This thesis explores the role of human behaviour in shaping livelihood resilience to covariate shocks at the individual and collective level from a multidimensional perspective using two case studies. In Chapters 2, 3, and 4, I study livelihood resilience in the case of rice smallholders cropping in flood-prone areas of Ecuador. In Chapter 5, I study livelihood resilience in the case of banana smallholders facing the threat of Banana Xanthomonas Wilt disease in Rwanda. I use the Adaptive Cycle as the theoretical spine of this thesis to understand livelihood resilience. In Chapter 1, I introduce the overall research purpose, questions, and design. Concepts and theories related to social capital, risk, and social dilemmas are integrated into the research across the different chapters. In Chapter 2, we investigate what is resilient and to what is it resilient? through a participatory approach and disaster risk management theoretical lens. The specific research question is, how does local understanding of risk shape collective (or re-organization) actions, in terms of function and expression forms, to face a covariate shock at the household and community level? To answer this question, we developed and applied a participatory resilience assessment, from a risk perspective, where users define what is at risk, why it is important, and how its probability of being affected by a risk should be measured and interpreted. We found that the meaning of being at risk was directly associated with local (individual and collective) strategies to cope with a shock relied on social capital relationships. Further, local involvement in the assessment tool design resulted in a rapid engagement and understanding of the tool. The results showed that all groups evaluated were at high or very high risk. Despite the high degree of risk, rice cropping is a livelihood that has persisted through generations; therefore it can be said it exhibits some degree of resilience (either healthy or pathological). In Chapter 3, we explore why is it resilient. Specifically, we ask, what are the critical factors, related to social capital, that contribute to the strengthening, weakening, or hindering of collective actions to prevent, respond, or prepare for a covariate shock? We develop a conceptual framework that integrates concepts of resilience, disaster risk management (DRM), and social capital to explore this question. We focus on coping strategies that are based on social capital relationships. Then, we operationalize the framework through a tool that assesses if social capital contributions to transit from a shock to a recovery stage. Our results show that some forms of self-organization do not always lead to a healthy resilience, especially when there is a high dependency on external resources. Opportunistic and power-unbalanced dynamics take place when locals cannot sustain strategies of sharing and exchanging resources to cope collectively. Therefore, the more and diverse are the local resources, the healthier (in terms of resilience) are the dynamics to cope with the shock. In Chapter 4, we ask how do smallholders make sense of their cooperative or defective behaviour in a shock situation, and how does such a sense-making process link to their livelihood resilience? Specifically, we explore how people make sense of their cooperation toward a public good (a saving-box) that is meant to strengthen the resilience of rice livelihoods at a local level. We use a public goods game as a tool to create a shared experience, rather than an experiment. A focus group discussion followed the game to make sense of when and why farmers choose to cooperate in different idiosyncratic and covariate shock scenarios. We found that farmers consider there to be no ideal and safe situations, therefore their choices are based on the assumption that there could always be an idiosyncratic shock. Nonetheless, in that scenario, they choose to cooperate (through repaying a loan) to the conservation of the public good. However, under covariate shocks, such as floods, they choose not to cooperate toward the conservation of the public good but rather prioritized saving their resources to be able to participate in local exchanges of food, water, labour, or other resources that are critical to cope collectively with a generalized shock. In Chapter 5, we explore why a livelihood is resilient from a public bad perspective. The specific research question was, how does the dynamic interplay of socio-ecological factors of a livelihood system influence household collective action to prevent or control a threat in common that risk their livelihoods’ resilience? To answer this question, we develop a conceptual framework for analysing a public bad risk threatening livelihood resilience, in the context of a socio-ecological system from a risk and collective action problem perspective. Then, we design a dynamic socio-ecologic game methodology, which adds SES attributes and its emergent phenomena to the game experience. Next, we design a context-specific game, the Musa game, to explore farmers’ cooperation to prevent and control Banana Xanthomonas Wilt (BXW) disease spread. Finally, the game is tested with actual farmers, and results analysed through a developed spatial statistical tool. We found that factors such a technical knowledge, risk perception, social trust, and institutional trust influence decision making (in terms of what to do, where, and when), collective action, and public-bads management. Furthermore, this study also introduces the game as a powerful tool for studying technical aspects of the diseases and showing the importance of collective action to prevent the disease. In Chapter 6, I synthesize and discuss the main contributions of this thesis. In the context of shocks, this thesis suggests that risk perception, coping capacity, institutional environment, and social dilemmas all have important influences on individual choices, and therefore collective actions. At the same time, these decisions are influenced by (and influence in turn) the biophysical factor, from which different expressions of resilience emerges. Therefore, livelihood resilience can be seen as a consequence of the sum of dynamic, interdependent, and synergic individual actions, providing either a joining benefit (by producing or preserving a shared good) or harm (by failing in preventing or controlling a public-bad).