The First Postwar Census of Manufactures

Abstract

rrHE Census of Manufactures, taken decen1 nially from I809 to I899 and quinquennially thereafter to I9I9, became a welcome biennial event in the two interwar decades. It was relied upon as the source of the basic data, or as the statistical rudder, of many important manufacturing measures, like the Day-Thomas and Federal Reserve Board production indices, the National Bureau of Economic Research (Mills and Fabricant) production and employment series, the WPA National Research Project production, employment, and labor productivity series (later revised and extended by the Bureau of Labor Statistics), and the Bureau of Labor Statistics monthly chain indices of employment and payrolls. But the Census had to be bypassed in I94I and in the remaining war years. The statistical consequences of the suspension of the Censusand of the wartime instability of both the product assortment and industrial classification of manufacturing establishments -were serious indeed. Many measures had to be discontinued. Some, such as the Federal Reserve Board indices, were kept current by improvisations which could hardly have gained universal approval (like the introduction of man-hour series, generally with pseudo-productivity adjustments). The Bureau of Labor Statistics sought new benchmarks for its employment and payroll indices in the data of the Old Age and Survivors Insurance and State unemployment compensation programs. The quality of labor productivity measures deteriorated at a time when minor movements were interpreted as more significant for wage adjustments than ever before. In the absence of meaningful or reliable manufacturing statistics, traffic in conventional and mislabeled measures flourished. In I948, new legislation (Public Law 67I, 8oth Congress) authorized a quinquennial Census of Manufactures and annual and other interim surveys. These provisions and others concerning quinquennial nonmanufacturing canvasses contributed to the realization of the long-discussed objective of rationally scheduling the recurrent activities of the Bureau of the Census.2 The first Census of Manufactures undertaken in accordance with the new law pertains to I947; subsequent canvasses will refer to I953 and every fifth year thereafter. The first interim surveys pertain to I949 and