Economists Jim and Margaret Cuthbert explore the Edinburgh schools scandal, which led to the closure of 17 PFI-built schools schools after the collapse of a wall at Oxgangs Primary, and find clear signs that the defect was rooted in the flawed logic of PFI
WITH structural flaws identified in all 17 schools in the Edinburgh PPP1 schools project, and with pupils’ education having been subject to massive disruption, clearly something went badly wrong with the execution of this project. The cause of the technical failings with this particular project will no doubt be revealed in due course: and it would be idle to speculate on these in advance of the findings of the Cole enquiry which is currently underway. But what can be seen at this stage, from an examination of the draft Full Business Case, (FBC), for the project, is a flaw in the general logic of PFI as regards the transfer of risk from the public to the private sector.
In 2001, when the Edinburgh scheme was being decided, the key step in the decision to go ahead with any PFI scheme was the comparison with what was known as the Public Sector Comparator, (PSC). A PFI scheme could only be justified if the expected costs under PFI could be shown to be less than the costs the public sector would expect to bear if it went ahead with a similar scheme under traditional public sector procurement. This comparison is greatly complicated by the question of risk. Under PFI, certain risks which would otherwise be borne by the authority would be borne by the consortium undertaking the PFI: for a fair comparison, therefore, the costs of such risks should be added in to the PSC.
It became clear, early in the history of PFI, that the whole area of assessing risk transfer was very problematic. First of all, for a given risk, two steps have to be carried out – first of all, calculating what the cost to the public sector would be if the public sector itself were doing the build and the risk event happened: and secondly, estimating the probability of the risk event happening. The expected contribution to transferred risk would then be the product of the monetary cost times the probability. But both of these quantities are very subjective: and the suspicion arose that, in many PFI projects, the value of transferred risk being added to the PSC was being chosen to make the PFI appear just cheaper than the PSC – so justifying a go-ahead with the PFI option.
A second difficulty arose because of common risks: if a risk event, and its costs to the public sector, are similar both under the PSC and the PFI, then it represents a common cost under both options: and it should not affect the choice between the PFI and the PSC. However, the suspicion arose that, in several PFI projects, the extent of risk transfer to the PFI was being over-estimated: and that risks which were in effect common were being added in solely on the PSC side – so biasing the comparison.
“What if certain risks are actually transferred under PFI from the private sector to the public sector: in other words, what if certain risks for the public sector are actually greater under PFI than the corresponding risk the public sector would have borne in the absence of PFI?”
Both of these problems became clear fairly early in the history of PFI – and, while they may not have been adequately addressed, they were at least identified. But there is a third conceptual problem about risk transfer. What if certain risks are actually transferred under PFI from the private sector to the public sector: in other words, what if certain risks for the public sector are actually greater under PFI than the corresponding risk the public sector would have borne in the absence of PFI?
The Edinburgh schools PPP1 project illustrates how this problem could arise. The scheme was massive, with the FBC setting out plans for 13 new-build sites, and the refurbishment of others, together with the maintenance of these schools over the 30-year life of the project, and the provision of all support services, such as catering, cleaning and facilities management for the life of the project. Examination of the FBC indicates that the net present value of the total annual unitary charge payments to be made by the authority was £122.6 million. The net present value of the cost of the PSC was assessed at £129.9 million: so the PFI scheme did indeed come in as the cheaper option, by £7.3 million. However, £13.4 million of the net present value of the PSC was the expected cost of the risk transferred to the private sector operating consortium under the PFI: so the question of risk transfer was indeed critical in the decision to go ahead with this particular PFI project.
The FBC provides a detailed breakdown of how this estimate of transferred risk was built up. Given the defects which have recently been discovered, the element of risk transfer relating to latent structural building defects is particularly relevant. In the FBC, the possible cost of latent defects under the PSC was estimated at £1.5 million: and it was assumed the probability of such a defect occurring was 2.09%. So, (after discounting to prices at the start of the project), the expected contribution to the risk transferred to the private sector was just over £20,000.
What was not included in the risk transfer calculation was the cost, to the authority and to pupils, if a major latent defect were to occur. If the chances, and nature, of a major latent defect were the same under public sector provision as they would be under PFI, then such costs could be regarded as common – and could be justifiably ignored in the PFI comparison with the PSC. But, in fact, the chances are very unlikely to be the same – for a number of reasons.
“To maximise the chances of getting the contract “off the books”, and therefore to avoid controls on government capital expenditure, PFI projects were designed to be complex – sweeping up both the provision of the capital asset, and associated services. And with complexity comes size.”
One of these reasons relates to the scale of typical PFI projects. PFI projects tend to be very large – as indeed was the case with the Edinburgh schools project. Partly this relates to the very large fixed costs associated with setting up a PFI deal: the documentation associated with one contract we examined exceeded 10 thousand pages, so, naturally, there is a desire to minimise the burden of fixed costs by spreading them over large projects. But PFI contracts are also large because of a specific and perverse characteristic of PFI itself. To maximise the chances of getting the contract “off the books”, and therefore to avoid controls on government capital expenditure, PFI projects were designed to be complex – sweeping up both the provision of the capital asset, and associated services. And with complexity comes size.
The size and complexity of PFI deals biases the comparison with what would happen under a public sector approach. If Edinburgh had been adopting a public sector approach to modernising its school stock, it would have been highly unlikely to embark upon such a large single project as the PFI deal: or, indeed, to embark on a scheme with such a large component of new build. (Indeed, in the FBC for the Edinburgh schools scheme, the PFI project involves the rebuild of 2 secondary and 1 special school which, under the PSC, would have been refurbished.) Instead, the public sector would have been much more likely to embark on an incremental approach. Under such an approach, the occurrence of a common defect affecting all parts of the scheme would have been much less likely: and the chances of identifying such a defect before it was incorporated in all parts of the scheme would have been much greater.
“Essentially, a large PFI scheme is subject to an element of “eggs in one basket” risk which a standard public sector approach would not suffer from.”
Essentially, a large PFI scheme is subject to an element of “eggs in one basket” risk which a standard public sector approach would not suffer from. This doesn’t come out in the comparison between the PSC and the PFI because the PSC has, in effect, been distorted by the decision to examine PFI options. But the public sector is actually bearing extra “eggs in one basket” risk under PFI: and this should have been taken into account in an accurate assessment of the costs and benefits of PFI. And since the size of this particular risk increases with the size of the scheme, proper recognition of this risk would have acted as a powerful disincentive to overly large schemes.
Another important reason why risk may transfer from private to public sector under PFI is because the public sector is limited in the extent it can check on detailed progress during the construction phase of a PFI project. As guidance issued by the Treasury on PFI contracts states, “it will not be appropriate for the authority to adopt the type of overseeing role it might traditionally expect to have when procuring stand-alone construction or development services”: (source: Treasury guidance on standardising PFI contracts, April 2004). The rationale for this is understandable: if the public sector gets too involved in checking work during the construction phase, this will cloud the allocation of responsibility should things go wrong. But the effect is that the public sector will have to rely on the agreed mechanisms for checking the quality of construction which are built into the contract: and this may well increase the risk of defects going undetected relative to what would happen under traditional public sector procurement.
Overall, the effect is that certain risks for the public sector may well be greater under PFI than under a conventional public sector approach. The logic of this is that the expected value of such risks should be added in on the PFI side when the cost benefit comparison is being made with a public sector approach. The failure to recognise this possibility, and to make an appropriate adjustment to PFI costs, represents a major flaw in the methodology of PFI, which should be added to the list of known problems with PFI. The effect of this particular flaw is that more PFI schemes will have been undertaken than would have been embarked on under a proper assessment of risk.
Visit Jim and Margaret Cuthbert’s website with an extensive listing of PFI-related work at jamcuthbert.co.uk