In my dissertation, I investigate the strategic behavior of politicians engaging in corruption. Specifically, I consider how local government officials in Malawi choose between different forms of corruption, and how anti- corruption interventions such as those that increase transparency or sanctioning capacity condition these choices. I argue that the corruption space has two dimensions: a detectability dimension, capturing the degree to which a given form of corruption is detectable; and an immediacy dimension, capturing the likelihood that citizens will be affected by a given form of corruption enough to sanction it. My primary hypothesis is that corrupt political officials strategically respond to anti- corruption interventions by displacing their corruption along these dimensions to other forms, not by reducing their overall involvement in corruption. These displacement effects are not random, but occur in predictable patterns shaped by policy. In studying corruption dynamics, I employ four data sources from Malawi, a country with widespread low-level corruption yet active anti-corruption efforts. A citizen survey and media content analysis demonstrate that citizens in Malawi have diverse avenues for sanctioning corruption and that citizen willingness to enact sanctions varies across corruption forms. I generate an index of the sanctioning capacity of citizens across the districts of Malawi using data from the Anti-Corruption Bureau and National Statistics Office. A first person, nationally representative survey of local government officials provides a measure of corruption. Finally, I conduct a survey experiment on district officials that manipulates the level of top-down and bottom-up transparency they face. I provide support for the corruption displacement hypothesis that anti-corruption interventions cause some forms of corruption to become less prevalent and others to become more prevalent. Specifically, transparency interventions displace corruption to low-detectability forms. Officials exposed to a bottom-up transparency intervention experience greater displacement along the immediacy dimension. The displacement effect is greatest when the transparency intervention occurs in an area with high sanctioning capacity. This research sheds light on how transparency and sanctioning capacity can be used to influence government official corruption choices outside elections