Hostname: page-component-cd9895bd7-q99xh Total loading time: 0 Render date: 2024-12-24T17:56:49.507Z Has data issue: false hasContentIssue false
  • English
  • Français

Influences on the Distribution of Landholdings in Early Colonial North Carolina

Published online by Cambridge University Press:  03 March 2009

Robert E. Gallman
Robert E. Gallman
Affiliation:
Kenan Professor of Economics and History at the University of North Carolina, Chapel Hill.
Get access Rights & Permissions [Opens in a new window]

Abstract

The economy of North Carolina in the early colonial period was agrarian. Land was a central element in the wealth stock and it was distributed unevenly among households. This paper analyzes the distribution of land by means of multiple regression models employing measures of the principal life events of households. This paper analyzes the distribution of land by means of multiple regression models employing measures of the principal life events of households. The data are drawn from an eastern community, Perquimans County, and refer to the late seventeenth and early eighteenth century.

Type
Articles
Information
The Journal of Economic History , Volume 42 , Issue 3 , September 1982 , pp. 549 - 575
Copyright
Copyright © The Economic History Association 1982

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Lords Proprietors' Instructions to Governor Samuel Stephens, County of Albemarle, as printed in Powell, William S., ed., Ye Countie of Albemarle, A Collection of Documents (Raleigh, 1958), p. 22.Google Scholar

2 Peter Carteret's Account of the Years 1666–1673 in Albemarle, Ibid., pp. 62–63.

3 The following account is based chiefly on Lefler, Hugh T. and Powell, William S., Colonial North Carolina (New York, 1973);Google ScholarPowell, William S., The Proprietors of Carolina (Raleigh, 1963):Google ScholarMorgan, L. N., “Land Tenure in Proprietary North Carolina,” The James Sprunt Historical Publications, 12 (Chapel Hill, 1912);Google ScholarPaschal, Herbert Richard, “Proprietary North Carolina: A Study in Colonial Government,” (Ph.D. dissertation, University of North Carolina, Chapel Hill, 1961).Google Scholar

4 Paschal, “Proprietary North Carolina.”Google Scholar

5 Ibid., p. 420. That is, some of the land patented in 1693/94 had been occupied since the 1660s, but was not officially patented until 1693/94.

6 In 1713 Daniel Richardson, Receiver General of Carolina, came to the colony with instructions to accept only British sterling in payment of quitrents. He was convinced by the assembly to accept rated commodities as well. In 1716 the assembly voted o permit payment in provincial bills of credit. Ibid, pp. 427, 429.

7 Morgan, “Land Tenure,” p. 60. Sales were continued “‘until the Lords Proprietors' further pleasure be known…’” “No special effort was made to gain any statement of this ‘further pleasure’.”Google Scholar

8 Paschal, “Proprietary North Carolina,” Chap. 13.Google Scholar

9 Morgan, “Land Tenure,” pp. 56, 57, 60.Google Scholar

10 Computed from a data set called MALEVARS based on the tax lists of 1720 and 1721, as well as other sources. The data set is described further, below. Entries on the tax list are treated as evidence of ownership, although some may have referred instead to leaseholds. See below, footnote 17.Google Scholar

11 See Soltow, Lee, “Land Inequality on the Frontier,” Social Science History, 5 (Summer 1981).CrossRefGoogle Scholar

12 Ibid., pp. 285–86.

13 But see Jones, Alice Hanson, American Colonial Wealth (New York, 1977, 1978),Google ScholarWealth of a Nation to Be (New York, 1980)Google Scholar and Wolf, Jacqueline, “The Proud and the Poor: The Social Origins of Leadership in Proprietory North Carolina, 1663–1729” (Ph.D. dissertation, University of Pennsylvania, 1977),Google Scholar“Patents and Tithables in Proprietory North Carolina, 1663–1729,” The North Carolina Historical Review, 56 (Summer 1979).Google Scholar

14 See the work of Wolf, ibid.

15 See Kuznets, Simon, “Economic Growth and Income Inequality,” American Economic Review, 55 (03 1955).Google Scholar

16 “His wealth,” because in Perquimans the property a wife brought to marriage or acquired in it became the property of her husband.Google Scholar

17 Clark, Walter, ed., State Records of North Carolina, Vol. 23, Laws, 1715–1776 (Goldsboro, N.C., 1909), Laws of 1715, Chaps. 51 and 63, pp. 72–73, 90–92,Google Scholar and North Carolina State Archives, Colonial Court Records, Document Box 190. The tax lists give the location of land taxed, in most instances, from which it is clear that the tax base included land held outside Perquimans, but in the colony. Presumably land held in Perquimans by, e.g., taxpayers in neighboring Chowan County was recorded for tax on the Chowan tax lists. It is possible that the Perquimans list includes some absentee land owners who lived outside the colony, e.g., in Virginia, but their number is unlikely to be large. The means by which the land was acquired (deed, will, patent), is also sometimes given. But, unfortunately, land patented by e.g. a father and left by will to his son could legitimately be termed either “will” or “patent” land. No consistent naming scheme was followed, so that these data are not helpful. In a few instances the term “lease” appeared. Whether this meant that the person named as holding the land had leased it to or from another is not clear. I assumed the former, i.e., that the person named as holding the land for tax purposes actually owned it. In fact no one in this category formed part of the data base for the large models discussed in the last section of this paper.

18 There is at least one partial exception to this rule. A man named Oates appears twice on the list, once in the household of another—to whose account Oates's poll tax liability is set—and once as a landholder. I assumed that Oates was an indentured servant—since another person was charged with his poll tax—and that therefore he was part of that person's household.Google Scholar

19 They were also used for another purpose. Cross-section studies of wealth-holding suffer from the fact that wealth-holding in any given year is influenced by innumerable random shocks. This problem can be diminished if what amount to trend level estimates are substituted for wealthholding values for a given year. (See Kuznets, “Economic Growth and Income inequality,” p. 1Google Scholar). The earlier tax lists offered the opportunity to prepare trend level estimates. Extended experiments showed, however, that the substituted estimates performed no better than the data drawn from 1720/21 alone. Thus we abandoned this effort.

20 The law encouraged accurate returns by stipulating severe penalties on both taxpayers and tax assessors for errors. (Laws of North Carolina, 1715, Chap. 63, 6–9, in Clark, State Records, pp. 91–92.Google Scholar) Despite fines, more than one taxpayer refused to list in a given year, on the ground that the government owed him money and that he would not pay taxes until he had recouped. (We know this from notations on the tax lists, where the refusing taxpayers are named.) But no one refused for all the years during which land taxes were collected, so it proved easy enough to estimate missing values. The MALEVARS data set was built up from the data for 1720 and 1721. Where a different amount of land was recorded for a household in the two years, the larger amount was recorded in MALEVARS. If a household appeared in only one of the two tax lists, we inculded those data in MALEVARS. We dated MALEVARS to 1721.

21 The list is published, in slightly different forms, in two separate places: MrsWinslow, Watson, History of Perquimans Country (Raleigh, 1931), pp. 449, 450, andGoogle ScholarThe North Carolina Historical and Genealogical Register, Vol. 1, No. 2 (04 1900), pp. 301–04.Google Scholar Winslow's version was used, corrected in a few places where the Register seemed clearly to have the better account. (The original list is no longer among the records of the country.) Both Winslow and The Register date the list to “before 1700.” Comparisons with data from the demographic file (subsequently described) indicate that it could not refer to a date earlier than 1693 or later than 1695 and that in all likelihood it refers to the latter date.

22 See Gallman, James M., “Determinants of the Age of First Marriage in Colonial North Carolina,” William and Mary Quarterly, 39 (01 1982).CrossRefGoogle Scholar

23 For example, age was established by subtracting the birth date from the date of the tax list. Most of the birth dates, of course, refer to births in Perquimans. Thus we have no true age data for most immigrant household heads. We attempted to deal with this problem by estimating the birth dates of these individuals from their marriage dates or the dates at which their first children were born. (See Gallman, James M., “Mortality Among White Males: Colonial North Carolina,” Social Science History (Summer 1980).Google ScholarPubMed) Obviously the procedure has some weaknesses: (1) It assumes that the average age of marriage was the same for immigrants as for the native-born, which may not have been the case. (But since immigrants were chiefly from nearby colonies, not from overseas, this assumption may not be bad.) (2) It makes no allowance for variations in marriage ages among individuals. (3) So far as estimates from the date of birth of the first child are concerned, the shortcomings described in (1) and (2) hold. Furthermore, there is the added assumption that the interval between marriage and the birth of the first child was the same for everyone (an assumption that is not likely to be a source of serious error). Finally, some uncertainty surrounds the identification of the first child, in the absence of information on the marriage date of the individual.

The specific assumptions made were that the typical male married for the first time at age 23 and was 24 when his first child was born. These assumptions are based on the work of Gallman, James M., “Determinants.”Google Scholar

24 MacDonald, Ross, The Moving Target (New York, 1970), p. 82.Google Scholar

25 See Kearl, J. R., Pope, Clayne, and Wimmer, Larry, “Household Wealth in a Settlement Economy: Utah, 1850–1870,” this JOURNAL, 40 (09 1980).Google Scholar

26 There is another possibility: Perhaps immigrants before 1695 were abler and wealthier (when they migrated) than were those who came later. There is one piece of evidence that hints at such a history: The level of literacy actually fell in Perquimans after 1695 and there is reason to believe that this development was at least partly due to the fact that immigrants after that date were less literate than those who came before. (Gallman, Robert E., “Notes on Literacy in Colonial Perquimans,” manuscript.)Google Scholar There is also some indication, however, that the children of the pre1695 generations were also less literate than their parents. This suggests that the forces at work on literacy were general and that the lower levels of literacy among the later immigrants may not have indicated that they were poorer human material than their predecessors.

27 But see footnote 26. The generation of the household head must be impounded in ceteris paribus for this to hold.Google Scholar

28 Thucydides, , The Peloponnesian Wars, translated by Jowett, Benjamin, revised by Brunt, P. A. (New York, 1963), p. 49.Google Scholar

29 Gallman, Robert E., “Family Size, Inheritance Systems and the Distribution of Land in Early Colonial North Carolina,” paper delivered to the Organization of American Historians, 04 1979.Google Scholar

30 One would expect the coefficient to be positive and to reflect the value (in acres of land) of gifts at marriage. If only the well-to-do married, however, then the coefficient would also capture the extent to which these people were better off, before the fact, than those who failed to marry. This point seems not to be relevant to Perquimans, where marriage was well-nigh universal. If the rich married earlier than the poor, the marriage dummy variable might also pick this up. But, again, the point seems irrelevant to Perquimans. J. M. Gallman, “Determinants,” has shown that, in fact, the rich did not marry younger than the rest of the population. A regression we have run, which attempts to explain age of first marriage by landholding, supports his position. The coefficient on land is actually the wrong sign, is very small (a 1000-acre difference in landholding accounting for only about a two-year difference in age of marriage), and is not statistically significant at a 10 percent level of significance. The test is imperfect, since land ownership refers to 1720/21, not the date of marriage. We introduced age and age squared on the right-hand side of the equation, to deal in a rough way with this problem, but the statistics for the variable, land, changed little, while the results in other respects deteriorated. For recent evidence of an association between marriage and earnings, see Griliches, Zri and Mason, William, “Education, Income, and Ability,” Journal of Political Economy, 80 (05/06 1972). I owe this reference to David Galenson.CrossRefGoogle Scholar

31 See also below, where the economic consequences (to the household head) of launching a son into independent life are considered.Google Scholar

32 Gallman, Robert E., “Family Size”. Daughters usually got personal property—slaves, household goods, livestock—although on occasion they received land.Google Scholar

33 The variable selected was number of male children born to the taxpayer's father. Thus a household head with no brothers has a value of 1, a household head with one brother has a value of 2, and so on. This appeared to be the most appropriate way to describe the impact of family size on the distribution of land at the death of the landholder, on the assumption that land was typically divided among sons. The number of male children born was used rather than the number of male children surviving, because it is the more reliable measure.Google Scholar

34 I also experimented with two birth-order variables. The use of these variables sharply restricts the number of observations, so that they could not be used in the comprehensive models. But the results obtained in the small equations (of which the following is a good example) are of some interestGoogle Scholar:

(Landholders with holdings in excess of 1000 acres are omitted.) The equation asserts that the farther down the birth order a son was, the more land he held, after taking into account his age and the vital status of his father (a dummy variable taking a value of I if the father was dead in 1720/21, and 0 if he was alive). How one should interpret this is not entirely clear. Did younger sons get the home plantation at the death of the parents and was this plantation the largest of the family holdings? Were they given more when they were launched on the world because the parents had more to give at that time, being older? Or were they given more because they had stayed at home to care for the parents? Or were they given the poorer land or undeveloped land, as latecomers, but given more of it, to compensate for its relatively low value per acre? I do not know. The difference between one place and the next, in terms of land held, was net large, of course, but across several positions, it was by no means negligible.

Incidentally, notice that all the other coefficients in this equation have the right signs and plausible values and one has at value indicating a high level of significance. The adjusted R2 is also good, for a cross-section regression.

35 I owe this idea to David Galenson.Google Scholar

36 It also probably includes a few who had died and whose deaths went unrecorded, and possibly even some who were simply missed by the tax assessors, perhaps because they were temporarily out of the county.Google Scholar

37 Servants and slaves were valued by available price data, so that they could be combined.Google Scholar

38 See Easterlin, Richard A., “Population Change and Farm Settlement in the Northern States,” this JOURNAL, 36 (03 1976).Google Scholar

39 Hemingway, Ernest, “The Snows of Kilimanjaro,” in The Fifth-Column and the First Forty-Nine Stories (New York, 1939), p. 170.Google Scholar

40 Since 7 of the 54 households held no land, we reran the calculations using a variant of TOBIT, called LIMDEP. On balance the results improved somewhat, although they changed littleGoogle Scholar:

(t values in parentheses.)

41 One-tail test, in the cases of the age variables and tithable sons living at home, since in each case 1 had strong prior expectations as to sign, which were realized in the event.Google Scholar

42 For a discussion of the significance of this point in the context of colonial social and demographic history, see Gallman, J. M., Determinants.”Google Scholar

43 One possibility is that the two sets of sons were treated equally (ceteris paribus) with respect to gifts of wealth, but sons staying at home received their gifts chiefly in the form of land, and those leaving, chiefly in the form of portable wealth. But such an explanation suggests a very passive attitude on the part of the family with respect to the structure of its asset portfolio. (After all, regardless of the form in which gifts were given, the portfolio could be restored to its original—or any other desired—structure by appropriate sales and purchases.) Another possibility is that sons who left were less in the good graces of the family than were sons who stayed, and therefore were given less to start them in life. (I thank David Galenson and Jane Gallman for these suggestions.)Google Scholar

A final possibility that cannot be excluded is that the measurement prablems attending the construction of this variable have not been successfully overcome. As indicated previously (footnote 36), some of the sons treated as emigrated out of the country may, in fact, have died prior to 1721, perhaps in childhood. Measurement errors of this type would bias the coefficient in the direction of a small value, positive or negative. In any case, the number of such sons involved in the model calculations is very small. See Appendix.

44 Three of 34 households owned no land. Once again we ran LIMDEP (see footnote 40) and obtained slightly improved resultsGoogle Scholar: