The authors are Professor of Economics and Associate Professor of Economics, respectively, at Ohio University. All calculations were made at the Ohio University Computer Center. We are indebted to John Kern, Michael Hamant, and Kathleen Gallaway for their assistance in carrying out the necessary data processing tasks.

¹ In The Immigrant in American History (Cambridge, Mass.: Harvard University Press, 1940)Google Scholar, Marcus Lee Hansen notes in the first sentence of chapter III that “The most persistent theme in American history is the westward march of the people.”

² Hansen, pp. 61–66, provides an interesting description of the nature of individuals who settled in the American West at various stages of its development, commencing with the frontiersman, and proceeding from there to the pioneer farmer, and, finally, to the permanent farmer and the commercial and non-agricultural types.

³ Possibly the most famous document in American historical writing is Turner's essay “The Significance of the Frontier in American History.” At the heart and core of his “frontier thesis” is the premise that the mobility of individuals to the frontier was a decisive influence in shaping the nature and course of American economic and social development.

⁴ For example, as of the 1880 census, between 20 and 25 percent of Americans born in the area comprising the present West North Central states (Minnesota, Iowa, Missouri, Kansas, Nebraska, and the Dakotas) and living outside their state-of-birth were residents of states lying east of the Mississippi River.

⁵ Hamsun, Knut, Fra det moderne Amerikan aandslive (Copenhagen, 1889), p. 2.Google Scholar

⁶ Wealth of Nations, Cannan Edition (Modern Library), p. 75.

⁷ There is a very substantial amount of recent literature dealing with the geographic mobility phenomenon. A complete listing in one footnote is impossible. However, a fairly complete bibliography of the more recent material is contained in Lowell E. Gallaway, “Age and Labor Mobility Patterns,” Southern Economic Journal, October 1969, pp. 171–180. In a more historical vein, we would simply call attention to the major study Population Redistribution and Economic Growth in the United States 1870–1950 (Philadelphia: American Philosophical Society, 1957–1964): Vol. IGoogle Scholar, Methodological Considerations and Reference Tables, by E. S. Lee, A. R. Miller, C. P. Brainerd, and R. A. Easterlin; Vol. II, Analyses of Economic Change, by Simon Kuznets, A. R. Miller, and R. A. Easterlin; and Vol. Ill, Demographic Analyses and Interpretations, by H. T. Eldridge and D. S. Thomas. It deals quite extensively with geographic movements of population. Hereafter, Vol. I of this study will be cited as Easterlin, et. al.

⁸ Hicks, J. R., The Theory of Wages (London: Macmillan Press, 1940).Google Scholar

⁹ See, e.g., Gallaway, “Age and Labor Mobility Patterns,” Greenwood, Michael J., “An Analysis of the Determinants of Geographic Labor Mobility in the United States,” Review of Economics and Statistics, LI (May 1969), pp. 189–194CrossRefGoogle Scholar, and Sjaastad, L. A., “The Costs and Returns of Human Migration,” Journal of Political Economy, LXX (Oct. 1962), Supplement, pp. 80–93.CrossRefGoogle Scholar

¹⁰ An example of a work emphasizing the job opportunity framework is Lowry, Ira S., Migration and Metropolitan Growth (San Francisco: Chandler Publishing Co., 1966).Google Scholar

¹¹ This is argued in Gallaway, Lowell E., “Industry Variations in Geographic Labor Mobility Patterns,” Journal of Human Resources, II (Fall 1967), pp. 461–474.Google Scholar

¹² Hansen, pp. 62–63, makes the following statement concerning the type of agriculture practiced by American farmers: “The pioneer farmers burned down the trees because it was the easiest way to make a clearing. They planted the same crop season after season, largely because they were mentally too sluggish to experiment with new products. As a result, the soil became impoverished and no attempt was made, either by scientific rotation or by fertilizing to restore or maintain its productivity. The land was mined, not farmed, and when the surface treasures had been skimmed off, the process was repeated in another place where Nature's bounty was as yet untouched.” Of course, it could be argued that this was simply the economical way to farm at this time.

¹³ Ibid., p. 61.

¹⁴ For a critique of the state-of-birth data see Easterlin, et. al., pp. 57–64.

¹⁵ Alternatively, the movement data could be expressed as the number of people born in state i living in state j at time t minus the number born in i living in j at time t-1. For example, the movement matrix for one census could be subtracted from the matrix for the succeeding census. This solves some of the time dimension problems but introduces other problems. In particular, it limits the temporal scope of the analysis somewhat since pre-1850 movement is automatically excluded.

¹⁶ David, Paul, “New Light on a Statistical Dark Age: U. S. Real Product Growth Before 1840,” American Economic Review, LVII (May 1967) pp. 294–306.Google Scholar

¹⁷ The availability of income data is the major constraint here.

¹⁸ Easterlin et. al., pp. 753–55.

¹⁹ These are reported in Easterlin, Richard A., “Interregional Differences in Per Capita Income, Population, and Total Income, 1840–1950,” Trends in the American Economy in the Nineteenth Century, Studies in Income and Wealth, XXIV (Princeton: Princeton University Press, National Bureau of Economic Research, 1960) 73–140.Google Scholar

²⁰ The problem with the 1850 data is that they relate only to real wealth.

²¹ As reported in the 1962 Statistical Abstract of the United States.

²² For the 1880, 1900, and 1920 censuses, these data are included in Easterlin, et. al., pp. 609–21.

²³ As reported in the 1850 and 1960 decennial censuses. The use of free male workers in 1850 may present some problems of interpretation since the immigration data are for all members of the population. However, we would expect the number of free male workers to be significantly correlated with the total number of free workers. Due to the type of statistical technique we will use to empirically evaluate the importance of the various factors influencing mobility (multiple regression equations) and the form of the expressions we will hypothesize (log-linear) the impact of this problem will be felt largely in the constant terms of the regressions and not in the estimates of the coefficient associated with the job opportunity measure.

²⁴ The regions are defined as follows: Northern—Maine, New Hampshire, Vermont, New York (because of the up-state part), Michigan, Wisconsin, Minnesota, North Dakota, South Dakota, Montana, Idaho, and Washington; Temperate—Massachusetts, Rhode Island, Connecticut, New York, Pennsylvania, New Jersey, Delaware, Maryland, West Virginia, Ohio, Indiana, Illinois, Michigan, Wisconsin, Iowa, Missouri, Nebraska, Kansas, Oklahoma, Wyoming, Colorado, Utah, Nevada, Oregon, and California; Southern Humid—Virginia, North Carolina, South Carolina, Georgia, Florida, Kentucky, Tennessee, Alabama, Mississippi, Arkansas, Louisiana, and Texas; and Southern Arid—Texas, New Mexico, Arizona, and California. Note that some states appear in more than one category. This is because it is felt that parts of them possess characteristics that would place them in more than one classification.

²⁵ As an alternative, in-migration to each state could be analyzed. We did not do this because of the problems implicit in comparing the behavior of individuals born in different states. By analyzing out-migration we deal with the behavior of only individuals born in the same state.

²⁶ In one instance, 1900, the zero order correlation coefficient between population size and the number of jobs is 0.995.

²⁷ These are the Northeast (consisting of the New England and the Middle Atlantic states), the Southeast (the South Atlantic states), the North Central (East North Central and West North Central states), the South Central (East South Central and West South Central states), and the West (Mountain and Pacific states).

²⁸ In all fairness it should be recognized that this pattern might also be the product of regional biases in the data. On a priori grounds we have no strong reason for suspecting this but it is a possibility.

²⁹ It might also be a product of the home environment of the trans-Appalachian, since his parents were also probably movers.

³⁰ This could reflect the growing importance of the rural-urban movement relative to East-West migration.

³¹ Interestingly, the 1960 elasticities of migration with respect to income and distance are quite similar to those reported in Gallaway, “Age and Labor Mobility Patterns,” which employs data from the Social Security Administration records for the period 1957–1960. The income elasticity in that study is 2.02 while the distance elasticity is - 0.98.

³² It could also reflect better measurement of per capita income levels through time.

³³ Louisiana treated slaves legally as real property. The actual treatment of them for census purposes in the individual parishes varied but it appears that in about one-half the parishes the legal definition was employed by censustakers in recording estimates of real and personal property. This becomes obvious from an examination of census manuscript records for this period. See, for example, Menn, Joseph K., The Large Slaveholders of Louisiana—1860 (New Orleans: Pelican Press, 1964).Google Scholar Therefore, we excluded Louisiana from the calculations used to develop a value for slaves. To obtain an estimate of per capita non-slave wealth for Louisiana we simply accepted the average value of a slave developed using the data excluding Louisiana. This should not produce any significant distortion.

³⁴ This estimate is somewhat lower than that frequently used in studies of the profitability of slavery in the ante bellum South. For example, in “The Profitability of Ante Bellum Slavery—Revisited,” Southern Economic Journal, XXXI (Apr. 1965), pp. 365–77Google Scholar, Richard Sutch suggests a value of $1,221 per slave. However, a more recent study of the 1860 value of slaves in Charleston, S.C., arrives at an average value of $799 which is quite consistent with our calculations. See Soltow, Lee, “The Value of Slaves in 1860 and the Value of Free Men in 1860 and 1968,” Research Paper No. 105, Department of Economics, Ohio University, 1971, p. 13.Google Scholar