Académique Documents
Professionnel Documents
Culture Documents
Abstract – Despite the highly significant flows involved, development aid has turned
out to be ineffective in a large number of recipient countries. This paper2 explores the
reasons behind aid failure. It starts from an investigation of donors’ motivations for aid
allocation and examines their potential implications on aid policies. The main findings
are that aid seems to be heavily impacted by inertia. Along with self-interest, this may
explain the time-inconsistency problem facing aid donors. Next, the paper examines
strategies toward mitigating such a time-inconsistency, including reputation signaling
and changing aid modalities. This provides guidance on how aid should be optimally
allocated among recipient countries.
Keywords: development aid, effectiveness, aid modalities, inertia, dynamic panel,
targeted infrastructure
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
1 W
HAT IS BEHIND THE INEFFECTIVENESS
OF AID?
Widespread conceptual and empirical literature suggests that foreign aid is inef-
fective (Alesina and Dollar, 2000; Bourguignon and Sundberg, 2007; Faye and
Niehaus, 2012; Qian and Nunn, 2014). The efforts of the international community
translated in the Monterrey consensus (2002) and the Paris Declaration on Aid
Effectiveness (2005), further supported by the Accra Agenda for Action (2008)
and the series of U.N. Summits for Financing Development (the latest took place
3. Monterrey Consensus (2002) has become the major reference point for development co-opera-
tion. It was the outcome the United Nations International Conference on Financing for Develop-
ment at Monterrey 2002. The Paris Declaration (2005) is an action-oriented road-map to improve
the quality of aid and its impact on development. The Accra Agenda for action (2008) is designed
to strengthen and deepen implementation of the Paris Declaration. It monitors progress and sets
the agenda for a sustained implementation of the Paris targets.
4. In this paper, I use interchangeably the terminology of development aid, foreign assistant or for-
eign aid to refer to official development assistance (ODA). OECD defines ODA as those flows of
official financing administered by official agencies with the promotion of the economic develop-
ment and welfare of developing countries as the main objective, and which are concessional in
character with a grant element of at least 25 percent.
5. Data on aid are collected from the OECD aid statistics portal. Those on GDP are collected from
the World Bank database.
62
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
In Sraieb (2015a), three categories of variables are used as proxies for the
donor motivations: (i) variables representing the need of recipient countries. these
translate the altruistic motivations of the donor, (ii) variables representing the do-
nor’s self-interest/strategic concerns, and (iii) variables proxying the merit-based
motivation of the donor. Each of the motivation categories makes use of a num-
ber of proxies as summarized in Table 1:
Interestingly, the merit-based variables are broken down into proxies for the
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
6. Recall that the “Washington Consensus” refers to a set of broadly free market economic prescrip-
tions supposed to be in line with policy advices by Washington, D.C.-based international organi-
zations.
7. A description of the content of each variable is provided in the appendix.
63
Mohamed Mounir Sraieb
1.2 W
hat is New?: Introducing Dynamics
and Economic Freedom Concerns
Accounting for inertia in the model and adding market liberalization concerns in
the merit-based motivations of the donor are central to this study and are a con-
tribution of my paper.
It is worth noting that most of the empirical studies on aid effectiveness are
cast in a static framework although it is widely understood that there may be
inertia in aid allocation. According to McGillivray and White (1993), a commonly
identified influence in this context is the tendency for aid bureaucracies, like other
spending agencies, to use the preceding year’s allocation as a benchmark for
the current year’s aid allocation in a process of marginal incrementalism or bu-
reaucratic inertia. In order to take into account such inertia, I introduce the lagged
dependent variable as a covariate in the model.
From an econometric perspective, introducing dynamics into the model is a
major advantage, since this allows to avoid a sizeable difficulty associated with
static models. Indeed, virtually most of the studies on aid motivations are based on
static models making use of one of three alternatives: a Tobit model, a Heckman
method or a two-part procedure. These techniques proceed in two stages:
–– the aid eligibility stage, in which the donor country decides whether a poten-
tial recipient country receives any aid at all; and
–– the level stage, in which it is decided how much aid to allocate to the coun-
tries that have been selected as eligible in the first stage.
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
64
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
view, more aid should be allocated to countries that perform better in terms of
quality of governance.
Surprisingly, the literature poorly measured the interest of donors for the
recipient’s implementation of sound economic policies and market-oriented re-
forms (market liberalization policies). This is anecdotal since market liberalization
policies were anchored in the missions’ statement of most Aid Agencies and are
a cornerstone of official declarations concerning the rational behind aid provision.
To address this shortcoming, I decompose the merit-based variables into
proxies for the quality of governance in the recipient country and proxies for mar-
ket liberalization in terms of the “Washington Consensus” requirements.
8. Aid conditionality means that donors attach conditions for granting aid. These conditions are
meant to ensure an environment that is favorable to aid effectiveness. Examples of such policies
include fighting corruption, promoting civil liberties and political rights, simplifying administrative
procedures, etc.
65
Mohamed Mounir Sraieb
Robust standard errors in parentheses (*** p < 0.01, ** p < 0.05, * p < 0.1)
Most striking here is that democratizers (recipient countries that perform well
in terms of promoting democracy as a perquisite for receiving aid) are not al-
located more aid. However, economic liberalizers (those who promote free mar-
ket reform are rewarded by the donor in terms of aid received. This puts a limit
on what a particular recipient country can expect from going more democratic
(promotion of human rights, civil and political liberties, etc.). These are among the
few factors over which the recipient has easy command in the short to medium
term.
66
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
1.4 M
ore on the results: the weight
of motivation categories in the donor’s decision
Next, I move beyond existing studies and investigate the relative importance of
the motivation vectors (need, merit and self-interest). The idea is to assess which
of the vectors is the most valued by the donor when allocating its aid across re-
cipient countries.
The exercise must be carried with due caution since the models considered
have different numbers of observations and different number of covariates. This
exercise should not be seen as a comparison between models in terms of their
goodness of fit. It is rather an appreciation of how close to the observations is the
prediction of each of the model. This gives a rough idea about the contribution
of the concerned vector of motivation to the fit of the model. That is, its contribu-
tion to the precision of the prediction on observed aid. It is in this sense that I can
conclude about the importance given by the donor to each of the merit, need or
self-interest vectors.
To proceed, I start first, with the basic model (containing only the time dum-
mies, “past AID” and the population variable). I then add sequentially the different
motivation vectors to the basic model. I therefore consider the following specifi-
cations:
–– the basic model augmented by the need variables (GDP per capita);
–– the basic model augmented by the merit variables (military expenditure,
democracy, PTS, trade freedom, financial freedom, and freedom from
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
Table 3 states that the basic model (past AID, population and time dummies)
returns a BIC of 403.166 (AIC: 354.933). Introducing the need vector, decreases
the BIC to 340.824 (the AIC decreases to 288.233) suggesting that the resulting
model fits better the data. Alternatively, adding the self-interest vector reduces
67
Mohamed Mounir Sraieb
further the BIC to 186.862 (the AIC further decreases to 126.527). I obtain a
model with a higher explanatory power. Finally, introducing the merit vector to the
base specification gives an even better model in terms of explanatory power. The
BIC drops to 125.91 (while AIC decreases to 54.829).
It is in this sense that I can claim that the merit-based motivation of the donor
country has more weight in the decision to allocate aid than any of the other mo-
tivation vectors. This is a finding that contrasts with the literature. In most models
on aid allocation, the donor’s self-interest motivation has a prominent role in the
aid allocation decision. I argue here that the merit dimension is the dominant vec-
tor of motivation for allocation and that the bias found in the literature toward the
self-interest motivation of the donor seems to be mainly due to a poor measure-
ment of recipients’ “merit”.
The point I make here is that the result made in the literature on the su-
periority of self-interest motivations of donors is surely biased since it is based
on several complementary proxies of donor interest (bilateral trade, votes in the
U.N., military assistance, etc.) compared to incomplete measurements of the
donor’s merit-based motivations. This is precisely the pitfall addressed in the
first part of these notes. I complement the measurement of merit-based motiva-
tions, by proxies for market liberalization policies, in the spirit of the «Washington
Consensus» requirements.
The results show that market liberalization performances explain more the
distribution of aid than does the need of the recipient or human rights concerns.
Furthermore, self-interest of the donor ranks high, far above donor’s altruism,
in the motivations driving aid allocation. This may explain the week degree
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
1.5 E
conomic Policy Implications:
Donors’ Reputation Building
Ultimately, the main message behind the investigation of aid motivations is two-
fold. For recipients, the results suggest that the best way to attract aid is to go
for market liberalization (trade freedom, financial freedom and economic free-
dom, in general). As for the donor, the conclusion is that there is no reason for aid
to be effective if it is based on pure inertia, on excessive altruism, or on self-
interest (if it disregards democratization or good governance and rewards geo-
strategic partners and allies). This creates a time-inconsistency problem on the
side of the donor and destroys incentives for effort provision on the side of the
recipients.
One efficient way to address the problem and retrieve effectiveness is for the
donor to invest in reputation. The extent to which this may solve the problem and
the conditions under which the solution is beneficial to the agents are questions
investigated in Sraieb (2015b).
68
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
2 W
HAT IS NEXT: A NEW INNOVATIVE APPROACH
TOWARD EFFECTIVENESS
At the heart of the analysis above, there is the time-inconsistency problem. This
leads to a commitment problem on the side of the donor: ex-post, the donor of
conditional aid is tempted to deliver aid regardless of the reform implementation
or the recipient involvement. Anticipation that this will happen, in turn, destroys
the recipient’s incentive to carry out costly reform policies ex-ante.
This would not be a problem had the recipients used aid efficiently.
Unfortunately, this is not always the case. The reason is that the donor faces
a disbursement pressure that acts toward an unconditional release of aid. The
literature has investigated the sources of such disbursement pressures.9 These
can be classified into three main categories: altruistic motivations of the donor
(warm glow), opportunistic behavior or pure bureaucratic inertia.
All these reasons point to the difficulty for donors to resist the pressure
toward aid release. As such, there is little hope that solutions to the time-incon-
sistency problem come from donor actions on aid volumes (i.e., decreasing or
withdrawing aid to recipient following a non-contracted use of aid).
Efficiently mitigating time-inconsistency may be implemented through an-
other instrument at the hands of the donor. This has to do with changing the
composition of aid. Indeed, aid can be given under lump sum transfers that enter
directly the general budget of the recipient country or as an investment via a
financing of a project therein. These forms of aid are called respectively budget
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
69
Mohamed Mounir Sraieb
should be allocated among the different agents in the recipient country (the poor
or the rich). Importantly, the donor has to rely on the local government for per-
forming this allocation. Because the donor has no control on the way the recipi-
ent will allocate monetary transfers between the rich and the poor, the donor may
favour targeted investments in projects like infrastructure building as opposed to
fungible non-targeted budget support.
The results in Legros and Sraieb (2015) suggest that it is often optimal for
the donor to offer both budget support and infrastructure assistance, in line with
practices of most donors.
Under perfect information about the ability of the recipient country to engage
in income redistribution toward the poor, the aid package involves a fixed project
and positive transfer to recipient countries that exhibit a low willingness or ability
to redistribute to the poor. For countries with a larger willingness to redistribute,
larger infrastructure projects are offered but those countries are required to co-
finance the project.
When the recipient country’s ability to distribute is private information, the
aid contract will involve both a positive monetary transfer to the recipient and
a project when the expected ability to engage in income redistribution is small.
If, however, this expected ability is high, the aid contract involves a positive co-
payment from the recipient country and therefore precludes some low-income
countries from receiving aid.
Our main finding is challenging in that it stipulates that while it is possible to
separate types via contracting, the donor prefers to use a pooling contract.
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
3 CONCLUDING REMARKS
Investing in reputation signaling and changing aid modalities are undoubtedly ef-
ficient strategies toward mitigating the time-inconsistency problem on the side of
the donor and therefore, strengthening aid effectiveness.
Yet, an objection against these strategies is that they assume a demand-
ing organizational design. Indeed, so far, the analysis considered the donor as a
black box, an institution whether bilateral (for instance a national aid agency) or
multilateral (ex. the International Development Association (IDA), the International
Bank for Reconstruction and Development (IBRD), etc.), where the decision on
whether to provide aid and enforcing the contract is centralized. In a number of
institutions, this may not be the case and the task of enforcing the contract is
delegated to departments.
Decentralization may harm reputation. On the one hand, existence of career
concerns for the staff (whether they are driven by intrinsic motivation for promo-
tion or in terms of existence of ex-ante monitoring and checks and balances on
70
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
the staff’s decisions) may lead to excessive toughness and too much reputation
signaling. On the other hand, these incentives may be also ineffective if it is the
staff that decides on success/failure of a recipient to comply with all (or part) of
the conditions linked to the contract. Actually, there is scope for distortions of dif-
ferent types that makes all the situations likely. Investigating these effects is out
of the scope of this paper. However, they are interesting lines for future research.
Another sizeable challenge to aid effectiveness is the failure of coordination
among donor countries. This is a constant theme that ranks high in the donors
community debates and declaration of intentions. The Paris Declaration (2005)
and the Accra Agenda for Action (2008) thus mentioned aid coordination as one
of the key mechanisms for enhancing aid effectiveness. The signatory donors
committed to coordinate, simplify procedures and share information to avoid
duplication of aid (the harmonization principle) and put effort in view to make their
activities as cost-effective as possible.
However, records on coordination among donors are quite disappointing.
Thiele et al. (2010) find persistence in aid duplication. Frot and Santiso (2011) find
evidence for herding among donors. Regression results of Nunnenkamp et al.
(2015) indicate that coordination among donors has even weakened since the
Paris Declaration (2005).
Considering coordination among donors enriches the analysis and would in-
form policies about its adverse impact on the strategies discussed above.
REFERENCES
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
71
Mohamed Mounir Sraieb
Qian, Nancy & Nunn, Nathan (2014). US Food Aid and Civil Conflict. American Econo-
mic Review, 104(6), 1630-1666.
Sraieb, Mohamed Mounir (2015a). An Empirical Model for U.S. Foreign Aid Allocation.
Ecares Working Paper 2015-48, Université libre de Bruxelles.
Sraieb, Mohamed Mounir (2015b). Development Aid Effectiveness: when is reputation
worth an investment? Mimeo, Université libre de Bruxelles.
Sraieb, Mohamed Mounir (2015c). Policies for Development Aid. Thesis in Economics,
Mimeo, Université libre de Bruxelles.
Thiele, Rainer, Nunnenkamp, Peter, & Inaki, Aldasoro (2010). Less Aid Proliferation and
More Donor Coordination? The wide gap between words and deeds. Journal of
International Development, 22(7), 920-940.
APPENDIX
Explanatory Variables and their Categories
In accordance with the literature, I relate aid flows to a set of explanatory
variables presented below:
–– Past Aid: this is aid lagged one year. The expected sign for the variable is
positive.
–– GDP per capita: measured in purchasing power parity in constant 2009.
–– Democracy: this is an indicator of good governance and the quality
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
72
What is Behind Foreign Aid Ineffectiveness? What is Behind Foreign Aid Ineffectiveness?
–– Military aid: this is the military assistance provided by the donor to partner
countries. This variable proxies the country’s strategic importance to the do-
nor’s security interests. Data are expressed in millions, constant 2009 U.S.
Dollars.
–– U.N. vote: this is a variable translating the strategic motivations of the donor
in terms of the voting allegiance of the recipient country in the U.N. General
Assembly (UNGA). For a given recipient, it represents the percentage of
votes in line with the donor in the UNGA. Abstentions and absenteeism are
discarded.
© De Boeck Supérieur | Téléchargé le 08/07/2022 sur www.cairn.info (IP: 212.147.59.126)
73