Reliability.
(ii) Reliability. A controlled participant's RAB share must be determined by using the most reliable estimate. In determining which of two or more available estimates is most reliable, the quality of the data and assumptions used in the analysis must be taken into account, consistent with 1.4821(c)(2)(ii) (Data and assumptions). Thus, the reliability of an estimate will depend largely on the completeness and accuracy of the data, the soundness of the assumptions, and the relative effects of particular deficiencies in data or assumptions on different estimates. If two estimates are equally reliable, no adjustment should be made based on differences between the estimates. The following factors will be particularly relevant in determining the reliability of an estimate of RAB shares: