While protected areas (PA) remain a key conservation strategy globally, their performance is likely shaped by the socio-political context in which they exist. Although decentralization is a good example of such a contextual phenomenon in multiple locations globally, it is rare to find quantitative empirical analyses of how it moderates PA effectiveness. We evaluate how the decentralization in Indonesia (proxied by the introduction of direct elections, district head (bupati) change, and district splitting) influenced PA effectiveness between 2000 and 2008. Focusing on three outcomes - deforestation, forest fragmentation, and fires, we apply a quasi-experimental approach to a carefully constructed spatially explicit village-level panel dataset, combined with geospatial biophysical and physio-geographic data. We hypothesize that the moderating influence of decentralization on PA effectiveness depends on whether decentralization increased threats to forests, strengthened local accountability, or weakened enforcement. On average, we find direct elections improved the PA impact in terms of reducing deforestation in protected villages, but had no statistically significant effect on forest fragmentation, fires, or leakage. On average, we find district splitting increased forest fragmentation in the recently protected villages, but had no statistically significant impact on deforestation and fires. On average, we find the bupati change had no statistically discernible influence on PA impacts on deforestation, fragmentation or fires. Given the increasing threats to forests due to decentralization, these results imply that district splitting and the bupati change weakened enforcement inside PAs with regards to deforestation and forest fragmentation, in contrast to direct elections. By highlighting the potential channels through which decentralization in Indonesia impacted forests, we offer insights into the effectiveness of a common conservation policy in the country. Broadly, we contribute to the conservation impact evaluation literature by quantitatively examining how political economy influences the performance of conservation policies.

虽然保护区（PA）仍然是全球范围内的一项重要保护战略，但其表现可能受其所处的社会政治环境影响。尽管权力下放是在全球多个地点发生这种情况的一个很好的例子，但很少能找到定量的经验分析以了解权力下放如何影响PA的有效性。我们评估了印度尼西亚的权力下放（通过引入直接选举，地区负责人（布帕蒂）变化和地区分裂）在2000年至2008年之间影响了PA的有效性。针对三个结果（森林砍伐，森林破碎和火灾），我们将准实验方法应用于精心构建的空间清晰的村级面板数据集，并结合了地理空间生物物理和自然地理数据。我们假设，权力下放对PA有效性的适度影响取决于权力下放是否增加了对森林的威胁，加强了当地问责制或削弱了执法力度。平均而言，我们发现直接选举在减少受保护村庄的森林砍伐方面改善了巴勒斯坦权力机构的影响，但对森林破碎，火灾或渗漏没有统计学上的显着影响。平均而言，我们发现地区分裂在最近受保护的村庄中，森林破碎化加剧，但对森林砍伐和大火没有统计学上的显着影响。平均而言，我们发现布帕蒂人的变化对PA对森林砍伐，破碎或火灾的影响没有统计学上可辨别的影响。鉴于权力下放对森林的威胁越来越大，这些结果表明，与直接选举相比，地区分裂和人为改变都削弱了巴勒斯坦权力机构内部在森林砍伐和森林破碎方面的执法。通过强调印度尼西亚的权力下放影响森林的潜在渠道，我们提供了对该国共同保护政策有效性的见解。从广义上讲，我们通过定量检查政治经济学如何影响保护政策的执行情况来为保护影响评估文献做出贡献。