On 27 October 2023, Li Keqiang 李克强, the former Chinese premier, died from a heart attack at 68 years of age.Footnote 1 An official obituary described him as a dedicated official loyal to General Secretary Xi Jinping 习近平, one who had urged the country to rally around Xi by “turning grief into strength.”Footnote 2 Li's sudden death prompted spontaneous expressions of mourning both online and offline. Despite the censorship efforts of authorities, which included removing a video in which Li Keqiang promised that China would remain open to the outside world, memorial comments surfaced on social media and numerous bouquets were laid in his native city of Hefei in Anhui province.Footnote 3 This reaction was considered a reflection of the public's sense of loss for the beacon that Li Keqiang represented: a time of great economic possibility and openness to the global economy. It also implied dissatisfaction with Xi Jinping, the hard-line Party leader who had overshadowed the premier's policymaking roles, making Li Keqiang “possibly the least powerful premier in contemporary Chinese history.”Footnote 4
By examining leadership dynamics in economic policymaking and their economic outcomes, this study aims to place the late Premier Li Keqiang and his policy legacy in a proper political and historical context. As the primary agent responsible for formulating and implementing economic policy, the premier in China must navigate a constellation of political, institutional and economic constraints. Perhaps most critical to our assessment of Li Keqiang's legacy is understanding his relationship with Xi Jinping as general secretary of the Chinese Communist Party (CCP), and the Party leader's increasingly critical role in economic policymaking. Notably, since assuming office in 2012, Xi has augmented his personal influence in the overall policy process. He has headed numerous policymaking bodies, including the newly established and powerful Leading Small Group on Comprehensively Deepening Reforms (CDRLSG), earning him the popular moniker “chairman of everything.”Footnote 5 As Xi Jinping consolidated his power, the economic policies initially brought into the spotlight by Li Keqiang, known as “Likonomics,” began to lose their prominence as early as 2014.Footnote 6
Given Xi Jinping's apparent dominance over the entire policymaking system, it is noteworthy – and somewhat puzzling – that numerous accounts have suggested potential discord between Xi and Li Keqiang, particularly in the realm of economic policy.Footnote 7 A frequently cited instance involves Li's comments regarding “600 million people with monthly incomes of less than 1,000 yuan” at the press conference following the National People's Congress on 28 May 2020. This statement contrasted with Xi's emphasis on declaring the elimination of extreme poverty (quanmian tuopin 全面脱贫).Footnote 8 Another revealing instance pertains to the management of economic activities during the COVID-19 pandemic.Footnote 9 On 25 May 2022, Premier Li held a video conference with local officials nationwide, advocating for a balance between COVID-19 measures and preordained growth targets, while Xi continued to emphasize the necessity of pursuing a zero-COVID-19 policy.Footnote 10 Additionally, in earlier attempts to address economic challenges presented by COVID-19, Xi reportedly opposed Li's efforts to mitigate employment pressures with the street-vendor economy (ditan jingji 地摊经济).Footnote 11 A more fundamental divergence between the two leaders lies in their approaches to state-owned enterprise (SOE) reforms and the role of market mechanisms. Among the many policy areas where top leaders compete for influence, the state–market relationship has been a central issue throughout the reform era. Beyond the familiar debate of “state advances, private sector retreats” (guojin mintui 国进民退), this issue also intersects with broader topics, including the shifting balance between national security and economic development.
Although sporadic, these accounts of potential divergence between Xi and Li reflect what regime insiders have termed the “dispute between north and south houses” (nanbeiyuan zhi zheng 南北院之争).Footnote 12 This term describes the conflict between China's two most important policymaking institutions: the State Council (which is housed to the north of Zhongnanhai 中南海, the centre of Chinese politics) and the Party centre (sited to the south of Zhongnanhai). Observers further suggest that this conflict, which reportedly persisted throughout Xi's tenure with significant consequences for policy implementation, has been alleviated with the appointment of Premier Li Qiang 李强 in March 2023.Footnote 13 However, few studies have systematically examined whether, and to what extent, there has been a divergence in policy orientation within Zhongnanhai and the subsequent ramifications on the effectiveness of economic governance. Given the long-standing observation of a potential rift between the leaders of China's two most critical political institutions, addressing these inquiries is of profound significance beyond the current period under Xi's rule.
This study uses textual datasets of leadership speeches and activities from 2012 to 2023 to examine leadership divergence in policy orientation, particularly regarding state–market relations. In reassessing Li Keqiang's representation, this study demonstrates how his policy position evolved in relation to Xi Jinping's policy position during Li's tenure as premier of the State Council of the People's Republic of China (PRC) from 15 March 2013 to 11 March 2023. The findings reveal an initial significant divergence, followed by a convergence during Xi's second term. Li Keqiang's market-oriented rhetoric was especially prominent during “Mass innovation week” during which he promoted private and grassroots innovation. Additionally, by estimating Li Qiang's policy orientation, the study demonstrates reinforced preference convergence between the general secretary and premier as Xi enters his third term.
This paper's findings make significant contributions to the literature on China's elite politics and political economy. First, we provide empirical evidence of potential divergences in economic policymaking within the upper echelons of the CCP, the discussion of which is often discussed based on qualitative observations. By quantifying policy divergences, we attempt to critically examine the prevailing notion of Xi Jinping's immediate and comprehensive control over policymaking and implementation, offering a more nuanced understanding of the evolving leadership dynamics and their policy ramifications. Second, our findings contribute to the literature on the Chinese political economy by elucidating how policy signals from the central leadership influence market behaviour. The analysis shows that discursive discrepancies between the two leaders, interpreted by the market as “mixed signals,” were associated with a substantial decline in firm-level investment. This suggests that some perceived gaps in central policy discourse may generate political uncertainty, giving rise to suboptimal market reactions. To our knowledge, this is the first data-driven attempt to quantify the mixed signals from China's top leaders and to measure their market outcomes. The remainder of this paper is organized as follows. The next section discusses the existing literature and debates about the sources and outcomes of policy divergences among top leaders regarding economic policy. It highlights several instances where political uncertainty, engendered by “mixed signals” from the centre, influenced market behaviour. The paper then presents the data and methods used in the study before reporting the main findings related to the measurement of policy orientations and the impacts on firm-level investment behaviour. It continues with a discussion of the broader implications of the analysis before concluding.
Background of Policy Divergence
The reform era
Despite the authoritarian and hierarchical nature of its political system, the PRC is no exception in that policymaking authority has been shared among a group of top leaders.Footnote 14 At the onset of the reform era, the “elders,” led by Deng Xiaoping 邓小平 and Chen Yun 陈云, were determined to prevent a recurrence of the disasters that they attributed partly to the concentration of power in a single leader. Deng intentionally devised a new division of responsibility, with a central focus on the separation of Party and government roles. Particularly in economic policy, a significant amount of authority is delegated to the government, and specifically to the premier. While the CCP retains ultimate authority over economic affairs (as it does in all areas), it delegates primary control over the economy to the premier, a pattern that has been fairly stable and consistent since the 1980s.
However, neither the motivations for power sharing nor subsequent institutional arrangements have ensured the elimination of policy divergence at the top. One source of conflict is institutional instability. The delegation of authority to the premier has never been absolute and can always be revoked, since the Party remains the ultimate authority. An earlier example involves a disagreement between Premier Zhao Ziyang 赵紫阳 and General Secretary Hu Yaobang 胡耀邦 concerning the pace of economic growth. This discrepancy required the direct personal intervention of Deng Xiaoping to establish the principle that the premier is responsible for economic policy.Footnote 15 Ironically, when Zhao assumed the position of general secretary following Hu's downfall, he remained as head of the Leading Small Group on Finance and Economy (FinEcon LSG) and continued to wield a strong influence over economic policy, while Li Peng 李鹏, the new premier, was never able to establish an effective policy regime.Footnote 16
Another source of policy divergence among top leaders is factional competition. Although factions in Chinese politics generally mean a personal network of senior leaders and followers, they are oftentimes endowed with distinct organizational resources that can shape their policy preferences and political strategies. For instance, Victor Shih identifies what he labels the generalist factions which are led by senior Party functionaries who exert the strongest control over Party apparatuses.Footnote 17 Then there are the technocratic factions that spring from specific bureaucratic groups, with those led by senior economic officials wielding by far the greatest influence owing to their direct oversight of monetary and investment policies. It is worth noting that, despite occasional sharp disagreements, these two types of factions primarily “coexist in a symbiosis,” as the technocratic factions do not aspire to gain ultimate control of the regime.Footnote 18 The differing stances towards mounting inflation pressures in the early 1990s between Jiang Zemin 江泽民 and Zhu Rongji 朱镕基 exemplify this factional dynamic, with its roots traceable to the conflict between Deng Xiaoping and Chen Yun throughout the 1980s.Footnote 19 What, then, are the outcomes of policy divergence outside of Zhongnanhai? One plausible outcome is that policy conflict at the very top may convey mixed signals to subnational governments, resulting in a pattern of inaction or resistance among lower-level officials. An extensive literature on central–local relations in China has noted the correlation between conflicting signals emanating from the central leadership – and subsequent ambiguity in policy directives – and ineffective policy implementation.Footnote 20 Indeed, the anti-inflationary “Sixteen measures,” issued in 1993 by the State Council following the conflict between Jiang Zemin and Zhu Rongji, encountered significant local resistance.Footnote 21 Moreover, perceived divergence within the leadership can have substantial ramifications for market actors. A growing body of research has documented that uncertainty regarding the timing, content and potential impact of economic policy decisions can yield suboptimal economic consequences.Footnote 22 Despite enduring methodological challenges in measuring political uncertainty, reform-era China presents considerable evidence that suggests that firms react to political or policy uncertainty stemming from certain types of elite disunity by, for example, delaying investments.Footnote 23
The Xi era
It has now become widely accepted that since assuming power, Xi Jinping, as general secretary, has enacted significant changes to the policy process both within and beyond Zhongnanhai. This observation is particularly pertinent in the sphere of economic policy. Indeed, as many analysts note, economic policymaking has become the focus of Xi's “institutional innovations,” through which the locus of policymaking is believed to have shifted markedly from the State Council to the Party.Footnote 24 The groundwork for this shift was laid early in the Xi administration with the creation of new leading small groups (LSGs) endowed with expanded executive authority.Footnote 25 According to Barry Naughton, this shift represents a clear departure from reform-era norms, which consistently placed the premier in charge of economic policy.Footnote 26
However, despite Xi Jinping's ostensibly near-complete control over the policy process, there are several reasons to posit that economic policymaking continued to exhibit a degree of policy divergence within the leadership, particularly between the general secretary and the premier. Institutionally, one significant factor is the legacy of collective leadership, which ensured the premier's primary control over economic policy. Given the significance and complexity of economic governance, it is implausible that the long-standing norms and practices of division of responsibility for economic policy, which were operating in arguably the most pronounced way during the preceding administration, would abruptly cease to function.Footnote 27 Moreover, as previously noted, the LSGs lacked the formal authority to impose decisions on the State Council – at least before they were formalized and given legal authority following a massive restructuring of government and Party bodies in March 2018. Thus, it remains unclear how effectively the new LSGs created by Xi were able to undermine, if not replace, the power of the premier, particularly during the early years of Xi's tenure.
Factional dynamics may provide another source of policy divergence among central leaders. It is plausible that the aforementioned conflict between the generalist and technocratic camps, rooted in differences in organizational resources and membership, persisted into the Xi era. Additionally, Li Keqiang may have inherited from Hu Jintao, his factional patron, what Cheng Li referred to as the “populist” policy preference, which contrasted with the “elitist” policy posture that was expected to be adopted by Xi.Footnote 28 Although not necessarily involving an open challenge to the Party's institutional supremacy and not homogenous in terms of specific policy orientations, these factional differences, both innate and acquired, would likely continue to shape economic policy debates within Zhongnanhai.
Finally, external shocks may lead to new policy debates or the manifestation of latent policy differences. For example, previous accounts have pointed to the 2015 stock market crash as one of the earlier instances when cleavages between the two leaders emerged and subsequently expanded into overt conflict, particularly over the pace and direction of SOE reforms.Footnote 29 Despite Xi Jinping's further consolidation of power following the 19th Party Congress, the policy divergence between Xi and Li continued to capture media attention, a situation that only intensified with the onset of the COVID-19 crisis. As noted earlier, the pandemic period was marked by sporadic yet notable disagreements between the two leaders concerning how to balance pandemic control measures with macroeconomic management.Footnote 30 With regard to economic outcomes of the apparently continued presence of divergence in Zhongnanhai, scholars have observed that shifts in structures and processes in economic policymaking have actually intensified political and policy uncertainty, further complicating the environment for policy implementation.Footnote 31 Crucially, there remains uncertainty about who is making specific decisions. To be sure, shifting economic decision making to the Party means that Xi Jinping ultimately makes the crucial decisions. However, as premier, Li Keqiang still oversaw the day-to-day operations of the government, which enabled him to retain significant influence over policy interpretation and the prioritization of policies competing for resources and attention. To what extent the lack of implementation can be attributed to perceived policy divergence within the leadership remains understudied.Footnote 32
State–market orientation
The preceding discussion suggests that elite politics in China remain contentious, which may result in mixed signals for economic actors. However, the scope and intensity of contention can vary across different issue areas, such as the relationship between central planning and the market, responses to inflation, or growth targets. This study focuses on state–market orientation (SMO) by delineating a spectrum where the state-centric pole advocates increased government and Party involvement in economic governance, and the market-centric pole emphasizes the role of market mechanisms and minimizes state intervention. A notable manifestation of this spectrum is the debate over the “state advances, the private retreats” phenomenon.Footnote 33 The crux of the controversy lies in the relative stagnation of the private sector following extensive policy interventions after the global financial crisis, contrasted with the expansion of the state-owned sector.
Despite the administration's earlier commitment to strengthening the role of the market, as indicated in Xi Jinping's ambitious reform agenda at the Third Plenum in November 2013, there were signs of divergence within the leadership regarding the state-sector reform. An illuminating episode was a national conference on SOEs held on 4 July 2016. Xi emphasized that SOEs are “an important force for strengthening the comprehensive strength of the country and safeguarding the common interests of the people” (guoyou qiye shi zhuangda guojia zonghe shili, baozhang renmin gongtong liyi de zhongyao liliang 国有企业是壮大国家综合实力、保障人民共同利益的重要力量), and called for “stronger, better and bigger” (zuoqiang zuoyou zuoda 做强做优做大) SOEs. In contrast, Li emphasized the importance of “adhering to and promoting SOE reforms” (jianchi buxie tuidong guoyou qiye gaige 坚持不懈推动国有企业改革) and stressed the need to “follow the laws of the market, slim down and improve quality and efficiency, eradicate excess and backward production capacity” (zunxun shichang guilü, shoushen jianti tizhi zengxiao, taotai guosheng luohou channeng 遵循市场规律，瘦身健体提质增效，淘汰过剩落后产能).Footnote 34
To a certain extent, the leadership divergence revealed in the July conference, particularly the premier's expressed willingness to adhere to SOE reforms, suggested that the reform programme outlined at the Third Plenum was already beset with significant difficulties, with the issue largely residing in the policy process. SOE reforms, like other critical elements of the Third Plenum, were supposed to be coordinated by the newly founded CDRLSG chaired by Xi Jinping. The CDRLSG delegated the drafting of different elements of the reform agenda to different parties, allowing the policy process to be “highjacked” by internal interest groups including, but not limited to, the control agency overseeing SOEs, the State Asset Supervision and Administration Commission (SASAC).Footnote 35 Consequently, when a fragmentary, disappointing and in some respects incoherent state enterprise reform plan was promulgated in September 2015, it became evident to many observers that the fate of SOE reforms was sealed.Footnote 36
Closely related to the faltering state of SOE reforms was a policy initiative led by Li Keqiang, “Mass entrepreneurship, mass innovation” (dazhong chuangye, wanzhong chuangxin 大众创业, 万众创新), which underscored the role of small start-ups in fostering innovation.Footnote 37 While Xi Jinping is generally perceived to have concentrated on large, state-owned enterprises, Li frequently visited small and medium-sized enterprises and start-ups. The premier actively promoted supportive policies through the organization of large-scale exhibitions in major cities nationwide during the “National mass entrepreneurship and mass innovation activity week” (quanguo dazhong chuangye wanzhong chuangxin huodong zhou 全国大众创业万众创新活动周, “Mass innovation week” hereafter). During “Mass innovation week,” which was held annually from 2015 to 2022, market-oriented messages were consistently conveyed through events and speeches.
The fervent atmosphere surrounding the initiative and the political connotations of the term “mass” (dazhong 大众) infused the event with a sense of campaign-style mobilization.Footnote 38 The governance literature has long noted the use of campaign-style mobilization in instances such as the one-child policy, crisis management (particularly pandemic control) and targeted poverty alleviation.Footnote 39 In the context of elite politics, it has been documented that when top leaders hold divergent views, they may undertake inspections of local areas, launching policy mobilizations to promote their own policy objectives. Historical examples include Jiang Zemin's visit to Guangdong and Deng Xiaoping's Southern Tour. Indeed, one of the most significant events during the eight years of “Mass innovation weeks” was the Global Entrepreneurial Leaders Forum, held in Shenzhen on 12 October 2016. Hosted by Li Keqiang, the forum was attended by prominent figures such as Apple's CEO Tim Cook, Alibaba Group founder Jack Ma, Tencent Holdings’ Pony Ma and Foxconn's Terry Gou, and attracted extensive media attention.Footnote 40
In short, previous studies indicate that SOE reforms, despite initial expectations, began to falter shortly after the Third Plenum. This faltering was in part owing to a radical shift in the structure and processes of economic policymaking, resulting in sporadic displays of divergent preferences from the leadership regarding the focus and strategies of reforms. Our discussion in the preceding sections predicts that this divergence could engender mixed signals to economic actors, particularly state and private firms in this case. Few studies, however, have attempted to systematically measure and analyse how economic policy orientations at the top evolved over time and what ramifications they brought to the real economy. Estimating the SMO of the two most important economic leaders and examining the consequences of divergent orientations contribute to two specific literatures: the literature on contentious elite politics and its policy outcomes (including the strategy for policy implementation) and the literature on the economic consequences of political uncertainty, including mixed signals and campaign-style mobilization.
Data and Methods
To measure the SMO of the top leaders, this paper utilizes textual data, a traditional data source in contemporary China studies.Footnote 41 With advancements in machine learning and computing capacity, the use of text data has been increasing in recent years. Various topics, including public diplomacy,Footnote 42 censorship and propaganda,Footnote 43 central–local dynamicsFootnote 44 and economic policy uncertainty,Footnote 45 have been analysed using text data, producing novel insights into various aspects of governance in China. These studies employ diverse types of text data, ranging from national and local newspapers, official policy documents, social media posts and speeches given by Party and government leaders at different levels. Besides the increased availability of information in digital form, the main drivers behind the recent growth in the use of text data are the widespread adoption of quantitative text analysis methods in social sciences.Footnote 46
We obtained text data from the database of Xi Jinping's “Important speech series” (Xi Jinping xilie zhongyao jianghua shujuku 习近平系列重要讲话数据库) for Xi Jinping,Footnote 47 the official website of the State Council for Li Keqiang,Footnote 48 and the “Li Qiang activity report special page” at the Chinese Communist Party News Network for Li Qiang.Footnote 49 The Xi Jinping database is a web-based source of related speeches and reports. Coverage begins on 15 November 2012, the day that Xi Jinping took office as the general secretary of the CCP. Items include speeches, activity reports and field visits, among others.Footnote 50 The State Council website and Li Qiang's Special Page collect the premier's political meetings, speeches and other activities, including local and foreign visits.
Given that the central motivation of this study is to measure economic policy orientation, we selected articles containing the following words: economy (jingji 经济), finance (jinrong 金融), commerce (shangye 商业), firm (qiye 企业), enterprise (gongsi 公司), industry (chanye 产业), employment (jiuye 就业) and income (shouru 收入). After screening, we obtained 5,184 articles for Xi Jinping from 15 November 2012 to 31 October 2023; 3,537 articles for Li Keqiang from 15 March 2013 to 11 March 2023; and 95 articles for Li Qiang from 12 March 2023 to 17 October 2023.
Using these datasets, we measure SMO using latent semantic scaling (LSS), a semi-supervised machine learning method.Footnote 51 In LSS, a small number of seed words are set, and word-level polarity scores are calculated based on the semantic distance between words as measured by cosine similarity. Using the word-level polarity score, the document-level polarity score is obtained by summing the frequency of appearance of all words in an article divided by the number of the words in the article. In recent years, analyses using LSS have spread to a wider range of social sciences.Footnote 52 The advantages of the LSS method are as follows: (1) it can be implemented in fields without representative dictionaries; (2) it does not require document-level training data (a requirement of supervised learning); and (3) it does not require posterior interpretation (a challenge for unsupervised learning). The first point is particularly beneficial for the current study because no existing dictionaries measure SMO.
We follow the standard procedures and set the seed words for the LSS (Table 1).Footnote 53 The SMO is designed to highlight divergent policy orientations, specifically in three key areas: the ownership dichotomy between SOEs and private enterprises; the contrast between state intervention and market-oriented mechanisms with lean government;Footnote 54 and the policy objectives of scale expansion versus promoting new start-ups and efficiency. The primary objective is to assess the temporal evolution of policy orientations by employing opposing keywords associated with Xi and Li Keqiang in their remarks on SOEs.
Table 1. Seed Words for the Latent Semantic Scaling Method
We refer to this indicator as SMO; moreover, it can be interpreted as gauging the degrees of Xi Jinping-ness (positive score) and Li Keqiang-ness (negative score). The online Appendix Figure 1 shows the word-level polarity scores before calculating the SMO, presenting reasonable variations among words. Appendix Table 1 presents article-level SMO scores for the top (most state-oriented) and bottom (most market-oriented) 20 articles. The majority of the state-oriented articles are from Xi (13/20), whereas the most market-oriented articles are predominantly by Li Keqiang (18/20).
Before proceeding to estimate the economic impact of policy divergence, the issue of media bias warrants a brief discussion.Footnote 55 Previous studies have shown that bias is more pronounced in media outlets under strong Party or state control. However, the presence of bias does not necessarily preclude the extraction of useful information from texts. Yun Huang and Paul Luk investigated how the variation in data sources affects the construction of an index of economic policy uncertainty, using 114 local Chinese newspapers.Footnote 56 They categorized the newspapers into those with significant bias and those with minor bias. The results showed a high degree of similarity, with a correlation coefficient of 0.93 between the benchmark index and the high-bias group and 0.94 between the benchmark index and the low-bias group.Footnote 57 This assessment suggests that media bias does not significantly impact the construction of an economic index.
After obtaining the SMO, we examine the influence of the divergence between Xi and Li Keqiang on the real economy by running firm-level investment models. As discussed earlier, the perceived gap in policy orientations between the general secretary and the premier suggests potentially divergent policy directions, which generate a degree of political uncertainty. Previous studies indicate that heightened political uncertainty typically exerts a negative impact on corporate investment behaviour and stock prices.Footnote 58 In the Chinese context, SOEs particularly exhibit pronounced responses to political uncertainty owing to their alignment with government policies.Footnote 59 The online Appendix notes describe the detail of data and variables.
Results
Figure 1 presents the estimated LSS score over time. At the onset of the Xi Jinping administration in 2013, the disparity in policy orientations between Xi and Li Keqiang was relatively small. However, the divergence between the two leaders became most apparent in 2016–2017. While Xi's stance remained consistent, Li Keqiang's tone exhibited significant fluctuations, indicating his periodic engagement in a time-limited, campaign-style policy mobilization. If Li Keqiang had used a stronger market-oriented tone to counterbalance Xi Jinping's emphasis on state-owned enterprises, a negative correlation would be expected between their tones. However, we found no compelling evidence of a negative correlation between the tones at the monthly or quarterly level.Footnote 60 This finding suggests that Li Keqiang's wave of market-oriented tones is not explicitly intended to counteract Xi's concurrent state-oriented tone. Post-2018, during Xi's second term, Li Keqiang's tone adopted a relatively moderate stance. Beyond 2020, although discernible differences persist in tones, a certain level of convergence appears to exist. Overall, during the second term of the Xi administration, Li Keqiang's tone became more neutral and moderate. It is noteworthy that Premier Li Qiang, who assumed office in March 2023, exhibited a pronounced state-oriented tone, although this is based on a limited sample size. If the differences in tone could be explained by the position, such a disparity between Li Keqiang and Li Qiang would not occur.
Figure 1. State–Market Orientation by Leader
Table 2 presents the descriptive statistics for the estimated LSS scores of the leaders. Across this period, the mean LSS scores for Xi Jinping and Li Keqiang are 0.24 and −0.36, respectively, which indicates a substantial difference.Footnote 61 In the first period, a large disparity between Xi and Li Keqiang is evident (mean gap: 0.71; 0.27–(−0.44)). Although the gap narrows in the second period as their tones converge, it remains substantial at 0.45 (0.21–(−0.24)). Notably, Xi's average orientation tends towards neutrality, which shifts from 0.27 in the first term to 0.14 in the third term. The F-test results, which assess statistical differences in tone between them, underscore a notable divergence during the first two terms. In the third term, however, after Li Qiang took office, the statistical significance in the difference between the general secretary and the premier diminished. We cannot reject the null hypothesis, which suggests that the mean difference in tones between Xi and Li Qiang is zero. This outcome supports the widely held belief that Li Qiang, as one of Xi's loyalists, lacks a distinctive policy orientation in terms of his SMO.Footnote 62
Table 2. Summary Statistics of State–Market Orientation
Note: Statistical significance markers represent * p < 0.1; ** p < 0.05; and *** p < 0.01.
Li Keqiang's campaign-style policy mobilization is manifested in his signature initiative, “Mass innovation week.” Notably, in the period 19–23 October 2015, the inaugural “Mass innovation week” witnessed the initiation of market-oriented outreach efforts.Footnote 63 The relevant policies led to a notable shift in the overall tone towards the market-oriented side. To assess the impact of these campaign weeks, we conducted weekly-level ordinary least squares estimations (Table 3). The results indicate that during the “Mass innovation weeks” (from 2015 to 2022), Li Keqiang's tone shifted approximately by 0.45 towards a market-oriented direction, while no discernible effect was observed on Xi's tone. The significant shift in Li Keqiang's tone is unsurprising, given his engagement in activities to promote the development and establishment of private enterprises during these weeks. If Xi and Li Keqiang shared a common policy objective, then Xi's coefficient of the week dummy should also be negative, as there would be no reason to disregard the premier's initiative. However, our findings suggest a divergence between the two leaders both in terms of policy orientation and activity.
Table 3. Campaign-style Policy Promotion by Li Keqiang
Notes: Standard errors are reported in parentheses. Statistical significance markers represent p < 0.1; **p < 0.05; and ***p < 0.01. The mass innovation week dummy is a dummy variable that takes the value of 1 for the following periods: 19–23 October 2015; 12–18 October 2016; 15–21 September 2017; 9–15 October 2018; 5–11 June 2019; 15–21 October 2020; 19–25 October 2021; 13–19 September 2022.
Finally, Table 4 presents the results of an analysis on panel data of Chinese listed firms, examining the impact of the discursive difference of the two leaders (Xi_Li_Gap) on the investment ratio (IoverK) of firms in the subsequent quarter. In all estimation models, we control for firm-level fixed effects and quarterly fixed effects. Additionally, Models 1 through 3 include Xi_Li_Gap as a key explanatory variable. Model 1 estimates the impact for all firms, Model 2 examines state-owned enterprises and Model 3 focuses on non-state-owned enterprises. In each of Models 1 through 3, the coefficient for Xi_Li_Gap is negative. These findings consistently reveal a negative association between tone difference and investment ratio, suggesting that perceived leadership divergence in policy orientation negatively impacts real economic activities. Among the subgroup analyses, we find a statistically significant negative effect on SOEs. In Models 4 through 6, additional control variables, including firm-level financial indicators and macroeconomic variables, are added. However, the estimation results remain qualitatively robust. These results are consistent with a previous study and suggest that state-owned enterprises are more sensitive to political uncertainty.Footnote 64
Table 4. Xi and Li Gap and Firm Investment
Notes: The dependent variable (IoverK) is the investment ratio and the independent variable L.xi_li_gap is the logarithm of the difference between Xi and Li Keqiang's index as estimated by LSS. Samples are Chinese listed companies in the manufacturing sector. Standard errors are clustered at the firm-quarter level, and reported in parentheses. Statistical significance markers represent * 0.10, ** 0.05, and *** 0.01.
Discussion
Examining the discursive evolution of policy orientations regarding SMO during their ten-year tenure reveals a pattern where the differences between Xi Jinping and Li Keqiang were relatively small shortly after the administration took office, then widened and eventually narrowed again. Comparing our results with the general observation that the distinctive elements of Li Keqiang's policies, known as Likonomics, began to fade as early as 2014, it appears that Premier Li Keqiang might have attempted to assert his own policy position more strongly than previously assumed until around early 2016. Although intermittently, Li Keqiang seems to have endeavoured to introduce his own policy initiatives even during the period leading up to the 19th Party Congress in October 2017. However, as Xi Jinping entered his second term and further institutionalized his personal control over the policy process, it became clear that Li Keqiang's assertiveness in promoting market-oriented policy initiatives gradually yet undeniably weakened.
The overall pattern and the timing of shifts in policy orientations of the two leaders have several implications for leadership dynamics in economic policymaking and in how the Xi era departs from previous administrations. First, the analysis illuminates the weakness of the institutions designed to ensure the delegation of operational control over economic policy to the premier, which proved to be particularly susceptible to the general secretary's individual ambitions and exogenous shocks. For example, Li Keqiang's convergence with Xi's policy stance after 2016 suggests that the premier's position may have relatively weakened as the stock market crisis helped to amplify the general secretary's agenda-setting power. Second, the analysis questions the factional basis of leadership divergence on economic policy. Li Keqiang's market-oriented policy orientations do not necessarily align with both the “technocratic” and “populist” preferences that he would have inherited from his predecessors and patrons. Similarly, Xi's consistent emphasis on the state's strengthened role in the economy reinforces the extant observation that factional dynamics have ceased to function, at least in their pre-existing form, in the Xi era.
Another point to consider is the role of policy campaigns. The “Mass innovation week,” which aimed at promoting innovation through start-up companies, is a notable example. As our analysis indicates, during the campaign weeks Li Keqiang's tone was significantly more market oriented. While this result is not surprising, we believe that understanding the Chinese political economy in the 2010s involves recognizing the significance of Premier Li Keqiang's efforts to convey market-oriented messages both domestically and internationally through extensive media mobilization. This continuous delivery of market-oriented messages by top leaders, even if not always resulting in concrete outcomes, had a symbolic importance in contrast to General Secretary Xi Jinping's emphasis on state control and national security. Conversely, Premier Li Qiang, who succeeded Li Keqiang, has not continued with “Mass innovation week” and other similar initiatives. Many elements included in Premier Li Keqiang's version of the “State Council work regulations” have been removed in the revised version issued by Premier Li Qiang in March 2023.Footnote 65 Although Premier Li Qiang occasionally emphasizes the importance of the private economy, his overall policy tone aligns more closely with that of Xi Jinping.Footnote 66
Finally, the firm-level estimation results suggest that the policy divergence between Xi Jinping and Li Keqiang may have been perceived as a source of political uncertainty.Footnote 67 SOE managers, in particular, who are highly attuned to national policy directions, were sensitive to mixed signals from the central leadership. Despite the locus of policymaking shifting to newly established LSGs controlled by Xi Jinping, Li Keqiang continued to express his own policy orientations through campaign-style mobilizations, which peaked around 2015–2016. Notably, Li Keqiang's active engagement with the “Mass innovation week,” characteristic of campaign-style mobilizations, may have introduced additional political uncertainty. As Sebastian Heilmann and Elizabeth Perry note, while institutions generally work to contain uncertainty, what they call “guerrilla policy style” embraces uncertainty.Footnote 68 Our analysis regarding the “Mass innovation” campaign and corporate investment provides evidence for a linkage between policy campaign and uncertainty. Li Keqiang's fluctuating policy focus thus underscores the continued relevance of campaign-style mobilization, particularly as a strategy of policymakers facing institutional barriers when promoting their individual policy agenda.
Conclusion
This study examines the evolution of leadership dynamics in economic policymaking and their economic consequences to situate the late premier, Li Keqiang, and his policy legacy within a proper historical and political context. Utilizing a semi-supervised text analysis, it demonstrates that Li Keqiang clearly favoured market-oriented policies but gradually converged towards Xi Jinping's preference for more state-led approaches in economic governance. In other words, the divergence in preference between China's two most important decision makers diminished as Li Keqiang gradually shifted away from his original stance. Our analysis further indicates that Li Keqiang's distinct policy orientation, partly shaped by his institutional role as premier, has not been carried forward by his successor in Xi's third term. Although we remain cautious in overinterpreting this finding due to the limited sample size for Li Qiang, this study represents, to the best of our knowledge, the first attempt to quantify the differences in policy orientation between Xi Jinping, Li Keqiang and Li Qiang, and, more broadly, to measure the economic impact of leadership contentions in the Xi era.Footnote 69
Our analysis has several limitations. First and foremost, the range of Li Keqiang's policy orientation extends well beyond the state–market dimension. In addition to promoting greater marketization in economic management, Li Keqiang's policy priorities encompassed advocating openness to the international economy and implementing reforms in broad areas of governance, such as domestic migrants and economic digitalization.Footnote 70 Second, our analysis only discusses the short-term consequences of political uncertainty. While Li Qiang's close alignment with Xi's policy orientation may help to mitigate political uncertainty in the short term, the absence of a premier (or central leader) with a distinct policy stance could have significant long-term implications for economic policymaking and the broader economy. Historically, as Ezra Vogel noted, Chen Yun's conservative stance and adjustment policies “paved the way for Deng to accelerate growth and reform” by stabilizing the macroeconomy.Footnote 71 Moreover, the negative impacts of the State Council's diminished role in economic management are already becoming apparent, as evidenced by market actors’ scepticism regarding the central government's commitment to recent policies aimed at promoting the private sector.Footnote 72 More broadly, further research is required to assess the outcomes of the rapid shift towards and entrenchment of an excessively centralized policymaking system under Xi's leadership. Last but not least, we concede that a comprehensive evaluation of Li Keqiang's preferences and legacies will require a deeper and extended examination of his entire political career.Footnote 73
Simultaneously, this study makes several contributions to the literature. First, it presents empirical evidence of leadership divergence over economic policy, offering a nuanced understanding of the elite politics in the Xi era and its policy outcomes. In doing so, this study proposes a novel approach for the real-time analysis of policy orientations among key policymakers. Second, this study demonstrates that a measure of political uncertainty can affect market behaviour, thus contributing to the literature on the Chinese political economy. Specifically, our analysis reveals that discursive gaps in the central leadership, interpreted by the market as mixed signals, were associated with a substantial decline in corporate investment. Third, this study provides evidence of Premier Li Keqiang's efforts at pursuing market-oriented reform agendas – albeit not always yielding tangible results – despite an increasingly challenging political and institutional environment. Combined with the campaign-style mobilizations undertaken by Li Keqiang, the overall evidence suggests a serious reconsideration of the popular portrayal of the late premier as the “least powerful premier in contemporary Chinese history.” While nostalgically labelling him as a “reformer” may oversimplify his legacy, a more balanced and thorough assessment of Li Keqiang's role and legacy will enhance our understanding of China's political economy in the years ahead.
Supplementary material
An appendix is available online at https://doi.org/10.1017/S0305741025000165
Acknowledgements
We thank participants of the Cambridge China Politics Research Workshop, the Chinese Economists Society North America Conference and Tomoki Kamo for their valuable comments and suggestions. This work was supported by the Japan Society for the Promotion of Science KAKENHI, grant number JP24K03158.
Competing interests
None.
Asei ITO is an associate professor in the Institute of Social Science at the University of Tokyo. His research focuses on Chinese business and economy. His recent work has been published in Social Science and Medicine, International Journal of Disaster Risk Reduction, China Economic Review and East Asia: An International Quarterly.
Jaehwan LIM is a professor in the School of International Politics, Economics and Communication at Aoyama Gakuin University, Japan. His research focuses on Chinese politics, especially Party–military relations. His recent work has been published in Journal of Contemporary East Asia Studies, Asia-Pacific Review and China Economic Review.
Hongyong ZHANG is a senior fellow and policy adviser at the Research Institute of Economy, Trade and Industry, Japan. His research focuses on international trade and investment and the Chinese economy. His recent work has been published in European Economic Review, Journal of Monetary Economics, Journal of Comparative Economics, Review of World Economics, The World Economy and China Economic Review.
Open access
