¹ This paper was stimulated by the discussion between Robert Gallman and Alfred D. Chandler, Jr. concerning structural change during the Civil War era. The discussion appeared in Economic Change in the Civil War Era: Proceedings of a Conference on American Institutional Change, 1850-73, and the Impact of the Civil War Held March 12-14, 1964, Gilchrist, David T. and Lewis, W. David, eds., (Greenville, Del., 1965)Google Scholar; see 147-48 and 160-61. Gallman suggested that within manufacturing there was a marked shift toward heavy industry during the Civil War decade; Chandler suggested that structural shifts in the Civil War decade were merely a continuation of developments which began apart from the war during the 1850's.

² The number of manufacturing industries included in the census stood between 400-600 during the latter half of the nineteenth century. There was no attempt at aggregation by major industry groups until 1880, and this was only provided for selected industries.

³ The census manufacturing data were obtained from the following sources: “Digest of the Statistics of Manufactures in the United States in 1850,” Senate Documents, 35th Cong., 2d Sess., Vol. 10, 1858-1859, pp. 137–42Google Scholar; Eighth Census of the United States, Manufactures in the United States in 1860, 733-742; Ninth Census of the United States, 1870, The Statistics of the Wealth and Industry of the United States, Vol. III, 394-98; Twelfth Census of the United States, 1900, Manufactures, Vol. VII, Part I, 3-17. The industry conversion was made on the basis of the definitions in the 1963 Census of Manufactures, Industry Descriptions, U.S. Department of Commerce.

⁴ This particular technique of converting census data to the SIC two digit categories was first demonstrated by this author in “The Development of Industrial Structure in Southern New England,” Journal of Economic History, XXX (September, 1970)Google Scholar; also see “Structural Shifts in Southern Manufacturing, 1849-1899,” Business History Review, XLV (Spring, 1971), 79–84.Google Scholar

⁵ In order to make an appraisal of structural developments within manufacturing, deflation for price change is needed due to the sectoral variations in price movements; for example, one need only glance at the sources of price data in notes 6 and 7 to become aware of the great industry differences in absolute levels and trends in prices during the latter half of the nineteenth century.

⁶ Warren, George F. and Pearson, Frank A., Prices (New York, 1933), 24–30.Google Scholar

⁷ The prices used from the report of Senator Aldrich, “Wholesale Prices, Wages, and Transportation,” appeared in Senate Reports, 52d Cong., 2d Sess., 1892-1893, Special Session, March 4, 1893, Vol. 3, Report 1394, Part 1, 6-52; Bezanson, Anne, Wholesale Prices in Philadelphia, 1852-96 (Philadelphia, 1954).CrossRefGoogle Scholar

⁸ Callman, Robert E. used output deflators for a portion of his work in “Commodity Output, 1839-99,” Trends in the American Economy in the Nineteenth Century, National Bureau of Economic Research, Studies in Income and Wealth, Vol. 24 (Princeton, N.J., 1960).Google Scholar At present, much of the U.S. Department of Commerce's series on real gross national product by major industry (Survey of Current Business, April, 1967, 18-24) is estimated through the use of output deflators rather than via double deflation. According to John W. Kendrick and C. Milton Jaycox, double deflation is used by the Commerce Department only in agriculture, construction, manufacturing, electric utilities, railroads, and the major portions of finance and insurance; see “The Concept and Estimation of Gross State Product,” Southern Economic Journal (October, 1965)Google Scholar. In addition, Kendrick and Jaycox point out that due to the fact that much of intermediate input in manufacturing represents output from related manufacturing industries, the common procedure of using output deflators is probably most defensible for manufacturing.

⁹ Use of output deflators involves the assumption that prices paid for intermediate inputs behaved similarly to prices received for output. Paul David points out some of the theoretical problems involved in the use of output deflators in “The Deflation of Value Added,” Review of Economics and Statistics (May, 1962), 148–155Google Scholar.

¹⁰ A partial test of the reliability of using New York prices for the nation can be achieved by using the New York based Aldrich Report prices and the Philadelphia based Bezanson price data to construct the percentage margin of error as follows:

where P = price index, Q = current dollar value added, i = the i^th industry, B refers to Bezanson's prices, and A refers to the Aldrich Report prices. (Warren and Pearson price data were not available at the disaggregated industry level found in Bezanson and therefore the Aldrich Report was used for New York prices). Because of data limitations, the margin of error was constructed for only two groups of industries, foods and metals. Included in foods were lard, cheese, butter, wheat flour, rye flour, cornmeal, molasses, and linseed oil; included in metals were bar iron, pig iron, copper sheeting, rolled copper, nails, and pig lead. Some rather wide differences developed due to changes in the geographic price base; however, these differences were always in the same direction for both foods and metals and were generally of the same order of magnitude. As a result of the similarity of error across industries for given time periods, it did not appear that use of the New York based Warren and Pearson prices would seriously affect the estimates of industrial structure.

¹¹ At the end of the nineteenth century, a small amount of electrical machinery production arose; this has been included in SIC 35.

¹² One should not attribute undue significance to the shifts within the metals and machinery groups during 1869-1879. It appears from Table 2 that primary metals declined and machinery rose over this decade; however, this is partially due to the inclusion, beginning in 1879, of the census category of foundry and machine shop products in SIC 35. Prior to 1879 this group had been disaggregated and portions were included in SIC 33.

¹³ Kendrick, John W., “Productivity Trends: Capital and Labor,” Review of Economics and Statistics (August, 1956), 248–257CrossRefGoogle Scholar. For a more detailed statement, see Kendrick's, Productivity Trends in the United States, (Princeton, N.J., 1961).Google Scholar Several other authorshave also cited the strong relationship between real output growth and growth in productivity. See Schmookler, J., “Changing Efficiency of the American Economy, 1869-1938,” Review of Economics and Statistics (August, 1952), 214–231CrossRefGoogle Scholar and Mills, F. C., “The Role of Productivity in Economic Growth,” American Economic Review, Papers and Proceedings (May, 1952), 545–557Google Scholar.

¹⁴ See Kendrick's discussion of productivity and real output growth in Productivity Trends, 182-87 and 206-209.

¹⁵ The estimates of capital invested in buildings and machines involved a conversion similar to that used in the estimation of value added. The basic data were obtained from the Eleventh Census of the United States, Report on Manufacturing Industries, Part I, 92-113 and the Twelfth Census of the United States, Manufactures, Part I, 20-50. The current dollar estimates were deflated to a base of 1879 prices by constructing an implicit (current/constant) deflator on the basis of the current and constant dollar estimates of capital in major branches of manufactures provided in Historical Statistics of the United States (Washington, 1960), 411–12.Google Scholar The real value added and real value added per employe series were redefined on a base of 1879 prices using the sources in Technical Notes, section 2.