Governments increasingly govern new policies through pilots: temporary, bounded trial programs whose results inform later nationwide policy. Song and Li (2024) built a framework in which two attributes of a pilot — ambiguity of means and compatibility with existing policy systems and adopter beliefs — map onto four management strategies (experimentation, refinement, upscaling, institutionalization), and illustrated it with China's New Rural Cooperative Medical Scheme (NRCMS, 2002–2013). The framework, however, rests on a single longitudinal case, admits that the same attribute configuration can support alternative responses, and leaves open how further pilot attributes (relative advantage, salience, ease of installation, incentive structure, political context) interact with the two framework attributes. The problem is to establish, through systematic evidence across multiple policy-pilot cases rather than a single case, the conditions under which governments select, combine, or sequence each management strategy.

Background

A policy pilot is a deliberately bounded, usually temporary program in which a government trials a new policy design in selected localities before deciding whether and how to extend it; in China this mode of 'experimentation under hierarchy' (Heilmann 2008) has been a core governance technique since the reform era, and the national policy process has repeatedly grown out of prominent local experiments. The New Rural Cooperative Medical Scheme (NRCMS), launched from 2002, is a canonical example: a rural basic health-insurance pilot that was progressively extended nationwide and eventually merged into the regular medical-insurance system by around 2013.

Pilot management sits at the intersection of two research traditions. Implementation research explains policy implementation by the ambiguity of a policy (how many competing understandings of goals and means exist) and the conflict it provokes — Matland's (1995) ambiguity–conflict model is the classic statement, predicting different implementation 'varieties' from the two dimensions. A second tradition studies pilots and other temporary organizations (projects, task forces, pilot programs) as organizing settings of their own, where learning, evaluation, and scaling decisions replace routine administration. Song and Li (2024) — whose earlier work had traced how the celebrated Sanming healthcare-reform pilot was upscaled (Li and Song 2023) — joined these traditions by adapting Matland's model for pilots: they replace conflict with compatibility — the fit between the pilot and existing policy systems and the beliefs of potential adopters — and cross it with ambiguity of means (many open questions about the instruments needed to reach a given goal). The four cells yield four ideal-type management strategies: experimentation (high ambiguity, low compatibility — promote varied local trials to find a workable design), refinement (high ambiguity, high compatibility — continuously adjust instruments and rules using accumulating evidence), upscaling (low ambiguity, low compatibility — push geographic extension with subsidies, hierarchy, and training), and institutionalization (low ambiguity, high compatibility — convert the pilot into a formal, routine public policy). Their NRCMS chronology maps the four strategies onto four successive stages between 2002 and 2013.

The framework's own limitations define the open problem. First, a two-attribute model may not capture pilot governance: management can also depend on relative advantage (the perceived benefit of the pilot relative to alternatives), salience (public and stakeholder attention), ease of installation, incentive structures, and the political context, and how these interact with ambiguity and compatibility is unknown. Second, within any single quadrant, several management choices remain possible — postponement, coercive rollout, administrative improvisation, freezing — so the framework describes what a government did in one case rather than predicting what it will choose; a systematic explanation of strategy selection is missing. Third, strategies can overlap, revert, or combine rather than following a linear sequence, and how to analyze such mixed configurations is unsettled. Fourth, the evidence base is one longitudinal case built from roughly 200 documentary sources and 16 interviews, so the framework's generalizability across pilots is untested; establishing it would require cross-case evidence, for instance from qualitative comparative analysis (QCA — a set-theoretic, small-to-medium-N method that identifies which configurations of conditions are sufficient or necessary for an outcome) or large-N studies. QCA is already established for neighboring pilot questions — pilot site selection in China's public hospital reform and innovation-city program, and pilot scaling-up in the water sector — but has not been applied to the management-strategy framework itself.

Problem Statement

Establish a systematic, generalizable account of policy-pilot management-strategy selection, going beyond the single longitudinal case through which Song and Li (2024) built their framework. Concretely: (1) determine under what conditions — in terms of a pilot's ambiguity of means and compatibility with existing policy systems and adopter beliefs, and of further attributes such as relative advantage, salience, ease of installation, incentive structure, and political context, both individually and in interaction — a government managing a policy pilot adopts experimentation, refinement, upscaling, or institutionalization, or instead an alternative response such as postponement, coercive rollout, administrative improvisation, or freezing; (2) specify how these conditions behave when strategies overlap, revert, or run in parallel rather than in the framework's linear sequence; and (3) test the resulting selection account across policy-pilot cases with systematic comparative evidence so that the ambiguity–compatibility framework's predictions are corroborated, refined, or falsified beyond the New Rural Cooperative Medical Scheme. These three components jointly serve one objective: replacing a one-case interpretive framework with a validated explanation of how governments steer pilots.

References

  1. Song, Yunpeng, and Li, Yanwei (2024). How are policy pilots managed? Findings from the New Rural Cooperative Medical Scheme in China. Journal of Public Policy 44(1):143-163. DOI: 10.1017/S0143814X23000338. https://doi.org/10.1017/S0143814X23000338 (published online 23 October 2023; print issue 2024)
  2. Song, Yunpeng, and Li, Yanwei (2024). Upscaling policy pilots: Four tales in Chinese health governance. Public Policy and Administration 41(2):202-222. DOI: 10.1177/09520767241247987. https://doi.org/10.1177/09520767241247987
  3. Yang, Xiao, and Yang, Guang (2026). What determines the success or failure of policy experiments?—A qualitative field study of China's healthcare reform pilots. Frontiers in Public Health 13:1748984. DOI: 10.3389/fpubh.2025.1748984. https://doi.org/10.3389/fpubh.2025.1748984 (published online 29 December 2025; formal publication date 29 January 2026)
  4. Ye, Zhipeng, Gao, Hongyi, and Wu, Weixu (2024). Local government capacity and response strategies to central policy piloting goals: a case study of China's new-type urbanization plan. Humanities and Social Sciences Communications 11(1):1648. DOI: 10.1057/s41599-024-04170-3. https://doi.org/10.1057/s41599-024-04170-3
  5. He, Alex Jingwei, Fan, Yumeng, and Su, Rui (2024). Unpack the black box of pilot sampling in policy experimentation: A qualitative comparative analysis of China's public hospital reform. Governance 37(3):803-824. DOI: 10.1111/gove.12804. https://doi.org/10.1111/gove.12804
  6. Jiang, Linxiu, and Li, Quan (2025). Dynamic selection of pilot: decoding the underlying logic of China's innovation policy experimentation. Journal of Chinese Governance 10(4):532-557. DOI: 10.1080/23812346.2025.2547516. https://doi.org/10.1080/23812346.2025.2547516
  7. Jiang, Haochen, and Ramesh, M. (2025). Steering health reform through policy stewardship: experience from Sanming, China. Health Economics, Policy and Law 21(3):349-363. DOI: 10.1017/S1744133125100285. https://doi.org/10.1017/S1744133125100285
  8. Jing, Shiying (2026). Protection without employment: the policy logic, implementation, and reform of China's occupational injury protection pilot for platform workers. Trends in Social Sciences and Humanities Research 4(4):60-71. DOI: 10.61784/tsshr3238. https://doi.org/10.61784/tsshr3238
  9. Matland, Richard E. (1995). Synthesizing the Implementation Literature: The Ambiguity-Conflict Model of Policy Implementation. Journal of Public Administration Research and Theory 5(2):145-174. DOI: 10.1093/oxfordjournals.jpart.a037242. https://doi.org/10.1093/oxfordjournals.jpart.a037242
  10. Heilmann, Sebastian (2008). From Local Experiments to National Policy: The Origins of China's Distinctive Policy Process. The China Journal 59:1-30. DOI: 10.1086/tcj.59.20066378. https://doi.org/10.1086/tcj.59.20066378
  11. Li, Yanwei, and Song, Yunpeng (2023). State-led pilot innovation upscaling? A tale of the Sanming healthcare reform in China. Local Government Studies 49(2):375-396. DOI: 10.1080/03003930.2021.2009804. https://doi.org/10.1080/03003930.2021.2009804
  12. Nair, Sreeja, and Howlett, Michael (2015). Scaling up of Policy Experiments and Pilots: A Qualitative Comparative Analysis and Lessons for the Water Sector. Water Resources Management 29(14):4945-4961. DOI: 10.1007/s11269-015-1081-0. https://doi.org/10.1007/s11269-015-1081-0
  13. Nair, Sreeja (2021). Rethinking Policy Piloting. Cambridge University Press (Cambridge Elements). DOI: 10.1017/9781108885867. ISBN 9781108840392. https://doi.org/10.1017/9781108885867