Competition and Buck-passing in the Political Economy of Aid

• Background and Motivation
• Concepts and Measures
• Data and Methods
• Results

• Builds on a broader theoretical research program.
• Theoretical understanding and empirical investigation of strategic interdependence in aid allocation.
• Here, focus on empirically identifying the conditions under which donor governments compete or pass the buck in bilateral aid allocation.

“Don't tell me what you value. Show me your budget, and I'll tell you what you value.” - Joseph R. Biden Jr.

• How donor governments distribute aid says more about their interests than leaders relate in speeches or memos.
• This extends to the question of how donors respond to each other as they target aid to developing countries.
• How donors respond to the aid given by others sheds light on the rival or common benefits donors obtain through their aid allocation.

• We still know very little about the conditions under which donors compete for rival foreign policy gains or pass the buck to others when pursing goals through aid.
• Existing evidence remains mixed.
• Some find evidence for competition (Dudley 1979; Mosley 1985; Barthel et al. 2014).
• Others find evidence of buck-passing (Rahman & Sawada 2012).
• And some find different results by donor or delivery channel (Steinwand 2015; Mascarenhas & Sandler 2006).
• Prior efforts fail to take steps to tease out when and where donor governments have rival or common objectives in allocating aid.

• I develop novel, theory-driven, measures to help identify variation in the degree of rivalry or commonality of donor goals.
• I use machine learning to conduct robust tests of heterogeneity in donor responses to each other on the basis of their goals in giving aid.
• I use classical econometrics to further identify how donor interests shape when they compete and pass the buck in giving aid to certain recipient countries.

• These goals imply different responses among donor governments.
• Development aid promotes a collective good \( \to \) free-rider problem \( \to \) negative response to where donors give aid.
• Aid-for-policy exchange promotes rival interests \( \to \) competition \( \to \) positive response to where donors give aid.
• How to tease out when and where one objective dominates the other?

• Donor Interest (DI)
• Recipient Need (RN)
• DI \( \times \) RN = Donor Responses

• Create composite measures of each.
• DI = Trade + Proximity + Alliance + Colonizer
• RN = Poverty + Size + Rights & Liberties + Conflict + Disasters
• I use an algorithm that finds a linear combination that is maximally correlated with each component.

• CRS data 1995-2014
• DAC ODA commitments across 24 sectors
• PWT, ATOP, CEPII, UCDP/PRIO

Hypotheses

• H1: Donors respond to the aid commitments of others.
• H2: Donors' responses systematically vary with respect to DI and RN.

Exploratory

• How do DI and RN interact to shape donor responses?

• Hypothesis testing:

  • Diagnostics on random forest predictions.
  • aid = f(Peer Aid, DI, RN, Donor, Aid Budget).
  • Permutation Importance (Fisher et al. 2019) and H-test (Friedman & Popescu 2008).

• Exploratory:

  • Linear model with:
  • Three-way interaction between Peer Aid, DI, RN.
  • Peer Aid instrumented with aggregate Peer DI.
  • Donor and Year FE

Do donors respond to peer aid?

Do strength of foreign policy interests and depth of recipient need condition this response?

How do these measures interact to shape donor responses?

• Question:

  • When and where do donor governments compete or pass the buck in aid allocation?

• Answer:

  • Donors compete where their foreign policy interests are strongest and recipient need is greatest.
  • Donors pass the buck where their foreign policy interests are minimal and recipient need is greatest.