The Andersen report was produced in 2000. However in 2003 the UK Treasury in its Green Book lowered its recommended discount rate to 3.5% pa. What it did was to account for risk separately and not include it in the interest rate (Treasury 2003, 26)15. It is worth looking at the difference this makes using the Andersen model.
It needs to be noted that when the Treasury reduced its discount rate to 3.5% pa, it not only separated out the risk of over-shooting the cost ('optimism bias') but it also separated out the effects on public finance of taxes paid by the private sector16. As far as the latter is concerned, the Treasury expected the size of the adjustment to be between 2% and 10% of cost for the majority of projects across the whole of government (see Department of Health, February 3 2003, page 5). However it is worth noting that for twelve PFI schemes looked at by Shaoul et al 2007, the tax payable averaged only 2.7% of their income (calculated from table 5 of Shaoul et al 2007) and even this would seem to be an over-estimate of the tax actually paid because of the companies' ability to defer payment (see Shaoul et al 2007, 13)17. Nevertheless here 2.7% is assumed to be the amount that has to be deducted from the PFI cost to allow for the 'tax effect' and again we are being kind to the PFI case.
The effect of the lowering of the discount rate is shown in table 4.4 below. The present value of the construction cost for the PFI version is now £181.6 mn compared to £108.9 mn for the PSC. We now need to knock the 'tax effect' off the PFI cost so that we get a net figure for the PFI capital cost of £181.6 mn less 2.7% = £176.7 mn. The excess cost of the PFI alternative is now £176.7 mn less £108.9 mn = £67.8 mn. This is higher than the previous figure of £56.4 mn, but lower in terms of the percentage of the PSC. The percentage is now 62% of the PSC compared to an 'excess' of 67% when using the higher 6% pa discount rate.
| Table 4.4; the Andersen model; a comparison between the capital costs to the public sector of PFI with a PSC; | |||||||||||||||
| using annual real interest rates of 4.3% for the PSC and 10% for the PFI alternative; | |||||||||||||||
| and using a discount rate of 3.5% pa | |||||||||||||||
| The model | Present | Public sector comparator…………….. | PFI example……… | ||||||||||||
| Year | Capital & | value (PV) | PV of | Capital cost | Capital cost | ||||||||||
| operating | factors | operating | Repayments | Capital | Interest | PVs at 3.5% pa ---- | Repayments | Capital | Interest at | PVs at 3.5%pa
| |||||
| cost | at 3.5%pa | costs | at 4.3%pa | balance | at 4.3%pa | Repayments | Interest | at 10% pa | balance | 10% pa | Repayments | Interest | |||
| (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | (£mn) | ||||
| 0 | 100 | 1.00 | 100.00 | 100 | 100.0 | ||||||||||
| 1 | 10 | 0.97 | 9.66 | 2.31 | 97.7 | 4.30 | 2.23 | 4.15 | 1.02 | 99.0 | 10.0 | 1.0 | 9.7 | ||
| 2 | 10 | 0.93 | 9.34 | 2.41 | 95.3 | 4.20 | 2.25 | 3.92 | 1.12 | 97.9 | 9.9 | 1.0 | 9.2 | ||
| 0003 | 10 | 0.90 | 9.02 | 2.51 | 92.8 | 4.10 | 2.26 | 3.70 | 1.23 | 96.6 | 9.79 | 1.1 | 8.8 | ||
| 4 | 10 | 0.87 | 8.71 | 2.62 | 90.2 | 3.99 | 2.28 | 3.48 | 1.35 | 95.3 | 9.66 | 1.2 | 8.4 | ||
| 5 | 10 | 0.84 | 8.42 | 2.73 | 87.4 | 3.88 | 2.30 | 3.26 | 1.49 | 93.8 | 9.53 | 1.3 | 8.0 | ||
| 6 | 10 | 0.81 | 8.14 | 2.85 | 84.6 | 3.76 | 2.32 | 3.06 | 1.64 | 92.2 | 9.38 | 1.3 | 7.6 | ||
| 7 | 10 | 0.79 | 7.86 | 2.97 | 81.6 | 3.64 | 2.33 | 2.86 | 1.80 | 90.4 | 9.22 | 1.4 | 7.2 | ||
| 8 | 10 | 0.76 | 7.59 | 3.10 | 78.5 | 3.51 | 2.35 | 2.67 | 1.98 | 88.4 | 9.04 | 1.5 | 6.9 | ||
| 9 | 10 | 0.73 | 7.34 | 3.23 | 75.3 | 3.38 | 2.37 | 2.48 | 2.18 | 86.2 | 8.84 | 1.6 | 6.5 | ||
| 10 | 10 | 0.71 | 7.09 | 3.37 | 71.9 | 3.24 | 2.39 | 2.30 | 2.4 | 83.8 | 8.62 | 1.7 | 6.1 | ||
| 11 | 10 | 0.68 | 6.85 | 3.51 | 68.4 | 3.09 | 2.41 | 2.12 | 2.64 | 81.2 | 8.38 | 1.8 | 5.7 | ||
| 12 | 10 | 0.66 | 6.62 | 3.66 | 64.7 | 2.94 | 2.42 | 1.95 | 2.90 | 78.3 | 8.12 | 1.9 | 5.4 | ||
| 13 | 10 | 0.64 | 6.39 | 3.82 | 60.9 | 2.78 | 2.44 | 1.78 | 3.19 | 75.1 | 7.83 | 2.0 | 5.0 | ||
| 14 | 10 | 0.62 | 6.18 | 3.99 | 56.9 | 2.62 | 2.46 | 1.62 | 3.51 | 71.6 | 7.51 | 2.2 | 4.6 | ||
| 15 | 10 | 0.6 | 5.97 | 4.16 | 52.8 | 2.45 | 2.48 | 1.46 | 3.86 | 67.7 | 7.16 | 2.3 | 4.3 | ||
| 16 | 10 | 0.58 | 5.77 | 4.34 | 48.4 | 2.27 | 2.50 | 1.31 | 4.25 | 63.4 | 6.77 | 2.4 | 3.9 | ||
| 17 | 10 | 0.56 | 5.57 | 4.52 | 43.9 | 2.08 | 2.52 | 1.16 | 4.67 | 58.8 | 6.34 | 2.6 | 3.5 | ||
| 18 | 10 | 0.54 | 5.38 | 4.72 | 39.2 | 1.89 | 2.54 | 1.02 | 5.14 | 53.6 | 5.88 | 2.8 | 3.2 | ||
| 19 | 10 | 0.52 | 5.20 | 4.92 | 34.3 | 1.69 | 2.56 | 0.88 | 5.65 | 48.0 | 5.36 | 2.9 | 2.8 | ||
| 20 | 10 | 0.50 | 5.03 | 5.13 | 29.2 | 1.47 | 2.58 | 0.74 | 6.22 | 41.8 | 4.80 | 3.1 | 2.4 | ||
| 21 | 10 | 0.49 | 4.86 | 5.35 | 23.8 | 1.25 | 2.60 | 0.61 | 6.84 | 34.9 | 4.18 | 3.3 | 2.0 | ||
| 22 | 10 | 0.47 | 4.69 | 5.58 | 18.2 | 1.02 | 2.62 | 0.48 | 7.52 | 27.4 | 3.49 | 3.5 | 1.6 | ||
| 23 | 10 | 0.45 | 4.53 | 5.82 | 12.4 | 0.78 | 2.64 | 0.36 | 8.28 | 19.1 | 2.74 | 3.8 | 1.2 | ||
| 24 | 10 | 0.44 | 4.38 | 6.07 | 6.3 | 0.53 | 2.66 | 0.23 | 9.10 | 10.0 | 1.91 | 4.0 | 0.8 | ||
| 25 | 10 | 0.42 | 4.23 | 6.33 | 0.0 | 0.27 | 2.68 | 0.12 | 10.02 | 0.0 | 1.00 | 4.2 | 0.4 | ||
| Totals | 264.82 | 61.19 | 47.69 | 56.08 | 125.50 | ||||||||||
| Present values of costs at 3.5% pa; | |||||||||||||||
| ---- Repayment of capital | 61.2 | 56.1 | |||||||||||||
| ---- Interest on capital | 47.7 | 125.5 | |||||||||||||
| ---- Total capital cost | 108.9 | 181.6 | |||||||||||||
| --- Operating cost | 164.8 | 164.8 | |||||||||||||
| --- Total cost | 273.7 | 346.4 | |||||||||||||
| Interest cost as % of the total | 17 | 36 | |||||||||||||
| Capital cost as % of the total | 40 | 52 | |||||||||||||
| The public service comparator | |||||||||||||||
| The capital cost is £100mn. At a real 4.3% pa rate of interest, the capital recovery rate over the 25 years of the | |||||||||||||||
| project is 0.06606. The annual capital cost is therefore 0.06606 * £100 mn = £6.606 mn split between repayments and interest | |||||||||||||||
| charges as above | |||||||||||||||
| The PFI alternative | |||||||||||||||
| The interest rate is assumed to be 10% pa in real terms. The capital cost is the capital recovery rate over the 25 | |||||||||||||||
| year period of the contract. The capital recovery rate at 10%pa over 25 years is 0.110168 | |||||||||||||||
| The capital cost is therefore assumed to be 0.110168 multiplied by 100 = £11.017 million pa split between repayments and interest | |||||||||||||||
Now the construction cost of the PFI alternative needs to be 38% (£62 mn compared to £100 mn) lower than the PSC to compensate for the higher interest rate of the private sector18.
This chapter has compared a PFI version with a publicly-financed alternative using the project model given in the Andersen report but making more realistic the assumptions about interest rates. It has shown that for the PFI to match a scheme financed by the public sector and having a construction cost of £100 mn., the construction cost of the PFI scheme would need to be as little as £62 mn.
15 . The new Treasury guidance came into effect from 1 April 2003 (Department of Health, February 3 2003, para 5). The Department of Health pointed out that "Other factors which have been implicitly bundled up in the old 6% figure are dealt with explicitly and separately. The most important of these is optimism bias" (para 20). It is worth noting that the change in the discount rate came in 2003, three years after the Andersen report was produced.
16 . Many thanks to Peter Cockett of the Department of Health for pointing out the 'tax effect' to me.
17 . On March 4, 2008, the Guardian reported that many PFI companies have been transferred to offshore tax havens to reduce their tax.
18 . It is worth noting that if the project's construction cost is not substantially lower as a result of being privately financed, it is unlikely that private sector would come out best when all the non-clinical costs are taken into account. Not only would the private sector have to provide catering and other non-clinical operating costs at about two-thirds the cost of the public sector comparator but it would need to be shown that these lower costs are the result of the private finance initiative. Otherwise the lower costs could be captured by a private sector contract for the catering and other services alone and the PFI contract for the building finance would be unnecessary. Note that the gains need to be true efficiency gains and not simply the result of inferior terms of conditions of employees which a Retention of Employment Model and other regulations are designed to guard against (LRD 2003, 8).