Projects first. Zhuang's team, with over 30,000 project records from more than 2,000 villages, counts nearly 40,000 projects in 2009 to 2012 and about 365.5 million USD over 2009 to 2011, around 22 USD per villager per year; two thirds of the money went to infrastructure (village roads, drainage, irrigation, water), 27 percent to welfare services, about 5 percent to training, education and recreation, with remoter villages choosing more roads. Cabannes 2018 reports over 100,000 projects and 1.2 billion USD since 2008. On governance, the early-years reports converge: Zhuang and Wang describe money that no longer depends on cadres' lobbying of upper bureaus, villagers who see the fund as theirs, and transparency above and inside the village; Chen Yiyuan's cadres call the council a shield that lets majority decisions proceed against holdouts, and recall villagers clearing land for roads unasked; Frenkiel finds a new, triply legitimate layer of village representatives who constrain the party secretary, the committee and township influence, at the price of leaders' discretion, with consequences confined to the village. The same authors report the reverse. Zhuang (2014) saw many councils retrograde: convened by the party branch, questionnaires filled by cadres, project details unpublished, elites allying with representatives. Chen Yiyuan (2018 fieldwork in one township) found 60 to 80 percent of each fund locked into sanitation, patrols, reserve cadres and street lighting, assessed tasks from above crowding out roads villagers still asked for, and a governance team of over 30 in a village of 2,000; He and Gui (2022) say the council's decisions became a procedure and the fund lost its mobilising character. Reduced conflict and petitioning is the stated purpose and is claimed by the vice-mayor and by Jiang and Huang (no instability during eight years of reform) rather than measured. The only causal evidence is the 2021 experiment in 349 communities, urban and rural, with a designed deliberation-plus-vote process and an app: six months in, treated residents were 12.2 points more likely to volunteer, 10.7 to donate, 7.0 to vote in neighbourhood committee elections and 4.7 to attend policy meetings, with no change in petitioning or speaking to officials, effects largest for low-risk behaviours; they wanted more improvement from the central government (1.80 versus 1.61 on a 4-point scale) but not from local government, and were more satisfied with national policy (4.91 versus 4.79 on 7), prouder of their city and more sure the government hears people like them. Those findings are about the 2021 community design, not the village scheme as it ran from 2009.