## Donor Herding and Domestic Debt Crisis

## Source details

**Canonical URL:** [Donor Herding and Domestic Debt Crisis](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06109.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06109.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06109.pdf.json)

---

### Model overview
- Purpose: Present a modified loan-pushing model (based on Basu (1991)) to show how donor herding can trigger a domestic debt crisis in a low-income country (LIC).
- Herding mechanism emphasized: rational herding due to payoff and information externalities (following Devenow and Welch (1996) and Vives (1993)).
- Key insight: Donor herding can cause a reversal of aid flows; if actual domestic debt exceeds the government’s repayable threshold, a domestic debt crisis can follow.

### Basic model details
- Borrower announcement: The LIC announces (Lt, it) each period t, where Lt is the desired loan amount and it is the interest rate offered.
- Lender supply unit: Each lender supplies either one unit of credit or nothing.
- Expected excess supply and reservation rate:
  - rj = rj(StE). (equation (1))
- Total supply specification:
  - St = St(StE, it). (equation (2))
  - Signs: ∂St/∂StE ≥ 0 and ∂St/∂it ≥ 0.
- Rational-expectations supply:
  - St = St(etY – Lt, it). (equation (3)) where etY is expected supply of funds.
- Lender’s expected return:
  - etρ = (it − πtθ − γtλ)e. (equation (4))
  - Parameters and domains:
    - πt is the default parameter for the LIC and lies between 0 and 1.
    - θ is a given constant (πtθ represents default costs to the lender).
    - γt ∈ (0,1) is the cost of acquisition of information to the lender (lower γt → higher expected return).
    - λ is a given constant.
- Supply function in rational-expectations form:
  - St = St(etY − Lt, (it − πtθ − γtλ)e). (equation (5))
- Threshold/corner nature of supply:
  - The supply of new loans follows the discontinuous specification in equation (6): supply occurs only if both etY and (it − πtθ − γtλ)e exceed thresholds Y and ρ* respectively; otherwise supply is zero.
  - Y is the threshold level of the supply of funds; ρ* is the threshold level of the lender’s return.
  - Model implication: a small decrease in lender return can cause supply to drop to zero (extreme/corner herding outcome).

### Generating a domestic debt crisis (donor-herding scenario)
- Donor withdrawal scenario:
  - Consider T-zero aid flows to the LIC (T is the number of periods without aid after herding). From period t+1 to t+T, there is no supply of new loans: St+1, St+2, ..., St+T = 0.
  - The operative condition for zero supply from equation (6): etY and (it − πtθ − γtλ)e must satisfy YYori e t e tt t <<−−*)(ρλγθπ (as stated in the text).
- Domestic interest-rate response:
  - Government borrows domestically in response, inducing rising real domestic interest rates dkt ri+ and higher accumulated domestic debt ψt+k.
- Accumulated domestic debt and repayment threshold:
  - Three possible period outcomes (from equation (7); notation preserved as in source):
    - Positive value is if kt+ ψ is above * kt+ π,
    - Negative value is if kt+ ψ is below * kt+ π,
    - Equals if kt+ ψ = * kt+ π.
  - Interpretation: a debt crisis occurs once kt+ ψ > * kt+ π; no crisis if kt+ ψ ≤ * kt+ π.
- Determinants of actual domestic debt accumulated:
  - kt+ ψ = f(dkt ri+, (DS t+k x dkt ri+ )2). (equation (8))
    - dkt ri+ is the real interest rate prevailing on the domestic financial market in period t+k.
    - DS t+k is a dummy equal to 1 if there is no supply of foreign aid in period t+k and 0 otherwise.
    - The square of the interaction (DS t+k x dkt ri+ )2 enters positively: 0)(/>∂∂ ++ dktktkt xriDSψ when DS t+k = 1.
    - When DS t+k = 0, kt+ ψ is a decreasing function of dkt ri+: 0/≤∂∂ ++ dktkt riψ.
- Policy response role:
  - If the government reduces fiscal expenditures and borrows a well-calculated sum domestically, the first two outcomes (no crisis) can obtain.
  - If the government does not reduce fiscal expenditures in face of donor withdrawal, the third outcome (domestic debt crisis) is likely.

### Empirical research implications and data needs (conclusion)
- Main empirical questions proposed:
  - Are changes in domestic debt stocks linked to sudden shifts in external aid inflows exacerbated by donor herding?
  - Has donor herding occurred (i.e., are changes in one donor’s flows heavily influenced by other donors)?
  - Does the growth rate of domestic debt beyond a certain threshold get explained by a decline in donor aid below a certain threshold, after controlling for macroeconomic variables?
  - Why, when faced with imminent or actual cuts in donor assistance, did the government not cut expenditure or raise revenues? (A suggested reason: belief in the “Good Samaritan’s” dilemma.)
  - What are the consequences of government behavior for domestic economic outcomes?
- Data challenges highlighted:
  - Need for data on individual donor financing commitments and disbursements with sufficient frequency (monthly or quarterly) and over a sufficiently long period.
  - Domestic debt data by stocks and flows, by type and holder, are required but often unavailable or incomplete in LICs (beyond banking system credit, treasury bills, and bonds).

*Source: IMF Working Paper WP/06/109 — "Donor Herding and Domestic Debt Crisis" (April 2006).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp06109.pdf_