The Public Sector Comparator showed that the PFI deal was cheaper

5.13  To measure the value of the deal against an appropriate benchmark, GCHQ estimated the costs of a conventionally financed alternative to the PFI deal. For most of the competition, the comparator assumed building on a single site, in line with the PFI bid. It did not include the costs of technical transition, as these were not part of the PFI deal and GCHQ had not at that time considered them in detail.

5.14  Having calculated a Public Sector Comparator on that basis, GCHQ used it as a benchmark to assess the increases in bid price following IAS's selection as preferred bidder. For example, after the preferred bidder stage, GCHQ increased the size of the building in the Comparator by some 30 per cent and the provision for staff places by some 7.3 per cent to make it compliant with its requirements. This increase was in line with the increase in the size of the PFI building after selection of the preferred bidder.

5.15  When Treasury became aware of the large increase in technical transition costs in late 1999, it suggested that GCHQ look in greater detail at a two site comparator. Both parties thought that this would use existing buildings to a greater extent and, at the time, that it would require less complex equipment moves, thus reducing transition costs. GCHQ considered that a return to the publicly funded Cheltenham Building Programme would be a more likely best alternative to the PFI deal, so the final judgement on costs reasonably used this as a comparator.

5.16  The estimated basic construction costs in the final Comparator were increased by 24 per cent in line with Treasury advice on historical cost overruns on large scale public sector projects. GCHQ did not build into the comparator an estimate of space inefficiencies which would result from continuing to work in multiple buildings on two sites; these inefficiencies were estimated at some 15 per cent and GCHQ considered that the construction costs elements in the comparator were a conservative estimate overall. As in other PFI cases, the adjustment for risk on construction costs of the public sector alternative more than accounts for the estimated cost difference between the comparator and the PFI deal.

5.17  In the final months before contract signature, GCHQ for consistency added an estimate for technical transition to the comparator and to the bid price. The Public Sector Comparator finally included technical transition costs of £68 million. These were much lower than in the PFI figure because, as explained above, it was assumed at the time that some of the existing buildings would be used at the two sites and that this would reduce costs. This was not a robust estimate, however, and it was not re-visited for comparison purposes as the lifetime costs of the adopted, extended, technical transition plan began to emerge.

5.18  Figure 12 shows the final comparison between the signed deal and the two site comparator that was reported to the Treasury during 2000. This estimated that the IAS bid, leaving aside broadly comparable services to be retained by GCHQ, would cost some £71 million less than the two site comparator in net present value terms, some seven per cent less, including technical transition. In this comparison technical transition costs for the PFI deal covered the first five years only but were included at the high cost which had by then been estimated.

The exercise also showed broadly similar services to be retained by GCHQ under each option of £99 million (PFI) and £94 million (comparator). These are not significant to the comparison made.

The figure (which excludes retained services as described above) shows that the final IAS bid cost some £71 million less than the Public Sector Comparator, including estimates for technical transition.

The technical transition costs were much lower in the Public Sector Comparator at £68 million because GCHQ assumed it would remain on two sites and use some of the existing buildings if the PFI deal did not go ahead. Technical transition would, therefore be less complex and costly. The PSC technical transition would deliver significantly less benefit to GCHQ and result in later additional expenditure to modernise its infrastructure.

The technical transition costs for the PFI deal shown above cover the first five years only.