This research was supported by National Science Foundation Grant GS-2457, National Endowment for the Humanities Grant RO 258–70–4429 and, at an early stage, a grant-in-aid from the Council on Research in Economic History. Robert W. Fogel bears the onus of having encouraged the entire undertaking. Stephen E. Fienberg and F. Kinley Larntz, Jr. advised on sampling and other statistical procedures.

The following registers of Probate furnished copies of documents for the indicated county: John J. Costello, Essex; Walter Gilday, Plymouth; F. Joseph Donohue, Worcester. For Hampshire a microfilm was obtained through Juliette Tomlinson of the Connecticut Valley Historical Museum, with whom I was put in touch by Register of Probate Frank E. Tuitt, II. Connecticut document copies were obtained from the Connecticut State Library, where Frances Davenport was helpful.

Searches for age, occupation and other supplementary data were made by Harold S. Burt, John E. Miller, Mrs. Clinton Ober, Jimmie B. Parker, Mrs. Ralph L. Thresher.

I am indebted to colleagues, especially Theodore C. Bergstrom, John M. Murrin and Frederic Q. Raines, and students, the Computing Facilities and the Social Science Institute of Washington University; also to Chester McArthur Destler, Linda Hoffmaster, Donald L. Kemmerer, Stanley Lebergott, Glenn Weaver and many others who have answered queries or given criticism. The errors remain mine.

¹ Alice Hanson Jones, “Wealth Estimates for the American Middle Colonies, 1774,” Economic Development and Cultural Change, XVIII, 4, Pt. 2 (July, 1970). This supplement may be ordered separately from the University of Chicago Press. It is referred to hereafter as EDCC. See also, Jones, Alice Hanson, “La fortune privée en Pennsylvanie, New Jersey, Delaware, 1774,” Annales: Economies, Sociétés, Civilisations, XXIV, 2 (Paris, France: Armand Colin, March-April, 1969), 235–49.Google Scholar

² Alice Hanson Jones, “Wealth Distribution in the American Middle Colonies in the Third Quarter of the Eighteenth Century,” paper presented at annual meeting of the Organization of American Historians, New Orleans (April, 1971), mimeographed. Available at Economics Department, Washington University, St. Louis, Mo., 63130. Referred to hereafter as OAH Paper.

³ With the exception of Essex, where I stopped with 102 cases at the letter P in an alphabetical listing of surnames, and in Hampshire where a few 1773 cases were included to bring the number over the minimum target of 25 cases.

⁴ Appendix available at Economics Department, Washington University, St. Louis, Mo. 63130.

⁵ Inclusion of servants and slaves as wealth would increase the figure for New England by one-half of one percent to £36.8 sterling or $948 in 1969 dollars.

⁶ £,1 sterling = 1.33 local pound of Massachusetts or Connecticut and was the equivalent in purchasing power of $25.77 in 1969. £ 1 of Massachusetts or Connecticut = £0.75 sterling and was the equivalent in purchasing power of $19.33 in 1969. See EDCC, Table 48, Price Trends 1774–1967, pp. 124–26; OAH Paper, Table 3, Price Trends 1774–1969, p. 2.

⁷ In the Middle Colonies such entries appeared only in one-tenth of 1 percent of the inventories; the average for the entire sample was less than one-half local pound and less than 0.05 percent of gross portable wealth.

⁸ See citations and discussion in EDCC, pp. 129–32. Raymond Goldsmith calls to my attention that it might be argued from the well known income estimates of Gregory King in 1688 and Arthur Young in 1770 for England that average wealth in England was higher than that indicated by my colonial figures. Yet, the problem remains that the variance was probably much greater in England (King shows over half the population running an annual deficit), so that higher average wealth in England could still be compatible with my estimates of a relatively higher level of living for a large bulk of the colonial population.

⁹ EDCC, pp. 128, 135–40. The 1966 figure was expressed there as $11,032 and the 1805 one as $926 in terms of preliminary 1967 dollars.

¹⁰ Blodget, Samuel, Economica: A Statistical Manual for the U.S.A. (Washington, D.C, 1806).Google Scholar

¹¹ Goldsmith, Raymond W., Historical and Comparative Rates of Production, Productivity and Prices. Hearings before the Joint Economic Committee, 86th Congress, 1st session, 2, April 7, 1959 (Washington, D.C: Government Printing Office, 1959).Google Scholar See also Taylor, George R., “American Economic Growth before 1840: An Exploratory Essay,” The Journal of Economic History, XXIV, 4 (Dec, 1964), 427–44.CrossRefGoogle Scholar

¹² David, Paul A., “The Growth of Real Product in the United States before 1840: New Evidence and Controlled Conjectures,” The Journal of Economic History, XXVII, 2 (June, 1967), 151–97.CrossRefGoogle Scholar

¹³ An item classification code, which runs to over 50 mimeographed pages, is available on request to the author.

¹⁴ In New England we assigned the total value for a lumped group of items, to the category containing the item of presumably greatest value in the mixture.

¹⁵ Correction would be a time-consuming matter of restudying and rejudging the proper classification of each entry, and perhaps making more allocations of lumpedvalues.

¹⁶ The size of the decedent samples in New England and in the Middle Colonies justifies the use of the central limit theorem and the expectation that successive sample means would be approximately normally distributed. The difference between average physical wealth for the age-adjusted probated decedents in the two regional samples cannot be said to be statistically significant at the .95 level of confidence, but it is statistically significant at the .64 level of confidence; that is, the null hypothesis can be rejected at the latter level. A summary of the relevant numbers for the mean of the age-adjusted decedent sample in each of the two regions is given below. S. D. = Standard Deviation.

The 95 percent confidence interval for the difference between the means for New England and the Middle Colonies (Sterling) is ( —22, 66).

A larger sample size would, of course, decrease the width of the confidence interval and provide more accurate estimates for the age-adjusted probate-type living samples. In general, to double the precision, that is, to halve the width of the confidence interval, a sample approximately four times as large as the present one would be required. Since I have used all the 1774 cases in all the sample counties except Essex, more cases could have been obtained only by adding more years or going to additional counties.

Another method of cutting variability would involve sampling every county, but costs precluded this alternative.

¹⁷ The number of cases was smallest and hence the averages for the youngest group are subject to a wider range of sampling error; yet even within the small sample, the consistency of the pattern of rising wealth with age seems undeniable. Similar results were found for the Middle Colonies.

¹⁸ See citations in OAH Paper.

¹⁹ Mechanically, to minimize operations, these weights and the weights required to keep the county contributions equal are combined into a single factor for each decedent.

²⁰ Since the preferred figure of number of adult deaths which actually occurred is unknown, numbers of probable deaths were approximated, using the same death rates for adult age classes as I used for the Middle Colonies for this purpose. As additional data become available, it is possible that these death rates will be revised downward, hence reducing somewhat my estimated proportion of nonprobate types in the living population. This is one reason why I term my present estimates provisional.

²¹ The 67.3 figure could also be revised downward if a re-check of some counties would show that we overlooked counting some of the 1774 probate cases. As the records were kept in differing ways in various counties, some with good and some with poor or no indexes, it is difficult to be sure exactly how many probate cases were filed in a given year and hence to make consistent counts from county to county. Sometimes a case of administration granted, or probate inventory filed turns out, after genealogical search, to pertain to a person dead several or even many years earlier, and hence such case should not be counted. Such a re-check is in process in Essex County, and if justified there, may be made for other counties. This possibility is a second reason for terming my present estimates provisional.

²² To pass on legal title to land in New England, probate was required, and this is the reason that sometimes, many years after death, a probate inventory was filed, at a time when it was desired to convey the land to a new owner who required a legal title. Land, apparently at least, sometimes descended to heirs informally and was sometimes purchased informally without legal title.

²³ The greater the Gini coefficient, the greater the wealth inequality. Its range is from 0, if everyone had equal wealth to 1 for maximum inequality if one person had all the wealth.

²⁴ For the Middle Colonies, the distribution of net worth was more skewed than that for physical wealth, with a Gini coefficient of .50 for the probate types and .54 or .57 for all living free adult potential wealth holders. For Philadelphia County alone, using assumption B (one-fourth) for nonprobates’ wealth, it was the most skewed of all, with a Gini coefficient of .68. Taxed wealth in Philadelphia County was even more highly skewed with a Gini coefficient of .83. Supporting tables and graphs appear in the OAH Paper. The net worth distribution is not yet available for New England, but I anticipate that it will be more unequal than that for physical wealth.

²⁵ Gallman, R. E., “Trends in the. Size Distribution of Wealth in the Nineteenth Century: Some Speculations,” in Soltow, Lee, ed., Six Papers on the Size Distribution of Income and Wealth, NBER Conference on Research in Income & Wealth, Studies in Income and Wealth, 33 (New York and London: Columbia University Press, 1969), pp. xiii, 6.Google Scholar

²⁶ Projector, D. S. and Weiss, G. S., Survey of Financial Characteristics of Consumers (Washington: Board of Governors of Federal Reserve System, 1966), p. 30.Google Scholar

²⁷ The value of servants and slaves is included in physical wealth for this comparison; cash and all other financial assets are excluded.

²⁸ I estimate at over 4 percent for New England and over 7 percent for the Middle Colonies the proportion which servants and slaves formed of all adults, in 1774.

²⁹ Values are stated in local money in the inventories. In Massachusetts pounds, the figure for the widow is £.4.8, for Gerry £5,584; in Pennsylvania pounds, the shoemaker had £6.5, Neave £14,176. Reduced to equivalent pounds sterling the figures are: widow £ 3.5, Gerry £4,188, shoemaker £3.9, Neave 8,335.

³⁰ Since “town” in New England comprised a wide area, similar to a township in some other regions, “town” population is not a clear guide to whether places within the “town” were urban or rural, and occupations were used for marginal decisions.