Earlier versions of this article were presented at the 1986 Cliometrics Conference, the 1986 University of Georgia Department of Economics Summer Seminar, the Economic History Association Meetings and Clemson University in 1987, and the University of Texas at Austin and University of New Mexico in 1988. The authors would like to thank Scott Atkinson, John Brown, Louis Cain, Cletus Coughlin, Christian Dustmann, Mason Gerety, Caren Ginsberg, Claudia Goldin, Robert Higgs, Gary Libecap, Hugh Nourse, Anthony O'Brien, Raymond Sauer, Dan Slesnick, Richard Sutch, Joseph Terza, Ron Warren, Paul Wilson, and anonymous referees for helpful comments and criticisms on earlier drafts. Financial support for the completion of this project was provided by the Earhart Foundation.

¹ Meeker, Edward, “The Improving Health of the United States 1850–1915,” Explorations in Economic History, 9 (Summer 1972), pp. 353–73.CrossRefGoogle ScholarPubMed

² Whipple, G., “Fifty Years of Water Purification,” in Ravenel, M., ed., A Half Century of Public Health (1921, reprint New York, 1970), p. 166.Google Scholar

³ U.S. Public Health Service, “Sanitation of 123 Communities in the Bituminous Coal Districts of 9 States,” Report of the U.S. Coal Commission, 68th Congress, 2nd session (Washington, D.C., 1925), PP. 1625–29. The water source is deemed the prevailing type if more than 60 percent of the population in the community used it.Google Scholar

⁴ The importance of the water source to sanitation choice in major cities is discussed in Louis Cain, “An Economic History of Urban Location and Sanitation,” Research in Economic History, 2 (1977), pp. 337–89.Google Scholar

⁵ U.S. Coal Commission, “The Bituminous Mine Workers and Their Homes,” Report of the U.S. Coal Commission, pp. 1439, 1474.Google Scholar

⁶ U.S. Coal Commission, “Bituminous Workers,” pp. 1445, 1473.Google Scholar

⁷ Lebergott, Stanley, The American Economy: Income, Wealth, and Want (Princeton, 1976), pp. 271–72,Google Scholar estimated that, in 1930, 51 percent of American households had indoor flush toilets, 85 percent of urban households had flush toilets, and 8 percent of farm households had flush toilets. Using Lebergott's figures and data on the number of households from U.S. Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970 (Washington, D.C., 1975), pp. 11, 43, we calculated that 6.7 percent of rural nonfarm households had indoor flush toilets in 1930. Since only 20 percent of all U.S. households had flush toilets in 1920, the percentage of households with flush toilets was probably similar in company towns and other towns of similar size in 1922.Google Scholar

⁸ U.S. Public Health Service, “Sanitation of 123,” pp. 1616, 1630–35.Google Scholar

⁹ Lumsden, L. L., “Rural Sanitation, A Report on Special Studies Made in 15 Counties in 1914, 1915, and 1916,” U.S. Public Health Service Bulletin No. 94 (Washington, D.C., 1918), P. 43.Google Scholar

¹⁰ White, Joseph H., “Houses for Mining Towns,” U.S. Bureau of Mines, Bulletin No. 87 (Washington, D.C., 1914), p. 6.Google Scholar

¹¹ The house rent (r) might be endogenous to the system in two ways. First, employers might recover some of the costs of sanitation by charging higher house rents. Second, if the company has a local monopoly on housing, rent becomes a function of the number of miners. The comparative statics for the remaining parameters are not affected by either of these changes. In the empirical section we experimented with omitting rent from the regressions and by replacing the wage variable with a net wage variable, the monthly wage minus the monthly rent. The qualitative results in the alternative regressions were basically the same as those reported in Tables 1 and 2. We chose to report the equations with rent included because we could explicitly test the extent to which wages adjusted to compensate for higher rents.

¹² Descriptions of the substantial movement of miners between towns, particularly in nonunion districts can be found in Price Fishback, V., “Employment Conditions of Blacks in the Coal Industry, 1900–1930” (Ph.D. diss., University of Washington, 1983), pp. 60–65;Google ScholarCorbin, David, Life, Work, and Rebellion in the Coal Fields: The Southern West Virginia Miners, 1880–1922 (Chicago, 1981), pp. 40–43;Google Scholar U.S. Senate, U.S. Immigration Commission, Report on Immigrants in Industries, Part I: Bituminous Coal Mining, 61st Congress, 2nd session (Washington, D.C., 1911), vol. 1, P. 164; and vol. 2, pp. 27, 149, 151–52, 155, 158, 218.Google Scholar

¹³ The LOCALE rating also contains a measure of accessibility.

¹⁴ U.S. Coal Commission, “Bituminous Workers,” pp. 1433, 1426–39, 1480–94, 1587–98.Google Scholar

¹⁵ The original schedules are located in the U.S. Coal Commission Records, Record Group 68, Boxes 24–32 at the National Archives in Suitland, MD. LOCALE was measured with a rating system, giving high marks to broad valleys with plenty of space and good connections with other communities.

¹⁶ Copies of the rating form are published in U.S. Coal Commission, “Bituminous Workers,” pp. 1481–88, 1594–98.Google Scholar

¹⁷ The hourly wage and union data are from Fisher, Waldo and Bezanson, Anne, Wage Rates and Working Time in the Bituminous Coal Industry, 1912–1922 (Philadelphia, 1932), pp. 248–53.CrossRefGoogle Scholar The hourly wage was transformed into a monthly wage by multiplying by an estimate of the average number of hours worked per month. The average of 135 hours is based on the U.S. average of 8.3 hours worked per day and 195.7 days of operation in 1921. U.S. Bureau of Labor Statistics, Bulletin No. 316, “Hours and Earnings in Anthracite and Bituminous Coal Mining” (Washington, D.C., 1922), pp. 8, 16–27.Google Scholar

¹⁸ U.S. Bureau of Mines, Mineral Resources of the United States, 1925, Part II-NonMetals (Washington, D.C., 1925), p. 86.Google Scholar

¹⁸ To obtain the monthly wage in West Virginia, the piece-rate wage was multiplied by 92 tons, which was the average amount of pick mining output per pick miner per month in West Virginia in the year ending June 30, 1922. West Virginia Department of Mines, Report for the Year Ending 06 30, 1922, pp. 175, 252–55.Google Scholar

²⁰ The frequency distributions of the sanitation ratings are reported in the Appendix, Section II. Since the ratings are bounded by zero and 100, we also ran regressions where we used a logit transformation of the sanitation ratings. The results were very similar to those reported in Tables 1 and 2.

²¹ The population variable is treated here as an exogenous variable affecting the company's costs of providing sanitation. The size of the town's population might be endogenous because the population is strongly correlated to the number of workers at the mine. We ran additional regressions where we replaced the population variable with population predictions from an instrument equation. The basic qualitative results are similar to those reported in Tables 1 and 2.

²² U. S. Coal Commission, “Bituminous Workers,” p. 1427.Google Scholar

²³ Fishback, Price, “Did Coal Miners ‘Owe Their Souls to the Company Store’? Theory and Evidence from the Early 1900s.” this JOURNAL, 46 (12 1986), p. 1013.Google Scholar

²⁴ Woodbridge, Dwight E., “Sanitation at Mining villages in the Birmingham District, Alabama,” U.S. Bureau of Mines, Technical Paper No. 33 (Washington, D.C., 1913), p. 21.Google Scholar

²⁵ Veillor, Lawrence, “Housing and Health,” The Annals of the American Academy of Political and Social Science, 37 (03 1911), p. 260.Google Scholar

²⁶ The INCORP variable was constructed by comparing the list of towns rated by the Public Health Service with a list of all incorporated communities in the United States in U.S. Bureau of the Census, Fourteenth Census of the United States taken in the Year 1920, Volume I, Population (Washington, 1921), PP. 178–314. All towns listed as incorporated communities were given a value of one; the remaining communities were given a value of zero. The Public Health Service in “Sanitation of 123,” p. 1611, felt that with some exceptions incorporated communities were generally not company towns. The decision-making process in the incorporated company towns was probably more like that of independent incorporated towns than of unincorporated company towns.Google Scholar

²⁷ U.S. Public Health Service, “Sanitation of 123,” pp. 1625–46.Google Scholar

²⁸ Fishback, , “Did Miners Owe Their Souls,” pp. 1011–29;Google Scholar and Fishback, Price V., “Workplace Safety during the Progressive Era: Fatal Accidents in Bituminous Coal Mining, 1912–1923,” Explorations in Economic History, 23 (07 1986), pp. 269–68.CrossRefGoogle ScholarPubMed

²⁹ U. S. Public Health Service, “Sanitation of 123,” pp. 1625–46.Google Scholar