The central issue in Taiwan’s politics revolves around the island’s relationship with China. The three major parties—the Kuomintang, Democratic Progressive Party, and Taiwan People’s Party—take different positions on this, and their supporters align with these respective positions. However, there is a scarcity of research on the foreign policy inclinations of individuals who do not align with any political party. We find empirically that self-reported independents in Taiwan exhibit a preference for bandwagoning, driven by concerns for tangible material benefits; cultural similarity, economic interests, and the power dynamics between the US and China also play substantial roles here. These findings shed light on the unique role and dynamics of independents in Taiwan’s political landscape. Our theoretical innovation lies in showing that in a democracy, different factors come into play in the formation of the foreign policy preferences of different groups.
Public opinion has a significant influence on government policies in Taiwan (Wu 2019), one of the few democracies in Asia. The major division within the Taiwanese public is the unification/independence issue, which is intertwined with Taiwanese/Chinese identity. Formerly a staunch advocate for Chinese unity, the century-old Kuomintang (KMT) now promotes closer ties with China as a means to attain economic benefits and mitigate the looming threat of military actions from China. In contrast, the Democratic Progressive Party (DPP) embraces Taiwanese nationalism and advocates autonomy from China. In pursuit of this objective, the DPP places great importance on fostering collaboration with the United States.
Other parties, including the New Party (NP), People’s First Party (PFP), New Power Party (NPP), and Taiwan People’s Party (TPP), have emerged alongside the two main parties since democratization. However, some 40% of Taiwan’s electorate remains unaffiliated with any party. Given Taiwan’s use of the first-past-the-post system for presidential elections, the policy preferences of independents have the potential to sway election outcomes. Furthermore, these election results could wield significant influence over the geopolitics of Asia (Wang 2023). Thus analysis of independents’ political attitudes can offer insights into voter preferences and the formulation of broader policy decisions.
This study draws on the literature on foreign policy postures (Hurwitz and Peffley 1987). The first posture is realism. There are many kinds of realism, but most ultimately contend that, given the uncertainties created by an anarchical international system, foreign policy preferences focus on maximizing material security interests, including the economic sources of security.1 In the case of Taiwan, the prediction derived from core realist claims, as held by individual members of the public (“folk realism”), is that the perception of China’s growing influence in the region and economic interests in cross-Strait trade tends to produce a preference for closer ties with China. Here we borrow from Schweller’s (1994) argument that when the immediate security of the state is not at stake, states prefer to “bandwagon for profit” most of the time—in other words, to reap security and economic gains by leaning toward stronger powers. At the level of individual preferences, then, realism suggests that people are likely to prefer improving economic relations with stronger neighbors.
The other major posture, distinct from realism’s assumptions about rationally maximizing material interest, is social identity. Here, in-group identity determines who the out-group is and how threatening it is considered. Thus, in the case of Taiwan, whether people think of themselves as Taiwanese, Chinese, or somewhere in between helps determine these things. The DPP’s strong sense of Taiwanese-ness has increasingly framed China as a threat to Taiwan’s autonomy, particularly since the 2019 anti-extradition protests in Hong Kong. However, the KMT’s tendency to see identity as combining elements of Taiwanese-ness and Chinese-ness has downplayed these concerns, contributing to the party’s support for leaning toward China for stability and prosperity (Fell 2021).
Our theoretical innovation is to suggest that when analyzing the foreign policy preferences of different political groups in a democracy, these postures may apply differently to different groups. For example, social identity may not apply to assessments of the relationship with China, because independents often feel alienated from political parties and may not embrace certain motivations parties emphasize, such as cultural identity. Instead, folk realism suggests that when low-information individuals face uncertainty, they tend to favor prudent policies that enhance self-interest, such as security or economic well-being (Kerzter and McGraw 2012). Without party cues, independents may rely more on information from their everyday lives. Therefore, in setting their foreign policy preferences we expect that independents are more likely to consider valence issues such as cross-Strait trade and China’s military coercion. In this context, independents will tend to prefer alignment with the growing power, which, in the case of Taiwan, refers to China.
We analyzed data from three survey waves to estimate the impact of identifying as an independent on foreign policy attitudes. Independents’ preferences differed from those of the comparison group, comprised of supporters of the DPP.2 In the next section, we review three strands of literature that might help explain the sources of independents’ foreign policy preferences.
Background
Partisanship and Information Processing
Partisanship, a central concept in American politics, refers to a long-term emotional attachment to a specific political party (Campbell et al. 1960). It guides voting behavior and fosters a lasting affective connection to politics as a whole (Green, Palmquist, and Schickler 2002). The Survey Research Center (in the Center for Political Studies, University of Michigan, USA) measures party identification using a two-part question. Respondents are first asked whether they identify with a political party. Partisans are then asked to indicate the strength of their affiliation, leading to a seven-point partisanship scale. Studies have shown that partisanship not only affects voting decisions but also provides cues for evaluating candidates and issues (Campbell et al. 1960; Clarke et al. 2009; Converse and Dupeux 1962; Miyake 1991). Recent studies underscore the stability of partisanship (Green and Platzman 2022; Kinder and Kamore 2017), although Kollman and Jackson (2021) noted that shifts within parties can alter partisan loyalties.
Despite the strong influence of party identity, many scholars argue that rational voters can transcend it. Key (1966) posited that voters can comprehend political issues that are presented clearly, a notion echoed by Nie, Verba, and Petrocik (1976). Some scholars suggest that political campaigns can increase voters’ knowledge and affect their voting intentions (Andersen 2003; Andersen, Tilly, and Heath 2005; Gelman and King 1993; Lazarsfeld, Berelson, and Gaudet 1944). It is now widely acknowledged that citizens maintain consistent policy positions and ideologies and can apply ideological labels to political parties (Abramowitz and Saunders 2006; Jacoby 1995).
While research highlights the significance of voter cognition, there is considerable evidence for the role of partisanship in information processing. Zaller (1992) identified partisan bias in how information is received, leading to biased issue positions. Individuals with stronger partisanship tend to more readily accept favorable information about their party’s candidates. Rahn (1993) also found that partisan stereotypes, which include policy information, group membership, trait judgments, and performance assessments, affect information processing. Green, Palmquist, and Schickler (2002, 136–37) emphasized that partisanship is a strong predictor of people’s opinions on important issues. And Delli Carpini and Keeter (1996) demonstrated that strength of party identification, sense of civic duty, and political efficacy influence political knowledge. In other words, partisanship not only impacts issue positions but also affects the cognitive process of preference formation.
Despite limited policy knowledge, individuals can make informed choices using accessible cognitive heuristics, as proposed by Lupia and McCubbins (1998). Parties that focus on specific policies can gain “issue ownership” (Petrocik 1996), allowing voters to use “gut reasoning” and simplified assumptions, such as “my enemy’s enemy is my friend” (Popkin 1991, 214), to evaluate complex information. Alvarez (1998) found that independent voters tend to be better informed about candidates’ policy positions and use partisanship as a filter to sift through information, while strongly ideological individuals are more likely to attribute such positions to their ideological group (Brady and Sniderman 1985). In contrast, less informed voters may rely on group heuristics—the views of in-group members they know or like—rather than ideology to form their issue positions (Cohen 2003; Sniderman et al. 1986).
In established democracies like the United States, there has been a noticeable decline in partisanship. This trend is primarily associated with candidate-centered politics and weakening party identification (Dalton 2016; LeBlanc and Merrin 1979; Wattenberg 1986). Consequently, the number of individuals without party identification has increased, a phenomenon known as party dealignment (Dalton 2012). Several factors contribute to this phenomenon, including low interest in politics (Tanaka and Martin 2003), the growing influence of political leaders in candidate evaluation by voters (Garzia, Ferreira de Silva, and De Angelis 2020), and party polarization, which alienates some voters (Beck 1999; Smidt 2017). Despite the increasing polarization among issue activists, donors, and campaign contributors, since the 1990s independents have remained the largest group (Fiorina 2016), underscoring the importance of considering them in public opinion research.
One of the main debates surrounding independents is whether they consistently occupy the ideological center on issues or tend to align with one of the major parties. The distinction between independents, weak partisans, and strong partisans may not be as clear-cut as the seven-point scale suggests. Craig (1985) and Valentine and Wingen (1980) identified similarities between weak partisans, independents, and no-preference nonpartisans. Klar and Krupnikov (2016) found that independents intentionally maintain a distance from political parties despite still having partisan preferences. In terms of voting behavior, there is substantial evidence that independents may base their choices on party evaluations (Tsai and Chao 2008), factors beyond partisanship (Hajnal and Lee 2011), and implicit party identity (Hawkins and Nosek 2012). Overall, there are mixed findings in the literature on the distinction between weak partisans, leaners, and independents.
Research has found that independent voters in the United States have more negative attitudes toward parties in general and place more emphasis on cognition in partisan attitudes than party identifiers do (Greene 2000, 515). In Japan, independent voters tend toward an apolitical or apathetic stance on political affiliations. In Taiwan, social movements and progressive forces may contribute to alienation from the DPP (Qi 2019). Some moderate partisans may become nonpartisan as parties become more polarized (Wang 2019). These studies suggest that in the absence of party affiliation, independents’ decision-making is less constrained by ideology or party allegiance; it may be related more to self-interest than to symbolic or ideological values (Sears, Hensler, and Speer 1979; Sears et al. 1980). Independents tend to be politically alienated (Keith et al. 1992, 198), yet they frequently point to the economy as the most important issue (Saad 2022).
In the present study, individuals were categorized, following standard partisanship measurements, into four groups: independents, pan-Blue, pan-Green, and TPP. Citizens who identified as “independent” or who expressed no partisan preference were classified as independents, a categorization that helps avoid inflating the number of party loyalists (Miller and Wattenberg 1983). The pan-Blue group consisted of the KMT, NP, and PFP, while the pan-Green group consisted of the DPP, NPP, and Taiwan Statebuilding Party (TSP).3 We examined the differences in foreign policy attitudes between independents and partisans from the two major political camps and one emerging party.4 As Taiwanese/Chinese identity and cross-Strait relations are inextricably intertwined with party positions, partisans are likely to view foreign policy through the lens of partisanship and identity. However, independents, who are alienated from political parties, may focus on tangible benefits and future well-being, as suggested by the concept of folk realism. Given the strong economic ties between China and Taiwan, independents might be more inclined to support closer ties with China, particularly for economic development reasons.
Foreign Policy and Public Opinion
The consent of ordinary citizens is seen as a crucial requirement for shaping foreign policy in democracies, including decisions regarding war and peace with other nations (Holsti 1996). With increasing economic, social, and political interconnectedness across borders, recent research has redefined foreign policy in democracies as collective actions taken by ordinary citizens, either through existing state institutions or other collective means, beyond their country’s borders (Headley and van Wyk 2012, 4). However, earlier studies expressed skepticism about the public’s level of sophistication in foreign policy matters. In particular, there is little consistency between ordinary people’s domestic and foreign policy attitudes (Converse 1964, 31). Still, the public is likely to be informed about specific policies, such as the Vietnam War (Verba et al. 1967, 331).
Recent studies have proposed a hierarchical model that includes a vertical relationship between various abstract constructs and specific foreign policy attitudes (Hurwitz and Peffley 1987; Peffley and Hurwitz 1985). It has been observed that foreign policy attitudes are well organized and linked to core values and general postures (Hurwitz and Peffley 1987, 1105). News commentary plays a significant role in shaping opinions on foreign policy (Jordan and Page 1992, 234). Religious beliefs and elite polarization also have a notable influence (Cavari 2013). This line of research also suggests that values hold more sway than self-interest, as proposed by Lau, Brown, and Sears (1978).
Additional evidence indicates that the general public holds meaningful attitudes toward foreign policy. Aldrich, Sullivan, and Bordiga (1989) argued that voters can differentiate between candidates’ stances on defense and foreign policy issues, influencing their voting decisions. Berinsky (2009, 210) posited that “partisan politics and group affiliations drive the public’s decision to support or oppose military conflicts.” Johns’s (2009) experimental research demonstrated that individuals may use heuristics from political parties, international actors, or news sources to form their foreign policy opinions. Kertzer and McGraw (2012) and Kertzer and Zeitzoff (2017) found that the public develops general foreign policy orientations, or “folk theories,” about international relations, based on intuitions, information, and experiences acquired from daily interactions with others. However, Todorov and Mandisodza (2004) found that widespread misperceptions about foreign policy in public opinion were linked to misunderstandings of specific policies.
Over the past three decades, the most significant issue in Taiwanese politics has been the question of unification versus independence (Hsieh and Niou 1996a, 1996b; Sheng and Liao 2017). While supporters of the KMT and DPP are polarized on the unification/independence issue (Achen and Wang 2017), the political participation of independents could lead to unpredictable outcomes (Tan et al. 2000). The proportion of independents favoring independence under different scenarios lies between Blue and Green partisans (Wang 2013). Therefore, studying the “moderate middle” is key to understanding how Taiwanese/Chinese identity and views on unification/independence influence election outcomes.
Bandwagoning with China or Balancing against China
As Yu (2022) demonstrates, Taiwan’s foreign policy can be understood through a balancing-versus-bandwagoning framework. Investigating how people in the “defender” states like Taiwan select balancing against or bandwagoning with China has broader implications for other states, including those in Southeast Asia (Roy 2005). Realist theories suggest that in an anarchic international system, states seeking to maximize security have two main options when faced with an emerging great power. If the rising power constitutes an acute security threat, states may opt for balancing, forming alliances to counter it (Waltz 1979). But if the rising power offers economic benefits that could also enhance security, a smaller state might choose to bandwagon instead (Schweller 1994).5 While balancing prioritizes security concerns, bandwagoning aims to acquire benefits beyond security. On the other hand, constructivism argues that states’ interests are socially constructed and stem from their identities. This perspective suggests that states seek closer relationships (bandwagoning or allying) with states they consider part of their identity in-group, but may pursue conflictual relations (balancing or leaning against) states they consider members of their identity out-group. In this view, anarchy is less relevant in explaining these preferences (Wendt 1999, 7–10); alliances are formed among states with similar identities.
Since 1949, China has claimed Taiwan as part of its territory. China regularly issues threats of military action if Taiwan were to declare independence. To enhance Taiwan’s defense capabilities, the US provides the island with defensive weapons in accordance with the Taiwan Relations Act of 1979. Meanwhile, China has been putting military pressure on Taiwan and blocking its participation in international organizations.6 China also frequently employs coercive measures against Taiwan’s companies, tourism industry, and agriculture as a means to push for “peaceful reunification” (Lai 2022). Yet amid the persistent tensions involving these three countries, China and the US are Taiwan’s largest and second-largest trading partners, respectively. And Taiwan consistently maintains significant trade surpluses with both of these countries.
As the bilateral relationship between the US and China fluctuates, these two strategies, balancing and bandwagoning, represent the basic choices that Taiwanese people must make at this critical juncture. As noted earlier, we draw on the folk realism and social identity theory to explain an individual’s choice between China and the US. From the perspective of folk realism, bandwagoning is preferred, to gain temporary access to the benefits offered by the more powerful state (Kirshner 2022; Schweller 1994). In contrast, social identity theory focuses on the roles of group identity and historical interactions between groups in shaping interests. This theory posits that individuals are inclined to support their in-group while disparaging the out-group (Hogg 2016; Hurwitz and Peffley 1987; Tajfel and Turner 2004). Based on these theoretical perspectives, an individual’s foreign policy orientation may be shaped by the presence or absence of party cues and the ability to cognitively process these cues (partisanship and information processing), by their social identities, or by their beliefs about how to maximize state security and economic gains. To test these hypotheses, this study employed empirical data and ordered probit regression models to explore the factors associated with the foreign policy positions of independents and partisans.
Data and Hypotheses
To create a larger sample, we combined data from three original surveys conducted by the Election Study Center at National Chengchi University in November 2019, November 2020, and November 2021, respectively. These samples were randomly selected from household telephone directories, with stratification proportional to the distribution of landline telephone numbers in each county/city in Taiwan (excluding Penghu, Kinmen, and Matsu, following the conventions of public opinion polling in Taiwan).7 The 2019 survey yielded 1,222 completed interviews from 5,646 contacts; the 2020 survey, 1,215 completed interviews from 5,646 contacts; and the 2021 survey, 1,120 completed interviews from 6,146 contacts. Merging the three data sets gives a total of 3,557 cases, of which 1,380 are self-labeled independents.8 Given that a new random sample was collected each year, we included the year as a variable in our analysis to assess its effect on the dependent variable.
Drawing from the literature above, we formed three hypotheses.
H1: Independents are likely to choose alignment with China rather than the US.
Our research assumes that independents in Taiwan are inclined to favor a better relationship between China and Taiwan for national security and economic development because the economy is an overarching concern for many Taiwanese people, including independents. As well as comparing the alignment preferences of independents with those of DPP partisans, we estimated the impact of perceived cultural similarity, China’s influence, and cross-Strait trade on foreign policy attitudes. We expect that if independents do not rely on partisan cues, the interaction term between independent status and perceived cultural similarity will be insignificant. However, we expect that the interaction terms between being an independent and China’s influence, and between being an independent and cross-Strait trade, will affect alignment preferences. If this is the case, it strongly suggests that independents make decisions about foreign policy based on folk realism.
H2a: There are no significant interaction effects between identifying as an independent and cultural similarity.
H2b: Significant interaction effects exist between identifying as an independent and perceptions of China’s influence and between identifying as an independent and cross-Strait trade.
As mentioned, social identity theory suggests that people tend to favor their in-group while exhibiting bias against out-groups. Political action is not only about evaluating alternatives based on their outcomes; more importantly, it is about the psychological benefits of aligning with the norms of the group (March 1994, 57–58). The content of a particular identity is defined as the collective expectations of the members of an identity group. For instance, the defining elements of European identity include a market economy, a democratic polity, and respect for human rights (Abdelal et al. 2006, 697). Consequently, we argue that perceptions of shared cultural identity shape foreign policy attitudes. In other words, individuals who see Chinese and Taiwanese culture as similar are more likely to choose alignment with China, while those who perceive them as fundamentally different are more likely to choose balancing. Given the security and economic tensions between Taiwan and China, individuals who identify as Taiwanese may view China as a threat rather than an opportunity. Also, Taiwanese identity represents an in-group preference that opposes Chinese identity. Consequently, self-identification as Taiwanese is expected to decrease the likelihood of choosing to align with China.
As we saw in the literature review, realist theories underscore the importance of both economic gains and the maximization of security. Considering Taiwan’s continued reliance on cross-Strait trade, it is logical to assume that Taiwanese people are incentivized to align with China. Thus we propose that respondents who prioritize cross-Strait economic interests are likely to align with the country that offers greater economic benefits.9 Following the logic of folk realism, we also expect that individuals who perceive China’s influence in Asia as stronger than that of the US are more supportive of bandwagoning with China.
The final proposition of our study posits that one’s stance on the issue of independence or unification is related to preferences regarding alignment with China. Supporters of unification with China are likely to prefer alignment with China, while individuals who believe that Taiwan should maintain its status as an independent country are likely to consider political and economic alignment with China less attractive.
The study investigates the preferences of respondents regarding relations with either China or the US, based on considerations of national security or economic development. In line with folk realism, we take perceived power distribution as a crucial influence on foreign policy attitudes. To measure this, respondents were asked whether the US or China has more influence in Asia (with answers on a five-point scale). Economic preferences were measured by asking about the relative importance of cross-Strait trade and sovereignty, with the premise that individuals who prioritize trade over sovereignty might be more inclined to align with China. Perceptions of cultural similarity were measured using a question about whether Chinese and Taiwanese culture are similar or dissimilar (with answers on a four-point scale). To assess the combined effect of cultural similarity, trade, and power, as well as respondents’ identification as independents, we created an interaction variable. Table 1 presents the specific wording used for each variable in the study, and Table A1 in the appendix provides descriptive statistics.
Wording of Variables
| Variable | Wording | Options |
|---|---|---|
| Alignment preference on national security | Regarding Taiwan’s national security, do you think Taiwan should focus on strengthening its relationship with the US or China? | The relationship with the US is very important = 1 The relationship with the US is important = 2 The same = 3 The relationship with China is important = 4 The relationship with China is very important = 5 |
| Alignment preference on economic development | Regarding Taiwan’s economic development, do you think Taiwan should focus on strengthening its relationship with the US or China? | The relationship with the US is very important = 1 The relationship with the US is important = 2 The same = 3 The relationship with China is important = 4 The relationship with China is very important = 5 |
| Military power | Do you think the US or China has more influence in Asia today? | The US has a lot more influence = 1 The US has more influence = 2 The same = 3 China has more influence = 4 China has a lot more influence = 5 |
| Cultural similarity | Are Taiwanese and Chinese culture similar or different? | Completely different = 1 Mostly dissimilar = 2 Mostly similar = 3 Very similar = 4 |
| Trade vs. sovereignty | If we cannot find a balance between sovereignty and cross-Strait trade, which one do you think is more important? | Cross-Strait trade = 1, otherwise = 0 |
| Party identification | There are many political parties in Taiwan. Which party do you identify with? | KMT, PFP, NP = pan-Blue DPP, TSP = pan-Green Don’t identity with any party = Independent |
| Independence/unification | Regarding cross-Strait relations, what do you prefer? | Independence as soon as possible = 1 Maintain the status quo and then move toward independence = 2 Maintain the status quo forever = 3 Maintain the status quo and decide later = 4 Maintain the status quo and then move toward unification = 5 Unification as soon as possible = 6 |
| Taiwanese identity | Do you think of yourself as Taiwanese, Chinese, or both? | Taiwanese = 1 |
| Gender | ||
| Age | ||
| Education |
| Variable | Wording | Options |
|---|---|---|
| Alignment preference on national security | Regarding Taiwan’s national security, do you think Taiwan should focus on strengthening its relationship with the US or China? | The relationship with the US is very important = 1 The relationship with the US is important = 2 The same = 3 The relationship with China is important = 4 The relationship with China is very important = 5 |
| Alignment preference on economic development | Regarding Taiwan’s economic development, do you think Taiwan should focus on strengthening its relationship with the US or China? | The relationship with the US is very important = 1 The relationship with the US is important = 2 The same = 3 The relationship with China is important = 4 The relationship with China is very important = 5 |
| Military power | Do you think the US or China has more influence in Asia today? | The US has a lot more influence = 1 The US has more influence = 2 The same = 3 China has more influence = 4 China has a lot more influence = 5 |
| Cultural similarity | Are Taiwanese and Chinese culture similar or different? | Completely different = 1 Mostly dissimilar = 2 Mostly similar = 3 Very similar = 4 |
| Trade vs. sovereignty | If we cannot find a balance between sovereignty and cross-Strait trade, which one do you think is more important? | Cross-Strait trade = 1, otherwise = 0 |
| Party identification | There are many political parties in Taiwan. Which party do you identify with? | KMT, PFP, NP = pan-Blue DPP, TSP = pan-Green Don’t identity with any party = Independent |
| Independence/unification | Regarding cross-Strait relations, what do you prefer? | Independence as soon as possible = 1 Maintain the status quo and then move toward independence = 2 Maintain the status quo forever = 3 Maintain the status quo and decide later = 4 Maintain the status quo and then move toward unification = 5 Unification as soon as possible = 6 |
| Taiwanese identity | Do you think of yourself as Taiwanese, Chinese, or both? | Taiwanese = 1 |
| Gender | ||
| Age | ||
| Education |
Source: Authors.
Results
In general, we see a moderately strong preference for alignment with the US regarding both national security and economic development. Asked whether Taiwan should foster a better relationship with China or the US for national security reasons, 51% of respondents chose the US, while 35% favored China. With respect to economic development, 47% preferred the US, while 41% preferred China (Figure 1). It appears that Taiwanese citizens are almost evenly divided regarding whether to align with the US or China to promote economic development.
Alignment preference regarding national security and economic development.
Source: Authors.
Alignment preference regarding national security and economic development.
Source: Authors.
We also compared the preferences of independents on these two issues with those who identify with a political party (Figure 2). Independents were fairly evenly distributed in their preferences regarding national security alignment. In contrast, pan-Green and pan-Blue identifiers exhibited a clear preference for the US and China, respectively. TPP supporters were the most evenly distributed across the five responses. Regarding economic development, most independents adopted a moderate stance, with only a few strongly favoring a closer relationship with either the US or China. Pan-Blue and pan-Green respondents, while less polarized in their economic alignment preferences compared to their views on national security, still tended to lean toward either China or the US, respectively. Notably, for TPP partisans, China appears to be more attractive than the US in terms of economic development. These findings suggest that the four groups hold distinctive alignment preferences, with independents displaying some ambivalence, pan-Blue supporters leaning toward alignment with China, pan-Green identifiers strongly favoring the US, and TPP supporters tending to align with independents on national security issues and with pan-Blues on economic development.
Alignment preference of partisans and independents regarding national security and economic development.
Source: Authors.
Alignment preference of partisans and independents regarding national security and economic development.
Source: Authors.
Alignment preferences are influenced by factors such as perceptions of China’s influence, cultural similarity, and the trade-off between cross-Strait trade and state sovereignty. More importantly, the interaction between these factors and being an independent can help us understand how independents differ from other groups when it comes to alignment choices.10 Being an independent was associated with a 0.59 standard deviation increase (compared to partisans and other respondents) in the latent scale of alignment with China (Table 2, column 1). Similarly, pan-Blue supporters and TPP partisans were more inclined than DPP partisans to bandwagon with China. Perception of China’s influence, emphasis on trade, Chinese cultural similarity, and unification position positively influence preference for bandwagoning with China. Having an exclusively Taiwanese identity has the opposite effect. The interaction terms between being an independent and perceiving China’s influence as greater and between being an independent and prioritizing trade were not statistically significant. The findings in Model 3 align with the results of the previous models, further supporting our initial hypotheses. These results suggest that independents do not base their alignment preferences on cultural identity, material interests, or regional influence. Overall, the findings in Table 2 support H2a but not H2b.
Ordered Probit Regression Models for Security Alignment (DV: bandwagoning)
| Coefficients | Model 1 | Model 2 | Model 3 |
|---|---|---|---|
| Partisanship (Reference: Pan-Green) | |||
| Independent | 0.593* (0.125) | 0.488* (0.070) | 0.759* (0.166) |
| Pan-Blue | 0.810* (0.074) | 0.810* (0.075) | 0.803* (0.074) |
| Taiwan People’s Party | 0.415* (0.097) | 0.415* (0.098) | 0.411* (0.097) |
| China’s influence | 0.277* (0.021) | 0.262* (0.017) | 0.261* (0.017) |
| Trade vs. Sovereignty | 0.380* (0.050) | 0.413* (0.066) | 0.380* (0.050) |
| Cultural similarity | 0.190* (0.032) | 0.191* (0.032) | 0.241* (0.041) |
| Identity (reference: dual identity and Chinese identity) | |||
| Taiwanese identity | –0.535* (0.054) | –0.532* (0.054) | –0.535* (0.054) |
| Unification | 0.082* (0.019) | 0.081* (0.019) | 0.082* (0.019) |
| Independent × China’s Influence | –0.041 (0.034) | – | – |
| Independent × Trade | –0.075 (0.096) | – | |
| Independent × Culture Similarity | –0.120 (0.062) | ||
| Year 2020 | 0.001 (0.054) | –0.002 (0.054) | 0.002 (0.054) |
| Year 2021 | 0.455* (0.056) | 0.453* (0.056) | 0.456* (0.056) |
| Cutpoints | |||
| US very important | US important | 1.275*** (0.187) | 1.233*** (0.182) | 1.344*** (0.193) |
| US important | Equally important | 2.171*** (0.189) | 2.070*** (0.185) | 2.240*** (0.195) |
| Equally important | China important | 2.626*** (0.191) | 2.517*** (0.186) | 2.696*** (0.196) |
| China important | China very important | 3.441*** (0.193) | 3.334*** (0.189) | 3.511*** (0.199) |
| AIC | 6,614.238 | 6,615.102 | 6,661.986 |
| Maximized likelihood R2 | 0.757 | 0.757 | 0.757 |
| Adjusted McFadden R2 | 0.349 | 0.349 | 0.349 |
| n | 2,540 | 2,540 | 2,540 |
| Coefficients | Model 1 | Model 2 | Model 3 |
|---|---|---|---|
| Partisanship (Reference: Pan-Green) | |||
| Independent | 0.593* (0.125) | 0.488* (0.070) | 0.759* (0.166) |
| Pan-Blue | 0.810* (0.074) | 0.810* (0.075) | 0.803* (0.074) |
| Taiwan People’s Party | 0.415* (0.097) | 0.415* (0.098) | 0.411* (0.097) |
| China’s influence | 0.277* (0.021) | 0.262* (0.017) | 0.261* (0.017) |
| Trade vs. Sovereignty | 0.380* (0.050) | 0.413* (0.066) | 0.380* (0.050) |
| Cultural similarity | 0.190* (0.032) | 0.191* (0.032) | 0.241* (0.041) |
| Identity (reference: dual identity and Chinese identity) | |||
| Taiwanese identity | –0.535* (0.054) | –0.532* (0.054) | –0.535* (0.054) |
| Unification | 0.082* (0.019) | 0.081* (0.019) | 0.082* (0.019) |
| Independent × China’s Influence | –0.041 (0.034) | – | – |
| Independent × Trade | –0.075 (0.096) | – | |
| Independent × Culture Similarity | –0.120 (0.062) | ||
| Year 2020 | 0.001 (0.054) | –0.002 (0.054) | 0.002 (0.054) |
| Year 2021 | 0.455* (0.056) | 0.453* (0.056) | 0.456* (0.056) |
| Cutpoints | |||
| US very important | US important | 1.275*** (0.187) | 1.233*** (0.182) | 1.344*** (0.193) |
| US important | Equally important | 2.171*** (0.189) | 2.070*** (0.185) | 2.240*** (0.195) |
| Equally important | China important | 2.626*** (0.191) | 2.517*** (0.186) | 2.696*** (0.196) |
| China important | China very important | 3.441*** (0.193) | 3.334*** (0.189) | 3.511*** (0.199) |
| AIC | 6,614.238 | 6,615.102 | 6,661.986 |
| Maximized likelihood R2 | 0.757 | 0.757 | 0.757 |
| Adjusted McFadden R2 | 0.349 | 0.349 | 0.349 |
| n | 2,540 | 2,540 | 2,540 |
***p < 0.001, *p < 0.05 (two-tailed test).
Notes: Coefficients of demographic variables are not shown here. The demographic variables are dummy variables for female, 30–39 years old, 40–49 years old, 50–59 years old, 60+ years old, junior high school, senior high school, some college, and university degree. None is significant statistically across the three models. The full results and data are available on request.
Source: Authors.
In terms of economic development, independents generally favored alignment with China over the US (Table 3). Pan-Blue supporters were also more inclined to align with China. Cultural similarity, perceptions of China’s influence, and cross-Strait trade were positively correlated with preference for a closer relationship with China. The interaction between trade and being an independent also had a significant coefficient. This underscores that on economic matters, independents are highly inclined to endorse bandwagoning with China.11
Ordered Probit Regression Models for Economic Alignment (DV: bandwagoning)
| Coefficients | Model 4 | Model 5 | Model 6 |
|---|---|---|---|
| Partisanship (Reference: Pan-Green) | |||
| Independent | 0.569* (0.123) | 0.529* (0.069) | 0.664* (0.164) |
| Pan-Blue | 0.745* (0.073) | 0.718* (0.074) | 0.742* (0.073) |
| Taiwan People’s Party | 0.462* (0.096) | 0.441* (0.097) | 0.461* (0.096) |
| China’s influence | 0.301* (0.021) | 0.288* (0.017) | 0.287* (0.017) |
| Trade vs. Sovereignty | 0.412* (0.049) | 0.499* (0.064) | 0.412* (0.049) |
| Cultural similarity | 0.169* (0.031) | 0.171* (0.031) | 0.205* (0.040) |
| Identity (reference: dual identityand Chinese identity) | |||
| Taiwanese identity | –0.503* (0.053) | –0.503* (0.053) | –0.504* (0.053) |
| Unification | 0.075 (0.019) | 0.074 (0.019) | 0.074 (0.019) |
| Independent × China’s Influence | –0.035 (0.033) | – | – |
| Independent × Trade vs. Sovereignty | –0.197* (0.095) | – | |
| Independent × Culture Similarity | –0.085 (0.062) | ||
| Year 2020 | –0.006 (0.053) | –0.011 (0.053) | –0.006 (0.053) |
| Year 2021 | 0.321* (0.055) | 0.317* (0.055) | 0.320* (0.055) |
| Cutpoints | |||
| US very important | US important | 0.967* (0.182) | 0.948* (0.178) | 1.005* (0.186) |
| US important | Equally important | 1.863* (0.184) | 1.842* (0.180) | 1.898* (0.189) |
| Equally important | China important | 2.256* (0.185) | 2.236* (0.181) | 2.292* (0.189) |
| China important | China very important | 3.198* (0.188) | 3.179* (0.184) | 3.235* (0.192) |
| AIC | 6,757.472 | 6,615.102 | 6,661.986 |
| Maximized likelihood R2 | 0.763 | 0.764 | 0.763 |
| Adjusted McFadden R2 | 0.352 | 0.352 | 0.352 |
| n | 2,578 | 2,578 | 2,578 |
| Coefficients | Model 4 | Model 5 | Model 6 |
|---|---|---|---|
| Partisanship (Reference: Pan-Green) | |||
| Independent | 0.569* (0.123) | 0.529* (0.069) | 0.664* (0.164) |
| Pan-Blue | 0.745* (0.073) | 0.718* (0.074) | 0.742* (0.073) |
| Taiwan People’s Party | 0.462* (0.096) | 0.441* (0.097) | 0.461* (0.096) |
| China’s influence | 0.301* (0.021) | 0.288* (0.017) | 0.287* (0.017) |
| Trade vs. Sovereignty | 0.412* (0.049) | 0.499* (0.064) | 0.412* (0.049) |
| Cultural similarity | 0.169* (0.031) | 0.171* (0.031) | 0.205* (0.040) |
| Identity (reference: dual identityand Chinese identity) | |||
| Taiwanese identity | –0.503* (0.053) | –0.503* (0.053) | –0.504* (0.053) |
| Unification | 0.075 (0.019) | 0.074 (0.019) | 0.074 (0.019) |
| Independent × China’s Influence | –0.035 (0.033) | – | – |
| Independent × Trade vs. Sovereignty | –0.197* (0.095) | – | |
| Independent × Culture Similarity | –0.085 (0.062) | ||
| Year 2020 | –0.006 (0.053) | –0.011 (0.053) | –0.006 (0.053) |
| Year 2021 | 0.321* (0.055) | 0.317* (0.055) | 0.320* (0.055) |
| Cutpoints | |||
| US very important | US important | 0.967* (0.182) | 0.948* (0.178) | 1.005* (0.186) |
| US important | Equally important | 1.863* (0.184) | 1.842* (0.180) | 1.898* (0.189) |
| Equally important | China important | 2.256* (0.185) | 2.236* (0.181) | 2.292* (0.189) |
| China important | China very important | 3.198* (0.188) | 3.179* (0.184) | 3.235* (0.192) |
| AIC | 6,757.472 | 6,615.102 | 6,661.986 |
| Maximized likelihood R2 | 0.763 | 0.764 | 0.763 |
| Adjusted McFadden R2 | 0.352 | 0.352 | 0.352 |
| n | 2,578 | 2,578 | 2,578 |
*p < 0.05 (two-tailed test).
Notes: Coefficients of demographic variables are not shown here. The demographic variables are dummy variables for female, 30–39 years old, 40–49 years old, 50–59 years old, 60+ years old, junior high school, senior high school, some college, and university degree. None is significant statistically across the three models. The full results and data are available on request.
Source: Authors.
To compare the effects of each variable across the three models and two dependent variables, we used R’s DAMisc package (Armstrong 2022). This package allows us to calculate the effect of changing each independent variable by one standard deviation (or other amounts) while holding quantitative variables at their median values and qualitative variables at their mode values. Thus we obtained predicted probabilities at the minimum, median, and maximum values of each variable.
Figure 3 presents the predicted probability of preference for national security alignment with China resulting from a one-standard-deviation change in 24 major variables. For instance, in the upper left corner, we can observe the predicted probability at the minimum, median, and maximum values of cultural similarity between Taiwan and China. These predicted probabilities remain relatively stable across the five categories of security alignment and across the three models with different interaction terms. The second column illustrates the predicted probability resulting from a one-standard-deviation change in self-identification as independent. This had a greater impact than change in perceived cultural similarity. Moreover, a one-standard-deviation change in pan-Blue partisanship had a greater effect than a one-standard-deviation change in other parties or independent. To summarize, there is a distinct basis for the alignment preferences of independents, and self-identification as independent had a greater effect on alignment preferences than either cultural similarity or Taiwanese identity.
Predicted probability of national security alignment preference at a one-standard-deviation change in each independent variable.
Source: Authors.
Predicted probability of national security alignment preference at a one-standard-deviation change in each independent variable.
Source: Authors.
In Figure 4, we observe the change in predicted probability of preference for economic development alignment with China for each explanatory variable. The most significant change occurred when respondents identified as pan-Blue. Preference for alignment with China was also associated with independents and TPP partisans. Cultural affinity, perceptions of China’s influence, and preference for trade over sovereignty did not have substantial effects.
Predicted probability of economic development alignment preference at a one-standard-deviation change in each independent variable.
Source: Authors.
Predicted probability of economic development alignment preference at a one-standard-deviation change in each independent variable.
Source: Authors.
H2b said that self-identification as an independent and cross-Strait trade preference, as well as self-identification as an independent and perception of China’s influence, have interaction effects. Model 5 in Table 3 confirms the former (coeff. = –0.197, s.e. = 0.095). As Model 5 is a nonlinear model, the magnitude and statistical significance of the interaction depend on the covariates in the model (Ai and Norton 2003, 129). To further illustrate the effect of the interaction between trade on independents versus partisans, we created a graph based on Model 5 in Table 3, using the ggeffects package in R (Lüdecke 2018). This graph (Figure 5) computes predicted probabilities while holding other variables constant for each category of the dependent variable, based on the interaction between independent and preference for trade over sovereignty.
Interaction between the Independent and Trade/Sovereignty Attitude variables.
Source: Authors.
Interaction between the Independent and Trade/Sovereignty Attitude variables.
Source: Authors.
Prioritizing trade over sovereignty had a notable impact on the probability of independents choosing “China is important” or “China is very important,” while decreasing the perceived importance of the US. For instance, among those who prioritized trade over sovereignty, the predicted probability of independents choosing “China is very important” was 0.16, whereas for partisans, it was 0.09. This implies that as independents become more aware of the significance of cross-Strait trade, they are more likely to favor deeper economic integration with China.
Conclusion
Independents in Taiwan are more likely to base their foreign policy preferences on economic considerations rather than cultural identity. This means their behavior cannot easily be explained by the theories typically used to account for the preferences of party loyalists. In other words, the characteristics of independents open the door to explanations (e.g., folk realism) that may be inconsistent with explanations for party loyalists (e.g., social identity theory). In the case of Taiwan, because economic development is a key valence issue, independents use cross-Strait trade as an information shortcut.
We used descriptive statistics, ordered probit regression models, and plots of predicted probability to show that independents in Taiwan tend to align with China for both national security and economic development, paralleling the preferences of pan-Blue supporters. Furthermore, self-identification as independent and support for cross-Strait trade had a significant interaction effect on preference for economic alignment with China. On the other hand, factors such as identity and perceptions of China’s relative influence did not affect independents’ alignment preferences.
These findings add to the body of knowledge on public opinion. One significant observation is that self-identified independents in Taiwan have foreign policy attitudes comparable to pan-Blue partisans. However, consistent with Converse’s argument (1964), independents are not constrained by abstract principles or ideology. The stability of independents’ issue positions and their responsiveness to new information acquired over time remain areas for further exploration (Green, Palmquist, and Schickler 2002). If their issue positions remain stable over time, the argument that actively engaged independents may not rely on party cues or ideology to guide their decision-making process becomes more plausible.
The second theoretical implication underscores the comparable influence of identity and self-interest on foreign policy attitudes, supporting the propositions of Lau, Brown, and Sears (1978). This is consistent with Berinsky’s (2009) argument that public opinion on both foreign and domestic policy may arise from a common attitude structure. Given the substantial importance of national identity in Taiwan, it is not surprising to find that it has a substantial influence on foreign policy attitudes.
By integrating insights from public opinion research and international relations theory, our study suggests new ways to investigate foreign policy orientation in Taiwan and beyond. Most recent research on foreign policy attitudes in Taiwan has focused on information, strategic considerations, economic coercion, and the US–China–Taiwan relationship, often treating party attachment and nonpartisanship as demographic variables. In contrast, our study explores the foreign policy attitudes of independents, who are not constrained by partisan affiliation. This enables us to better understand the formation of and influences on public opinion on foreign policy.
We should acknowledge that we have not established a causal link between general attitudes and specific policy positions. Survey experiments could help uncover the underlying mechanisms linking these aspects. By providing different cues and stimuli about international affairs, survey experiments can test how individuals respond and make decisions in specific scenarios. Cross-national survey experiments would be especially useful in testing the generalizability of the hypotheses we have tested here in the Taiwan case.
Published online: January 23, 2024
Acknowledgments
An earlier draft of this paper was presented at the 2022 annual conference of the American Political Science Association in Montréal, Québec, Canada. We are grateful for Austin Horng-en Wang’s comments, and to Hui-ling Chen for her superb research assistance. We also extend our thanks to the four anonymous reviewers. All of the survey data for this article were collected by the Election Study Center, National Chengchi University. Data collection was sponsored by the National Science and Technology Council, Taiwan.
Notes
Classical, neoclassical, and “folk” realism all theorize about the preferences of leaders and the public (Kertzer and McGraw 2012; Kirshner 2022). Structural realists have claimed that theirs is a theory not of foreign policy preferences but of general patterns in state behavior (e.g., balancing versus bandwagoning). However, in a world where great powers dominate (an oligopolistic system), an obvious empirical implication of realist theory is that decision-makers must actively choose whether to balance against or bandwagon with these powers. In other words, embedded in structural realism is a theory of foreign policy preferences at the group and individual levels (Elman 1996).
We use DPP partisans as the baseline category because the DPP has been in power since 2016. Thus we can observe how KMT partisans and independents react to the DPP’s foreign policy by comparing them with DPP partisans.
Pan-Blue refers to a group of parties associated with the KMT and is named for the predominant blue color of the KMT flag. Pan-Green consists of parties led by the DPP that advocate Taiwan’s independence and a distinct Taiwanese identity. The TPP was established in 2019 by former physician Ko Wen-je, who also served as mayor of Taipei between 2014 and 2022. The TPP emphasizes clean politics and pushes for greater dialogue between Taiwan and China.
Batto and Cox (2016, 12) argue that political parties are motivated to coordinate their efforts to win executive office. This dynamic leads to the bipolarization of politics into the Blue camp versus the Green camp, paralleling the quadrille bipolaire in France’s Fifth Republic. However, Chen (2021) suggests that the Blue–Green cleavage and ideological landscape might be undergoing restructuring. As recently as 2017, most of Taiwan’s parties were still primarily organized along the one-dimensional independence/unification issue (Wang 2019). Despite the rise of the NPP and TPP as the third-largest party in 2016 and 2020, respectively, no new cleavages have yet emerged. Therefore, we argue that the foreign policy stances of partisans are predominantly influenced by the independence/unification issue.
Hedging is sometimes seen as a third option. In the absence of an immediate threat, ideological confrontation, or zero-sum competition between great powers, hedging emerges as an alternative. This strategy encompasses risk management (Wu 2016), engagement, and limited bandwagoning, along with other non-balancing behaviors (Kuik 2008). Hughes (2016) argued that states may choose balancing if conditions deteriorate but opt for hedging otherwise. Given the increasing rivalry between China and the US and the perceived imminent threat from China, the conditions for hedging seem less viable for Taiwan. However, we believe that both bandwagoning and balancing include elements of limited hedging, because neither precludes engagement or future cooperation. For example, Chong, Huang, and Wu (2023, 3) defined hedging as “a preference to cooperate with all major power actors to open options, preserving possibilities for isolation, bandwagoning, and balancing.” Wang and Tan (2021, 68) define hedging as “a hybrid strategy of combining both balancing and bandwagoning by which a country counters the threatening power through strengthening its own defense and/or forming an alliance with other countries while at the same time engaging with the menacing state.” Roy (2005, 306) succinctly pointed out that “hedging implies a present condition of strategic uncertainty. Balancing is sometimes the response, for example, to a certain and compelling threat, in which case hedging is not involved.”
In 2022, China received 25% of Taiwan’s exports (Economist 2023). Taiwan had a trade surplus of USD 100.4 billion with China and Hong Kong, compared to USD 29.4 billion with the US.
The use of landline phone numbers may reduce the representation of younger respondents, who mostly use cellular phones. Despite our efforts to mitigate age-related bias through household sampling and weighting, weighting could exaggerate the impact of respondents with specific demographic characteristics. It is essential to keep this in mind when interpreting the effects of age on the outcome variables.
We analyzed the three data sets with the same model separately and obtained similar results. Therefore, we only report the results from the combined data.
Scholars such as Grieco (1988) and Powell (1991) have discussed the concepts of absolute and relative gains. Grieco argued that states pay attention to both absolute gains and the gains achieved by others. Powell introduced the concept of relative loss, suggesting that states may cooperate if relative loss is costly. The intensifying competition between the US and China exemplifies this, as each side strives to prevent the other from securing an advantage. As a smaller actor, Taiwan’s focus is on absolute gains. Therefore, we hypothesize that individuals are sensitive to absolute gains when choosing between the US and China. Unfortunately, our data do not enable us to assess how cross-Strait trade affects individual welfare. While members of some groups may capture a larger share of benefits, others could be at a disadvantage. A definitive account of these disparities requires further research.
No variable had a variance inflation factor greater than 5, except independent and its interaction terms. Therefore, multicollinearity is not a concern.
We further analyzed the data by changing the reference group to pan-Blue and TPP for both security and economic alignment preferences. In both scenarios, independents differed from pan-Blues when the interaction between independents and trade was included. However, we did not find a significant difference between independents and TPP supporters. This is perhaps due to the small sample of TPP supporters (198), resulting in a large standard error for the coefficient. But note that the surveys were conducted before the TPP emerged as a high-profile political party.
References
Appendix
Descriptive Statistics of Variables
| Variable | Mean | SD | Min. | Max. |
|---|---|---|---|---|
| Independent | 0.492 | 0.529 | 0 | 1 |
| Pan-Blue | 1.209 | 0.407 | 0 | 1 |
| Pan-Green | 1.294 | 0.456 | 0 | 1 |
| TPP | 0.229 | 1.410 | 0 | 1 |
| Influence of China | 3.334 | 1.441 | 1 | 5 |
| Trade vs. sovereignty | 1.437 | 0.496 | 0 | 1 |
| Culture similarity | 2.536 | 0.777 | 1 | 4 |
| Taiwanese Identity | 0.587 | 0.492 | 0 | 1 |
| Independence / Unification | 1.410 | 0.492 | 1 | 6 |
| Female | 0.529 | 0.499 | 0 | 1 |
| 30–40 years old | 0.130 | 0.336 | 0 | 1 |
| 40–50 years old | 0.213 | 0.409 | 0 | 1 |
| 50–60 years old | 0.217 | 0.412 | 0 | 1 |
| 60+ years old | 0.323 | 0.468 | 0 | 1 |
| Middle school | 0.073 | 0.260 | 0 | 1 |
| High school | 0.272 | 0.445 | 0 | 1 |
| Some college | 0.142 | 0.349 | 0 | 1 |
| University | 0.409 | 0.492 | 0 | 1 |
| Alignment preference on national security | 2.735 | 1.445 | 1 (US) | 5 (China) |
| Alignment preference on economic development | 2.899 | 1.438 | 1 (US) | 5 (China) |
| Variable | Mean | SD | Min. | Max. |
|---|---|---|---|---|
| Independent | 0.492 | 0.529 | 0 | 1 |
| Pan-Blue | 1.209 | 0.407 | 0 | 1 |
| Pan-Green | 1.294 | 0.456 | 0 | 1 |
| TPP | 0.229 | 1.410 | 0 | 1 |
| Influence of China | 3.334 | 1.441 | 1 | 5 |
| Trade vs. sovereignty | 1.437 | 0.496 | 0 | 1 |
| Culture similarity | 2.536 | 0.777 | 1 | 4 |
| Taiwanese Identity | 0.587 | 0.492 | 0 | 1 |
| Independence / Unification | 1.410 | 0.492 | 1 | 6 |
| Female | 0.529 | 0.499 | 0 | 1 |
| 30–40 years old | 0.130 | 0.336 | 0 | 1 |
| 40–50 years old | 0.213 | 0.409 | 0 | 1 |
| 50–60 years old | 0.217 | 0.412 | 0 | 1 |
| 60+ years old | 0.323 | 0.468 | 0 | 1 |
| Middle school | 0.073 | 0.260 | 0 | 1 |
| High school | 0.272 | 0.445 | 0 | 1 |
| Some college | 0.142 | 0.349 | 0 | 1 |
| University | 0.409 | 0.492 | 0 | 1 |
| Alignment preference on national security | 2.735 | 1.445 | 1 (US) | 5 (China) |
| Alignment preference on economic development | 2.899 | 1.438 | 1 (US) | 5 (China) |
University of California Press