Convergence, Polarization, or Both? John Fuh-sheng Hsieh University of South Carolina

Convergence, Polarization, or Both? John Fuh-sheng Hsieh University of South Carolina Hsieh@sc.edu Introduction There are heated debates in the United States about whether American society is becoming more polarized. 1 As one listens to the radio, watches cable news, or reads about congressional bickering, partisan furor can hardly go unnoticed. However, for those who know something about Taiwan politics, these are familiar scenes almost from the outset of democratic transition in the late 198s and early 199s: the uncompromising tone from the underground radio programs, the unapologetic pro-blue or pro-green rhetoric on cable TV talk shows, and the never-ending quarrels or even fist fights in the legislatures. These are all too familiar to people living on the island. As will be made clear later, a major source for polarization in Taiwan is the divergent views on the national identity issue: Some prefer Taiwan independence, others favor unification with China, and still others stand somewhere in between (or the so-called status quo). Yet, oddly, if we take a look at survey data, there seems to be convergence of views on this issue with a majority or close to a majority of the respondents support the status quo (depending, of course, upon how it is defined). Moreover, as evidenced by the 216 presidential and parliamentary elections, both major political parties, the Kuomintang (Nationalist Party or KMT) and the Democratic Progressive Party (DPP), talked about the maintenance of the status quo. It seems 1 See, for example, Morris P. Fiorina with Samuel J. Abrams and Jeremy C. Pope, Cultural War? The Myth of a Polarized America, 3 rd ed. (Boston: Longman, 211), and Alan I. Abramowitz, The Disappearing Center: Engaged Citizens, Polarization, and American Democracy (New Haven, Conn.: Yale University Press, 21). 1

that there is a trend for both the elites and the masses to move toward the middle. So, is there a trend toward convergence or does Taiwan remain polarized? Or is it possible that polarization and convergence can coexist in some way? This is the topic I am going to address in this paper with the help of the data provided by the Taiwan Elections and Democratization Studies (TEDS) and Taiwan National Security (TNS) surveys. 2 In the next section, I will review the salient cleavages in Taiwan, followed by a section on polarization or lack thereof on the island. And then, I will explore its political impact. The last section concludes. Salient Cleavages in Taiwan In any society, conflicts of interests and/or values are inevitable. In a market economy, for instance, workers would like to have higher wages so that they can better feed their families, but at the same time, business owners may want to cut down the wages so as to lower their production costs and make their products cheaper in the market. This is inherent in the division of labor in a market economy, and is not surprising at all. In a similar vein, we can imagine the emergence of a variety of other issues which may divide the society into different groupings with positions on one side or the other. Some of these issues may become politicized, and play an important role in the political process. 2 The TEDS surveys are conducted by a consortium of political scientists from various academic institutions in Taiwan and sponsored by Taiwan s Ministry of Science and Technology (formerly National Science Council). Professor Chi Huang has been the coordinator of these surveys since 2. For the three surveys analyzed in this paper, the 24LB, 28P and 212 surveys, the principal investigators are Professors I-chou Liu, Ching-hsin Yu and Yun-han Chu, respectively. The TNS surveys were conducted by the Election Study Center of National Chengchi University, and the principal investigator is Professor Emerson Niou. 2

In their seminal work, Party Systems and Voter Alignments, Seymour Lipset and Stein Rokkan argued that the major cleavages in Western democracies were, to a large extent, shaped by two historical events: National Revolution and Industrial Revolution. 3 The former led to the emergence of state-church and center-periphery divisions while the latter to the advent of ownerworker (or class) and rural-urban differences as exemplified by the formation of Christian democratic parties, various types of regional parties, socialist parties, agrarian parties, and the like. Many of these cleavages remain salient almost half a century after the publication of their work. Indeed, to a large extent, class and religion continue to be prevalent in the Western world. 4 Undoubtedly, new issues may emerge and become politically salient across time and space. One example is post-materialism characterized by such values as environmentalism, human rights, and women s rights, giving rise to, for instance, green or ecological parties in many societies. 5 Of course, not all cleavages are created equal. Some are more salient; some are more enduring; some may entail a high degree of emotional reactions; and from time to time, several cleavages may cluster together and reinforce each other. More often than not, different societies may come up with diverse constellations of cleavages. And what concerns us here is the type of cleavage or cleavages that emerge and buttress the party system in Taiwan. In particular, we will explore whether its party configuration is polarized to such an extent that people stand at the 3 Seymour M. Lipset and Stein Rokkan, Cleavage Structures, Party Systems and Voter Alignments: An Introduction, in Seymour M. Lipset and Stein Rokkan (eds.), Party Systems and Voter Alignments (New York: The Free Press, 1967), pp. 1-64. 4 Arend Lijphart, Patterns of Democracy: Government Forms and Performance in Thirty-Six Countries (New Haven, Conn.: Yale University Press, 1999), pp. 78-89. 5 See Ronald Inglehart, Silent Revolution: Changing Values and Political Styles among Western Publics (Princeton, N.J.: Princeton University Press, 1977), and Russell J. Dalton, Citizen Politics: Public Opinion and Political Parties in Advanced Industrial Democracies, 6 th ed. (Washington, D.C.: CQ Press, 214). 3

opposite poles of the cleavage(s) with very few people clustering around the center, and whether the issue(s) are so emotional that compromises become very difficult, if not impossible. In order to find out the salient cleavages in Taiwan s party politics, a set of four questions had been included in academic surveys since 1992, each tapping into a potential cleavage in this new democracy. One of the questions is related to class or the traditional left-right issue. The respondents were asked to choose between a score standing for more social welfare and 1 for economic growth (or less social welfare). Another question deals with environmentalism or the new left-right issue. The respondents were similarly asked to pick a score between and 1 with representing environmental protection and 1 economic development. What else are potentially salient in Taiwan politics in addition to the old and new leftright issues found in many Western democracies? For anyone who knows something about Taiwan, he/she cannot fail to notice the importance of the national identity issue there. This issue refers to a voter s position on the future political association between Taiwan and mainland China: unification of Taiwan and Mainland China; an independent Taiwan to be separated from China permanently; or somewhere in between. To get a snapshot of voters positions on the issue, the respondents were asked to choose between Taiwan independence (a score ) and unification (a score 1). Besides, given the fact that Taiwan embarked on democratization not long ago, a fourth question was added to the list: a tradeoff between (reform) and 1 (stability). This was the talking point many people associated with the old authoritarian regime had maintained during or before democratic transition much like what the Communist regime in China is doing now. This set of questions had been included in a number of surveys conducted by Opinion Research Taiwan, Election Study Center of National Chengchi University, and several TEDS 4

surveys. Table 1 shows the logistic regression results coming out of two TEDS surveys: 24LB (conducted in 25) and 28P. Unfortunately, the 28P survey was the last TEDS survey to include all these four questions with the same format. In Table 1, I also run a truncated model with the 212 TEDS survey to show the trend. [Table 1 about here] The dependent variable in the regression analysis shown in Table 1 is voters support for (1) the Pan-KMT or Pan-Blue camp, referring to the long ruling KMT and its allies, as against (2) the Pan-DPP or Pan-Green camp, including the DPP, the ruling party from 2 to 28 and the winner of the 216 presidential and parliamentary elections, along with other political forces closely allied with it. This is a binary variable with standing for support for the Pan-DPP camp and 1 for the Pan-KMT camp. As for the independent variables, the four potential cleavages noted above are included in the analysis of the 25 and 28 data, but only the welfare and national identity issues are included for 212. 6 There are also clean government issue, 7 ethnicity, 8 interaction terms between ethnicity and national identity, 9 and such demographic 6 The question on social welfare is not ideal. Instead of asking the respondents to make a tradeoff between social welfare and economic growth, the question in the survey looks into a choice between the maintenance of the current level of social welfare (a score ) and the expansion of it (a score 1). But, unfortunately, this is all we have. And to conform to the format of other cleavage questions, I change the scores so that stands for more welfare a position normally regarded as on the left and 1 for less welfare a position on the right. 7 The attitude on the clean government issue is coded from voters responses to the question about whether corruption is one of the three (in 25) or two (in 28) most important issues. However, in the 212 TEDS survey, the respondents were asked only about the most important issue instead of two or three most important ones as in earlier surveys and very few people actually chose corruption or other related issues, so this variable has to be deleted from the analysis of the 212 survey. 8 There are four major ethnic groups in Taiwan: Minnan Taiwanese, Hakka Taiwanese, Mainlanders, and aborigines. As of December 215, the aborigines account for only 2.3 percent of the total population. See Ministry of the Interior, Monthly Bulletin of Interior Statistics, 5

variables as gender, age, and education. As can be seen from the table, among the four potential cleavages, national identity and stability issues are consistently significant except that there is no question on reform vs. stability in 212. The welfare and environmental issues are significant occasionally. Generally, those who are more pro-unification, less pro-environment, and more concerned about stability are more likely to support the Pan-KMT camp. Support for less welfare seems to negatively affect the support for the Pan-KMT camp, but it is not always significant and the signs are different in various years. Moreover, since these variables are measured on the same scale, their coefficients can be compared. Obviously, the national identity issue overwhelms all others. And this is indeed the dominant issue shaping Taiwan s party politics. 1 In addition, clean government is significant in 28, and interestingly, those who believe that corruption is an important issue lean toward the Pan-KMT camp. The sign is different from a lot of earlier surveys (including the 25 figure shown in the table, which is nonetheless insignificant), reflecting probably the scandals involving President Chen Shui-bian, the leader of the Pan-DPP camp, and his associates during his term in office. It is also noteworthy that the ethnicity variables are mostly insignificant, a fact signifying that national identity may, in some way, diminish the salience of ethnicity as a factor affecting Taiwanese voters partisan attachments (see the following discussion). Gender and age matter somewhat. And education is consistently significant: The better educated tend to support the Pan-KMT camp. <http://sowf.moi.gov.tw/stat/month/elist.htm >. So it is not included in the analysis. And with three main ethnic groups, two dummy variables are constructed. 9 These interaction terms try to capture the possibility that national identity is formed in different ways for different ethnic groups. 1 Tse-min Lin, Yun-han Chu, and Melvin J. Hinich, Conflict Displacement and Regime Transition in Taiwan: A Spatial Analysis, World Politics, Vol. 48, No. 4 (July 1996), pp. 453-481, and John Fuhsheng Hsieh and Emerson M.S. Niou, Salient Issues in Taiwan's Electoral Politics, Electoral Studies, Vol. 15, No. 2 (May 1996), pp. 219-235. 6

Thus, among the potential cleavages, national identity is undoubtedly the dominant issue in shaping Taiwan s party politics. Stability, welfare and environmentalism may matter, but play a secondary role. Ethnicity, with a couple of exceptions, is generally not significant after controlling for other variables. And clean government issue may be important, but it reflects a judgment call, which may change as a consequence of voters' assessment of corruption associated with various political forces, very much like voters' evaluation of, say, economic performance under the government of a particular party. It is, therefore, not truly a long lasting cleavage that affects voters' relatively durable partisan attachments. A few words need to be said about ethnicity that has often been seen as an important dividing line between the Blue and Green camps. It is generally believed that Minnan Taiwanese are more likely to be in favor of the Pan-DPP camp while the mainlanders and, to a lesser extent, Hakka Taiwanese tend to support the Pan-KMT camp. It is partially born out by the regression results shown in Table 1. Table 2 exhibited the association between ethnicity and partisan attachments in a different way on the basis of the 212 TEDS data. As can be seen from the table, there is indeed some association between the two. Yet it should be pointed out that the association is not perfect. A lot of Minnan Taiwanese support the Pan-KMT camp, too. Some Hakka Taiwanese or even mainlanders may turn to the Pan-DPP camp. As a matter of fact, anecdotal evidence may suggest that the direct factor affecting a person s partisan attachment is not ethnicity per se; rather, it is his/her position on the national identity issue. For instance, if a Minnan Taiwanese indicates that she is for unification between Taiwan and the mainland, we know that she is likely to be a supporter of the Pan-KMT camp. Similarly, if a mainlander says he favors Taiwan independence, we can surmise that he is very likely a Pan-DPP supporter. It is the national 7

identity issue, not ethnicity per se, that directly affects an individual s partisanship. That is why as we control for the national identity issue, the significance of ethnicity mostly disappears as the regression models indicate. Table 3 displays this tendency with the 212 data. Here, to simplify the matter, I regroup the respondents into three categories based upon their responses to the 11- point scale question noted above: (1) pro-independence if they pick a score 4 or less on the national identity issue; (2) pro-status quo if they choose 5; and (3) pro-unification if they select 6 or above. It can be seen that Minnan Taiwanese are no more likely to support the Pan-DPP camp if they are in favor of the status quo or unification. Hakka Taiwanese tend to support the Pan- DPP camp if they prefer independence, but lean toward the Pan-KMT camp otherwise. To a certain extent, a similar trend can be detected for the mainlanders as well. [Tables 2 and 3 about here] How Polarized is Taiwan? For many people in Taiwan, polarization is obvious, and is a nightmare. If Taiwan politics is indeed severely polarized, it must have something to do with national identity since this is the dominant issue underpinning Taiwan s party politics. Then, how polarized is it? Is it polarized at the mass or the elite level? Let us begin with the masses. Figure 1 shows the distribution of voters positions on the 11-point scale on this particular issue in 212 (see All in the figure for the moment), and it exhibits a somewhat multimodal picture but with a big mode at the center. That is, it is not quite polarized given that a lot of respondents cluster in the middle. So, can we assert that voters in Taiwan are not that polarized 8

and the kind of polarization we witness lies mostly at the elite level? This sounds a bit puzzling since for many people living in Taiwan, they understand that they have to be careful when they, for instance, discuss politics with their friends or cab drivers, particularly during the campaign season. [Figure 1 about here] Actually, Figure 1 is somewhat misleading. The positions voters may take include not only their true or sincere beliefs but also their strategic calculations. For example, a lot of people may support the status quo not because they truly believe in it but because they sense that their most preferred position (unification, independence, or whatever) may not be feasible, so they end up choosing something less preferable. Thus, the distribution of voters sincere positions on this issue is not readily known by looking solely at Figure 1. In order to find out voters true intentions, I will turn to the 212 and 215 TNS surveys which are more up-to-date and include two sets of questions on the national identity issue. On one, respondents were asked whether they support (1) independence as soon as possible; (2) status quo for now, but independence later; (3) status quo permanently; (4) status quo for now, then moving toward independence or unification depending on the situation; (5) status quo for now, but unification later; and (6) unification as soon as possible. To simplify the discussion, I regroup the respondents into three categories: (1) The first two are recategorized as independence supporters; (2) the third and fourth as status quo supporters; and (3) the last two as unification supporters. Another set of questions tap into respondents conditional preferences. They were asked 9

about their support (1) for independence if peace can be maintained between China and Taiwan after Taiwan declares independence and (2) for unification if Taiwan and the mainland are similar in economic, social and political conditions. Obviously, these two preconditions are seen by most people to be unrealistic. Under the current circumstances, a war is likely if Taiwan declares independence, and of course, economic, social and political conditions between the two differ markedly. Thus, for those who are willing to select independence or unification even under the current unfavorable conditions in response to the question on the original national identity issue, I assume they are, to a certain degree, sincere in their responses. 11 Consequently, I will focus, in the following discussion, on those who support the status quo, and check whether they may shift to either independence or unification if conditions are favorable. 12 But before turning to the status quo supporters, let us take a look the distribution of all the respondents in the 212 and 215 TNS surveys on the two questions with preconditions. As Table 4a shows, in 212, 72.1 percent of all the respondents respond positively to the question about declaring independence if peace can be maintained, and 32.9 percent agree that Taiwan and the mainland should be unified if the conditions are right. Together, 21.3 percent of the respondents show flexibility by saying yes to both questions while 46.3 percent of the respondents would support independence but not unification even if the two sides are similar in economic, social and political conditions, and 9.2 percent would accept unification but not independence even if peace can be maintained. Some respondents (8.5 percent) would favor 11 Actually, there are two other questions regarding conditional preferences in these surveys. The respondents are asked if they support (1) independence even if China will attack Taiwan or (2) unification even if the two sides are far apart in economic, social and political conditions. These correspond mostly to the current conditions, and will be treated as such here. 12 For a similar treatment, see John Fuh-sheng Hsieh and Emerson M.S. Niou, Measuring Taiwanese Public Opinion on the Independence-Unification Issue: A Methodological Note, China Quarterly, No. 181 (March 25), pp. 158-68. 1

neither independence nor unification even under favorable conditions. The picture for 215 is similar except that the support for unification even under favorable conditions drops significantly. Now, only 23.1 percent of the respondents support unification conditionally, and a mere 13.4 percent responds positively to both questions. [Table 4 about here] Then, how many of those choosing the status quo in responding to the national identity question without preconditions truly believe in the status quo and how many opt for it due to strategic considerations? To answer these questions, I first reclassify the original status quo supporters on the basis of the following criteria: (1) if they favor independence if peace can be maintained, but do not support unification even if both sides are similar in economic, social and political conditions, they will be reclassified as independence supporters; (2) if they favor independence if condition is right, but do not respond to the question about unification, they will also be reclassified as independence supporters; (3) if they support unification if both sides are similar in economic, social and political conditions, but do not favor independence even if peace can be maintained, they will be reclassified as unification supporters; (4) if they support unification if condition is right, but do not respond to the question about independence, they will also be reclassified as unification supporters; (5) if they say yes or no to both conditional questions, they are regarded as status quo supporters; (6) if they do not respond to both questions, or they say no to one but do not respond to the other, they will not be reclassified; (7) for those who do not respond to the original national identity question, they are reclassified in accordance with their responses to the conditional questions in a similar fashion. Table 5 shows the results. 11

[Table 5 about here] As can be seen from the table, the distribution of voters on the national identity issue is flatter, that is, more people are now at either end of the spectrum. This is, of course, a no-brainer since this is exactly what we have been trying to do. However, it does tell us that there are indeed a lot of people who support the status quo for strategic reasons. And if the conditions are favorable, they may shift their positions. It also tells us that Figure 1 may disguise the degree of polarization in Taiwan it should be more polarized than what Figure 1 suggests if we take into account only the sincere attitudes of the voters. Moreover, even though it makes sense to look at the national identity issue as multimodal, in reality, the picture may be more complicated. Indeed, as Table 6 indicates, at least in 212, the distribution of voters with regard to the association between national identity and partisan attachment is, to a certain extent, dichotomous: Those who favor Taiwan independence tend to support the Pan-DPP camp; those who prefer unification lean toward the Pan-KMT camp; and those who would like to maintain the status quo are also inclined to support the Pan-KMT camp. 13 That is, it is independence pitted against anything else. And this is true no matter whether we are looking at voters summary positions, including both sincere and strategic calculations, or only their adjusted sincere attitudes. In 215, the situation changed somewhat. The DPP made inroads into the status quo territory which was KMT s traditional 13 The reason that status quo supporters may turn to the Pan-KMT camp, particularly the KMT, has a lot to do with the fact that the KMT has been seen as the status quo party. This is understandable since it is the KMT which, to a large extent, defines the status quo that involves the name of the country, the Constitution, the national flag, the national anthem, and so on. These are the very things opposed by the pro-independence forces. 12

stronghold as far as the original attitudes were concerned. This shows, to a certain extent, the DPP s success in campaigning on the maintenance of the status quo leading up to the 216 elections. (However, it is interesting to watch whether the position of the new DPP government on the status quo is sustainable given China s insistence on the one-china principle and many Green supporters antipathy to such an idea.) [Tables 6 about here] Another piece of evidence is the stability of vote shares up until 212 in Taiwan s elections. As seen in Table 7, between 1989 and 212, the Legislative Yuan elections have been relatively stable in terms of popular vote received by the two major camps. In general, the Pan- KMT camp has been able to capture around 5% of the total popular vote while the Pan-DPP camp about one-third in the 199s and over 4% since 2. Presidential elections are a bit different since these are single-seat (or more precisely single-ticket) elections and some idiosyncratic events such as the missile crisis in 1996 or the two bullets in 24 might tilt the election results one way or the other. (By contrast, there are many seats in the legislative elections, and idiosyncratic events in different districts may cancel each other out, thus better reflecting the underlying division in the electorate.) But overall, it can still be detected that the relative strength between the two major camps is, on average, reasonably stable (see Table 8). That is to say, the dividing line between the two major camps has been, to a certain extent, rigid stable and enduring. [Tables 7 and 8 about here] 13

Of course, the 216 elections changed such a pattern. Now, the DPP was able to capture a much higher proportion of the votes. Apart from such medium- to short-term factors as the KMT government s handling of the economy, pension reform for public servants, educational reform with regard to senior high schools, cooking oil scandals, and so on, some long-term factors may be at work as well. 14 More specifically, the young generation are now turning to Taiwan independence, and very few support unification. This does not mean that a new consensus is emerging, but that the traditional advantage enjoyed by the Pan-KMT camp may now shift to the Pan-DPP camp. Therefore, the situation may look more like bipolar than multipolar as shown in Figure 1. Then, if so, does bipolarity necessarily mean serious polarization? Hard to say. It is polarized only if it is difficult to cross the line dividing the two poles and people feel so strongly that they do not want to make any compromises. Indeed, the anecdotal evidence seems to suggest that national identity is, at least for some people, a highly emotional life-and-death issue, similar to what has happened in Northern Ireland. But the situation is somewhat more manageable in Taiwan than in Northern Ireland since there is a readily available compromised position (i.e., the status quo) in Taiwan. 15 So Taiwanese society is polarized, but the situation has been less serious than that in Northern Ireland. Then, are elites more polarized than the general public in Taiwan? To answer this question, I examine the attitudes toward the national identity issue for those who responded 14 See John Fuh-sheng Hsieh, Taiwan s 216 Elections: Critical Elections? American Journal of Chinese Studies (forthcoming). 15 In the case of Northern Ireland, it is difficult to find a compromise between the two alternatives: remaining in the United Kingdom and being part of the Irish Republic. Fortunately, after years of efforts, a compromised position, the Good Friday Agreement of 1998, finally came up, laying the foundation for the solution of the Troubles since the 196s. 14

positively in the 212 TEDS survey that they had been engaged in such activities as campaigning, donating money to candidates, reading election bulletins, watching televised debates among the candidates, and so on. Altogether, 79.5 percent of the respondents have participated in at least one of the 14 activities on the list. As exhibited in Figure 1, these activists (labeled as Activist1) are no more polarized than the ordinary voters. But these 14 activities include some which only show the interests of the respondents in the elections a very low threshold for being labeled as activists. So I exclude some activities like reading election bulletins, watching TV debates, and so forth, from the list. As a result, only 17.5 percent of the respondents have been engaged in at least one of the remaining ten activities. Interestingly, the distribution of this small group of activists (labeled as Activist2 in Figure 1) is only slightly more polarized than the general public or the larger group of activities. The differences among the three groups are very small. Political Impact Then, is polarization good or bad for democratic governance in Taiwan? Intuitively, polarization may increase the chances of gridlock in the policy-making process, particularly if the executive and legislative branches of government are controlled by different political forces as seen from 2 to 28 when the DPP controlled the executive branch while the KMT and its allies maintained a majority in the Legislative Yuan. And down the road, divided government cannot be ruled out completely. As a general rule, the majority camp in the population is likely to capture a majority of parliamentary seats under the current mixed-member majoritarian 15

system; 16 however, under certain conditions like what happened in 2 and 24 the minority camp may be able to win the presidential race. Consequently, a divided government may ensue, 17 and given the polarized nature of the political scene, a certain degree of stalemate is very likely. In terms of specific policies, the stalemate may be particularly acute in policy areas related to China. On the issue of trade and investment in China, for instance, the Pan-KMT camp generally sees the economic interactions between Taiwan and the mainland in a relatively positive light, but the Pan-DPP camp often worries that these will hollow out Taiwan s economy, and may endanger Taiwan s security since China may hold Taiwan s business interests hostage. This becomes a hotly contested issue, rendering compromises difficult, if not impossible. Also, it may raise the stakes in political competition since political gains and losses are considered by some to be a life-and-death choice, that is, it deals with the kind of country they can identify and live in. Thus, election campaigning often brought about heightened tensions between different groups in the society. Is polarization always bad? Not necessarily so. Of course, if there is an extremely high degree of polarization, the society may plummet into such a deep division that compromises become untenable, and stalemate or even chaos may turn out to be the norm. This is certainly not optimal. But it can be argued that a certain degree of division may sometimes serve as a kind of checks and balances among different groupings in the society. As far as democracy 16 John Fuh-sheng Hsieh, The Origins and Consequences of Electoral Reform in Taiwan, Issues & Studies, Vol. 45, No. 2 (June 29), pp. 1-22. 17 This may not necessarily be the case if the president decides to appoint someone acceptable to the Legislative Yuan to head the cabinet. Indeed, President Ma Ying-jeou stated a couple of times that if his party, the KMT, lost the parliamentary majority, he would appoint someone from the majority party to be the premier. And indeed, he kept his promise after the 216 elections by asking the DPP to form the cabinet immediately, but the proposal was turned down by the DPP. 16

is concerned, this may not always be bad. It depends upon whether the division is excessive. As noted earlier, Taiwan is polarized, but, fortunately, not excessively so as compared to, say, Northern Ireland. Conclusion So, although, on surface, there seems to be a trend toward convergence at least in terms of campaign rhetoric in the recent elections, a certain degree of polarization remains in Taiwanese politics. Is polarization a transitory phenomenon that will fade away as Taiwan s democracy matures? Possibly not. It will be present as long as neither independence nor unification is an acceptable and feasible solution. People may continue to lament the seemingly never-ending bickering between the two major political camps. And it is a phenomenon at both the mass and elite levels. So, does it affect public policies? Of course, particular in those areas related to China. Will it eventually adversely affect the very survival of democratic rule in Taiwan? I do not think so since Taiwan s democracy is reasonably well-established as shown in the voters responses to the questions in the 212 TEDS survey about (1) whether democracy is always better than dictatorships and (2) whether they are satisfied with democratic rule in Taiwan. In response to the former question, 53.1 percent of the respondents think that democracy is always the best political system. As for the latter, 67.6 percent of the people are very or somewhat satisfied with democratic rule in Taiwan. Nevertheless, a certain degree of polarization is a reality, and people in Taiwan may have to learn how to live with it. 17

Table 1: Logistic Regression for Leaning toward the Pan-KMT Camp Variable 25 28 212 B S.E. B S.E. B S.E. Intercept -3.392**.747-4.45**.74-2.249**.64 Less Welfare -.57**.2 -.29.25.15.22 Less Environment.42.35.88**.31 Unification.315**.97.499**.1.51**.14 Stability.195**.38.88**.28 Clean Government -.1.22 1.283**.391 Minnan Taiwanese -1.728**.588 -.745.549 -.585.532 Mainlander.956 1.2 1.373.921 2.86**.968 Minnan x Unification.174.115.68.111 -.65.111 Mainlander x Unification.56.194.198.27 -.99.23 Female -.29.179.472**.156.226.139 Older.152*.85.89.7.66.64 Better Educated.236**.86.228**.73.149**.64 Number of Cases 768 1,126 1,187 Nagelkerke R 2.422.492.358 Source: 24 LB (conducted in 25), 28P, and 212 Taiwan Election and Democratization Studies surveys. The reference category is the Pan-DPP camp. * indicates p <.1. ** indicates p <.5. 18

Table 2: Ethnicity and Party Attachments Pan-KMT Camp Pan-DPP Camp KMT NP PFP DPP TSU Total Minnan Taiwanese 449 (48.5%) 5 (.5%) 7 (.8%) 448 (48.4%) 16 (1.7%) 925 (99.9%) Hakka Taiwanese 15 (67.3%) (.%) 4 (2.6%) 47 (3.1%) (.%) 156 (1.%) Mainlanders 15 (84.3%) 8 (4.5%) 8 (4.5%) 12 (6.7%) (.%) 178 (1.%) Source: 212 Taiwan Elections and Democratization Studies survey Note: Cell entries are numbers of respondents, with row percentages in parentheses. Row percentages may not sum to 1 due to rounding Political Parties: KMT: Kuomintang DPP: Democratic Progressive Party NP: New Party PFP: People First Party TSU: Taiwan Solidarity Union 19

Table 3: National Identity, Ethnicity and Party Attachments (a) Independence Supporters Pan-KMT Camp Pan-DPP Camp KMT NP PFP DPP TSU Total Minnan Taiwanese 52 (17.9%) 1 (.3%) (.%) 229 (78.7%) 9 (3.1%) 291 (1.%) Hakka Taiwanese 1 (31.2%) (.%) (.%) 22 (68.8%) (.%) 32 (1.%) Mainlanders 6 (54.5%) (.%) (.%) 5 (45.5%) (.%) 11 (1.%) (b) Status Quo Supporters Pan-KMT Camp Pan-DPP Camp KMT NP PFP DPP TSU Total Minnan Taiwanese 239 (56.9%) 3 (.7%) 7 (1.7%) 166 (39.5%) 5 (1.2%) 42 (1.%) Hakka Taiwanese 62 (73.8%) (.%) 3 (3.6%) 19 (22.6%) (.%) 84 (1.%) Mainlanders 93 (84.5%) 5 (4.5%) 7 (6.4%) 5 (4.5%) (.%) 11 (99.9%) 2

(c) Unification Supporters Pan-KMT Camp Pan-DPP Camp KMT NP PFP DPP TSU Total Minnan Taiwanese 125 (79.1%) 1 (.6%) (.%) 3 (19.%) 2 (1.3%) 158 (1.%) Hakka Taiwanese 27 (87.1%) (.%) 1 (3.2%) 3 (9.7%) (.%) 31 (1.%) Mainlanders 44 (95.7%) 1 (2.2%) (.%) 1 (2.2%) (.%) 46 (1.1%) Source: 212 Taiwan Elections and Democratization Studies survey Note: Cell entries are numbers of respondents, with row percentages in parentheses. Row percentages may not sum to 1 due to rounding 21

Table 4: Approval of Independence or Unification under Favorable Conditions (a) 212 Unification If Both Sides Differ Little Yes No No Response Total Yes 229 (21.3%) 499 (46.3%) 49 (4.5%) 777 (72.1%) Independence If Peace Can Be Maintained No 99 (9.2%) 92 (8.5%) 11 (1.%) 22 (18.8%) No Response 26 (2.4%) 31 (2.9%) 41 (3.8%) 98 (9.1%) Total 354 (32.9%) 622 (57.8%) 11 (9.4%) 177 (1.%) (b) 215 Unification If Both Sides Differ Little Yes No No Response Total Yes 143 (13.4%) 546 (51.%) 46 (4.3%) 735 (68.6%) Independence If Peace Can Be Maintained No 71 (6.6%) 18 (1.1%) 17 (1.6%) 196 (18.3%) No Response 33 (3.1%) 31 (2.9%) 76 (7.1%) 14 (13.1%) Total 247 (23.1%) 685 (64.%) 139 (13.%) 171 (1.%) 22

Source: 212 and 215 Taiwan National Security surveys Note: Those who respond that they agree or agree strongly are categorized as Yes, and those who disagree or disagree strongly are categorized as No. Cell entries are numbers of respondents, with row percentages in parentheses. Row percentages may not sum to 1 due to rounding 23

Table 5: Respondents Attitudes toward the National Identity Issue (a) 212 Independence Status Quo Reunification No Response Total Original 248 (23.1%) 645 (6.%) 113 (1.5%) 69 (6.4%) 175 (1.%) Adjusted 568 (52.9%) 283 (26.4%) 188 (17.5%) 35 (3.3%) 175 (1.1%) (b) 215 Independence Status Quo Reunification No Response Total Original 23 (21.5%) 699 (65.3%) 81 (7.5%) 61 (5.7%) 171 (1.%) Adjusted 613 (57.3%) 289 (27.%) 139 (13.%) 29 (2.7%) 171 (1.%) Source: 212 and 215 Taiwan National Security surveys Note: The adjusted figures include those who are in the No Response category in accordance with their original responses to the 11-point scale question, but are now reclassified according to the questions about conditional preferences. Cell entries are numbers of respondents, with row percentages in parentheses. Row percentages may not sum to 1 due to rounding 24

(a) 212 Original Table 6: Respondents Attitudes toward the National Identity Issue and Their Party Identification Pan-DPP Camp Pan-KMT Camp DPP TSU KMT PFP NP Other Total Independence 122 48.8% 15 6.% 43 17.2% 5 2.%.% 65 26.% 25 1.% Status Quo 131 2.3% 8 1.2% 199 3.9% 3 4.7% 2.3% 275 42.6% 645 1.% Unification 12 1.6%.% 5 44.2% 9 8.% 5 4.4% 37 32.7% 113 1.% Other 1 14.7% 3 4.4% 23 33.8% 2 2.9%.% 3 44.1% 68 1.% Total 275 25.6% 26 2.4% 315 29.3% 46 4.3% 7.7% 47 37.8% 176 1.% (b) 212 Adjusted Pan-DPP Camp Pan-KMT Camp DPP TSU KMT PFP NP Other Total Independence 24 35.9% 23 4.% 126 22.1% 12 2.1% 1.2% 23 35.7% 569 1.% Status Quo 48 17.% 3 1.1% 93 32.9% 14 4.9% 1.4% 124 43.8% 283 1.% Unification 21 11.2%.% 86 45.7% 18 9.6% 5 2.7% 58 3.9% 188 1.% Other 2 5.7%.% 1 28.6% 2 5.7%.% 21 6.% 35 1.% 25

Total 275 25.6% 26 2.4% 315 29.3% 46 4.3% 7.7% 46 37.8% 175 1.% (c) 215 Original Pan-DPP Camp Pan-KMT Camp DPP TSU NPP KMT PFP NP Other Total Independence 115 5.% 6 2.6% 3 1.3% 21 9.1% 2.9%.% 83 36.1% 23 1.% Status Quo 25 29.3% 4.6% 4.6% 163 23.3% 19 2.7% 2.3% 33 43.3% 7 1.% Unification 11 13.6%.%.% 32 39.5% 3 3.7% 2 2.5% 33 4.7% 81 1.% Other 9 14.5% 2 3.2% 1 1.6% 4 6.5%.%.% 46 74.2% 62 1.% Total 34 31.7% 12 1.1% 8.7% 22 2.5% 24 2.2% 4.4% 465 43.3% 173 1.% (d) 215 Adjusted Pan-DPP Camp Pan-KMT Camp DPP TSU NPP KMT PFP NP Other Total Independence 258 7 6 82 9 1 25 613 42.1% 1.1% 1.% 13.4% 1.5%.2% 4.8% 1.% Status Quo 61 2 1 82 1 132 288 21.2%.7%.3% 28.5% 3.5%.% 45.8% 1.% Unification 18 1 1 54 6 2 57 139 26

12.9%.7%.7% 38.8% 4.3% 1.4% 41.% 1.% Other 2 1 1 26 3 6.7%.% 3.3% 3.3%.%.% 86.7% 1.% Total 339 1 9 219 25 3 465 17 31.7%.9%.8% 2.5% 2.3%.3% 43.5% 1.% Source: 212 and 215 Taiwan National Security surveys Note: Cell entries are numbers of respondents, with row percentages in parentheses. Row percentages may not sum to 1 due to rounding 27

Table 7: Vote Shares of Major Political Parties in Legislative Yuan Elections, 1969-216 (%) KMT DPP NP PFP TSU NPSU NPP 1969 76. 1972 73.9 1975 79.4 198 73.7 1983 73.1 1986 69.2 22.2 1989 6.2 28.2 1992 53. 31. 1995 46.1 33.2 13. 1998 46.4 29.6 7.1 21 28.6 33.4 2.6 18.6 7.8 24 32.8 35.7.1 13.9 7.8 3.6 28 (Dist.)* 53.5 38.2..9 2.4 (PR) 51.2 36.9 4. 3.5.7 212 (Dist.)* 48.2 43.8.1 1.3 1.3 (PR) 44.5 34.6 1.5 5.5 9. 216 (Dist.)* 38.9 44.6.6 1.3.8.2 2.9 (PR) 26.9 44.1 4.2 6.5 2.5.6 6.1 *The district figures for 28, 212, and 216 elections include the votes for both singlemember districts and the two three-member districts reserved for aborigines. Political Parties: KMT: Kuomintang DPP: Democratic Progressive Party NP: New Party PFP: People First Party TSU: Taiwan Solidarity Union NPSU: Non-Partisan Solidarity Union Political Parties NPP: New Power Party 28

Table 8: Vote Shares of Major Candidates in Presidential Elections, 1996-212 (%) 1996 2 24 28 212 216 KMT DPP 3 rd Party or Independent Candidates 54. 21.13 14.9* 9.98* 23.1 39.3 36.84* 49.89 5.11 58.45 41.55 51.6 31.4 45.63 56.12 2.76** 12.83** *KMT-turned independent **PFP Candidate 29

Figure 1: The Distribution of Voters on the National Identity Issue in Taiwan 6 5 4 3 2 All Activist1 Activist2 1 1 2 3 4 5 6 7 8 9 1 Source: 212 Taiwan Elections and Democratization Studies survey Note: refers to independence and 1 unification. All stands for all respondents; Activist1 for those who choose any of the 14 activities on Question B8 in the survey; and Activist2 for those who choose the 1 more demanding of the 14 activities. 3