¹ de Tocqueville, Alexis, L'Ancien Régime, trans. Patterson, M. W. (Oxford 1947), 185–86Google Scholar.

² The pervasiveness of medieval rebellions conducted by better-off peasants is alone sufficient to destroy the novelty of Tocqueville's finding. See Coulton, G. C., Medieval Village, Manor and Monastery (New York 1960)Google Scholar, esp. chaps, II, 24, 25.

³ Brinton, Crane, The Anatomy of Revolution (New York 1965), 60Google Scholar.

⁴ An explicit statement of this position is given in Secretary of Defense Robert Mc-Namara's address before the American Society of Newspaper Editors in Montreal, published in the New York Times, May 19, 1966.

⁵ Russett, Bruce M., “Inequality and Instability: The Relation of Land Tenure to Politics,” World Politics, xvi (April 1964), 442–54CrossRefGoogle Scholar; , Ivo. K. and Feierabend, Rosalind L., “Aggressive Behaviors Within Polities, 1948–1962: A Cross-national Study,” Journal of Conflict Resolution, x (September 1966), 249–71Google Scholar. The latter study suggests (pp. 256–57), but does not find, that the most traditional and least developed nations may be significantly more stable than the transitional or developing nations.

⁶ Cited in fn. 4 above.

⁷ This point of view is attributed to Ambassador Lodge by Richard Critchfield in a Washington Star article entered approvingly into the Congressional Record, February 17, 1966, by Senator Jacob Javits. It is also affirmed by George McT. Kahin in a Memorandum published in the Congressional Record, April 13, 1967.

⁸ This study is based upon Mitchell, Edward J., Land Tenure and Rebellion: A Statistical Analysis of Factors Affecting Government Control in South Vietnam, The RAND Corporation, RM-5181 (abridged) (Santa Monica, June 1967)Google Scholar.

⁸ Some recent observations on peasants in India may be found in Bailey, F. G., Politics and Social Change: Orissa in ig$g (Berkeley 1963), 88Google Scholar; and Nair, Kusum, Blossoms in the Dust (London 1961), 192–93Google Scholar. A more extensive discussion of this matter and others relating to the interpretation of the statistical results can be found in Mitchell, 22–30.

¹⁰ Richard Baxter, Autobiography, cited in Hill, Christopher, The English Revolution, 1640 (London 1959), 13Google Scholar.

¹¹ Reliquiae Baxterianae, cited Ibid., 13.

¹² This is standard practice in econometric work. See, for example, Malinvaud, E., Statistical Methods of Econometrics (Chicago 1966), 55–56Google Scholar. Numerous works describe and analyze the model to be presented here, but Malinvaud's treatment is among the most complete.

¹³ Ibid., 172–95, 250–73.

¹⁴ Los Angeles Times, December 26, 1965, Opinion Section, 1. The Times was kind enough to provide the author with copies of the published maps.

¹⁵ A landholding is merely a farm—ownership is not necessarily implied. Thus, one variable measures only inequality in ownership, whereas the other also reflects inequality in the sizes of tenant holdings. Alternative measures of inequality were considered, but were found to be less significant statistically and to provide no additional explanation of control.

¹⁶ Agricultural Economics and Statistics, Department of Rural Affairs, Republic of Vietnam (Saigon 1961). Strictly speaking, this was not a census but rather a sample of about ten percent of the hamlets.

¹⁷ Ibid., 17–18.

¹⁸ Gittinger, J. Price, Studies in Land Tenure in Vietnam, Division of Agriculture and Natural Resources, United States Operations Mission to Vietnam (Saigon, December 1959), 59–60Google Scholar; G. Peautonnier, “Contribution à l'étude des conditions de l'exploitation de la rizière en Cochinchine,” L'Information d'Indochine économique et financièrs (December 1946).

¹⁹ (Hanoi 1932).

²⁰ The areas of Vietminh control were determined from a map facing p. 200 of Giap's, Vo NguyenPeople's War People's Army (Washington 1962)CrossRefGoogle Scholar.

²¹ The details of this program may be found in many places, among which are Gittinger; Ladejinsky, Wolf, “Agrarian Reform in the Republic of Vietnam,” in Fishel, Wesley R., ed., Problems of Freedom: Vietnam Since Independence (New York 1961)Google Scholar; and Scigliano, Robert, South Vietnam: Nation Under Stress (Boston 1963)Google Scholar.

²² Division of Agriculture and Natural Resources, United States Operations Mission to Vietnam, Vietnamese Agricultural Statistics (Saigon, March 1959), 47Google Scholar.

²³ See Ladejinsky. A recent USIS publication, Land Tenure in South Vietnam, indicates that in Ba Xuyen, which contained almost thirty percent of all land subject to transfer, about eighty-four percent had been expropriated by 1963.

²⁴ The two-thirds figure is attributed to Diem by Scigliano (p. 123). The one-third figure appears in Shaplen's, RobertThe Lost Revolution (New York 1963)Google Scholar. In personal communications Wolf Ladejinsky has given a one-third, and John Donnell a one-half, estimate.

²⁵ Ladejinsky, in a personal communication.

²⁶ USIS, Land Tenure.

²⁷ Cooper, John L., Land Reform in the Republic of Vietnam, Agency for International Development (Washington, March 1966), 3Google Scholar, mimeographed.

²⁸ There are presumably ways in which these large Vietnamese and French estates differ, other than the degree of redistribution carried out. However, a study of the limited historical literature on their development reveals no other important differences that might have some subsequent effect on control. But this could be due to the paucity of information rather than the absence of important differences. For the earlier history, see Hall, D. G. E., A History of South-east Asia (New York 1955), 654–56Google Scholar; Buttinger, Joseph, The Smaller Dragon: A Political History of Vietnam (Ne w York 1958), 431–32Google Scholar; Lindholm, R. W., “An Economic Development Oriented Land Reform Program for Vietnam,” in Froelich, Walter, ed., Land Tenure, Industrialization and Social Stability: Experience and Prospects in Asia (Milwaukee 1961), 183–86Google Scholar; Robequain, C., The Economic Development of French Indo-China (London 1944), 83–85Google Scholar, 181–94.

²⁹ One could not have, for example, a province with seventy percent government-owned land and fifty percent owner-operated land.

³⁰ P. 4.

³¹ PP. 81, 3.

³² A variable is called significant here if its coefficient has a t—statistic of 2.1 or larger; otherwise, it is insignificant. The level of significance of a t-statistic is the probability that a value of t so large (in absolute size) could have been obtained by chance when in fact the coefficient of the particular variable is zero. A t-value of 2.1 indicates significance at the 5 percent level; that is, in only one case in twenty would one find a coefficient so large when the true coefficient is zero. Higher t-values imply greater significance. For example, if t is 2.8, die level of significance is 1 percent and when t is.3.9 the level of significance is one-tenth of 1 percent. These statements are of course dependent upon the approximate validity of certain assumptions about die random errors. See Malinvaud, 172–208, 250–63.

³³ The regression also has the property that it has the highest

of any regression examined. It can be shown that a correctly specified regression has a higher

on the average than any odier regression. Thus the regression equation (2) is more likely to be the “true” relationship than any of the alternatives. See Theil, H., Economic Forecasts and Policy (Amsterdam 1961), 211–14Google Scholar.

³⁴ This finding is supported by a statement in the Cooper report that “a larger proportion of the former French holdings are in the secure areas than Ordinance No. 57 lands [Vietnamese lands subject to transfer] or state-owned lands” (p. 6).

³⁵ It should be stressed again that the presence of other factors distinguishing French and Vietnamese land could imply a reinterpretation of this finding. At this time we are not aware of what these factors might be. Because of this uncertainty, however, our interpretation of the coefficients of the French and Vietnamese land variables should not be held with a high degree of confidence.

³⁶ That is, in Figure I the control variable, C′, is given by

where the bars indicate mean values and e is the estimated value of the residual. Figures 2, 3, and 4 were constructed in a similar manner. As pointed out above, there are limits to the extent to which one can choose arbitrary values for the independent variables. One consequence of this is the negative predicted values of control in Figure 3. The values of the independent variables that give rise to this phenomenon could almost never have been observed. They imply too high a value of owner-operated land, given the land occupied by large French and Vietnamese estates. There is also the possibility of large deviations from linearity of the control equation as one approaches extreme values.

³⁷ Tilly, Charles, The Vendée (Cambridge, Mass., 1964)Google Scholar, passim.

³⁸ Coulton, 130.