3. Aid and Conict [30 points] You can make additional assumptions to clarify your interpretation of any of the questions if you wish, but state your assumptions clearly. (a) Nunn & Qian (2014) show that an increase in food aid from the U.S. causes an increase in civil conict in recipient countries. Explain this nding using (i) the opportunity cost theory of conict, and (ii) the state as prize theory of conict. You may consider the possible effects on the income and incentives for different members of the population (e. g., farmers, non-farmers living in cities, the government, and rebel group leaders). (b) Now, suppose Nunn & Qian are wrong, and food aid has no effect on conict. Politicians in Wheat-growing areas of the United States argue that the government should reallocate foreign aid from building new roads to providing more food aid to poor countries. Using the Solow model, what do you expect will happen to the level and growth rate of income per capita in the short in response to a persistent reduction in road investment (all else equal)? Which type of aid do you think is more likely to increase incomes in poor countries: infrastructure investment or humanitarian assistance like food aid? (c) There is a lot of controversy among economists as to whether aid actually causes economic growth or whether the relationship is biased due to omitted variables and reverse causality. You notice that the World Bank, which is a major aid donor, stops providing foreign aid to countries when their GDP per capita cross a threshold of $3,000. How could you use this rule to identify a causal effect of foreign aid on economic growth? ((1) Some donors argue that foreign aid will help reduce conict because as countries get richer, they are less likely to engage in civil war. Suppose you are asked by the US Aid Agency to estimate the effect of income on conict in poor countries. Someone suggests using the Ebola outbreak in Africa as an instrumental variable for income. Provide two reasons Why an Ebola outbreak might not be a good instrument for income in such a study. (e) A foreign aid agency strongly suspects that poor farmers will be less likely to participate in violence if they had higher income. The agency offers to fund a randomized controlled trial offering cash transfers in order to test their hypothesis. They have a limited budget and argue that you do not need a baseline (pre-intervention) survey to recover a causal effect in the randomized evaluation design. Do you agree? Why or why not