[eng] Fighting poverty and protecting the environment are two of the most urgent challenges facing the international community at the start of the 21st century (United Nations, 2015). One aspect that will be crucial in this challenge is aligning environmental, energy and development policy in order to create a triple dividend between the three policy fields. Additionally, integrating environmental protection policies with poverty reduction strategies is by no means a new concept, but as we continue to think about these challenges, it is important to link environmental and economic components together through a type of development that is economically feasible, socially desirable and environmentally benign. Within each of these challenges lies different possible interventions, but space must be created for public policy to be at the center of these challenges in order to generate the best available evidence and guide our decision making. This dissertation consists of three independent chapters that represent responses to these challenges, and provides empirical evidence that can guide specific policy recommendations in the fields of development economics, and environmental and resource economics. Universal access to modern energy services is central to the international sustainable development agenda. According to the International Energy Agency (2010, 2017), 1.4 billion people across the world did not have access to electricity, and 2.7 billion people still rely on traditional use of solid biomass. The international development agenda has placed an emphasis on improving access to electricity as well as made large investments in the energy sector throughout developing countries. It is argued that the universalization of access to electric energy in the world is of fundamental importance for the eradication of poverty and the reduction of social inequality (Kaygusuz, 2011). Despite the large investments made in the energy sector to increase rural electrification and the diffusion of modern cooking systems, relatively little is known about the impact such policies have on household well-being (Bonan et al., 2017). From an individualistic point of view, energy is a material prerequisite to achieving valued capabilities (Nussbaum, 2001). Energy is interconnected with the socio-economic and human development of the individual, and deprivations of energy interact with health and education in different ways to undermine an individual’s well-being. In recent years there has been a movement towards multidimensional methodologies built upon the foundations set forth by Alkire and Foster (2007, 2011) and Alkire and Santos (2010) as well as research focused specifically on multidimensional energy poverty (Nussbaumer et al., 2012). Thinking about energy poverty through a multidimensional lens, examines the underlying problems of different types of ‘poverties’ rather than just poverty, and asks the question of ‘poor in what’ in addition to ‘who is poor’. With the recent movement towards multidimensional analysis, parallel with resources being allocated towards energy programs, understanding how these two areas interact is imperative. Building upon the multidimensional framework set forth by Nussbaumer et al. (2012), the premise of the second chapter of this dissertation is to jointly analyze how multidimensional energy poverty deprivations affect different measures of education in Uganda. The first goal builds upon previous work by Nussbaumer et al. (2012) who postulate six energy components: 1) modern cooking fuel, 2) indoor air pollution, 3) access to electricity, 4) household appliance ownership, 5) radio or television, and 6) mobile phone. The sample population is then classified as to whether the respondent has access to electricity or not, and then whether the respondent has achieved other possible energy components within the MEPI. Exploiting the data in this manner demonstrates that access to electricity is unable to capture whether other household energy domains have been achieved. After motivating the need for the use of a multidimensional framework through descriptive statistics, the study moves to weight each of the energy domains. This is done through a Factor Analysis, which searches for inter-correlations amongst the variables within the index. Using the derived weights, the MEPI is constructed for each individual in the sample. Last the study applies the MEPI in a regression analysis and compares its effects to that of access to electricity on measures of education. In order to address methodological challenges of attribution, this study draws upon an IV strategy employed by Dinkelman (2011) that uses the land gradient of the community. Where this study differentiates from the previous empirical strategy is it will use the land gradient of rural households. This will be the first time that this IV strategy is employed at the individual/household level. I find that the MEPI improves upon our understanding as to the effects energy poverty has on measures of education. This is seen in the precision of the estimated coefficients and smaller standard errors as well as the ability of the MEPI to estimate significant results that access to electricity is unable to. The results further demonstrate that there are other important energy mechanisms beyond access to electricity that must be considered within an individual’s set of energy capabilities, and this may explain the insignificant or inconsistent findings of previous studies based on simpler indicators (like access to electricity). Analyzing household energy deprivations from a multidimensional perspective makes several important contributions to the literature on energy poverty. The first is the ability to quantitatively characterize the type of energy poverty an individual suffers from. The methodologies employed in this chapter better illustrate complementary input mechanisms beyond the singular dimension of whether an individual has access to electricity. Furthermore, considering only access to electricity may lead to a misdiagnosis of the true problem, which is that individuals may have access to electricity, but are unable to realize the benefits of the energy through appliances or modern cooking technologies. From a methodological point of view this misdiagnosis may introduce measurement bias into the analysis. The developed MEPI framework attempts to reconcile this problem and expand the literature’s ability to measure energy poverty. The second contribution comes from the use of Factor Analysis, which is a model based approach that seeks to reproduce the inter-correlations amongst variables and is focused on explaining the common variance across indicators instead of total variation. This methodology is employed to assign context specific weights to each of the components within the MEPI. This is the first time that energy poverty has been quantitatively characterized in such a manner in order to be used in causal modeling. Additionally, insights gained from the weighting analysis can be used by policy makers to prioritize different energy deprivations. Third a contribution is made through the use of an innovative IV that builds-on the work by Dinkelman (2011), who uses the local land gradient to instrument for the MEPI access to electricity. This analysis uses the land gradient of the household as an instrument, which provides a more accurate depiction of actual electricity access as the average local land gradient can hide the fact that some households in a community may or may not have electricity. Chapters 3 and 4 of this dissertation study the interrelated dynamics of development economics, environmental and natural resource economics. In Chapter 3, I focus on how development impacts the environment through the effects that a community-driven development program of small-scale infrastructure projects in the Philippines had on forest coverage. In Chapter 4, I analyze the opposite direction by focusing on how a changing environment impacts development. Specifically, Chapter 4 analyzes the impact of a policy intended to revitalize the mining sector in the Philippines in terms of an unintended increase in malaria cases. The loss of forest coverage is a global as well as a local and regional environmental concern. Globally deforestation represents around 9 percent of anthropogenic carbon emissions (Le Quéré et al., 2015). Local and regional impacts from deforestation can lead to land degradation such as a reduction in soil fertility, increased runoff into fisheries as well as a loss of biodiversity. Additionally, stemming deforestation in low-income countries is viewed as one of the most cost-effective solutions to reducing global CO2 emissions (Nabuurs et al., 2007; Stern, 2007). As developing countries continue to build infrastructure in parallel with their development needs, one challenge to confront is how to meet the development needs in a sustainable manner. One solution could lie in community-driven development (CDD) programs, which can best be characterized by the movement of responsibility over resources and planning decisions. This type of program supports a bottom-up approach to development by decentralizing the decision making process to the local level. Where the issue of deforestation and increasing frequency of CDD programs throughout the world come together, is that international donors and multilateral organizations are targeting CDD programs as a strategy for climate change mitigation and adaptation (Arnold et al., 2014). The objective of Chapter 3 is to examine whether the goals of fighting poverty and protecting the environment are in contradiction by asking whether development aid has unintended environmental effects in regards to deforestation, focusing on a CDD program in the Philippines. This question is derived from several areas. First is in regards to the scarcity of evidence on the environmental impacts of actions designed to reduce poverty, as well as in the opposite direction of the poverty impacts of actions designed to protect the environment (Alpízar and Ferraro, 2020). Second there is surprisingly very little evidence on the impact international aid has on the environment and more specifically forest coverage. Third, even as international donors and multilateral organizations position CDD programs with the parallel strategies of poverty reduction and climate change mitigation and adaptation, little empirical evidence exists on the effects CDD programs have on the environment and forest coverage. This chapter will address each of these areas by providing rigorous empirical evidence of the effects international aid has on deforestation. In this study, I utilize satellite-generated forest coverage data to analyze the effects development aid has on deforestation through a large-scale CDD program in the Philippines called KALAHI-CIDSS (KC). In order to disentangle this relationship, I exploit the manner in which the CDD program was allocated to municipalities through a regression discontinuity design (RDD) and a randomized control trial (RCT). The former leverages quasi-experimental variation, where the main identifying assumption is that municipalities on either side and close to the eligibility cut-off are systematically similar except that one received the program and the other did not. The later strategy leverages experimental variation between municipalities that were randomly selected via lottery to either be treated by the KC program or remain a part of the control for three years. I find evidence that indicates the KC program had strong and statistically significant effects on deforestation. Eligible municipalities in the RDD period experienced an average of 236 percent more deforestation and treated municipalities in the RCT period experienced an average of 265 percent more deforestation than the control. The paper makes several important contributions to the literature and the under-standing of the effects development aid has on the environment. First, each of the empirical strategies have the ability to overcome traditional concerns stemming from the non-randomness of aid allocation, in order to uncover a causal estimate of the effects development aid has on deforestation, and more specifically the effect that CDD programs have on deforestation. Second, each of the empirical strategies analyzes the same development program, but from two different time periods in which the Philippines experienced different levels of deforestation. Third is the scale of the program, since this study represents the largest evaluation of the environmental effects that a CDD program has had on deforestation. Additionally the Philippines offers a context that has substantial spatial heterogeneity in terms of economic, social and ecological diversity. Fourth, a rich dataset on subproject characteristics is exploited with information on the type of implemented subproject, the number of household beneficiaries, construction duration and the subproject cost. This information has yet to be exploited for a CDD program and additionally offers a unique dimension to understand how different types of subprojects such as infrastructure, education/health facilities, water/electricity, water protection, and support projects differentially impact the environment. Last, a contribution is made to whether higher levels of poverty lead to more deforestation relative to less poor areas. This refers to the environmental Kuznets curve or the contrasting view of the poverty-environment hypothesis, which remains an open debate in the environmental literature. The latter suggests that as income grows, even at low income levels, the surrounding environmental quality improves, while the former indicates the existence of a non-monotonic relationship where rising living standards first increases pressure on the environment and then later improve them. By exploiting the structure of the RDD, I am able to find limited support for the environmental Kuznets curve. Finally, Chapter 4 focuses on the unintended consequences of changes in development policy. It reverses the previous Chapter’s direction of thinking, from how development may impact the environment, to how policy-driven environmental changes may impact development and more specifically health outcomes. In this regard, land transformation and land clearance activities are likely to increase diseases, and roughly one quarter of the global burden of disease can be attributed to environmental changes (Prüss-Üstün et al., 2008). One disease that is susceptible to such environmental changes is malaria. This is due to the fact that cleared lands are generally more exposed to sunlight and prone to puddle formation with more neutral pH levels that can favor Anopheline larvae development (Patz et al., 2000) as well as a loss of biodiversity can reduce or eliminate species that prey on Anopheline larvae or Anopheles mosquitoes (Laporta et al., 2013; Yasuoka and Levins, 2007). Deforestation is one form of land transformation that has been shown to alter the disease ecology of malaria (Berazneva and Byker, 2017; Chakrabarti, 2018; Garg, 2019; Keesing et al., 2010; MacDonald and Mordecai, 2019; Pattanayak and Pfaff, 2009; Tucker et al., 2017). Another form of land transformation that could be potentially linked to the emergence and proliferation of malaria is through mining activities. These have received much less attention in the literature, which has focused on small geographical areas and localized effects of malaria (De Santi et al., 2016; Rozo, 2020; Valle and Lima, 2014) or on a corollary relationship between gold mining and malaria (Barbieri et al., 2005; Caselli and Tesei, 2016; De Santi et al., 2016). With this in mind, this chapter aims at analyzing whether there is an ecological response from mining activities, by investigating how a change in extractive resource policy in the Philippines led to more cases of malaria. In January 2004, the government of the Philippines launched the Minerals Action Plan (MAP) with the goal of revitalizing the mining sector. As a result, the policy change led to a reduction in the mining permit process between application and the grant of a permit from 3-5 years to 6 months in 2005 (Fong-Sam, 2005). Using the MAP reform, I exploit two sources of variation in the timing of the MAP reform as well as spatial variation in the distribution of mineral endowments through a difference-in-difference (DID) approach that compares provinces with and without gold deposits before and after the reform. The basic pathway in which gold mining is hypothesized to accelerate the reproductive environment of the Anopheles mosquito is through the process of leaving behind slow-moving bodies of water, which happen to be the common location of many gold mines. If the stagnant pools of water are left open, they can provide an ideal breeding site for the Anopheles mosquito to reproduce. I find evidence that is consistent with an ecological response, where the policy change to a more extractive resource position had a statistically significant effect on the incidence of malaria. More specifically, after the MAP reform, provinces with gold deposits had 32 percent more malaria cases relative to provinces without gold deposits. There are three main contributions this chapter intends to make. First, it moves beyond corollary results and provides causal estimates by exploiting the timing of the reform as well as the spatial distribution of geological endowments. Second, it provides the first nation-wide causal estimates of the impact gold mining has on the incidence of malaria. Third, it analyzes a policy that encouraged the expansion of the mining sector. This differentiates from the Rozo (2020) context in Colombia, where the bulk of the argument is placed on the fact that illegal gold miners do not comply with mining regulations and have limited knowledge on measures needed to protect themselves from malaria or safety measures to prevent the reproduction of malaria. Evidence from this chapter indicates that it is not necessarily a legal versus illegal issue, as a legal expansion of the resource extraction sector through the MAP reform also led to an increase in the incidence of malaria.