Chapter 1 studies the electoral response to the Ghost Buildings program, a nationwide anti tax evasion policy in Italy which used innovative monitoring technologies to target buildings hidden from tax authorities. The difference-in-differences identification strategy exploits both variation across towns in the ex ante program scope to increase enforcement as well as administrative data on actual building registrations. Local incumbents experience an increase in their reelection likelihood as a consequence of the policy. In addition, these political returns are higher in areas with higher speed of public good provision and with lower tax evasion tolerance, implying complementarity among enforcement policies, government efficiency, and the underlying tax culture. Chapter 2 uses a road-level regression discontinuity design in Sierra Leone to study the impact of improvements in rural road infrastructure on agricultural markets.

We show that the improved roads reduced the market prices of local crops. These price effects are stronger in markets that are further from major urban centers and in less productive areas. We also find that these price effects are reversed in areas with better cell phone penetration. We show that our empirical findings are consistent with a search cost framework a la Mortensen, but inconsistent with other models, such as Bertrand competition, bilateral bargaining, and Cournot oligopsony. Chapter 3 present results from a randomized controlled experiment designed to study the multiple margins through which value is passed from traders to agricultural producers in the presence of interlinked transactions. Consistent with other studies, we find limited price pass-through in response to an increase in the trader resale price. However, there is a large response in credit provision.

We develop a model of interlinked transactions that highlights the substitutability of price and credit pass-through across markets, and verify its predictions empirically. Calibration suggests that to ignore margins of pass-through other than price has substantial implications for welfare analysis.