TY - THES A1 - Palaric, Enora T1 - From symbolic politics to policy learning? A comparative case study of policy changes in the context of impact assessment in France, Germany and Italy N2 - This thesis examines the causal process that led to policy change in the field of impact assessment (IA) in France, Germany and Italy since the mid-2000s. I seek to understand whether changes to the goals and/or ends of IA may be explained by symbolic politics or by a domestic learning process. Additionally, I aim to find out how and under which conditions policy learning results in policy change, and how the policy process affects the type of policy change that is adopted. Based on the policy diffusion literature, the policy learning framework and the Multiple Streams Framework, I operationalise three different causal mechanisms: symbolic politics, problem-oriented learning and power-oriented learning. Using document analysis and expert interviews, I trace the causal process that resulted in nine selected cases of policy change. I find that in most cases policy changes resulted from a domestic learning process, whereby domestic entities defined policy problems and/or goals, acquired and diffused knowledge, and adopted measures to implement and enforce the policy change. This shows that France, Germany and Italy have gone beyond the symbolic phase in the implementation of IA. In Germany and Italy, recent policy changes may be explained by problem-oriented learning. Domestic entrepreneurs identified problems in the implementation of IA and aimed at improving the performance of the instrument. In these two countries stable organisations with a mandate to evaluate IA were the main policy entrepreneurs. They facilitated the acquisition of knowledge on IA, acted as learning fora, and largely contributed to disseminating knowledge and setting the policy agenda. With respect to how and under which conditions policy lessons may be translated into policy change, findings show that this depends on the ability of policy entrepreneurs to use windows of opportunities such as changes of government to translate the acquired knowledge into policy change. I find that because the type of causal mechanism affects who learns and what is learnt, problem-oriented learning tends to result in incremental change of the means of IA, while power-oriented learning is more conducive to redefining the ends of IA. T3 - Dissertations submitted to the Hertie School - 18/2021 Y1 - 2021 UR - https://opus4.kobv.de/opus4-hsog/frontdoor/index/index/docId/4200 UR - https://nbn-resolving.org/urn:nbn:de:kobv:b1570-opus4-42006 N1 - Shelf mark: 2021D018 + 2021D018+1 ER -