Memorandum submitted by Elisa Van Waeyenberge
SOAS (University of London)
1. This submission seeks to shed light on
two related aspects of conditionality as practiced both by the
World Bank (WB) and DFID:
— first, the principle of selectivity
to allocate aid flows that has become increasingly popular across
the donor community; and
— secondly, the way in which the selectivity
practice has been applied, with a particular focus on the assessment
tool at the heart of both the WB's and DFID's aid allocation mechanisms,
the Country Policy and Institutional Assessment (CPIA).
It asserts that both the practice of selectivity
and the criteria on the basis of which the latter proceeds are
ill-conceived. This has potentially severe consequences for the
prospects for development and poverty reduction, particularly
in those low-income countries (LICs) that remain dependent on
aid for their access to foreign resources.
2. Under a performance-based allocation
of aid or "selectivity", the conditionality accompanying
aid no longer reflects the flow of reforms, but the state of the
policy and institutional environment of a particular country.
When aid flows are allocated selectively, donors set conditions
that identify environments judged beneficial for development and
poverty reduction, and aid is allocated accordingly. Conditions
relate to past rather than future actions (policy-level versus
policy-change conditionality). Selectivity could be seen as some
form of "pre-emptive development", where funds are withheld
until demands made by the donors are met.
3. The selectivity discourse became increasingly
formalised in the late 1990s, abetted by the appearance of a set
of analytical and empirical arguments. These came to constitute
a new aid paradigm, which was heavily promoted by the WB and through
which the WB sought to encourage other agencies to emphasise prior
actions rather than future policy promises when allocating aid
flows. A paper by Burnside and Dollar (2000) was central in providing
analytical foundations to the new paradigm, with the Bank policy
report Assessing Aid (WB 1998) built around its core premise
of conditional aid effectiveness. In a nutshell: aid only affects
the growth rate positively if a certain set of policies/institutions
are characteristic of a country; aid does not affect the policy
environment; and, hence, aid should be (re-)allocated towards
those countries characterized by a "good" policy/institutional
environment. Such "good" policy/institutional environments
broadly reflect the Country Policy and Institutional Assessment
(CPIA) ratings (see below), or more narrowly focus on the "core"
macroeconomic policy stances of budget surplus, low inflation
and trade openness. Collier and Dollar (2002) extended the Burnside-Dollar
results that aid has no impact on growth except in a "good"
policy environment and that aid does not affect policy reform,
into a prescriptive model of what are "poverty-efficient"
inter-recipient aid allocations. It was argued that, traditionally,
aid has been used to induce policy reform and, as a result, has
been targeted on "weak" policy and institutional environments.
Increasing poverty reduction efforts then does not necessarily
require an increase in aid, but, more importantly, a change in
the existing allocation of aid towards those countries that are
characterized by "good" policy.
4. Without wanting to elaborate on the substantial
academic debate that was engendered by the Burnside-Dollar-Collier
proposition, it needs to be observed that the debate has settled
on the acknowledgment that there are no analytical or empirical
grounds upon which the policy stance of selectivity as advocated
by Burnside-Dollar-Collier can be defended. It has been systematically
illustrated that the productivity of aid does not depend on the
sets of policies focused upon in the Burnside-Collier-Dollar proposition.
The latter represents a biased research effort characterised by
poor theoretical and empirical practice and is not representative
of the broader aid impact literature.[162]
Yet, notwithstanding the serious criticism that exists regarding
the reliability and relevance of the proposition, the selectivity
practice it advocates has been keenly adopted in the donor community
with now an apparently "scientific" rationale underpinning
donors' attempts to increase their leverage in poor countries
to make these comply with a set of what are perceived to be "common
values".
5. The core of the WB's performance-based
allocation system, which applies to the distribution of its concessional
resources through the International Development Association (IDA),
is the Country Policy and Institutional Assessment (CPIA). The
latter measures a country's performance on a set of macroeconomic,
structural, social and governance criteria and then feeds into
an allocation formula for IDA's resources that is sixteen times
more sensitive to changes in policy/institutional variables than
to changes in income per capita (as a proxy for poverty).
6. The CPIA-steered selectivity framework
has been married to the recognition of the importance of ownership
for the development effectiveness of aid programmes through the
"negotiation" framework introduced in 1999 by the WB
and the International Monetary Fund (IMF), the Poverty Reduction
Strategy Paper (PRSP). The CPIA conditions the scope for alternative
development and poverty reduction strategies articulated in the
PRSP as it is expected that the implementation of PRSP policies
reflects in a country's CPIA ratings. This reveals an implicit
assumption that the former are necessarily in line with the imperatives
embedded in the latter and WB documents indicate how the CPIA
effectively serves as a filter between a country's PRSP and the
operational realities of WB concessionary assistance.
7. Since 2000, the WB has placed the CPIA
in the public domain. As such, it has sought to promote the CPIA,
which is rapidly becoming a standard in the broader donor community.
Both the African Development Fund and the Asian Development Fund
use a very similar, but independently estimated CPIA. In addition,
the Debt Sustainability Framework (DSF), the framework newly formulated
by the International Financial Institutions (IFIs) and which will
also be used by other creditors and fora such as the Paris Club,
has the CPIA at its core in determining debt distress thresholds.
8. DFID formally draws on the CPIA scores
to allocate its aid resources. Following DFID's allocation formula,
the starting point for determining country allocations is determined
as a result of: the country's population multiplied by its (WB-provided)
CPIA score (with a weighting of 2 attached to the CPIA score)
and divided by its Purchasing Power Parity (PPP) income per capita.
This feeds into a decision tree that takes additional relevant
elements into account, which can include: MDG indicators; social
inequality/income distribution; conflict/post-conflict status;
political concerns; private flows; DFID's comparative advantage;
etc. Note that DFID's allocations for India, China, South Africa
and Indonesia are fixed outside the model.
9. With selectivity, the donor increases
its emphasis on the quality of the policies and institutions of
a country, as measured against a uniform ideal model captured
in the CPIA, relative to need-based criteria when allocating its
aid resources. This is allegedly meant both to direct resources
to environments where they are expected to be relatively more
effective, as well as encourage LIC governments to "improve"
their policy performance (through some form of "demonstration
effect").
10. The allocation mechanism, however, is
riddled with serious problems with pernicious implications for
development and poverty reduction prospects in LICs. First, the
operational reality of selectivity threatens attempts in poor
countries to raise investment rates and/or to protect pro-poor
expenditures. Aid constitutes a crucial resource for least developed
countries dominating their investment and budgetary processes
(see Attachment 1). As a result, a decrease in aid allocations
to a country may have pernicious implications for the nature of
the adjustment process a country will have to engage in. Allocating
aid resources to those that have already done a minimum of stabilization
or structural adjustment (as captured in the CPIA) risks jeopardizing
attempts in poor countries to raise their investment rates. Further,
it is, on average, easier to implement pro-poor expenditures within
the constraints of a particular policy environment than it is
to implement an economic reform program that would include pro-poor
policies (Morrissey 2001).
11. Second, the deployment of the CPIA is
built on the presumption of government control over policy outcomes.
The WB (2001, p 3) explicitly emphasizes that: "The CPIA
intentionally measures policies and institutional arrangements
rather than actual outcomes (growth, poverty reduction)—in
other words, the key elements within a country's control
that determine growth and poverty reduction". Such an approach
blatantly fails to take into account the various structural parameters,
both domestic and international affecting domestic policies and
institutions. These typically include the state of a country's
productive capacity, the skill base of its economy, its debt,
its trade relations, etc, and tend to be worse the poorer the
country. A simple calculation reveals that the average GNI per
capita for the countries in the top quintile of the CPIA ranking
is at least three times (and at times even four times) the size
of the average GNI per capita for countries in the bottom quintile,
illustrating the tendency of the CPIA to be biased in favor of
better-off poor countries.
12. Third, the nature of the aid delivery
system strongly affects the macroeconomic stability of countries
characterized by large aid ratios. The selectivity proposition
on the basis of the CPIA may, however, be sensitive to small changes
in scores. This raises the specter of uncertainty and volatility
of aid flows with negative effects on debtor countries, for instance
through negative effects on investment levels and efficiency (Lensink
and Morrissey 2000).
13. Fourth, there is of course the primary
question of the particular policy/institutional agenda the CPIA
seeks to promote and what is its relationship to economic development
and poverty reduction, an issue with which the rest of this submission
will be concerned.
14. Currently, the CPIA encompasses sixteen
criteria grouped in four different clusters. Under "economic
management" are: macroeconomic management, fiscal and debt
policy. Under "structural policies": trade, financial
sector, and business regulatory environment. Under "policies
for social inclusion/equity" we find: gender equality; equity
of public resource use; building human resources; social protection
and labour; and, policies and institutions for environmental sustainability.
And, finally, the categories constituting "public sector
management and institutions" are: property rights and rule-based
governance; quality of budgetary and financial management; efficiency
of revenue mobilisation; quality of public administration; and
transparency, accountability and corruption in the public sector
(WB 2006). WB staff provide scores (between 1 and 6) for each
of these criteria; these scores are averaged per cluster; and
the CPIA score is obtained as the average of the scores of the
four respective clusters.
15. Staff on WB country desks are provided
with a CPIA Questionnaire which explicitly states which policy/institutional
environment merits a particular score ("narrative guidelines").
These narrative guidelines for each criterion are supplemented
with guideposts consisting of both economic indicators and diagnostic
reports which seek to assist staff further in assigning specific
scores to the various elements of a country' s policy and institutional
environment.
16. When the distinct guidelines regarding
how to rate a particular country on a specific criterion are more
closely scrutinised, the following transpires. The economic core
of the CPIA is built around a set of particular precepts including:
low inflation; an implicit preference for a surplus budgetary
position; minimal restrictions on trade and capital flows; "flexible"
goods, labour and land markets; market-determined interest rates;
prohibition of directed credit; competition policies guaranteeing
equal treatment of foreign and domestic investors ("national
treatment"); "virtually" complete capital account
convertibility; protection of shareholder rights ("good corporate
governance"); and no restrictions on public sector procurement.
17. The development record of these various
policy prescriptions (including trade liberalisation; capital
account liberalisation; "national treatment" of foreign
investment) has been particularly poor (and often negative), and
these policies fail to capture the actual historical experience
of development. In this regard, the work of Ha-Joon Chang of Cambridge
University plays an important role, as it draws on extensive empirical
material to illustrate how the actual experience of development
has been widely divergent from the currently prevailing policy
prescriptions—as also embodied in the CPIA.[163]
The economic policies promoted through the CPIA, in effect, eliminate
the possibilities for strategic interventions along which specific
sectors of an economy can be promoted and the importance of which
to the economic success of the now-developed countries and the
East Asian "miracle" economies has been repeatedly pointed
out (Amsden 1989; Wade 1990). Selective allocation of aid flows
to LICs on the basis of the CPIA hence risks locking in an extensive
economic policy agenda with ambiguous, if not outright negative,
repercussions for growth.
18. Furthermore, there are contradictions
and inconsistencies between the imperatives defended in the economic
core of the CPIA and those that constitute the social cluster
of the CPIA. Apart from considering social issues as an "add-on"
to economic issues rather than acknowledging that the former are
intricately bound up with the latter, the trade-offs between the
prescriptions entailed in the economic clusters and those put
forward in the social cluster are ill-appreciated. For instance,
the various specifications of good policy in such areas as building
human resources or social protection sit awkwardly with the stringent
fiscal and monetary order embodied in the economic management
cluster.
19. Also, the relationship to development
or growth (and thus aid productivity) of the governance issues
incorporated in the CPIA remains dramatically ill-understood.
The imperatives embedded in the Public Sector Management-cluster
stem from a particular (and inadequate) prescriptive approach
regarding the role of the private sector and the state, and are
informed by a preoccupation with corruption as a source of (static)
welfare loss rather than that corruption be assessed in the historical
and dynamic context of development—with the latter entailing
complex and shifting underlying political-economic processes.
The processes that drive development can not be understood as
an unfortunate deviation from a particular norm of liberal governance,
but emerge as strategies of adaptation and survival in contested
settings. The implications of specific governance arrangements
for growth and development crucially depend on: the particular
constellation of the political-economic forces both within the
state and society (and the nature of the relationships between
these); the state of development; the nature of the international
relations of the country; etc. The policy/institutional imperatives
embedded in the CPIA matrix, whether touching upon the property
rights regime, corruption, budgetary and financial management
processes, tax regimes, quality of public administration or transparency
in the public sector, further, at most describe what certain advanced
economies could look like. Yet, there is significant institutional
diversity even among industrial countries and imposing a common
standard on all countries, with widely varying conditions is likely
to counterproductive.
20. Rather than imposing a fixed set of
policies/institutions on a widely diverging set of countries through
such a mechanism as the selective allocation of aid flows on the
basis of the CPIA, the imperative transpires for the "policy
space" of a developing country to be left wide open so that
sufficient scope for discretion exists for a government that would
seek to move a country's economy away from low productivity activities.
A recent UNCTAD Report (2006, p. 288) restates the need to put
the development of productive capacities at the heart of national
and international policies to promote economic growth and poverty
reduction in LICs. For such a purpose, policies would need to
focus on promoting capital accumulation, technological progress
and structural change. Such imperatives require a host of interventions
such as trade tariffs, import substitution, export promotion,
the extensive use of performance requirements on both domestic
and foreign investment, the selective promotion of industries,
massive investment in skill creation, infrastructure and support
institutions. Most of these measures, however, incur a penalty
under the current CPIA scoring exercise and hence are actively
discouraged by the donor community.
21. In addition, the defining features of
the least developed countries, including per capita income and
the extent of economic vulnerability, provide sufficiently satisfactory
criteria to steer aid allocations on both efficiency and
equity grounds. Aid needs scaling-up rather than that it be preoccupied
with policies that project to increase its effectiveness—with
the latter often hampered by an inadequate understanding of the
dynamics (domestic and international) of aid and its conditionality.
At its core, donor countries' parliaments should critically examine
their government's real (rather than rhetorical) commitment to
financing poverty reduction efforts around the world, with
Official Development Assistance (ODA) net of what are called "special
purpose grants" (including debt relief, technical assistance
and emergency relief) currently at a low of 0.15 percent (in 2005)
relative to (OECD) donors' national income, well below the 0.25
percent attained in the early 1990s.
22. On the basis of these observations a
set of questions arise including the following:
— On the basis of what evidence did
DFID decide that a performance-based aid allocation process will
deliver better results in terms of development and poverty reduction?
— Has there been any assessment within
DFID regarding the relationship between the various constituent
criteria of the CPIA and development or poverty reduction? If
not, why are these policies/institutions considered important
for aid effectiveness, development or poverty reduction?
— To what extent does the historical
experience of development inform the discussions around "necessary"
prerequisites for development and poverty reduction?
— How does DFID seek to address the
possible contradictions between the social and economic criteria
embodied in the CPIA?
12/10/2007
REFERENCES
Amsden, A. (1989), Asia's Next Giant: South Korea
and Late Industrialisation, New York: Oxford University Press.
Burnside, C. and D. Dollar (2000), "Aid, policies
and growth", American Economic Review 90(4): 847-68.
Collier P. and D. Dollar (2002), "Aid allocation
and poverty reduction", European Economic Review 46(8):
1475-1400.
Lensink, R. and O. Morrissey (2000), "Aid instability
as a measure of uncertainty and the positive impact of aid on
growth", Journal of Development Studies 36(3): 31-49.
Morrissey, O. (2001), "Pro-poor conditionality
for aid and debt relief in East Africa", paper prepared for
the WIDER Development Conference on Debt Relief, Helsinki 17-18
August.
Moss, T. and A. Subramaniam (2005), "After the
big push? Fiscal and institutional implications of large aid increases",
Center for Global Development Working Paper 71, available at:
http://www.cgdev.org/content/publications/detail/4436.
UNCTAD (2006), The Least Developed Countries Report
2006. Developing Productive Capacities, Geneva: UNCTAD.
Wade, R. (1990), Governing the Market, Princeton:
Princeton University Press
World Bank (1998), Assessing Aid. What Works,
What Doesn't and Why, New York: Oxford University Press.
World Bank (2001), Review of the Performance-Based
Allocation System, IDA10-12, Operations Evaluation Department,
Washington, DC.
World Bank (2006), "Country Policy and Institutional
Assessment 2006. Assessment Questionnaire", Operations Policy
and Country Services, September, Washington, DC.
Attachment 1:
AID INTENSITY INDICATORS FOR LEAST DEVELOPED
COUNTRIES, 2004, PERCENTAGES IN BRACKETS
| Country |
| ODA as share of GNI |
| > 50% | Burundi (54,3), Liberia (52,8), Sao Tome and Principe (61,9), Solomon Islands (47,1)
|
| 20% < 50% | Afghanistan (36,7), Democratic Rep. Congo (29,1), Eritrea (28,4), Malawi (25,6), Madagascar (26,8), Guinea-Bissau (29,4), Mozambique (22,0), Rwanda (26,0), Sierra Leone (34,6), Timor Leste (31,7), Zambia (21,6)
|
| 10% < 20% | Bhutan (10,9), Burkina Faso (12,7), Cambodia (10,3), Cape Verde (14,7), Ethiopia (18,9), Mauritania (11,1), Mali (12,1), Lao PDR (11,3), Kiribati (12,7), Gambia (16,5), Niger (17,6), Senegal (14,1), Tanzania (15,5), Uganda (17,3), Vanuatu (12,4)
|
| 5% <10% | Angola (6,6), Benin (9,4), CAR (8,0), Chad (8,8), Comoros (6,7), Haiti (6,3), Guinea (7,5), Nepal (6,4), Samoa (8,6)
|
| < 5% | Maldives (3,7), Sudan (4,4), Togo (3,0), Yemen (2,1)
|
| ODA as share of GCF |
| > 100% | Burundi (386,6), Comoros (217,8), Eritrea (122,7), Malawi (163,9), Madagascar (116,9), Liberia (346,8), Guinea Bissau (213,9), Mozambique (100,6), Niger (111,16), Rwanda (124,5), Sao Tome and Principe (169,5), Sierra Leone (318,0), Solomon Islands (134,1), Timor Leste (158,6)
|
| 50% < 100% | Angola (63,5), Benin (51,2), Burkina Faso (66,2), Cape Verde (72,3), Ethiopia (88,2), Mauritania (54,5), Mali (61,1), Lao PDR (62,3), Guinea (68,6), Gambia (55,8), Senegal (59,0), Tanzania (83,9), Uganda (75,5), Zambia (76,8)
|
| 10% < 50% | Bhutan (17,5), Cambodia (37,9), Chad (29,7), CAR (45,6), Maldives (10,3), Nepal (24,0), Sudan (18,2), Togo (16,6), Yemen (11,5)
|
ODA as a share of GXP
|
| > 100% | Democratic Rep. Congo (592), Guinea Bissau (170), Sierra Leone (128)
|
| 50% < 100% | Burundi (88), Cambodia (67), Chad (64), Eritrea (53), Ethiopia (79), Gambia (54), Haiti (56), Lao PDR (85), Malawi (71), Mozambique (88), Niger (91), Rwanda (78), Sao Tome and Principe (74), Solomon Islands (61), Tanzania (77), Uganda (64)
|
| < 50% | Angola (10), Benin (40), Bhutan (29), Burkina Faso (49), CAR (34), Comoros (39), Guinea (44), Madagascar (46), Nepal (36), Togo (16), Yemen (6), Zambia (48)
|
Source: World Development Indicators on-line (2006 edition) for
ODA as share of GNI and GCF; Moss and Subramaniam (2005) for ODA
as share of GXP. GNI: Gross National Income; GCF: Gross Capital
Formation; GXP: total Government Expenditure.
162
For a good account, see Beynon, J. (2001), "Policy implications
for aid allocations of recent research on aid effectiveness and
selectivity", paper presented at the Joint Development Centre/DAC
Experts Seminar on Aid Effectiveness, Selectivity and Poor Performers,
OECD, available at: www.oecd.org/dataoecd/15/62/2664833.pdf. Back
163
See most recently: Chang, H-J. (2007), Bad Samaritans: Rich
Nations, Poor Policies and the Threat to the Developing World,
Random House. Back
|
