Editor's Note: The first half of this article was published on the International Network earlier this month. To view it visit our Archives at http://database.corrections.com/news/

Services purchased by UK public authorities under the Private Finance Initiative (PFI) are estimated to be yielding on average 17% savings from many of the early contracts. This is a key finding from a study by Arthur Andersen and Enterprise LSE commissioned by HM Treasury's Private Finance Taskforce and published on its website: www.treasury-taskforce-projects.gov.uk

This article briefly summarises the findings of the Arthur Andersen/Enterprise LSE study and then, after a review of the history of PFI prison contracts, considers one of the study's main area of focus: public sector comparators (PSCs) as an aid to deciding whether to let a PFI contract against the data from this series of contracts. It also highlights the value of the prison programme as a model of serial PFI procurement and long term contractual management.

For how long will the estimates of savings in the PSC be valid?

The Arthur Andersen/LSE study drew attention to the critical importance of the good public sector contract management. When the prison has been in operation for a year or so, forecast savings are likely to change as a result of external factors. For example, as the average cost of the comparator reflected overcrowding in local prisons, while the PFI cost per place is at a higher 'uncrowded' cost per place. Prolonged accommodation pressures would mean that such a comparator would underestimate potential savings. Internal public sector management could have a significant impact.

• Inevitably during a 25-year contract period, the nature of the service required will undergo changes. These could provide opportunities to improve value for money (eg. by sharing benefits on refinancing after service has commenced) and risks (eg. if there are penal policy and legislative changes to be implemented).

• By creating benchmarks, PFI and other contractually managed prisons act as a spur to create greater efficiency within directly managed prisons. In addition, the introduction noted by the NAO of a PFI developed modular houseblock technique into the Prison Service's own building program contributed to improving value obtained by the use of conventional (ie. publicly funded) capital.

None of these potential changes can be quantified in the initial assessment of likely savings prior to contract signature. Moreover, where they reduce savings because benchmarking helps raise standards in the public sector, the latter savings, they may not have been achievable without the contract going ahead and these, indirect savings (ie. some 130 directly managed prisons compared with the current 11 contracts) are likely to be greater than the direct savings identified in the PSC. As noted above, however, the tranche 3 and 4 comparators assumed substantial cost improvements to directly funded services. This balancing factor adds significantly to the robustness of the results over time.

Does the PSC distort the vfm assessment away from quality of service?

The comparator's clear cost focus could blind decision makers. An important question with the prison PSC was whether they could taken as evidence of likely value for money, given and the concerns of a number of penologists regarding the potential quality of regime in contractually managed services?

Internationally, Harding has demonstrated subsequently that these concerns and similar misgivings expressed abroad are not supported by available evidence, as:

• Congruent evidence of significant cost savings (as well as in the UK, 11-14% in Louisiana, 13-17% in Arizona and 9-13% in Queensland).

• Migration of good practice between sectors (especially US public sector prisons following the example set by contractually managed counterparts in seeking American Correctional Association accreditation).

• An important study in Florida suggesting that recidivism rates can be improved significantly when the quality of anti-recidivism programs is incentivised by the fee structure.

The potential benefits (in terms of the cost and impact of crime) of the latter finding are not quantified for the above estimated savings.

Within the UK, the experience of the contractual management of publicly funded new prisons since 1992 by the same operators as PFI prisons, provided confidence about acceptable service delivery likely to be delivered, coupled with increasing competence with the HMPS unit concerned in the arts of tender specification and bid evaluation:

While the initial difficulties of new contractually managed prisons have been well documented (eg. Wolds and Doncaster), equally serious problems have been encountered in new directly managed prisons (eg. Moreland and Whitmore). Of the thirteen incidents of fee deductions for contractually managed/PFI prisons between February 1994 and February 1999, nine related to failures to achieve service standards by the first twelve months or so of operations . However, during its first year of operation, Fazakerley Prison operated at an average 11% overcrowding level , exceeded only at two of the newer directly managed local prisons which were not exposed to the additional operational risks inherent in overcrowding for several years after they opened .

For contractually managed prisons that have been in operation for a number of years now, the latest detailed analyses by the Prison Service of assault statistics in the research cited above suggests that there is no significant difference. The evidence of independent inspections by the HM Inspector of Prisons and his staff and the Prison Service's research analyses demonstrate generally higher levels of service in terms of domestic arrangements (eg. mealtimes and evening/weekend visits) and training, education and work for prisoners. Should these results cease to hold good, the Prison Service has the comparative data (both between and intra sectors) depth of contract management experience and contractual powers to take effective remedial action. Indeed, from Tranche 3 onwards, there are benchmarking conditions in the contract that would allow recompetition by the fourteenth year, rather than at the end of the 25-year contract if the price for custodial services under the PFI contract ceases to represent good value.

How is value for money achieved?

Noting that what public sector project managers said in response to a survey for the Arthur Andersen/Enterprise LSE study, I would suggest that cost savings reflect bidders' assumptions, in a highly competitive market, when responding to the output specification in the tender documents about their scope to produce an attractive bid and their subsequent ability to control costs by the following practices:

• Core (ie custodial) and ancillary staff conditions at market rates (output specification).

• More effective and flexible use of staff time, achieved both through design and operational management (output specification/innovation).

• Commercial pressure to avoid both overspecification in design and excessive space standards (output specification/innovation).

• Incentivisation for the contractor to achieve better whole life building costs (in terms of reduced maintenance expenditure, time accommodation has to be out of use for refurbishment, more efficient use of utilities) and synergy between design and operation. (long term contract/output specification/innovation).

• Similar incentivisation for contractors to effectively control over construction costs (in the context of historical average cost overruns identified by the NAO of 18% for new prisons) (private sector management skills).

Of the risks retained by the public sector, the most significant are demand risk and residual value. Historical trends and the ability to move prisoner populations over considerable distances, suggests that there is little risk both during the 25-year contractual term and thereafter, in terms of residual value after the reversion of ownership to the Prison Service, of such highly secure and flexible penal accommodation provided in PFI prisons not being required. (This is consistent with the principles for mitigating those risks when retained by the public sector reflected in Treasury Taskforce guidance about PFI accounting treatment.) While, under the contract, general inflation risk and prison specific legislation risk was taken by the Prison Service, the finance risks stemming from a general change in law was accepted by the contractors.

Further comfort is obtained by the circumstances under which these prices and risk allocations were negotiated. These reflect an unprecedented ten years of market development and acquisition of contractual competence by a single specialist unit within the Prison Service able to tap a globally competitive market:

• The internal market was deepened by the arrival of US prison contractors (also active in Australasia and subsequently South Africa) who brought new thinking, better management processes (especially self-audit) and a relatively deep pocket to cope with initially abortive bid costs (most competitions resulted in 3 or 4 bidders being unsuccessful). UK and US service providers whose bids were successful were not subordinate to construction only contractors, hence the bidding led by entities with a long-term perspective.

• The management only contracts, though of much shorter duration and much less complexity than PFI contracts, ensured that Prison Service staff began the PFI procurement process with some prior commercial experience and three tranches of PFI competition resulted in the consolidation of this in-house knowledge together with project management skills.

Strategic decisions were taken by the Prison Service that increased the chances of greater operational efficiency (cost neutral as far as the PSC is concerned but highly relevant to any public sector strategy irrespective of the source of finance), for example:
• by whenever possible, opting for larger capacity prisons (the cost of a secure perimeter, gate and control centre is largely fixed irrespective of size)
• by not going ahead with a long-planned separate therapeutic prison, with an intrinsically much higher cost per place, but replacing this with a self-contained unit within a large training prison complex with balancing economies of scale.

Moreover, with three successive PFI competitions, it was possible for the prison service management to move increasingly away from expectations that bidders' solutions would be similar to established design and operational specifications, especially by creating opportunities for bidders to test more radical ideas directly with senior operational management. This undoubtedly contributed significantly to potential savings as the programm developed.

Conclusions
This article indicates how the Arthur Andersen/LSE Enterprise study can be used as a starting point within the public sector for reviewing the results of PFI procurement. The PSC's produced by the Prison Service show the care that went into the value for money evaluations. Further grounds for confidence in the judgments made to date are the openness with which the Prison Service has been managed its PFI and contractual management programms, the congruent evidence from the management only contracts and international comparators, and the evidence of genuine competition from the significant movement in unit costs between the first and subsequent tranches of competition for directly comparable prisons.

End note
Prison names, competitions and tranches
In this article, I have used the names used in earlier publications to which readers may wish to refer. Further studies are likely to use the names adopted when the prisons began operations. This information is tabulated below.

BOX 7: THE FIRST SEVEN PFI PRISONS

Note: all were built to Category B security standards and HMP Altcourse adapted to Category A (ie. respectively high and maximum security standards)

References:
2Richard Harding Australian Institute of Criminology: Trends and Issues in Crime and Criminal Justice No 4, “Private Prisons in Australia: The Second Phase”, April 1998 I am grateful to Professor Harding for drawing my attention subsequently to a debate about the Florida recidivism study: A Comparative Recidivism Analysis of Releasees from Private and Public Prisons by Lonn Lanza-Kaduce, Karen F Parker and Charles W Thomas – 1999 Crime and Delinquency, Volume 45, pp. 28-47; Private Prisons, Criminological Research and Conflict of Interest: A Case Study by Gilbert Geis, Alan Mobley and David Shichor – 1999 Crime and Delinquency, Volume 45, pp 372-386 and The Devil in the Details: The Case against the Case Study of Private Prisons, Criminological Research and Conflict of Interest, by Lanza-Kaduce, Parker and Thomas - 2000 Crime and Delinquency, Volume 46, pp 92-136.

3Official Report of the House of Commons, 28 April 1999, Cols 161-162
4HMPS Annual Report and Accounts for 1997-98
5Mutual comfort about controlling disorder is provided by reciprocal arrangements for the nearest prisons to support each other irrespective of management. This was first tested in February 1997 when contractually managed prison staff were called with HMPS colleagues to a HMPS managed (long-term high security) prison.
6Treasury guidance on public sector comparators has been published in TTF Policy Statement No 2 (Public Sector Comparators and Value for Money), TTF Technical Note No 4 (How to appoint and work with a preferred bidder) and Technical Note No 5 (How to Construct a Public Sector Comparator). These are available on the Taskforce website at: www.treasury-projects-taskforce.gov.uk
7 In early 2000, the Prison Service announced its next competition for new prisons at Ashford, Middlesex and Peterborough, Cambridgeshire