation hasn’t saved taxpayers one cent. A former minister, the late Lynne Kosky, admitted as much ten years ago. We do save money relative to Sydney by not having train guards, a decision that predates privatisation.
Those labour savings were part of large reductions in Melbourne rail staff, from 18,000 to 8,400, during the 1990s. This was done in a corporatisation process that preceded privatisation.
Taking account of the sources cited above and Melbourne’s experience, as well as the sources PwC cites, an assumption that rail privatisation will deliver cost savings of 25-32% by the end of Term 2 tenders is, to say the least, heroic.
We would see even the “conservative” figure of 25% as a high estimate. Melbourne’s experience suggests many of the mooted savings might also be achievable by a thorough corporatisation process. Infrastructure Australia, by simply taking the competitive tendering pathway, does not do justice to the choices available.
In a 2007 review of bus privatisation, one of us reported cost reductions from the first round of tenders as follows:
Great Britain 50-55%
Scandinavia most in the range 20-30% but with a range of 5-34%
Australia: Perth 22%; Adelaide 38%
New Zealand: 40% ex-public operators; about 5% private operators
The savings for Norway are lower than elsewhere because the local bus industry had been improving its efficiency over some time, mirroring NZ experience. The Norwegian study suggested the threat of competitive tendering was important in driving savings.
These various bus findings suggest the cost savings range of 30-35% selected by PwC and reported by Infrastructure Australia might be possible, if an efficient private operator replaces an inefficient public operator. We make no comment on the efficiency of the public bus operators reviewed by PwC.
But such savings will not be achievable in Melbourne, where private operators already provide service. Cost savings of less than 5% are more realistic there.
As we argue below, achieving the latter savings does not rely on competitive tendering. Importantly, all such savings, in rail and bus, are a one-off windfall gain. Our research commonly shows evidence of significant real cost increases after the initial tender round.
What about the alternatives?
Wallis, Bray and Webster, pioneers of the Adelaide privatisation experience, recognised the tendency for costs to increase in subsequent tendering rounds. They proposed that performance-based contracts negotiated with incumbent operators should replace the city’s fourth round of bus tenders.
It is surprising that neither the Infrastructure Australia report nor PwC technical paper mention such contracts as an alternative to competitive tendering.
Our research suggests that negotiated performance-based contracts with efficient operators deliver cost outcomes in line with competitive tendering. And this approach provides two other benefits. Transaction costs are lower and it provides opportunities for closer partnering between purchaser and provider on service planning.
Importantly, given competition law in Australia, if it can be shown that the benefits of a particular contracting regime, such as competitive tendering or negotiated performance-based contracts, outweigh the costs, then the compliance case is strong.
It is surprising that Infrastructure Australia is either not aware of, or more likely has chosen for some reason to ignore, the opportunities afforded by negotiated performance-based contracts, particularly for bus services. This narrowing of the policy options put forward for public consideration risks loss of benefits from important transport reforms.
In short, we believe that Infrastructure Australia, and its advisers PwC, have overestimated the achievable cost savings from competitive tendering of publicly provided rail services. They have also ignored the potential for savings within existing publicly provided public transport services. Competitive tendering of such services will deliver cost savings, but it would be surprising if they were as big as suggested.
For Melbourne bus services, any operating cost savings from competitive tendering are likely to be much less than forecast, most likely under 5%. Such savings will be equally achievable under negotiated performance-based contracts.
Although we recognise that the first round of any competitive tendering will produce a windfall gain, this will be partly at the risk of service quality. Once we consider this, the one-off financial benefits are diluted.