dissenting:
Our instructions on remand went beyond the narrow limitations ascribed to them by the majority and permitted the district court to declare Chemung the successful bidder upon a finding that “the FAA acted arbitrarily and capriciously in improperly recalculating costs, not made part of the solicitation, in its re-evaluation and in then recognizing the NFTA as the successful offeror____” Chemung I, 781 F.2d at 971. Since the FAA purported to undertake the recalculation of telecommunications costs in accordance with the guidelines it had adopted at the urging of the GAO, the issue to be resolved on this appeal simply is whether or not those guidelines were followed. An examination of the record persuades me that the FAA’s computation of telecommunications costs violated its own guidelines, that its decision to award the flight station contract to NFTA therefore was arbitrary and capricious, and that the judgment of the district court should be affirmed.
Specific standards for estimated telecommunications costs are established by the guidelines. First, the “most economical *224mix” of telephone lines must be used for pilot access. Any mix of lines — local, foreign exchange (“FX”), or in-WATS — may be employed, so long as the most economical mix is obtained to meet current levels of demand. Second, the number of lines required for each type of “service” (here, “service” means type of telephone trunk line) should be based on the level of demand and the point of origin of demand. Third, the guidelines provide a complex formula as a “rule of thumb” to estimate the number of telephone lines required to serve a given level of demand. They also provide that “[allocating the demand to the different trunk circuits may be estimated by operations personnel if line counter data is not available in the facility.” Joint App. at 754 (emphasis added). Further, although use of total “services” (in this context, yearly flight plans and weather briefings) is contemplated in the formula, the figure to be derived is the total number of calls during the busy hour, which then essentially is divided by ten to arrive at the total number of lines required. Finally, the FAA guidelines also provide that when an FSS station has a mix of lines, as both Elmira and Buffalo did, calculations must be made for each type of “trunk service” provided.
Applying these guidelines, it becomes apparent that the FAA’s use of “services” (flight plans and weather briefings) to calculate demand at Buffalo, see Joint App. at 604-05, was unwarranted. While the guidelines formula is designed to calculate demand levels by use of the “services” factor, there is no need to employ that factor where actual call data is available. The guidelines make it clear that allocation of demand to the different trunk lines may be estimated only if counter data is not available. Common sense dictates the same approach. Since the number of “busy hour” calls were reflected on the line counters, the remainder of the formula computations could have been undertaken in reliance on the accurate information at hand. Instead of proceeding in this fashion, the FAA inserted “services” numbers into the equation, ignoring the Buffalo counter evidence of actual use. The FAA procedure produced skewed results. Acceptance of the distorted product engendered by imput of the “services” factor led to the obvious conclusion that Buffalo offered a competitive advantage over Elmira, which would need some form of long distance line to substitute for each of Buffalo’s local lines.
In her closing statement, the attorney for the FAA said: “[P]utting in counter figures into that formula is irrational, especially in the situation like this where only one of the flight service stations in a particular automated flight service station flight plan area has a counter.” Joint App. at 605. Curiously, in its point of origin determinations, the FAA considered it quite rational to use available counter data for the Buffalo station together with estimates for all the other stations. In any event, I am of the opinion that it was arbitrary and capricious, as well as irrational, not to follow the guidelines and use actual data, where available, to determine demand levels.
Also violative of the FAA guideline was the FAA’s use of 1984, rather than 1985, telephone rates in its re-evaluation. Although the initial recalculation was based on 1985 rates, which were then available, the FAA ordered a further recalculation using 1984 tariffs. Dennis DeGaetano, the author of the guidelines, testified in a deposition that the guidelines call for use of “the rates in effect at the time of the calculation.” Joint App. at 353. It was his opinion that this referred to the time of the original calculation. Actually, the guidelines are not so specific as Mr. DeGaetano would have them. However, they do speak of rates in the present tense, e.g., the “current 260 tariff” for interstate rates and “intrastate rates are standard,” and therefore require the latest available data. Moreover, it defies logic to conclude, as has the FAA, that “the most economical mix of lines” can be determined without using the most current rate information.
*225There is a claim that use of the 1984 tariffs was favorable to Elmira. NFTA attempted to introduce into evidence an undated set of calculations purporting to show that use of the 1985 rates caused Elmira’s bid to descend in the ranking of telecommunications costs. That exhibit was not put into evidence, however, because all of the supporting documentation for 1985 was willfully destroyed by the FAA, supposedly because having such information in the file would “muddy up things.” Joint App. at 202. The district court correctly observed that such admitted and willful destruction could give rise to a strong inference that production would have been unfavorable to the spoliator. Joint App. at 736 (citing Dow Chemical Co. (UK.) v. S.S. Giovanella D’Amico, 297 F.Supp. 699, 701 (S.D.N.Y.1969)). The district court chose, however, merely to exclude the proffered exhibit from evidence. Accordingly, there is no way to determine whether the 1985 rates favored Elmira or not, Mr. Kearick’s testimony on this issue having been predicated on the rejected exhibit. In my view, the use of the older rates, when combined with the intentional destruction of supporting material consisting of calculations based on more current rates, leads inexorably to the conclusion that the FAA could not have determined the most economical mix of lines, primarily because the information supplied was outdated. Combining this conclusion with the fact that the FAA projected future costs for telecommunications without accounting for future rate changes, I am constrained to conclude that the FAA’s use of 1984 rates was arbitrary and capricious.
For the foregoing reasons, I respectfully dissent.