This dissertation studies three major macroeconomic development policies: (1) Reducing internal and international barriers to trade, (2) The sustainability of cost-recovery in microfinance programs, and (3) The impact of foreign aid assistance. I divide my dissertation into three chapters. The first chapter quantifies the size of intra-country versus inter-country trade barriers and assesses the impact on patterns of trade and welfare. I develop a quantitative multi-sector international trade model featuring non-homothetic preferences in which states trade both within country and internationally. I discipline the model using rich micro data on foreign and domestic trade flows at the Indian state level and price dispersion both within and across states. I find that: (1) state-wise price data predict internal trade flows well; (2) internal trade barriers make up 30% of the total trade cost on average, but vary substantially by state depending on the distance to the closest port; and (3) the welfare impacts of domestic integration are substantial; reducing trade costs across states to the U.S. level increases welfare by 15% compared to a welfare gain of 7% when fully eliminating international import barriers. The second chapter studies the scalability and financial sustainability of microfinance programs. It provides a theoretical and experimental evaluation of a cost-reducing innovation in the delivery of 'Self-Help Group' microfinance services, in which privatized agents providing services earn payment through membership fees. Under the status quo, agents are paid by an outside donor and offer members free services. Theoretically, we show that membership fees can improve performance without sacrificing membership by mitigating adverse selection. In our randomized control, the innovation provides similar levels of services over time, but cost the donor less. Moreover the innovation provides greater benefits (borrowing, saving, and investing in business) by catering to more business-oriented households. The third chapter assesses the impact of untied versus tied aid on the governance and GDP per capita of developing countries. I develop a principal-agent model with a conflict of interest between the donor and the recipient. The donor provides foreign aid for development purposes whereas the recipient prefers to retain some foreign aid for unproductive private benefits. To overcome this conflict of interest, aid can be tied to public capital investment, although this is costly. Due to its irreversibility, tied aid can lead to lower corruption levels and higher per capita consumption in developing countries. I find that tied aid is especially effective for corrupt governments. In terms of optimal aid delivery, tying aid is more cost-effective for corrupt countries while untied is more cost-effective for benevolent countries.