"Machinery has completely taken over": The diffusion of the mechanical cotton picker, I949-I964

Journal of Interdisciplinary History

Volumn 39, Issue 1, 2008, Pages 65-96

  Heinicke, Craig ^a   Grove, Wayne A ^b  

^a UNIVERSITY OF RICHMOND   (United States)

^b LE MOYNE COLLEGE   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 50949126820     PISSN: 00221953     EISSN: None     Source Type: Journal    
DOI: 10.1162/jinh.2008.39.1.65     Document Type: Review

References (74)

1. The quotation in the title was spoken by Hughley M. Jones of Panola County, Mississippi, in Goin' to Chicago, a California Newsteel, produced by George King and Associates in 1994. After World War II, the South played a key role in income convergence. See Francesco Caselli and Wilbur John Coleman II, "The U.S. Structural Transformation and Regional Convergence: A Reinterpretation," Journal of Political Economy, CIX, (2001), 584-616;
2. Kris Mitchener and Ian McLean, "U.S. Regional Growth and Convergence, 1880-1960, Journal of Economic History LIX, (1999), 1016-1042;
3. Michelle Connolly, "Human Capital and Growth in the Postbellum South: A Separate but Unequal Story," ibid., LXIV (2004), 363-399;
4. Gavin Wright, Old South, New South (New York, 1986).
5. On rural change, see Pete Daniel, Breaking the Land: The Transformation of Cotton, Tobacco, and Rice Cultures since 1880 (Urbana, 1985);
6. Jack Temple Kirby, Rural Worlds Lost: The American South 1920-1960 (Baton Rouge, 1987).
7. On paternalism, see Lee Alston and Joseph Ferrie, Southern Paternalism and the American Welfare State (New York, 1999).
8. In Wright's view, federal agricultural programs led to tenant displacement, and minimum wage laws reduced employment growth in labor-intensive industries ("The Economic Revolution in the American South," Journal of Economic Perspectives, 1 [1987], 161-178).
9. See also Warren Whatley, "Labor for the Picking, the New Deal in the South," Journal of Economic History, XLIII (1983), 909-929.
10. On schooling as a cause of migration, see Robert Margo, Race and Schooling in the South, 1880-1950 (Chicago, 1990);
11. on civil rights and economics, Wright, "The Civil Rights Revolution as Economic History," Journal of Economic History, LIX (1999), 267-289.
12. Acemoglu and James Robinson, Persistence of Power, Elites and Institutions (Cambridge, Mass., 2006), 49, credit the "collapse" of southern equilibrium to changes in incentives favoring previous political strategies.
13. Douglass North, "Economic Performance Through Time," American Economic Review, LXXXIV (1994), 360.
14. North's definition of institution distinguishes the concept from physical (and sometimes more transitory) manifestations, such as specific governments or their branches. Some economic historians prefer a more expansive definition - for example, Avner Grief, Institutions and the Path to the Modern Economy (New York, 2006)), 15-24 - although usage in this literature tends to follow North more closely.
15. See Daron Acemoglu and Simon Johnson, "Unbundling Institutions, "Journal of Political Economy, CXIII (2005), 950.
16. The social consequences of cotton-picker diffusion include effects on wage convergence between the South and North, the civil rights movement, the migration of African Americans, and changes in racial economic inequality. See Wright, Old South;
17. Daniel M. Johnson and Rex R. Campbell, Black Migration in America (Durham, N.C., 1981);
18. Heinicke, "African-American Migration and Mechanized Cotton Harvesting, 1950-60," Explorations in Economic Histor)', XXXI (1994) 501-520;
19. idem, "One Step Forward: African-American Married Women in the South, 1950-1960," Journal of Interdisciplinary History, XXXI (2000), 43-62;
20. Richard Day, "The Economics of Technological Change and the Demise of the Sharecropper," American Economic Review, LVII (1967), 427-449;
21. Donald Holley, The Second Great Emancipation: The Mechanical Cotton Picker, Black Migration, and How They Shaped the Modern South (Fayetteville, 2000);
22. Willis Peterson and Yoav Kislev, "The Cotton Harvester in Retrospect: Labor Displacement or Replacement?" Journal of Economic History, XLVI (1986), 199-216;
23. Alston and Ferrie, Southern Paternalism;
24. idem, "Paternalism in Agricultural Labor Contracts in the U.S. South: Implications for the Growth of the Welfare State," American Economic Review, LXXXIII (1993), 852-876.
25. Wright, "Economic Revolution, "164-171.
26. James Street, The New Revolution in the Cotton Economy (Chapel Hill, 1957), 34. The focus of this article is on diffusion, but institutions may well have blunted the incentive to invent the machine.
27. Alston and Ferrie, Southern Paternalism.
28. See Holley, Second Great Emancipation, 77, for the view of mechanization as ending small-scale farming in the South.
29. Connelly, "Human Capital," argues that racial discrimination in schooling undermined human-capital formation of both African Americans and whites in the South. Street, New Revolution, 238, asserts that the South was more "tradition bound" than the West.
30. Moses Musoke and Alan Olmstead, "The Rise of the Cotton Industry in California: A Comparative Perspective," Journal of Economic History, XLII (1982), 385-412,
31. cite capital-market imperfections and numerous other factors. On institutions and cotton-seed quality, see Paul Rhode and Olmstead, "Hog-Round Marketing, Seed Quality, and Government Policy: Institutional Change in U.S. Cotton Production, 1920-60," ibid., LXIII (2003), 447-488.
32. Wright, Old South.
33. Alston, "The Wright Interpretation of Southern U.S. Economic Development: A Review Essay of Old South, New South by Gavin Wright," Agricultural History, LXI (1987), 65-67,
34. questions the relative strength of the New Deal vs. outmigration. Musoke and Olmstead, "Rise of the Cotton Industry," argue for the importance of environmental conditions in determining cotton mechanization. Whatley, "A History of Mechanization in the Cotton South: Institutional Hypothesis," Quarterly Journal of Economics (1985), 1191-1215, focuses on share tenancy, although his argument applies mainly to invention.
35. On the New Deal, see Bruce Schulman, From Cotton Belt to Sunbelt: Federal Policy, Economic Development, and the Transformation of the South, 1938-1980 (New York, 1991).
36. For detail about invention, see Holley, Second Great Emancipation 93-158; Street, New Revolution, 104-127.
37. Maier, "Economic Analysis," uses the 10% figure to designate the initial adoption of the cotton harvester. Zvi Griliches, "Hybrid Corn: An Exploration in the Economics of Technological. Change," Econometrica, XXV (1957), 501-522, also uses 10% of the "ceiling acreage" on which hybrid corn was planted to denote when the experimental stage for this innovation was passed.
38. Note that differential rates of adoption by state based on a 20% mechanization rate showed a longer delay of the Delta relative to the West and a smaller lag within the South than the 10% threshold implies. For the 20% threshold, the lag is about equal between the West and the Delta, as well as between the Delta and the Southeast (four years). The speed of diffusion, however, is best suited to addressing this question. At least one Delta state experienced initial adoption before a western state (Louisiana vs. New Mexico); the ordering is preserved at 20% mechanization.
39. In a study of 265 diffusion cases by Arnulf Grübler, "Diffusion, Long-Term Patterns and Discontinuities," Technological Forecasting and Social Change, XXXIX (1991), 159-180, a 10-90 lag range of fifteen to thirty years bounds the modal cell for diffusion lags in cells of fifteen-year periods; the mean 10-90 lag is forty-one years. The automobile diffused faster, the tractor a good deal more slowly
40. (Olmstead and Rhode, "Reshaping the Landscape: The Impact and Diffusion of the Tractor in American Agriculture, 1910-60," Journal of Economic History, LXI [2001], 663-669). The cotton-harvester diffusion pattern lacked the often-observed S-shape. See Griliches, "Hybrid Corn"; Bronwyn Hall and Beethika Kahn, "Adoption of New Technology," National Bureau of Economic Research, working paper 9730 (Cambridge, Mass., 2003)).
41. For the 1957 crop year, see Clay Lyle, "Labor and Technology on Selected Cotton Plantations in the Delta Area of Mississippi, 1953-1957," Mississippi Agricultural Experiment Station. Bulletin 575 (April 1959), 6, 17. For government cotton programs and mechanization, see Maier, "Economic Analysis." Note that in many USDA publications, Arkansas, Louisiana, and Mississippi were categorized as "Delta states," although only parts of these states fell within the Mississippi Delta sub-region proper. About half of the cotton in Arkansas and Mississippi and about 80% in Louisiana were produced within the fertile Mississippi Delta. The state of Tennessee defies clear categorization. For the Bracero Program that temporarily slowed western mechanization, see Grove, "The Mexican Farm Labor Program, 1942-1964," Agricultural History, XXV (1996), 302-320.
42. Margo, Race and Schooling, notes that parental demand for schooling was instrumental in African-American, progress.
43. According to John Donahue, James Heckman, and Petra Todd, "The Schooling of Southern Blacks: The Roles of Legal Activism and Private Philanthropy," Quarterly Journal of Economics, CXVII (2002), 225-262, schooling had improved by 1960, but regional differences remained stark.
44. Maier, "Economic Analysis," 28-37; Musoke and Olmstead, "Rise of the Cotton Industry," 387-391. After the turn of the century, the boll weevil spread throughout the high moisture areas of the Southeast, the Delta, and Texas but not to the semi-arid regions. Western growers also benefited from the federal government's Mexican Farm Labor, or Bracero, Program. See Heinicke and Grove, "Labor Markets, Regional Diversity, and Cotton Harvest Mechanization in the Post-WWII United States," Social Science History, XXIX (2005), 269-297.
45. An operator in the North Carolina Piedmont gave two reasons for returning a picker to the dealer, each, related to the steepness of slopes or narrowness of terraces on slopes. See J. Gwyn. Sutherland and.Brooks James, Mechanical Cotton Harvesting in North Carolina (Chapel Hill, 1947), 7. Early implementation of the "one variety" system helped to produce great cotton yields in the far West (Olmstead and Rhode, "Hog-Round," 457-459).
46. It is misleading to think in terms of an exogenous "threshold" size needed for adoption. International Harvester's assessment of threshold size was hardly stable. Operation size was also endogenous with respect to tractor adoption, which led to a change in the size of operations. See Gilbert Fite, Cotton Fields No More (Chapel Hill, 1984), 184-185. On that issue nationally, see Olmstead and Rhode, "Reshaping."
47. Holley, Second Great Emanicpation, 111-118;
48. Street, New Revolution, 120-123. The machine-harvest cost estimates in Maier's "Economic Analysis" explicitly include yields, given that higher output per unit of land, with a given velocity of the machine through the field, resulted in lower average costs (113-114, 116, Appendix G). Owners may have used the picker in their own ways (sharing and renting), but because cotton could remain in the fields only for a certain amount of time before it deteriorated, the length of time for which fixed costs could be spread over output was limited.
49. See the following three memos located in the International Harvester Company archives, State Historical Society of Wisconsin, Madison: J. A. Hamilton, "Cotton Picker Potential Market," March 10, 1947, Market Analysis and Consumer Research Department, Farm Equipment Division, International Harvester Company; M. J. Steitz, "The Market for Cotton Pickers," March 20, 1951, Market Research Section, Consumer Relations Department, Farm Equipment Division, International Harvester Company; idem, "The Total Industry Potential Market for Cotton Pickers and Strippers," January 20, 1953, Market Research Section, Consumer Relations Department, Farm Equipment Division, International Harvester Company.
50. U.S. Dept. of Commerce, Historical Statistics of the U.S., Colonial Times to 1970 (Washington, 1970), Part 1, Series K 458, 500. Olmstead and Rhode referred to this increase in yield as the "other revolution" ("Hog-Round," 480). For contemporary under-predictions, see "Appraisal of Agricultural Productive Capacity in Arkansas, 1955," Form 3 National Archives RG 164, Records of the Office of Experiment Stations, Arkansas Agricultural Experiment Station;
51. Raymond P. Christensen and Ronald O. Aines, Economic Effects of Acreage Control Programs in the 1950s, (1962) USDA, ERS, Agricultural Economic Report 18.
52. Data on the relevant variables, except for the percentage of cotton harvested mechanically, are unavailable after 1964.
53. See U.S. Department of Agriculture, Statistics on Cotton and Related Data 1920-1973 (Washington, D.C. 1974), Tables 48-61, 64-77.
54. Maier, "Economic Analysis," 104-162, estimated machine harvest costs for the spindle picker from 1949 to 1964 based on (1) fixed and variable costs, such as purchase prices, depreciation schedules, interest rates, storage costs, insurance, taxes, tractor costs, repairs, lubricants, fuel, and labor; (2) estimates of the hours of potential use by state during a typical harvest season; (3) field waste relative to hand picking; and (4) grade loss relative to hand picking. Correction to these data came from Whatley, "New Estimates of the Cost of Harvesting Cotton: 1949-1964," Research in Economic History, XIII (1991), 199-225, who added the cost of the machine operator's "helper," and Grove, "The Economics of Cotton Harvest Mechanization in the United States, 1920-1970," unpub. Ph.D. diss. (University of Illinois, Urbana-Champaign, 2000), 25-137, who added ginning costs associated with machine harvesting. Grove also provides a time series of hand-harvest costs to match the Maier machine-cost data that (1) converts piece-rate wages to cash wages per lb. of lint and (2) estimates non-wage costs for resident laborers, day-haul workers, domestic migrants, and foreign contract workers. Annual state hand-harvesting costs are obtained by combining state wage and non-wage labor expenses, weighted according to the type of labor as estimated by the state for each year. Cash wages (used in other studies), paid per 100 lb. of unprocessed "seed-cotton," do not reflect growers' total hand-harvest costs because they do not reflect the unit costs per pound of lint and because they ignore other employer expenses. No measure of operation size is available for the full panel. Farm size, which is recorded only for agricultural census years, is misleading, since sharecropper plots are entered as farms in the agricultural census. Hand costs are measured per lb. of lint - the same unit as machine-harvest costs.
55. Grove and Heinicke, "Better Opportunities or Worse? The Demise of Cotton Harvest Labor in the US, 1949-1964," Journal of Economic History, LXIII (2003), 753; Wright, "Economic Revolution," 171. Olmstead and Rhode, "Reshaping," 682-691, find that substitute draft-animal prices are endogenous with respect to tractor diffusion. Despite the fact that the wage rates measured herein are lagged one year, a high degree of autocorrelation in this series suggests that testing for endogeneity is in order.
56. -4). The average annual increase in machine harvesting was 4.3 percentage points.
57. The estimates hold up reasonably well with different specifications, including one with a time trend. One exception is that the version with the log of the odds ratio as the dependent variable results in a positive machine-cost coefficient in the full (year and state) fixed-effects model, although inclusion of a time trend restores the negative coefficient. There is no reason to suggest that the functional form is logistic, even with respect to time, however (see Figure 1).
58. Regressing each potentially endogenous variable on the exogenous variables, with the nonagricultural wage rate (lagged) added as a regressor for identification, finds the F-statistic for these tests to be 24.98 (df = 1, 184; p-value = 0.0018) for harvest wages, and 13.9 (1, 184; p-value = 0.0003) for overhead costs.
59. See Day, "Technological Change"; Grove and Heinicke, "Better Opportunities"; Wright, Old South. The variable federal grants to states is problematical because it is a weak instrument for harvest wages; the correlation coefficient with cotton-picking wages is 0.2. Given that these data include a time-series element, there is the possibility that the variables are nonstatiouary. Insufficient time periods exit to perform formal tests for unit roots. Along the lines of recommendations by A. Levin and C. F. Lin, "Unit Root Test in Panel Data: Asymptotic and Finite-Sample Properties," Economics Working Paper 93-56 (University of San Diego, 1993);
60. Levin, Lin, and C.J. Chu, "Unit Root Tests in Panel Data: Asymptotic and Finite Sample Properties," Journal of Econometrics, CVIII (2002), 1-24;
61. Badi H. Baltagi, Econometric Analysis of Panel Data (New York, 1995), 235-238, the variable was regressed on its lag with fixed effects (including an intercept, and with or without a time trend). Whereas Levin and Lin suggest a formal-hypothesis test, we simply compared the standard errors to see whether the coefficient on the lagged value is "close" to 1, with inconclusive results. Despite the lack of evidence for a nonstationary harvesting variable, we ran a first-differenced equation one (available upon request). The machine-cost coefficient is negative but not statistically significant (10% or lower) with that specification.
62. Regressions of the log of cotton-harvest wages on a time trend show a negative and significant (10 percent or lower) coefficient for our period except for Alabama, Arizona, Georgia, and South Carolina.
63. -6.57), controlling productivity with the U.S. manufacturing productivity index.
64. -1.26. If this figure were statistically significant at usual levels, the southern yield growth would be almost three-quarters of a percentage point per year lower than that in the West.
65. The measure for educational attainment in the West is likely to register values for large urban centers to a much greater extent than in the South, given the population distributions. In 1950, the West's urbanization rates were 70% for whites and 90% for blacks; the South's were 49% for whites and 48% for blacks (Johnson and Campbell, Black Migration, 132). Hence, the differences between the West and the South may be overestimated. Working in the other direction, the quality of southern schools was poor, despite improvements by 1960.
66. The combination of machine costs and schooling attainment "over-predicts" the difference in machine harvesting between these two states in 1951. Since the schooling-attainment figures for many of the states in the South show little variation, the predicted differences across states due to that variable would be close to zero, and most of the difference in mechanization, would be due to other factors, such as machine-cost differences.
67. In the 1950s, the federal government's Soil Bank displaced labor in the South but not in the West (Heinicke and Grove, "Labor Markets," 288-289).
68. Peterson and Kislev, "Cotton Harvester in Retrospect," 213-215, and Holley, Second Great Emancipation, 171-175, find that the decline in labor supply was greater than demand, and Heinicke, "African-American Migration," 513-517, finds that harvest mechanization was not necessarily the cause of migration in the 1950s. Contrast Grove and Heinicke, "Better Opportunities," 754-761, and Day, "Technological Change," 441-443, for whom decreasing labor demand was predominant.
69. Daniel, Breaking the Land, 239;
70. Fite, Cotton Fields No More.
71. Peterson and Kislev, "Cotton Harvester in Retrospect," 213-215; Holley, Second Great Emancipation, 171-174;
72. Grove and Heinicke, "Better Opportunities," 761; Alston and Ferrie, Southern Paternalism, 117-121.
73. Street, New Revolution.
74. Musoke and Olmstead, "Rise of the Cotton Industry," 386-389.