blatant departure from recognized administrative standards. This wise advice was apparently accepted by the judge to whom this case was originally assigned. He granted full discovery into the exercise of the Secretary's discretion even beyond a "peek." The resulting extensive record demonstrates with crystal clarity that the Secretary conscientiously exercised his discretion and that if that exercise is subject to review, his decision not to depart from regular Census data in an attempt to adjust for the undercount by estimate was unassailable under any standard.
The administration of the complex Revenue Sharing program involved allocations to many different state, county, and local governments. The Secretary through his delegee, Graham W. Watt, Director of Revenue Sharing, carefully considered the possibility of adjusting for the undercount when the Siegel paper was brought to his attention. A method of estimating the undercount at the state and local level was needed to translate the national figure, itself an estimate, to localities in aid of allocation. This was recognized as a highly technical and intricate undertaking: because factors such as population composition and trends and census procedures
varied widely among localities, no flat, across-the-board estimate would have been equitable or proper. Seeking a statistically valid comprehensive technique, Watt had a series of discussions with top technical experts at Census. Although as early as April 1973 Census had advised that it was not practical to make allocations of the undercount to local levels, Watt continued to pursue the matter. Stanford Research Institute was contracted to consider whether there were statistically valid data available in or out of Census. The Institute reported there was no such data. This report as well as Census material were reviewed by the Data and Demography Section of Watt's office. Still no solution was found. A workshop was held with interested groups seeking solution. Nothing satisfactory to the Secretary was developed. It was determined not to attempt an estimate for the some 38,000 entities involved. This was a rational decision based on an informed judgment.
Central to plaintiffs' position is their claim, based on the affidavit of their own expert, that a statistically reliable procedure was available in April 1973 for adjusting the 1970 Census population data at the state, county, and local level to reflect the undercount. The motion papers raise doubts as to the reliability of plaintiffs' expert's procedure, but this dispute is not at the core of the controversy. Rather, the question is whether or not the Secretary abused his discretion in determining not to adopt a particular mathematical formula to reflect the undercount, preferring instead, after soliciting and obtaining advice from his own experts and retained consultants, to rest his allocations on the regular Census data as directed by the Act. The Act does not require the Secretary to adopt a statistical method for estimating even if one exists, and he was satisfied after consultation with Census and outside experts that an equitable adjustment could not be achieved by estimates attempting to reduce a national estimate to the local level. Recognizing these considerations plaintiffs belabor their position by suggesting that the Secretary was too cautious in evaluating possible estimation procedures, requiring "near absolute demographic accuracy and purity." This, however, he had every right to do.
The Secretary's discretion under 31 U.S.C. § 1228(a)(7)(B) is not reviewable. Alternatively, if reviewable, his decision not to estimate the undercount for Baltimore and for Newark was neither arbitrary nor capricious.
Summary judgment is granted defendants and denied plaintiffs.