Opportunities for Reducing the Cost of the 1982 Agricultural and Economic Censuses

GAO reviewed the agricultural and economic censuses, examining costs already incurred and the Bureau's estimated future costs to aid the fiscal year 1983 budget reconciliation process. In examining costs, GAO: (1) compared the 1982 censuses to previous censuses to identify changes and their expected benefits; and (2) determined the usefulness of work already performed if portions of the censuses were reduced or eliminated and the effect of the changes on estimated costs.

If budget cuts arise, the Bureau of the Census is considering reducing the size and scope of the agricultural and economic censuses. For the agricultural census, the Bureau could save $13 million by eliminating an area sample, raising the threshold for defining a farm, and eliminating the censuses of areas outside the United States. For the economic censuses, the Bureau is considering eliminating the data that are collected on a sample basis, which include data on transportation, minority-owned businesses, and major retail and central business districts. With these actions, the Bureau could save about $6.7 million spent on the economic censuses and not affect the data gathering from establishments included in the censuses for retail and wholesale trade, selected service industries, mining and manufacturers, and construction. If the budget cuts are made within the next few months, the Bureau could save about $19.7 million. The Bureau plans to conduct the 1982 censuses similarly to the previous censuses, except for the introduction of a changed method of data entry using computer technology. Bureau data show that, aside from increases due to inflation, there will be no additional increases in the cost of taking the 1982 censuses. The Bureau believes that the cost of the new technology, including the use of a newly acquired computer, will be offset by reduced personnel costs. Although proposed reductions would affect the completeness of the agricultural census, the specific effect on the utility of the data has not been determined.