Institutionalized parties and party systems have traditionally been viewed as necessary conditions for democracies to function effectively. Although this area of research is germane to all democracies, most analyses have been divided by regional investigation. Seeking to bridge the gap, this article applies concepts and measures of institutionalization from the study of Latin America to Pacific Asia's two most prominent cases of democratic transition, South Korea and Taiwan. An effort is made to apply the approaches of Dix and Mainwaring and Scully on party and system institutionalization in Latin America to South Korea and Taiwan. Cross-national comparison reveals a curvilinear relationship between institutionalization and consolidation. Taiwan's path to consolidation has been predicated on a pattern very similar to those taken by Latin American cases, whereas South Korea, theoretically, should not be as close to consolidation as it is. POLITICAL PARTIES, PARTY SYSTEMS, AND DEMOCRACY IN EAST ASIA Lessons From Latin America HANS STOCKTON University of St. Thomas As a result of the proliferation of democratic transitions during the past decade and a coinciding decline in the number of new cases, scholars have turned increasingly toward addressing issues concerning the consolidation of new democratic regimes. On making the transition from authoritarian rule, new democracies face the immediate dilemmas of establishing political institutions that deal effectively with "authoritarian legacies" while also being capable of addressing new problems that arise from the introduction of democratic government. In addition, because elections lie at the heart AUTHOR'S NOTE: I would like to thank Michelle Taylor-Robinson for her encouragement and helpful comments during the preparation of this manuscript. I would also like to acknowledge the continuing support of the Government Information Office in Taiwan and the Taipei Economic and Cultural Office in Houston, Texas. COMPARATIVE POLITICAL STUDIES, Vol. 34 No. 1, February 2001 94-119 © 2001 Sage Publications, Inc. 94 Stockton / CROSS-NATIONAL ANALYSIS 95 of democratic politics and parties are viewed as the appropriate vehicles to power, the presence of a strong party system has long been viewed as a necessary step toward strengthening the performance of democratic rule (Powell, 1982; see also Duverger, 1954; Huntington, 1968; Sartori, 1976). A necessary condition for a strong system of parties, therefore, is that political parties be institutionalized as well (Dix, 1992; Huntington, 1991; Mainwaring, 1992; Mainwaring & Scully, 1995). The consensus has been that two particular changes have to occur for democracy to consolidate in newly transiting polities. First, politics must be a relatively free competition between parties within a regularized system of competitive elections. Second, political parties must adapt themselves to the new rules of the game, such as electoral laws and legislative-executive jurisdictions, which structure the who, what, when, and how of elite behavior in democratic systems. This article builds on the research presented in the edited volume Building Democratic Institutions by Mainwaring and Scully (1995). While developing the kernel for future research into comparative party system institutionalization in relatively new democracies in Latin America, the authors directed less effort to systematically exploring the subsequent implications for the quality of a country's democracy. This shortcoming is minor given the main thrust of the volume, but it is a gap that this article attempts to fill. In the Mainwaring and Scully volume, party system institutionalization is treated as a necessary, but not sufficient, condition for democratic consolidation. The authors are correct in their assertion that the process of institutionalization is neither a unilinear nor irreversible process.1 But if not linear, then what is the nature of the relationship? This article seeks an answer to this question. The goals for this research are two-fold. First, using Dix (1992) and Mainwaring and Scully (1995) as a foundation, assumptions are tested about the direction of the relationship between political party and competitive party system institutionalization and quality of democracy in Latin American presidential systems relative to each other.2 Second, the challenge put forth by Mainwaring and Scully (1995) to extend this investigation into another region of the world by examining transitions in East Asia is pursued. 1. Indeed, the curvilinear nature of institutionalization has been addressed elsewhere (Diamond, 1997; Huntington, 1968; Powell, 1982). 2. Newly installed democracies are not created equal. The persistence of authoritarian institutions and actors may place a considerable drag on the rate of consolidation. Thus, comparison is a complicated issue. Some countries have experienced prior democratic intermissions, whereas others have not, and thus many of the political elite have had some experience with the corresponding institutions. Political parties, especially those in Chile, Colombia, Honduras, and 96 COMPARATIVE POLITICAL STUDIES / February 2001 PARTY AND PARTY SYSTEM INSTITUTIONALIZATION Few can deny that democracy at the state level cannot operate effectively for very long, or at all, without the organizing and channeling capabilities of political parties. This importance is recognized by Duverger (1954), Huntington (1968), Lijphart (1984), Lipset (1981), Powell (1982), and Sartori (1976, 1987), to list a few examples. In fact, research by Zhang (1994) has indicated that cases of mass-driven regime change have never resulted in democracy in the posttransition period. Standard definitions of democracy revolve around three primary characteristics: (a) There must be meaningful and regularly scheduled competition for government power, free from coercion by force; (b) the selection of political leaders must result from political participation through regular and free elections that do not exclude any major group; and (c) a minimum level of political and civil liberties such as freedom of expression, assembly, press, and association is necessary to sustain the integrity of the political process. Typically, we see direct references to a system of elections in standard definitions of democracy but no mention of the necessity of political parties. However, when we expand questions about democracy to include concepts such as stability, efficiency, and consolidation, the important impact of the actors on the system becomes plain. Improved democracies are perpetuated by the routinization of a political process based on political parties that are consistent to a large degree in their ideological and policy positions, parliamentary and extraparliamentary discipline, and organization. When discussing the concept of an institutionalized party system, confusion frequently arises around an important distinction. For example, are we referring to an institutionalized system of gaining power based around political parties or a system, not necessarily routinized, revolving around institutionalized political parties? These two concepts are frequently interchanged and treated as synonymous, although most often referencing a routinized system of elections based on political parties. Thus far, analyses of party and systemic institutionalization, although acknowledging a nonlinear relationship, have failed to treat these two conceptually correlated variables as ultimately distinct. This is highlighted in the two primary studies (Dix, 1992; Mainwaring & Scully, 1995) from which we draw in this article. Paraguay, have celebrated or are close to celebrating centennials. Therefore, their ability to "lay low" during a succession of authoritarian regimes has given them time to refine their organizational and mobilization skills. Stockton / CROSS-NATIONAL ANALYSIS 97 Political elites and masses must acknowledge that the system of gaining political power through elections has become the legitimate institution for gaining political power. This, of course, does not prevent personalities from using the party system for electoral gain. Accordingly, the central actors (political parties) within that system must also accept the normative and organizational value of routinization and themselves institutionalize it within their own political behavior, expressed through political parties. This is not to say that parties must become mechanical and inflexible but that the behavior of political parties must be self-regulated within some predictable set of limits. The political party system is the set of rules that governs the externaliza-tion of political parties to the mass public. A party system is "a set of interactions in the competition among parties" (Mainwaring & Scully, 1995, p. 4). Note that the standard emphasis on political parties as the central actors in the system is maintained. Whereas a party system cannot be conceptualized without mention of parties, it must be operationalized with an eye to excluding indicators that measure aspects of party and not system. An institutionalized party system is one that is distinguished by a routinized election calendar and relatively consistent election rules that determine who votes and how winners are determined. Such characteristics determine the regularized pattern of interactions between political parties and which parties enter or exit political competition. These are key components to the system. Another aspect of party system institutionalization stresses the positioning of parties within the system from one election to the next. "Party system institutionalization entails the stabilization and social embeddedness of the major party alternatives and their relative policy positions, hence a regularity in the patterns of interparty competition" (Toka, 1997, p. 96). An in-depth discussion of the definition of political party will not be presented here because the important distinction between parties and party system is that the parties are the actors that operate within the encompassing structure of a party system. Institutionalized parties are those that demonstrate consistent patterns of internal organization, mass mobilization, and leadership succession. In addition, such parties must become autonomous from their founding personalities and, for example, survive their departure from the political scene. To this point, the standard arguments that set rules and procedures for political parties are important for democracy has been repeated. In addition, the presence of a competitive electoral system consisting of parties as the primary actors is also important for democracy. Yet, how are these two variables related to each other and, together, to democracy in general? Given our assumptions about the importance of parties and party systems in shaping democracy, we should usually find a positive correlation between 98 COMPARATIVE POLITICAL STUDIES / February 2001 the development of an institutionalized party system and positive development in political and civil freedoms. However, as we illustrate below in the empirical analysis, the relationship is neither linear nor the inverted u-curve that most often comes to mind. Can democracy exist and, better yet, flourish under an institutionalized system of elections that are contested by uninstitutionalized political parties? Can the same conditions be prompted when the actors are de-institutionalizing political parties? Conversely, can democracy exist and flourish within a noninstitutionalized party system in which its actors are considered to be institutionalized? Following from above, this article tests three propositions that examine the quality of a democracy at a given point in time. Because the analysis presented here is not dynamic in the sense that change over time, within a country, is not conducted, framing the following discussion in terms of consolidation would not be appropriate. The propositions are: (a) as the degree of party system institutionalization increases, the quality of democracy increases; (b) as the degree of political party institutionalization increases, the quality of democracy increases; and (c) as the degree of both political party and party system institutionalization increases, the quality of democracy increases. METHOD The methodology of this article combines indicators used by Dix (1992) for measuring political party institutionalization and those used by Mainwaring and Scully (1995) for measuring party system institutionalization. I seek to replicate their party and system institutionalization scores for the cases of South Korea and Taiwan. I then attempt to build on their work by testing for the effect that party and system institutionalization have on overall quality of democracy by examining each country's score in conjunction with Freedom House's freedom scores of political and civil rights. Because party and system scores are not available for every unit in every year, a time series analysis will not be attempted. Although capturing the dynamic element of change over time would be enlightening and will eventually be essential, conducting a static cross-sectional analysis at least allows us to answer an important question: Do prior levels of institutionalization act as a predictor for the level of democracy at a given point in time? Due to the small sample size, a more sophisticated regression technique will not be used in this article. Cross-national comparisons will be made on the basis of the relative system and party institutionalization rankings of each country with their relative quality of democracy rankings in the year following the final Stockton / CROSS-NATIONAL ANALYSIS 99 Table 1 Electoral Volatility Across Legislative Elections Country Time Span Electoral Periods Mean Volatility Colombia 1970-1990 5 8.5 Uruguay 1971-1989 2 9.1 Argentina 1983-1993 5 12.7 Chile 1973-1993 2 15.8 Venezuela 1973-1993 4 17.7 Costa Rica 1970-1990 5 18.2 Mexico 1982-1991 3 22.4 Paraguay 1983-1993 2 25.8 Ecuador 1978-1992 5 32.5 Bolivia 1979-1993 4 33.0 Taiwan 1983-1992 4 33.3 Brazil 1982-1990 2 40.9 Peru 1978-1990 3 54.4 South Korea 1981-1992 4 72.3 Source: Latin American scores from Mainwaring and Scully (1995). year in which data were collected on the Mainwaring and Scully (1995) data sets.3 Thus, the Freedom House score for Argentina, for example, is assessed at its 1992 level because its system institutionalization scores were collected up to 1991. One additional year has been added to the freedom scores to allow for the various timing of elections among the sample cases. Dix's (1992) measures indicating improvement or decline in party institutionalization represent parties as of 1989. As in the system indicator, the countries are ranked according to their relative improvement or decline in party institutionalization. For the independent variables, election, case study, and survey data (when available) will be collected for South Korea (1985-1995) and Taiwan (1986-1995) to measure the following indicators for party system institutionalization from Mainwaring and Scully (1995): stability in the rules of interparty competition, the development of societal support for parties, and emphasis on elections as the means of selecting rulers. A fourth criterion used by Mainwaring and Scully, development of strong party organization, is excluded from analysis of system institutionalization because it is explicitly concerned with the internal organization of political parties, not party systems. See Table 1 for system institutionalization scores with the additions of South Korea and Taiwan. 3. Judging from the data presented in Mainwaring and Scully (1995), the authors used the most recent congressional or presidential election at the time they were writing Building Democratic Institutions. 100 COMPARATIVE POLITICAL STUDIES / February 2001 Indicators used by Dix (1992) for measuring political party institutionalization are adaptability (indicated by party age and generational succession and change in opposition and/or ruling status), complexity (indicated by control of the party by an individual and structural organization and differentiation), autonomy (indicated by the extent to which parties are independent of other social groupings), and coherence (indicated by Rae's Index of Fractionalization). A problem arises with the comparability of the data between the Mainwaring and Scully (1995) and the Dix (1992) studies. Mainwaring and Scully (1995) presented measures of system institutionalization accounting for a single time period from the 1970s into the early 1990s to arrive at a final aggregate institutionalization score for each polity. The goal of Dix's (1992) presentation, however, was to offer a comparison of political party institutionalization between two time points, 1969 and 1989. Consequently, his summary measures are presented as improving, declining, or not changing in the specified number of categories. The former article leaves us in question as to how each country fared relative to the overall sample at individual points in time, whereas the latter article leaves us uncertain as to how each country fared in terms of where they started relative to the overall sample. The test of the relationship between variables is a simple one, based on finding the best line of fit through the sample universe. Three lines are attempted: linear, quadratic, and cubic; the line with the best fit best expresses the direction of the relationship. The value of this simple test is that it provides us with a visualization of the empirical relationship between cases. Once the sample size is increased sufficiently, regression techniques can be applied that will verify the significance of the independent variables. POLITICAL PARTIES AND PARTY SYSTEM INSTITUTIONALIZATION IN ASIA This section explores the prospects for continued entrenchment of democratic politics in South Korea and Taiwan by drawing on the above analyses of Latin American cases. These two cases have been chosen because of the abundance of existing research from which to draw and because they are two of the only three cases of democratization in presidential systems in Third Wave Pacific Asia. I attempt to replicate measures used by Mainwaring and Scully (1995) for identifying levels of party system institutionalization and Dix's (1992) measures of party institutionalization in Latin American democracies and apply them to the new democracies of Pacific Asia. Stockton / CROSS-NATIONAL ANALYSIS 101 Examining the development of party and party system institutionalization has been a major field of inquiry in case studies on South Korea and Taiwan. Applying these cases to the first criterion of institutionalization, regular patterns of interparty competition, is relatively easy, conducted by measuring Pedersen's index of electoral volatility in lower chamber seats. By measuring the volatility of seat distribution, it is assumed that low levels of volatility are best, whereas high levels are not conducive to institutionalization. As illustrated in Table 1, we see that both Taiwan and South Korea exhibit high degrees of volatility in the change in percentage of seats held by parties from one election to the next. This statistic is quite misleading, however. Taiwan, for example, began a transition from single-party dominance by the Nationalist Party (KMT) in 1986 when the Democratic Progressive Party (DPP) was allowed to run in legislative elections. Thus, with its beginning percentage of seats at nearly 85% in the Legislative Yuan as of 1983, the KMT could only lose seats to a newly legalized opposition. Conversely, with its beginning percentage of seats at 0%, the opposition could only gain seats. A moderately successful opposition gain in seats across elections would indicate a certain stability in the number of seats held by the same party from one election to the next. But in the case of Taiwan, a low level of change in seat distribution would not have been reflective of democracy in any sense, but rather would reflect a continued domination by the KMT. In South Korea, we see a particularly high Pedersen score of 72.3 %, which surpasses the highest score in the Latin American sample for Peru (54.4%). This measure truly captures the seemingly chaotic entry and exit of political parties in the Korean political scene from one election to the next. Of interest, however, is the fact that this high score represents the changing party representation in the National Assembly almost exclusively for opposition parties. The ruling Democratic Justice Party lost only an average of 8.5% of its seats from 1981 to 1992.4 We see the Pedersen scores for the case universe represented in Table 1. What do these volatility scores tell us about regularized patterns of party competition and system institutionalization? Although high in Taiwan, they reflect increasingly healthy political competition between the persistent party actors and a move away from a single-party dominant system toward a multiparty system. In South Korea, the Pedersen score indicates a system in flux revolving around constantly changing political parties. Competition in Korea has been based on "winning it all" as opposed to securing a percentage of support from one election to the next. 4. This analysis only accounts for elected seats, not those assigned by a proportional representation allocation. 102 COMPARATIVE POLITICAL STUDIES / February 2001 Table 2 Differences Between Presidential and Legislative Voting Number of Latest Mean Country Time Span Elections Difference (%) Difference (%) Uruguay 1971-1989 3 0.6 0.6 Costa Rica 1970-1990 6 4.0 3.3 Argentina 1983-1989 2 2.9 3.3 Colombia 1974-1990 5 13.6 6.8 Mexico 1988 1 6.9 6.9 Venezuela 1973-1993 5 8.5 7.8 Taiwan 1996 1 8.3 8.3 Bolivia 1979-1993 5 12.0 9.8 South Korea 1987-1992 2 8.9 10.05 Peru 1980-1990 3 12.8 11.7 Chile 1989-1993 2 14.0 14.7 Paraguay 1993 1 14.8 14.8 Ecuador 1984-1992 3 48.8 31.3 Source: Latin American cases from Mainwaring and Scully (1995). The second criterion proposes that parties develop "stable roots in society . . . [i.e.] linkages between parties, citizens, and organized interests" (Mainwaring & Scully, 1995, p. 9). The main indicator used by Mainwaring and Scully to measure mass identification with party labels is the difference between the percentage of a party's votes between its legislative and presidential candidates. These scores are presented in Table 2. Both South Korea and Taiwan are in the mid-range of cases for this indicator, illustrating average electorate party label voting. Korea's presidential elections in 1987 and 1992 were won by a simple plurality of 36% (Ro Tae-woo) and 41% (Kim Young-sam), respectively, yet reflected the degree of geographically fractionalized voter support among three primary candidates. In this instance, a low percentage difference between legislative and presidential voting only reflects the development of stable roots in particular regions by particular parties or, more appropriately, party leaders. Taiwan presents a more dispersed pattern of voting while also exhibiting an 8.3% difference between legislative and presidential votes in 1996. Survey data on mass attitudes and values toward the political system would be a more direct indicator for the entrenchment of the party system into societal norms yet are not always systematically available for this collection of cases.5 A survey conducted in South Korea by Shin (1994) measures sup- 5. For additional surveys on party identification in South Korea, see Lee and Glasure (1995). For additional surveys on Taiwan, see Hsieh and Niou (1996) and Hsieh, Niou, and Paolino (1996). Stockton / CROSS-NATIONAL ANALYSIS 103 port for a democratic party system. The results show that Koreans are still cautious in their support of the party system that exists in Korea. In their attitudes toward political party performance, only 27% felt that "parties serve the interest of the public" and only 57% felt that parties "provide opportunity to participate in politics" (Shin, 1994, p. 54). A 1985 survey asked respondents to mention the first thing that came to mind when hearing the word democracy. Only 1.1% of those responding answered "political parties; responsible politics," and around 19.2% responded with parliamentary politics as their first answer (Ahn, Kil, & Kim, 1988, p. 241). Thus, the hold that the party system has on society in general is fairly weak, and that of the parties themselves is weaker. The high degree of personalism and regionalism of the political parties creates barriers to the mass acceptance of the efficacy of the party system. The Social Change Surveys conducted in Taiwan in 1985, 1990, and 1991 asked, "Do you believe that many political parties lead to bad politics?" In 1985, 66% agreed that many parties were bad, whereas only 22% felt so in 1991 (Parish & Chang, 1996). A survey conducted by the Election Study Center immediately after the 1992 Legislative Yuan elections found an almost perfect relationship between those voters claiming a strong party identification and their votes. Those identifying themselves as "strong KMT" supporters voted for the KMT candidates (99%), while those identifying themselves as "strong DPP" supporters voted for the DPP candidates (100%). Thus voters, or at least a certain class of voters, identify themselves with a party and support that party electorally. Another indicator for the entrenchment of parties in society is that parties should have strong connections with organized interests such as labor, student groups, and neighborhood associations. This, of course, overstates the institutionalization of party systems within corporatist states where the ruling party has, over an extended period of time, controlled the creation, introduction, and behavior of social organizations. Such a tactic has been a hallmark of the KMT rule in both mainland China and Taiwan. A key distinction proposed by Mainwaring and Scully (1995) is that countries be divided according to linkages "under democratic conditions" (p. 12). Thus, the integration of social organizations into Taiwanese politics has been along the lines of Mexico and Paraguay. The relationship between societal sectors and parties has been different in South Korea due to the high mortality rate of political parties. Typically, the ruling party has courted the military and industrial elites while being far from entrenched in unions (illegal during martial law), student groups (waged street warfare), and civic associations. 104 COMPARATIVE POLITICAL STUDIES / February 2001 A final indicator attempts to reflect the proposition that a system of parties has become entrenched in society as the age of the parties holding 10% or more of legislative seats increases. Higher average ages of such parties should indicate some stability in the number and type of major parties within the system. After the 1992 National Assembly elections in South Korea, the oldest party with 10% or more of seats (Democratic Liberal Party) was only 2 years old.6 Even if considered to be an organizational continuation of the old Democratic Justice Party of Presidents Chun and Ro, its age of 12 would rank 27th in comparison to the 37 Latin American parties that qualify for this coding. The Democratic Party and United People's Party were the only other parties to win more than 10% of the seats, and these parties had been established in 1991 and 1992, respectively. Thus, the average age of parties in the Korean party system after the 1992 elections was a very young 1 year old. As mentioned above, it is important to remember that political parties in South Korea may be very young, but these parties are led by political personalities that have been active in Korean politics since before the Korean War. This average age ranks last in the aggregated sample. The next to last case is that of Brazil, where the average party age is 12 years. In comparison to South Korea, Taiwan presents a much higher average party age after the 1992 legislative Yuan elections. This, of course, is due to the KMT's age of 81 (at that time) skewing the average because the only other party to hold 10% of seats was the 6-year-old DPP. Therefore, the average age of the parties in Taiwan, as of 1992, was 43.5 years. This ranks Taiwan below Mexico (59) and Costa Rica (44) and above Chile (37) in the average age of parties holding 10% or more of lower chamber seats. The third criterion emphasizes that "citizens and organized interests must perceive that parties and elections are the means of determining who governs" (Mainwaring & Scully, 1995, p. 14). Essentially, the indicators for this are presented as being the presence of coups, regular and meaningful elections, and low degrees of personalism. Neither South Korea nor Taiwan has had coup attempts during the time period studied, whereas both have witnessed regular legislative elections (South Korea every 4 years and Taiwan every 3). South Korea had two presidential elections (1987 and 1992), whereas Taiwan conducted its first in the spring of 1996. These two cases differ dramatically in terms of the penetration of personalities into politics. South Korea's political system has been most often characterized as person- 6. The Democratic Liberal Party was created when the previous ruling party (Democratic Justice Party) merged with Kim Young Sam's Reunification Democratic Party and Kim Jong Phil's New Democratic Republican Party on January 22, 1990. Stockton / CROSS-NATIONAL ANALYSIS 105 Table 3 Party System Institutionalization in Latin America and Asia Aggregate Country Criterion 1 Criterion 2 Criterion 3 Score Uruguay 3 3.0 3.0 9 Costa Rica 2.5 3.0 3.0 8.5 Chile 2.5 3.0 3.0 8.5 Colombia 3 3 2.5 8.5 Taiwan 2.5 2.5 3 8 Venezuela 2.5 2.5 2.5 7.5 Argentina 2.0 2.5 2.5 7 Mexico 1.5 2.5 1.5 5.5 Paraguay 1.0 2.5 1.0 4.5 Bolivia 1.0 1.0 2.0 4 South Korea 1.0 1.0 2.0 4.0 Ecuador 1.0 1.0 2.0 4 Brazil 1.0 1.0 2.0 4 Peru 1.0 1.0 1.0 3 Source: Mainwaring and Scully (1995). ality based or driven (Cotton, 1989; Han, 1989; Kim, 1989). Taiwan's party system, on the other hand, does not suffer such extensive personalism and revolves more around party factions and localization of politics (Chao, 1992; Chen, 1996; Tien, 1989; Wu, 1995). The system criteria scores for Latin American cases and South Korea and Taiwan are presented in Table 3. Based on their own analysis and that of the contributors to their volume, the summary table is reproduced from Mainwaring and Scully (1995, p. 17). Each of the indicators for institutionalization is scored from 1 (low) to 3 (high). This table is reproduced with the additions of South Korea and Taiwan. Mainwaring and Scully originally aggregated the indicators to form a 4-point to a 12-point index for Latin American countries. Countries with high scores (9-12) are considered to have institutionalized party systems. These cases include Costa Rica, Chile, Uruguay, Venezuela, Colombia, and Argentina. Midrange polities (7-9) are labeled "hegemonic in transition" and include Mexico and Paraguay. Countries with the lowest scores (4-6) are inchoate party systems, or systems characterized by shifting factions, personalism, and regionalism. These cases include Bolivia, Ecuador, Brazil, and Peru. Because the fourth criterion has been deleted, the aggregate scores do not reflect the original scoring and a modification must be made. Despite the deletion, none of the cases lose or gain too much ground, and the ordering of 106 COMPARATIVE POLITICAL STUDIES / February 2001 cases on which the point cutoffs were originally based is used to identify the three types of systems. Now, scores of 7 to 9 are institutionalized, scores of 4.5 to 5.5 qualify as the hegemonic in transition category, and the remainder are considered to be inchoate. POLITICAL PARTY INSTITUTIONALIZATION IN SOUTH KOREA AND TAIWAN In his 1992 study, Dix posed the question of whether the democracies installed during the 1980s will have a better chance at consolidation than those installed during the last period of democratization 25 years earlier. Based on the premise that "parties are typically major vehicles for the recruitment of political leadership, the structuring of electoral choice and peaceable political competition, and the framing of policy alternatives" (Dix, 1992, p. 489), Dix used the indicators proposed by Huntington (1968) to compare levels of party institutionalization for the years 1969 and 1989. These indicators are adaptability, complexity, autonomy, and coherence. Using an identical approach is impossible in the case of Taiwan due to the fact that there has never been an earlier democratic or even multiparty period with which to draw temporal comparisons. In Taiwan, formalized political opposition was illegal until 1986. A case can be made that Korea experienced a mild form of free party-based politics during the interim government of President Ho Chong (1960), the Second Republic of Premier Chang Myon (1960-1961), and, despite its having been installed as the result of a coup, the first 10 years of the Third Republic under General Park Chung Hee from 1962 to 1971 (Han, 1985; Kihl, 1984). The time points for comparison in Korea are therefore 1971 and 1992. In Taiwan, the selection rule must be stretched to simply include the year in which political opposition was legalized (1986) and 1992, the last first legislative election of the 1990s. The first indicator is that of party adaptability, essentially a party's ability to adapt to a changing political environment. Parties equipped with this skill as a function of either leadership or structure have the most likely chances of survival and thus contribute an element of continuity to the political system. As such, the most likely indicator would be the average age of principal parties in a given year—that is, those with 10% or more of the vote during the legislative election of the period. As a result of the 1971 National Assembly elections, President Park's ruling Democratic Republican Party (DRP) won 47.7%, or 86 of the elected seats. The main opposition party, New Democratic Party (NDP), won 42.4%, or 65 of the elected seats. As of 1971, the Stockton / CROSS-NATIONAL ANALYSIS 107 DRP was 8 years old (founded in 1963). The NDP was organized in 1967 and thus was only 4 years old.7 By 1992, the NDP and DRP had been long abandoned and their members and supporters absorbed into other parties. The discussion above on system institutionalization presented the pertinent data on primary parties as of 1992, and the comparison of party ages shows a decline between the two periods. There can be no question that the KMT has dominated politics in Taiwan since 1949. Many studies on political change in Taiwan make the case that the KMT, faced with domestic and international political change, had to liberalize the political arena. Under the rule of Chiang Ching-kuo, Taiwan began to relax its "hard authoritarianism" (1972-1981), entered a stage of "soft authoritarianism" (1982-1985), and then saw the beginning of democratization (1986-1988) (e.g., Wu, 1995). In addition, the internal dynamics of the party began to liberalize as the highly structured and private Leninist-style party began to face dissent from within and public criticism for its closed decision making. This culminated in part with the KMT's adoption of a system of party primaries in 1992. The DPP has survived and prospered electorally since its inception, and it too has adapted to the changing political arena. One such adaptation has been to tone down its reliance on the independent Taiwan platform to appeal to more moderate voters, characterized by their support for maintaining the status quo of Taiwan's international position while also supporting multipartyism.8 A second indicator proposed by Dix is that of generational succession after the passing of party founders. The value of this is seen in the new generation leaders "having significantly different organizational experiences and demonstrating the party's ability to survive the inevitable inter-generational conflicts" (Dix, 1992, p. 491). In Korea, only the Democratic Justice Party can be said to have gone through a semblance of a generational succession, with Ro Tae-woo's appointment as party leader by Chun Doo-hwan prior to the presidential election in 1988 (Han, 1988, 1989). Ro was younger and, although dependent on the traditional support base of the military, was considerably more amenable to compromise with Korea's leading opposition leaders, Kim Dae-jung and Kim Young-sam (Cotton, 1989; Han, 1988). By 7. Information on party founding and ages is taken from Fukui (1985), Kihl (1984), and Han (1985). 8. My conclusion is drawn from interviews conducted among Democratic Progressive Party members andrelated independence activist groups in Taipei, Taiwan, during the springs of 1992 and 1996. This tactic has prompted extensive intraparty debate and, ultimately, a party splinter in the spring of 1997. 108 COMPARATIVE POLITICAL STUDIES / February 2001 1992 (and up to the present), the major political parties and competition still revolved around the rivalry between the two Kims. The issue of leadership succession is clearly captured in the case of Taiwan's ruling KMT. From the time of its retreat to Taiwan after losing the 1945-1949 civil war to the communists, the KMT and Taiwan were governed by a paramount leader in Chiang Kai-shek. On Chiang's resignation from active rule, a dynastic succession occurred with the appointment of his son, Chiang Ching-kuo, to the presidency. As mentioned above, differences between father and son were reflected in the manner in which the party conducted its authoritarian rule. A clear succession took place with the appointment of Taiwan-born Lee Teng-hui to the presidency. Lee began the process referred to as the "Taiwanization" of the KMT with his appointments of Taiwanese, as opposed to mainland Chinese, to important party and governmental positions. It has been under President Lee's rule that the KMT and Taiwan have experienced the most change. A final measure of adaptability is a party's ability to shift roles from one of opposition to rule. This has been the case in Korea only through the merger of the Democratic Justice Party and Kim Young-sam's Reunification Democratic Party in January 1990. Kim later won the party nomination in May 1992 and effectively became party leader (Hong, 1993). Taiwan has seen no such shift in party roles. Although the DPP continues to make gains in legislative, national assembly, and provincial assembly seats, its representation is still insufficient to gain majorities in any body. In addition, the disappointing margin of loss in the 1996 presidential election by candidate Peng Ming-min was a major setback in the move toward gaining national power. The second indicator of party institutionalization is complexity. The value of complexity is seen to lie in the reasoning that a party "having multiple functions, or a multifaceted organization, is more likely to endure, and to adjust to change, than a less complex organization" (Dix, 1992, p. 498). Complexity is measured by the level of personalism within the party and by the multiplication of organizational subunits (hierarchically and functionally) and the differentiation of separate types of subunits. In his analysis, Dix handles the second measure in a blanket fashion by stating that most Social and Christian Democratic parties, as well as some Marxist parties and Mexico's Partido Revolutionary Institutional, could be considered complex because of their reliance on "permanent party committees or cells in most communities and a variety of functional sections" (Dix, 1992, p. 499). The KMT, with its Leninist structure, is similar in many respects to Mexico's Partido Revolutionary Institutional. The organizational structure of the DPP has also adopted a similar structure. As alluded to Stockton / CROSS-NATIONAL ANALYSIS 109 earlier, the KMT is undergoing change, but whether it is toward or away from an ideal level of complexity is unclear. Perhaps changes are more related to matters of efficiency. Little research has been conducted on the internal organization of South Korea's political parties, with most concentrated on election-seeking behavior. Thus, arriving at a score for the level of complexity based on the second measure is difficult. Making a judgment as to decline or improvement is very subjective, but based on the proposition that complexity in organization promotes persistence, levels of complexity in parties in South Korea in 1971 and 1992 must not have varied a great deal. The extent of personalism over politics has been mentioned in the discussion above. Although seemingly repetitive, personalism may be viewed as having different importance when considering its impact on a party as opposed to a system. Whereas personalism harms the system by forcing voters to put their support of the party system in the fortunes of individuals, personalism harms the party itself by placing the future of the party in the fortunes of individuals. The effect that personalism has on the party system in Korea is reflected by regional patterns of voting; that is, it has been shown to hold across all elections that candidates for national office receive extreme pluralities from their home provinces, whereas candidates from outside the province receive minimal voter support. The effect that personalism has on the party itself is that parties in Korea do not survive the departure of their primary founder(s) past one election (Korea Annual, 1994). Has personalism worsened from 1972 to 1992? Based on electoral returns for the two periods, regionalism has remained an important factor in elections and party life spans have not improved, so a loose assertion would be for no change. In contrast, recent party splits in the KMT and DPP illustrate a growing conflict of personalities and appeal to personalism as a basis for support. However, as of 1992, personalism could not be considered to be on the rise within the KMT despite a hotly contested struggle for power between President Lee and Premier Hau Pei-tsun (representing the more conservative party elements) for control over the party' s direction pertaining to political liberalization. The result was the formalization of two major factions: the mainstream faction led by Lee and the nonmainstream faction led by Hau (Wu, 1995). The third indicator for institutionalization is autonomy, the "extent to which political parties exist independent of other social groupings" (Dix, 1992, p. 500). In Huntington's (1968) words, "Political organization that is the instrument of a social group—family, clan, class— lacks autonomy and institutionalization" (p. 20). This indicator raises a potential contradiction when compared to the necessity for parties to become entrenched in society, 110 COMPARATIVE POLITICAL STUDIES / February 2001 as discussed in the above section on system institutionalization. Yet, where parties encompassing social groups are seen as beneficial to the system, it is conceivable to view social groups controlling parties as negative within the context of cases examined in this article.9 Although student movements and labor, for example, have been closely entwined with political development and political parties in South Korea, they have been unable to exert significant degrees of influence over inner-party processes. They have, however, provided a significant impact on the timing and extent of regime concession over liberalization (Billet, 1992). The influence of the military on the ruling parties was quite heavy until the "civilianization" of Korean politics with the election of Kim Young-sam in 1992. In that sense, one could argue that parties, or at least the ruling party, have become more autonomous. The degree of political control of the KMT over all societal groups has prevented nonmilitary groups from exerting any degree of control over party processes. However, ethnic divisions have played a major role in Taiwan since the arrival of the KMT in 1949. Up to Lee's succession, mainlander interests overwhelmingly dominated the direction the polity took in terms of political, social, and economic development. The creation of the DPP was in part a response to the exclusion of native Taiwanese from politics. A result of President Lee's "Taiwanization" of the KMT has been a slight decrease in this tension and a move away from a perception of the KMT as purely a Chinese party. The final measure of party institutionalization presented by Dix (1992), coherence, is excluded from this analysis because of its more appropriate application to the study of the relationship between parties and the electoral system, not coherence within parties. Table 4 presents the results of Dix's survey with the additions of South Korea and Taiwan. PARTY SYSTEM INSTITUTIONALIZATION AND DEMOCRACY Typologies contribute an ordering to our thinking about classes of phenomena, and their value cannot be underestimated. However, another important element of typologizing is the expectation of certain implications arising from the ordering. The most obvious implication arising from a typology of political parties and party systems is the assumption that these two concepts have something to do with improving the quality of democracy. This link is a 9. However, this line of reasoning borders on contradicting much of the work done on corporatism and consensual government among the European social democracies. Stockton / CROSS-NATIONAL ANALYSIS 111 Table 4 Changes in Political Party Institutionalization Between the 1960s and 1980s Adaptability j, . , Complexity Party General _1 1 Age Succession Turnover Personality Structure Autonomy Change Costa Rica + + 0 + + 0 4 Chile - 0 0 0 0 -2 Uruguay - - 0 0 0 -3 Venezuela + + + + 0 0 4 Colombia + 0 0 + 0 0 2 Argentina + + 0 + + + 5 Mexico + 0 0 0 0 0 Paraguay + 0 0 + 0 0 2 Bolivia + + 0 - + 1 Ecuador 0 + + 0 0 1 Brazil 0 - 0 0 0 -2 Peru + + + 0 0 0 3 South Korea 0 + 0 0 0 0 Taiwan + + 0 0 0 + 3 Source: Latin American cases from Dix (1992, p. 506). constant thread throughout any discussion of institutionalization. Therefore, the rest of this article will attempt to provide a rough empirical analysis of the relationship between parties and party systems and quality of democracy. All of the countries are assigned ranks based on their relative scores on party and party system institutionalization as well as ranks based on their relative Freedom House scores in the final year in which data are presented. Compiled annually by Freedom House, the freedom scores reflect levels of political and civil freedom on a 14-point scale, with lower scores reflecting more freedom. These scores are presented in Table 5. The best line of fit is then assessed to identify the type of relationship that explains the most variance between cases. From Figure 1, we see moderate support for the curvilinear relationship between system institutionalization and democracy rankings for the sample. The strongest line of fit is cubic, which indicates that cases in the middle range of democratic quality rankings were fairly well distributed along the x axis, thereby creating a dip in the line. Thus, countries with the highest system institutionalization rankings (1-3) also reflect, as a group, a somewhat linear relationship with "democraticness" (Uruguay, Costa Rica, Chile, Colombia, and Venezuela). Countries in the midrange of system institutionalization rankings (4-7) exhibit the greatest range in freedom rankings. Perhaps this is a reflection of 112 COMPARATIVE POLITICAL STUDIES / February 2001 Table 5 Relative Party, System, and Democracy Ranks Freedom Freedom System System Party Party System Country Score (year) Rank Score Rank Score Rank Average Costa Rica 2 (1990) 1 8.5 2 5 2 2 Uruguay 3 (1990) 2 9 1 -4 10 5.5 Chile 4 (1993) 3 8.5 2 -2 9 5.5 South Korea 4 (1992) 3 5 7 0 7 7 Bolivia 5 (1993) 4 4 9 -1 8 8.5 Ecuador 5 (1992) 4 4 9 1 6 7.5 Brazil 5 (1990) 4 4 9 0 7 8 Argentina 5 (1991) 4 7 4 6 1 2.5 Colombia 6 (1990) 5 8.5 2 3 4 3 Venezuela 6 (1993) 5 7.5 3 5 2 2.5 Mexico 7 (1991) 6 5.5 6 1 6 6 Paraguay 8 (1990) 7 3 10 4 3 6.5 Peru 8 (1990) 7 3 10 4 3 6.5 Taiwan 8 (1992) 7 6 5 3 4 4.5 Source: Party scores from Dix (1992); system scores from Mainwaring and Scully (1995); freedom scores from Freedom in the World (1991, 1992, 1993, 1994). Note: Numbers in parentheses are the years for which the freedom scores are recorded. the brinkmanship brought on by the uncertainty that pervades systems at the halfway point in the development of their party systems. Both of the cases introduced in this study fall within this midrange. It is interesting that whereas Taiwan ranks ahead of South Korea in system institutionalization, South Korea's freedom ranking is more than twice as good as Taiwan's. This can be explained in part by the fact that the system cannot be so easily separated from the party in Taiwan due to the KMT's prolonged role as the system caretaker. On the other hand, the relative lack of institutionalization of South Korea's party system has perhaps allowed more flexibility in the entry and exit of political actors while still constricting political competition to minimum norms. Finally, cases with the lowest rankings (8-10) also reflect a within-group curve. The last-ranked case for institutionalization is also the least democratic of the sample (Peru). The nature of the curve in Figure 1 suggests that systems pass through threshold points that have a detrimental effect on democracy just prior to passing through, followed by a marked improvement in democracy just after passing through. Throughout the sample, this is repeated three times. Perhaps countries pass through more than one threshold toward democracy when considering the effect of party system institutionalization instead of the more common view that there exists only a single threshold that when passed will lead to the persistence of democracy. Stockton / CROSS-NATIONAL ANALYSIS 113 System Ranking R2 = .432 Figure 1. Political party institutionalization and democracy. Source: Party rankings based on Dix (1992); freedom rankings based on Freedom in the World (1991, 1992, 1993, 1994). Note: TAI = Taiwan; PER = Peru; MEX = Mexico; PRG = Paraguay; COL = Columbia; VEN = Venezuela; ARG = Argentina; CHI = Chile; SKO = South Korea; BOL = Bolivia; BRA = Brazil; URG = Uruguay; ECD = Ecuador; CR = Costa Rica. POLITICAL PARTY INSTITUTIONALIZATION AND DEMOCRACY When examining Figure 2, we find a similarly successful cubic line of fit between party institutionalization and freedom rankings. In contrast to the double-peaked fit discussed above, the relationship appears to be single peaked. Although suggesting again that midranges of party ranking have a more disbursed effect on democracy, there appears to be moderate support for a negative relationship between the two variables. Three of the four most democratically ranked cases (1-3) rank in the bottom half of party rankings. If the cases are dichotomized into halves along both axes, we find that six of the seven cases ranked in the top half of party rankings fall in the bottom half of democracy rankings. 114 COMPARATIVE POLITICAL STUDIES / February 2001 1 Pi S o 7- PER(90) TAI(92) ■ PRG(93) MEX(91) VEN(93^C ■ / ■ ARG(9t? ECD(9X BOL(93) ■ / BRA(9a) CHI(93) SKO(92) \. URG(90) / CR(90) / ■ 10 12 Party Ranking R2 = .446 Figure 2. Political party institutionalization and democracy. Source: System rankings based on Mainwaring and Scully (1995); freedom rankings based on Freedom in the World (1991, 1992, 1993, 1994). Note: TAI = Taiwan; PER = Peru; MEX = Mexico; PRG = Paraguay; COL = Columbia; VEN = Venezuela; ARG = Argentina; CHI = Chile; SKO = South Korea; BOL = Bolivia; BRA = Brazil; URG = Uruguay; ECD = Ecuador; CR = Costa Rica. It is possible that this seemingly negative relationship is an artifact of the confusion that the values of change and stability provoke in the study of political development. Democratic change is most often approached with the idea that democratic governments should eventually converge toward the Anglo-European models. However, although this requires a great deal of change within young democracies, it calls for relative static within the consolidated group of democracies. Thus, when we examine cases within one group or another, different values are applied to change and stability. This is reflected in Figure 2, in which very low levels of party change (regardless of direction) are correlated with higher relative levels of democracy, whereas cases with the most change toward party institutionalization (with the exception of Costa Rica) exhibit lower freedom rankings. This dilemma is aggravated by the static nature of the analysis because we are not informed of the Stockton / CROSS-NATIONAL ANALYSIS 115 point on the democratic scale from which each case has had to traverse. Can it be argued that cases illustrating simultaneously low progress toward party institutionalization and high levels of democracy simply did not have too far to travel toward improving democracy and vice versa? Again, there is no clear, generalizable answer. South Korea and Chile emerged from extreme military authoritarianism, whereas Costa Rica, ranked second in party change, has not had comparative dangers on the road to its current democracy. PARTIES, SYSTEMS, AND DEMOCRACY Mainwaring and Scully (1995) conclude that countries that have developed institutionalized party systems are more likely to consolidate democracy. They define consolidation as a condition in which a party-based democracy will survive into the near future because (a) political interaction between elites and masses follows established and somewhat predictable rules of party competition, (b) political parties develop stable roots within society, (c) all actors perceive party-based elections to be the crucial factor in determining who governs, and (d) party organizations are firmly structured, rou-tinized, and territorially comprehensive (Mainwaring & Scully, 1995). Dix (1992) concludes that Latin American political parties at the beginning of the 1990s were more institutionalized than they were during their first democratic episodes in the 1960s. Therefore, because "[parties] are almost certainly necessary for the long-term consolidation of broad-based representative government" (Dix, 1992, pp. 489,508), chances for democratic consolidation are more likely. The dependent variable to this point has been democratic quality, for which the Freedom House scores are clearly an appropriate measure. Whereas quality deals with degree of the democratic entrenchment, consolidation places more of an emphasis on the temporal dimension of survival into the indefinite future. Nevertheless, any conceptualization of consolidation has to include a qualitative element, such as those that are captured by Freedom House. Freedom House scores do not inform us if a state is consolidated, but one can arrive at general predictions about the progress toward consolidation based on the quality of political and civil freedoms. Thus, when hypothesizing that increasing levels of both party and party system institution- alization correlate with increasing levels of democratic quality, we can capture some elements of consolidation. Figure 3 illustrates the relationship between the average of the rankings for party and system institutionalization and democratic rankings for South Korea, Taiwan, and 12 Latin American polities. 116 COMPARATIVE POLITICAL STUDIES / February 2001 o ■ TAI(92) PER(90) 1- ■ ■ MEX(91) PRG(93) 6- ■ ■ VEN(93) COt 5- ■ ■/ ARG(9/f) \. ECD(92) BOL(93K 4- ■ / —■__-m*^ m CHI(93) SKO(92) BRA(90) 3- ■ ■ URG(90) 2- CR(90) ■ 1 ■ / ■ 0 123456789 Averaged Party and System Rankings R2 = .229 Figure 3. Political party and system institutionalization rankings and democracy. Source: Party rankings based on Dix (1992); system rankings based on Mainwaring and Scully (1995); freedom rankings based on Freedom in the World (1991, 1992, 1993, 1994). Note: TAI = Taiwan; PER = Peru; MEX = Mexico; PRG = Paraguay; COL = Columbia; VEN = Venezuela; ARG = Argentina; CHI = Chile; SKO = South Korea; BOL = Bolivia; BRA = Brazil; URG = Uruguay; ECD = Ecuador; CR = Costa Rica. Although the best line of fit is still cubic (double peaked), the .R-squared is dramatically smaller in comparison to the tests based on separate party and system institutionalization rankings. The cases are fairly well clustered in the four x-y quadrants, with the exception of the lower left in which only Costa Rica is located. However, half of the total cases (7) are located in the region that would suggest a vague linear, yet negative, relationship between overall system institutionalization and democracy; that is, as system institutionalization deteriorates, democracy improves. If we trace a path backward from upper to lower right and then from upper to lower left, the progress toward the best level of democracy (Costa Rica) can be highlighted in an analysis similar to that presented in the discussion on the correlation between party system institutionalization and democracy. Stockton / CROSS-NATIONAL ANALYSIS 117 CONCLUSION From this analysis the traditional conclusion can be repeated that party and party system institutionalization are both important to democracy, but a caveat can be added. There appears to be no single barrier to overcome in pursuit of the party and system institutionalization, which are seen to complement democracy. Getting it right the first time is a high expectation when contemporary democratic transitions are occurring in an increasingly complex time. Repeatedly, we find that the path toward greater democracy is one of progress followed by recession. We can expect not only that more than one threshold of system institutionalization toward democracy will exist but also that changes will be sharp in the midrange of cases traveling from least democratic to most democratic. This article illustrates that the path toward democratic deepening is not a linear one in which party and system institutionalization follow hand in hand toward consolidation. Instead, new democracies in Latin America have exhibited the potential for at least one more "bump in the road" on the way to consolidation. Perhaps democratic collapse was an essential part of democratic deepening within the Third Wave countries of Latin America. Through learning, political actors can alter behaviors that proved detrimental to democratic governance during prior periods of democracy. What does this mean for democratic quality and consolidation in South Korea, Taiwan, and Asia in general? Should we anticipate or even expect a return to authoritarianism as a prerequisite for democracy to become more entrenched in the future, or would such relapses merely be hiccups in democratization? Both cases have steadily improved their respective democracies in the years since 1992, and an authoritarian relapse seems highly unlikely, especially given the great opportunity for authoritarian relapse due to the persisting Asian economic crisis.10 As stated earlier, the next step in analysis should be extending the case size to allow for regression analysis as well as applying a more dynamic methodology. This study has intentionally ignored numerous potential variables in pursuit of a simple analysis of the bivariate correlation between party and party system institutionalization and democracy at a single point in time. Although the results can be challenged on a number of theoretical and meth- 10. Indeed, one may argue that Kim Dae Jung's victory in the 1997 presidential elections in South Korea was a final step in that country' s consolidation process. Also, the March 2000 election of Chen Shui-bien marked the first alternation of power in the 50 years that the Republic of China has been on Taiwan, indicating that consolidation has taken a firm root. 118 COMPARATIVE POLITICAL STUDIES / February 2001 odological points, this article has presented an easily understood and visually interpretable analysis of the relationships between two crucial variables in the construction of democracy. Perhaps more questions were raised than answered; however, the issues addressed here hold the potential for providing a fruitful continuation of exploration on the topic of institutionalization and democracy. REFERENCES Ahn, B. M., Kil, S.H.,& Kim, K. W. (1988). Elections in Korea. Seoul: Seoul Computer Press. Billet, B. (1992). The history and role of student activism in the Republic of Korea: The politics of contestation and conflict resolution in fledgling democracy. Asian Profile, 20, 23-34. Chao, Y. M. (1992). Local politics on Taiwan: Continuity and change. In D. Simon & M.YM. Kau (Eds.), Taiwan: Beyond the economic miracle (pp. 43-68). New York: M.E. Sharpe. Chen, Ming-tong. (1996). Local factions and elections in Taiwan's democratization. In H. M. Tien (Ed.), Taiwan's electoral politics and democratic transition: Riding the third wave (pp. 174-192). New York: M.E. Sharpe. Cotton, J. (1989). From authoritarianism to democracy in South Korea. Political Studies, 37, 244-259. Diamond, L. (1997). Introduction: In search of consolidation. In L. Diamond, M. F. Plattner, Y. H. Chu, & H. M. Tien (Eds.), Consolidating the third wave democracies (pp. xii-xlvii). Baltimore, MD: Johns Hopkins University Press. Dix, R. H. (1992). Democratization and the institutionalization of Latin American political parties. Comparative Political Studies, 24, 488-511. Duverger, M. (1954). Political parties. New York: John Wiley. Freedom in the world. (1991). New York: Freedom House. Freedom in the world. (1992). New York: Freedom House. Freedom in the world. (1993). New York: Freedom House. Freedom in the world. (1994). New York: Freedom House. Fukui, H. (Ed.). (1985). Political parties of Asia and the Pacific. Westport, CT: Greenwood. Han, S. J. (1985). Democracy in South Korea: 1961-1985. Paper presented at the Conference on Democracy in Developing Nations, Stanford, CA. Han, S. J. (1988). South Korea in 1987: The politics of democratization. Asian Survey, 25(1), 52-61. Han, S.J. (1989). South Korea in 1988: Arevolutionin the making. Asran Survey, 29(1), 29-38. Hong, Y. L. (1993). South Korea in 1992: A turning point in democratization. Asian Survey, 33(3), 32-42. Hsieh, F. S., & Niou, E.M.S. (1996). Issue voting in the Republic of China on Taiwan's 1992 Legislative Yuan election. International Political Science Review, 17, 13-27. Hsieh, F.S., Niou, E.M.S., & Paolino, P. (1996). Strategic voting in the 1994 Taipei City mayoral election (Working papers in Taiwan studies). Durham, NC: American Political Science Association Conference Group on Taiwan Studies. Huntington, S. (1968). Political order in changing societies. New Haven, CT: Yale University Press. Huntington, S. (1991). The third wave: Democratization in the late twentieth century. Norman: University of Oklahoma Press. Stockton / CROSS-NATIONAL ANALYSIS 119 Kihl, Y. W. (1984). Politics and policies in divided Korea: Regimes in contest. Boulder, CO: Westview. Kim, H. N. (1989). The 1988 parliamentary election in South Korea. Asian Survey, 29,480-495. Korea annual 1994. (1994). Seoul: Hapdong News Agency. Lee, A. R., &Glasure, Y. U. (1995). Party identifiers in South Korea. Asian Survey, 25,367-376. Lijphart, A. (1984). Democracies: Patterns of majoritarian and consensus government in twenty-one countries. New Haven, CT: Yale University Press. Lipset, S. M. (1981). Political man: The social bases of politics. Baltimore: John Hopkins University Press. Mainwaring, S. (1992). Transitions to democracy and democratic consolidation: Theoretical and comparative issues. In S. Mainwaring, G. O'Donnell, & J. Valenzuela (Eds.), Issues in democratic consolidation (pp. 3-45). Notre Dame, MI: University of Notre Dame Press. Mainwaring, S., & Scully, T. R. (1995). Building democratic institutions: Party systems in Latin America. Stanford, CA: Stanford University Press. Parish, W. L., & Chang, C. C. (1996). Political values in Taiwan: Sources of change and constancy. In H. M. Tien (Ed.), Taiwan's electoral politics and democratic transition: Riding the third wave (pp. 27-41). New York: M.E. Sharpe. Powell, G. B. (1982). Contemporary democracies: Participation, stability, and violence. Cambridge, MA: Harvard University Press. Sartori, G. (1976). Parties and party systems. New York: Cambridge University Press. Sartori, G. (1987). The theory of democracy revisited. Chatham: Chatham House. Shin, D. C. (1994). Political parties and democratization in Korea: The mass public and the democratic consolidation of political parties. Paper presented at the International Political Science Association 155th World Congress, Berlin. Tien, H. M. (1989). The great transition: Political and social change in the Republic of China. Taipei, Taiwan: SMC Publishing. Toka, G. (1997). Political parties in East Central Europe. InL. Diamond, M. F. Plattner, Y. H. Chu, & H. M. Tien (Eds.), Consolidating the third wave democracies (pp. 93-134). Baltimore: Johns Hopkins University Press. Wu, J. J. (1995). Taiwan's democratization: Forces behind the new momentum. New York: Oxford University Press. Zhang, B. (1994). Corporatism, totalitarianism, and transitions to democracy. Comparative Political Studies, 27, 108-136. Hans Stockton is a lecturer at the Center for International Studies at the University of St. Thomas in Houston, Texas, and is also a visiting lecturer at the Department of Political Science at the University of Houston's central campus. His research focuses on the role of political parties in democratic transition and consolidation and the political economy of democratic consolidation. He has published in the journal Asian Perspectives.