Results for tendering times

3  Figure 15 sets out the results showing all variables of interest. This model explained 38 per cent of the variation in the length of the tendering period (adjusted R2). The table indicates that the only factor that was statistically significant (at the five per cent level) was project sector. Health PFI deals took longer all else equal than PFI schools, and the 'other' PFI deals took longer still. Project preparation time was not significantly associated with reduced tendering time, though the variable had a negative coefficient. The Prior Experience variables were similarly not significant at the five per cent level, but are suggestive of a tendency for greater experience to be associated with shorter tendering times. The size variables were not significant, (net present value providing a better fit than capital value). This may be consistent with the greater use of resources on larger PFI deals.

15

Multiple regression coefficients and significance values to explain PFI tendering lengths

Model

Unstandardized Coefficients

Standardized Coefficients

 

 

 

B

Std. Error

Beta

t

Sig.

(Constant)

108.718

38.458

 

2.827

0.014

 

Other sector

25.861

8.944

0.771

2.891

0.013

 

Hospital

16.848

6.128

0.720

2.749

0.017

 

Net present value

2.98E-009

0.000

0.455

1.972

0.070

 

One or two prior deals, but no SRO experience

- 7.781

7.534

- 0.194

-1.033

0.321

 

One or two prior deals and SRO experience

- 6.160

5.343

- 0.294

-1.153

0.270

 

Three or more prior deals

- 25.402

14.661

- 0.460

-1.733

0.107

 

Project preparation time (months)

- 0.356

0.258

- 0.385

-1.380

0.191

 

Strength of competition (number of bidders at ITN)

- 2.510

4.599

- 0.109

- 0.546

0.594

Source: National Audit Office