For useful comments and advice, I wish to thank Price Fishback, Claudia Goldin, Robert Margo, Bruce Fallick, David Buffum, Laura Huntoon, Martin Eisenberg, Gerald Friedman, Richard Sutch, David Roediger, Josh Rosenbloom, and seminar participants at the University of Pennsylvania, University of Kansas, Social Science History Association Meetings, and Economic History Association Meetings. Work on this article was performed while I was an Andrew Mellon Fellow with the Program for Assessing and Revitalizing the Social Sciences at the University of Pennsylvania.Google Scholar

¹ President Wilson, Woodrow, Aug. 29, 1916, endorsing the Adamson Act.Google ScholarLink, Arthur s., ed., The Papers of Woodrow Wilson, vol. 38 (Princeton, 1982), p. 97.Google Scholar

² Cahill, Marion Cotter, Shorter Hours: A Study of the Movement since the Civil War (New York, 1932)Google Scholar; and Roediger, David and Foner, Philip, Our Own Time: A History of American Labor and the Working Day (Westport, 1989), give overviews of the shorter-hours movement.Google Scholar

³ According to the Census of Manufactures, the average “normal” work week fell from 57.31 hours (1909) to 55.51 hours (1914) to 51.26 hours (1919). It is the census concept of normal hours which I use. The frist twentieth-century industry-level data on hours are for 1909; city-level data on hours begin in 1914.Google Scholar

⁴ Hunnicutt, Benjamin, Work Wothout End: Abandoning Shorter Hours for the Right to work (Philadelphia, 1988), p. 35.Google Scholar

⁵ Immigration figures are from U.S. Department of Commerce, Historical Statistics of the United States: Colonial Times to 1970 (Washington, DC, 1976), series C-89.Google ScholarLabor force figures are from Lebergott, Stanley, Manpower in Economic Growth: The American Record Since 1800 (New York, 1964), tables A-3, A-5.Google ScholarPeak military strength is from Ayres, Louis, The war with Germany: A Statistical Analysis (Washington, DC, 1919).Google Scholar

⁶ Figures are from Lebergott, Manpower, tables A–3, A–5, A–15, A–16, A–18.Google Scholar

⁷ The “basic eight-hour day” refers to a flat rate of pay for the first eight hours, with time-and-a-half for hours beyond eight. For state maximum-hours laws, see U.S. Women's Bureau, Chronological Development of Labor Legislation for Women in the United States, Bulletin No. 66–11 (Washington, DC, 1931).Google Scholar

⁸ See Wolman, Leo, The Growth of American Trade Unions, 1880–1923 (New York, 1924), on the growth of unionsGoogle Scholar; and Conner, Valerie, The National War Labor Board: Stability, Social Justice, and the voluntary State in World War I (Chapel Hill, 1983), on the role of union leaders in the government.Google Scholar

⁹ For a survey, see Whaples, Robert, “The Shortening of the American work Week: An Economic and Historical Analysis of its Context, Causes and Consequences” (Ph.D. Diss., University of Pennsylvania, 1990)Google Scholar; and Owen, John, Working Hours: An Economic Analysis (Lexington, MA, 1979).Google ScholarNarrower studies which focus on the period around World War I include Shiells, Martha, “Collective Choice of Working conditions”, in this issue of the journal, pp. 379–92,Google Scholarand Goldin, Claudia, “Maximum Hours Legislation and Female Employment in the 1920s: A Reassessment”, Journal of Political Economy, 96 (02. 1988).CrossRefGoogle Scholar

¹⁰ Cahill, Shorter Hours, p. 27.Google Scholar

¹¹ This model is presented in Bernanke, Ben, “Employment, Hours, and Earnings in the Depression,” American Economic Review, 76 (03. 1986).Google Scholar

¹² The U.S. Bureau of the Census, Fourteenth Census of the United States, vol. 9, Manufactures, 1919 (Washington, DC, 1923), reports only one level of hours for all workers in a single firm, stating, “no attention was paid to the fact that a few wage earners might have hours different from those of the majority. In most establishment, however, all the wage earners work the same number of hours, so that the figures give a substantially correct representation of the hours of labor,” p. 17.Google Scholar

¹³ The weights used are the number of manufacturing workers divided by the average number of manufacturing workers in the city or industry in 1919. Unweighted regressions show essentially the same patterns as the weighted equations, except that the independent variables SHIP and ELECTRIFICATION lose significance.Google Scholar

¹⁴ The hours of salaried employees were set seperately from those of operatives. However, salaries probably moved along with operatives' wages, since both were tied to the tightness of the local labor market and salaried workers received wage premia in industries in which operatives received wage premia. This anomaly has been noted for the postwar period in Krueger, Alan and Summers, Lawrence, “Efficiently Wages and Interindustry Wage Structure,” Econometrica, 56 (Mar., 1988), and is verified using the industry-level data from 1909 and 1919.CrossRefGoogle Scholar

¹⁵ This does not capture the interaction between unionization and strikes, since I cannot link strikes to industries and unionization to cities. Nor does it capture the threat effect that unionization had no nonunionized cities and industries. However, the increased unionization and strikes themselves may be a function of the labor market tightness and the actions of the federal government, so it may also overstate the role of unionization.Google Scholar

¹⁶ Piore, Michael, Birds of Passage: Migrant Labor and Industrial Societies (New York, 1979)CrossRefGoogle Scholar; Brody, David, Steelworkers in America: The Nonunion Era (New York, 1960)Google Scholar; and Shiells, “Collective Choice”.Google Scholar

¹⁷ See, for example, Dunlevy, James and Gemery, Henry, “The Role of Migrant Stock and Lagged Migration in the Settlement Patterns of Nineteenth-Century Immigrants”, Review of Economics and Statistics, 59 (May 1977), who show that the past flow of immigration is the key predictor of its future flow.Google Scholar

¹⁸ One indication that these laws were binding is seen in New England textile cities like Fall River, MA. In 1914, 90.1 percent of workers worked the legal weekly maximum for women (54 hours), in 1919, 97.2 percent worked the new maximum (48 hours).Google Scholar

¹⁹ This estimate is very close to that of Jones, Ethel, “State Legislation and Hours of Work in Manufacturing,” Southern Economics Journal, 41 (Apr. 1975).CrossRefGoogle Scholar

²⁰ Soule, George, Prosperity Decade: from War to Depression, 1917–1929 (New York, 1947), notes: “One of the main causes of the reduction in hours was a law that had been passed in 1912 making the eight-hour day … mandatory for those engaged in government contracts”Google ScholarDay, Edmund, “The American Merchant Fleet: A War Achievement, A Peace Problem,” Quarterly Journal of Economics, 34 (Aug. 1920), notes, “government control of merchant shipping … was absolute”.CrossRefGoogle Scholar

²¹ See Conner, The National War Labor Board. The board made 384 industrial awards directly to about 600,000 manufacturing workers. Not all awards dealt with hours. The official summary and analysis of NWLB awards notes: “The umpires usually have awarded the basic eight-hour day.”Google ScholarSee “National War Labor Board,” in U.S. Department of Labor, Bulletin of the Bureau of Labor Statistics, No. 287 (Washington, DC, 1921), pp. 19, 71–85. Since the 48-hour week was more than seven hours below the average work week of 1914 and about three hours below the 1919 average, the awards probably accelerated the decline in hours. The Bethlehem, PA, case represents almost 10 percent of the workers covered by NWLB awards. Multiplying Bethlehem's effect by 10 implies that the NWLB's total impact explains about 3.4 percent of the hours decline.Google Scholar

²² Conner, National War Labor Board, p. 133.Google Scholar

²³ Roediger and Foner, Our Own Time, p. 206.Google Scholar

²⁴ Jacoby, Sanford, Employing Bureaucracy: Managers, Unions, and the Transformation of Work in American Industry, 1900–1945 (New York, 1986).Google ScholarThe personnel boom of the war era was largely induced by labor market tightness, but as Wright, Gavin, ” in Field, Alexander, The Future of Economic History (Boston, 1987), argues, it was part of a longer-term shift toward a “high-wage-high-quality-low-turnover” employment relation.Google Scholar

²⁵ Hicks, John, The Theory of Wages (New York, 1962), suggests just such an effect.Google Scholar

²⁶ This is not just an effect of government contracts going to larger firms; substituting a measure of the percentage of employees in large establishments yields an insignificant effect. Another indication that productivity enhancement was a goal of employers in reducing hours is that, controlling for all else, hours fell by a greater percent when they were initially above 60 per week. Goldmark, Josephine, Fatigue and Efficiency (New York, 1912), was very influential in showing the link between hours and efficiency, as were government studies made during the war.Google ScholarA number of economic historians have found an inverted-U shape between hours and productivity, including Atack, Jeremy and Bateman, Fred, “How Long Did People Work in 1880?” (paper presented at Economic History Association meetings, 1989).Google Scholar

²⁷ See Devine, Warren, “From Shafts to Wires: Historical Perspective on Electrification”, this JOURNAL, 43 (06 1983).Google Scholar

²⁸ The auto industry reduced nonmilitary output by 50 percent in 1918 under an agreement with the War Industries Board (WIB), and home construction fell by two-thirds due to the WIB's allocation of scarce raw materials. Although the WIB was virtually dismantled by the beginning of 1919, consumer credit was rationed in 1919, so auto and new home sales remained relatively low. Baruch, Bernard, Industry in the War (New York, 1941)Google Scholar; and Soule, Prosperity Decade, p. 103.Google Scholar

²⁹ Horowitz, Daniel, The Morality of Spending: Attitudes toward the Consumer Society in America, 1875–1940 (Baltimore, 1985), p. 120. An example of this patriotic guilt is that over two million people took part in the drive to sell Liberty Bonds. Community pressure was applied so ubiquitously that nearly every household purchased bonds in lieu of spending, even though the real return on this investment was negative. This phenomenon lasted well into 1919. Twelve million subscribers, half of all American households, bought bonds in the Victory Loan drive from March to June 1919, and many of the subscribers to the loan drive of late 1918 paid in installments during the first half of 1919.Google ScholarBogart, Ernest, War Costs and Their Financing (New York, 1921), pp. 216–28.Google Scholar

³⁰ Friedman, Milton, Price Theory (Chicago, 1962), p. 204Google Scholar; and Kyrk, Hazel, A Theory of Consumption (Boston, 1923), pp. 255–58.Google Scholar