"Suppose the amount expended in the purchase of a given service to be a certain sum, and that an individual equally capable of rendering this service should offer to render it at less expense. Is there any good reason for refusing such an offer?"
- British philosopher Jeremy Bentham (1748-1832)
An unassailable proposition, surely? An idea whose time has come? These words were uttered 200 years ago by the British philosopher Jeremy Bentham, and yet there is still discomfort at the idea of market-testing public services. The notion that competed-for services are more efficient than monopolies (and more responsive to their customers) has instinctive appeal. We all understand the virtues of competition over monopoly from the everyday experience of life. And yet some people are uncomfortable with the idea that providers might make a profit out of public services, even if they are delivered at much lower costs. Some fear that, by accident or design, the public might get less for less.
One of the most common sources of discomfort with market testing lies in the difficulty of knowing whether services are actually being rendered at lower cost. Since public services are (in general) not traded in the marketplace, the challenges of measurement are complex. In a recent study of contractor logistics support in the US Air Force, the American think tank RAND noted that statistical analysis of the comparative efficiency of in-house and outsourced provision is difficult because the data required to standardise the diverse characteristics of different programmes does not exist. There were problems with comparing the same programme over time because of changes in key characteristics such as age, usage and fleet size. Contrasting public and private is compounded by the way in which overheads are allocated in the public sector, and performance metrics are often not directly comparable. Of the studies that do exist, many are based on the findings of an artificial ‘public sector comparator’ and the estimated cost of delivery by a contractor (rather than the actual cost). Writing of public-private partnerships, Professor Paul Grout of the UK’s Bristol University has repeatedly argued that, “Private suppliers are damaged by a dearth of information… no-one ever keeps a record of the whole-life cost of the public alternative.”
However, there is still a considerable body of evidence showing that, if done well, competition and contracting can deliver much greater productivity. A survey of competitive tendering in English local government throughout the 1980s and 1990s reported that costs were lower as a result of competition, with most studies showing no change in quality. A more recent survey, undertaken by Paul Grout in 2008 for the Julius Review of Britain’s public service industry, reported savings of up to 30%, noting that while the evidence on quality change was weaker, the few rigorous studies available pointed to similar or improved service quality. An international study conducted in 1996 by Australia’s Industry Commission (a government think tank) also reported savings of 10 to 30%, and an academic survey four years later concluded that there were real cost savings of around 8 to 14%. In 2004, two more Australian academics found that, “While there is still some ongoing debate about the magnitude of the cost change associated with outsourcing, there seems to be some consensus about the direction of the change.” Recent European surveys have also reported savings in the order of 10 to 16%.
Contestability
One of the more comprehensive surveys in recent years was undertaken by The Serco Institute, drawing on 198 studies from 12 differentcountries over 30 years, across five sectors – defence support, hospital services, prisons management, municipal services and refuse collection. One of the ways in which this 
surveydiffered was that, where possible, studies were used that considered the impact of competition and contestability, rather than public versus private ownership.The results differed from sector to sector and from country to country. In defence support, studies from Australia, New Zealand and the United States reported savings in the range of 20 to 30%. Financial benefits in excess of 20% were reported from the competition of health support services in England, Australia and Denmark, but were lower in other jurisdictions. Of the ten methodologically robust studies of prison contracting in the US, all but one found savings, mostly in the range of 5 to 15%. In England, the first-round savings from competition in prison management were 11 to 15%, but a series of competitions based on design, construction and management brought further cost reductions (in capital and operating costs) of 38%. Studies of refuse collection, across several countries, have sparked debate, and the financial gains cluster around 20%. On the other hand, the evidence from the contracting for other municipal services is mixed, ranging from 5 to as much as 25%.
Quality matters
It would be helpful to know what accounts for these differences, and while the evidence on this is not extensive, we have enough to extract some insight. Unsurprisingly, the quality of procurement and contract management matters. For example, savings tend to be somewhat less when there is only one bidder and government discloses its comparator, as Kentucky discovered with the tender for the Otter Creek Correctional Facility in 2005. That much is obvious. Less obvious, but just as real, are the benefits that come from specifying outcomes or high-level outputs. Where commissioners use detailed input measures, they are demanding that providers deliver the service in the way it has always been delivered, denying themselves the opportunity for innovation.
Market design also matters. In America, there are two different kinds of prison market: regional markets based on a single government buyer procuring services from a multitude of suppliers through competitive tendering, and a national market where public and private sector providers build prisons on a speculative basis, and sell places to state governments urgently in need of additional facilities to relieve their overcrowding. The available evidence suggests that the latter spot market is less efficient than the markets where services are commissioned, in part because facilities built on speculation are smaller and thus more expensive, in part because it has been a seller’s market, and providers have been able to dictate monitoring levels and thus influence quality standards.
The way in which services are organised prior to competition will also have an effect. This has been true of contracting for defence support: uniformed personnel are significantly more expensive to deploy than civilians, partly because they are more highly trained and partly because their military duties require them to be diverted into training or transferred more quickly to other jobs. In this case, contracting has been a vehicle for the civilianisation of support services. It also matters whether the service in question has already been subjected to competition. Unless the scale or scope of the service changes over time, or there has been technological innovation, it will be difficult for future tenders to capture additional savings of the same magnitude. However, the UK Ministry of Defence has progressively expanded the size and range of its base support contracts over several decades, allowing providers to exploit scale economies and explore synergies between a wider variety of functions. In prison contracting, the expansion of contractual responsibilities under the ‘Private Finance Initiative’ enabled providers to deliver new operational efficiencies through innovation in design and the introduction of new technologies.
Gaining success
All of this suggests that, over the medium-term, governments have the potential to bring down the cost of delivering public services considerably. Whether or not providers succeed in reducing costs will be heavily influenced by such matters as market design, the nature of performance specification and the customers’ willingness to explore alternatives in scale and scope. In Britain, it seems that the new government is sensitive to these considerations, although, as always, the devil will lie in the detail of execution. Success will also depend, in part, on the ability of providers to meet the challenges associated with new markets and innovative new contracting models such as Payment by Results.
Studying Contracting Reforms in Real Time
When vehicle maintenance and warehousing services were outsourced at Trentham Army Base, New Zealand in 1997, researchers from the Australian Graduate School of Management were given unprecedented access to data from both the government and contractor to study the impact of competition and contracting. The task was made easier by the availability of detailed information on cost and performance prior to competition. After service levels and contract variations were taken into account, the researchers indicated cost savings of 24%.
Of greater interest, however, were insights about the sources of these cost reductions. In the workshops, savings came from reductions in capital and labour slack, indicating reductions in inefficiency. In the warehouse, however, they came through technological innovation, reducing the time and labour required to deliver the same outputs. The contractor also made a significant reduction in administrative overheads. While wages were somewhat lower after competition, well over half of the savings came from productivity improvements, using fewer staff to perform the required services, and around a quarter of them came through technical change.
The debate over evaluation
The one public service where outsourcing has been well studied over many years and across a variety of different countries is household refuse collection. In part, this is due to a relatively simple service where performance is easily measured. It is also because competition and contracting have been widely used, so thatdata sets are large. And there has been a robust methodological debate, so that researchers are confident about the data.
When refuse collection in the UK was first studied in the mid-1980s by economist Simon Domberger and others, they found that these services were around 22% cheaper after competition, with no evidence that this had been achieved by compromising quality. Greater productivity of labour seemed to provide the explanation.
Their methodology came under criticism, including the claim that their results had been distorted by a few high-performing councils. But even when these outliers were omitted (and it was not clear to Domberger et al that they should be) average savings still amounted to 15%. Later studies using a longer time series also found savings around 20%. Studies in Ireland, Switzerland, Catalonia and the Netherlands have found similar results,
as have surveys in the United States and Canada.
