The analysis of military regimes, as opposed to military coups, has attracted comparatively little attention. This paper examines whether the political, military, and economic performance of systems which have experienced a military regime differs from the performance of systems which have not. The comparison between the performance of these two types of system is then used to examine the validity of the occurrence of a military regime as an indicator of instability. The population consists of all independent countries of the world. The time span examined is 1961–70. The comparison is made across a number of political, military, and economic variables. The basic comparison is elaborated by introducing controls for GNP, area, number of coups, and duration of the military regime. While the controls show a number of variations, the main summary finding is that it is easy to differentiate military and non-military regime systems in political terms, but not in military and economic terms. This finding seriously questions the utility of the occurrence of a military regime as an indicator of instability.
1 The works vary substantially both in scope and method. The most common form consists either of individual case studies or anthologies of case studies. Examples are Welch, C. E., ed., Soldier and State in Africa (Evanston: Northwestern University Press, 1970); Bienen, H., ed., The Military Intervenes (New York: Russell Sage Foundation, 1968); Hurewitz, J. C., Middle East Politics: The Military Dimension (New York: Praeger, 1969); Lieuwen, E., Arms and Politics in Latin America (New York: Praeger, 1961); Johnson, J. J., ed., The Role of the Military in Underdeveloped Countries (Princeton: Princeton University Press, 1962); Van Doorn, J., ed., Armed Forces and Society (The Hague: Mouton, 1968). There is also a smaller collection of synthesizing works. See, for example, Janowitz, M., Military in the Political Development of New Nations (Chicago: University of Chicago Press, 1964); Finer, S. E., The Man on Horseback (New York: Praeger, 1962); Gutteridge, W., Military Institutions in the New States (New York: Praeger, 1965); Luttwak, E., Coup of État (New York: Knopf, 1969).
2 See, for example, S. E. Finer (above); Pye, L. W., “Armies in the Process of Political Modernization,” European Journal of Sociology, 2, no. 1 (1961), pp. 82–92. Hopkins, K., “Civil-Military Relations in Underdeveloped Countries,” British Journal of Sociology, 17 (06, 1966), 165–182; Zolberg, A. R., “Military Rule and Political Development in Tropical Africa,” in Military Profession and Military Regimes, ed. Van Doorn, J. (The Hague: Mouton, 1969).
3 Examples of case studies are Luckham, R., The Nigerian Military (Cambridge: Cambridge University Press, 1971); Be'eri, E., Army Officers in Arab Politics and Society (New York: Praeger, 1970); Stepan, A., The Military in Politics: Changing Patterns in Brazil (Princeton: Princeton University Press, 1971). Examples of more general statements are Feit, E., “Pen, Sword and People,” World Politics, 25 (01, 1973), 251–273; Nordlinger, E. A., “Soldiers in Mufti: The Impact of Military Rule Upon Economic and Social Change in the Non-Western States,” American Political Science Review, 64 (12, 1970), 1131–1148; Willner, A. R., “Perspectives on Military Elites as Rulers and Wielders of Power,” Journal of Comparative Administration, 2 (11, 1970), 261–276; Pinkney, R., “The Theory and Practice of Military Government,” Political Studies, 21 (06, 1973), 152–166; Bienen, H., ed., The Military and Modernization (Chicago: Aldine, Atherlon, 1971).
4 Communist Countries were excluded because of non-comparability of economic data.
5 Some of the recent standard data compilations such as Taylor, C. L. and Hudson, M. C., World Handbook of Political and Social Indicators (New Haven: Yale University Press, 1972) and Banks, A., Cross-Polity Time Series Data (Cambridge: M.I.T. Press, 1971) were not adequate, given the wide range of variables used.
6 The main data sources for the political variables were Statesman's Yearbook (London: Macmillan, 1950–1970); Europe Yearbook (London: Europa Publications, 1958–1970); Stebbins, Richard P. and Amoia, Alba, eds., Political Handbook and Atlas of the World (New York; Simon and Schuster, 1970); Whitaker's Almanac (London: Whitaker's Publications, 1950–1970); and Keesing's Contemporary Archives (London: Keesing's Publications, 1950–1970).
7 The main data sources were The Military Balance (London: Institute for Strategic Studies, 1960–1972); Stockholm International Peace Research Institute, Yearbook of World Armaments and Disarmament (Stockholm: Almquivst, 1968–1972); Booth, Richard, “The Armed Forces of African States,” Adelphi Papers, #67 (London: Institute for Strategic Studies, 1970).
8 The main data sources were United Nations, Statistical Yearbook (New York, 1951–1972); United Nations, Yearbook of International Trade Statistics (New York, 1950–1972); United Nations, Yearbook of National Account Statistics (New York: 1964–1972); and International Financial Statistics (Washington: International Monetary Fund, 1964–1972).
9 The main area not covered is monetary transfers involving transactions of deposit money banks and central banks.
10 Per capita GNP is divided into the following categories: $0–$150, $151–$400, $401–$900, and $901 +. This last category is not used as it contains no military regime systems. The geographic areas are Central America and the Caribbean, South America, The Middle East and North Africa, Sub-Saharan Africa, Central and Southeast Asia. Duration is categorized as 0–1 year, 1–2 years, 2–5 years and 5 + years. The number of coups is categorized as 0, 1, 2, and 3 +.
11 The scores for each of the clusters are contained in the first appendix.
12 The basic comparison used throughout this paper is between all military regime systems and nonmilitary regime systems. In most comparisons we have removed the high-income systems. A high-income system is defined as one which has a mean per capita GNP of $900 or more for the period between 1961–70. The eighteen high-income systems were removed primarily because they differ so markedly in economic terms from the rest of the world. This was confirmed by the cluster analysis. Since no military regime systems fall into the high income group the high-income system would distort the picture of the nonmilitary regime system. In the aggregate comparisons data on all nonmilitary regime systems and on low-income nonmilitary regime systems were included, but wherever comparisons using the various controls were done, the high-income systems were excluded.
13 The independent variables used in the regression analyses were: a dummy variable for whether a military regime had occurred, the percentage of the cabinet that was made up of military personnel, size of military, expenditure on the military, per capita GNP, rate of growth of GNP, and trade balance. All regression equations are significant at the .01 level. In the case of constitution ban 37 per cent of variance is explained; for legislature ban 53 per cent of variance is explained; and for military in the cabinet 58 per cent is explained. It is interesting to note that percentage military in the cabinet is a significant estimator of constitution and legislature ban; military size is a significant estimator of constitution ban and military in the cabinet; and military expenditure is a significant estimator of legislature ban. Thus a number of military variables, independent of whether or not there has been a military regime, seem to be related to restrictions on levels of political activity.
14 T-tests on the constitution ban show that the ban in military regime systems (0–2 years) is significantly lower than the ban in military regime systems (2–5 years) at the .01 level and that the ban on the latter is significantly lower at the .05 level than the ban in military regime systems (5 +). The same results hold for the percentage of the cabinet made up of military personnel. Difference tests on the legislature ban show the ban in military regime systems (0–2 years) is significantly lower at the .01 level than the bans in military regime systems (2–5 years and 5 +).
15 Difference of means tests show that there are significant differences between noncoup systems and each of the one, two and three coup categories at the .01 level on legislative ban, percentage military in the cabinet and number of constitution changes. For the constitution ban the differences are significant at the .01 level between noncoup systems and the one- and two-coup systems and at the .05 level between non-coup systems and the three-coup systems. On none of these variables are there any differences that are significant between the one-, two-, and three-coup systems. On number of executives there are significant differences between the noncoup and each of the different coup categories at the .01 level, but there are also some differences between the coup categories. The two-coup category has more executives than the one-coup category. This is significant at the .05 level. The three-coup category has more executives than the one- or two-coup categories (significant at the .01 and .05 levels respectively).
16 The multiple regressions contained the same variables as did the earlier ones except that the dummy variable of whether a military regime had occurred was replaced by the variables of duration and number of coups. All the regression equations were significant at the .01 level with 43 per cent, 60 per cent, and 67 per cent of the variance of constitution ban, legislative ban, and military in the cabinet being explained.
17 The simple correlations of number of coups with number of constitution changes and number of executive changes are .62 and .51 respectively, both significant at the .01 level. When duration is controlled, these correlations become .49 and .52, respectively, both still significant at the .01 level. The simple correlation between duration and number of constitution changes and number of executives are .45 and .17, the first of which is significant at the .01 level. When number of coups is controlled, these correlations become .14 and −.19, only the latter of which is significant at the .05 level.
18 In Africa there are significant differences on all political activity and political change variables. In Asia the same holds true with the exception of the ban on political parties. In the Middle East and North Africa significant differences occur only on the percentage military in the cabinet and party bans. In both Central and South America there are significant differences on the percentage military in the cabinet, party ban, and legislation bans.
19 All regression equations are significant at the .01 level, but the amounts of variance explained for each of the rates of growth are not impressive. The independent variables used for each of the rates of growth are the dummy variable of whether a military regime has occurred, diversification, size, and expenditure of the military, GNP, rate of growth, budget, and international liquidity. The variance explained in each case is 21 per cent. The amount of variance explained for size and expenditure are 65 per cent and 59 per cent respectively. The independent variables for size are the dummy variable, diversification, expenditure, GNP, rate of growth of GNP, budget and international liquidity. The variables for expenditure are identical to the above except that size of the military is substituted for expenditure.
20 The differences are the following: Sub-Saharan Africa, where the low-income nonmilitary regime systems rate of growth of expenditure is significantly higher at the .05 level; Central America, where the rate of growth of size is significantly higher in the military regime systems at the .10 level; South America, where the size in the low-income nonmilitary regime systems is significantly higher at the .10 level.
21 For example, military size in military regime systems and low-income nonmilitary regime systems ranges from 9.6 and 9.2 in Africa to 91.1 and 94.0 in Asia; expenditure ranges from 0.87 and 1.40 in Central America to 6.86 and 6.30 in the Middle East and North Africa. Diversification ranges from 2.34 and 1.94 in Africa to 29.0 and 32.0 in South America. Rates of growth of size range from 4.30 and 1.35 in Central America to 13.8 and 11.7 in Africa. Rates of growth of expenditure range from 3.10 and 1.70 in Central America to 12.3 and 19.3 in Africa.
22 In Africa, military regime systems have lower GNP, lower GDFCF and fewer doctors. Each difference is significant at the .10 level. In Asia, military regime systems have lower GNP, smaller budgets and fewer doctors. The differences are significant at the .05 level.
23 The partial correlations between type of regime and exports and type of regime and imports controlling for GNP are −.23 and −.19, respectively, significant at the .01 and .05 levels.
24 The regressions on trade balance and direct investment balance are both significant at the .01 level. In the former, 58 per cent of variance is explained and in the latter, 40 per cent. The independent variables are the dummy of whether a military regime has occurred, GNP, rate of growth of GNP, exports, international liquidity, central government capital balance. Direct balance and trade balance were alternatively used where appropriate.
25 The performance rates of the African military regime systems are consistently lower than their nonmilitary counterparts with the rate of growth of GNP being lower at the .10 level. The Asian military regime systems, which had a similar profile to the African military regime systems in the two previous categories, have poorer performance rates on cost of living and exports than the low income nonmilitary regime systems but a better rate of growth of GNP, but the differences are not significant. The military regime systems of Latin America have significantly higher export growth rates (at the .01 level) and lower cost of living increases (significant at the .05 level for Central America). The military regime systems of the Middle East and North Africa have better performance rates than the nonmilitary counterparts, but the differences are not significant.
26 Difference of means tests show that the military regime systems of 5+ years differs from low income nonmilitary regime systems, military regime systems (0–2 years), and military regime systems (2–5 years) at the .05, .10, and .01 levels respectively.
27 The regression is significant at the .01 level and explains 34 per cent of variance of the rate of growth of exports. The independent variables are the number of coups, duration, GNP, rate of growth of GNP, exports, international liquidity, and GDFCF.
* The authors wish to thank the Social Science Research Council (England) which provided funds for this project, and Anthony Mughan who acted as research assistant.
Email your librarian or administrator to recommend adding this journal to your organisation's collection.
Email your librarian or administrator to recommend adding this journal to your organisation's collection.
* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.
Usage data cannot currently be displayed