Ruochen Yi

Working Paper

Dual-head Leadership Structure and Balance of Power, Evidence from Mid-level officials’ promotion in China (Job Market Paper)

This paper answers three questions: 1) How do local Chinese officials manipulate GDP growth for promotion, depending on the crucial junctures in their careers? 2) How does the central government strategically react to such manipulation in terms of promotion decisions? 3) How does the interplay between the Mayor and Secretary under the distinct dual-head leadership structure affect GDP growth manipulation? To address these questions, I collected a comprehensive dataset on the career trajectories of 536 prefecture Party Secretaries and 583 Mayors. I begin by characterizing how the promotion prospects of Mayors and Secretaries depend on their ages. I build a game theory model to generate testable predictions about the dynamic relationship between Mayors’ and Secretaries’ ages and their decisions to manipulate GDP growth. I find that both Secretaries and Mayors manipulate GDP growth at their last opportunity for promotion to primary vice-provincial positions with real authority. Aware of these manipulations, the higher authority is found to have set a higher promotion standard for these Secretaries. Secretaries at their critical age must achieve a 1.45% higher nominal GDP growth rate on average (12.5% of the average GDP growth in my data) than Secretaries at other ages to get promoted. Additionally, I have discovered that Secretaries at their critical ages manipulate less when paired with Mayors who are qualified to succeed them as Secretaries, since short-term policy measures used by Secretaries to "stimulate" economic development may hinder Mayors' future pursuit of high GDP growth, highlighting how the presence of Mayors can act as a check on Secretaries, reducing short-sighted economic policies aimed solely at promotion and ensuring long-term development.

Does Social Media Increase Polarization? A Reassessment of the 2020 Facebook-Instagram Election Study Using Multiverse Analysis  (with Olivier Bergeron-Boutin and Neil Malhotra)

During the 2020 U.S. presidential election, an independent team of academic researchers collaborated with Meta to study the effects of Facebook and Instagram on politics and society. The results of four experiments have so far been published, and they overwhelmingly report null effects of various platform interventions and policy changes on a plethora of attitudinal outcomes, including political polarization. In this paper, we reanalyze the data from these four experiments given their unique value and their prominence in debates on social media and politics. We investigate the possibility that the limited pre-registered analyses in the original studies produced null results by chance. We focus on two outcome variables – affective polarization and issue polarization – and employ multiverse analysis, which seeks to identify a set of justifiable data manipulation and analysis choices and re-estimate treatment effects under each combination of those choices. We then use permutation tests to perform formal hypothesis tests on this “multiverse” of results. We conclude that the null results originally reported by the four published articles are generally robust to a variety of researcher degrees of freedom. Consistent with the small effect sizes originally reported, however, we find suggestive evidence that deactivating Instagram led to a substantively small decrease in affective polarization, which appears to be driven mainly by less negative views of partisan opponents. More generally, this paper provides a model for future researchers to incorporate multiverse analyses within pre-registered experiments. 

Endogenous Rebellion Benefits and Political Survival (with Zhaosong Ruan)

Presented at MPSA 2023

Governments and politicians often face survival threats from multiple actors, where the success of a rebellion depends on the coordination and perceived benefits of these actors. While these actors share a common goal of overthrowing the government, they have conflicts regarding the distribution of post-rebellion benefits. This paper investigates how a government can strategically choose institutions to exploit these conflicts, ensuring its survival. By initially implementing institutions that disadvantage certain "weaker" challengers and later transitioning to institutions that empower them, the government can delay rebellions. The timing of this transition is crucial; transitioning too early or late can render the strategy ineffective. This framework sheds light on political dynamics in various contexts, such as the economic and political shifts observed in China. The government's ability to manipulate perceived benefits, as seen with China's economic reforms and subsequent reversals, can play a pivotal role in delaying potential rebellions and ensuring governmental longevity. 

Work in Progress

Government Incentives in China’s Land Market Auction (with Yifan Yang)

This paper studies three government incentives in China's land market auctions that diverge from revenue maximization: corruption incentives, financial pressures, and long-term development objectives.  Delving deep into China's land auction dynamics, we explore how local governments balance immediate financial gains with broader economic and developmental goals. Utilizing a diff-in-diff approach, we evaluate the impact of central inspection teams introduced during the 2013 anti-corruption campaign. Our findings indicate that their presence results in more transparent auctions, characterized by lower deposit rates and higher success rates. Furthermore, we characterize a positive correlation between local government debt and final auction prices, suggesting a shift in auction strategies based on financial imperatives. Lastly, our analysis underscores the pivotal role of major developers, such as Wanda Plaza, in shaping subsequent auction prices, highlighting the government's incentives to guarantee the future developmental potential of their administrative lands.

Civil Conflicts and Ethnic Partitioning at Subnational Boarders, Evidence from Africa (with Yifan Yang)

Does ethnic partitioning within states reduce conflicts? Drawing from the unique historical context of Africa's European-designed borders, we leverage this exogenous factor to examine the implications of dividing ethnic homelands on conflict dynamics. We build a theoretical model to show that 1) when an ethnic group is split by a country's national border, it increases investment in military power by both the incumbent and opposition groups; 2) when the same ethnic group is further partitioned by the subnational border,  the opposition group is deterred from investing heavily in violence. Furthermore, we empirically show that inefficient border design leads to an increase in conflicts primarily when an ethnic homeland is split by a country border but NOT further divided by a provincial border. This result remains consistent across various measures of civil conflicts, including the number of incidents, fatalities, and the overall intensity of civil wars.