My thanks, without implication of responsibility for remaining errors, go to Galen Burghardt, Ronald Ehrenberg, Stanley Engerman, Robert Fogel, Claudia Goldin, and an anonymous referee of this Journal.

¹ There are numerous anthologies and commentaries on the issue of the cause of the Civil War, but the following are useful, representative statements containing references for further reading: Pressly, Thomas J., ed., Americans Interpret Their Civil War (Princeton: Princeton University Press, 1964)Google Scholar; Rozwenc, Edwin C., ed., The Causes of the American Civil War (Boston: D. C. Heath, 1961)Google Scholar; Stampp, Kenneth M., ed., The Causes of the Civil War (Englewood Cliffs: Prentice Hall, 1965)Google Scholar David M. Potter, “The Literature on the Background of the Civil War” in Potter, David M., The South and the Sectional Conflict (Baton Rouge: Louisiana State University Press, 1968)Google Scholar; and (a review of recent literature) Lee Benson, “Explanations of American Civil War Causation: A Critical Assessment and a Modest. Proposal to Reorient and Reorganize the Social Sciences” in Benson, Lee, Toward the Scientific Study of History (Philadelphia: J. B. Lippincott, 1972).Google Scholar

² A good statement of the conflicting objectives which each of the parties faced and therefore the necessity of recognizing some trade-off or compromise between them is contained in David M. Potter, “Why the Republicans Rejected Both Compromise and Secession” in Knowles, George Harmon, ed., The Crisis of the Union (Baton Rouge: Louisiana State University Press, 1965).Google Scholar

³ An excellent summary and explanation of this subject is contained in Fogel, Robert W. and Engerman, Stanley. L., Time on the Cross (Boston: Little, Brown and Co., 1974).Google Scholar

⁴ Fogel and Engerman, Time on the Cross, p. 96.

⁵ Goldin, Claudia Dale, “The Economics of Emancipation” The Journal of Economic History, XXXIII (March 1973).Google Scholar This estimate compares very closely with the assessment of slaves (adjusted for the included populations) made by the Confederate government in July of 1861. See Todd, Richard Cecil, Confederate Finance, (Athens: Univ. of Georgia Press, 1954) p. 199.Google Scholar

⁶ An important corollary of this discussion is that the primary and predominant effect of the abolition of slavery (without compensation) would most likely be a transfer of earnings from slave owners to the freemen. Third parties, such as northerners, would be affected little on net. Slavery is implicitly a method of distributing the returns from labor differently than the distribution of benefits which result from free labor. In either case, however, the value of the labor effort is derived from the resulting output. The demand for cotton or any other antebellum southern crop should have been independent of the form of labor input except insofar as it affected the price of the final product (cotton textiles, for example). One might conjecture that slavery—as opposed to free labor—would produce cotton much cheaper because the grower would need only to provide the subsistence for slaves while he would be required to pay the (much higher) prevailing wage rates for free labor. That view, however, is very misleading. The real cost of slave labor to its employer was slave maintenance plus the opportunity cost of the investment in acquiring him. (The same phenomenon appeared when slaves were hired by the year rather than purchased. The hire rate was at least several times as large as room and board expenses for the slave.) Furthermore the opportunity cost of using slaves tended to rise to the level of free labor because, when it was less, potential employers had reason to purchase slaves to capture those extra profits thereby bidding up the price of the slaves. Thus the cost of producing cotton tended to be equalized whichever labor system was employed in providing it and northerners had little economic stake in the institutions which the South utilized to organize its economy. This argument should be hedged, however, by the possible existence of economies of scale in the utilization of slaves not otherwise possible with free labor. See Engerman, Stanley L.“Some Considerations Relating to Property Rights in Man,” The Journal of Economic History, XXXIII (March 1973). Engerman's paper is the best comprehensive treatment currently available of the role of slavery in its broader context as a property right. See also Fogel and Engerman, Time on the Cross, pp. 232–246.Google Scholar

⁷ The basic framework is presented in Davis, Lance E. and North, Douglass C., Institutional Change and American Economic Growth (New York: Cambridge University Press, 1971).CrossRefGoogle Scholar

⁸ Richard A. Easterlin, “Interregional Differences in Per Capita Income, Population, and Total Income, 1840 to 1950,” Trends in the American Economy in the Nineteenth Century (Studies in Income and Wealth, Volume Twenty-four) by the Conference on Research in Income and Wealth (Princeton, 1960) and Gallman, Robert E., “Gross National Product in the United States, 1834–1909,” Conference on Research in Income and Wealth, Output Employment and Productivity in the United States after 1800, (Studies in Income and Wealth, Vol. 30, New York: Columbia University Press, 1966).Google Scholar

⁹ Evans, Robert Jr. “The Economics of American Negro Slavery, 1830–1860,” Aspects of Labor Economics (Princeton: Princeton Univ. Press, 1962), p. 216Google Scholar, also available in Robert W. Fogel and Stanley L. Engerman, “The Economics of Slavery,” p. 325, in Fogel, Robert W. and Engerman, Stanley L., eds., The Reinterpretation of American Economic History (New York: Harper and Row, 1971).Google Scholar Robert W. Fogel and Stanley L. Engerman, Time on the Cross, pp. 67–78. Because Fogel and Engerman's earning structure was calculated for the period approximating the 1850 level of slave income, all net positive incomes of slaves were increased by a factor of 1.18 (the ratio of Evan's hire rates for 1860 to the average of those for 1846–50 and 1851–55).

¹⁰ Gunderson, Gerald, “Southern Antebellum Income Reconsidered,” Explorations in Economic History, X (Winter 1973).Google Scholar

¹¹ These figures are strong evidence against the “slave-power conspiracy” hypothesis. They state that the aggregate vested interest in slavery for each of the states which initially seceded is sufficient to explain their actions; that is, it is not necessary to resort to an explanation of a contriving minority. This will be formalized in inequality (7) below.

¹² The obvious advantage of using numbers to express these interests is that it bypasses the morass of explaining whether some, a majority, most, or all individuals believed and acted in such a way. Hereafter when such groups as the North or the South are used they refer to the weighted averages of their preferences or costs as revealed in their actions. It should definitely not be taken as implying that all included parties agreed with a given decision or held preferences or equal intensity about it.

The very strong relation between the ownership of slaves and the vote for secession within states has been demonstrated by Wouster, Ralph A. in his The Secession Conventions of the South (Princeton: Princeton University Press, 1962).Google Scholar Another important observation relevant here was made by Olsen, Otto H. when he showed that by any reasonable relative standard, slave ownership was widespread in the South. See his “Historians and the Extent of Slave Ownership in the Southern United States,” Civil War History, XVIII (June 1972).Google Scholar

¹³ Railroad bonds were yielding 7½ percent in January of 1860. Call rates on the New York stock exchange for all of 1860 averaged 6 percent. See Conrad, Alfred H. and Meyer, John R., The Economics of Slavery and other Studies in Econometric History (Chicago: Aldine, 1964), p. 55.Google Scholar The sensitivity of the results of this study to this particular rate is reduced by using the same value for other capitalizations such as the respective costs of the war.

¹⁴ Using the income level of 1860 as the fixed base for calculating the various annuity costs raises the possibility that growth after 1860—or decreases, for that matter—might affect the results of this study. While a general evaluation is very complicated, an over-all appraisal suggests that this is not likely to be an important factor within the range of (known) post-war economic change. For this complication to exist, increases in output must be dependent on the existing level of income. If, for example, growth results from maximizing behavior in the face of foreseen opportunities for innovation independent of the level of income, post-war growth rates could be unaffected by wartime losses. Even when the growth and the incremental gains to output are positively correlated, however, the probable amount of the effect appears to be insufficient to affect the results of the study. (The following explanation will be more comprehensible after reading through the appropriate sections of the text below.) All the parameters of the study with the exception of the cost of the war are stated as fractions of the income level and therefore automatically adjust to any changes in it. It is only changes in income relative to the cost of the war that could endanger the results of the analysis. In the (most likely) case of increasing incomes, that effect would be most likely to appear in inequality (6). Generally, paired combinations of F and D in which F is relatively small and D relatively large are most vulnerable. For “middle values” of the pair, however, it would require relatively large growth rates after 1860 to violate this inequality. For example, for the mid-point of the locus of F and D combinations, D≤13.3%,F ‹11.4% (See Figure 1), it would require a growth of real per capita income in excess of two percent to reverse the behavior modeled in inequality (6). This is greater than any long-term growth in per capita income in the latter half of the nineteenth century which begins at 1860—the relevant initial year for this study.

¹⁵ Suppose compensated emancipation was financed by taxing only northern citizens. Now T as seen by the North would rise from .06 to .09 and as seen by the South would fall from .06 to zero. Clearly the North would find other alternatives, such as war, relatively more attractive how while the South would obviously favor compensated emancipation. This would now. establish conditions which would encourage the South to compensate or “pay off” the North to move it away from its favored action, such as war, toward the course favored by the South, compensated emancipation. In other words bargaining between the two regions could adjust the terms of such a program if there was any interest by either side in doing so.

Another emancipation program occasionally suggested would have freed the slaves and then taxed them by the amount necessary to pay off their purchase price. One problem with this procedure would be to induce the slaves to work sufficiently to pay such taxes. After the war the freed men showed a strong preference for the leisure which had been previously denied them and such a program would only reinforce their incentive not to work. Probably more important, however, this program would simply switch ownership of the slave from his former master to the federal government. If the price of the slave approximates the present value of his expected future earnings net of subsistence, the federal government would need to extract that identical stream of earnings to pay off his purchase price. Such a program would not alter the basic nature of the institution or the primary source of objections to its propriety.

¹⁶ The implications of a value of T greater than .06 for the South will become more obvious below as the paper unfolds. An increase in T first increases the upper bound values of the pair of D and F, the expectation of the South being subjugated and the size of the war respectively. This, in turn, (through inequality (3) implies an increase in S_L the reduction in the present value of slaves that continue to be Kept in the Union. The sum of these changes do not affect any of the conclusions of this study other than increasing the indicated parameter values by a scaler factor.

¹⁷ A test for extra-economic motivation by the South in maintaining slavery can be developed in the following way. Below in the study some possible independent measures of the value of S_L are suggested. These should suggest whether S_L is adequately explained by direct losses in slave values or is sufficiently large to indicate other factors as well.

¹⁸ Two good descriptions of this important period are Stampp, Kenneth M., And the War Came (Binghamton, New York: Louisiana State University Press, 1950),Google Scholar and Catton, Bruce, The Coming Fury (New York: Doubleday, 1961).Google Scholar

¹⁹ Inequality (9) below will show that the North's primary motivation in going to war was its antislavery sentiment. Several simulations on inequalities (9) and (10) and variants thereof yield similar implications; namely, that in the probable range of N (the North's antislave preference) of .015 to .04, .the feasible values of D are bounded by about 20 percent to 50 percent.

²⁰ These values are inferred at a time when the war has begun. Thus C, the foreseen probability of war resulting from Lincoln's election, is set at one. The value of C prior to the beginning of the conflict will be estimated later in the paper.

²¹ See, for example, William, and Catton, Bruce, Two Roads to Sumter (New York: McGraw Hill, 1963).Google Scholar

²² Current, Richard N., “God and the Strongest Battalions” in Donald, David, ed., Why the North Won the Civil War (Binghamton, New York: Louisiana State University Press, 1960).Google Scholar

²³ The rationale for that assumption is demonstrated in inequality (5) below. The assumption is not strictly necessary, however, in that there is some “surplus” in inequality (4) so that D can rise above any values given in Figure 1 for a given value of F. For example, in inequality (4) where F < 12%, D could rise as high as 16 percent whereas in Figure 1 at that F value it could be no higher than 12.5 percent. Thus, the original seven states of the Confederacy could “afford to go it alone” in that they could accept a somewhat higher probability of subjugation if the lower border states did not join them.

²⁴ This position can be viewed as a statement of fact and/or a bargaining stance on the part of these states. In the latter case, the threat is aimed at the North in that it clearly indicates a substantial increase in the South's military capabilities in the event of war. It also suggests (to the extent that the position is being used as a bargaining tool) that these states viewed it more likely that war would be initiated by the North rather than the South. The border states were the one section of the nation to whom war did not seem the best alternative and thus one of their major interests would be in preventing it, if possible.

²⁵ Wright, Quincy, A Study of War (Chicago: University of Chicago Press, 1965).Google Scholar

²⁶ North, Douglass C., Growth and Welfare in the American Past (Englewood Cliffs: Prentice Hall, 1974), 2nd. ed., chapter 9.Google Scholar

²⁷ The basic distinction between the North's possible motives of abolishing slavery or the preservation of the Union (in the broad context of preserving national institutions) is fairly widely recognized. Attributing the motivation to an aversion to slavery is generally acknowledged to have begun, with Rhodes, James Ford. See his Lectures on the American Civil War, (New York, Macmillan, 1913).Google Scholar The best known modern articulation of this hypothesis is Schlesinger, Arthur M. Jr's “The Causes of the American Civil War: A Note on Historical Sentimentalism,” Partisan Review, XVI (1949).Google Scholar

Modern historians have given far less emphasis to the importance of preserving the Union or national institutions than contemporaries at the beginning of the Civil War. This possible motivation continues to be mentioned in discussions of the question even though no specific well-known interpretation gives it substantial emphasis. A variant of this general explanation has recently emerged, however, underlining a frequent observation about the question of the Civil War's causation; namely, that because no definitive answers have been available, old explanations continue to revive and adapt to new concerns. See Paludon, Phillip S., “The American Civil War Considered as Crisis in Law and Order,” The American Historical Review, LXXVII (October 1972).Google Scholar

²⁸ Stampp, Kenneth, The Peculiar Institution: Slavery in the Ante-Bellum South (New York: Knopf, 1956).Google Scholar

²⁹ These estimates of emancipation rates are constructed by assuming that the slave population would have grown at an average rate of twenty percent per decade approximating that for all slaves in the nation during that period. Any shortfall from this forecast would be explained by exports of slaves or emancipation. The division between these two alternatives can be made from the census figures for free blacks reported in Negro Population, 1790–1915, Department of Commerce, Bureau of the Census (Washington: G. P. O., 1918) p. 57. This pattern of relative emancipation among states appears to continue in the 1830 to 1860 period, but because free blacks then constituted a large portion of the Negro population, possible variations in their population growth on migration make such an estimation procedure less reliable. This finding is verified by Calderhead, William, “How Extensive was the Border State Slave Trade?” Civil War History, XVIII (March, 1972).Google Scholar

³⁰ See, for example, Table 5 by Robert Evans Jr. as quoted in Fogel and Engerman, “The Economics of Slavery,” p. 325.

³¹ The same phenomenon was also evident during the war. When Union armies occupied Confederate territory, slavery in that locality simply began to dissolve. Slaves deserted their masters and clung to the military as a refuge. This occurred from the very first hostilities, well before the Emancipation Proclamation and it occurred even in the slave states such as Maryland and Missouri which officially remained in the Union. Whatever motivated the North to subjugate the South, either the desire to end slavery or the preservation of the Union or some combination thereof—an issue we shall examine below—northern armies marching across the South would inevitably destroy slavery in the process.

³² Todd, Richard Cecil, Confederate Finance (Athens: University of Georgia Press, 1954): Loans, pp. 82–84; notes, pp. 119–120; tariffs and taxes, p. 156.Google Scholar

³³ Todd, Confederate Finance, pp. 72, 114.

³⁴ Coulter, E. Merton, The Confederate States of America 1861–1865(Binghamton, N.Y.: Louisiana State University Press, 1950), p. 172.Google Scholar

³⁵ Coulter, Conferedate States, pp. 251–254; Todd, Confederate Finance, pp. 165–171.

³⁶ Coulter, Confederate States, p. 199.

³⁷ Moore, Albert Burton, Conscription and Conflict in the Confederacy (New York: Macmillan, 1924), p. 30.Google Scholar

³⁸ Moore, Conscription, p. 358. The true economic cost of military conscription is the difference between the amount the military pays the draftee and the amount it would be required to pay him to volunteer for the military in the absense of the compulsion of conscription. The price of a substitute approximates this difference because it expresses the amount an individual is willing to pay to retain his civilian status.

³⁹ Coulter, Confederate States, chps. X, XII and IX.

⁴⁰ Historical Statistics of the U.S. (Census Bureau, 1960), Series Y 350–356, p. 719.

⁴¹ Historical Statistics, p. 115.

⁴² Ibid., p. 735.

⁴³ Goldin, “The Economics of Emancipation.”