Long-term trends in agricultural specialization in the United States: Some preliminary results

Agricultural History

Volumn 70, Issue 1, 1996, Pages 90-101

  Gregson, Mary Eschelbach ^a  

^a KNOX COLLEGE   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

AGRICULTURAL HISTORY; CENSUS; CROP CHOICE; HISTORICAL STUDY; SPECIALISATION TREND; TRANSPORT COST;

USA, (MIDWEST);

EID: 0029799212     PISSN: 00021482     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (22)

1. Additional data on railroad growth and relative prices after the Civil War appear in Mary Eschelbach Gregson, "Rural Response to Increased Demand," Journal of Economic History 53 (June 1993): 332-45, and Mary Eschelbach Gregson, "Strategies for Commercialization: Missouri Agriculture, 1860-1880" (Ph.D. diss., University of Illinois, Urbana-Champaign, 1993), chapter 2.
2. Additional data on railroad growth and relative prices after the Civil War appear in Mary Eschelbach Gregson, "Rural Response to Increased Demand," Journal of Economic History 53 (June 1993): 332-45, and Mary Eschelbach Gregson, "Strategies for Commercialization: Missouri Agriculture, 1860-1880" (Ph.D. diss., University of Illinois, Urbana-Champaign, 1993), chapter 2.
3. My emphasis is on livestock, grain, and soybean production in the Mississippi River valley states of Illinois and Missouri. Consideration of other regions involves the addition of other crop categories, most notably dairy products, to the analysis. In this paper I limit geographical scope to accommodate analytical depth. See also Morton D. Winsberg, "Agricultural Specialization in the United States Since World War II," Agricultural History 56 (October 1982): 692-701. In a paper on the importance of agricultural land from a national income accounting perspective, T. W. Schultz emphasizes the role of inputs such as motorized equipment, fertilizer, and lime in maintaining agricultural productivity, despite the small contribution of agricultural land per se in "The Declining Importance of Agricultural Land," Economic Journal 61 (December 1951): 725-40.
4. My emphasis is on livestock, grain, and soybean production in the Mississippi River valley states of Illinois and Missouri. Consideration of other regions involves the addition of other crop categories, most notably dairy products, to the analysis. In this paper I limit geographical scope to accommodate analytical depth. See also Morton D. Winsberg, "Agricultural Specialization in the United States Since World War II," Agricultural History 56 (October 1982): 692-701. In a paper on the importance of agricultural land from a national income accounting perspective, T. W. Schultz emphasizes the role of inputs such as motorized equipment, fertilizer, and lime in maintaining agricultural productivity, despite the small contribution of agricultural land per se in "The Declining Importance of Agricultural Land," Economic Journal 61 (December 1951): 725-40.
5. German rural economists, following the lead of Johann Heinrich von Thünen, were among the first to recognize that as transport costs fall inherent soil suitability becomes a more important factor in the farm-level crop mix decision. More recent discussions of specialization and diversification in the economic geography literature include Howard Gregor, Geography of Agriculture: Themes in Research (Englewood Cliffs, N.J.: Prentice-Hall, 1970); Michael Chisholm, Rural Settlement and Land Use (London: Hutchinson, 1979);
6. German rural economists, following the lead of Johann Heinrich von Thünen, were among the first to recognize that as transport costs fall inherent soil suitability becomes a more important factor in the farm-level crop mix decision. More recent discussions of specialization and diversification in the economic geography literature include Howard Gregor, Geography of Agriculture: Themes in Research (Englewood Cliffs, N.J.: Prentice-Hall, 1970); Michael Chisholm, Rural Settlement and Land Use (London: Hutchinson, 1979);
7. and Earl O. Heady and Uma Srivastava, Spatial Sector Programming Models in Agriculture (Ames: Iowa State University Press, 1975).
8. Other models proposed for understanding nineteenth-century crop choice include Gavin Wright and Howard Kunreuther's safety-first model in "Cotton, Corn and Risk in the Nineteenth Century," Journal of Economic History 35 (September 1975): 526-51, and Robert McGuire's income variability model in "A Portfolio Analysis of Crop Diversification and Risk in the Cotton South," Explorations in Economic History 17 (October 1980): 342-71, and "Economic Causes of Late Nineteenth Century Agrarian Unrest: New Evidence," Journal of Economic History 41 (December 1981): 835-52.
9. Other models proposed for understanding nineteenth-century crop choice include Gavin Wright and Howard Kunreuther's safety-first model in "Cotton, Corn and Risk in the Nineteenth Century," Journal of Economic History 35 (September 1975): 526-51, and Robert McGuire's income variability model in "A Portfolio Analysis of Crop Diversification and Risk in the Cotton South," Explorations in Economic History 17 (October 1980): 342-71, and "Economic Causes of Late Nineteenth Century Agrarian Unrest: New Evidence," Journal of Economic History 41 (December 1981): 835-52.
10. Other models proposed for understanding nineteenth-century crop choice include Gavin Wright and Howard Kunreuther's safety-first model in "Cotton, Corn and Risk in the Nineteenth Century," Journal of Economic History 35 (September 1975): 526-51, and Robert McGuire's income variability model in "A Portfolio Analysis of Crop Diversification and Risk in the Cotton South," Explorations in Economic History 17 (October 1980): 342-71, and "Economic Causes of Late Nineteenth Century Agrarian Unrest: New Evidence," Journal of Economic History 41 (December 1981): 835-52.
11. The six townships were originally chosen as part of the Bateman-Foust samples. See Fred Bateman and James D. Foust, "A Sample of Rural Households Selected from the 1860 Manuscript Censuses," Agricultural History 48 (January 1974): 75-93. They are Rocky Fork (Boone County), Jackspn (Camden County), Freeborn (Dunklin County), Prairie (Montgomery County), Salt River (Rails County), and Mt. Pleasant (Scotland County). The data used in this section of the paper were collected with the support of the National Science Foundation (SES 91-00517). For further details, see Mary Eschelbach Gregson, "Specialization in Late-Nineteenth-Century Midwestern Agriculture: Missouri as a Test Case," Agricultural History 67 (Winter 1993): 16-35, and Gregson, "Strategies for Commercialization," chapters 3-4.
12. The six townships were originally chosen as part of the Bateman-Foust samples. See Fred Bateman and James D. Foust, "A Sample of Rural Households Selected from the 1860 Manuscript Censuses," Agricultural History 48 (January 1974): 75-93. They are Rocky Fork (Boone County), Jackspn (Camden County), Freeborn (Dunklin County), Prairie (Montgomery County), Salt River (Rails County), and Mt. Pleasant (Scotland County). The data used in this section of the paper were collected with the support of the National Science Foundation (SES 91-00517). For further details, see Mary Eschelbach Gregson, "Specialization in Late-Nineteenth-Century Midwestern Agriculture: Missouri as a Test Case," Agricultural History 67 (Winter 1993): 16-35, and Gregson, "Strategies for Commercialization," chapters 3-4.
13. The six townships were originally chosen as part of the Bateman-Foust samples. See Fred Bateman and James D. Foust, "A Sample of Rural Households Selected from the 1860 Manuscript Censuses," Agricultural History 48 (January 1974): 75-93. They are Rocky Fork (Boone County), Jackspn (Camden County), Freeborn (Dunklin County), Prairie (Montgomery County), Salt River (Rails County), and Mt. Pleasant (Scotland County). The data used in this section of the paper were collected with the support of the National Science Foundation (SES 91-00517). For further details, see Mary Eschelbach Gregson, "Specialization in Late-Nineteenth-Century Midwestern Agriculture: Missouri as a Test Case," Agricultural History 67 (Winter 1993): 16-35, and Gregson, "Strategies for Commercialization," chapters 3-4.
14. See Gregson, "Specialization in Late-Nineteenth-Century Midwestern Agriculture," 16-35, for more about the index. The crops included are cotton, wheat, rye, corn, oats, tobacco, pulse, Irish potatoes, sweet potatoes, barley, buckwheat, orchard produce, market gardens, butter, hay, flaxseed, sorghum molasses, maple molasses, value added to cattle, and value added to hogs as reported on the census manuscripts for each farm in the sample. The index is calculated as the sum of squares of the shares of crops in gross farm income.
15. For more information on the derivation of the soil type variable, see Gregson, "Rural Response to Increased Demand," 336-38.
16. See Gregson, "Rural Response to Increased Demand," 340 for a similar ANOVA analysis. ANOVA, like ordinary least squares regression, describes the relationship between the dependent variable (here the index of specialization) and independent variables (here the number of soil types on the farm and other characteristics of farm and farmer). ANOVA is preferred when the main independent variables are categorical (soil types) rather than continuous (age). The ANOVA results indicate that the index of specialization for farms with few soil types is different from the index for farms with several soil types. The difference for 1880 is significant at the 10 percent level (F-ratio = 2.37 for 1880 but F-ratio = 1.36 for 1860). A complete set of ANOVA results is available from the author.
17. Jeffrey Finke and Earl R. Swanson, "Diversification in Illinois Crop Production: 1938-1970," Illinois Agricultural Economics 13 (January 1973): 8-11.
18. In my dissertation and in "Specialization in Late-Nineteenth-Century Midwestern Agri-culture," I criticize the use of aggregated data to make farm-level conclusions. I have not changed my opinion. In this paper I hope only to show that there are good reasons to pursue farm-level research into the determinants of twentieth-century specialization, including the role of fertilizers in that process.
19. U.S. Department of Commerce, Bureau of the Census, Fourteenth Census of the United States Taken in the Year 1920, Agriculture, vol. 5 (Washington, D.C.: Government Printing Office, 1922), 702. I would like to use the census figures on the share of each crop in gross farm income, but the numbers in the 1910 and 1920 volumes on agriculture are not consistent. The 1910 census (vol. 5, p. 540) reports that the value of corn was 26.2 percent of the total value of production. A similar table in the 1920 report (vol. 5, p. 702), which is meant to compare the value of specific crops in the two census years, says that corn accounted for 50.9 percent of the value of production in 1909. No reconciliation of the figures is offered in the notes.
20. U.S. Department of Interior, Office of the Census, Twelfth Census of the United States Taken in the Year 1900, Agriculture, Part I: Farms, Livestock, and Animal Products, vol. 5 (Washington, D.C.: Census Office, 1902), 496.
21. The data used here were digitized by Thomas Cooley, Stephen DeCanio, and M. Scott Matthews as part of the Agricultural Time Series-Cross Section Dataset (ATICS). See Cooley, DeCanio, and Matthews, "ATICS: An Agricultural Time Series-Cross Section Dataset," National Bureau of Economic Research Working Paper No. 197 (Cambridge, Mass.: NBER, 1977). I am deeply indebted to Steve DeCanio who graciously dug out a 20-year-old data tape and uploaded the data onto the Internet for me. (Copies of the data are available from me in dBase format.)
22. The correlation coefficient between five-year averages for fertilizer use and five-year averages for share of corn and soybeans in grain income was significant at the 95 percent level. For a recent history of soybean growing and marketing, see John C. Hudson, Making the Corn Belt (Bloomington: Indiana University Press, 1994), chapter 10.