¹ Although Kenya requires all such organizations to register with the government before engaging in self-help activities, no one knows precisely how many self-help organizations exist in Kenya. The figure of 15,000 to 20,000 organizations is the authors' estimate based on their review of registration rolls at the district level. Kenya is divided into 40 administrative districts with an estimated average of 375–500 self-help organizations in each.

² We define “local community” in spatial terms to refer to all people residing in an area of two to five square kilometers, which is roughly the area of a “sublocation,” the smallest territorial unit in the Kenyan administrative system. There are over 4,000 sublocations in the country.

³ Barkan, Joel D., “Bringing Home the Pork: Legislative Behavior, Rural Development and Political Change in East Africa,” in Smith, Joel and Musolf, Lloyd, eds., Legislatures in Development Durham, NC: Duke University Press, 1979), 265–88Google Scholar; Barkan, , “Legislators, Elections, and Political Linkage,” in Barkan, Joel D., ed., Politics and Public Policy in Kenya and Tanzania, rev. ed. New York: Praeger, 1984), 71–101Google Scholar; Ngethe, Njuguna, “Harambee and Development Participation in Kenya,” Ph.D. diss. Ottawa: Carleton University, 1979)Google Scholar.

⁴ Barkan, Joel D., Holmquist, Frank, Gachuki, David, and Migot-Adholla, S. E., “Is Small Beautiful?” Occasional Paper No. 15, Comparative Legislative Research Center, University of Iowa, 1979Google Scholar; Barkan, , “Development Through Self-Help: The Forgotten Alternative,” Rural Africana 19–20 (Spring-Fall 1984), 115–29Google Scholar; Holmquist, Frank, “Implementing Rural Development Projects,” in Hyden, Goran, Jackson, Robert, and Okumu, John G., eds., Development Administration: The Kenyan Experience Nairobi: Oxford University Press, 1970), 201–22Google Scholar; Frank Holmquist, “Class Structure, Peasant Participation, and Rural Self-Help,” in Barkan (fn. 3, 1984), 171–97; Holmquist, , “Self-Help: The State and Peasant Leverage in Kenya,” Africa 54 (No. 3, 1984), 72–91CrossRefGoogle Scholar; Mbithi, Philip and Rasmuson, Ramus, Self-Reliance in Kenya: The Case of Harambee Uppsala: Scandinavian Institute of African Studies, 1977)Google Scholar; Njau, Peter M., “Tensions in Empowerment: The Experience of the Harambee (Self-Help) Movement in Kenya,” Economic Development and Cultural Change 35 (April 1987), 523–38Google Scholar; Thomas, Barbara P., Politics, Participation, and Poverty: Development Through Self-Help in Kenya Boulder, CO: Westview Press, 1985)Google Scholar; Winans, Edgar V. and Haugerud, Angelique, “Rural Self- Help in Kenya: The Harambee Movement,” Human Organization 36 (No. 4, 1977), 334–51CrossRefGoogle Scholar.

⁵ Amsden, Alice H., “A Review of Kenya's Political Economy since Independence,” Journal of African Studies 1 (Winter 1974), 417–40Google Scholar, at 423. See also Hunt, Diana, The Impending Crisis in Kenya: The Case for Land Reform Brookfield, VT: Gower Publishing Co., 1984)Google Scholar; Patrick O. Alila, Gatheru Wanjohi, and Kabiru Kinyanjui, “Rural Landlessness Situation in Kenya,” a report for the FAO Expert Consultation on “Landlessness: Dynamics, Problems and Policies” (Rome: October 1985); and Livingstone, Ian, Development, Employment and Incomes in Kenya Brookfield, VT: Gower Publishing Co., 1986)Google Scholar.

⁶ See Bates, Robert, Markets and States in Tropical Africa Berkeley: University of California Press, 1981)Google Scholar; Hyden, Goran, Beyond Ujamaa in Tanzania Berkeley: University of California Press, 1980)Google Scholar; and Popkin, Samuel J., The Rational Peasant Berkeley: University of California Press, 1979)Google Scholar.

⁷ The data reported in this article were collected in 1980 by a survey conducted by Joel D. Barkan of the University of Iowa and David Gachuki of the University of Nairobi. Interviews were obtained from 2,075 adults living in seven parliamentary constituencies in rural Kenya: Embu-South, Githunguri, Kajiado-North, Laikipia-West, Ikolomani, Mbita, and Kitutu- East. Separate samples of 300 respondents were drawn for each constituency based on age-sex quotas computed from the 1979 Kenyan census. Interviews for each sample of 300 respondents were randomly assigned to 30 sampling plots one kilometer in diameter, which were themselves randomly distributed across the constituency. Ten interviews were thus attempted in each plot; all were conducted in the local language of the area or in Swahili.

⁸ Several explanations for the weak relationship between income and the size of landholdings have appeared in the literature. Gavin Kitching has presented data that demonstrate that smallholders farm their land more intensively and place a greater proportion of their land under cultivation than do large farms. Ecological conditions also account for lower productivity per acre and hence lower farm income per acre on large holdings, many of which are located in less fertile regions. A substantial proportion of personal income, moreover, is derived from off-farm sources. See Kitching, , Class and Economic Change in Kenya New Haven: Yale University Press, 1980), 330–74Google Scholar, and Lijoodi, J. L. and Ruthenberg, Hans, “Income Distribution in Kenya's Agriculture,” Zeitschrift für Ausländische Landwirtschaft 17 (1978), 124Google Scholar, as cited by Peterson, Stephen, “Neglecting the Poor: State Policy Toward the Smallholder in Kenya,” in Commins, Stephen K. et al. , eds., Africa's Agrarian Crisis: The Roots of Famine Boulder, CO: Lynne Rienner, 1986), 59–83Google Scholar, at 62.

⁹ While 11% of teachers and 17% of civil servants said they owned more than 20 acres, 19% and 28% respectively reported they owned none. Similarly, shopkeepers ranked first in average size of landholdings, but fifth in average income. These figures also make us question the thesis advanced by Kitching that the level of income derived from off-farm employment is the determinant of the size of landholdings in Kenya. Off-farm income may, but does not necessarily, lead to control over large landholdings—a qualification that Kitching himself concedes. See Kitching, ibid., 371.

¹⁰ We attempted just such an analysis by first excluding the respondents who did not fall into these three categories, and then comparing those who did in terms of their level of participation in self-help, their values, and their level of political involvement. Even after dividing our sample into these “ideal” types, class distinctions were not evident.

¹¹ Although the statistical relationship between the amount of land owned and measurable income is weak, the extent of people's landholdings or access to land greatly determines the level of overall wealth for the large segment of the rural population that farms for subsistence in addition to farming for the market. Income figures reported by our respondents do not include estimates of the cash value of their subsistence production, but estimates of their landholdings do provide an indirect measure of this component of individual wealth. la Central Bureau of Statistics, Integrated Rural Survey,1974–75 Nairobi: Government Printer, 1977), 44Google Scholar.

¹³ In order of importance, the eight types are nursery schools, primary schools, secondary schools, health clinics, cattle dips, water projects, mutual assistance groups, and production groups.

¹⁴ Michael Bratton reaches the same conclusion to explain why peasant farmer organizations in Zimbabwe also draw a disproportionate amount of their membership from “middle peasant” households. See Bratton, “Farmers' Organizations and Food Production in Zimbabwe,” World Development 14 (No. 3, 1986), 367–84CrossRefGoogle Scholar, at 373.

¹⁵ Because self-help projects require collective action, they are always susceptible to the problem of “free riders”; most peasants are cautious about joining such efforts. That they do so anyway suggests that they have no other option. Our finding is thus consistent with Popkin's discussion of the conditions under which peasants are prepared to engage in collective action. See Popkin (fn. 6), 252–58.

¹⁶ Only 11% of our respondents reported that they had at some time been coerced into contributing to a self-help project; there was little variance across landholding groups. Yet, when asked whether it was sometimes necessary to force people to contribute to Harambee projects for the good of the local community, 56% of the small peasant respondents and 52% of the middle peasants said “yes”; only 40% of the large landholders answered affirmatively. Of the landless respondents, 49% also favored coerced contributions. Barbara Thomas reports similar findings; see fn. 4, 154–56.

¹⁷ Barbara Thomas reports a similar figure of 88%. Ibid., 164.

¹⁸ A 1971 survey of Nairobi wage earners indicated that 21% of their incomes were remitted to the rural areas. The figure is suggestive, although no breakdown was given for remittances to self-help as distinct from those to family members. Cited in Collier, Paul and Lai, Deepak, “Why Poor People Get Rich in Kenya, 1960–79,” World Development 12 (No. 10, 1984), 1007–18, at 1016CrossRefGoogle Scholar.

¹⁹ Republic of Kenya, Sessional Paper No. 1 of 1986 Nairobi: Government Printer, 1986), 29–32Google Scholar.

²⁰ Ministry of Planning and National Development, Economic Survey 1987 Nairobi: Government Printer, 1987), 68, 178–82Google Scholar.

²¹ Peterson (fn. 8), 59–83.

²² Recent research by Robert Bates in Bungoma, Kakamega, Nandi, and Trans Nzoia Districts of Kenya indicates that smallholders, via their elected M.P.s, have become increasingly successful in securing higher official prices for their principal crop, maize. Between 1982 and 1986, the average producer price for maize increased 85% compared to 81% for coffee and 74% for tea—crops that are normally associated with relatively affluent smallholders. See Economic Survey 1987 (fn. 20), 107.

²³ Sessional Paper No. 1 of 1986 (fn. 19), 29–32. See also Cohen, John M. and Hook, Richard M., “District Development Planning in Kenya,” Development Discussion Paper No. 229 (Cambridge: Harvard Institute for International Development, 1986), 60–61, 65–67Google Scholar.

²⁴ The official name of the policy of decentralization is District Focus for Rural Development. It seeks to control all development efforts in the rural areas by requiring that all projects funded or partially supported by the state be reviewed by a series of local development committees culminating in the District Development Committee.

²⁵ Cohen and Hook (fn. 23), 77.

²⁶ The prospects for such events cannot be ruled out. In 1982, Kenya experienced an attempted coup following the amendment of the country's constitution which made Kenya ade jure one-party state. Since the early 1980s, President Moi has intervened periodically in the process by which funds are raised for self-help by holding large public rallies for large showpiece projects which he controls, thus undercutting local initiatives for smaller projects under local control. In 1988, the President announced several controversial changes in the procedures for holding parliamentary elections. These included the holding of primary elections in which voters were required to queue behind pictures of the candidates they supported instead of casting secret ballots. Turnout in the primary elections was extremely low (about 13% of those eligible by age). The primaries and the general elections that followed were marked by an unusually high number of allegations of fraud.