Application of the Performance Based Allocation System (PBA) to Fragile States Edward Anderson, Karin Christiansen and Rowan Putnam Poverty and Public.

Application of the

Performance Based Allocation System (PBA)

to Fragile States

Edward Anderson, Karin Christiansen and Rowan PutnamPoverty and Public Policy Group, Overseas Development Institute

Meeting of EU+ IDA Deputies, June 19th 2007, Berlin

Summary of Presentation

• Background• Definitions and categories of fragile states• Overview of IDA aid to fragile states, 2003-2006• Does the PBA allocate enough IDA resources

to fragile states? • What would be the implications of adjusting

or ‘fine-tuning’ the PBA? • Conclusions and recommendations

Background

• The PBA is the mechanism used to allocate IDA resources across eligible countries;

• Under 14th IDA replenishment (FY2006-FY2009):

– US$31.5bn (SDR 21 billion) to be allocated

– across 81 low and lower-middle income countries

• As with systems used by other donors, PBA allocates IDA resources on the basis of ‘performance’ and ‘need’.

Background (cont.)

• Key features of the PBA:

– use of Country Policy and Institutional Assessment (CPIA) scores to measure performance;

– large weight attached to performance relative to need;

– additional emphasis given to governance component of CPIA in measuring performance;

– additional assistance given to post-conflict countries, for limited period, over and above ‘normal’ PBA allocation.

Definitions of fragile states

• Fragile states:

– poor countries + weak governance, conflict and/or insecurity

• More specific ‘LICUS’ definition (World Bank):

– low-income countries + low Country Policy and Institutional Assessment (CPIA) scores*:

• CPIA less than 2.5 = severe LICUS/fragile state• CPIA 2.5 to 3.0 = core LICUS/fragile state• CPIA 3.0 to 3.2 = marginal LICUS/fragile state

Relevant categoriesof fragile states

• Post-conflict: “countries emerging from violent and protracted conflict” (IDA definition);

• Conflict-affected: “recently experienced, are experiencing, or are widely regarded as at risk of experiencing, violent conflict” (World Bank definition);

• Not conflict-affected: all other fragile states;

• Non-accrual: repayments to IDA are overdue by more than six months;

• Other distinctions, e.g. on the basis of development performance, natural resources, prospects for reform.

Categories of fragile states,FY2003-2005

Post-conflict Conflict-affected Not conflict-affected

Afghanistan Cambodia Nigeria Cameroon Pap. N. Guinea

Angola C. African Rep. Solomon Is. Chad S. Tome e Pr.

Burundi Comoros Myanmar Djibouti Togo

Congo (DRC) Cote d’Ivoire Somalia Gambia Tonga

Cong (Rep.) Guinea-Bissau Sudan Guinea Uzbekistan

Eritrea Haiti Tajikistan Lao PDR Vanuatu

Sierra Leone Liberia (13) Mauritania Yemen

Timor-Leste Niger Zimbabwe

(8) (16)

Pop = 126m Pop = 282m Pop = 135m

Notes: Under-lined countries were in non-accrual status for at least part of period. Population data refer to 2005.

IDA net disbursementsto fragile states, 2003-2005

(1)

US$ million*% of total

IDA netdisbursements

US$* per capitaper year

All fragile states 3,888 19 2.4

- post-conflict 1,844 9 5.0

- conflict-affected 748 4 0.9

- not conflict-affected

1,296 6 3.3

All IDA countries 20,256 100 2.6

Notes: *Constant 2004 prices. Source: OECD-DAC (2007)

Net disbursementsto fragile states, 2003-2005

(2)

% of total multilatera

l

% of total multilatera

land

bilateral

IDA as % of total

multilateral

IDA as % of total

bilateral and multilateral

All fragile states 20 20 31 8

- post-conflict 9 9 34 8

- conflict-affected 6 7 23 4

- not conflict-affected

6 4 36 12

Source: OECD-DAC (2007)

Disbursements from World Bank Trust Funds to fragile

states

US$ million,

2003-2006

% of totalIDA to fragile states

Countriesbenefiting

Post-Conflict Trust Fund

24.2 0.6

Burundi, Comoros, Congo (DRC), Congo

(Rep.), Cote d’Ivoire, Haiti,

Nigeria, Sierra Leone, Somalia, Sudan,

Tajikistan, Timor-Leste (12)

LICUS Trust Fund 37.7 0.8

Central African Rep., Comoros, Cote d’Ivoire, Haiti, Liberia, Somalia,

Sudan, Togo, Zimbabwe (9)

Source: World Bank staff

Estimated IDA allocation ‘norms’ under IDA14

Notes: *India, Indonesia and Pakistan; **includes regional projects, arrears, tsunami and Iraq. Source: Authors’ simulations.

 % of all IDA

resources

% of IDA resources, not including capped-blend or

special purposes

All fragile states 17 24

- post-conflict 10 14

- conflict-affected 6 8

- not conflict-affected 2 3

Capped-blend* 20 -

All other countries 55 76

Special purposes** 8 -

Total 100 100

The PBA and fragile states:key questions addressed

1) Does the PBA allocate enough of total IDA resources to fragile states, given:

– IDA’s objectives, i.e. “to raise standards of living in the less-developed areas of the world” (IDA Article 1)

– evidence of the effectiveness of aid to fragile states in achieving those objectives.

2) If not, how the PBA could be adjusted or ‘fine-tuned’ to bring about a more appropriate allocation?

Insights from poverty-efficiency approach

• Approach proposed by Collier and Dollar (2002): – allocate aid so as to achieve largest possible reduction in

poverty;

• Assumes that:– aid affects poverty through its impact on economic growth;– effect of aid on economic growth higher in countries with

higher CPIA scores and in post-conflict countries;

• Approach justifies higher allocations to countries with higher CPIA scores, and to post-conflict countries – as in the PBA;

• But PBA still under-allocates to fragile states in comparison with a Collier-Dollar ‘poverty-efficient’ allocation.

Further considerations

Issue Key question Assessment

Incentives Allocate on the basis of performance to increase incentives for improving performance?

Limited evidence and ‘one size fits all critique’

Take-up Would higher allocations to fragile states lead to higher actual disbursements?

No strong evidence to suggest not, but needs research

Absorptive capacity

Are fragile states able to absorb additional disbursements?

No strong evidence to suggest not, but again needs research

Other donors

Should IDA allocations be adjusted to take account of other donors?

Other donors allocate similar shares of aid to fragile states – suggests no adjustment

Sources of evidence

Are econometric estimates of aid effectiveness reliable?

Not the only source of evidence, but not obvious that they over-estimate

Further considerations(cont.)

Issue Key question Assessment

Aid-growth relationship

Are the Collier-Dollar findings reliable? Do factors other than CPIA scores matter?

Not unreasonable to use Collier-Dollar findings, although more research required

Direct aid-poverty links

Can aid reduce poverty directly, as well as through growth?

Probably, but evidence remains limited

Windows of opportunity

Can aid be used to promote policy turnarounds?

Possibly, but evidence preliminary and main effects are from TA

Country disadvantages

Should IDA give additional resources to countries facing inherent disadvantages?

Perhaps, but would have mixed implications for fragile states

Allocations to post-conflict countries

Issue Key question Assessment

Eligibility Which countries should qualify for exceptional post-conflict allocations?

Need for open discussion about which indicators and thresholds to use

Allocation How should aid be allocated among post-conflict countries?

Use of different performance indicators raises problems; alternative is to use a post-conflict enhancement factor

Duration For how long should post-conflict countries receive exceptional allocations?

Evidence that current period of assistance (4+3 years) is too short

Implications of adjusting or ‘fine-tuning’ the PBA

• Advantages of the PBA: simple formula, transparent, and linked to empirical evidence

• But still possible to adjust or ‘fine-tune’ the PBA to achieve a more desirable allocation

• Three main sorts of adjustments considered here:

– changing the weights attached to existing indicators in the PBA formula;

– adding new indicators and/or enhancement factors to the PBA formula;

– altering the way post-conflict countries are treated.

1. Changing the weights attached to existing indicators

Share of IDA resources to fragile states* (%)

AExisting PBA formula (baseline)

24

BSmaller governance factor (exponent reduced from 1.5 to 1)

26

CNo governance factor (exponent reduced from 1.5 to 0)

29

DSmaller weight to country performance (exponent reduced from 2 to 1)

29

C+D

No governance factor and smaller weight to country performance

32

*Not including resources allocated to capped-blend countries or special purposes

2. Adding new indicators and/or enhancement factors

Share of IDA resources to fragile states* (%)

C+D

No governance factor and smaller weight to country performance (baseline)

32

EAdd infant mortality as indicator of need(exponent set at 1)

38

FMeasure GNI per capita in US$ PPP(exponent set at minus 1)

35

GEnhancement for small-island states (enhancement factor set at 1.5)

32

HEnhancement for land-locked countries (enhancement factor set at 1.5)

31

*Not including resources allocated to capped-blend countries or special purposes

3. Changing the way post-conflict countries are treated

Share of IDA resources to post-conflict countries* (%)

I Existing PBA formula (baseline) 14

JAdd a post-conflict enhancement factor(factor set at 1.25)

7

KAdd a post-conflict enhancement factor(factor set at 2)

17

*Not including resources allocated to capped-blend countries or special purposes.

Summary and conclusions

• ‘Fragile states’ currently receive around 20% of total IDA resources, while post-conflict countries receive around 10%;

• There is evidence to suggest that these shares are too low;

• Potential ways to raise the share of IDA resources allocated to fragile states include:

– reducing the size of the ‘governance factor’ in the PBA formula;– reducing the weight attached to performance relative to need;– adding additional indicators of need.

Summary and conclusions

(cont.)

• There is a case for exceptional allocations to post-conflict countries:– a post-conflict enhancement factor (PCEF) provides a simple and

transparent method to achieve this, although care needed when setting the level of a PCEF;

– evidence that existing duration of exceptional allocations to post-conflict countries appears too short.

• Areas where further research is required:– research on aid effectiveness and absorptive capacity in post-

conflict countries, and in countries with high economic vulnerability;

– research on links between IDA aid allocations and actual net disbursements.

Thank you

for listening

Edward AndersonPoverty and Public Policy Group,

[email protected]

June 19th 2007