¹ See Geddes, Barbara, “What Do We Know about Democratization after Twenty Years?” Annual Reviews of Political Science 2 (1999), 124Google Scholar, for a classification of regimes in which the military plays a role; and Linz, Juan J. and Stepan, Alfred, Problems of Democratic Transition and Consolidation: Europe, South America, and Post-Communist Europe (Baltimore and London:Johns Hopkins University Press, 1996Google Scholar), Table 3.1, 44–45, for a classification of nondemocratic regimes. See also Crouch, Harold, “Patrimonialism and Military Rule in Indonesia,” World Politics 31 (July 1979), 587CrossRefGoogle Scholar, in which the author describes Suharto s New Order Indonesia as a patrimonial regime “built on nonpatrimonial foundations.”

² See, for example, Crouch, Harold, The Army and Politics in Indonesia (Ithaca, N.Y.:Cornell University Press, 1978Google Scholar); and Jenkins, David, Suharto and His Generals: Indonesian Military Politics, 1975–83 (Ithaca, N.Y.:Cornell University Press, 1984Google Scholar).

³ The distinction between transition and consolidation can be traced back to Rustow, Dankwart, “Transitions to Democracy: Toward a Dynamic Model,” Comparative Politics 2 (April 1970), 357CrossRefGoogle Scholar–63. We use the broad definition of transition found in O'Donnell, Guillermo and Schmitter, Philippe C., Transitionsfrom Authoritarian Rule: Tentative Conclusions (Baltimore and London:Johns Hopkins University Press, 1986Google Scholar) as an interval between two regimes. A transition begins with the process of termination of an authoritarian regime, and ends with the establishment of an alternative regime-be it democratic, authoritarian, or some alternative form. For a classification of nondemocratic regimes and the types of agendas that must be accomplished for a successful transition to democracy to take place, see Linz and Stepan (fn. 1).

⁴ See O'Donnell, Guillermo, Schmitter, Philippe C., and Whitehead, Laurence, Transitionsfrom Authoritarian Rule: Comparative Perspectives (Baltimore and London:Johns Hopkins University Press, 1986Google Scholar), and Diamond, Larry, Developing Democracy: Toward Consolidation (Baltimore:Johns Hopkins University Press, 1999Google Scholar). This is, of course, not always the case. In the transitions of the former Warsaw Pact states, for example, the role of the military was generally small.

⁵ Agiiero, Felipe, “Democratic Consolidation and the Military in Southern Europe and South America,” in Gunther, Richard, Diamandouros, P. Nikiforos, and Puhle, Hans-Jurgen, eds., The Politics ofDemocratic Consolidation: Southern Europe in Comparative Perspective (Baltimore and London:Hopkins University Press, 1995), 165Google Scholar.

⁶ For work in the social sciences that harmonizes individual rationality with altruism and loyalty, see, for example, Becker, Gary S., “Altruism in the Family and Selfishness in the Market Place,” Eco nomica 48 (February 1981Google Scholar).

⁷ While these may no doubt have been factors in the transition, as will be shown, they complemented important structural forces.

⁸ See Huntington, Samuel P., The Third Wave: Democratization in the Late Twentieth Century (Norman and London:University of Oklahoma Press, 1991Google Scholar); and Geddes (fn. 1).

⁹ See, for example, Janowitz, Morris, The Professional Soldier: A Social and Political Portrait (Glencoe, Ill.:Free Press, 1960Google Scholar); van Doom, Jacques, ed., Armed Forces and Society: SociologicalEssays (The Mouton, 1968Google Scholar); and Bienen, Henry, ed., The Military and Modernization (Chicago and New York:Al-dine and Atherton, 1971), 4Google Scholar.

¹⁰ See Casper, Gretchen and Taylor, Michelle M., Negotiating Democracy: Transitionsfrom Authori tarian Rule (Pittsburgh:University of Pittsburgh Press, 1996), 19Google Scholar.

¹¹ The importance of incentives operating at the individual level has been observed in the context of a number of militaries, including those of Bolivia, Panama, Nigeria, and Thailand. See, for example, Geddes (fn. 1), 126. To this list must be added the case of Indonesia, where, in the Suharto era, the military was considered a safe means of individual enrichment and accrual of power.

¹² This is now made explicit in definitions of democracy. See, for example, Karl, Terry Lynn, “Dilemmas of Democratization in Latin America,” Comparative Politics 23 (October 1990), 2CrossRefGoogle Scholar.

¹³ See Agiiero (fn. 5), 127.

¹⁴ Stepan, Alfred, The Military in Politics: Changing Patterns in Brazil (Princeton, N.J.:Princeton University Press, 1971), 9Google Scholar.

¹⁵ Agiiero (fn. 5), 139.

¹⁶ See Maniruzzaman, Talukder, Military Withdrawalfrom Politics (Cambridge:Ballinger Press, 1987Google Scholar)

¹⁷ See Peck, Mitchell B., “Assessing the Career Mobility of U.S. Army Officers: 1950–1974,” Armed Forces & Society 20 (Winter 1994CrossRefGoogle Scholar).

¹⁸ See Porch, Douglas, The Portuguese Armed Forces and the Revolution (London:Croom Helm, 1977Google Scholar); and Maxwell, Kenneth, “Regime Overthrow and the Prospects for Democratic Transition in Portugal,” in O'Donnell, Guillermo, Schmitter, Philippe C., and Whitehead, Laurence, eds., Transitions from Authoritarian Rule: Southern Europe (Baltimore and London:Johns Hopkins University Press, 1986Google Scholar).

¹⁹ See Baynham, Simon, The Military and Politics in Nkrumah's Ghana (Boulder, Colo, and London:Westview Press, 1988Google Scholar); Kammen, Douglas and Chandra, Siddharth, A Tour of Duty: Changing Patterns of Military Politics in Indonesia in the 1990s, Monograph No. 75 (Ithaca, N.Y.:Cornell Modern Indonesia Project, 1999Google Scholar); Luckham, Robin, The Nigerian Military: A Sociological Analysis (Cambridge:Cambridge University Press, 1971Google Scholar and 1980); Porch (fn. 18); and Stepan (fn. 14).

²⁰ The “transition” is broadly defined as the period between the last months of Suharto's presidency and the months following the election of Abdurrahman Wahid and it includes the interim government of Bacharuddin Jusuf Habibie from May 1998 to October 1999. The “consolidation” begins with the selection of Abdurrahman Wahid as president; the end of the “transition,” therefore, overlaps with the beginning of the “consolidation.”

²¹ See van Dijk, Kees, A Country in Despair: Indonesia between 1997 and 2000 (Leiden:KITLV 2001Google Scholar).

²² Indonesia's experience is entirely in keeping with the emerging consensus, based on comparative study of dozens of transitions, that poor economic performance raises the probability of a transition from authoritarian regimes. For research on this, see Diamond, Larry, Linz, Juan, and Lipset, Seymour, Democracy in Developing Countries: Latin America (Boulder, Colo.:Lynne Rienner, 1989Google Scholar); Bermeo, Nancy, “Rethinking Regime Change,” Comparative Politics 22 (April 1990CrossRefGoogle Scholar); and Adam Przeworski, and Limongi, Fernando, “Modernization: Theories and Facts,” World Politics 49 (January 1997Google Scholar).

²³ Between 1964 and 1999 the Indonesian military was called Angkatan Bersenjata Republic Indonesia (Armed Forces of the Republic of Indonesia), abbreviated ABRI. On April 1, 1999, this was changed to Tentara Nasional Indonesia (Indonesian National Army, TNl), a name dating from the Indonesian Revolution (1945–49). We will employ these as appropriate for the period under discussion.

²⁴ This sentiment prevailed despite the kidnapping of political activists in early 1998 by Special Forces (Kopassus), the shooting of four student protesters at Trisakti University in Jakarta on May 12, and the instigation of massive riots in Jakarta and several provincial cities on May 13–14.

²⁵ For example, while the military wished to buy the latest military equipment from international suppliers, it was stuck with the inferior goods produced by the state munitions manufacturer, directly under Habibie's control. When the military wished to purchase new naval craft, Habibie brokered a deal for thirty-nine used East German warships that the navy had no choice but to accept.

²⁶ Habibie agreed to hold new national elections in June 1999 and to allow the People's Consultative Assembly (MPR Rl) to meet late that year to select a president. The election saw the Indonesian Democratic Party of Struggle (PDI-P) come in first with 34 percent of the vote, followed by Golkar (the old ruling party) and two Islamic parties, the National Unity Party (PKB) and the National Mandate Party (PAN). Most observers believed that PDI-P's victory would be translated into the selection of party chairwoman Megawati Sukarnoputri, daughter of Indonesia's first president, as the PAN chairperson Amien Rais and several of the smaller Islamic parties secured the selection of Wahid as president.

²⁷ In November 1999, more than one hundred officers were rotated, including the appointment of a number of officers who had supported reforming the military's political role. In February 2000, another reshuffle was announced involving seventy-four senior personnel, including the replacement of outspoken commander of the Army Strategic Reserve (Kostrad)-Major General Djamari Chaniago- by the reform-minded Major General Agus Wirahadikusumah. In June 2000 there was another rotation of 122 officers. On August 1 yet another reshuffle was announced, which, though somewhat smaller, included changes in a number of strategic posts.

²⁸ Recent reshuffles and promotional patterns among the Indonesian military elite have been analyzed in detail in the journal Indonesia. See Editors, “The Indonesian Military in the 1990s: Political Maneuvering or Structural Change?,” Indonesia 63 (April 1997Google Scholar); idem, “Current Data on the Indonesian Military Elite: October 1,1995 to December 31,1997,” Indonesia 65 (April 1998); idem, “Current Data on the Indonesian Military Elite: January 1, 1998 to January 31, 1999,” Indonesia 67 (April 1999); idem, “Changes in Civil-Military Relations since the Fall of Suharto,” Indonesia 70 (October 2000); and idem, “Current Data on the Indonesian Military Elite: January 1, 1999 to January 31, 2001,” Indonesia 71 (April 2001). The data used in this paper draw in part on the data published in these articles.

²⁹ See Kammen and Chandra (fn. 19), 30–36.

³⁰ The only institutional outlet for excess officers was secondment (kekaryaan) to positions as local and regional executives and as legislators, and to posts in the civil service, even though this was only common for officers in their mid-forties and older. See Kammen, Douglas, “Military Politics and Kekaryaan in Post-Suharto Indonesia,” in Tremewan, Chris, comp., Indonesia after Suharto (Auckland:New Zealand Asia Institute, University of Auckland, 1999Google Scholar).

³¹ See Appendix 1 for details.

³² The finding that the overall decrease in command tenure is best explained by the changing size of the officer corps suggests two related hypotheses. First, if command tenure is a function of the size of the officer corps, it would seem likely that variation in tenure would be observable between classes of different sizes, with officers from large classes enjoying short tours of duty while officers from small classes receiving comparatively longer tours of duty. Surprisingly, however, there is no direct correspondence between class size and tenure per class in a given position. Officers from large academy classes do not necessarily serve for shorter lengths of time than do officers from smaller classes; the data in Table 1 suggest that tenure is inversely related to the size offive-yearcohorts. Classes 1–4, who commanded the various Kodam during the second half of the 1980s, had an average tenure of 650–850 days. The larger Class 5 1965 that followed these small and highly successful classes suffered in terms of both total number of officers appointed as well as mean tenure.

The second cohort of graduates, Classes 6–9, had nearly identical tenures—at roughly 525 days per officer—despite considerable differences in class size. The third cohort, Classes 10–14, is larger still, and here we see a further decline in overall tenure as well as in rough parity between classes, with mean tenure 250–450 days. It is important to note that there are officers from Classes 10, 11, and 12 currently serving as Kodam commanders who have not been included in these calculations, and that Classes 13 and 14 are likely to receive additional appointments. The second corollary hypothesis is that there is a correlation between class size and the relative success of each class in terms of the total number of appointments received. In other words, with no change in the total number of available billets, the percentage of officers appointed to command a particular type of unit should be higher for small classes and lower for large classes. Once again, however, there is no direct correspondence between class size and the relative success of a particular academy class. Rather than showing an even distribution of command billets by class, Table 1 shows three “peaks."The first peak represents the combination of Classes 1,2 and 3, which received 10,13, and 12 Kodam commands, respectively, and which saw over 10 percent of their officers appointed to these posts. The success of these three classes served to block the promotion of officers from Classes 4 and 5, accounting for the sharp decrease in the total number of Korem (for Class 5) and Kodam (for both classes) commands received. The second peak represents Class 6 1965, which received an extraordinary fifty-nine Korem commands (13.6 percent of all officers) and fourteen Kodam commands (for a more modest 3.2 percent of all officers). Again, the exceptional success of Class 6 1965 prevented the collective promotion of officers from the next three graduating classes. The third peak occurs for Kodam appointments for Classes 10 1970 and 11 1971, and for Korem appointments for Classes 13 1973 and 14 1974.

³³ See, for example, Crouch (fn. 1); Shiraishi, Takashi, “The Indonesian Military in Politics,” in Schwarz, Adam and Paris, Jonathan, eds., The Politics of Post-Suharto Indonesia (New York:Council Foreign Relations Press, 1999Google Scholar); and Honna, Jun, “Military Ideology in Response to Democratic Pressure During the Late Suharto Era: Political and Institutional Contexts,” Indonesia 67 (April 1999CrossRefGoogle Scholar).

³⁴ Theoretically, it is possible to employ this method to analyze all staff and command positions in the army and armed forces headquarters, though differences in rank and in the significance of billets makes such an exercise impractical. For present purposes it is sufficient to limit the analysis to the two largest units in the Army's territorial structure by analyzing the forty subregional military commands and then moving up the chain of command to the eleven regional military commands.

³⁵ This was partially compensated for by the appointment of several of these officers to the most senior military posts, as seen, for example, in the appointment of General Wiranto, a Class 9 1968 graduate, to the post of ABRI commander in chief.

³⁶ These models are presented and discussed in Appendix 2.

³⁷ “ABRI, Perombakan di Luar Dugaan” (Unexpected shake-up), Forum Keadilan, September 1, 1994.

³⁸ “Penvira angkatan 70 hams berperan sesudah Pemilu nanti” (Class '70 officers must have a role after the next election), Angkatan Bersenjata, June 16, 1996.

³⁹ The most notable of these were Bambang Yudhoyono (son-in-law of General Sarwo Edie), who was appointed to command the Sriwijaya Kodam in 1997, replacing an officer seven classes his senior; Prabowo Subianto (son-in-law of Suharto and a Class 14 1974 graduate), who was appointed commander of Special Forces in 1995-replacing an officer four classes his senior-and made commander of the Army Strategic Command in April 1998-replacing an officer three classes his senior; Sjafrie Syamsoedin (a Class 14 1974 graduate with close ties to Prabowo) who was appointed commander of the Jakarta Kodam, replacing an officer five classes his senior; and Ryamizard Ryacudu (son of General Ryacudu) who was appointed commander of the First Infantry Division (Army Strategic Reserve) in April 1998, replacing an officer three years his senior.

⁴⁰ Shiraishi (fn. 33).

⁴¹ See Harold Crouch, “Wiranto and Habibie: Military-civilian relations since May 1998,” in Budiman, Arief, Hatley, Barbara, and Kingsbury, Damien, Reformasi: Crisis and Change in Indonesia (Clayton, Victoria:Monash Asia Institute, 1999), 137Google Scholar–39.

⁴² These included former TNI Commander in Chief General Wiranto (coordinating minister for politics and security), Lieutenant General Susilo Bambang Yudhoyono (minister of mines and energy), Lieutenant General Agum Gumelar (minister for transportation), Rear Admiral Freddy Numberi (minister of administrative reforms), and Lieutenant General Surjadi Soedirdja (minister of home affairs).

⁴³ In 1999 Agus held the powerful post of assistant for general planning and in early 2000 was appointed commander of Kodam VII Wirabuana (covering the island of Sulawesi). As of June 2000 he was the commander of the powerful Army Strategic Reserve Command (Kostrad).

⁴⁴ See Wirahadikusumah, Agus, “Hubungan Sipil-Militer: Visi, Misi dan Aksi” (Civil-military relations:Vision, mission and action) (Manuscript, May 1999Google Scholar).

⁴⁵ See Wirahadikusumah, , Indonesia Baru dan Tantangan TNI: Pemikimn Masa Depan (Jakarta:Pus-taka Sinar Harapan, 1999Google Scholar).

⁴⁶ This is particularly true of Major General Djaja Suparman, who served as commander of Kodam Jaya Jakarta from July 1998-November 1999, then as commander of the Army Strategic Reserve from November 1999-March 2000, only to be replaced by his nemesis Agus Wirahadikusumah and shuffled off to the unimportant post of commander of the Military Staff and Command School (Sesko TNI).

⁴⁷ See, for example, TNI Watch!, “Orang Dekat Suharto Masih Dapat Tempat di TNI-AD” (Officers close to Suharto still have a place in the army), posted on SiaR@minihub.org. dated June 21,2000.

⁴⁸ Editors (fn. 28,2000), 133.

⁴⁹ During this period the post of TNI chief of staff was held by Lieutenant General Fachrul Razi (10 1970), Lieutenant General Sugiono (10 1970), Lieutenant General S. Marasabessy (11 1971), and Djamari Chaniago (11 1971); that of army chief of staff was held by General Subagyo Hadisiswoyo (10 1970), General Tyasno Sudarto (10 1970), and General Endriartono Sutarto (11 1971); and that of deputy army chief of staff was held by Lieutenant General Sugiono (10 1970), Lieutenant General Johnny Lumintang (10 1970), Lieutenant General Djamari Chaniago (11 1971), Lieutenant General Endriartono Sutarto (11 1971), and Lieutenant General Kiki Syahnakri (11 1971).

⁵⁰ Note, however, that this may not be possible for particular cohorts of officers. A large number of officers from Classes 16–20 (1976–81) are tainted by their involvement in East Timor, particularly during 1999, and officers from Classes 21–23 (1982–84) are now deeply implicated in the ongoing violence in Aceh.

⁵¹ These cases reflect a broader pattern in which attention is paid to the structure of the military only when the military stages a coup. A unique aspect of the current study is the detailed analysis of structure in a case in which the military did not intervene decisively in national politics even when, by all accounts, it had the means to do so. In terms of the choice of cases, Alfred McCoy's excellent account of the role of the Philippine military in the transition of 1986, with its unique methodology and emphasis on military socialization and class solidarity, also makes for an interesting comparison with the Indonesian case; see Alfred McCoy, Closer than Brothers: Manhood at the Philippine Military Acad emy (New Haven and London: Yale University Press, 1999).

⁵² Luckham (fn. 19).

⁵³ Ibid., 132, also chap. 5 (“Peer Groups: Solidarity and Cleavage”), in particular, sections in that chapter titled “Cohesion: 'Friends and Class-mates'” and “Cleavage: The Differentiation of Peer Groups.”

⁵⁴ Ibid., 36–37.

⁵⁵ Baynham (fn. 19), 158, Table 8.1,161 and Table 8.2,162.

⁵⁶ Ibid., 163.

⁵⁷ Samudavanija, Chai-Anan, The Thai Young Turks (Singapore:Institute of Southeast Asian Studies, 1982Google Scholar).

⁵⁸ Ibid., 46–47. Similarly, senior officers from Class 5 staged the subsequent coup in 1991. See Suchit, Bun-bongkarun, State of the Nation: Thailand (Singapore:Institute of Southeast Asian Studies, 1996Google Scholar).

⁵⁹ Ibid., 47. A quotation from a statement by the chief spokesman of the “Young Turks” group runs “we are the Class of 1960 .. . we went to fight in Laos .. . in the jungle with the Communist terrorists. ... We were closely united, all of us determined in our pursuit of the same objective” (p. 35).

⁶⁰ Ibid., 41. On the related subject of intercohort dynamics, the author also points out the importance of “elder-younger brother relationships,” such as those between the members of Class 7 and those of Classes 12,13, and 14 (pp. 64–65).

⁶¹ Porch (fn. 18).

⁶² Ibid., 73.

⁶³ Maxwell (fn. 18), 111. The MFA (Movimento das Forcas Armadas or Armed Forces Movement) was the group that these officers formed.

⁶⁴ Luckham (fn. 19), 174. The author's argument is based in part on his Table 32 (Ibid., 175), which shows clearly the unusual speed with which the 1957–59 cohort advanced through the lower ranks.

⁶⁵ Ibid., 174.

⁶⁶ See Ted R. Gurr, Why Men Rebel (Princeton: Princeton University Press, 1970).

⁶⁷ Chai-Anan (fn. 57), 69.

⁶⁸ Ibid., chap. 5. “This term, adopted from statistical methodology, is the bias of coefficient estimates in a regression in which a possibly important variable has not been included among the independent variables. If that variable is truly unimportant and it is included in the analysis, it will not affect the estimates. If it is truly important, however, its omission renders the results of the incompletely specified regression invalid.

⁷⁰ See Porch (fn. 18); and Maxwell (fn. 18).

⁷¹ This was the famous Decree Law 353–73. See Porch (fn. 18), 64–66.

⁷² See Kammen and Chandra (fn. 19).

⁷³ Porch (fn. 18), 68, 74.

⁷⁴ Baynham (fn. 19), Table 6.1,114.

⁷⁵ Ibid., 260.

⁷⁶ Ibid., 176.

⁷⁷ The analysis of the Thai case should, for this reason, be considered incomplete. This is perhaps justifiable on the grounds of scarcity of data, a common problem when it comes to analyzing the structure of a corps of military officers.

⁷⁸ Agiiero (fn. 5).

⁷⁹ See, for example, O'Donnell, Schmitter, and Whitehead (fn. 4); and Diamond (fn. 4).

⁸⁰ Bienen (fn. 9).

⁸¹ Finer, as quoted in McCoy (fn. 51), 5.

⁸² For statistical evidence of the relationship, see Siddharth Chandra, “Regional Commanders in the Indonesian Army: An Analysis of Promotional Logjam in the 1990s” (Manuscript, Graduate School of Public and International Affairs, University of Pittsburgh, Pittsburgh, June 2000).

⁸⁴ Recall that larger cohort size leads to shorter command tenure and higher turnover, which, in turn, leads to higher numbers of appointees from larger classes.

⁸⁵ This latter effect is a function of class size, suggesting that the variables LCLASSIZE. and CLAS-SIZEt may be correlated with LTENURE. Note that if two independent variables are correlated, the parameter estimates are still unbiased, though the standard errors may be large. In any case, since the phenomenon of interest is the effect of being in a particular class after controlling for the effects of class size and tenure, the specific testing of hypotheses with respect to these variables is of secondary interest.

⁸⁶ Because the age variable is a discrete number of years and therefore inexact in terms of number of days, it was not added to predecessor tenure to yield age at exit for the predecessor, which would have been a more robust structural variable.

⁸⁷ These include Kodam I, II, Jaya, III, IV, V, VI, VII, and VIII. In order to eliminate the linear dependence of the KODAM. vector with the intercept, Kodam IX was treated as the default Kodam. Insufficient data on the recently created Kodam XVI precluded the use of Pangdam data for that Kodam.

⁸⁸ For reasons of linear dependence, dummy variables for AMN classes 2,5, and 8 were excluded from the estimation. The substance of the results remained the same if alternative AMN class dummies were dropped.

⁸⁹ For details of the computation of these statistics, see SAS Institute, Inc., SAS/STAT User's Guide, vol. 3 (Raleigh:SAS Institute, Inc., 2000Google Scholar).

⁹⁰ For a discussion of the AMN 6 monopoly phenomenon, see Kammen and Chandra (fn. 19), 48. While the contribution of membership in AMN 9 is high, it is statistically insignificant.

⁹¹ Similar analysis of succession patterns for Korem commanders revealed a monopoly effect for Class 6. For lack of space, however, the Korem commander results are not presented.