Approaches to the study of poverty and environmental impacts of conservation interventions

Reducing poverty and halting biodiversity loss are two crucial global goals. Protected areas (PAs) are an important example of how goals to reduce poverty and halt biodiversity loss interact. PAs aim to conserve biodiversity, but also have impacts on poverty. In this thesis I focus on the environmental and social impacts of protected areas using a suite of large datasets. Specifically, I focus on how our understanding of PA impacts can be improved by (1) assessing heterogeneity in more detail, (2) comparing impacts relative to impacts of other land uses, and (3) by using better data to study impacts in countries with currently insufficient data. In chapter 2 I assess how Nepali PAs influence poverty, extreme poverty, and inequality. I find that Nepali PAs reduced overall poverty and extreme poverty, and crucially, did not exacerbate inequality. I also find that tourism was a key driver in poverty reductions, but PAs also reduced extreme poverty in areas with few tourists. In chapter 3 I compare PA impacts relative to competing land uses and find that sustainable use PAs, agriculture and mining have led to different outcomes in forest cover and poverty in the Brazilian Amazon. I also show that PAs were effective in reducing deforestation compared to larger-sized landholdings, but not smallholders. I also show evidence that mining sites had more deforestation, but that mining sites also raised local income. In chapter 4 I test whether machine learning methods can be informative to estimate poverty in Tanzania using publicly available satellite imagery. I find that our machine learning methods can be used to estimate household consumption fairly accurately, but cannot be used to measure poverty change or multidimensional poverty. Combined, my findings highlight that PAs can reduce poverty and protect forests although impacts are highly heterogeneous and further scrutiny of PA impacts is needed in more countries. Novel methods using publicly available secondary data show promise to drastically improve the evidence base of PA impacts in data-poor countries where poverty is most prevalent.