¹ Watanabe, Tsuneo, Habatsu: Hoshuto no Kaibou (Factions: Anatomy of the Conservative Party) (Tokyo: Kobundo, 1964Google Scholar); Thayer, Nathaniel B., How the Conservatives Rule Japan (Princeton: Princeton University Press, 1969Google Scholar); Fukui, Haruhiro, Party in Power: The Japanese Liberal Democrats and Policy Making (Berkeley and Los Angeles: University of California Press, 1970Google Scholar); idem, “The Liberal Democratic Party Revisited: Continuity and Change in the Party's Structure and Performance,” Journal of Japanese Studies 10 (Summer 1984Google Scholar); Sato, Seizaburo and Matsuzaki, Tetsuhisa, Jiminto Seiken (LDP rule) (Tokyo: Chuo Koronsha, 1986Google Scholar); Baerwald, Hans, Party Politics in Japan (Boston: Allen and Unwin, 1986Google Scholar); Curtis, Gerald L., The Japanese Way of Politics (New York: Columbia University Press, 1988Google Scholar).

² Immediately after its establishment, the LDP was administered by a collective leadership, which consisted of two ex-Liberals and two ex-Democrats. Ichiro Hatoyama, the former president of the Democrats, was subsequently chosen as the first LDP president, after the death of his rival, Taketora Ogata, the former Liberal president, in early 1956. Hatoya-ma's decision to retire in late 1956 led to a competition for the leadership among these eight powerful politicians (Nobusuke Kishi, Ichiro Kohno, Takeo Miki, and Tanzan Ishibashi of the former Democrats; and Hayato Ikeda, Eisaku Sato, Banboku Ohno, and Mitsujiro Ishii of the former Liberals).

³ Fukui (fn. 1, 1970), 116–17; Uchida, Kenzo, Habauu (Factions) (Tokyo: Kodansha, 1983), 26–30Google Scholar.

⁴ There has been a tendency for smaller groups to depart from thefivefactions when there is a change in leadership, but these defecting groups also tend eventually to disappear or merge with the major factions. For a detailed lineage of LDP factions, see Sato and Matsu-zaki(fn. 1), 241.

⁵ It is extremely difficult to identify exactly who belongs to which faction and thus to determine the precise number in each faction. Although, technically, factional membership includes only elected members of the Diet, there are some who are temporarily out of office but are strongly affiliated with their factions. However, having accounted for the short-term fluctuations due to such electoral turnover, the recent increase in the average factional membership remains an undeniable trend.

⁶ Iseri, Hirofumi, Habatu Saihensei (Factional realignment) (Tokyo: Chuo Koronsha, 1988Google Scholar).

⁷ Sato and Matsuzaki (fn. 1); Curtis (fn. 1).

⁸ Muramatsu, Michio and Krauss, Ellis S., “Bureaucrats and Politicians in Policymaking: The Case of Japan,” American Political Science Review 78 (March 1984CrossRefGoogle Scholar); Inoguchi, Takashi and Iwai, Tomoaki, ”Zok≥ Giin” no Kenfyu (A study of “policy tribes”) (Tokyo: Nihon Keizai Shimbunsha, 1987Google Scholar).

⁹ Ishikawa, Masumi, Deeta Sengo Seiji-shi (Postwar political data in history) (Tokyo: Iwa-nami Shoten, 1984), 219Google Scholar–21.

¹⁰ Sato and Matsuzaki (fn. 1), 71–72.

¹¹ Ohira eventually appointed Kunikichi Saito, another senior member of his own faction. This appointment, however, was not a result of overriding the principle of separation of powers; rather, it should be seen as an exceptional face-saving compromise, because Saito was universally respected by Liberal Democrats for his earnest personality.

¹² Nakane, Chie, TateShakpi no Ningen Kunkei (Human relations in vertical society) (Tokyo: Kodansha, 1967Google Scholar); Doi, Takeo, Amae no Kozo (The structure of dependence) (Tokyo: Kobundo, 1971Google Scholar).

¹³ Curtis (fn. 1), xii; emphasis added.

¹⁴ Thayer (fn. 1) describes the evolution of an LDP faction as follows: Take, for example, the Kono faction. … The drive for membership in the faction started with a series of banquets in Mukojima, a geisha district, … to which Kono's lieutenants invited Dietman who, they thought, would rally to Kono's cause. The site was carefully chosen. It was far enough outside the usual haunts of the Dietmen so that any guest would not have to explain his attendance to other Dietmen, but not so distant that the invitees would have to make an undue effort to attend.… The Dietmen ate, drank, joked with the geisha, listened to the raconteurs, and at the end of the evening, signed a book to commemorate the occasion.… Kono's only factional membership list was the banquet books, and the only yardstick of factional loyalty was the distance the Dietmen were willing to travel to eat and drink with him.… In this period, any faction could be defined as a group of politicians who travelled to the resorts and hot springs for “seminars” and met for an occasional meal in Tokyo, (pp. 22–23)

¹⁵ Thayer (fn. 1), 41.

¹⁶ Richardson, Bradley M. and Flanagan, Scott C., Politics in Japan (Boston: Little, Brown, 1984), 102Google Scholar.

¹⁷ Baerwald (fn. 1), 17.

¹⁸ Curtis (fn. 1), 88–89.

¹⁹ Sato and Matsuzaki (fn. 1), 41–44.

²⁰ For the latter, see Leiserson, Michael, “Factions and Coalitions in One-Party Japan: An Interpretation Based on the Theory of Games,” American Political Science Review 62 (September 1968CrossRefGoogle Scholar).

²¹ The two exceptions are Ishibashi's appointment of party officials in 1956 and Ikeda's appointment in 1964. Interestingly enough, in both cases Takeo Miki was appointed as the secretary-general. Moreover, both Ishibashi and Ikeda became ill and resigned soon after making the appointment.

²² Curtis (fn. 1), 81.

²³ Downs, Anthony, An Economic Theory of Democracy (New York: Harper and Row, 1957), 28Google Scholar.

²⁴ Implicit in this assumption is that these institutional constraints are exogenous in the sense that all political actors must take the legal regulations as “givens” in making actual organizational choices; they are thus unable, at least in the short run, to change these regulations in any significant way.

²⁵ Thus, the impact of political dynamics at an interparty level is considered negligible and/or irrelevant as a source for the strategic behavior of LDP politicians and LDP supporters. This assumption is justifiable, given the expectation of these actors that the domination in the Diet by the LDP will continue for the foreseeable future.

²⁶ Petnpel, T. J., Policy and Politics in Japan (Philadelphia: Temple University Press, 1982), 36Google Scholar.

²⁷ Since 1955 there has been only one single-member district (in the Amami Island district). In the most recent, 1986 redistricting one six-member district (Hokkaido) and four two-member districts (Ishikawa, Hyogo, Kagoshima, and Nigata prefectures) were created as exceptions.

²⁸ Since there are usually more candidates wanting to run for the LDP than the optimal number that the party nominates, the nomination process produces a strange phenomenon in many electoral districts: some candidates run with support from LDP factions but without the official nomination of the party. While this may split the conservative vote and thus seem to work against the LDP, the net effect remains ambiguous. Should the officially nominated candidates lose and the unofficial candidates with factional support win, the LDP can always then nominate the winners in order to ensure its majority in the House.

²⁹ Kohno, Masaru and Nishizawa, Yoshitaka, “A Study of the Electoral Business Cycle in Japan: Elections and Government Spending on Public Construction,” Comparative Politics 22 (January 1990CrossRefGoogle Scholar).

³⁰ One can also argue that from the perspective of LDP supporters LDP factions have a rational foundation. Suppose a voter is calculating his “expected candidate differential” among two or more LDP candidates. The most important element in this calculation is his judgment about the relative ability of each local candidate to influence the central policymaking authority, such as high-ranking Japanese bureaucrats and/or influential members of policy tribes within the LDP. Whether his vote will be translated into tangible benefits for his own welfare depends almost entirely on this ability, which in turn depends on the candidate's background, especially her connections with the relevant ministries and bureaus, her relationship with senior LDP members, her experience, and her overall influence in the policy-making process. Obviously, it is extremely difficult and costly for the voter to collect such information for all candidates. The information may not be available to the public, and even if it were, it might not be credible. The voter knows, however, that affiliation with a major LDP faction almost guarantees the candidate a network of personal ties with the established elites and thus direct access to the policy-making processes in important issue-areas. In this sense, factional affiliation plays a role similar to Downs's “political ideology”: it helps voters make their voting decisions without knowing about every candidate specifically and in detail. See Downs (fn. 23), 96–113; and John Ferejohn, “The Spatial Model and Elections” (Paper presented for the Downs Conference, University of California, Irvine, 1989).

³¹ There is ample evidence that these disincentives actually determine factional affiliation. A well-known recent example is the case of candidate Hayashi in Kochi prefecture. Hayashi had been an elected member of the House of Councillors, but he lost his seat in the July 1989 election (in which the LDP lost its majority in the upper house for the first time since 1955). In 1990 the House of Representatives was dissolved, and Hayashi decided to run in the February general election. Although he was originally a member of the Miyazawa faction, he knew that he would hot receive factional endorsement because that faction already had representation in the lower house in the district. Hayashi therefore left the Miyazawa faction for the Takeshita faction before the election.

³² Many examples of testimony by factional leaders during the 1950s and 1960s not only point to the existence of a certain optimal factional size but also confirm the basic logic of the explanation developed here. For example, Shojiro Kawashima, who led a minor faction, was once quoted as saying, “The appropriate size of a faction is about twenty-five, given the financial constraints”; quoted in Sato and Matsuzaki (fn. 1), 61. Banboku Ohno, who led a major faction, is reported to have said that the optimal number was about forty. Beyond this number, he explained, factions would face three major problems: the possibility that members would have to compete within the same electoral district, the difficulty of soliciting sufficient contributions, and the shortage of cabinet and party positions for promotion; see Watanabe (fn. 1), 2–3; and Curtis (fn. 1), 83.

³³ AsahiShimbun, December 12, 1987.

³⁴ Iwai, Tomoaki, ”Seiji Shikjn” no Kenkyu (A study of “political funding”) (Tokyo: Nihon Keizai Shimbunsha, 1990), 106Google Scholar.

³⁵ Duverger, , Political Parties: Their Organization and Activity in the Modern State (New York: Wiley, 1963Google Scholar).

³⁶ Reed, , “Structure and Behavior: Extending Duverger's Law to the Japanese Case,” British Journal of Political Science 20 (July 1990), 336CrossRefGoogle Scholar.

³⁷ See Douglas Rae's treatment of Canada as an exception to the simplest version of Du-verger's proposition that plurality formulas cause a two-party system; Rae, , The Political Consequences of Electoral Laws (New Haven: Yale University Press, 1967), 94Google Scholar.

³⁸ The static cross-sectional analysis (and the replication of such analyses) that I have conducted may not constitute a sufficient proof for the actual selection process, which is dynamic in nature. Ideally, one would need a large-scale panel survey in order to verify more directly that LDP politicians and LDP supporters act rationally and that they really abandon the losing factions over the long run. However, given that such a survey has never been conducted in Japan, I believe that the data and analysis presented here remains the best available evidence.

³⁹ In reality, cabinet shuffles occur about once a year, and a new set of party officials is usually selected at about the same time.

⁴⁰ Goto, Motoo, Uchida, Kenzo, and Ishikawa, Masumi, Sengo Hoshu Seiji no Kisekj: Yoshida Naikaku kara Suzuki Naikaku made (History of the postwar conservative politics: From Yo-shida cabinet to Suzuki cabinet) (Tokyo: Iwanami Shoten, 1982), 139Google Scholar.