This study examines whether populist incumbency mitigates the crises that often facilitate populists’ rise or instead contributes to their reproduction through governance. We employ a mixed‑methods design combining cross‑national panel analysis of 23 countries (2000–2022) with structured within‑case process tracing in Hungary, Brazil, and the United States. Using difference‑in‑differences and event‑study models with country and year fixed effects, the quantitative results show that populist incumbency is associated with substantively large declines in political institutional quality and with more diagnostically sensitive evidence of deterioration in social-liberal conditions, including weaker constraints on executive power and reduced protections for media freedom and civil liberties. In contrast, average macroeconomic performance does not exhibit robust improvement during populist rule, undermining claims of a systematic prosperity‑for‑democracy tradeoff. The qualitative evidence traces how these political, social, and economic outcomes can emerge: populist incumbents deploy delegitimation of oversight actors, crisis framing, and polarization as governance strategies that lower the political costs of institutional confrontation and enable concrete interventions that weaken accountability and narrow pluralism. Cross‑case comparison further indicates that the depth and durability of these effects vary with the strength of veto points, civil society counterweights, and informational pluralism. Overall, the findings contribute to debates on populist governance and democratic backsliding by showing how populist incumbency is associated with deterioration in political and social‑liberal outcomes without consistent economic gains, while also identifying scope conditions that shape heterogeneous trajectories across cases.
Using a comprehensive panel dataset of 48 sub-Saharan African countries from 1970 to 2024, this essay study employs a mixed-methods research design, triangulating regression analysis with fuzzy-set Qualitative Comparative Analysis (fsQCA), to investigate the causal impact of foreign aid on state-building. Transcending unidimensional correlational analyses, the article advances a “comprehensive multi-causal” theoretical framework that integrates probabilistic statistical trends with configurational logic. Departing from the reductive “aid-is-harmful” thesis, our findings delineate a nuanced and heterogeneous impact: while aggregate foreign aid exerts a statistically significant negative effect on state capacity, this relationship is highly bifurcated by types of aid. Specifically, concessional loans are robustly associated with enhanced state capacity, whereas grant aid correlates with its erosion. This divergence is driven by two distinct governance logics: concessional loans catalyze an “accountability logic”, wherein repayment obligations incentivize sovereign debtors to bolster revenue extraction, streamline bureaucratic coordination, and improve policy implementation. Conversely, grant-based aid reinforces a “rent-seeking logic” that compromises domestic accountability and attenuates endogenous institutional incentives. Crucially, these causal pathways are context-contingent; the magnitude and direction of these effects are moderated by structural variables, most notably ethnic fractionalization, regime type, and natural resource rents.
With the increasing use of artificial intelligence (AI) in public administration and political science, discussions about the future of state governance have re-emerged as a prominent topic of debate, as has been the case with previous major technological developments. In this context, the main question of this study is whether artificial intelligence can govern the state. Since answering such a broad main research question is quite challenging, four additional sub-research questions have been identified to narrow the scope of the study: (i) Can artificial intelligence replace governments or heads of state? (ii) Should artificial intelligence be used as a support tool to improve and facilitate public services provided by the state? (iii) A hybrid model (iv) Neither of the above. The study employs a multi-layered qualitative and quantitative research design to answer these broad and specific research questions: “Multi-Schema Abstract-Level Thematic Content Analysis”; “Bibliometric Mapping and Thematic Structuring”; “Qualitative Document Analysis”, and “Hypothetical Reasoning and Normative Thought Experiment”. When these layers of analysis are brought together, they provide a comprehensive view of how AI is represented in philosophical, institutional, legal, and academic contexts and enable judgments about the current state, limits, and trajectories to answer the research question. The analysis reveals that artificial intelligence is viewed as a tool for enhancing capacity and supporting decision-making, particularly in bureaucratic processes and the delivery of public services; however, it still has a long way to go before it can become a decision-maker with political authority.
This study examines the comparative influences of economic and technological drivers on corruption perception in lower- and lower-middle-income countries of the Belt and Road Initiative consortium. The study uses panel data regression with 539 observations (from 2012 to 2022) and a fixed-effects model to assess how perceived corruption is impacted by economic growth, the E-Government Development Index (EGDI), information and communication technology (ICT) infrastructure, inflation, and foreign direct investment. Economic growth shows a modest negative relationship with the Corruption Perception Index, while foreign direct investment and inflation have no significant impact, indicating a limited direct effect on perceived corruption. In contrast, EGDI and ICT infrastructure significantly influence the Corruption Perception Index, suggesting that advancements in digital governance and technology enhance transparency and reduce corruption perception. Critically, interaction effects reveal that technological advancements mitigate economic limitations. Specifically, the EGDI–growth and ICT–growth interactions yield negative (albeit diminishing) returns, suggesting that digital transparency compensates for weak economic spillovers. These findings challenge traditional growth-centric anti-corruption approaches, advocating instead for a dual-track policy model in which ICT and e-government investments are primary levers for transparency, even in economically constrained Belt and Road Initiative contexts.
Whole-process people’s democracy has become a major theme in the research on Chinese politics and the advancement of China’s political system, thus bearing significant value for academic research. However, academic debates within the study on Chinese democratic politics remain diverse and ongoing. This study integrates existing research to establish an analytical framework for elucidating democratic governance in China, with the objective of explaining the democratic principle and governance orientation inherent to whole-process people’s democracy. It argues that, in contrast to Western democracy, whole-process people’s democracy emphasizes its principles in subject, relation, object, method, mechanism, and value. It also exhibits a distinct governance orientation, thus forming a composite structure of democracy and governance. This study elucidates China’s balanced strategy of democracy and governance, demonstrates the unique democratic governance structure of Chinese democracy, and contributes to the academic understanding of democracy in China and study on the relationship between democracy and governance.
Agenda democracy emphasizes citizens’ effective influence on policy agendas as central to democratic quality. Traditional electoral democracy typically overlooks policy substance, while deliberative democracy faces challenges in linking discourse to political systems, thereby limiting agenda democratization. This study analyzes China’s consultative democracy, which is embedded within its institutional framework and systematically transforms public agendas into policy agendas, ensuring a continuous integration of popular will into official decision-making and government action. Through institutionalized consultative mechanisms tailored to diverse public concerns, consultative democracy fosters broad participation, equal dialogue, and dynamic agenda adjustment across policy stages. This process enhances citizens’ substantive influence on policy formulation and implementation, addressing structural inequality and improving governance responsiveness. Consultative democracy serves as a viable model for advancing agenda democratization by bridging public opinion and political decision-making, thereby contributing to an inclusive and adaptive democratic governance.
Enterprise trade unions in China exhibit distinct adaptive features in democratic practice, moving beyond traditional binary analytical frameworks such as capital dominance versus state control or resistance versus dependency. As economic growth slows and external uncertainties intensify, the party-state places a growing emphasis on social stability. Consequently, the reform of enterprise trade unions has been further advanced, rendering their integrative functions increasingly prominent while the space for collective mobilization has correspondingly narrowed. Amid these institutional environment changes, the diversity of democratic practices within enterprise trade unions has become increasingly evident. Based on the interaction between democratic will and democratic capacity, four types of action strategies emerge, namely, proactive accommodation, symbolic response, tentative advancement, and passive avoidance. Empirical evidence drawn from case studies of a large private enterprise, a state-owned enterprise, a foreign-invested enterprise, and a small private enterprise substantiates the abovementioned differentiated adaptation. These cases reveal that enterprise unions adopt distinctive strategies to creatively adapt to their institutional environments. Further analysis suggests that these strategies are shaped by the interplay of three factors: institutional boundaries and gray areas; enterprises’ rational calculations for survival; and ownership structures. By revealing the differentiated practices of enterprise trade unions within an integration-oriented institutional environment, this study challenges prevailing one-sided views of Chinese enterprise unions and makes a crucial contribution to labor politics research.
This study investigates whole-process people's democracy (WPD), focusing on its integration of direct and indirect democracy and provides a potential answer to the question of what constitutes democracy today by proposing a new paradigm for modern democracy. Modern democratic theory predominantly equates democracy with representative democracy, thereby reducing democratic practice to voting and competitive elections and confining it to a set of indirect democratic practices. In contrast, WPD highlights direct democratic practices, such as public participation in the management of state and public affairs, to remedy the deficiencies of representative democracy. This effort embodies the Party's commitment to safeguarding and upholding the people's right to run the country. Furthermore, WPD integrates direct and indirect democracy by adopting consultation (xieshang (sic)(sic)) as its fundamental technical procedure. Rooted in China's socialist democracy, consultation has evolved from a guiding principle to a formalized technical procedure. It will continue expanding into an extensive, multilevel, and institutionalized technical procedure that effectively integrates direct and indirect democracy within WPD practice.
The difficulty of collecting social conflict data has caused its study to focus on recent events predominantly in the West. This paper shows how to use artificial intelligence to generate social conflict data regardless of language, location, or date. Digitizing tertiary books, lightly cleaning their text representation, and submitting that text to a large language model (LLM) produces accurate date, location, and event descriptions. These capabilities and results are demonstrated with books on Latin America after 1492, Imperial Russia, and Tokugawa Japan. Using LLMs requires a larger fixed cost than working a team of research assistants, but it produces results more quickly for most sources. It is less accurate for the Tokugawa Japan source; whether it or human translation is cheaper depends on the cost of correcting the LLM’s work. These results represent the floor of accuracy for artificial intelligence, suggesting researchers should soon use it for creating social conflict data from tertiary sources.
Economists typically model power through prices or outside-option wedges. This paper develops a formal model in which economic power operates instead through control over the feasible set of economic choices. Political scientists have long understood power as control over feasibility—the ability to determine which markets and technologies remain accessible. This paper formalizes power as control over the domain of economic choice. Economic coercion is modeled as a conditional restriction on the set of feasible inputs. When inputs exhibit complementarity—low elasticity of substitution—access restrictions generate large output losses, forcing compliance by eliminating alternatives rather than through price adjustment. Vulnerability depends on technological structure, not trade volume: an economy importing 2
Significant academic and public attention has been paid to how generative AI tools can be used in political persuasion and microtargeting. A growing body of recent research finds that in many cases, effort to customize political messaging using an LLM yields little persuasive benefit. In this project, we explore the way LLMs construct persuasive messages, how such statements vary when the LLM is asked to microtarget individuals, and the degree to which the changes induced by microtargeting lead to increased persuasion of human readers. We find variation in the degree to which different LLMs successfully comply with the microtargeting task. Furthermore, even for LLMs that produce more distinct messaging, the strategies are rarely systematically aligned with users’ background features and do not increase the persuasiveness of the message in the aggregate. We discuss the implications of these findings for research on persuasion and generative AI.
It has become common to distinguish between two types of consociations. Liberal consociations, which leave open which groups will share power and enjoy autonomy, are generally considered “good”. In contrast, corporate consociations are viewed as “bad” because they institutionalize ethnic politics. This paper challenges these assumptions. It shows that the advantages of liberal consociations have been overstated and the downsides of corporate consociations have been exaggerated. Moreover, there is little liberal about liberal consociations. The paper recommends to go back to the earlier terminology of self-determination vs. pre-determination, to accept that consociational politics is group-based, and to do more systematic empirical research on the effects of different institutional arrangements.
Geographic Regression Discontinuity Designs (GRDDs) are used to estimate long-run historical effects, despite well-established results showing that RDD requires a discontinuous assignment rule rather than merely plausibly exogenous spatial variation. This article evaluates Henn et al.’s (Comparative Political Studies. 10.1177/00104140251369335, 2025) study in Comparative Political Studies, which treats Catholic diocesan boundaries as quasi-random cutoffs. Archival church records indicate that these boundaries were endogenous to colonial infrastructure, disease ecology, settler geography, and ecclesiastical administration. Spatial-econometric simulations demonstrate that standard GRDD specifications produce statistically significant effects in 64–71
Artificial intelligence (AI) technologies are reshaping national security and the future landscape of warfare. Alongside this transformation, academic discourse on military AI has gained significant traction in recent years. In response to the rapidly expanding yet fragmented scholarly discourse on military AI, this paper employs BERTopic modeling, an advanced natural language processing technique, to conduct a systematic review of military AI literature spanning over a decade (2014-2025). Based on 505 papers screened from the Web of Science (WOS) and moving back and forth between the literature and the topic modeling results, we identified 12 topics and summarized them into three spectra and six sub-themes: governance: national strategy, rules and ethics; application: regional security and conflicts, social and cultural influences; and technology: tactics, logistics. Our study generates a novel knowledge map of military AI research that provides scholars with a systematic understanding of the field's intellectual structure and evolution and offers policymakers valuable empirical insights into emerging military AI technological trends and critical governance challenges.
Can GenAI evaluate scientific evidence? Large language models are becoming increasingly embedded across the research process, yet we lack direct tests of their ability to judge how empirical findings inform beliefs in a theory. I introduce a new benchmarking centered on Bayesian likelihood estimation, drawing from Bayesian process tracing, a case-based method in which researchers assign explicit values to how likely we would expect to observe a specific empirical finding given a theoretical hypothesis. Using 289 evidence–hypothesis pairs, I elicit likelihoods from Open AI’s GPT models. Com-pared to expert estimates, LLMs show substantial average distance, 30–40
This article examines the transformation of Japan’s security and foreign policy against the backdrop of China’s rise. It argues that Japan’s policy exhibits a paradoxical and complex posture that is “neither pure balancing nor absolute buck-passing,” and that mainstream international relations theories such as realism and liberalism face limitations in explaining Japan’s behavior. To overcome this impasse, the article introduces the Preference-for-Change model, advancing “soft revisionism” and the “domestic-international two-level game” as core analytical lenses, contending that Japan fundamentally seeks incremental adjustment within the existing order, rather than wholesale overturning. The empirical analysis proceeds on two levels—international and domestic. Internationally, it focuses on China’s military rise and uncertainty surrounding U.S. security commitments, and domestically, it examines Japanese public attitudes toward constitutional revision and security policy, as well as the constructed perceptions of the “China threat.” The findings indicate that Japan’s security strategy is jointly shaped by multiple domestic and international factors, bearing pronounced features of soft revisionism, and is constrained both by domestic public opinion and structural dependence on the U.S.-Japan alliance. Japan’s policy may continue to evolve with changes in the strategic environment based on the logic of soft revisionism, thus warranting sustained attention.
China’s national image is integral to its soft power and has crucial implications for its overseas interests. By explaining China’s national image with a country’s specific characteristics (e.g., ideology and economic level) and its relationship (e.g., trade relations) with China, previous studies have implicitly assumed that each country’s perception of China is independent of other countries’ views. This study challenges this assumption and argues that an international flow network exists in which one country’s perception of China can influence another’s. We perform a Granger causality analysis using sentiments expressed in English Twitter data from 51 countries between September 2011 and August 2021, which yields yearly international influence networks concerning China’s image. Temporal exponential random graph models show that gross domestic product (GDP) is positively correlated with a country’s ability to influence others and its susceptibility to external influences. A shared colonial history predicts the existence of influence. In addition, lower susceptibility to external influences is predicted by a greater number of high-level official visits with China, but not by measures of economic ties, including the proportions of foreign direct investment and imports from China in the country’s GDP.
This article examines the strategic interplay between Russia and China in Central Asia, a region pivotal for its resource wealth and geopolitical position, through a comparative geopolitical and strategic analysis framework. China’s Belt and Road Initiative (BRI) and the Shanghai Cooperation Organization (SCO) drive its economic and security engagement, contrasting with Russia’s post-Soviet reassertion through military presence and energy dominance via the Collective Security Treaty Organization (CSTO) and Gazprom. The study identifies areas of convergence in their efforts to limit Western influence, alongside divergence in economic and security approaches, shaped by external actors like the United States and Central Asian agencies. Future challenges include economic sustainability, security threats, and regional autonomy, all of which have implications for Eurasian connectivity. Drawing on new literature, this analysis highlights Central Asia’s role in global power dynamics, offering insights into the balance of cooperation and competition between these powers.
This paper develops a Cognitive-Network Framework (CNF) that integrates cognitive theory, observable behavioral manifestations in discourse and interaction, and computational indicators derived from NLP and network analysis. The CNF distinguishes theoretical mechanisms (motivated reasoning, identity protection), their meso-level behavioral signatures (selective sharing, sentiment homogeneity, bridge erosion), and the computational measurements (transformer classification, GNN dynamics, time-series features) that serve as proxy indicators of those signatures. XLM-RoBERTa models are used for multilingual classification and feature extraction, while temporal GNNs and VAR/LSTM architectures are used for dynamic modeling and short-horizon forecasting within validated time-series frameworks. The framework identifies four mechanisms of civic culture degradation—civil society restriction, media capture, opposition harassment, and institutional erosion—and shows how civic-electoral divergence (declining civic culture despite continued electoral competition) emerges from interactions among cognitive, network, and technological processes. The approach combines large-scale behavioral data, survey validation, experimental replication, and qualitative tracing to support inference from computational indicators to broader civic outcomes. Construct validation demonstrates consistent relationships between computational proxies and survey measures (example correlation r = 0.67 for partisan identity strength and within-community sentiment homogeneity), and cross-level interaction models show that cognitive × network effects substantially increase explanatory power for civic culture deterioration.
