Taiwan and the “One-China Principle” in the Age of COVID-19: Assessing the Determinants and Limits of Chinese Influence

Abstract During the current global COVID-19 crisis Taiwan has portrayed itself as both an example for other countries to follow and as a country willing to assist others in their own efforts with the virus. Taiwan has also renewed efforts to participate in the World Health Organization (WHO), an organisation from which it is currently excluded. Although some countries have supported Taiwan's efforts to participate in the WHO or have praised its COVID-19 response, others have been silent or even critical, sometimes citing commitments to a “one China policy.” In this paper, we use newly collected data to explore cross-national variation in support for Taiwan during the current pandemic. We find that a country's level of economic development and security ties with the US are strongly correlated with support for Taiwan while a country's economic ties to China is a less consistent predictor.

Introduction vicinity in early 1996. While at Cornell, Lee delivered a lengthy speech in which he outlined Taiwan's economic achievements and democratization, and highlighted the disconnect between these achievements on the one hand and Taiwan's growing international marginalization on the other. At the time, Taiwan had already been excluded from the United Nations for nearly a quarter of a century, and only 29 countries maintained formal diplomatic ties with the Republic of China (ROC). Pointing to Taiwan's economic and political triumphs, Lee argued in his speech that the "Republic of China on Taiwan does not enjoy the diplomatic recognition that is due from the international community," that Taiwan's people were "not happy with the status accorded our nation by the international community," and that Taiwan was "determined to play a peaceful and constructive role among the family of nations." 1 In the 25 years since Lee's speech, and 50 years after the "Shanghai Communiqué," it would appear on the surface that Taiwan has only become more marginalized internationally. The number of countries maintaining diplomatic ties with the ROC has continued to dwindle, and Taiwan remains excluded from the United Nations and most of its affiliated organizations. The détente of the Ma Ying-jeou 馬英九 years (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) brought some respite, including a tacit diplomatic truce and observer status in the World Health Organization's (WHO) World Health Assembly. But since the election of Tsai Ing-wen 蔡英 文 as Taiwan's president in 2016and her refusal to accept any version of a "one-China principle"the PRC has ratcheted up the pressure on the island, both militarily and diplomatically. Beijing again sought to block Taiwan's participation in the WHO, and the PRC picked off several more diplomatic allies of the ROC; today only 14 countries maintain formal diplomatic ties with Taipei. Taiwan, it seems, is more boxed in than ever.
Beneath the surface, however, is considerable variation in how countries, even among those without formal diplomatic ties with the ROC, effectively interpret and operationalize their relationships with Taipei. As Liff and Lin point out in this special section, although the PRC today actively promotes a "myth of consensus," there is in fact an important distinction to be made between Beijing's self-asserted "one-China principle," which claims that Taiwan is an inseparable part of the PRC, and other countries' "one China" policies, regarding which there is considerable variation globally. 2 Some governments adopt Beijing's official position verbatim; others adopt far more ambiguous positions. Official positions and effective policies not only vary across cases but also within them (i.e. over time). The US's ambiguous 1979 position on Taiwan's status and its forward-leaning support of Taiwan in terms of policy is the most famous example of the distinction between Beijing's "one-China principle" and many other countries' effective "one China policies." 3 Indeed, some describe current US-Taiwan relations as being "better than at any time before," 4 reflecting in part recent decisions in Washington to sell advanced fighter jets to Taiwan, undertake frequent freedom of navigation patrols through the Taiwan Strait and send highlevel officials, including the secretary of health and human services, to Taiwan. As Liff and Brown observe in this special section, many other countriesespecially US democratic allieshave similarly vague (and fluid) policies towards Taiwan. 5 More broadly, even as most countries have severed formal diplomatic ties with Taiwan as part of their "one China policies," many of these countries have nevertheless maintained considerable informal ties with the island. And some countries have at times been willing to risk Beijing's ire by expressing support for Taiwan in particular contexts. For instance, when the PRC passed the 2005 Anti-secession Law that euphemistically threatened the use of "non-peaceful means" in the event of Taiwan's "secession" from China, several countriesincluding the US and Japancriticized the law for raising tensions in the Taiwan Strait. 6 In 2021, Lithuania angered the PRC by agreeing with Taipei to establish a representative office that would include "Taiwan" in its name. 7 Cross-national variation in support for Taiwan has also been evident during the global COVID-19 crisis. Throughout the crisis, Taiwan portrayed itself as both an example for other countries to follow and as a country willing to assist others in their own efforts with the virus. Taiwan also renewed efforts to participate in the WHO. Although many countries have been silent or even critical of these efforts, sometimes citing commitments to a "one China policy," others have supported Taiwan's push to participate in the WHO or have praised its COVID-19 response.
How can we explain cross-national variation in the willingness of countries to express support for Taiwan? In this article, we use newly collected data to explore cross-national variation in support for Taiwan during the COVID-19 pandemic. More specifically, we examine which factors determine whether countries advocated for Taiwan's participation in the WHO or praised Taiwan's COVID-19 response. To preview: we find that a country's level of economic development and security ties with the US are the most consistent predictors of support for Taiwan, whereas economic ties to China are inconsistent predictors. 3 Romberg 2003. 4 These were the words of Taiwan's foreign minister, Joseph Wu, in 2019. See "Foreign minister: Taiwan-US relations probably at their best ever," VOA News, 30 August 2019, https://www.voanews.com/a/usa_ foreign-minister-taiwan-us-relations-probably-their-best-ever/6174867.html. Accessed 21 April 2022. 5 Liff, this issue; Brown, this issue. 6 "US, Japan criticize 'anti-secession' law for raising tensions," Taipei

Theory and Hypotheses
Why would countries express support for Taiwan when doing so will likely trigger Beijing's ire? What explains cross-national variation in this regard? In this section, we develop three separate causal logics that could help to explain why some countries are more willing to express open support for Taiwan than others. First, China's rise as an economic power has corresponded to a rapidly growing global economic footprint. In turn, countries with extensive economic ties to China might be less likely to risk Beijing's ire by expressing open support for Taiwan, fearing that doing so could put those ties in jeopardy. Second, Taiwan has become a vibrant democracy in recent decades, and the island has become an advanced economy. It may be, in turn, that more democratic and developed countries simply identify more with Taiwan and are thus more likely to show support for Taipei. Finally, the US is Taiwan's primary security partner; countries with closer security ties with Washington may be more likely to follow the US lead in expressing support for Taiwan.

China's economic influence
After more than four decades of rapid growth and economic transformation, China has emerged as a global economic powerhouse. Today China is the world's second-largest economy, the largest trading state, and an important source of foreign direct investment and foreign aid. Against this backdrop, an item of conventional wisdom has taken hold that China's burgeoning foreign economic ties have become a major source of the PRC's growing international influence. Indeed, China is seen as using economic ties intentionally to strengthen its international influence, i.e. as a tool of economic statecraft rather than merely supporting commercial activities. According to this narrative, foreign governments, wary of alienating an important economic partner, are more reluctant to challenge China on issues ranging from human rights to maritime disputes in the South China Sea; companies that do business in China avoid actions that might trigger PRC boycotts; and foreign business groups lobby for policies likely to keep relations with Beijing on an even keel. The PRC's foreign economic ties might translate into influence abroad via at least three broad causal mechanisms. First, economic ties can be a source of direct bargaining power, giving Beijing more tools through which to reward compliance or punish non-compliance with China's preferred policies. As countries become more economically tied to China, they potentially become more vulnerable to the PRC's coercive economic statecraft, wherein Beijing either threatens or imposes economic sanctions as a way to shape target behaviour. Economic ties also give Beijing the opportunity to pursue inducement strategies, or "carrots," in which favourable target behaviour is rewarded in anticipation of or subsequent to favourable actions. 8 Second, economic ties can generate interest groups in countries that come to depend on China economically, in what Kirshner has called a "Hirschmanesque" effect. 9 In countries with extensive economic ties with China, a coalition of economic interestssuch as businesses that export to or invest in China, or that rely on Chinese products, investment or aidshould strongly prefer stable and peaceful relations with China. 10 For these actors, political turbulence with China could be bad for business. As the size and wealth of this coalition grows, in turn, it should become more able to influence policy debates relating to Chinain effect steering policy to be more accommodating towards China. 11 Finally, economic ties with the PRC also can have a transformative effect on how even non-vested actors might view their interests pertaining to China, leading to greater identification with China and Chinese interests. For instance, through explicit or implicit threats to market access, individuals (e.g. pop performers 12 ) and firms may be incentivized to portray China in a favourable light and to avoid focus on issues, such as human rights, that might damage views of the PRC. Similar effects could occur via media that are owned by Chinese firms or firms with a stake in China. 13 In turn, if these sorts of processes lead to more favourable public opinion towards China, a country's political leaders may be more inclined to seek favourable relations with Beijing (or, at least, may be less likely to be punished for doing so).
Each of these causal processes point in the direction of our initial hypothesis: Hypothesis 1: As a country's economic ties with China increase, it will be less likely to express support for Taiwan.

Political affinity with Taiwan
The choice to express support for Taiwan likely also reflects some level of political affinity with Taipei. After decades of martial law and highly authoritarian rule, Taiwan underwent a transition to democracy in the late 1980s and 1990s. Since the ROC's first direct presidential election in 1996, the island has accomplished three peaceful transfers of power. Today, Taiwan scores a 94 out of 100 on Freedom House's "Freedom in the World" score, higher than the US, and well above the threshold to be classified as a "free" society. 14 Meanwhile, Taiwan's economy enjoyed rapid and sustained growth in the decades after 1960, and the island today is considered to be an advanced economy by the International Monetary Fund. 15 Political leaders in other countries, in turn, sometimes cite Taiwan's political and economic accomplishments when expressing support for the island. For instance, the co-chairs of the US Congressional Taiwan Caucus praised Taiwan as a "beacon of democracy" when congratulating Tsai Ing-wen after her 2020 re-election as Taiwan's president. 16 In a speech on Taiwan given later that year, the then US assistant secretary of state for East Asian and Pacific affairs, David Stilwell, emphasized Taiwan's role as a "vibrant democracy" and "global economic engine" in making the case for closer US-Taiwan ties. 17 And President of the Senate of the Czech Republic Miloš Vystrčil in 2020 travelled to Taiwan where he highlighted shared democratic values between the two countries. 18 We expect that these sentiments likely generalize to some degree, such that more democratic and developed countries will be more willing to express support for Taiwan, on average.
Hypothesis 2: Countries that are more democratic are more likely to express support for Taiwan.
Hypothesis 3: Countries that are more economically developed are more likely to express support for Taiwan.

Security ties with the United States
Finally, we expect that a country's security ties with Washington are likely to influence that country's approach to Taiwan. Although US commitments to Taiwan have been both informal and to some degree ambiguous since the US abrogated its alliance treaty with the ROC after recognizing Beijing in 1979, it is nevertheless the case that the US maintains extensive security ties to Taipei, and the US has signalled considerable interest in Taiwan's security. Indeed, US interest in Taiwan's security appears to be increasing in recent years. In addition to the arms sales and naval transits noted earlier, the US engages in extensive security dialogue with Taiwan, and as Glaser, Bush and Green note, "the mechanisms and channels used to manage the relationship have not evolved considerably in recent decades, but the breadth of issues addressed within them has broadened by orders of magnitude." 19 Given US security interests in Taiwan, we expect US security partners to be, on balance, more supportive of Taiwan than other states.
Hypothesis 4: Countries that have close security ties with the United States are more likely to express support for Taiwan.

Research Design
A key challenge to testing our hypotheses quantitatively centres on measuring our dependent variable: country-level support for Taiwan. Simply put, it is difficult to code with confidence the degree to which individual countries are more or less supportive of Taiwan across a large sample of countries. However, the COVID-19 pandemic offers a useful opportunity in this regard. Even though the PRC views Taiwan as an important national interest and as rightfully part of China, a number of countries have been willing to risk angering Beijing by expressing support for Taiwan during the pandemic. In this section, we elaborate on how we plan to test our hypotheses using new data.

The dependent variable: support for Taiwan during the COVID-19 pandemic
Background: Taiwan as a "core" national interest. Chinese leaders have framed Taiwan as a "core" national interest and have viewed unification with the island as an important objective since the Kuomintang (KMT)along with their government, the ROCretreated to Taiwan at the end of the Chinese Civil War in 1949. For decades after 1949, both the ROC and PRC governments claimed to be the legitimate government of "China," and both viewed Taiwan as unambiguously a part of "China." The basic PRC position remains the same today: Taiwan is rightfully a part of "China," the PRC is the sole legitimate government of "China," and Taiwan must ultimately be reunified with the remainder of the country. In Taiwan, on the other hand, the island's sovereign status became contested politically after martial law was lifted and Taiwan democratized during the late 1980s and 1990s. In recent years, most of Taiwan's citizens have not selfidentified as Chinese, and there is little interest in unification with the PRC. Against this broader backdrop, Taiwan has faced considerable international marginalizationwhich has grown more pronounced as China's international power continues to grow. The number of countries maintaining diplomatic ties with the island has dwindled to 14 (the PRC requires that countries sever diplomatic ties with Taiwan if they wish to establish ties with Beijing), and Taiwan is excluded from most international organizations. A brief respite from these broader trends occurred during the Ma Ying-jeou administration. Unlike his immediate predecessors, Ma was willing to endorse a "one-China principle" (albeit in weak form via the "1992 Consensus"), opening the door for increased dialogue with Beijing and a tacit diplomatic truce. The number of countries recognizing the ROC remained stable for most of Ma's time in office, and Taiwan was able to participate in a limited capacity in some international organizations, such as the WHO. However, Ma's KMT was routed in 2016 elections, and his successor as Taiwan's president, Tsai Ing-wen of the Democratic Progressive Party (DPP), has refused to endorse a "one-China principle." 20 Since 2016, cross-Strait relations have deteriorated, and the PRC has ratcheted up coercive pressure against Taiwan. Beijing ended all cross-Strait dialogue, and the PRC has frequently dispatched military aircraft into airspace near Taiwan. China has again blocked Taiwan's participation in international forums such as the WHO, and has convinced several of Taiwan's diplomatic allies to break ties with the ROC and recognize the PRC instead. In early 2020, Tsai was re-elected in a landslide.
Coding cross-national variation during the COVID-19 crisis. The global COVID-19 crisis, then, has occurred in an environment where Taiwan has been excluded from the main international organization dealing with the crisis (WHO), and where Beijing pressures other countries to limit recognition of Taiwan and encourages them to endorse the PRC position that Taiwan is a part of China. Taiwanese officials, on the other hand, have argued that the crisis demonstrates the importance of Taiwan's participation in the WHO, and the Taiwan government has sought to contribute to the global supply of public health goods during the crisis byamong other thingssending large quantities of masks abroad. The PRC, for its part, has criticized these efforts as seeking "independence under the pretext of the pandemic." 21 Despite Beijing's condemnations, however, a number of countries have responded favourably to Taiwan's overtures. Some countries have openly supported Taiwanese participation in the WHO in some capacity. 22 For instance, the US Department of Health and Human Services released a statement expressing disappointment over Taiwan's exclusion from the WHO, emphasizing that "Taiwan must be welcomed to contribute to, not be excluded from, WHO meetings." 23  that "it is necessary for Taiwan to join the WHO." 24 Many additional countries have praised Taiwan's pandemic response and its efforts to assist other countries, or have publicly thanked Taiwan for its willingness to provide aid. Numerous other countries, however, were silent on the issue, and a few countries publicly opposed Taiwan's participation in the WHO.
We have collected data on each UN member country's public response to Taiwan's COVID-19 diplomacy from March to October 2020. To do so, multiple coders undertook country-by-country English-and Chinese-language searches using internet search engines such as Google. The coders also undertook searches in NexisUni, on social media such as Twitter, and Taiwan's Ministry of Foreign Affairs webpage. Each coder used multiple search terms, such as "Taiwan" and "France" and "coronavirus" or "Taiwan" and "France" and "WHO." We coded two specific features of each country's response. First, we coded whether a country explicitly indicated that it supported Taiwan's participation in the WHO in some capacity. Second, we coded whether a country publicly praised Taiwan's "mask diplomacy" (such as by publicly thanking Taiwan for shipping masks) or its general management of the COVID-19 pandemic. For each country, we only coded official statements coming from the executive function of a government (for instance, statements by individual legislators were not counted). 25 Altogether, we coded 26 countries as supporting Taiwan's participation in the WHO, and 45 countries as praising Taiwan's COVID-19 response or COVID-19 diplomacy in some way. Table 1 lists each of these countries.
Note that nearly all 26 countries in the left-hand column also appear in the right-hand column, meaning that nearly all countries that supported Taiwan's participation in the WHO also expressed support for Taiwan in other ways. Only a few countries publicly expressed some opposition to Taiwan's participation in the WHO or otherwise criticized Taiwan's pandemic response. 26 As such, we do not include such actions as a separate category in our analysis. Democracy. We use Freedom House's 2019 Freedom in the World scores to measure a country's level of democracy (Democracy). The scores are an index, such that the closer a country's score is to 100, the more democratic the country is. 31

A note on sample and method
We coded a dependent variable that has three values: no support for Taiwan on its COVID-19 response/diplomacy or WHO participation (zero), praising Taiwan's COVID-19 response/diplomacy, but not advocating participation in the WHO (one), and advocating Taiwan's participation in the WHO (two).
Since advocating for participation in the WHO is a stronger indicator of support than simply praising Taiwan's COVID-19 diplomacy, we utilize an ordered logit model. Our initial sample includes 191 UN member countries. 33 We also run a separate model where we exclude countries that maintain diplomatic ties with the ROC government. 34 All regressions were estimated using R. Summary statistics for each variable used in our analysisfor the 191 countries in our final sampleare provided in Table 2. A correlation matrix of all variables used in the analysis is included in the appendix (Table A4). 2022. Note that we code this variable as equal to one for the United States. 33 We exclude China since some of our independent variables code economic ties with China. The Marshall Islands was excluded due to economic data discrepancies. Note that one other country recognizing Taiwan, the Vatican, is not a UN member and so is also excluded from the analysis. 34 All ROC-recognizers in our sample expressed support for Taiwan's participation in the WHO, so we cannot include a control variable for ROC-recognizing countries. Although we run the regression both ways (with and without the ROC-recognizers), our own view is that it makes most sense to keep ROC-recognizers in the sample. To see why, consider an analogy. Suppose we want to understand who supports a particular bill that restricts abortions after the first trimester. Suppose our sample includes members of a pro-life group that opposes abortion in all circumstances; all members of this group also support the bill that interests us. We can't include a control for membership in the group (since there is no variation among group members concerning the bill in question), and it would make little sense to exclude members from the sample (since doing so would remove from the sample a set of people whose behaviour we are interested in explainingthose who are most strongly prolife). In our case, removing ROC-recognizers means taking out of the sample the set of countries that is most ardently pro-ROC, precisely the sort of behaviour we seek to explain. Table 3 displays the results of our ordered logit regressions. Model 1 includes the full sample of 191 countries, while Model 2 excludes countries recognizing the ROC. Support for our hypothesis on the effect of economic ties (Hypothesis 1) is mixed. Two measures of economic ties with China, Chinese investment (log) and BRI, are significantly and negatively correlated with support for Taiwan in Model 1 (as Hypothesis 1 anticipates), but neither variable is significant in Model 2. And although Trade with China (log) is a significant predictor in both models, it has a positive coefficient (contrary to the expectations of Hypothesis 1). Figures 1A-1C show the substantive effects of the variables measuring economic ties to China. 35 As shown in Figure 1A, as trade with China as a percentage of GDP increases (from one standard deviation below the mean to one standard deviation above), the likelihood that a country stays silent declines even as the likelihood of supporting Taiwan's entry into the WHO increases markedly. Again, this is contrary to expectations and possibly a spurious finding. 36 Figure 1B shows that Chinese investment in a country has more modest effects: as investment increases, the likelihood that a country stays silent also increases while the likelihood of praising Taiwan's pandemic response or supporting Taiwan's participation in the WHO decline. And Figure 1C shows quite large substantive effects for BRI membership (where membership increases the likelihood of staying silent by 32 per cent and decreases the likelihood of expressed support for Taiwan's participation in the WHO by 24 per cent). Turning to Hypothesis 2, although Democracy is positively correlated with support for Taiwan in both models, it is not statistically significant. As shown in Figure 1D, its substantive effect is also small. On the other hand, we find 35 We use an observed-value approach in calculating substantive effects of the independent variables. Put simply, we manipulate the change of the variable of primary interest (e.g. from the mean to one standard deviation above the mean) and hold all the other independent variables at their observed values in each case (i.e. country). We then calculate the predicted probabilities of each country, and then take their average. For a detailed theoretical justification of the observed-value approach in interpreting substantive significance of logit models, especially its advantage over the average-case approach, see Hanmer and Kalkan 2013. 36 Although the finding may be spurious, we note that it is consistent with conclusions drawn by Morgan 2019. In a study of African perceptions of China's economic engagements, she finds that negative perceptions of China are primarily linked to trade-related interactions, whereas foreign direct investment or foreign aid are tied to more positive assessments of China. strong support for Hypothesis 3: Log PCI is highly significant in both Models 1 and 2, and Figure 1E suggests the substantive effects are also quite substantial. For instance, a country with per capita income one standard deviation above the mean was roughly twice as likely to support Taiwan's participation in the WHO as a country with per capita income one standard deviation below the mean. Finally, we also find support for Hypothesis 4: US Alliance is weakly significant in Model 1 and strongly significant in Model 2, although in substantive terms its effect is somewhat modest (see Figure 1F).

Alternative Specifications
We adopted several alternative specifications of our model to probe the robustness of our findings. To begin, the ordered logit models used above are essentially a group of binary logistic models 37 with the assumption that the slope coefficients remain the same while the intercepts change across regressions. Although the assumption is rarely examined in practise, it is possible that some independent variables can increase the probability of being in the second category relative to the first and yet have a null relationship at other levels. 38 We therefore turn to a more flexible approach, estimating those binary logistic models directly. First, because relatively few non-ROC-recognizer states supported Taiwan's participation in the WHO, we created a dichotomous version of the dependent variable where countries that either praised Taiwan's COVID-19 response or supported Taiwan's participation in the WHO are coded one, and other countries are coded zero. Table A1 in the appendix reports binomial logit results using this dependent variable. Findings are similar to those reported above, although the Notes: *p < 0.1; **p < 0.05; ***p < 0.01. Model (1) uses the full sample. Model (2) restricts the sample to non-ROC recognizers.
37 In this case, the ordered logit model contains two binary logistic models, with the first one estimating the probability of being in the "Participate in the WHO" group relative to the other two and the second one estimating the probability of being in the "Participate in the WHO" or the "Public praise" group relative to the other. 38 Ward and Ahlquist 2018. significance level of US alliance and Chinese investment (log) drop slightly in the full-sample model.
Second, because we regard supporting Taiwan's participation in the WHO as a much stronger signal of support than praising its COVID-19 response, we also created a dichotomous variable equal to one only for those countries that supported Taiwan's WHO participation, and zero for all other countries.  Table 3. Confidence intervals are calculated from 1,000 simulations. Table A2 in the appendix reports binomial logit results using this dependent variable. Here, the alliance variable is significant in both the full-sample model and the model that excludes ROC-recognizers, and Chinese investment and BRI participation remain strong predictors in the full sample model only. However, in this specification the per capita income variable is no longer significant in either model (though its coefficient remains positive).
Finally, although we have no prior expectations of possible confounding variables, we ran separate models that include several control variables that could be correlated with support for Taiwan and other independent variables, including the log value of a country's population ( population, log), the amount that a country spends on its military (military expenditure), and whether a country is located in Asia (Asia). 39 We speculate that countries with larger populations and military expenditures may, on balance, be more likely to support Taiwan since they have more ability to resist PRC influence. Countries located in Asia might be less supportive of Taiwan because they are more susceptible to China's influence. Data on military expenditures are limited to 155 countries in our sample, so we run two separate ordered logit regressions, reported in Table A2 in the appendix. In the first, only population (log) and Asia are included as controls, and in the second all three control variables are included. Results closely parallel findings reported in Table 3 above, although the significance level of US Alliance drops slightly (and below the 90 per cent threshold).

Conclusion
Taiwan, in many ways, remains trapped in what Shelley Rigger has called an "international birdcage." 40 The number of countries maintaining diplomatic ties with the ROC today stands at only 14. At the behest of the PRC, Taiwan remains excluded from most major international organizations, 41 and since 2017 Beijing has blocked Taiwan from even observer status in the WHO's World Health Assembly. The only country that continues to sell arms in significant quantity to Taiwan is the US. 42 Yet this apparent marginalization belies considerable variation in how even countries that have severed all formal ties with the ROC approach their relationship with Taiwan. Other studies in this special section highlight interesting variation over time in Japan, the US and the European Union. In this article, we show as well that there exists considerable cross-national variation in how countries across the world have approached Taiwan during the COVID-19 pandemic, and the degree to which they have been willing to publicly support Taiwan. To make sense of this variation, we proposed three broad explanations, which centred on a country's economic ties to China (Hypothesis 1), its affinity to Taiwan (Hypothesis 2, Hypothesis 3) and its security ties with the US (Hypothesis 4). We found only mixed support for our first hypothesis. While Chinese investment and participation in the BRI were both correlated with reduced support for Taiwan in our full sample, the effect of these variables appeared quite sensitive to the sample of countries used (and disappears in Model 2, which excludes countries recognizing the ROC). Moreover, trade with China is actually positively correlated with support for Taiwan, in contrast to the expectations of Hypothesis 1. These findings are consistent with earlier studies that found only limited or conditional influence of economic ties with China on countries' foreign policy stances on political issues. 43 We also found little support for our second hypothesis: democracy was generally not a statistically significant predictor of support for Taiwan. On the other hand, we found stronger support for our third and fourth hypotheses. A country's level of economic developmentwhich we measured using per capita incomehad a large and significant impact on support for Taiwan in our main findings and most of our robustness tests, and a country's security ties to the US was also a significant predictor in most of our models.
Given the strong influence of Beijing's preferences on Taiwan's international standing, it is interesting to consider why so many countries with which China has strong tiesagain, with the noted exceptionsremained silent rather than actively voicing opposition to Taiwan's mask diplomacy or WHO participation. Our thoughts here are highly speculative. We note that, despite Beijing's often assertive behaviour to diminish Taiwan's independent international standing, in the COVID-19 era Beijing's efforts to signal to or mobilize other countries were relatively tepid. 44 The silence of many countries could therefore reflect a lack of at least overt pressure from China. Alternatively, host countriesdealing with the impact of COVID-19 in their own countriesmay be distracted, or know that they can rely on China to handle the matter unilaterally. Internal forces that might otherwise lobby in China's interests also may be otherwise distracted. Finally, we speculate that, while Taiwan's international status is a highprofile issue for Beijing, it is of low salience for other countries, such that when not pressed to respond by China (or the US for that matter), countries may choose to remain silent. This study has obvious shortcomings. Though it was well timed to capture controversy over an issue of importance to China -Taiwan's international statusour data collection covered only a short time period. Given that governments had much to deal with during this period, and given the low salience of Taiwan for most countries, it is perhaps not surprising that most countries stayed silent on this issue. Moreover, since we only focused on a single issue area, we need to be very cautious in making any claims of generalizability. On the other hand, it is interesting that several dozen countries have been willing to risk angering Beijing by expressing some level of support or praise for Taiwan during the pandemic, and understanding what sets these countries apart from other countries is an important undertaking. We believe, moreover, that our research approach offers a rich line of inquiry to pursue for understanding both cross-Strait relations and China's overseas influence more generally. It is clear that Taiwan's status is of deep concern to the PRC leadership, and countries with which China deals are well aware of Beijing's preferences. Further research should explore other situations in which countries may have opportunities to "take a side" over Taiwan, including in a less fraught context than the global pandemic. It is clear that support for Taiwaneven among countries endorsing a "one China policy"varies considerably, and future research should continue to explore the determinants of that variation.
Joseph YINUSA is a PhD student in the Department of Government at Cornell University. He studies international relations and political theory.    Notes: *p < 0.1; **p < 0.05; ***p < 0.01. Model 1 uses the full sample. Model 2 restricts the sample to countries whose military expenditure data are available.