¹ The new nations are very numerous and too short-lived to warrant definitive judgments about the durability of their political systems. Analysis of polities in older nations should provide a baseline of generalizations which can be used later to assess the prospects of the new nations. The numbers of states included, by world region, are:

Europe 49^a

Central and South America 23

Africa 4

Asia 15

^aIncluding Australia, Canada, New Zealand, the Orange Free State, South Africa, and the United States.

² Eckstein, Harry, “Authority Patterns: A Structural Basis for Political Inquiry,” American Political Science Review, 47 (December, 1973), pp. 1153Google Scholar. For a survey of conventional academic definitions of “authority” see Peabody, Robert L., Organizational Authority (New York: Atherton Press, 1964)Google Scholar, chap. 2.

³ In Eckstein, Harry and Gurr, Ted Rober, Patterns of Authority: A Structural Basis For Political Inquiry (New York: Wiley-Interscience, forthcoming)Google Scholar, Part II.

⁴ At the time this study was begun we treated Decision Constraints as a simple and distinct dimension. We later came to regard it as a synthetic concept, i.e., a multiple function of a number of other dimensions. In either conceptualization it is a dimensional concept and can be used, as it is here, to characterize and distinguish among political entities.

⁵ Eckstein, Harry, The Evaluation of Political Performance: Problems and Dimensions (Beverly Hills: Sage Professional Papers in Comparative Politics, 1971), Vol. 2, No. 01–017Google Scholar. Gabriel Almond has identified five categories of system performance, which have to do with how well such functions as regulation, extraction, and distribution are performed; see his “Political Development: Analytical and Normative Perspectives,” Comparative Political Studies, 1 (January, 1969), pp. 460–467Google Scholar.

⁶ See Hurwitz, Leon, “Contemporary Approaches to Political Stability: A Review Essay,” Comparative Politics, 5 (April, 1973), pp. 449–463CrossRefGoogle Scholar.

⁷ See, for example, Buckley, Walter, Sociology and Modern Systems Theory (Englewood Cliffs, N.J.: Prentice-Hall, 1967)Google Scholar and Ashby, W. Ross, Introduction to Cybernetics (New York: Wiley, 1956)CrossRefGoogle Scholar.

⁸ Gurr, Ted Robert and McClelland, Muriel, Political Performance: A Twelve-Nation Study (Beverly Hills: Sage Professional Papers in Comparative Politics, 1971, No. 018), pp. 72–85Google Scholar. The dependent variable of this earlier study was labeled Durability but was defined as Persistence is in this study.

⁹ Some of these problems are discussed by Gurr, Ted Robert, “Vergleichende Analyse von Krisen und Rebellionen,” in Janicke, Martin, ed., Herrschaft und Krise (Opladen: Westdeutscher Verlag, 1973), pp. 68–69Google Scholar, and by Eckstein, , The Evaluation of Political Performance, pp. 69–73Google Scholar. Other approaches to crises and crisis-response are to be found in Binder, Leonard et al. , Crises and Sequences in Political Development (Princeton: Princeton University Press, 1971)Google Scholar.

¹⁰ These have been used in collateral research. See Eckstein and Gurr, Parts II and IV, and the studies cited there for use of these dimensions in collateral research.

¹¹ I have resorted to a neologism, suggested by Harry Eckstein, because no conventional term describes well the kind of state I have in mind: one which has minimal functions, an uninstitutionalized pattern of political competition, and executive leaders constantly imperiled by rival leaders. The term has the same Greek root as anarchy, which means literally the absence of any sort of rule. Anocracy, on the same basis, means literally the absence of power or control, though of course I use it to signify states which approach but do not reach the extreme condition. Archetypical examples in history include the Holy Roman Empire and many lesser medieval principalities.

¹² The literature includes a number of indicators of “democracy.” One is Fitzgibbon's set of judgmental indicators of the extent of democracy in Latin American countries; see Fitzgibbon, Russell H., “Measuring Democratic Change in Latin America,” Journal of Politics, 29 (February, 1967), 129–166CrossRefGoogle Scholar. An index of “democratic development” based on electoral characteristics is used by Neubauer, Deane E., “Some Conditions of Democracy,” American Political Science Review, 61 (December, 1967), 1002–1009CrossRefGoogle Scholar. Arthur S. Banks has recently assessed empirically the extent of what he calls “democratic performance” in a large number of nations in the 19th and 20th centuries, in “Correlates of Democratic Performance,” Comparative Politics, 4 (January, 1972), 217–230CrossRefGoogle Scholar. A critical review of this literature is May, John D., Of the Conditions and Measures of Democracy (Morristown, N.J.: General Learning Press, 1973)Google Scholar. But there seem to be no attempts to index degrees of “autocracy” or any variable similar to “anocracy.”

¹³ Uganda is not among the countries included in the analyses of this paper; it was coded in an extension of the study to “new nations.”

¹⁴ Singer, J. David and Small, Melvin, “The Composition and Status Ordering of the International System, 1815–1940.” World Politics, 18 (January, 1966), 236–282CrossRefGoogle Scholar.

¹⁵ Preliminary results of analyses by Robert Harmel of Northwestern University, using more restrictive definitions of what constitutes a “polity change,” are consistent with most of the findings reported below. An important exception is his finding that some sets of incoherent polities, especially in Europe, have been as durable as coherent democracies and more so than coherent autocracies.

¹⁶ Langer, William L., comp. and ed., An Encyclopedia of World History (Boston: Houghton Mifflin, 4th ed., 1968)Google Scholar; Banks, Arthur S., comp., Cross-Polity Time-Series Data (Cambridge, Mass.: MIT Press, 1971.)Google Scholar

¹⁷ The following materials cannot be published for reasons of space but are available from the author and the Inter-University Consortium for Political Research: the coding guidelines and category definitions; frequency distributions of coded authority data; list of countries coded; and dates of polity transformation in each. Available from the Consortium only are the full set of coded data for all polities and narrative summaries of the political events and changes for each polity, with lists of sources used.

¹⁸ See Gurr and McClelland. This is a convergent validation test rather than a test of intercoder reliability because different instruments and sources were used in the two studies. Carryover from the first to the second studies was minimized by two circumstances: the coder in the present study did not examine the results of the first study before or during the coding, and the author compared the results of the studies only after the completion of data collection in the second study.

¹⁹ Banks, Cross-Polity Time-Series Data, codes “type of regime” in four categories, civilian, military-civilian, military, other; and also rates the effectiveness of legislatures. Despite the lack of discrimination in these sets of categories, they permit some general comparisons with the coded data on polities. Michael Ward of Northwestern University has used the Banks data to construct approximately-comparable indicators of five authority dimensions. Scores on his indicators, taken at the midpoints of polity lifespans, have these Pearson r coefficients with the scores used in this paper: Openness .50; Decision Constraints .77; Participation .61; Directiveness .52; Complexity .81.

²⁰ The “life expectancy” estimates are of total polity lifespan, including attained age. For example, a 50-year-old Third-World polity would be expected to survive another 50 years for a 100 year total. The figures are equivalent to medians, not averages: they are the age beyond which less than half the polities of a cluster (including historical and contemporary polities) are known to have survived. The inclusion of contemporary polities increases the medians for polities at year 0 over the medians which are listed for regional and historical clusters of polities in the text, above.

²¹ Note that the independent variables are structural properties of the political system, not properties of other social institutions: data are not available here to test directly Harry Eckstein's proposition that congruence between nongovernmental and governmental authority patterns determines the performance and durability of political systems. See his “Authority Relations and Governmental Performance: A Theoretical Framework,” Comparative Political Studies, 2 (October, 1969), 283Google Scholar.

²² Huntington, Samuel P., Political Order in Changing Societies (New Haven: Yale University Press, 1968), pp. 1–92Google Scholar.

²³ Eckstein, , “Authority Relations and Governmental Performance,” pp. 300–307Google Scholar.

²⁴ The life expectancy values for all polities are used rather than those of the regional or temporal clusters because, first, the N's for the subsets are too small, and second, because the subsets are categorical variables whose effects on persistence are to be assessed separately.

²⁵ A logarithmic transformation is conceptually less appropriate because too strong: it compresses the upper range of persistence valves more than is justified given this study's special focus on the determinants of longevity.

²⁶ Five polities are double-counted: three are above both the autocratic and anocratic threshold of 3.5, two are in both the democratic and anocratic subsets.

²⁷ “Authority Relations and Governmental Performance,” pp. 315–322.

²⁸ It is assumed here that the variables represented by “region” and “era” control the effects of authority traits on durability, and that there is no consistent relation between “region” or “era” and durability. If these assumptions are not warranted, then other causal sequences are also possible. See, for example, Blalock, Hubert M. Jr.,, Causal Inferences in Non-experimental Research (Chapel Hill: University of North Carolina Press, 1964)Google Scholar, and Aiker, Hayward R. Jr.,, Mathematics and Politics (New York: Macmillan, 1965)Google Scholar, chap. 6.

²⁹ See Hurwitz, Leon, “Democratic Political Stability: Some Traditional Hypotheses Reexamined.” Comparative Political Studies, 4 (January, 1972), 476–490CrossRefGoogle Scholar; and “An Index of Democratic Political Stability: Methodological Note.” Comparative Political Studies, 4 (April, 1971), 41–68CrossRefGoogle Scholar.

³⁰ “Authority Relations and Governmental Performance.”

³¹ There are other activist polities in Latin America now than Mexico, of course; their ideologies and approaches are as diverse as those of Brazil, Peru, and Cuba. But they are all young, and whether they will endure and adapt as has the Mexican system is problematical; the “life expectancy” of Third-World polities at their inception is 16 years (Table 6).

³² The Imperial Order (Berkeley and Los Angeles: University of California Press, 1967), p. 443Google Scholar.

³³ Wesson, The Imperial Order, preface (unnumbered).