Micro effects of aid projects in Afghanistan: A pre-analysis plan

Title	Micro effects of aid projects in Afghanistan: A pre-analysis plan
Post date	10/10/2018
C1 Background and Explanation of Rationale	A growing literature in the social sciences (with great interest from policy makers) has investigated the impacts of aid projects in conflict zones on state-building, violence and economic growth. Building on the existing work, there remain numerous unanswered questions in this literature, primarily related to three issues. • Citizen attitudes in outlying, conflict-affected regions are often hypothesized about, but seldom analyzed in detail. • Previous studies focus on \$\$ spent rather than actual aid outputs. • The vast majority of aid spending and projects are non-randomized, leaving us with nasty observational data that are difficult to unpack. Although there have been a few RCTs, they have been restricted (appropriately, for ethical reasons) to the least intensely affected conflict areas. We endeavor to resolve each of these three by (1) leveraging new geo-coded citizen surveys to explore the impacts of projects on public attitudes, (2) using new data from the German BMZ, explore the impacts of verified aid outputs rather the projects or spending, and (3) deploying a pre-analysis plan, we can inoculate ourselves from charges of fishing, p-hacking or cheating more broadly.
C2 What are the hypotheses to be tested?	Existing theories of aid and insurgency suggest a few unconditional effects. A nearby aid output should generate, in the short run: • increased access to the relevant public service • increased citizen confidence in the state • improved economic well-being, real and perceived • decreased sympathy and support for armed anti-government groups • improved attitudes towards foreign intervenors Scholarship to date suggests a complicated relationship with violence outcomes: • Aid spending in contested or rebel-held areas should increase insurgent violence\footnote{This can be via predation, which can directly cause violence, or indirectly, by increasing insurgents capability for future violence.} o via project sabotage and retribution/coercion towards collaborators • Aid spending in government-held areas may reduce insurgent violence o via changes in citizen attitudes and expectations • Citizen perceived security likely moves in a similar way as the above measures of violence
C3 How will these hypotheses be tested? *	We specify a difference-in-differences design for the evaluation and provide justification for the parallel trends and exchangeability assumptions. We will only use outcome data accessed after the design has been formally registered with EGAP in order to strengthen the credibility of the analysis (i.e. no p-hacking or fishing). We write down our hypotheses and protocol for exploring heterogeneous or unexpected results.
C4 Country	Afghanistan
C5 Scale (# of Units)	148 clusters, 744 sampling points ~ 4,300 respondents
C6 Was a power analysis conducted prior to data collection?	Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval?	not provided by authors
C8 IRB Number	n/a
C9 Date of IRB Approval	n/a
C10 Will the intervention be implemented by the researcher or a third party?	Researchers
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research?	Yes
C12 If relevant, is there an advance agreement with the implementation group that all results can be published?	Yes
C13 JEL Classification(s)	F35, F51, F52, F63