Profiteering from Prisons

On the global evidence, private prisons are no more efficient, and are no less expensive for the taxpayer

by Rory MacKinnon

The government plans to open up prison construction and management to the private sector in order to drive innovation, and cut expenditure. But does anyone know the real cost?

When National won the 2008 general election by a landslide, the pundits were quick to attribute their popularity to two things: a promise to get tough on crime and an equally fervent promise to cut taxes by reducing government expenditure. But nearly two years on, the contradictory nature of these goals is increasingly obvious: it costs a lot to keep someone captive in a first-world nation, and the ideological commitment of recent governments to prison-as-punishment is doing more to create a Nanny State than any number of banned lightbulbs.

Corrections spending has more than doubled over the last ten years, with custodial services alone now costing taxpayers over $890 million a year – more than the entire budget for Youth Affairs, Maori Affairs, Pacific Island Affairs, foreign aid and senior citizens combined. With an average 8,000 people behind bars at any one time, it’s also one of the highest incarceration rates in the OECD – and according to Treasury predictions that incarceration rate is likely to rise another 40 percent over the next eight years. Even the most hard-hearted social conservative has to admit our current system of “lock’em up and throw away the cash” can’t continue. So what to do then about our staggering prison population?

If you’re Bill English, the answer is to pay someone else to handle it: either by contracting out the running of existing prisons, such as Auckland Central Remand Prison in Mt Eden, or, by building a brand-new, 1000-bed prison in the South Auckland suburb of Wiri — designed, financed, maintained and managed entirely by private companies in the belief that this will provide better services, and better value to taxpayers:

International experience suggests that building a new prison at Wiri using a public-private partnership will offer savings of between 10 and 20 percent over conventional methods over the 25 to 35-year life of the proposed contract.

“Those are substantial gains that will leave more money available for other vital infrastructure priorities like schools, hospitals and roads.

“This Government has provided a big infrastructure funding boost, but we are committed to spending that money wisely in all areas, including Corrections.

But is that money really spent wisely? The claimed efficiency of public-private partnerships, or PPPs, is almost an article of faith these days for economists and politicos alike – and in New Zealand, even those opposed to the privatisation of prisons often take the economic efficiency argument as read. Thus, they tend to focus on the broader social principles involved. As a consequence – in New Zealand as in many other countries – the debate about prison privatisation has largely concentrated on the ethical and social implications, and on the unseemliness of profiteering from what Corrections Minister Judith Collins once wryly described as a “growing industry”.

Yet very little research evidence exists on whether private prisons actually do deliver better services for less money. In practice, the commercially confidential terms of public-private partnership generally prevent this kind of research from being conducted. With so much at stake here, the rationales and the results seen overseas are vital in determining whether the push to open New Zealand’s prisons to the private sector is anything more than blind ideology at work.

Of course, most supporters of public-private partnership in New Zealand don’t think of themselves as ideologues, and understandably so. Debt is an unpleasant fact of life at the best of times; the thought of a government willingly taking on millions of dollars’ worth in the wake of a global recession seems positively reckless. But this rationalisation overlooks a crucial difference in the way lenders treat sovereign debt: governments basically don’t go bankrupt. Of course it’s possible – Iceland’s 2008 meltdown springs to mind; Argentina and Russia around the turn of the century – but these are exceedingly rare. Governments are still far more reliable lendees than any private institution, and consequently they can raise the money for public works far more easily and with lower rates of interest. This is not idle speculation; this is the entire premise behind the Official Cash Rate.

But that debt doesn’t magically disappear under a public-private partnership. The government simply outsources it to the contractor; in essence, paying the company to borrow money for them. The private operator still has to borrow the funds for the project – and at higher rates of interest – but then bills the government annually for the cost of repayments, plus interest, plus the contractor’s profit margin. Logically the government must end up spending more than it would if it directly financed the project, but by re-categorising the debt as an ongoing operational expense it creates the illusion of a smaller deficit and ensures a more favourable credit rating. So while this creative accounting appears to offer a reduced public debt, over time it actually costs the taxpayer far more.

In all fairness, there is an audacious counter-argument to all this: namely, that private contractors operate with such ruthless efficiency, they can reduce the principal sought to just a fraction of the amount needed by those bumbling bureaucrats in Wellington — a figure so lean it can accommodate a commercial interest rate and a tidy profit margin and still come out under the public sector’s budget. This proposition dovetails neatly with the mantra of “accountability”: the idea that an unerring focus on the bottom line drives private operators to use public funds more efficiently and with more scrupulous bookkeeping than the state would.

But again, little independent research exists to support this claim, and if the United Kingdom’s experience is any indication, privatisation tends to obscure the use of public funds – not expose it.

David Price would know: as a senior lecturer and researcher with the Centre For International Public Health Policy at the University of Edinburgh, he has had plenty of experience with these partnerships, known in the UK as ‘private finance initiatives’ or PFI. In 2005 he and his colleagues at the Centre contacted the Treasury to request the data from its first ever study of time and cost overruns in conventional and PFI procurement – a study that Treasury had already cited in a 2003 policy document and had promised to publish on its website six months ago. Price requested this information under the Freedom of Information Act, a piece of legislation similar to our own Official Information Act.

But there was a catch. Like our OIA, the Freedom of Information Acts of Scotland, England and Wales only cover information which is already possessed by the civil service — meaning private operators have absolutely no obligations whatsoever. Some basic operational information such as staffing levels has been made available under a “general right of access”, but the Acts still curtail that right where the revealing of a trade secret is likely to prejudice a person’s commercial interests or “the economic interests of the United Kingdom”. Unfortunately for Price, almost any information about a business can be construed as commercially sensitive information – meaning that in many cases, vital information about expenditure, performance and procedures has been withheld from independent researchers, members of the public, and even Members of Parliament who want to know whether these private operators are indeed delivering on their promises.

Such was Price’s plight. The Treasury replied that disclosure might “be detrimental to the commercial interests of specific PFI contractors or the financial interests of procuring authorities”, and only after some negotiation did Price obtain a spreadsheet of raw survey results. Of course, without the accompanying report there was no way to assess the study’s methodology — making the raw data completely useless.

It’s not just Price’s problem either. As recently as 2008, the Scottish Information Commissioner explicitly upheld a decision by the Scottish Prison Service to withhold Cabinet papers comparing the costs of public-private partnership prisons with those procured in the public sector. While the Prison Service acknowledged there was a public interest in releasing the papers, it argued the Service was already pursuing procurement and “it would be highly disadvantageous in discussing value for money figures associated with these projects if the alternative costs were to be made public”. The Commissioner agreed: there was a “strong case” for releasing as much information on prison procurement as possible – but in this instance ‘as much as possible’ amounted to three paragraphs of background information.

This lack of transparency is unavoidable in a public-private partnership due to the way the procurement process works. When the government offers a new contract, competitors tender their bids in strict secrecy – only those on the procurement team ever get to see those bids. Yet even once a provider is picked and a contract signed, all bids – winning and losing alike – remain secret. The upshot is that ministerially-appointed procurement panels pick bidders without ever revealing their criteria and alternatives to anyone but the ministers involved – and sometimes not even then. The problem this prevents for an informed public is obvious.

Price says this kind of chicanery is endemic and unavoidable so long as laws around official information are interpreted to exclude full contract disclosure — but such disclosure is hardly in the government’s interests, and provisions against it may even be written into the contract itself. Whatever the reasons, it all adds up to one massive fiscal blind spot.

“In all but a few cases commercial confidentiality has been used to withhold the financial model in PFI contracts. The model shows contractors’ returns. This means that 13-15 years into a policy – with a bill approaching 60 billion pounds – it is still not possible to say how much of the cost is due to profits or identify the rate of profit.”

But the secrecy that surrounded these deals only spurred Price and his fellow researchers at the centre, Allyson Pollock and Stewart Player, to investigate further. In 2007 the trio published “An Examination of the UK Treasury’s Evidence Base for Cost and Time Overrun Data in UK Value-for-Money Policy and Appraisal”; a paper which despite its dry-as-dust title held shocking revelations about the entire basis for the UK’s private procurement policy. From the abstract:

We review[ed] the studies cited by the Treasury in support of this claim and [found] that only one purports to compare PFI with traditional procurement. The results of this single study are uninterpretable because of selection bias, small sample size (only 11 out of 451 PFI projects are included) and fundamental flaws in the analysis. There is thus no evidence to support the Treasury cost and time overrun claims of improved efficiency in PFI. We conclude that Treasury appraisal guidance, the ‘Green Book’ which compares PFI with other methods of procurement, is not evidence based but biased to favour PFI.

Pollock then penned a column in the Guardian to explain in plain English the full extent of the scandal: the basis for 54 billion pounds’ worth of investment and over 200 billion in debt repayments boiled down to three – three – case studies.

The only report that contains any comparative data was commissioned from a consultancy and engineering firm called Mott Macdonald. This study is very curious. Full cost and time overrun details are provided for just three PFI schemes, although at the time of the study 451 PFI deals had been completed. Mott MacDonald claimed that it had difficulty getting data on other projects. The report then compares these three with 39 public-sector schemes, although very little public procurement was going on at the time. What is more, of the 39 public-sector schemes, 20 are “non standard” – complex and difficult projects – whereas the three PFI projects are all standard. You don’t have to be a statistical genius to conclude that this is hardly comparing like with like.

Mott MacDonald also used different starting points when comparing cases. They counted cost and time overruns for PFI projects from when the case was signed off, but they started counting at a much earlier stage for the public-sector projects. Yet Mott MacDonald well knew, as consultants to the PFI industry, that one of the most striking aspects of PFI was that costs escalated between the initial tendering and the contract being signed off. Take the Paddington hospital PFI scheme for example: the proposed cost rose in real terms from £411m in 2000 to £894m at the time of the scheme’s collapse in May 2005.

Yet the outcries and inquiries never came: Price suspects the paper may have prompted the National Auditory Office to tone down the claims in its PFI performance reports, but he says he cannot recall any response from the government. The silence speaks volumes.

At last count, the governments of Scotland and England and Wales have collectively opened 11 private prisons – housing 20 and 11 percent of their respective prison populations – while construction of public prisons has all but ceased. Nine of these new private prisons operate under a PPP model with a 25-year contract, just like the one proposed at Wiri. But if the overseas experience is any indicator, these partnerships prove far riskier for those purchasing the services than the operators who profit from it.

Take the cautionary tale of South Africa’s Mangaung and Kutama Sinthumule prisons, the country’s first foray into prison privatisation. Opened in 2001 and 2002 under a PPP model, the prisons soon became a headache for the Zuma government with exorbitant operating costs of 300 million rand each per year — nearly double the cost estimated for a public prison of the same size. These companies stood to draw massive returns of between 26% and 29% over the course of their 25-year contracts, and when the government sought to refinance the projects in 2006, the operators simply refused to negotiate.

Likewise the government of Costa Rica faced a US$20 million lawsuit from Management & Training Corporation in 2006 when it scrapped plans to contract out a new 1200 bed prison at Pococi for $73 million. The project had stalled since 2001 due to an Ombudsman’s legal challenge that contracting out correctional services would breach the country’s constitution. Finally in 2005 the courts ruled in the Ombudsman’s favour, making the Pococi deal impossible. Yet although the government’s hands were clearly tied, the MTC sued for US$20 million in compensation, ultimately settling for US$4 million. It is worth noting that the legal wrangle over Pococi did not stop construction of publicly-financed prisons, with 2600 new beds built in the intervening period at a cost of $US10 million. The Pococi deal by comparison would have housed fewer than half that number of prisoners, at more than seven times the cost.

Closer to home, the Victorian government of Australia in 2000 decided to break the contract on its first privately-financed prison, the Deer Park Metropolitan Women’s Correctional Centre – now known as the Dame Phyllis Frost Centre – after four years of poor operational performance. The buyout cost the state of Victoria AU$20.2 million in addition to the unexpected costs associated with essentially opening a new prison. Due to commercial confidentiality it is unclear exactly how much the original Deer Park contract was worth, but Victoria’s Corrections Minister at the time, Andre Haermeyer, was emphatic: “The cost to government of $20.2M for ending the agreement and buying the facility is less than the value of ongoing payments for use of the prison had the contract continued.” The numbers might be hazy, but the message is clear: public-private partnerships for building and/or running prisons should not be entered into lightly, if at all.

Such examples are conspicuously absent from the current government’s “international evidence”. When I asked Bill English’s office for evidence of savings under a PPP model, I was referred to three reports from the UK, a single 1997 article from the United States’ Quarterly Journal of Economics an assurance without citation that “the Scottish Prison Service has claimed savings of 32 percent” and a brief statement that analysis of the Australian experience carried out for Corrections had found savings of 11 percent in New South Wales and up to 32 percent in Victoria.

At the time this article went to print neither Corrections nor English’s office had supplied a source for the Australian figures, despite an Official Information request to Corrections in June requesting any documents presented to the minister in the past eighteen months which compared public-private partnerships with a public sector comparator.

But the massive savings claimed in Victoria are at odds with the tone of the Victoria Auditor-General’s own report on prison management under public-private partnerships, released just six weeks ago. Its conclusions are remarkably blunt:

Shortcomings in the management and administration of the prison accommodation services contracts by DOJ mean it is not possible to demonstrate whether the state has received value-for-money.

DOJ has used an overly complex contract governance structure for the pre- and post-2001 contracts with multiple reporting lines and duplication of responsibility. DOJ has however now decided to review and simplify this governance structure.

DOJ’s contract management systems are inadequate and contract administration manuals and risk management plans are missing. In addition, DOJ has poorly documented its decisions and contract variations. The complexity of these systems inhibits the states ability to obtain value-for-money from these contracts.

As for the journal citation, it comes courtesy of Harvard’s economics department, sourced from a single paragraph on page 21:

Private prisons are perhaps 10 percent cheaper, per prisoner, than public prisons. The major reason for the lower costs appears to be the roughly 15 percent wage premium for public guards over private guards {Donahue 1988}. Part of the labor cost difference is that private contractors do not pay the public union wage premium; another part is that they hire lower quality workers. Since labor accounts for two-thirds of the incarceration costs, the differences in labor costs can roughly account for the 10 percent cost saving from private prisons.

In other words, a rule of thumb based on another rule of thumb that presupposes a poorer quality of service: hardly inspiring stuff.

The reports from the UK appear, at first, to be considerably more substantial: a 2003 National Audit Office report on the operational performance of privately-financed prisons, a 2003 House of Commons report by the Committee of Public Accounts on the same, and a 2005 report on market-based approaches in general from the Department of Trade and Industry. According to English, these documents provide ample evidence of savings of between 10-20 percent and private sector innovation. So, what do they say?

The NAO report makes no claims of savings. In fact, it explicitly refuses to make any cost comparisons. Just seven pages in:

Comparing the performance and cost of PFI [Private Finance Initiative] prisons against publicly-managed prisons is difficult because of the different ways they are funded, the variable proportion of prisoners of different categories, the variations in design, age and function of the prisons, the ways they are measured and the different targets they are set. The difference in capital financing between the PFI prisons and other prisons adds another level of complexity when seeking to compare costs.

Then again in Appendix 3, “The Problems Of Comparing PFI And Public Prisons”:

There have been a number of recent attempts to analyse the components of the unitary charge paid to PFI prisons and compare it with the costs of public prisons. Academic research has highlighted the inherent difficulties of evaluating the comparative costs of prisons and this is a problem in other countries.

Even when limiting their focus to operational performance, the report’s authors are loathe to compare public with private:

Although there are features common to the two systems there is currently a lack of consistent data, both in terms of quality and in what it measures. It is not therefore possible to make a meaningful comparison of PFI and public prisons by simply comparing performance against KPTs [Key Performance Targets] even when those figures are available.

The report is riddled with such qualifications: differences in the contracts negotiated for each PFI prison; grey areas in how performance is defined, fines for non-compliance being an inaccurate indicator of performance, differences in the way prison controllers go about their work and so on. Even so, the report does say that the operational performance of these PFI prisons has been “mixed.” In an overall performance review of 21 prisons, four PFIs (Forest Bank, Lowdham Grange, Altcourse and Parc) were ranked amongst the top 10.

But the most poorly-performing PFI, Ashfield, was one of the worst prisons overall. The prison reported 279 proven assaults in the previous year – representing 74 percent of the prison population – and a 200% increase in “control and restraint” incidents on the year before. Both the Chief Inspector of Prisons and Youth Justice Board reported inmates too afraid to leave their cells, staff who did not wear name badges or other identification, dirty cells and bedding and few controls over prisoners’ whereabouts. Consequently, Ashfield was the only prison in the entire study to receive an “unacceptable” security rating.

It seems therefore, that private sector performance is something of a crapshoot. In the authors’ own words, “the use of the PFI is neither a guarantee of success nor the cause of inevitable failure…. A general verdict that the PFI is either good or bad in the case of prisons, or more generally, cannot be justified.”

Yet even that last clause is debatable. True, the report notes that PFI prisons generally performed better in areas related to decency, such as purposeful activity and respect shown to prisoners. Yet at each of the seven private prisons studied, staff turnover was at least double that of state prisons (and often as much as four times that), staffing levels were significantly lower and basic pay rates for private prison officers were on average £4000 less than in the public sector. Those prisons also had a significantly poorer safety record, with five reporting prisoner assault rates in the upper quartile for their category of prison.

The report concludes that “there is a difficult balance to be struck between the two” — but safety and security are not to be treated as trade-offs. They are core functions of the corrections sector. It is also worth noting that just three years after the NAO report was published, the Chief Inspector filed a damning annual report in which she felt “some concern that the four private adult prisons reported on had more negative than positive assessment, and only one out of four was assessed as performing satisfactorily on safety.”

The House of Commons report also makes no mention of cost comparisons or savings under a PPP model, although it does re-iterate many of the performance issues outlined in the NAO report. It also recommends “as much emphasis on the sustained delivery of an acceptable service as there is on contract price.” It again invokes the spectre of Ashfield: opened in 1999 by Premier Prison Services (and now part of the international Serco Group) this 300-bed young offenders’ facility was so critically and continually understaffed that the government had to intervene after the aforementioned inspectors’ reports and a full-scale riot in 2002. It is worth noting that Ashfield was returned to private management five months later and had made “significant improvement” according to the Chief Inspector’s report the following year, but the Committee nonetheless hammers its point home:

The Prison Service should not shy away from terminating prison contracts….Although the contractor suffered substantial financial penalties, there were also knock-on costs to the public sector as young offenders had to be moved to other, already overcrowded, institutions.

Yet as the case studies in South Africa, Costa Rica and Australia have shown, terminating these contracts carry their own heavy costs. Two reports in, and the evidence for any savings at all – let alone of 10-20%, as English has claimed – under a public-private partnership is non-existent.

Superficially, the UK Trade & Industry report also – at first – appears to support Bill English’s projections. It cites a British business lobby group which estimated savings in operating costs over an 11-year period of 10-15 percent, and construction costs reduced by 20 percent. But the report’s authors note that there is “some question” over whether the methodology used – cost per prisoner – is appropriate. The marginal cost of an extra prisoner is small, and this can make overcrowded prisons appear more cost-efficient than those that are actually operating at capacity. In fact, they write, a 2000 study by the Home Office actually found that if calculated on the basis of cost per available place, the gap between private and publicly financed prisons narrowed over time “to the extent that there are no longer any savings to be made from private sector involvement.”

The report concludes that the procurement process for prisons in the UK has been apparently successful, but not without a “heavy resource commitment, in both up-front and ongoing costs”: a quarter of a million pounds to pay the procurement team; another half-million or more for legal advice and insurance, plus the cost of ongoing monitoring and compliance. There is some potential for net savings through procurement, it suggests, but you get what you pay for.

Or perhaps not. The Trade & Industry report mentions in passing a 2001 study by the US Bureau of Justice Assistance which found quite the opposite; that “rather than the projected 20 per cent savings, the average saving from privatisation was only about 1 per cent, and most of this was achieved through lower labour costs.” The US Bureau continues in even blunter terms :

As pointed out by Gaes and colleagues (1998), a coherent theory of why privately operated prisons would outperform public facilities has yet to emerge. Instead, one could argue that the private sector has simply drawn upon the methods used by the public sector with respect to inmate management and staffing and only attempted to reduce the costs associated with that model. In effect, the private sector may be applying a more efficient model that is essentially mimicking the public sector. This is most obvious in facilities at which former public-sector managers have been recruited by the private sector to manage the private facilities. Should this approach be considered by policymakers, the future of privatization may be very limited as the public sector in turn copies the private sector’s methods. For these reasons it may be concluded that there are no data to support the contention that privately operated facilities offer cost savings over publicly managed facilities.

It then adds :

There is no evidence showing that private prisons will have a dramatic impact on how prisons operate. The promises of 20-percent savings in operational costs have simply not materialized.

There are also some familiar figures in the US Bureau’s report, with assaults on inmates increasing by 38 percent under private management while staff shrink by 15 percent. Yet the UK Trade & Industry team blithely push these findings aside: “The UK has avoided many of the problems seen in the US, as “quality has been an important part of the process for assessing tenders”. (Presumably, the deterioration of safety records and staffing levels at private prisons in the UK as in America is a mere statistical anomaly.)

Yet even in the face of such damning evidence, the authors of the US Bureau report do not reject the private sector entirely. They insist that the private sector does have a vital function in the corrections system, but that function is the testing of corrections policy – things like education and counseling services – not running the prisons themselves. The final line of the report is unequivocal: “If nothing else, the private sector has shown that it is as equally capable of mis-managing prisons as the public sector.”

Has Bill English even read the US Bureau report? It is possible. The Greens have been citing it ever since the Wiri facility was first announced, and English has said the Government will revert to public procurement if the bids are no better than existing costs. Meanwhile Corrections Minister Judith Collins has given lip service at least to performance issues. Cost is a factor, she has said, but “it is more important that we have effective and safe prisons than cheap ones.”

Such caution would be advisable. The international evidence has shown that not only is the cost of private prisons counted in dollars, but in lives as well. Meanwhile the companies which seek out these contracts can be relied on to lobby aggressively to expand their market – and they will fight just as hard to retain that market share, regardless of the public good. Collins’ quip about Corrections as a growth industry is true in more ways than one.

To borrow a phrase from our Prime Minister, New Zealand would do well to kick the tyres on this one for a good while longer.