Working Paper Number 116
April 2007

What Have IMF Programs With Low-Income
Countries Assumed About Aid Flows?
By David Goldsbrough and Ben Elberger


Background Note for the CGD Working Group on
IMF-Supported Programs and Health Expenditures




Abstract

This paper examines the nature of aid projections in IMF programs with low-income countries. On
average, IMF projections of net aid increased sharply in the first year of programs but tapered off in
subsequent years. Projections were also significantly more optimistic in countries with low initial
levels of aid but differed little across regions. Most notably, projections of net aid to countries in
Sub-Saharan Africa following the Gleneagles Summit are significantly more pessimistic than the
path implied by commitments to double aid to Africa by 2010. This pattern is strong throughout the
group with only two Sub-Saharan African countries showing increases in net aid consistent with the
Gleneagles commitments.

We argue that much greater clarity is needed about the role of the IMF in the aid architecture. In
addition to projecting likely aid flows based on detailed discussions with donors, the IMF should
utilize sector-level inputs to assess the macroeconomic effects of a significant scaling-up of aid in
programs with low-income countries. Such a scenario would help the international community and
the country itself judge whether there are any macroeconomic constraints to absorbing more aid. The
obvious benchmark to use for aid levels in such a scenario would be what donors have committed to
globally--i.e. a doubling of aid in the case of African countries. Finally, we conclude that the IMF
should be more transparent about what its collective program projections imply for the expected path
of global aid flows.




The Center for Global Development is an independent think tank that works to reduce global poverty and
inequality through rigorous research and active engagement with the policy community. Use and
dissemination of this working paper is encouraged, however reproduced copies may not be used for
commercial purposes. Further usage is permitted under the terms of the Creative Commons License. This
paper was made possible in part by funding from the Bill & Melinda Gates Foundation. The views
expressed are those of the author and should not be attributed to the directors or funders of the Center for
Global Development.

www.cgdev.org

---

What Have IMF Programs With Low-Income
Countries Assumed About Aid Flows?1

2
David Goldsbrough and Ben Elberger

The issue of what projections for aid the IMF should incorporate into its programs is
controversial. The IMF position has generally been that it should make the best possible
estimates of aid flows that are likely to be delivered, drawing on detailed discussions with donors
about their intentions, but that it would be counterproductive to base programs on overly
optimistic assumptions. For example, the Managing Director of the IMF, Mr. Rodrigo De Rato,
said in a recent speech: “it doesn’t do low-income countries any favors to pretend that they will
receive more aid than they actually will—in a world where commitments still usually exceed
disbursements.” 3 In contrast, some critics argue that the IMF is overly cautious in its
projections, thereby sending a signal to potential donors that lower volumes of aid would be, in
some sense, sufficient.4 These critics call for the IMF to play more of a catalytic role, by
incorporating more optimistic projections into its baseline programs as well as through
macroeconomic assessments of scenarios that incorporate an ambitious scaling-up of aid. These
differences of view reflect sharply contrasting perspectives about what signals the IMF should be
sending to the international community about aid. Indeed, there is still a lack of clarity within
the IMF itself about exactly what is expected of IMF staff in their work on aid-dependent
countries, an issue we will return to in the concluding section.

In the context of this broader debate, this paper examines what the IMF has actually assumed
about net aid flows in the programs it supports. We focus on the baseline projections (not any
alternative scenarios) and ask the following questions:

• What has been the profile of net aid flows assumed in IMF programs with low-income
countries?

• Has this profile changed over time or for particular groups of countries, especially Sub-
Saharan Africa?

• How do countries’ starting positions influence the IMF projections? (For example, is aid
to countries with relatively low per capita aid assumed to grow faster?)

1 This paper is one of the background papers prepared for the Working Group on IMF –Supported Programs and
Health Expenditures under the Center for Global Development’s Global Health Policy Research Network. A
description of the project is available at http://www.cgdev.org/section/initiatives/_active/ghprn/workinggroups . JEL
Codes: F33, F35, F37. Keywords: International Monetary Fund, foreign aid, ODA, projections, Gleneagles Summit,
macroeconomic frameworks, macroeconomic programs, Africa.
2 David Goldsbrough is a senior fellow and Ben Elberger a research assistant at the Center for Global Development.
3 Rodrigo de Rato. “Renewing the IMF’s Commitment to Low-Income Countries” (speech, Center for Global
Development, Washington, DC, July 31, 2006).
4 Oxfam International, The IMF and the Millennium Development Goals: Failing to deliver for low-income
countries, Report. September 2003.

1

---

• How do the projections of aid in recent IMF programs compare with donors’ global
commitments, such as at the Gleneagles G-8 Summit, and have these projections become
more optimistic post-Gleneagles?

Our focus is on what the IMF has assumed, not on actual outcomes. Earlier evidence suggests
that, compared to outcomes, IMF program projections have not been unduly pessimistic. For
example, the 2004 evaluation of the PRSP and PRGF by the Independent Evaluation Office of
the IMF concluded that, on average, program design allowed for larger external financing flows
than actually occurred.5 But more recent evidence for Sub-Saharan Africa suggests that initial
IMF programs may have under-predicted aid for the medium term.6 Of course, such comparisons
cannot answer the broader questions of whether the IMF should be playing a more proactive role
in mobilizing additional aid and what the impact might be on aid levels.

To address the questions listed above, we looked at projections of net aid over the medium term
in two sets of IMF programs: (i) all original IMF-supported programs under its concessional
lending facilities—the Enhanced Structural Adjustment Facility ( ESAF) and its successor, the
Poverty Reduction and Growth Facility (PRGF) during the period January 1997 through January
2007 (77 programs in 46 countries);7 and (ii) all original programs and any reviews (i.e., any
time the IMF formally revisited its aid projections) in the 18-month period before and after the
July 2005 G8 Gleneagles Summit.

Details of the data sources and the countries and programs included in each data set are given in
Appendix I.

For the purposes of this paper, net aid is defined as grants plus gross concessional loans less
amortization actually paid (i.e., after taking account of debt service flow relief). Extracting this
information from IMF reports proved more difficult than we had expected, since the treatment of
some components of aid flows—especially debt relief—is complex and varies substantially
among different country reports.

Aid Projections in Original Programs

The broad profile for projected aid has been remarkably similar across different groups of
countries. For all programs, aid is projected to expand substantially in the initial program year
(t0) (by 26 percent for the entire sample) before tapering off by the third year (t2) (Table 1 and
Figure 1). This pattern holds true for projections of aid in Sub-Saharan African countries. The
table shows both median and unweighted means (excluding three outliers)8 and the broad pattern
is the same for both measures.


5 Independent Evaluation Office, International Monetary Fund, Evaluation of the IMF’s Role in Poverty Reduction
Strategy Papers and the Poverty Reduction and Growth Facility, 2004. See Chapter 4, page 51.
6 Independent Evaluation Office, International Monetary Fund, An Evaluation of the IMF and Aid to Sub-Saharan
Africa, 2007. See especially Annex B.
7 Some programs could not be included either because the program documents have not been made public or
because the projections did not cover the full 3-year period on which we focused.
8 Sri Lanka (2003 PRGF), Central African Republic (1998 ESAF), Ethiopia (1998 ESAF)

2

---

Table 1. Net Aid Flow Projections (Normalized to Net Aid at t-1 of 100)
Region
Program Vintage

Average Projections


Average Projections

Category
t-1
t0
t1
t2
N
Category t-1
t0
t1
t2
N
All
100 130 129 118 74 ESAF 100 155 131 112 27
SSA
100 134 131 120 46 PRGF 100 114 128 122 47

Median Projections


Median Projections

Category
t-1
t0
t1
t2
N
Category t-1
t0
t1
t2
N
All
100 116 112 102 77 ESAF 100 126 110 100 29
SSA
100 117 115 102 48 PRGF 100 106 113 103 48
Figure 1. Projected Net Aid Growth Over All Programs
160
140
) 1
120
t
t-
0 a
100
to 10
ed
liz
80
ma
60
(Nor
Aid
t
40
Ne
20
0
t-1
t0
t1
t2


t-1 t0 t
Year
1 t2
Average
Median

Under the PRGF, projected aid grew more steadily through the second year of the projection but
also tapered off by the third year (Table 1). By the end of the program, aid projections for PRGF
programs generally exceeded those for ESAF programs.9 The increase in projected aid between
the two groups of programs was much larger for the means than the medians, indicating that the
increases in aid projections under the PRGF were not uniformly distributed.

We examined the influences of several factors on the medium-term program projections for aid,
using the percentage change in net aid over the 3-year period as the dependent variable. (PRGF

9 Our data set only includes the last two years of ESAF-supported programs. Comparisons using the full sample of
all ESAF arrangements have found a larger change between the ESAF and PRGF. (See, for example, IEO (2004),
op cit). Conclusions on any shifts in aid profiles between the ESAF and PRGF also depend substantially on whether
the changes are measured from the year in which the program was concluded (t0) or the preceding year (t-1).

3

---

arrangements are for an initial period of 3 years and this period is the focus of most program
projections for the “medium term”.) The results indicate that the IMF program projections
implicitly assumed that countries with low initial levels of aid per capita would benefit from
faster growth in aid (Table 3 and Figure 2). Dummy variables testing the importance of
geographic location (Sub-Saharan Africa) and shift from the ESAF to PRGF were not significant
at the 5% level.10 Quantitatively, a 1-percent lower initial net aid per capita was associated with
a 1-percentage point increase in projected growth in net aid over the 3-year period. This result
was, statistically, highly significant. In other words, IMF program projections implicitly assume
that countries with lower aid per capita will receive a gradually increasing share of global aid.

Figure 2. Net Aid per Capita and IMF Aid Projections
1200%
1000%
t 2
800%
to
t -1
r
om
600%
d f
400%
200%
Change in Net Ai
0%
0.01
0.1
1
10
100
1000
-200%
Ln (Aid per Capita at Time t-1)













10 As noted earlier, this may reflect the fact that our sample only includes the “late” vintage of ESAF-supported
programs.

4

---

Table 2. Multivariate Model Correlates of Projected Net Aid Growth

Percent Change in Net Aid Over Program Period
-0.9644
-1.0171
Ln (Aid per Capita)
(-6.03)***
(-6.32)***
-.5206
SSA Dummy

(-1.89)*
0.0574
PRGF Dummy

(.21)
β
3.4863
3.9461
(6.53)***
(6.32)***
R2
0.3267 0.3593
F
35.3850 13.6449
N
74 74
Notes: t-statistics in parentheses * Significant at the 10% level ** Significant at the 5% level *** Significant
at the 1% level



How Do IMF Aid Projections Compare with Gleneagles Commitments on Aid?

At the July 2005 Summit at Gleneagles, G8 leaders made a number of commitments that
indicated a more optimistic environment for official development assistance. For example, the
Chair’s summary of the Summit included the following statement, “We have agreed to double
aid to Africa by 2010. Aid for all developing countries will increase, according to the OECD, by
around $50 billion per year by 2010, of which at least $25 billion extra per year for Africa.” 11
This statement was part of a broader trend of global commitments to expand aid, in total and to
Africa in particular. To investigate how IMF program projections for aid have responded to this
changed environment, we examined all publicly available program projections for aid (in
original programs or reviews) made for countries with PRGF arrangements in the 18 months
before and after the Summit.12

The results indicate four key conclusions about IMF aid projections:

• Over the medium-term, the global profile of aid projections in IMF programs is
substantially less optimistic than the Gleneagles commitments. Weighted by the
initial level of aid, average growth in net aid in IMF programs is projected to be
broadly consistent with Gleneagles commitments in the first and second year of
projections but diverges starkly in the third year (See Figure 3 and Table 3). 13 This
pessimism over the medium term is attributable, in part, to declining aid projections

11 For further details, see the official website of the summit at www.g8.gov.uk
12 Excluding Lesotho (Sixth Review, Pre-G8 Summit) and Azerbaijan (Fifth Review, Pre-G8 Summit). Lesotho is
treated as an outlier due to its extremely low initial net aid flow in year t-1 (fiscal year 2003, $1.3 million).
Azerbaijan is treated as an outlier due to high aid volatility resulting from lumpy amortization payments.
13 These projections use the assumptions and figures of the OECD-DAC Secretariat Simulation of DAC Members’
Net ODA Volumes in 2006 and 2010 (http://www.oecd.org/dataoecd/57/30/35320618.pdf)

5

---

to Afghanistan but is also broad-based with only 37% of post-Gleneagles Summit
projections more optimistic than the path implied by the Gleneagles commitments.

Table 3. Weighted IMF Aid Projections and the Gleneagles Aid Path14

t-1
t0
t1
t2
Percent Change over 3-Year Period
N
Implied Gleneagles Aid Path: Global Aid
100 112 124 135
35%

Implied Gleneagles Aid Path: Aid to SSA
100 116 132 149
49%

All
Pre-G8
Summit
100 119 121 119
19% 44
All
Post-G8
Summit
100 114 122 121
21% 43
SSA Pre-G8 Summit
100
115
117
115
15% 30
SSA
Post-G8
Summit
100 114 123 121
21% 27
Non-SSA Pre-G8 Summit
100
136
141
138
38% 14
Non-SSA Post-G8 Summit
100
115
118
120
20% 16



Figure 3. Weighted IMF Net Aid Projections Prior to and
Following the 2005 Gleneagles Summit
140
130
) 1
t t-
0 a
120
ized to 10
110
r
mal
o
100
Aid (N
Net
90
80


t1-1
2
t0
3
t1
4
t2
Year
IMF Pre-Gleneagles
IMF Post-Gleneagles
Implied Gleneagles Path


• There is a wide distribution in aid projections. While the average profile of
projected aid in IMF programs is not consistent with the Gleneagles commitments,
there is a wide dispersion of aid projections across countries (Figure 4). Nearly 39%
of aid projections project country aid to decrease over the medium term and 24%

14 Weighting is conducted using the level of net aid at t-1 of the projection period. Unless otherwise noted, all
averages reported are unweighted.

6

---

project aid to increase but by less than implied by the Gleneagles path. Of those that
increase at a pace consistent with or faster than the Gleneagles commitments, a small
number are projected to receive large increases in aid (i.e. Moldova, Armenia,
Zambia).

Figure 4. Distribution of IMF Net Aid Projections Prior to and
Following the 2005 Gleneagles Summit
7
6
5
y
4
nc
r
e
que
3
F
2
1
0
0%
0%
0%
%
%
%
%
%
%
%
%
%
0%
%
-3
0
-2
-1
10%
to
20
30
40
50
60
70
80
90
10
150%
200%
200
an
to
to
to
to
th
0%
%
% to
% to
% to
% to
% to
% to
% to
% to
ss
0%
0%
-10%
10
20
30
40
50
60
70
80
than
90
0% to
0% to
Le
-3
-2
10
15
eater
Gr
Projected Aid Growth over Three-Year Projection Period
Pre-G8 Summit
Post-G8 Summit


• The profile of IMF aid projections for countries in Sub-Saharan Africa has not
changed much in the post-Gleneagles period and, on average, remains well short
of the path that would double aid by 2010. Mean projected growth of net aid over
the 3-year period increased by only 4 percentage points and remained well below the
implied growth of Gleneagles Summit statements (Table 3 and Figure 5). Of the 27
IMF programs and reviews in Sub-Saharan Africa that were completed in the post-
Gleneagles period, only two were more optimistic than the Gleneagles projections of
aid to Africa.15,16



15 The two cases were Mozambique (Fourth review in June 2006) and Zambia (Third review in December 2005).
16 We use OECD DAC-reported ODA to Africa listed in Roodman’s (2005) Net Aid Transfers dataset as the
baseline for the Implied Gleneagles Aid Path to SSA.

7

---

Figure 5. Weighted IMF Net Aid Projections in Sub-Saharan African
Countries Prior to and Following the 2005 Gleneagles Summit
160
150
) 1 140
t
t-
100 a 130

to
d
lize 120
orma
110
i
d (N
e
t
A
100
N
90
80


-1
t
0
1
2
-1 t0 t1 t2
Year
SSA Pre-Gleneagles
SSA Post-Gleneagles
Implied Gleneagles Path
Implied Gleneagles Path to SSA


• For countries outside of Sub-Saharan Africa, the weighted mean projected path
of aid is closer to a path consistent with the Gleneagles commitments, although
there was actually some deterioration in the levels of aid projected before and
after the Summit (Figure 6). This reflects the relatively small number of countries
involved and the particular circumstances of Afghanistan (the latter entered into a
new program with the IMF in 2006, but aid is projected to decline over the medium
term). On an unweighted basis, projected aid to countries outside Africa is projected
to increase substantially (Figure 7).


8

---

Figure 6. Weighted IMF Net Aid Projections in Countries Outside Sub-
Saharan Africa Prior to and Following the 2005 Gleneagles Summit
150
140
)

t -1 130
at
0
120
zed to 10
a
li
r
m 110

Aid (No 100
Net
90
80
1
2
3
4


t-1 t0 t1 t2
Year
Non-SSA Pre-Gleneagles
Non-SSA Post-Gleneagles
Implied Gleneagles Path
Figure 7. Unweighted IMF Net Aid Projections in Countries Outside Sub-
Saharan Africa Prior to and Following the 2005 Gleneagles Summit
180
170
160
) -1
150
at t
00
1 140
to
130
a
l
i
zed
120
(Norm
Aid 110
Net
100
90
80
-1
0
1
2


t-1 t0 t
Year
1 t2
Non-SSA Pre-Gleneagles
Non-SSA Post-Gleneagles
Implied Gleneagles Path

9

---

Document Outline

---