Abstract by Dean Sobczak
A Bayesian Regression Model for Foreign Aid Allocation
By many economic indicators, Malawi is one of the least developed countries in the world. As a result, it receives millions of dollars per year from many public and private donors. Researchers are interested in better understanding how efficiently these donations are dispersed throughout the country; i.e., they want to know how responsive these donors are to the actual needs of the people to which they are contributing. Gains in computing power in recent years has allowed for greater use of Bayesian statistics to model such problems. We use the multivariate normal and Dirichlet-multinomial distributions in conjunction with the Gibbs sampling method and Metropolis-Hastings algorithm to understand the relationship between seven economic indicators and the amount of money given by a donor to a particular district and project sector in Malawi.