I wish to thank Professors Ralph Beals of Amherst College and Charles Sackrey of Smith College, and two unknown referees for helpful comments on an earlier draft of this paper.

¹ Beard, Charles and Beard, Mary, The Rise of American Civilization one vol. ed. (New York: Macmillan, 1930)Google Scholar, ch. 18, stress the impact of the Civil War and Reconstruction on the political power of southern planters. Woodward, C. Vann, Origins of the New South, 1877–1913 (Baton Rouge: Louisiana State University Press, 1951)Google Scholar, ch. 11, stresses northern ownership of southern resources as do Stocking, George, Basing Point Pricing and Regional Development (Chapel Hill: University of North Carolina Press, 1954)Google Scholar, chs. 4–5, and Stover, John, The Railroads of the South (Chapel Hill: University of North Carolina Press, 1955).Google ScholarWoodman, Harold, King Cotton and His Retainers (Lexington: University of Kentucky Press, 1968), parts 4–6Google Scholar, discusses northern financing of the post-war cotton economy, while Sylla, Richard, “Federal Policy, Banking Market Structure, and Capital Mobilization in the United States, 1863–1913,” Journal of Economic History, XXIX (December 1969), 657–686, provides evidence on the role of the National Banking system in retarding southern rural development.Google Scholar

² Beard, Rise, ch. 18.

³ It should be emphasized that while Beard was concerned with how the war affected distribution between interest groups, classes, and regions, in this paper my focus is only on regions—the South versus die rest of the United States—and I ignore classes and interest groups, both within and between regions.

⁴ This paragraph is based heavily on Eugene Lerner, “Southern Output and Agricultural Income, 1860–1880,” and also his “Money, Prices and Wages in the Confederacy, 1861–1865,” both in Andreano, Ralph,, ed., The Economic Impact of the American Civil War rev. ed. (Cambridge, Mass.: Schenkman, 1967), pp. 31–60Google Scholar and 109–122. See also Engerman, Stanley, “Some Economic Factors in Southern Backwardness in the Nineteenth Century,” in Kain, John and Meyer, John, eds., Essays in Regional Economics (Cambridge: Harvard University Press, 1971), pp. 279–306.Google Scholar

⁵ Easterlin, Richard, “Regional Income Trends, 1840–1950,” in Fogel, Robert, ed., The Reinterpretation of American Economic History (New York: Harper and Row, 1971), pp. 38–49.Google Scholar

⁶ Computed using the value of southern commodity output presented in Stanley Engerman, “The Economic Impact of the Civil War,” in Andreano, pp. 188–209.

⁷ The argument in the text that for the South to recover, the cotton economy had to be prosperous, should not be misconstrued as an assertion that long-run specialization in cotton was necessarily desirable for the South. But given that cotton in fact dominated the southern economy in the immediate post-war years, prosperity for the South required prosperity in cotton. Whether income earned in cotton should have been reinvested there or elsewhere is a separate issue.

⁸ Friedman, Milton, “Price, Income, and Monetary Changes in Three Wartime Periods,” American Economic Review XLII (May 1952), 612–625, esp. 624–625.Google Scholar

⁹ This section draws heavily from Kindahl, James K., “Economic Factors in Specie Resumption: The United States, 1865–1879,” Journal of Political Economy, LXIX (February 1961), 30–48Google Scholar, and Williamson, Jeffrey G., American Growth and the Balance of Payments 1820–1913 (Chapel Hill: University of North Carolina Press, 1964)Google Scholar, chs. 2 and 4.

¹⁰ Nugent, Walter T. K., The Money Question During Reconstruction (New York: Norton, 1967), ch. 2.Google Scholar

¹¹ Kindahl, “Economic Factors,” 34–37.

¹² Taussig, Frank W., “International Trade Under Depreciated Paper—A Contribution to Theory,” Quarterly Journal of Economics, XXXI (May 1917), 380–403Google Scholar; Graham, Frank, “International Trade Under Depreciated Paper. The United States 1862–1879,” Quarterly Journal of Economics, XXXVI (February 1922), 220–273; Kindahl, “Economic Factors,” 30–48.Google Scholar

¹³ Kindahl, “Economic Factors,” 35.

¹⁴ Williamson, American Growth, pp. 124–125.

¹⁵ Ibid., pp. 118–134.

¹⁶ For export components, see Williamson, American Growth, Table B-7.

¹⁷ Rothstein, Morton, “America in the International Rivalry for the British Wheat Market, 1860–1914,” in Scheiber, Harry, ed., United States Economic History (New York: Alfred Knopf, 1964), pp. 290–308.Google Scholar

¹⁸ Computed from data in U.S. Department of Commerce, Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1957. For the 1850's, merchandise exports excluding gold and silver (Series U 1) were used, while total merchandise exports including gold and silver (Series U 64) were employed for the 1870's since under flexible exchange rates, gold is a commodity like any other.

¹⁹ Williamson, American Grototh, p. 32. Taussig, “International Trade,” 387, noted the possibility that wheat exports might be an independent (in his words “inherent”) force driving down the specie premium, but neither he nor Graham explored the phenomenon in any detail.

²⁰ Kindahl, “Economic Factors,” 30–35, argues that adjusted price ratios of unity were unnecessary for resumption, and even in the absence of capital inflows adjusted price ratios of from 1.08 and 1.18 would have been consistent with successful resumption. This was attributable, I would add, to the structural changes in world trade which had occurred since 1860 such as the expansion of the market for American exports discussed in the text.

²¹ To avoid confusion, it should be explicitly noted that the fact that the South was able to recover her pre-war share of foreign markets by the late 1870's, and that relative cotton prices during this period were not low by historical standards (see Engerman, “Some Economic Factors in Southern Backwardness,” p. 303) is irrelevant to my argument. Had gold not been driven below parity, ceteris paribus both cotton prices and the South's share of foreign markets would have been higher than they in fact were.

²² Data on farm prices and acreage are from U.S. Department of Commerce, Bureau of the Census, Historical Statistics, Series K 301 and 303. Data on production, in 500 pound bales, are from U.S. Department of Commerce, Bureau of the Census, Cotton Production and Distribution, Season of 1928–1929 Bulletin 166 (Washington: G.P.O., 1929), Table 31.Google Scholar

²³ For an attempt to explain the acreage devoted to various crops during the early years of the twentieth century using an equation similar to (2), see Nerlove, Marc, The Dynamics of'Supply: Farmers' Response to Price (Baltimore: Johns Hopkins, 1958), ch. 8Google Scholar, or his article “Estimates of the Elasticities of Supply of Selected Agricultural Commodities,” Journal of Farm Economics, XXXVIII (May 1956), 496–509.Google Scholar For similar supply and acreage equations estimated for the nineteenth century see Wright, Gavin, “An Econometric Study of Cotton Production, 1830–1860,” The Review of Economics and Statistics, LIII (May 1971), 111–120Google Scholar, and Fisher, Franklin and Temin, Peter, “Regional Specialization and the Supply of Wheat in the United States, 1867–1914,” The Review of Economics and Statistics, LIII (May 1970), 134–149.Google Scholar

²⁴ Exports and domestic consumption, in 500 pound bales, are from U.S. Department of Commerce, Bureau of the Census, Cotton Production, Table 31. Frickey's index is from Frickey, Edwin, Production in the United States, 1860–1914 (Cambridge: Harvard University Press, 1947), Table 4, Col. 2.Google Scholar

²⁵ Mitchell, B. R., Abstract of British Historical Statistics (Cambridge: Cambridge University Press, 1962), ch. 11, Table 8, col. 6, and Table 9, col. 3.Google Scholar

²⁶ New York cotton prices are from U.S. Department of Commerce, Bureau of the Census, Cotton Production, Table 31. The index of textile prices is from U.S. Congress, Senate Committee on Finance, Report by Mr. Aldrich on Wholesale Prices, Wages, and Transport 52d Cong., 2d Sess., Senate Report 1394, pt. 1 (Washington: G.P.O., 1893), Table 14, Col. 11.Google Scholar

²⁷ The index of British cloth prices is computed from the prices of British piece goods exported in Ellison, Thomas, The Cotton Trade of Great Britain (London: Frank Cass & Co., 1968), Table 2, col. 12.Google Scholar

²⁸ The model just described is a partial equilibrium model and consequently it suffers from the limitations common to all such models. Two of these limitations are important enough to note explicitly. First, the treatment of the gold price as entirely exogenous and unaffected by cotton exports is an oversimplification and it ignores the fact that cotton sales abroad helped to diminish the specie premium. To the extent that a falling gold price was partly attributable to cotton exports, the calculations in part V below overstate the effects of an exogenously determined decline in the gold price on cotton exports, prices, and incomes. Second, this model ignores the possibility of a “reversal of protection.” Appreciation of the dollar may have shifted the terms of trade against purchasers of American exports. If this decline in the terms of trade raised the demand for British textiles, it could have raised the demand for American cotton and thus offset the direct effects of a decline in PG on American cotton exports. For an interesting application of a general equilibrium model to the impact of the antebellum tariff on income distribution which concludes that a reversal of protection was unlikely, see Pope, Clayne, “The Impact of the Ante-Bellum Tariff on Income Distribution,” Explorations in Economic History, X (Summer 1972), 375–421, esp. p. 416.Google Scholar

²⁹ Equations (1) and (2) were estimated from 1869 on due to the absence of data on farm prices prior to that date.

³⁰ The presence of lagged endogenous variables in (1) and (2) biases the Durbin-Watson statistic toward a value of two, thus diminishing its usefulness in testing for serial correlation. (See Nerlove, Marc and Wallis, Kenneth, “Use of the Durbin-Watson Statistic in Inappropriate Situations,” Econometrica, XXXIV (January 1966), 235–238.Google Scholar) It is possible to do away with serial correlation directly by assuming that the error terms, u_i in each equation, follow an autoregressive scheme of the form u_i = ku_i-1+e_i, where the e_i are assumed uncorrelated, and fitting equations which minimize the sum of the squared e_i. (For a simple description of this procedure see Frank, Charles, Statistics and Econometrics (New York: Holt Rinehart and Winston, 1971), pp. 279–283.Google Scholar Equations using this technique were estimated, but since they yielded the same conclusions as the equations in the text, the results are not presented.

³¹ The F test revealed all of the following equations to be highly significant. The numbers in parentheses are t ratios; all values of R² have been adjusted for degrees of freedom.

³² Supply equations using prices lagged two years and using undeflated prices were estimated, and still the price coefficient remained insignificant.

³³ Equations using unlagged prices with acreage lagged one year, and using prices lagged two years and acreage lagged one year were also estimated, but with no better results.

³⁴ See Temin, Peter, “The Causes of Cotton-Price Fluctuations in the 1830's,” The Review of Economics and Statistics, XLIX (November 1967), 463–470, and Gavin Wright, “An Econometric Study of Cotton Production, 1830–1860.” Wright finds an insignificant price elasticity of supply for five southern states, and a significant but very low price elasticity for total southern cotton supply.CrossRefGoogle Scholar

³⁵ Temin, “Causes,” 468.

³⁶ The equation is H

³⁷ McGouldrick, Paul, New England Textiles in the Nineteenth Century (Cambridge: Harvard University Press, 1968), p. 144. Because of the poor performance of the TEX variable, an index of the output of the New England textile industry based on McGouldrick's output series (McGouldrick, pp. 142–143) was also employed in the home demand equations, but with no better results.Google Scholar

³⁸ British exports of cotton piece goods, taken from Ellison, Cotton Trade, Table 2, were substituted for variable BR in the export equations, but performed no better.

³⁹ The reader should be aware that the elasticities found in (10a) and (10c) are not comparable to those found by Wright, “Econometric Study,” pp. 112–113, for the antebellum period. Wright explained total British cotton imports and his price term was a weighted average of the prices of American and East Indian cotton, while our equation makes American exports depend upon their own price. Since American cotton had better substitutes than all cotton, the price elasticities indicated by (10a) and (10c) are somewhat higher than those found by Wright.

⁴⁰ To test the hypothesis that the American share of the British cotton market was sensitive to the price of American cotton relative to other cotton prices, a market share equation was estimated for the period 1866–1879. A number of different forms were tried, the most successful being:

where I is total British cotton imports, US is imports from the United States, T a time trend, and PA and PI are the prices of American and East Indian cotton respectively. All coefficients are significant, and the equation indicates that the American market share was indeed quite sensitive to the price spread between American cotton and that from other sources.

⁴¹ In 1880, the South contained about five percent of all cotton spindles in the United States. See Copeland, Melvin, The Cotton Manufacturing Industry of the United States (Cambridge: Harvard University Press, 1912), p. 34.Google Scholar

⁴² The reader should again be warned that the figures in Tables 2 and 3, because they are derived from a partial equilibrium model, are likely to overstate the effects of capital inflows and Western exports on the specie premium and hence on cotton prices and incomes. Some of the appreciation of the dollar must have been attributable to the expansion of cotton exports (although cotton fell from eightythree percent of the value of all merchandise exports in 1866 to twenty-three percent in 1879). Also, the changes in cotton prices and incomes depicted in Tables 2 and 3 may have been offset by additional expenditures on southern cotton due to an expansion of demand for British textiles which exchange appreciation could have generated.

⁴³ Thomas, Brinley, Migration and Economic Growth (Cambridge: Cambridge University Press, 1954)Google Scholar, chs. 7–11. Cairncross, A. K., Home and Foreign Investment, 1870–1913 (Cambridge: Cambridge University Press, 1953), chs. 7–8.Google Scholar