¹ Turner, Frederick Jackson, The Frontier in American History (New York, 1920)Google Scholar.

² The safety valve was to reduce cyclical movement in the U. S. economy and draw the poor out of the East. Evidence for the validity of the safety-valve thesis is in short supply. See Billington, Ray A., The American Frontier Thesis: Attack and Defense (Washington, D. C., 1971), pp. 20–25Google Scholar.

³ Curti, Merle, The Making of an American Community: A Case Study of Democracy in a Frontier Community (Palo Alto, 1959), p. 75ffGoogle Scholar.

⁴ George, Henry, Progress and Poverty, 4th ed. (New York, 1883), pp. 207–219Google Scholar.

⁵ Ibid., p. 217.

⁶ “The association of poverty with progress is the great enigma of our times. It is the central fact from which spring industrial, social and political difficulties that perplex the world, and with which statesmanship and philanthropy and education grapple in vain.” Ibid., p. 9.

⁷ Soltow, Lee, Men and Wealth in the U. S., 1850–1870 (New Haven, 1975), p. 80Google Scholar.

⁸ For example, see Campbell, Randolph B. and Lowe, Richard G., Wealth and Power in Antebellum Texas (College Station, Texas, 1977), pp. 59–61Google Scholar; Jeremy Atack and Fred Bateman, “The Egalitarian Ideal and the Distribution of Wealth in the Northern Agricultural Community,” University of Illinois Faculty Working Paper, #522; Lee , Soltow, Patterns of Wealth Holding in Wisconsin Since 1850 (Madison, 1971), pp. 34–40Google Scholar.

⁹ See Sahota, Gian S., “Theories of Personal Income Distribution: A Survey,” Journal of Economic Literature, 16 (03 1978), 1–55Google Scholar.

¹⁰ Ricardo, David, On the Principles of Political Economy and Taxation (London, 1817)Google Scholar.

¹¹ For a review of this literature, see Stigler, George, Essays in the History of Economics (Chicago, 1965), pp. 173–85Google Scholar.

¹² This change in distribution with less equality as the settlement economy matures is not a logical outcome of the process described above, but rather a plausible outcome. It is logically possible for the effect of settlement to be egalitarian.

¹³ Stigler, George, “The Economics of Information,” Journal of Political Economy, 69 (06 1961), 213–25CrossRefGoogle Scholar.

¹⁴ Note that if one argues that entry into an economy is a determinant of household income and wealth, then the boundary of the economy in question must be denned so that entry may be dated. Moreover, the spatial boundary of an economy will change with the economic activity in question. For example, the economy of academics in a particular field may be geographically dispersed so that moving from London to Princeton, New Jersey does not imply entry into a new economy occupationally, but it probably implies entrance into a new economy in terms of consumption. Information may also transfer between economies. A migrant from New York to Ohio in the nineteenth century may have been an entrant into a new local economy. However, the stock of information accumulated in New York should have transferred to Ohio better than information accumulated in Sweden. One would also expect migration patterns to be influenced by the expected loss of information. For example, European farmers are more likely to settle in regions similar to their own; the existence of technologically homogeneous industries is more conducive to migration. Previous migrants might help reduce the cost of acquiring information.

¹⁵ Curti, The Making of an American Community; Greven, Phillip, Four Generations: Population, Land and Family in Colonial Andover, Massachusetts (Ithaca, 1970)Google Scholar; Lockridge, Kenneth A., A New England Town: The First Hundred Years, Dedham, Massachusetts, 1636–1736 (New York, 1970)Google Scholar.

¹⁶ , Greven, Four Generations, pp. 46, 58Google Scholar. The Gini coefficient for the first-generation land distribution is.25, indicating relative equality. It seems likely that total wealth was less equally distributed.

¹⁷ Ibid., p. 129. Price increases are not in real terms and therefore are difficult to interpret, but it appears that the inventoried land (likely to be the land of an earlier settler) increased most in value.

¹⁸ Ibid., p. 225. The precise changes in the distribution are not apparent from Greven's table since the price level is changing and the upper tail of the distribution is not given. The fourth-generation distribution is less equal than the third-generation one because the proportion of estates above £2,000 is constant and the proportion below £400 has increased by 70 percent. The change in inequality between the second and third generations is unclear. The proportion of estates below £400 declines dramatically, but the proportion above £1,000 increases dramatically. The distribution in the first generation clearly is more equal.

¹⁹ Ibid., p. 63n.

²⁰ Lockridge, A New England Town.

²¹ Ibid., p. 73.

²² Ibid., pp. 147–59.

²³ , Curti, The Making of an American Community, Table 12, p. 142Google Scholar.

²⁴ Ibid., p. 179ff.

²⁵ Ibid., in Table 12, p. 142, the mean age in 1880 of those who appear in the 1880 Census only is 41 years, whereas the mean age of farmers who appear in all three censuses is S3 years.

²⁶ A multiple regression gives an estimate of marginal impacts of variables on the level of wealth only if there are no interactions between the independent variables. Unfortunately this is unlikely to be true. We try to examine some of the interactions in the Utah data later in the paper.

²⁷ Soltow, Men and Wealth.

²⁸ For the United States and Utah, see Table 1 of this article. For Wisconsin, see Soltow, Lee, Wealthholding in Wisconsin, p. 10Google Scholar.

²⁹ In the linkage process, tolerances must be assumed for age, spelling of name, and so forth. We found birthplace an important aid in linkage. We were conservative in linking, and created categories for doubtful links in order to reduce the bias in the single census category.

³⁰ In subsequent work, the polygamy problem will be resolved by using family group sheets contained in the Genealogical Society Library in Salt Lake City. The direction of bias that the polygamy problem imparts to our results is unclear. It is likely that polygamists were drawn from the wealthier segments of the population as well as from those who had been in Utah longer. For those polygamists who had multiple households in the census, we have an underestimate of their wealth, which probably biases downward the impact of time in Utah. A preliminary examination indicates that most polygamists listed all of their wealth in one household.

³¹ The life cycle in Table 5 is cross-sectional. It does not mean that an individual would experience this pattern of wealth holding over time unless the assumption is made that the cross-sectional life cycle is constant and there is no growth in the economy. The life cycle varies according to characteristics of the population such as occupation.

³² It is interesting to note that Campbell and Lowe found southern-bom individuals to have the highest wealth holdings, with those born in the lower South being more wealthy than those born in the upper South; see , Campbell and , Lowe, Wealth and Power, p. 59Google Scholar. It seems clear that birthplace is a proxy for duration in the economy.

³³ It is obviously difficult to make inferences in this case about causal relationships: low wealth causing moves to better one's position versus moves for other reasons resulting in low wealth. It is clear, however, that those who moved had a set of characteristics that led to substantial differences in observed wealth holdings. The less important impact early and the more important impact later are suggestive of the possibilities of movement to better one's position. At the very least, we can say that it is those who are relatively worse off who engage in internal migration.

³⁴ Merchants and professionals make up about 1.6 and 2 percent of the population, respectively. Thus, we find that for almost all of the population, occupational choice does not matter as much as does a characteristic related to wealth holdings. This cannot be interpreted that occupational choice per se does not matter, since it may be precisely that occupational mobility in response to differential quasi-rents or incomes leads to occupation not being related in a statistical sense to income. The interesting thing about these data is that, on this interpretation, such movements appear not to be too costly in terms of wealth holdings. These results are merely speculative, however, since we have not yet examined occupational mobility in our data, particularly mobility in response to economic incentives. Professionals are generally those with educational attainments (lawyers, doctors, and so forth) who differ from the rest of the population or households in positions to which there cannot be voluntary movement (ecclesiastical and so forth).

³⁵ See footnotes 7 and 8.

³⁶ A regression run with a full set of interactions did not significantly change the coefficient of the variables. This is probably due to the size of our sample (more than 19,000 observations).

³⁷ Time of entry has a large impact in all the regressions that have been run on various subgroups. For example, in the regression for foreign born only, entry before 1850 had a marginal impact of $1,893, and entry before 1860 increased wealth by $1,115 marginally.