In its indictment of Jacob Zuma the National Prosecuting Authority notes that up until the final decisions were made on the arms deal there were two separate and parallel processes. On the one hand there was the "The formal evaluation of the competing bidders ... conducted through an ostensibly rigorous and scientific evaluation process." On the other, there was an "informal process" through which "persons and entities interested in participating in the contracts sought to glean information about the process and exert influence, directly or indirectly, on formal decision makers."

From the evidence already gathered by German, British and South African investigations it seems that, with most elements of the arms deal, the "informal process" subverted the prescribed one. Choices were predetermined in backroom meetings, and the evaluation of the different bidders was then manipulated to ensure the desired outcome.

This seems to have been the case in the corvette contract as well. What other inference can one draw from the payment of $25m in bribes by the German Frigate Consortium (GFC) to "South African officials and members of cabinet" following its successful bid?

There is a contrary view that the GFC's Meko A200 won the corvette contract on legitimate grounds and if the Germans "did pay any bribes, they were wasting their money." Certainly, the GFC's bid was a good deal less bad than certain of the other winning contenders (most notably BAe's Hawk 100 - which won the jet trainer contract.) It is a matter of public record that the GFC beat Bazan of Spain on the strength of their National Industrial Participation (NIP) offer. However, there is evidence to suggest that if the integrity of formal evaluation process had been maintained Bazan would have come out ahead of GFC in the assessments - even with their lower rating on the NIP.

Each of the bidders for the corvette contract (as on the other elements of the arms deal) was supposed to be assessed on a formula combining price, military performance, financing cost, defence industrial participation (DIP), and NIP. The closing date for the Request for Final Offers was May 11 1998 and the assessments by each of the evaluation teams began very soon after. They were all completed by the end of June that year and consolidated at a SOFCOM meeting on July 1-2 1998. By the time of a meeting of the Armaments Acquisition Council on July 13 1998 the GFC's Meko A200 had come out ahead of the other bidders on the best value rankings. By this stage each of the process the bidders were assessed on the following formula:

Best Value (BV) = Military Value (MV) + Industrial Participation (IP) + Financing Index (FI)

Military Value was determined by dividing military performance over cost. Industrial Participation by a combination of each bidder's NIP and DIP ratings. If the information presented by the Auditor General in their report into the arms deal is correct, the result of the evaluations of the different bidders (in mid-July) would have been as follows:

It is clear then that by this stage the GFC bid had come out slightly, but not insubstantially, ahead of Bazan with 4.9 points between them. The normalised scores on "best value" were 100 for the GFC bid to 98.2 for Bazan's. Originally the value system that was supposed to be used to assess the different bids was:

BV = MV + IP
               FI

However, this was changed to BV = MV +IP + FI at the July 1-2 SOFCOM meeting. The minutes of the meeting recorded that "The Chairman [Chippy Shaik] emphasises the importance of showing the values of the three evaluation domains in a progression which culminates in a best value of Military value + IP value + Financing value."

The minutes of the special Armaments Acquisition Screening Board meeting on July 8 1998 further recorded:

"The Chief of Acquisition [Chippy Shaik] briefly reviewed the process, stressing the integration of the results of four independent evaluations per equipment undertaken by SOFCOM. Due to disproportionate influence of the financing result in the top-level value system, the SOFCOM accepted a modified equation prior to integration, i.e. Ranking = Technical + IP + Financing (each evaluation contributing one third to the final ranking)."

This decision to change the formula was highly irregular as SOFCOM was an ad-hoc committee with no formal decision making powers. In his interview with arms deal investigators in 2001 Erich Esterhuyse, the co-chair of SOFCOM, said that he had been on holiday at the time this decision was made. "When I came back, I was presented with this particular slide that says, this is now the value system."

"I looked at this and I said, this does not make sense from a mathematical point of view. You cannot change a multiplier into a plus-factor and I could not figure out why it was done. And I said, fine, if I look at the results of, from the project, the study teams, if this new formula changes the outcome from one contractor to the other, then I will object violently. I could not find a single case where the outcome was effectively changed by the change in approach."

At that time the main concern of Esterhuyse (and others) was with the efforts to push through the contract for the Hawk 100 - regardless of cost - and the change in formula did little to help BAe's bid in that regard. However, if the original formula had been used for rating the corvette contract the results would have been as follows:

Thus, although the GFC bid still came out ahead on the original value system, the differential between it and Bazan was much smaller.

In the NIP evaluations, carried out by at team at the Department of Trade & Industry (Minister: Alec Erwin), the GFC had been given over double the score (100) to Bazan (48). This was despite the fact that their offers were - in dollar terms - about the same. Bazan had however presented a far superior DIP offer to the GFC both qualitatively and quantitatively.

It was the only bidder to provide a detailed business plan for its counter-trade offer on the combat suite. It had promised 21.9% counter-trade on the platform component, as well as indirect defence counter-trade to the value of $406m. By contrast the GFC offered only 11.5% direct DIP on the platform, and only $6m indirect defence counter-trade. It had also failed to meet a series of minimum criteria on the DIP offer which should have led to its disqualification.

Although Bazan was given the best score on its DIP offer (100) the GFC was not all that far behind (81). This rating for the GFC was built partly on some questionable decisions and partly on miscalculation. When the Auditor General's Office re-examined the DIP calculations they uncovered a number of errors. According to their revised figures the GFC should only have received 71 points to Bazan's 100. What this meant is that Bazan's IP score should have been 86.5 rather than the 82 it was given.

If the correct DIP rating had been used - on the later formula - the differential between the two "best value" scores would have been reduced from 4.9 points to 0.4 points (out of 300). Bazan would have received a normalised best value score of 99.9 to the GFC's 100.

On their own the miscalculations in the DIP scores, and the change of formula, did not change the GFC's number one ranking (although they both came as close as damnit to doing so). But combined they certainly do. On the original formula, and using the correct IP ratings, Bazan comes out well ahead of the GFC. It has a normalised score of 100 to the GFC's 98.2. What this means - if the Auditor General's figures are correct - is that the R6,9bn corvette contract was swung from Bazan to the GFC on the basis of a combination of calculation errors (on the DIP) and a questionable change in the value system.

Further analysis of the evaluation results can be found here. The previous article in this series can be accessed here.