The principal problem is that PFI has been very much more expensive than public financing because of what in practice has been a higher cost of capital for PFI projects.
This might seem odd in the light of the written evidence of the National Audit Office. In the NAO's written evidence to this Committee (see NAO October 2009, paragraphs 2.38, 2.39 and 2.40), it is stated that, up to the credit crunch in September 2008, private finance for PFI projects was available at between 1 and 2 percentage points above the Gilt rate. Over the 10 years between 1999 and 2008 inclusive the average Gilt rate (the interest rate on government bonds) was 4.8 % pa (from an updating of table 4.2 in Edwards C, June 2009). This should imply that private finance would be available at an annual rate of between 5.8% and 6.8%. This is the rate at which senior debt would be available for PFI projects and since senior debt has generally accounted for more than 90% of total PFI capital, it might be expected that the average cost of capital for PFI projects would be close to the top of this range.
Unfortunately there does not seem to be any evidence provided by the National Audit Office (NAO) of the actual costs of capital on PFI projects. If this is the case (that such estimates do not exist), it might be sensible for the Committee to ask the NAO to provide such estimates.
In the meantime what is the evidence available from other sources?
In general terms, the cost of capital (senior debt, subordinate debt and equity) on PFI projects seems to have been at least one and a half times the government's cost of capital (see Edwards C, June 2009, 28). This general conclusion is backed up by the case studies of Cuthbert and Cuthbert, 2008 and by my own case study of the Norfolk and Norwich University Hospital (NNUH).
The largest of the three PFI hospitals looked at in detail by Cuthbert and Cuthbert was the New Royal Infirmary in Edinburgh (NRIE). The capital raised by Consort Healthcare for the NRIE was £189 million and the average cost of this capital was 11.9% pa (Cuthbert and Cuthbert 2008, 5).
I have looked in detail at the NNUH or at least in as much detail as permitted by the limited published information and by the Freedom of Information Act (restricted as it is by "commercial confidentiality"). The capital cost of the NNUH (including development and start-up costs) was £195 million and I estimate the nominal cost of capital to be 13.0% pa.
These two large PFI hospitals were both constructed over the years from 1998 to 2001 inclusive. Over these four years the average yield on government bonds was 5.2% pa. Therefore the PFI cost of capital is more than twice as high as the government's cost of capital in both cases (the ratio is 2.3 in the case of the NRIE and 2.5 in the case of the NNUH).
These ratios need to be adjusted downwards if tax is paid to the government on the interest and profit received by the PFI companies. This information is not available to me. The paper by Cuthbert and Cuthbert gives the Corporation Tax generated by the NRIE project as equal to £167 million or 22% of the rent payable but they state that this might be overstated since the ownership of several PFI schemes is registered in tax havens abroad and little or no tax may be paid (see Cuthbert and Cuthbert 2008, pages 6, 10). I do not have information on tax paid on interest in the case of the NNUH.
Clearly though, if the tax paid is 22% of the rent, then the cost of capital for the NRIE PFI project is reduced to 11.9% minus 22% % 9.3% pa and for the NNUH it would be 13.0% minus 22% = 10.1% pa.
However even after these tax deductions, the after-tax ratios of the PFI cost of capital to the government's cost of capital (of 5.2% pa) are 1.8 for the NRIE and 1.9 for the NNUH.
Given a ratio of 1.8, a PFI hospital project is highly unlikely to provide value for money (VFM) to the taxpayer. For the taxpayer to get VFM from the project, the PFI company would need to build a hospital for just over half the construction cost of one built through public financing. This is highly unlikely.
The upshot of this is that the cost of a PFI hospital (even if tax paid back on the interest and profits is deducted) will be very much higher than a publicly-financed one. The NNUH is a good example. In 1995-96-that is before the PFI contract had been signed-the Hospital Trust was paying an annual rent to the NHS of at least £8.4 million. This was sufficient (at the government's cost of capital and over the PFI contract period up to 2037) to meet the cost of the new hospital (£195 million).
Now move 10 years forward to the year 2005. The PFI hospital had been built and the rent paid in that year to the PFI company (Octagon Healthcare) was £27 million. If the old rent had continued to be paid, in 2005 prices it would have been equivalent to £10.9 million. The PFI rent was two and half times as much as the "rent cost" payable under a publicly-financed contract.
To put this another way, the additional cost of the NNUH (in 1995 present value terms, using the government's cost of capital and over the life of the project) will be £267 million. As a result of the PFI contract, the government is paying well over twice the construction cost of the new hospital.