Madagascar, a biodiversity hotspot, faces significant deforestation driven by poverty and food insecurity. The country's governance system is fragile, frequently experiencing political crises that exacerbate environmental challenges. This study addresses the urgent need to understand which conservation interventions, CFM or state-managed protected areas, are more resilient during and after political crises. Community forest management (CFM) has been promoted as a more equitable and effective alternative to state-managed protected areas, particularly in remote areas where local communities might have a stronger capacity for monitoring and enforcement. Conversely, state management might be superior when local institutions are weak or incentives for local conservation are insufficient. The existing literature presents mixed results on the relative effectiveness of CFM versus state-managed protected areas, with some studies suggesting that CFM can reduce deforestation more effectively, while others emphasize the role of various biophysical, economic, social, and political factors in determining which approach is better suited to particular contexts. The novelty of this study lies in its investigation of the relative performance of CFM and MNP during a specific political crisis (2009–2014), using a counterfactual approach to account for potential confounding factors influencing deforestation rates. This approach helps isolate the effect of the crisis and post-crisis period on the performance of each conservation intervention.

The literature on conservation interventions during political crises is limited. Existing studies primarily focus on the impact of armed conflict, with mixed results. Some research indicates that local institutions can effectively mitigate deforestation even during conflict, while others show an increase in deforestation during periods of instability. Studies examining the broader sociopolitical context and its influence on conservation outcomes highlight the role of governance, democratic institutions, and corruption in shaping deforestation rates. For instance, research suggests that deforestation rates are higher in less democratic nations with weak rule of law and higher levels of corruption. In contrast, other studies show a positive effect of elections on conservation efforts, but this effect isn't consistent across different countries or contexts. Previous research on Madagascar has shown increased deforestation during periods of political instability, underscoring the importance of this study's focus on the impact of the 2009–2014 crisis on the relative effectiveness of CFM and MNP.

This study utilizes a quasi-experimental approach combining statistical matching and an event study design to assess the relative performance of CFM and MNP in reducing deforestation during and after Madagascar's 2009–2014 political crisis. The researchers used remotely sensed data on forest cover change from 2000 to 2020, aggregated to 90m and 270m spatial resolutions. The analysis focused on CFM and MNP areas established before 2005 to ensure a sufficient pre-crisis baseline. Statistical matching was employed to create comparable groups of CFM and MNP forest grid cells, controlling for time-invariant confounding factors such as remoteness, agricultural suitability, elevation, slope, and proximity to roads, villages, and urban centers. The event study design, implemented using OLS regression with fixed effects for each forest grid cell, allowed the researchers to control for time-variant confounding factors, including annual rice prices, climate variables (precipitation, temperature, wind speed, drought severity), and population density. The interaction terms in the model allowed for the assessment of the differential impacts of the crisis on CFM and MNP performance over time. The study also examined the robustness of its findings by testing different spatial resolutions and considering the subset of CFM areas where contracts were renewed. Additionally, tests for heterogeneity of the impacts were performed based on remoteness, population density, development level, and security conditions.

The analysis revealed that annual deforestation rates were consistently higher in CFM than in MNP areas throughout the study period (2005–2020). Specifically, deforestation rates in CFM were approximately three times higher than in MNP. A significant acceleration in deforestation rates was observed immediately following the crisis (2014–2017), particularly within CFM. Statistical matching successfully controlled for many confounding factors; after matching, CFM and MNP areas exhibited a high degree of similarity in pre-crisis deforestation trends. The event study results showed that during the crisis (2010–2013), both CFM and MNP performed similarly poorly in terms of reducing deforestation. However, in the post-crisis period (2014–2017), CFM performed significantly worse than MNP, with annual deforestation rates 1.7–2.4% higher, even after accounting for confounding factors. This difference represents a substantial ecological impact, considering the size of the CFM areas. The difference in performance between CFM and MNP was less pronounced after 2017, suggesting potential variation in recovery rates among CFM. The study found that CFM areas further from urban centers performed relatively better in the post-crisis period. However, even in remote areas, CFM still performed worse than similarly remote MNP areas. Overall, the findings suggest that community-managed forests are less resilient to intensified deforestation pressures in the post-crisis period compared to state-managed protected areas.

This study's findings have significant implications for conservation strategies in politically unstable regions. The unexpected finding that both CFM and MNP performed similarly poorly during the crisis highlights the pervasive impact of the crisis on forest protection efforts. The lack of capacity and resources, combined with weak governance, likely hindered effective enforcement of conservation regulations in both types of areas. The significant difference in post-crisis performance between CFM and MNP, with CFM showing greater vulnerability to increased deforestation pressures, challenges the common assumption that community-based conservation is always more resilient. The study suggests that the absence of centralized financial support for CFM in Madagascar may have contributed to their heightened vulnerability during and after the crisis. The observed post-crisis increase in deforestation could be attributed to various factors, including a lagged response to the crisis, lingering effects of weak governance and corruption, and potential impacts of post-crisis elections. Future research is needed to disentangle the complex interplay between political instability, policy changes, and deforestation dynamics. The study's results emphasize the need to enhance the resilience of both state and community-managed conservation mechanisms to political and economic shocks. Specifically, CFM in Madagascar might benefit from increased external support to improve their capacity to manage deforestation pressures, especially during and after crises.

This study demonstrates the vulnerability of community-based forest conservation in Madagascar during and after a significant political crisis. While both CFM and MNP suffered during the crisis, CFM showed notably lower resilience in the post-crisis period when deforestation pressures intensified. The findings underscore the importance of considering the potential impacts of political instability on conservation efforts and the need for strategies that strengthen the resilience of both state and community-based conservation approaches. Future research should explore the role of factors such as tenure security, community engagement, and external support in enhancing the effectiveness of conservation interventions during times of crisis.

The study acknowledges limitations related to data availability and potential unobserved confounding factors. The remotely sensed data used might not fully capture selective logging, which could affect the accuracy of deforestation estimates. The study’s focus on de jure designations of CFM and MNP, rather than de facto management practices, might limit its ability to fully understand the nuances of on-the-ground implementation. The analysis also focused on a single political crisis in one country, limiting the generalizability of the findings to other contexts. Despite these limitations, the study’s rigorous methodology, employing statistical matching and an event study design, provides valuable insights into the relative resilience of different conservation approaches during and after political crises.