Despite Economics increasing interest in the effects of conflict, our empirical knowledge has lagged behind. Recent studies have highlighted important and varied long run costs to conflict. Due to a lack of research on how households adapt to conflict and insecurity, however, we have little understanding of how these costs emerge and little evidence with which to design policies. This dissertation contributes by examining the responses of households to insecurity, particularly during conflicts. While other authors point to the importance of insecurity, insecurity has never been measured. The dissertation introduces a methodology to create spatially disaggregated measures of conflict risk and uses these to present the first estimates of the relative contributions of insecurity and exposure to violence to the aggregate household costs of conflict. While the dissertation leverages several unique datasets, I show that more widely available data can be used to similar effect. Having established that insecurity accounts for at least half of the costs in rural Northern Uganda, I examine household livelihood responses along both intensive and extensive margins. Lastly, using a ten period panel with the first direct household subjective perceptions of insecurity, I investigate the impact of these perceptions on income and find that these effects differ throughout the year, likely reflecting the heterogeneity of income generating opportunities throughout the year. Moreover, I study the effect of prior perceptions in determining current beliefs and find that prior beliefs have little economic significance. Rather, agro-meteorological conditions, particularly pasture quality (NDVI), drive changes in subjective perceptions of insecurity. The primary contribution of the dissertation is to underline the importance of responses to insecurity in determining the costs to conflict. Despite the focus on exposure to violence, the majority of the aggregate costs of conflict arises from responses to insecurity and may occur irrespective of the presence of violence. More broadly, by focusing on areas and households which experience conflict, current studies conflate the effects of the risk and realization of violence. The results here suggest that the large and varied lasting costs of conflict likely can be traced back to household responses to insecurity.