Attridge, Samantha; Engen, Lars

Research Report
Blended finance in the poorest countries: The need
for a better approach

ODI Report

Provided in Cooperation with:

Overseas Development Institute (ODI), London

Suggested Citation: Attridge, Samantha; Engen, Lars (2019) : Blended finance in the poorest
countries: The need for a better approach, ODI Report, Overseas Development Institute (ODI),
London

This Version is available at:

http://hdl.handle.net/10419/206745

Standard-Nutzungsbedingungen: Terms of use:

Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Documents in EconStor may be saved and copied for your
Zwecken und zum Privatgebrauch gespeichert und kopiert werden. personal and scholarly purposes.

Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle You are not to copy documents for public or commercial
Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich purposes, to exhibit the documents publicly, to make them
machen, vertreiben oder anderweitig nutzen. publicly available on the internet, or to distribute or otherwise
use the documents in public.
Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen
(insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, If the documents have been made available under an Open
gelten abweichend von diesen Nutzungsbedingungen die in der dort Content Licence (especially Creative Commons Licences), you
genannten Lizenz gewährten Nutzungsrechte. may exercise further usage rights as specified in the indicated
licence.

https://creativecommons.org/licenses/by-nc/4.0/
 Report

Blended finance
in the poorest
countries
The need for a better approach
Samantha Attridge and Lars Engen
April 2019
Readers are encouraged to reproduce material for their own publications, as long as they are not being sold commercially. ODI requests due
acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views
presented in this paper are those of the author(s) and do not necessarily represent the views of ODI or our partners.

This work is licensed under CC BY-NC-ND 4.0.

Cover photo: Indian truckes queue up to leave from the Trade Gate on Pakistani side. Photo credit: Asian Development Bank CC BY-NC-ND 2.0
Acknowledgements

The authors would like to thank the peer reviewers who kindly gave their time to provide very helpful
critique and comments: Paddy Carter (CDC Group), Charles Kenny (Centre for Global Development),
Paul Horrocks (Organisation for Economic Co-operation and Development, OECD), Cécile Sangare
(OECD), Tomas Hos (OECD), Soren Andreasen (Association of European Development Finance
Institutions, EDFI), Morten Lykke Lauridsen (International Finance Corporation, IFC), Artur Karlin
(International Finance Corporation, IFC), Neil Gregory (IFC), Joan Larrea (Convergence), Marcus
Manuel (Overseas Development Institute, ODI), Jesse Griffiths (ODI), David Watson (ODI) and
Hannah Caddick (ODI).
The authors would also like to thank colleagues from the following institutions for engaging in
the research, providing comment and information: the Asian Development Bank (ADB), Agence
Française de Développement (AFD), African Development Bank (AfDB), CDC Group, the United
Kingdom Department for International Development (DFID), European Investment Bank (EIB),
EDFI, International Development Association (IDA), IFC, Multilateral Investment Guarantee Agency
(MIGA), OECD, Overseas Private Investment Corporation (OPIC), Proparco and Norfund.
The authors gratefully acknowledge the generous financial support from the Bill & Melinda
Gates Foundation, AFD and Australia’s Department of Foreign Affairs and Trade, which made this
research possible.
All views expressed are those of the authors’ alone and do not reflect those of the funder, ODI or
the institutions reviewed in this report.

3
Contents

Acknowledgements3

List of tables and figures 6

Acronyms and abbreviations 8

Executive summary 10

1 Introduction and overview 14

1.1 Introduction 14
1.2 Overview 16

2 Blended finance: no common conceptual framework 17

2.1 Numerous official definitions 17
2.2 Understanding the ‘output’: mobilising private finance 19
2.3 Understanding the ‘input’: ODF investment  21
2.4 The urgent need for better data and transparency  22

3 ODI approach: scope and methodology 24

3.1 Definition and scope 24
3.2 Data  25

4 The big blended-finance push  26

4.1 The argument for blended finance  26
4.2 Increased ODA investment in blended finance due to policy changes 27
4.3 Risk of leaving the poorest countries behind 29

5 The potential of blended finance: time for a reality check 31

5.1 How much private finance is being mobilised? 32
5.2 Low leverage ratios and the need to temper expectations  34

6 The current state of blended finance in LICs 40

6.1 How much private finance is being mobilised in LICs? 41
6.2 Who is mobilising private finance in LICs? 42
6.3 In which LICs does blended finance mobilise private finance?  45

4
 6.4 In which LIC sectors is blended finance mobilising private investment? 46
6.5 Which instruments are used to mobilise private finance in LICs?  47
6.6 Instrument use by sector in LICs  50
6.7 Blended-finance project size in LICs 51

7 The need for a more tailored approach for LICs 52

7.1 A limited use of concessional finance to blend in LICs 52
7.2 Poor investment climate and a lack of investable opportunities in LICs 53
7.3 Limited tailoring of the blended-finance toolkit in LICs 55
7.4 Limited blended private investment in infrastructure in LICs 55
7.5 The limited headroom and risk appetites of MDBs, DFIs and donors 56

8 Conclusions and recommendations 59

References61

Annex 1 Institutional selection process  65

Annex 2 Data-collection methodology  67

Annex 3 Scatter plots 75

5
List of tables and figures
Tables

Table 1 Blended finance – definitional differences 19

Table 2 Summary of mobilisation data sources and implied leverage ratio 36

Table 3 Largest blended-finance recipient countries by MDB and DFI 46

Table A1 Largest blended-finance actors 65

Table A2 Largest blended-finance actors in LICs 66

Table A1 Project-level databases, by institution 68

Table A2 Exchange rates 68

Table A5 Consolidated sector classifications 70

Table A6 Examples from the blended-finance assessment process 73

Figures

Figure 1 Comparison of ODA and blended finance (per capita and per person living in poverty) 30

Figure 2 Private finance mobilised by MDBs 33

Figure 3 Leverage ratios by category and sector, nine selected MDBs and DFIs, 2013–2015 37

Figure 4 Sector leverage ratios by country income group, nine selected MDBs and DFIs, 2013–2015 38

Figure 5 External financial flows: income-group comparison 41

Figure 6 Private finance mobilised through blended finance by institution, 2012–2015 42

Figure 7 Share of total private finance mobilised, overall and in LICs, 2012–2015 43

Figure 8 Private finance mobilised in LICs, 2012–2015 43

Figure 9 Average annual MDB and DFI commitments to mobilise private finance in LICs, 2013–2016 44

Figure 10 MDB and DFI commitments to mobilise private investment in LICs by destination, 2013–2017 45

6
Figure 11 MDB and DFI commitments to mobilising private finance in LICs (selected MDBs and DFIs),
2013–2016 46

Figure 12 Commitments to mobilise private finance by sector (selected MDBs and DFIs), 2013–2017 47

Figure 13 Blended-finance commitments by sector and MDB and DFI, 2013–2017 48

Figure 14 Commitments to mobilise private finance by instrument (selected MDBs and DFIs), 2013–2017 48

Figure 15 Limited variation in instruments use, by income classification (MDB and DFI group A) 49

Figure 16 Variation in instrument use by income classification (MDB and DFI group B) 50

Figure 17 Instruments used for private-sector mobilisation in LICs, by sector 50

Figure 18 Average project size by sector in LICs 51

Figure 19 Mobilised private finance flows to countries with a credit rating 54

Figure 20 MDB and RDB risk-adjusted capital-adequacy ratios, 2014–2016 58

Figure A1 Comparison of total annual commitments, project-level datasets versus annual reports 68

Figure A2 Breakdown of projects by mobilising status 74

Figure A3 Relationship between private-finance mobilisation and financial depth 75

Figure A4 Relationship between private-finance mobilisation and the Human Assets Index 75

7
Acronyms and abbreviations

ADB Asian Development Bank

AFD Agence Française de Développement (French Development Agency)
AfDB African Development Bank
AIIB Asian Infrastructure Investment Bank
CAF Development Bank of Latin America
CGIF Credit Guarantee and Investment Facility
CIF Climate Investment Funds
CIV collective investment vehicle
CPA country programmable aid
CRS creditor reporting system
DAC Development Assistance Committee
DEG Deutsche Investitions - und Entwicklungsgesellschaft mbH
DFI development finance institution
DFID Department for International Development (UK)
EBRD European Bank for Reconstruction and Development
EDFI Association of European Development Finance Institutions
EFSD European Fund for Sustainable Development
EIB European Investment Bank
EU European Union
FY financial year
GDP gross domestic product
GEEREF Global Energy Efficiency and Renewable Energy Fund
GNI gross national income
HIC high-income country
IADB Inter-American Development Bank
ICD Islamic Corporation for the Development of the Private Sector
ICT information communications technology
IDA International Development Association
IDFC International Development Finance Club
IFC International Finance Corporation
IsDB Islamic Development Bank
LIC low-income country
LMIC lower-middle-income country
MDB multilateral development bank

8
MIC middle-income country
MIGA Multilateral Investment Guarantee Agency
NDB national development bank
ODA official development assistance
ODF official development finance
ODI Overseas Development Institute
OECD Organisation for Economic Co-operation and Development
OOF other official flow
OPIC Overseas Private Investment Corporation
PIDG Private Infrastructure Development Group
PSI private-sector instrument
PSW private-sector window
RDB regional development bank
S&P Standard and Poor’s
SDG Sustainable Development Goal
SME small and medium enterprise
TOSSD Total Official Support for Sustainable Development
UK United Kingdom
UMIC upper-middle-income country
UN United Nations
UNCTAD United Nations Conference on Trade and Development
US United States (of America)
USAID United States Agency for International Development
US IDFC United States International Development Finance Corporation

9
Executive summary
The need to mobilise private finance is at the finance in plugging the SDG financing gap in
heart of international discussions on how to developing countries by:
finance the Sustainable Development Goals
(SDGs) and move the needle from ‘billions’ •• reviewing the amounts of private finance
of dollars in development aid to ‘trillions’ of mobilised and estimating leverage ratios to
dollars in investment (World Bank, 2015). assess the scale and potential of blended finance
With an estimated SDG financing gap of $2.5 •• analysing the blended-finance landscape of
trillion a year in developing countries alone country groups and economic sectors
(UNCTAD, 2014), the international development •• focusing on low-income countries (LICs),
community is placing an increasing emphasis on where the need for additional finance is
blended finance.1 greatest, and identifying factors likely to
Blended finance uses public-sector constrain blended finance’s potential there.
development finance to spur additional private
investment in a bid to generate economic The report examines in detail the investment
growth and create jobs, thus lifting people out portfolios of the largest and most important
of poverty. The notion of ‘billions to trillions’ blended-finance actors, which account for more
(World Bank, 2015), though originally broader than three quarters of the total private finance
in meaning, has become synonymous with the mobilised in LICs, according to the OECD.2 It
mobilisation of private finance for development. analyses the most recent four-year period for
However, policy-makers often have lofty which comprehensive mobilisation data are
aspirations, with limited appreciation of its available.3 The institutions included are:
potential and limitations. The more official
development assistance (ODA) is channelled to •• MDBs: IDA, IFC and MIGA
blended finance, and the more blended finance is •• regional development banks (RDBs): ADB,
scaled up, the more pressing the need for better AfDB and EIB)
understanding of its potential to bridge the SDG •• bilateral DFIs of developed countries: AFD
financing gap. (France), CDC Group (UK) and OPIC (United
This report aims to provide hard evidence States) and its private-sector financing arm –
to inform the discussion on the role of blended Proparco – and Norfund (Norway).4

1 There is no common official definition of blended finance. For the purposes of this report, we define it as ‘the strategic use
of official development finance (ODF) to mobilise additional private capital flows to developing countries to achieve the
SDGs’. We discuss definitions and why we chose this focus in section 3.

2 Our selected multilateral development banks (MDBs) and development finance institutions (DFIs) mobilised 77% of
total private finance mobilised overall and 85% of total private finance mobilised in LICs, as reported in the OECD 2016
mobilisation survey for 2012–2015 (which excludes China) (Benn et al., 2017).

3 We use the 2016 OECD mobilisation survey for the 2012–2015 period, as it contains the most comprehensive semi-
disaggregated data available.

4 Annex 1 outlines our institutional selection process.

10
Key findings As policy-makers task MDBs and DFIs with
mobilising ever more private finance for LICs,
Finding 1 they should be aware that leverage ratios may
fall even further, requiring even bigger public
Each $1 of MDB and DFI invested mobilises on subsidies to incentivise private investment in
average $0.75 of private finance for developing more marginal investment opportunities.
countries, but this falls to $0.37 for LICs.
Expectations that blended finance can bridge the Finding 3
SDG financing gap are unrealistic: ‘billions to
billions’ is more plausible than ‘billions to trillions’. The potential of blended finance in LICs is
There is much talk by policy-makers of the hindered by factors such as poor investment
potential of blended finance to mobilise climate, lack of investable opportunities, lack
significant sums of private finance. High financial of tailored approach and low risk appetites of
leverage ratios are at the core of their arguments MDBs and DFIs.
for investing ODA in blended finance, but our Private commercial finance will not flow freely
research shows that real leverage ratios are to countries where the investment climate is
actually very low. Our database shows that, on challenging, markets are not functioning and
average, for every $1 of MDB and DFI resources the risk-adjusted rate of return is uncompetitive.
invested, private finance mobilised amounts to In fact, we estimate that 96.3% of private
just $0.37 in LICs, $1.06 in lower-middle-income finance mobilised through blended finance flows
countries (LMICs) and $0.65 in upper-middle- to countries with a credit rating, which most
income countries (UMICs). Leverage ratios are LICs do not have. Blended finance may tip the
generally low across sectors, with a slightly balance, but it will not work if the economic
higher ratio in the social sectors and the lowest fundamentals are not in place. So, the push for
ratios in LIC and middle-income country blended finance should not eclipse the need for
(MIC) infrastructure. grants to boost local investment environments.
Using concessional finance to blend can
Finding 2 help pioneer and create new markets, foster
innovation and invest at the earliest stages of
We need to better understand the poverty and projects, when risk levels are at their highest and
development impact of blended finance, as well when private investors need a greater degree of
as its true costs, to ensure value for money and risk mitigation. The data suggest, however, that
effective policy-making and allocation of ODA. MDBs and DFIs are primarily using less risky
We need to better understand the development senior debt rather than instruments that are
impact of blended-finance investment and the more risk appreciative to take on early-stage or
value for money of official ‘subsidies’ if we are to ‘pioneer’ risk, such as subordinated debt, equity,
understand their comparative value over other risk-sharing facilities, guarantees or grants.
forms of development financing. Low leverage The data also show very little variation in the
ratios suggest ODF will have to play a major role instruments used in different country income
in blended-finance investment. Our estimates groups. This suggests that the current blended-
suggest that the public sector (the MDBs and finance approach in LICs is not tailored to the
DFIs) has on average picked up 57% of the cost risk requirements of private investors, which may
of blended-finance investments to date and as limit the potential of blended finance to mobilise
much as 73% of the cost in LICs. Given that private finance in LICs.
the public sector picks up much of the cost, and Conservative MDB and DFI financing models
that often blended finance does not mitigate and the returns required on blended concessional
risk but merely transfers it from the private finance are dampening risk appetite and the
to the public sector, we need to understand ability to engage in LICs. MDBs’ willingness to
better the development impact and value for assume risk is hindered by their need to maintain
money of investing ODA in blended finance. a AAA credit rating, while DFIs need to remain

11
profitable and financially sustainable. Both must denting public trust. This is at odds with
take on greater risk if they are to mobilise more the blended-finance principles agreed by the
private finance for LICs. The answer to date has international community (DFI Working Group,
been for donors to provide more concessional 2017; OECD, 2018b). There is a clear disconnect
finance, which they blend with their own- between high-level political commitments to
account resources, but this has not spurred a transparency and accountability and operational
material shift in the overall risk appetite in the policies and rules.
MDB and DFI system. Interestingly, bilateral
institutions appear to play a more significant Recommendations
role in mobilising private finance for LICs than
MDBs and RDBs. The comparative advantages Recommendation 1
of various types of institution, their approaches
and how they can best complement each other If blended finance is to be scaled up, MDBs and
warrant further exploration. DFIs will need to get better at using blending
to mobilise private finance while managing the
Finding 4 higher level of risk this implies.
If blended finance is to be scaled up, leverage
The big push for blended finance risks skewing ratios will need to increase significantly. MDBs
ODA away from its core agenda of helping and DFIs will need to make fundamental changes
eradicate poverty in the poorest countries. to their business models and take on riskier
ODA per capita is higher in LICs than other projects. Changes could include:
countries, but the opposite is true for blended
finance. Changing policy incentives at the bilateral •• making more use of concessional finance and
and multilateral levels to encourage greater use subordinate instruments to meet the risk-
of ODA to mobilise private finance means we mitigation needs of the private sector in LICs
will see increased investment of ODA in blended •• using more concessional finance to fund
finance. But it is easier to mobilise private finance project preparation and early-stage project
in MICs and in ‘hard’ economic sectors. So, it development, as well as to foster the use of
is not surprising that blended finance is heavily more innovative risk-appreciative instruments
concentrated in MICs and flows predominantly •• revisiting the required rates of return on
to the ‘hard’ economic sectors (infrastructure, concessional resources used in blending and
banking and financial services), with very the ‘hurdle rates’ of bilateral DFIs: in other
little to social sectors (health, education, social words, accepting higher levels of financial risk
protection). This underscores the risk that the big •• MDBs assessing capital adequacy in a more
push for blended finance may deflect ODA from uniform way, allowing greater transparency of
the crucial investment needed to eradicate poverty scope to take on risk and incorporate callable
in LICs. MDBs and DFIs will need to adopt a share capital into capital adequacy models.
more tailored approach to ensure that this will not
divert ODA away from LICs. Recommendation 2

Finding 5 Donors need to think carefully about the

allocation of ODA and the risks and trade-offs of
Effective policy-making has been thwarted investing ODA in blended finance.
by the lack of a common official blended- There may be other public policy interventions
finance framework and poor data availability, that are more transparent and effective
hindering transparency and accountability, and in achieving development objectives than
undermining public trust in this approach. providing a direct subsidy to the private sector.
A lack of transparency and accountability For example, MDBs, DFIs and donors could
undermines official efforts to build the case for make greater use of grant finance to strengthen
more investment of ODA in blended finance, the investment climates of LICs, focusing on

12
country-led programmes of policy reform, Recommendation 3
financial-sector development and capacity-
building. Given that blended finance has not There is an urgent need for better data
targeted well the poorest countries, and is not and transparency.
used equally in all sectors, donors need to Efforts should be made to align and harmonise the
manage the risk that increased investment of OECD and MDB blended-finance frameworks.
ODA in blended finance could further exacerbate All institutions should publish disaggregated
the poor targeting of ODA, neglecting the project-level data. The OECD Development
countries and sectors that need it most. Assistance Committee (DAC) needs to resolve
outstanding issues on the treatment of private-
sector instruments in the modernisation of ODA
and make efforts to publish the ‘grant equivalent’
of blended-finance transactions to the OECD
Creditor Reporting System (CRS). This should be
disclosed publicly at a semi-aggregated level to
overcome the commercial confidentiality concerns
of MDBs, DFIs and donors.

13
1 Introduction and
overview

1.1 Introduction the investment in health, education and social

protection needed to eradicate extreme poverty:6
The need to mobilise private finance is at the heart
of international discussions on how to finance the •• First, while the volumes of private finance
SDGs. International and domestic public finance, mobilised through blended finance are
alone, cannot plug the estimated $2.5 trillion growing every year, the amounts are very
annual SDG financing gap in developing countries limited compared with the estimated SDG
(UNCTAD, 2014), so private finance must play financing gap. The best estimates of private
a crucial role. The international development finance mobilised through blended finance by
community is placing growing emphasis on MDBs, DFIs and donors in LICs and MICs
blended finance, which uses public development range from $3.3 billion (DFI Working Group,
finance to mobilise additional commercial capital 2018) to $27 billion annually (Benn et al.,
to bridge the financing gap and spur private 2017). This rises to $59.4 billion if we use
investment for economic growth and job creation the latest data on total direct and indirect
to lift people out of poverty. Aspirations are mobilisation (World Bank, 2018a).
high, but there is limited political appreciation •• Second, the private finance mobilised is
of the potential of blended finance in specific heavily concentrated in MICs, with very
contexts. Policy-makers need to better understand little mobilised in LICs. The best estimates of
when, where and how to use a blended-finance private finance mobilised by MDBs, DFIs and
approach,5 and the circumstances in which it donors in LICs appear to be in the range of
represents value for money. $725 million (Benn et al., 2017) to $1.6 billion
Changing policy incentives at the bilateral annually (World Bank, 2018a). The upper end
and multilateral levels to encourage greater use of the range rises to $5.3 billion if we take
of ODA to mobilise additional private finance the total direct and indirect mobilisation most
means we will see increased investment of recently reported by MDBs (ibid.).
ODA in blended finance. However, three trends •• Third, blended finance predominately mobilises
underscore the risk that this big push for blended private finance in ‘hard’ economic sectors, with
finance may steer ODA away from LICs and very little going to the social sectors.

5 It should be noted that the DFI community has developed broad ‘operational’ blended concessional finance principles
(DFI Working Group, 2017) for DFIs to internalise in their operational procedures. The OECD DAC has also agreed
‘high-level’ policy principles (OECD, 2018b).

6 This risk is explored in section 4.2.

14
 Blended finance is mobilising private finance exercised in their emphasis and interpretation,
in three sectors: infrastructure, banking and it is useful to look at leverage ratios, as they
financial services,7 and the productive sectors. can help policy-makers gauge the potential of
blended finance. To date, only highly aggregated
This report aims to recalibrate the financing- leverage ratios are available (for the main
for-development discourse by focusing on MDBs, for example). We go further, estimating
blended finance in LICs. It seeks to go further disaggregated leverage ratios (for instance, by
than the current literature by looking inside income category and sector). We find that $1 of
the investment portfolios of the largest official public investment by MDBs and DFIs mobilises
blended-finance actors overall and those just $0.37 of private finance in LICs, $1.06 in
mobilising private finance in LICs. It focuses LMICs and $0.65 in UMICs. We also find that
on the top three actors in each of the following leverage ratios are low and fairly consistent
categories: (1) MDBs, (2) RDBs, and (3) bilateral across sectors.
DFIs. The institutions we focus on are IDA, IFC, Second, we present a disaggregated analysis
MIGA, EIB, ADB, AfDB, OPIC, CDC, AFD, of the blended-finance landscape, classified by
Proparco and Norfund.8 This disaggregated country income. To date, there has been no
analysis is based on a unique new ODI database analytical breakdown of the blended-finance
of the blended-finance commitments of these landscape by national income and by institution.
institutions. Our selected MDBs and DFIs We observe that, in contrast to the overall trend,
mobilised 77% of total private finance mobilised bilateral DFIs appear to play a more important
overall and 85% of total private finance role in LICs than MDBs and RDBs and that
mobilised in LICs between 2012 and 2015, private finance mobilised in LICs is concentrated
according to the (2016) OECD mobilisation in relatively richer LICs. We find that
survey (which excludes China) (Benn et al., infrastructure is the largest destination sector for
2017), so we have covered the main traditional blended investment and that very few MDBs and
international actors. Our analysis excludes DFIs make any blended-finance commitments to
national development banks (NDBs), which LIC social sectors. We also find that loans are the
are important blended-finance actors in their most commonly used instrument for mobilising
countries. ODI is exploring the role of NDBs in private finance in LICs and that the instruments
blended finance and will publish its findings on used by MDBs and DFIs vary little across
the subject once its research is completed. country income groupings. We also find that
The goal of this report is to help calibrate project size is significantly smaller in LICs.
the discussion on how and where ODA should Third, we identify factors that are probably
be deployed and temper expectations as to the constraining the potential of blended finance in
potential of blended finance in LICs. It does this LICs. We observe limited use of concessional
in three ways: finance to blend in LICs compared with
First, we review the amounts of private finance UMICs and HICs and question whether more
mobilised and the financial leverage ratios in of it should be used in LICs. We argue that
LICs, which suggest that the potential of blended the potential of blended finance in LICs is
finance in LICs is limited. At policy-making level, constrained by several factors, most notably
there is much talk of the potential of blended poor investment climates, a lack of investable
finance to mobilise significant sums of private opportunities, the limited use of subordinate
finance. High leverage ratios are at the core instruments by MDBs and DFIs and the limited
of these arguments. While caution should be risk appetites of MDBs and DFIs.

7 Much of the investment categorised as being in the banking and finance sector is lent to local financial institutions, which
lend on to local end-borrowers (mainly small and medium enterprises (SMEs) in a range of sectors, so we do not know
the ultimate beneficiaries.

8 Annex 1 outlines our institutional selection process.

15
We make several recommendations to ease these 1.2.2 Part 2
constraints. In section 4, we discuss the big push for blended
finance, examining the rationale behind it and
1.2 Overview the changing incentives, which mean that we can
expect to see a significant step up in investment
The first part of this report deals with the of ODA in blended finance. This underscores
definitional and methodological issues that the risk of ODA being diverted from LICs. In
underpin our analysis (sections 2 and 3). The section 5, we discuss the potential of blended
second part of the report is an analysis of finance by comparing various official surveys and
the policy landscape and the blended-finance present an overview of global leverage ratios,
landscape in LICs (sections 4 to 7). including our own disaggregated leverage ratios,
which are low overall and very low in LICs.
1.2.1 Part 1 We argue that the potential of blended finance
The report discusses the various definitions of to mobilise private finance in LICs is limited
blended finance and the different methodologies and that policy-makers need to temper their
used to measure the mobilisation of private finance, expectations accordingly.
highlighting the implications of a lack of a common Section 6 maps the landscape of blended
framework for data analysis, informed discussion finance in LICs. It focuses on the largest
and effective policy-making (section 2). The report traditional MDBs and DFIs engaging in blended
then explores how we address these challenges, by finance overall and in LICs and analyses their
outlining the scope and methodology underpinning blended-finance portfolios in LICs. We focus
our analysis. We summarise how we have defined our analysis on institutions for which data
blended finance for the purposes of this research are available, based on our unique dataset of
and give an overview of our approach to data blended-finance commitments. In section 7,
collection and data usage, including the creation we identify several factors that are probably
of a new ODI blended-finance commitment constraining the potential of blended finance in
database (section 3). A detailed description of the LICs, underlining the need for a more tailored
methodology used to build this unique database approach. We summarise our conclusions in
can be found in Annex 2. section 8.

16
2 Blended finance:
no common conceptual
framework

Key findings
•• There is no common definition of blended finance at the official level. The lack of a
common conceptual framework presents challenges in terms of data collection, analysis
and comparability.
•• Poor-quality data on the private finance mobilised and ODF invested, at both the aggregate
and disaggregated project level, result in a lack of transparency that hinders informed analysis
and understanding, informed discussion and, ultimately, effective policy-making.
•• There is no consistent picture of the size and scope of blended finance or its development impact.
•• A lack of transparency and accountability undermines official efforts to build the case for
increased investment in blended finance and undermines public trust in this approach.

2.1 Numerous official definitions As the development finance landscape and

thinking have evolved, so has the definition of
There are myriad definitions of blended finance.9 blended finance. Where the discourse appears to
This presents obvious challenges, as it means have settled for now, we can see several common
different things to different institutions and attributes that feature in the various definitions,
actors. At the official level, there is no common namely:
definition, so no common methodology for
measuring blended finance on a consistent •• the use of concessional development finance
and comparable basis, with obvious adverse •• the intent to mobilise additional finance,
implications for transparency, accountability and primarily private commercial finance11
effective policy-making.10 •• some form of development impact associated
with the investment.

9 See chapter 3 of OECD (2018b) and annex A of Development Initiatives (2016) for definitions and concepts.

10 There is an appreciation of these differences and nuances among practitioners at operational level, but far less
understanding elsewhere.

11 For many years, blending public concessional resources with public non-concessional resources dominated the blended-
finance approach. Much European Union (EU) blending was of this form, for example. EU grants were combined with
other public (non-concessional) and, to a limited extent, private-sector resources to support public, private or mixed
investment projects. The discourse has now shifted away from ‘public–public’ blending to ‘public–private’ blending;
namely, the use of public resources to leverage private commercial finance.

17
 Additionality is key. Two main forms of combining concessional finance from
additionality are generally understood in the donors or third parties alongside DFIs’
development finance context: (1) financial normal own-account finance and/
additionality, when public investment results or commercial finance from other
in private investment that would not have investors, to develop private-sector
materialised without it; and (2) development markets, address the Sustainable
additionality,12 whereby development impacts are Development Goals (SDGs), and
secured in a commercial investment that would mobilise private resources.
otherwise not have materialised. An assessment (DFI Working Group, 2017: 3)
of additionality, especially financial additionality,
is complex and challenging, not least because of MDBs and DFIs distinguish between
the lack of counterfactual, and lies outside the concessional finance provided by donors or third
scope of this research (Pereira, 2015; Carter et al., parties and their own-account non-concessional
2018). For the purposes of this research we adopt resources. Most MDB and DFI mobilisation of
a conservative approach and assume that all the private-financing operations is funded solely
mobilised private finance reported is additional. from own-account resources and is not identified
as blended finance by the institutions themselves.
2.1.1 OECD DAC versus MDB, DFI and The operations are identified as blended finance
United Nations definitions by the OECD, however.
In essence, there are two main definitions that The MDB and DFI definition is broadly
have gained traction in the development finance aligned with the definition adopted by the
discourse. First, the OECD DAC definition,13 UN in the Addis Ababa Action Agenda of the
which is much broader in scope compared to Third International Conference on Financing
the second, narrower definition adopted by the for Development, which focuses on the use of
MDBs, DFIs and the United Nations (UN). concessional finance and defines blended finance
The OECD DAC defines blended finance as financing that:
broadly as:
combines concessional public finance
the strategic use of development finance with non-concessional private finance
for the mobilisation of additional and expertise from the public and
finance towards the SDGs in developing private sector. (UNDESA, 2015: 24)
countries. (OECD, 2018a: 50)
The three definitions essentially chime on the
Development finance, in the context of this ‘output’ side of the equation (the mobilisation
definition not only includes ODF, both ODA and of commercial private finance), but differ on
other official flows (OOF) (for example, MDB and the ‘input’ side (what is invested to spur that
DFI own-account resources),14 but also private mobilisation), as summarised in Table 1.
finance governed by development mandates (such In a nutshell, what the OECD defines as
as philanthropic finance). Additional finance refers ‘blended finance’, the MDBs and DFIs refer to
to public and private commercial finance with a as ‘mobilisation’ of private finance. MDBs and
non-developmental purpose. DFIs do not consider private-sector financing
MDBs and DFIs have adopted a narrower, operations wholly funded by their own-account
more precise definition, which focuses on the use resources to be blended finance. MDBs and
of concessional finance: DFIs only consider own-account private-sector

12 Development additionality can also include other aspects of additionality, such as operational and institutional additionality.

13 The Business and Sustainable Development Commission concurs with and adopts the OECD DAC definition.

14 Own-account resources refer to the institutions’ normal operational financing resources.

18
Table 1 Blended finance – definitional differences

‘Input’ classified as blended finance OECD definition MDB and DFI, UN definition
MDB and DFI own-account resources Yes No
(invested on its own)
OOF Yes No
(with a development mandate)
Concessional ODA Yes Yes
(Donor or third-party concessional finance)
Philanthropic capital Yes No
(with a development mandate)
Impact funds Yes No
(Investment below market rate)

financing operations that are blended with discuss the OECD and MDB surveys later in
third-party or donor concessional resources to this report, here, we note the major
be blended finance. This is a very small subset methodological differences that affect their
of their total private-financing operations and interpretation.
a very small subset of what the OECD would
identify as blended finance.15 2.2.1 OECD DAC versus the MDB
We thus find ourselves in an unsatisfactory mobilisation measure
situation in which the OECD DAC defines MDB methodology differentiates between ‘direct’
blended finance in a way that prevents thorough and ‘indirect’ mobilisation. MDBs define direct
reporting16 and with which the MDBs and DFIs mobilisation as:
do not agree. Consequently, numerous MDBs
and DFIs do not report in full to the OECD DAC financing from a private entity on
and OECD CRS. commercial terms due to the active
and direct involvement of a MDB
2.2 Understanding the ‘output’: leading to commitment. Evidence of
mobilising private finance active and direct involvement include
mandate letters, fees linked to financial
Further confusing the issue is the fact that the commitment or other validated or
OECD and MDBs have developed two different auditable evidence of a MDB’s active
methodologies to measure the mobilisation of and direct role leading to commitment
private finance (the ‘output’), which yield vastly of other private financiers.
different results and are not comparable. As we (World Bank, 2018a: 8)

15 Only a very small percentage of the total volume of MDB and DFI private finance operations every year would be
classified as blended finance by MDBs and DFIs. This is calculated at approximately 5% during the period 2014 to 2016
(DFI Working Group, 2017) increasing to 8.8% in 2017 (DFI Working Group, 2018).

16 The rationale behind the adoption of such a broad definition is clear, however, as if it is defined it will be difficult for
the OECD DAC to report on the definition in its entirety as there is no obligation for providers of development finance
outside the DAC to report to the OECD (e.g. philanthropic actors).

19
 They define indirect mobilisation as: clear for MDB direct mobilisation, it is unclear
for indirect mobilisation (OECD, 2018c). In light
financing from private entities provided of these concerns, caution should be exercised
in connection with a specific activity in interpreting the MDBs’ indirect mobilisation
for which an MDB is providing figures. We report the MDB direct and indirect
financing, where no MDB is playing mobilisation figures separately.
an active or direct role that leads to Also problematic is the fact that DFIs, which
the commitment of the private entity’s have not previously reported to the DAC CRS,18
finance. (ibid.) have started to adopt the MDB methodology (the
EDFI adopted it in 2018).19 The MDBs are also in
There are differences between the MDB and talks with the International Development Finance
OECD methodologies, meaning the respective Club (IDFC) on adopting this methodology. As
survey results are not comparable. Essentially, the the number of DFIs using the MDB methodology
differences centre on two key issues. expands, it will be important to get further clarity
The first is causality. Although the OECD on the indirect mobilisation approach. Worked
does not distinguish between direct and indirect examples along the lines of those published by the
mobilisation, its methodology captures both, in OECD would help (OECD, 2018d).
effect, but only if causality can be demonstrated. Another important difference is the level of
The second is attribution. The MDB detail that is reported and the availability of
methodology for direct mobilisation attributes all public data underpinning the survey results.
the private finance mobilised to the lead MDB, The MDB survey reports on a collective basis,
so no attribution is made to other MDBs, DFIs, with data disaggregated by MDB, by country
or domestic public investors (such as NDBs) that income and region, and by infrastructure sector
co-finance the investment (World Bank, 2018b). (World Bank, 2018a). Disaggregated data
In contrast, the OECD methodology attempts at country and project level are not publicly
to consider the level of risk assumed by official available due to client confidentiality agreements,
investors. For certain instruments (such as shares which do not permit public disclosure of the
in collective investment vehicles (CIVs) and direct terms of the financing package. However, semi-
corporate investments), the OECD methodology aggregated data underpinning the OECD survey
attributes a higher share of the private finance are publicly available,20 although it should be
mobilised to official investors that are exposed noted that not all DFIs reported in full to the
to a higher level of risk. The MDB methodology OECD mobilisation survey, in part due to client
does not take this into account (ibid.).17 There is confidentiality, but also because of the reporting
also disagreement between the two ‘camps’ on burden, as the MDB and OECD definitions and
the issue of double counting. The MDBs are of methodologies are not aligned.
the view that their methodology provides a clear The challenges that this presents for a
basis for attribution that avoids double counting comprehensive and consistent picture of the
(ibid.). The OECD believes that, while this is landscape has been recognised and there have

17 The MDBs note that typically all parties invest on the same terms and there is no reason to think that the lead MDB bears
less risk (IFC, 2018, unpublished assessment).

18 The OECD notes that some of these DFIs have not been reporting to the DAC CRS system (ibid.).

19 Of the 15 EDFI members 12 contributed to the 2018 MDB mobilisation survey.

20 The OECD survey is noteworthy in this regard, as the only publicly available survey that attempts to present a more
disaggregated analysis accompanied by a semi-disaggregated dataset (Benn et al., 2017). See www.oecd.org/development/
mobilisation.htm. However, it should be noted that the level of disaggregation is limited, as the data are not disaggregated
by recipient and institution. Our understanding is that the survey respondents did not want data to be published at this
level, although it is not obvious what commercial confidentiality issues would be breached by doing so.

20
been calls to align and harmonise the two The OECD has made efforts to address this
methodologies. Discussions between the OECD issue, with a modernisation of ODA that will
and the MDBs have started, but the prospects enable the counting and reporting of investments
for alignment and harmonisation do not look in instruments to mobilise private finance (i.e.
promising (OECD, 2018c). At a technical level, private-sector instruments, or PSIs) as ODA.
the OECD is working on case studies with the Alas, its work has reached an impasse, as the
MDBs to better understand the differences DAC was unable to secure agreement on how
in methodological approach, with a view to to account for PSIs in ODA at its high-level
identifying where bridges can be built. These case meeting in October 2017. The outcome is an
studies should be made publicly available to help unsatisfactory situation, whereby donors can
inform independent analysis and understanding, report PSI investment as ODA, even though the
while efforts to align and harmonise frameworks reporting rules have not been finalised. Donors
should ideally involve a variety of stakeholders will be able to report using either an ‘instrument’
to inject independence into thinking on these or ‘institutional’ approach (OECD, 2018e).22
critical issues. This will result in inconsistent ODA figures
being reported by donors, with obvious adverse
2.3 Understanding the ‘input’: consequences for the quality of data and for
ODF investment transparency. These concerns, however, mask a
more pressing problem. Previously, much DFI
To understand the value for money of blended- activity did not qualify as ODA, mainly because
finance approaches, we need to understand what the investments were non-concessional. Now,
ODF has been invested to mobilise private finance however, these flows can be reported as ODA
(the ‘input’) and how much this has cost (the rather than OOF. This needs to be resolved as
subsidy). Data on what has been invested and the a matter of urgency, as it has the potential to
subsidy provided are not publicly available for undermine the very concept of ODA, the quality
the most part. For example, the aforementioned of ODA data and, hence, public trust.
OECD and MDB surveys do not provide any Although disaggregated data underpinning
data on what ODF has been invested (that is, the MDB survey are not disclosed, most
MDB and DFI own-account resource and donor (though not all) MDBs publish project-level
concessional funds). A review of the literature data. The quality of the published data varies
reveals that only one official survey, by the DFI significantly from institution to institution,
Working Group on Blended Concessional Finance, limiting the level of meaningful analysis that can
has attempted to estimate the aggregate amount be carried out. Furthermore, the data are not
of ODF21 invested and the private finance it has reported in standard form, as each institution
mobilised. The results of the group’s first survey has its own disclosure policy,23 hindering
were presented at a highly aggregated level, with comparability. As investment in blended finance
no supporting data made available (DFI Working is scaled up and these institutions increase their
Group, 2017). Its latest survey presents a more engagement in blended finance, a common
disaggregated analysis by sector, geography and international reporting standard for ODF
instrument, but supporting data are still not investment and mobilised private finance would
available (DFI Working Group, 2018). be extremely useful.

21 Split between MDB and DFI own-account resources, donor concessional funds and public contributions.

22 Under the ‘institutional’ approach, institutions can report on a cash-flow basis the ODA-eligible share of capital funding
to their DFI as ODA; under the ‘instrument’ approach, institutions can report individual loan and equity investments at
the transactional level on a cash-flow basis.

23 See Annex 2 for more discussion on this.

21
2.4 The urgent need for better data There have been hard-won gains on
and transparency transparency and accountability through the
aid effectiveness agenda and these need to be
Because of different definitions, varying survey preserved. ODA invested in blended finance
coverage with respect to participating institutions should be subject to the same transparency rules
and instruments, and different methods of as non-blended ODA. Ideally, blended-finance
measuring the mobilisation of private finance, flows (input and output) should be reported by the
we don’t have a clear and consistent picture of MDBs, DFIs and donors to the OECD CRS system
how much private finance is being mobilised at a transactional level25 (using the ‘instrument’
through the use of blended finance (‘the output’). approach) to allow analysis of blended-finance
Likewise, we do not have a clear understanding flows by provider, destination country, sector and
of how much ODF is being invested, its cost instrument, be they ODA or OOF. Furthermore,
(‘the input’) and what development impact the the OECD CRS should also facilitate analysis of
investment is having (‘the outcome’). these flows at investment level, as multiple actors
This situation is very much at odds with may participate in a single investment.
the blended-finance principles agreed by On the sensitive issues of pricing,
international donors at the OECD DAC24 subsidisation and commercial confidentiality,
(OECD, 2018b) and by the MDBs and DFIs (DFI several ideas deserve consideration. For MDBs
Working Group, 2017), which emphasise the and DFIs that report to the OECD using the
importance of transparency and accountability. ‘instrument’ methodology, the grant equivalent
The gap between high-level pronouncements and of individual transactions26 could be reported
practice needs to be urgently addressed by the using the OECD ‘instrument’ methodology, but
donor-shareholders of the MDBs and DFIs and publicly disclosed at a semi-aggregated level
those on the OECD DAC. (for example, by country and sector for each
Blended finance structures can be quite MDB and DFI). For MDBs and DFIs that
complex, comprising layered capital structures report to the OECD using the ‘institutional’
involving numerous institutions, several of which approach, the grant equivalent of transactions
are private, and the use of various financial semi-aggregated by country and sector using
instruments. This complexity, combined with the ‘instrument’ methodology could be reported
the fact that the transaction is commercial in to the OECD and publicly disclosed.27 In this
nature, presents a challenge for data collection way, the actual terms of the investment are
and publication, especially data on the embedded not reported or publicly disclosed, addressing
subsidy in each blended-finance investment. Many confidentiality concerns.
MDBs, DFIs and donors argue that commercial In order to allocate and invest ODF most
confidentiality restricts what disaggregated effectively (especially ODA), policy-makers
transactional data can be published, but a balance need to understand the value of blended-finance
needs to be struck, as ODA is being used to approaches versus other forms of financing
subsidise private finance and there needs to be to achieve the same objective. At the granular
accountability and transparency of use. level,28 policy-makers, the taxpayer and those

24 G7 ministers, at their meeting in Toronto, also emphasised the need for a common understanding of blended finance and
transparency in its use. They agreed to broaden awareness of the OECD DAC blended-finance principles (G7, 2018).

25 The ‘institutional’ approach currently allowed by the OECD removes transparency at the transactional level.

26 It would not be possible to calculate a grant equivalent for an equity investment.

27 MDBs and DFIs could be allowed to redact if there were only one transaction in a given country in a given year.

28 This report focuses on the perspective of public policy, but it is also critical for the private sector, to enable it to engage in
blended finance (performance data).

22
affected by a project need to know how much country, sector and instrument. This, we would
was invested, what it cost (the subsidy), what argue, is a basic data requirement and crucial
‘additional’ private finance was mobilised and its amid the sensitivities to, and risks associated
development impact,29 disaggregated by provider, with, subsidising private investment.30

29 So far, the evidence base on the development impact of blended finance is limited. Focus should also be turned to tracking
and measuring development impact, as the objective of blended finance is to enable private investment in the SDGs and,
hence, increase the overall development impact of public and private finance to deliver on global agendas, such as the
SDGs and 2015 Paris Agreement.

30 Prime risks include distorting markets, crowding out the private sector and prioritising financial returns over development
impacts.

23
3 ODI approach: scope
and methodology

3.1 Definition and scope rather than a conceptual view on the definition of
blended finance. Like the OECD approach, it will
The focus of this report is MDBs and DFIs, so we identify MDB and DFI private-sector investment
have tried to concentrate our analysis on blended operations funded purely by MDB and DFI own-
concessional finance using their definition, account resources as blended finance. From an
identifying projects in their public datasets that MDB and DFI perspective, this will overestimate
have blended third-party or donor concessional the amount of private finance mobilised through
resources with their own-account resources. blended finance.
However, data on concessional finance invested While we understand the rationale behind the
to mobilise private finance are almost impossible differing definitions, we believe the most useful
to obtain,31 so we have been unable to undertake definition of blended finance should be guided
a landscape analysis using the narrower MDB by public policy concern over the effective use of
and DFI definition. ODF and, therefore, focus on the public subsidy33
Because of this poor data availability, we provided to mobilise additional private finance
have had to focus on a broader definition, for development. This chimes with the MDB, DFI
closer to that of the OECD. Even so, we have and UN definition, which focuses on blended
limited our focus to the use of the official concessional finance (which subsidises private
component of development finance, due to investment below market terms) and gives a nod
the lack of comprehensive and consistent data to the OECD definition by including MDB and
on private development finance. We have, DFI own-account investment activity, where
therefore, confined our analysis to the strategic explicit and/or implicit subsidies are purely
use of official development finance to mobilise funded by MDB and DFI own-account resources.
additional private finance for development It stops short of including private development
purposes.32 finance, however. Data on subsidies to mobilise
It should be noted that by delineating our private finance are not disclosed by MDBs,
focus in this way, we are limiting the scope of our DFIs and donors, so it is not possible to analyse
focus to what is determined by data availability blended finance through this lens.

31 We were able to identify the IFC’s concessional projects. The IFC publishes a list of donor or third-party concessional
funds that it uses to blend, which we confirmed with the IFC. We then reviewed every project description in the IFC
database to identify whether any of these funds had been used at project level. However, our analysis was constrained by
the fact that project descriptions did not disclose the amount of concessional finance invested, just that it was used.

32 This is essentially the official-finance subset of the OECD definition of the input side and includes concessional and non-
concessional capital (MDB and DFI own-account resources). Our work excludes analysis of private development finance.

33 We borrow the definition of ‘subsidy’ from Carter (2015: 5), who defines a subsidy as ‘any intervention by a public
development agency, at the project level, that has the effect of raising expected risk-adjusted returns for private investors.

24
3.2 Data blended-finance universe, as we do not cover all
MDBs and DFIs, and exclude NDBs and China.
To analyse the private finance that has been Our selected MDBs and DFIs mobilised 77%
mobilised (the ‘output’), we rely on the OECD of the total private finance overall and 85% of
data,34 as this is the most disaggregated dataset total private finance in LICs between 2012–2015,
publicly available. Still, it is not ideal: it is not according to the (2016) OECD mobilisation
comprehensive, covering only five instruments, survey (which excludes China) (Benn et al.,
and the data are semi-disaggregated. 2017), so we have covered the main traditional
There are no readily available data to help us international actors.
understand the ODF that has been invested (the More information about the quality of the
input). We have, therefore, built our own database institutional datasets and the publicly available
of ODF committed by the largest MDBs and DFIs data we have used to build our database can
engaged in mobilising private finance (their own- be found in Annex 2. We recognise that our
account resources).35 We identified 11 institutions dataset is not perfect, but in the absence of
for study, but data were only available for 9 of publicly available, good-quality, disaggregated
them. Our data were drawn from those reported comparable data, we believe our approach is
publicly in the institutions’ project-level datasets. sensible and that the data on ODF committed
Unfortunately, the quality and availability of data are in the right general area.36 We also believe
vary considerably from institution to institution our analysis raises important policy issues for
and none has published a complete public dataset further analysis and discussion by practitioners
on private finance mobilised at project level. and policy-makers, highlighting the urgent need
It is important to note that our data for alignment of conceptual frameworks, better-
and analysis focus on a subset of the total quality data and transparency.

34 The public OECD mobilisation survey dataset (Benn et al., 2017) contains disaggregated data on recipients, but these are
not linked to the institutions. From communication with the OECD, we were able to obtain a more disaggregated dataset
(OECD, 2017), which does contain information on which institutions mobilise in which recipient countries. However,
these data are not available by year or by agency. Various sections of our analysis require the use of one dataset over the
other, leading to some discrepancies between numbers. Throughout the analysis, we have highlighted which dataset is
used by referencing Benn et al. (2017) or OECD (2017).

35 Data on disbursements and outstanding amounts are not generally available, so we have used commitment data.

36 We compared the sum of the project-level commitment data for each institution with the aggregate sums reported by the
institutions in their annual reports. The sums were broadly similar, so we were reassured about the completeness of our
dataset and our analysis of total figures. Standardisation of the datasets in terms of instrument, sector classification or
poor data may result in some misclassification of instruments and/or sectors.

25
4 The big blended-
finance push
Key findings
•• There is a strong conceptual underpinning, rooted in public economics, supporting the
argument for donors to increase ODF investment in blended finance. On balance, we would
expect to see more public subsidy funded by blended concessional finance in LICs than in
MICs, given the more pronounced existence of market failures in LICs.
•• Private investment is set to play a critical role in financing the achievement of the SDGs.
The international community has, therefore, shifted its emphasis to the use of ODA as a
catalyst for mobilising additional private finance.
•• At the bilateral level, ODA reporting rules are changing to allow donors to report as ODA
their capitalisation of DFIs or their investment in instruments for mobilising private finance.
In future, donors will be able to report significant amounts of private finance crowded in
using ODA.
•• At the multilateral level, MDBs are tasked with better utilising their resources to increase the
mobilisation of private finance, and targets have been set to increase mobilisation by the G20.
•• There is a risk, however, that this shift in the policy landscape and the ensuing rise in
investment of ODA in blended finance will not deliver for LICs. It may steer ODA away
from LICs, as it is easier to mobilise private finance in more stable and mature markets.
Mobilisation targets may also shift emphasis away from ensuring financial additionality and
prioritising development impact.
•• Donors need to understand and manage these risks to ensure that blended finance does not
exacerbate the poor targeting of ODA and that investments are better targeted to help support
LICs that cannot finance the eradication of extreme poverty from their own resources.

4.1 The argument for blended The rationale behind this approach is firmly
finance rooted in public economics. Markets are
‘imperfect’ and the presence of market failures,
Blended finance seeks to unlock private which are especially pronounced in developing
commercial investment in SDG outcomes that countries, results in sub-optimal37 levels of
would not happen otherwise. It does this by private investment and provision of goods and
using ODF to provide a subsidy to bring the services, so the public sector intervenes to correct
risk-adjusted rate of return on investment in for the market failure. One form of public
line with the market, increasing the allure of intervention to correct for the underprovision
the investment from a private commercial of socioeconomic beneficial goods and services
investor perspective. is a public subsidy. Blended-finance approaches,

37 From society’s point of view. This argument focuses on production externalities.

26
which subsidise the private sector, can be justified 4.2 Increased ODA investment
when they make high-impact development in blended finance due to policy
investments financially viable (essentially, those
projects where the socioeconomic returns exceed
changes
the private commercial returns).38 Blended-finance approaches and the use of PSIs
The level of the public subsidy will vary with have been used in the climate-finance arena and
the level of market development. At one end of the at national level (for example, by NDBs) for
spectrum, where markets do not exist, pioneering a number of years. The approach is not new
investments are made to create markets (for in and of itself. What is new is the increasing
example, in LICs and fragile and conflict-affected interest of donors in investing ODA in these
states) where there is a greater need for public approaches, and a shift in emphasis from ‘public–
subsidy, as the risks are likely to be too great public’ blending to ‘public–private’ blending,41
for private investors. As the market develops, which places a greater emphasis on mobilising
however, there is less need for public subsidies, commercial finance and the lofty ambition
so commercial forms of finance become more assigned to it in shifting the development
important.39 There is, thus, a dynamic logic to financing needle from ‘billions to trillions’.
blended finance, whereby blended finance should
successively enter new markets and sectors (when 4.2.1 Fundamental policy changes
socioeconomic returns exceed private returns) Fundamental shifts in the policy landscape
and then exit established blended-finance markets and supportive policy signals will result in
when commercial investors eventually take over increased investment of ODA in blended-finance
in full (when private returns exceed the benefits approaches.
to society). This rationale sits at the core of the The Addis Ababa Agenda for Action
argument made by proponents of blended finance (UNDESA, 2015: 27) recognised the central
for the use of subsidies.40 Viewed in this way, the role of private investment and, for the first
more pronounced the degree of market failure, time, the potential of blended finance, assigning
the higher the level of subsidy funded by ODF (in an important role to the catalytic use of
other words, greater concessionality). As market ‘international public finance, including ODA to
failures are more pronounced in LICs, one would mobilise private finance’. We see a fundamental
expect to see higher levels of subsidy in these shift in the development finance landscape,
countries (greater use of concessional funding) with the international community giving
than in MICs. MDBs and DFIs a key role in blending ODA ‘to
scale up financing for development’ (ibid).
The 2015 Paris Agreement (United Nations,
2015), where 195 countries submitted national

38 See Carter (2015) for a more detailed exploration of why donors may choose to subsidise the private sector and Warner
(2013) for an exploration of when this approach may be justified.

39 The idea that the level of public subsidy should fall in a given market over time is central to the DFI blended concessional
finance principles (principle three on commercial sustainability) (DFI Working Group, 2017) and the OECD blended
finance principles (principle two on designing blended finance to increase the mobilisation of private finance)
(OECD, 2018b).

40 This is the case when it is expected that full commercial markets can be developed over time and that the market will
supply the optimal level of goods and services. This will not always be the case. There are some markets where the
subsidy is more permanent, as demonstrated by the existence of widespread subsidies in many OECD markets, such as
public transport.

41 Public–public blending essentially mixes concessional ODF with non-concessional ODF to create a new form of slightly
less concessional ODF. Much EU blending was in this form, but the emphasis has shifted to public–private blending, as
evidenced by the new European Fund for Sustainable Development (EFSD) and the EU’s new external investment plan.

27
climate action plans outlining the investment and there have been calls to set MDB and DFI
required to keep global warming below 2°C, mobilisation targets (Blended Finance Taskforce,
saw developed countries reiterate their intention 2018b). Indeed, in 2017, the G20 endorsed a
to scale up climate financing to mobilise target to increase MDB mobilisation of private
$100 billion per annum by 2020, and extend finance by 25%–35% by 2020.43 MDBs have
this to 2025, to help developing countries deal taken up the gauntlet and formed a working
with climate change. Given the interdependence group, which has sought to harmonise blended-
of the climate-change adaptation and mitigation finance approaches across the MDB system, for
agenda and the SDGs, blended-finance example, through the adoption in October 2017
approaches are likely to play a significant of the Enhanced Principles for Blended Finance44
role in crowding in the additional private and through the publication of the MDB global
finance required to meet these internationally toolbox to advance private-sector investment
agreed goals. (IFC, n.d.). The group has also developed a
Changing incentives at the bilateral level. methodology to estimate the amount of private
Significant efforts are underway in the OECD finance the MDB system is mobilising and has
DAC to build the case for blended finance and published two reports, discussed in section 5.1.
institutionalise the approach. ODA reforms
by the OECD DAC (OECD, 2017) and the 4.2.2 Increased investment of ODA in
development of the new Total Official Support blended finance
for Sustainable Development (TOSSD)42 metric Increased bilateral investment in DFIs. At the
will incentivise increased investment of ODA bilateral level, blended-finance approaches are
in blended-finance approaches, in addition mainly channelled through DFIs and there is
to reporting investment in blended finance as growing interest in their role and how they can
ODA, as donors will be able to report significant mobilise ‘additional’ private commercial finance
amounts of private finance mobilised in TOSSD. above and beyond their own-account resources.
As traditional donors run up against budget We expect this increased interest and intention
constraints, investing ODA in this way will be to scale up investment in blended finance to play
attractive, as it will be easier to align to domestic out in the establishment of new DFIs, such as
political and economic interests in the context FinDev Canada, established in January 2018, and
of rising ‘aid nationalism’ (Gulrajani, 2017). the planned new US International Development
The OECD DAC has also agreed blended- Finance Corporation (US IDFC). We also see
finance principles, which aim to guide the use increased investment in DFIs – for example,
of ODF and which set out the steps required for the additional £3.5 billion capitalisation of
blended finance to achieve to scale and impact the CDC Group, announced in October 2017,
(OECD, 2018b). corresponding to around 8% of the UK’s ODA
Changing incentives at the multilateral level. budget for the next five years.45 As noted, donors
A more effective multilateral finance system is will count this investment as ODA using either
high up the G7 and G20 political agendas. MDBs the ‘institutional’ or ‘instrument’ method agreed
have been tasked with exploring balance-sheet by the OECD DAC as part of efforts to reform
optimisation operations to increase mobilisation ODA and incentivise its use as a catalyst.

42 See www.oecd.org/dac/financing-sustainable-development/tossd.htm.

43 The Hamburg Principles and Ambitions on crowding in private finance developed by the G20 International Financial
Architecture Working Group (G20 Germany, 2017).

44 See DFI Working Group (2017). This builds on IFC (2013).

45 The UK will report the £3.5 billion capitalisation as ODA to the OECD CRS under the ‘institutional’ approach. It is
estimated that CDC will ‘receive up to 8% of the UK ODA budget or GBP 600-700 million per year on average over the
next 5 years’ (OECD, 2018f: 3).

28
 Increased multilateral investment in blended Second, mobilisation targets run the risk of
finance. At the multilateral level, we expect a step diverting MDB and DFI focus away from financial
up in blended-finance investment, as demonstrated additionality, as these targets may encourage
by the IFC’s new $5.5 billion capital injection, MDBs and DFIs to provide unnecessary subsidy
announced in April 2018. We also see the and invest in projects that the private sector would
establishment of new multilateral windows, such have invested in on a standalone basis. Focus is
as the $2.5 billion IDA Private Sector Window also likely to be shifted away from development
(PSW) in July 2017,46 which aims to mobilise $6 impact, as it is generally harder to mobilise
billion–$8 billion in additional private investment commercial finance for projects that prioritise
in IDA-only countries, with a focus on fragile strong developmental benefits.
and conflict-affected states, and funds such as This risk is a real concern. The eradication
the EFSD in September 2017 and the new EU of extreme poverty lies at the core of the SDG
External Investment Plan, which aims to mobilise agenda, yet ODI’s latest research finds that the
€44 billion of extra public and private investment. allocation of Country Programmable Aid (CPA)48
is not well targeted at countries that cannot
4.3 Risk of leaving the poorest afford to finance the eradication of extreme
countries behind poverty from their own domestic resources (in
other words, all LICs, except Tajikistan, and
This increased investment of ODA in blended some LMICs). The research finds that even
finance runs the risk of diverting ODA allocation though UMICs have over 300 times the potential
away from LICs and the investment in health, revenue per person living in extreme poverty,49
education and social protection needed to they receive 10 times more CPA per person living
eradicate extreme poverty. This risk essentially in extreme poverty than an LIC. For LMICs,
emanates from the setting of mobilisation targets this falls to 10 times the potential revenue and
in the absence of a coordinated framework. 7 times more CPA per person living in extreme
First, mobilisation targets will be easier to poverty. Yet the financing gap is especially acute
achieve in more stable and mature markets, due in LICs, where 96% of the countries that are
to their lower levels of perceived or actual risk, ‘severely challenged’ and cannot afford even half
greater ease of doing business, etc. This is likely the investment needed in health, education and
to result in a further concentration of MIC and social protection from domestic resources, even if
hard-sector financing in MDB and DFI portfolios they raised taxation levels to the maximum level,
at the expense of investment in more challenging are LICs (Manuel et al., 2018).
countries and sectors with higher financial and Comparing ODA flows and blended finance in
development additionality. Some argue that this more detail, we can see that far less blended finance
risk can be reduced if there is careful calibration is aimed at the poorest countries (Figure 1). While
of targets across the MDB and DFI system, and ODA per capita is higher in LICs, the opposite is
within the institutions themselves, but there are true for blended finance. If we assume the purpose
potential pitfalls.47 of ODA is to eliminate poverty and look only

46 The IDA18 replenishment of $75 billion included a 3% allocation of $2.5 billion over three years to create the IDA18
IFC MIGA PSW, which started operation on 1 July 2017. The objective of the window is to support the mobilisation of
private-sector investment in IDA-only countries, with a focus on fragile and conflict-affected states, by supporting IFC-
and MIGA-led transactions.

47 For more on potential pitfalls of setting mobilisation targets, see Carter (2018).

48 CPA is used as a proxy for aid that is spent in recipient countries. As such, it excludes donor administration costs, debt
relief and humanitarian aid.

49 The report identifies 48 countries that are unable to fully fund core poverty eradication expenditure (such as health,
education and social protection) from their own domestic revenue; 31 are LICs and 17 are LMICs.

29
at those people living in poverty, the trend is the It is still early days, however, and data
inverse of what we would expect for both flows – limitations have meant that we have been unable
the poor populations in UMICs receive far more to quantify how much ODA has been invested
than the (larger) poor populations in LICs and in DFIs and PSIs, or the destination countries of
LMICs.50 This pattern is much more pronounced investment. We have, therefore, been unable to
for blended-finance flows. While a person living quantify the risk of increased investment of ODA
in poverty in MICs receives 2.4 times more ODA in blended finance diverting traditional ODA
than a person living in poverty in LICs – a ratio of investment away from LICs. Although blended
1:2.4 – they receive 47 times the amount of blended concessional finance constitutes a small part of
finance – a ratio of 1:47. MDB and DFI private-finance operations, it had
As donors increase their investment of ODA in grown from approximately 5% in 2014–2016
blended finance, they will need to be clear on the (DFI Working Group, 2017) to approximately
value for money of blended finance in different 9% in 2017 (DFI Working Group, 2018). This
countries and sectors. It will be crucial to ensure will be the subject of future research once ODA
that the resulting change in ODA allocation is reform by the OECD DAC has been agreed
not at the expense of those countries who need and fully implemented by all MDBs and DFIs –
ODA the most. something that is likely to take a few years yet.

Figure 1 Comparison of ODA and blended finance (per capita and per person living in poverty)

LIC LMIC UMIC

70 600

60 500
$ per person living in poverty

50
400
$ per capita

40
300
30
200
20

10 100

0 0
Mobilised private ODA per capita Mobilised private ODA per person
finance per capita finance per person living in poverty
living in poverty

Note: Poverty measured using the World Bank’s $1.90 per day at 2011 purchasing-power parity.
Source: Mobilised private finance data (2012–2015) (Benn et al., 2017); ODA (2015) data from OECD CRS; poverty data
(2013) from World Bank PovcalNet (World Bank, 2017a).

50 This reflects the fact that while (1) the total population in MICs is larger than that of LICs, and (2) the total number of
poor people living in MICs is larger than LICs, the share of the total population of LICs living in poverty is much higher
than in MICs.

30
5 The potential of
blended finance: time for
a reality check
Key findings
•• Estimates of private finance mobilised through blended finance by MDBs, DFIs and donors in
LICs and MICs range from $3.3 billion (DFI Working Group, 2018) to $27 billion annually
(Benn et al., 2017). This rises to $71.1 billion if we use total direct and indirect mobilisation
reported by MDBs (World Bank, 2017b).
•• This is based on investment of ODF of around $2.2 billion to $121 billion annually. The large
range in the mobilisation and commitment figures reflects the different scopes of the surveys.
•• Overall, global leverage ratios for MDBs and DFIs as a whole are very low, ranging from
1:0.14 to 1:1.3.51 This rises to 1:1.5 if we use total direct and indirect mobilisation reported
by MDBs.
•• ODI estimates that $1 of public investment mobilises just $0.37 of private investment in LICs,
$1.06 in LMICs and $0.65 in UMICs. Leverage ratios are lowest for infrastructure in LICs
and MICs.
•• Despite practitioners having the knowledge, we do not know how much it costs to mobilise
private finance or, consequently, the value for money of blended finance, as levels of MDB and
DFI subsidy are not reported. Ideally, leverage ratios would be calculated based on the grant
equivalent of the ODF invested. Coupled with metrics on impact, this would give policy-
makers a better understanding of the potential and value of blended finance.
•• Low financial leverage ratios raise three issues: (1) policy-makers need to recalibrate their
expectations of the potential of blended finance, (2) MDBs and DFIs need to get better at
mobilising private finance; fundamental changes are required, and (3) there is a crucial need to
understand the development additionality of blended-finance investment to assess its value for
money compared with other forms of financing to achieve stated development objectives.

‘Billions to trillions’ has become synonymous one of the best solutions to turn billions of ODA
with blended finance and the mobilisation of aid money into trillions of investment capital for
private finance for development, despite its the SDGs’.
broader original context (World Bank, 2015). Our research suggests that while there is
Indeed, according to the Blended Finance an appreciation of the limitations of blended
Taskforce (2018b: 13), ‘blending, done well, is finance at the operational level within the MDBs

51 Calculated as total private finance mobilised against ODF invested, expressed as a ratio of $1 of ODF invested. These
ratios have been calculated using the face value of the ODF investment.

31
and DFIs, this appreciation is more limited in knowledge transfer and demonstration effects.
the wider financing-for-development policy These ‘catalytic’ effects are challenging to
discourse. Experience to date suggests a reality measure in a consistent manner and are not yet
check is required to calibrate the policy debate reported by MDBs and DFIs, but they are likely
and temper expectations and bridge the current to be particularly important for LICs.
disconnection between policy rhetoric and the
operational reality: ‘billions to billions’ might be 5.1.1 OECD mobilisation surveys
a more plausible goal. The OECD has conducted multiple surveys
of DAC donors and multilateral institutions
5.1 How much private finance is to measure the total private finance mobilised
being mobilised? by the investment of ODF in blended finance.
The most recent 2016 survey estimates that
As discussed in section 2, the lack of definitional a total of $81.1 billion was mobilised in
and methodological harmonisation at the global 2012–2015, with $27 billion mobilised in
level makes it difficult to estimate the scale 2015 (Benn et al., 2017).55
of blended finance. Most attempts focus on The measurement methodology underpinning
the amount of private finance mobilised, with these surveys is a work in progress. The 2016
little attention paid to how much ODF has survey covered five instruments: guarantees,
been invested and its cost. Three recent official syndicated loans, shares in CIVs (such as private
surveys of note have been undertaken by the equity funds), direct investments in companies
OECD, a group of MDBs and the DFI Working and credit lines. The OECD is developing its
Group on blended concessional finance.52 These instrument methodology to include project
surveys estimate the range of private finance finance (public–private partnerships), standard
mobilised in LICs and MICs at anywhere loans and grants in private-sector co-financing
from $3.3 billion53 to $27 billion per annum, schemes. Once these methodologies are
increasing to $71.1 billion per annum if indirect approved, the survey coverage will be deemed
MDB mobilisation is included.54 This is modest comprehensive.
compared with ODA flows of $146.6 billion The OECD’s 2016 survey is based on the
in 2017 (OECD, 2018g) and positively tiny responses of 71 DAC donors and multilateral
compared with the estimated SDG financing gap institutions. It underestimates the amount
of $2.5 trillion per annum. of private finance mobilised, due to its
At this juncture it is important to note that instrument coverage and the fact that some
we are focusing on private finance mobilised important non-DAC institutions, such as the
and reported by MDBs, DFIs and donors. This Islamic Development Bank (IsDB), did not
is distinct from the wider ‘catalytic’ impact of respond. The IsDB mobilised a total of $4
MDB and DFI investment created, for example, billion in private finance in LICs and MICs in
through technical assistance, policy support, 2017 ($67 million in direct mobilisation and

52 Other sources of data on blended finance include the convergence dataset. The World Economic Forum (2016) also
conducted a survey in 2016.

53 Using the strict DFI Working Group definition which focuses on the use of concessional finance. This is at the lower end
of the scale, as it concentrates solely on blended concessional finance, which makes up a relatively small percentage of the
total volume of all MDB and DFI private-finance mobilisation activities.

54 In an effort to understand the potential of blended finance, further data and analysis on the type of private investor being
mobilised would be very useful. Unfortunately, this data is not available.

55 In February 2019 the OECD released high-level mobilisation data for the period 2012–2017. It reports that
$152.1 billion was mobilised from the private sector in 2012–2017. See OECD website (www.oecd.org/dac/financing-
sustainable-development/development-finance-standards/mobilisation.htm).

32
$3.9 billion in indirect mobilisation) (World (World Bank, 2017b).56 It should be noted,
Bank, 2018a). however, that most of this – $33.4 billion
($92.5 billion) – was mobilised in high-income
5.1.2 MDB mobilisation surveys countries (HICs).57 The amount mobilised in
MDBs have also carried out surveys to estimate LICs and MICs totalled $16.5 billion ($71.1
the amount of private finance they have billion) in 2016.
mobilised. As noted previously, MDBs would The second survey published in 2018 included
not consider the mobilisation they report to for the first time the results of 12 European
be blended finance, only that private finance DFIs.58 The report estimates that MDBs and DFIs
mobilised through blended concessional finance. directly mobilised $52 billion ($163.5 billion)
In the surveys, MDBs report total private of private finance in 2017 (World Bank, 2018a).
co-financing mobilised, divided into direct If we again exclude operations in HICs, which
and indirect mobilisation. In the following totalled $32.9 billion ($104.1 billion), they
paragraphs, we cite the reported direct directly mobilised $19.1 billion ($59.4 billion)
mobilisation figure and note the total including in 2017. If we strip out DFIs, the direct private
indirect mobilisation in italics and in brackets finance mobilised by MDBs was $18.1 billion
(for reasons explained in section 2.2). ($55.9 billion).
The first survey published in 2017 estimates As discussed in section 2.2, the MDB survey
that the MDBs directly mobilised $49.9 billion uses a different methodology to the OECD
($163.6 billion) of private finance in 2016 survey, covering all private-sector instruments.

Figure 2 Private finance mobilised by MDBs

OECD survey (2015) MDB joint report data – MDB joint report data –
direct mobilisation (2016) total mobilisation (2016)
25

20

15
$ billions

10

0
IFC MIGA EIB EBRD IDA/IBRD IADB ADB AfDB

Note: World Bank (2017b) data exclude investments in HICs. OECD survey data for EIB are for non-EU recipients only.
Source: Benn et al. (2017); World Bank (2017b).

56 Long-term financing (tenor greater than one year) excludes facilities such as trade finance and working-capital facilities.

57 Of this, $32 billion ($50.2 billion) was mobilised by the EIB in EU countries.

58 The EDFI has adopted the MDB methodology for measuring the mobilisation of private finance, with 12 of its 15
members reporting their results for 2017.

33
It includes the largest MDBs: the AfDB, ADB, The second survey, published in 2018, reported
Asian Infrastructure Investment Bank (AIIB), $1.2 billion in concessional finance invested
European Bank for Reconstruction and alongside $3.9 billion of MDB and DFI own-
Development (EBRD), EIB, Inter-American account finance in 2017, mobilising around
Development Bank (IADB), IsDB and the World $3.3 billion in private finance (DFI Working
Bank Group. Disaggregated data underpinning Group, 2018). The report is more useful for
the report are not publicly available, but analysis purposes than the 2017 report, as it breaks
the annex to the MDB survey does contain down the financing by volume, instrument, sector
aggregated data by institution, by country and geography, although the disaggregated data
income classification and by infrastructure sector. underpinning the report are not publicly available.
By way of illustration, Figure 2 plots the 2015
data of the OECD survey against the MDB 5.2 Low leverage ratios and the
survey. One can see material differences in the need to temper expectations
reported amounts of mobilised private finance.
However, the two surveys cannot be directly Leverage ratios are often at the core of donor
compared, as explained in section 2.2. arguments for increasing the investment of ODA
in blended finance. They are often used by policy-
5.1.3 DFI Working Group surveys makers as a proxy indicator of the success of
The DFI Working Group on Blended blended-finance investment (the higher the ratio,
Concessional Finance, consisting of eight MDBs the more development finance there is to invest
and the EDFI, has conducted two surveys using in the SDGs).60 The importance of leverage ratios
the narrower MDB and DFI definition of blended should not be overstated, however, as there are a
finance, which focuses on blended concessional number of interpretive issues to be aware of, so
finance. It should be noted that this is a very caution should be exercised when using them.
small percentage of all MDB and DFI private
finance-mobilisation operations. 1. Leverage ratios are simple arithmetic ratios
The first survey, published in 2017, found that do not imply causality, so financial
that $1.5 billion in concessional finance was leverage should not be interpreted as a proxy
invested alongside $5.2 billion of MDB and DFI measure of financial additionality. Indeed,
own-account finance in 2014–2016, mobilising assessment of financial additionality is
$8.5 billion in private investment, which complex and challenging, not least because of
translates into around $2.8 billion59 per year the lack of counterfactual, and is beyond the
(DFI Working Group, 2017). The report is highly scope of this report.61 High leverage ratios are
aggregated, reporting only total volumes of ODF not, therefore, automatically synonymous with
invested, split by concessional and own-account high levels of additional financing. Indeed, they
resources and total private finance mobilised. may signal little additional capital, as high
It also breaks down the ODF invested by private-finance multiples may suggest that the
instrument. The disaggregated data underpinning private investment would have materialised
these findings are not publicly available. anyway, without public support.

59 Calculated as total project value of $15.2 billion, less $6.7 billion of concessional finance and DFI own-account resources,
divided by three.

60 Some argue that blended finance does not necessarily increase the size of the development financing pie. For example,
Carter (2015) argues that in the absence of binding borrowing constraints on donors, donors could create the same
financing package at the same cost to the taxpayer, so no additional finance has been mobilised.

61 See Pereira (2015) for an overview of the literature on the additionality of using ODA to leverage private investment,
as well as Carter (2018), who examines potential evidence of additionality and concludes that concrete evidence of
additionality is elusive.

34
2. Financial leverage ratios say nothing at all leverage ratios by the Blended Finance Taskforce
about development additionality and nothing and the DFI Working Group on Concessional
about the broader catalytic effects of Blended Finance. Attempts to estimate global
blended finance. leverage ratios have used total MDB and DFI
3. Measurement is highly problematic and a commitments as a measure of public effort
challenge. It is more of an art than a science (in other words, investment of ODF) and are
given the lack of good-quality comparable calculated at highly aggregated levels.
‘input’ and ‘output’ data, the plethora of Table 2 summarises the main sources of data
definitions and the various measurement on mobilisation and shows the range of leverage
methodologies. Thus, it is not possible to ratios calculated at global level.
calculate financial leverage ratios with any The Blended Finance Taskforce (2018b) report
degree of precision. calculates individual MDB leverage ratios, as well
as a leverage ratio for the MDB system, using
Essentially, what we are interested in is the the total value of MDB commitments, which it
multiplication effect of ODF – in other words, estimates at around $207 billion in 2016, and
the ability of ODF to mobilise a larger multiple finds a low leverage ratio of 1:0.8. Using the
of private capital – and how much it costs to Taskforce’s figures, we strip out HIC operations,
mobilise this multiple (the public subsidy). This, adjusting for EIB EU operations,63 and estimate
together with metrics on impact, can help shed MDB system leverage ratios for direct and total
light on the potential and value for money of mobilisation (see Table 2), which range from
blended finance. There are several ways in which 1:0.14 to 1:0.7.
leverage can be defined and measured. The most The Blended Finance Taskforce report
informative metric for policy-makers would be also calculates leverage ratios for individual
a measure of the leverage effect of the public MDB PSWs using the value of dedicated PSW
subsidy (or how much private finance is being commitments only, which it estimates at a total
mobilised per dollar of public subsidy).62 This of $40 billion in LICs and MICs in 2016.64
would involve calculating the grant equivalent of The Taskforce finds that the leverage ratio
the investment as a proxy measure of the subsidy nearly doubles to 1:1.5. As discussed in the
and calculating leverage based on the grant methodology note to this report (Annex 2), this is
equivalent of the ODF invested. Unfortunately, likely to overestimate MDB efforts, as not all of
there are no publicly available data on the their activities target mobilisation, so the report
grant equivalents (subsidies) provided by MDBs and our estimates may slightly underestimate
and DFIs, so we focus instead on the ratio of system leverage ratios. It should also be noted
public investment (as proxied by MDB and DFI that this study only includes MDBs, so excludes
own-account commitments) to private finance any donor efforts on the part of bilateral donors
mobilised. So, for example, if $1 of MDB and or their DFIs.65
DFI investment led to $2 of private capital being Another attempt has been made by the DFI
invested, the leverage ratio would be 1:2. Working Group on Blended Concessional
Finance, which focuses on the leverage effect
5.2.1 Global leverage-ratio estimates of blended third-party or donor concessional
A review of the literature identifies two attempts finance (i.e. the amount of MDB and DFI
to calculate very rough estimates of global own-account and private finance leveraged as

62 For a technical discussion, please see Carter (2018).

63 As a proxy for HICs, 96% of the mobilisation reported in HICs was mobilised in the EU by the EIB.

64 EIB EU operations are excluded.

65 For a sense of scale, the 2016 OECD mobilisation survey estimates that around 36% of the private finance mobilised is
drummed up by bilateral donors and their DFIs (Benn et al., 2017).

35
Table 2 Summary of mobilisation data sources and implied leverage ratio

Source MDB/DFI own-account commitment Private finance mobilised Leverage ratio

(annual average for years (annual average for years (ratio of ODF to private
in brackets) in brackets) finance mobilised)
OECD mobilisation survey Not surveyed $20.2 billion (2012–2015) Not possible to calculate
(2016)
(MDBs and DAC donors)
MDB mobilisation survey Not surveyed $19.1 billion direct mobilisation (2017) Not possible to calculate
(2018) (MDBs excluding HIC operations)
(MDBs and EDFI)
$59.4 billion total mobilisation (2017)
(MDBs excluding HIC operations)
Blended Finance Taskforce: $121 billioni (2016) $17 billion direct mobilisation (2016) 1:0.14
better finance, better world (MDBs excluding EIB EU operations) (MDBs excluding EIB EU operations)
(2018b)
(MDBs only) $121 billioni (2016) $82.3 billion total mobilisation (2016) 1:0.7
(MDBs excluding EIB EU operations) (MDBs excluding EIB EU operations)

$39.6 billion (2016) $15.5 billion direct mobilisation (2016) 1:0.4

(MDB PSWs excluding EIB EU (MDB PSWs excluding EIB EU
operations) operations)

$58.9 billion total mobilisation (2016) 1:1.5

$39.6 billion (2016) (MDB PSWs excluding EIB EU
(MDB PSWs excluding EIB EU operations)
operations)
DFI Working Groupii $2.2 billioniii (2014–2016) $2.83 billion (2014–2016) 1:1.3
(MDBs and DFIs)
$5.5 billion (2017) $3.3 billion (2017) 1: 0.6
ODI dataset $23.1 billion (2013–2016) Collection not possible Not possible to calculate
(9 focus institutions)
ODI estimateiv $21.4 billion (2013–2015) $16 billion (2013–2015) 1:0.75
Notes: (i) We assume 10% of EIB operations are in non-EU countries. EIB works in over 150 non- EU states which receive
around 10% of EIB funding. See EIB website (www.eib.org/en/about/key_figures/index.htm). (ii) We calculate this as the ratio
of public investment (i.e. concessional and own-account resources) to private investment. (iii) Concessional resources and
own account. (iv) Combining our dataset with the OECD mobilisation survey for nine selected institutions.

a result of using the donor and or third-party 5.2.2 ODI disaggregated leverage-ratio
concessional finance). These ratios will be higher estimates
than those reported in Table 2 and what we have To have a more nuanced understanding of the
calculated, as they include MDB and DFI own- potential in different countries and sectors, we
account investment as finance leveraged. For need to move beyond these global estimates and
example, the DFI Working Group reports that understand leverage ratios at a more granular
$1 of concessional donor support mobilises $10 level. We can use our disaggregated blended-
from MDBs and DFIs and the private investor finance commitment database to explore this.
(in other words, a leverage ratio of 1:10) (DFI Caution should, however, be exercised in
Working Group, 2017). For 2018, this ratio falls interpreting our disaggregated leverage ratios,
to 1:6 (DFI Working Group, 2018). as two different data sets have been used to
compute it, with different methodologies for
collecting and analysing data. Furthermore, we
have only looked at a sample of nine MDBs

36
and DFIs. The exercise is presented by way of the annual average for 2013–2015 and limiting
illustration, to stimulate discussion about the the scope to our nine focus MDBs and DFIs,
potential of blended finance and to highlight we find that $64.2 billion of MDB and DFI
the need for better data and more transparency commitments mobilised a total of $47.9 billion
to enable more precise analysis. It is the picture in private finance, resulting in a leverage ratio of
the data paint in the context of the ‘billions to 1:0.75.67 Our overall ratio is roughly similar to
trillions’ agenda that is important, rather than that calculated by the Blended Finance Taskforce.
the necessarily imprecise methodology used. As both our dataset and the OECD
Our data show that commitments invested mobilisation survey are available in
in projects aimed at mobilising private finance disaggregated format,68 we can go one step
by our nine focus MDBs and DFIs totalled an further and tentatively look at leverage ratios
annual average of $21.4 billion in 2013–2015. within specific sectors and in different country
Combining the findings of our collated dataset income categories, as shown in Figures 3 and 4.
and the disaggregated mobilisation numbers of A striking observation is the low financial
the OECD mobilisation survey,66 we can get a leverage ratio of LICs (Figure 3), which
rough estimate of global leverage ratios. Using should not be surprising, as it is more difficult

Figure 3 Leverage ratios by category and sector, nine selected MDBs and DFIs, 2013–2015

1.2
mobilises $x of private investment
$1 of MFB and DFI investment

1.0

0.8

0.6

0.4

0.2

0.0
Overall

LMIC

UMIC

Banking and finance

Productive/extractive
sectors

Infrastructure

Other/multisector

Social sectors
LIC

Overall Income class Sector

Source: Authors’ calculations based on ODI dataset and OECD mobilisation survey.

66 Our commitment data are obtained from data reported publicly in the MDB and DFI project-level datasets.
We have assumed they underpin MDB and DFI estimates and OECD reported estimates of private finance mobilised
by these institutions.

67 We used our dataset to calculate the total commitment of our nine selected institutions and then calculated a simple
annual average per institution. We then used the OECD dataset to calculate the total mobilisation for our nine selected
institutions and then calculated a simple annual average mobilisation per institution. We combined the two to estimate
the overall leverage ratio. We then divided the data in both datasets by income classification, by instrument and by sector
for more disaggregated leverage ratios.

68 We use 2016 OECD mobilisation data (Benn et al., 2017), as they are available at a more disaggregated level. No MDB
and DFI mobilisation data are available. The calculated leverage ratios would probably be higher if that data were
available, as the aggregate mobilisation figures reported by MDBs and DFIs are vastly different to the numbers reported
by the OECD (Figure 2).

37
Figure 4 Sector leverage ratios by country income group, nine selected MDBs and DFIs, 2013–2015

LIC MIC Total

1.0

0.8
$1 of MDB and DFI investment
$x of private investment

0.6

0.4

0.2

0.0
Banking and finance Productive/extractive sectors Infrastructure Social sectors

Source: Authors’ calculations based on ODI dataset and OECD mobilisation survey.

to mobilise private finance in these markets and the lack of secure streams of positive cash
than in more stable ones, due to higher levels flow to repay investors in the social sectors, or
of perceived or actual risk, less conducive the limited opportunity to invest at scale in a way
investment climates, etc. We estimate a leverage that is efficient for structuring blended-finance
ratio of 1:0.37 in LICs, that is, $1 of public transactions. This reinforces the need to think
investment mobilises $0.37 of private finance. carefully about the allocation of ODA to
The ratio triples to 1:1.06 in LMICs, but falls ensure it is properly targeted at eliminating
to 1:0.65 in UMICs. This would suggest that extreme poverty.
blended finance is perhaps more suited to LMICs.
The other interesting finding is that leverage 5.2.3 Policy issues raised by low leverage
ratios appear to be low and fairly consistent ratios
across sectors, with a (surprisingly) slightly As mentioned, data limitations hinder the ability
higher ratio in the social sector.69 to estimate leverage ratios with any degree of
When we look at differences between country precision. Nevertheless, there is a clear emerging
income categories and sectors (Figure 4), we picture and, with it, several policy issues.
see that sector leverage is higher in MICs and First, policy-makers need to rein in their
at least double that of LICs across the board. expectations of the potential of blended finance.
Leverage ratios are lowest in infrastructure in ‘Billions’ would appear to be a more realistic
both LICs and MICs.70 Leverage is very low in goal than ‘trillions’. Our estimates suggest
infrastructure and the social sectors in LICs. that leverage ratios are low, which could
This may reflect their lack of bankable projects allay concerns about the issue of financial

69 Future ODI research will explore blended finance and leverage in the social sectors, as equity issues are likely to be raised,
alongside the question of whether this is just a different way of financing public expenditure on social services (in other
words, buying private provision).

70 We attempted to look at leverage ratios by instrument and country income classification, but this was not possible to
do for several reasons. For example, the DAC category of direct investment in companies includes both equity and debt
investment and several MDBs and DFIs do not report equity investment.

38
additionality. However, these low levels suggest it from the private to the public investor, it
there is a real need to temper the expectations is critical to understand the development
of policy-makers in relation to the potential of additionality and value for money of this
blended finance to shift the financing needle approach. For example, there may well be other
from ‘billions to trillions’ of investment for public policy interventions that could support
development, especially in LICs. It is unrealistic the achievement of the same stated development
to think that blended finance can mobilise the policy objective, which are more effective and
trillions of dollars of private finance needed to transparent than providing a direct subsidy to the
meet the SDGs in MICs and LICs. In the best- private sector.
case scenario, investment in blended finance Third, institutions need to get better at
would mobilise at most $220 billion a year, mobilising private finance. If blended finance
assuming an extreme 100% ODA reallocation is to be scaled up rapidly, these leverage ratios
into blended finance. And, for the purposes of need to increase significantly. Fundamental
illustration, even if 100% of ODA were invested system changes will be required to overhaul the
in mobilising private finance in LICs, it would development finance system (both public and
only generate a maximum of $54 billion a year. private). A recent Blended Finance Taskforce
Second, there is an urgent need to understand (2018a) report attempts to flesh out what this
the development additionality and value for means practically and offers concrete practical
money of blended-finance investment. Low recommendations to overhaul the system.
leverage ratios also raise a different set of In section 7, we argue that blended finance will
potential issues surrounding development need to be better tailored to the needs of LICs if
additionality and value for money. Low leverage the approach is to deliver and highlight the risk
ratios imply a large role for ODF in investment. that the blended-finance agenda in LICs may
For example, our system-average leverage ratio become more expensive amid a poor enabling
of 1:0.75 implies that the public sector (the environment and limited supply of commercial
MDBs and DFIs) has picked up approximately or close-to-commercial investment opportunities.
57% of the cost of the investment. As blended This reinforces the pressing need to increase the
finance does not often mitigate risk, but transfers supply of investment opportunities.

39
6 The current state of
blended finance in LICs
Key findings
•• The share of private finance mobilised by blended finance flowing to LICs is very small.
•• The flow of private finance mobilised through blended finance reflects a broader trend in the
flow of other external private capital flows to LICs and is commensurate with their relative
size in the global economy.
•• MIGA, France and the US are the largest mobilisers of private finance in LICs. Bilateral DFIs
appear to play a more important role in LICs than MDBs and RDBs.
•• The share of total private finance mobilised in LICs has fallen over time, as several large
LICs have transitioned to LMIC status. What’s more, flows are concentrated in the relatively
richer LICs.
•• Infrastructure is the largest sectoral recipient of blended finance in LICs, at almost 50% of
total commitments. More blended finance goes to infrastructure in LICs than in other country
income groups. Very little goes to the social sectors of LICs.
•• Loans are the most prevalent instrument for mobilising private finance in LICs, accounting for
56% of total commitments. These are followed by guarantees, accounting for 22% of total
commitments in LICs. Lines of credit are hardly used at all in LICs, in stark contrast
to UMICs.
•• We see limited use of equity in LICs or MICs, and little variation in the tailoring of
instruments in LICs.
•• Project size is much smaller in LICs than in MICs. The average project size in LICs in 2016 was
$14 million, compared with $32 million and $74 million, respectively, in LMICs and UMICs.

In the analysis that follows, we include private survey and exclude NDBs. Our selected MDBs
finance mobilised by MDBs and DFIs which has and DFIs mobilised 77% of total private finance
been financed 100% by MDB and DFI own- mobilised overall and 85% of total private
account operations. We also focus on a subset finance mobilised in LICs between 2012 and
of the total blended-finance universe, as we do 2015, according to the (2016) OECD survey, so
not cover all MDBs and DFIs in the OECD we cover the main traditional actors.71

71 Note that our focus sample of MDBs and DFIs only includes three DFI members of the EDFI. In total, EDFI members had
31% (€11.5 billion) of their total investment portfolio invested in sub-Saharan Africa at the end of 2017. EDFI members
invested about €2.2 billion in sub-Saharan Africa in 2017 (EDFI, 2018).

40
6.1 How much private finance is LICs in 2017 ($5.3 billion in total). The DFI
being mobilised in LICs? Working Group on Blended Concessional
Finance reported that MDBs and DFIs mobilised
Blended finance is mainly mobilising private $145.1 million in private finance in LICs in 2017
finance in MICs, with very little mobilised using blended concessional finance (DFI Working
in LICs. Analysis of OECD data finds that Group, 2018).
3.6% ($2.9 billion) of the total private finance We should not be surprised by the
mobilised in 2012 to 2015 flowed to LICs, concentration of blended finance in MICs. It
equivalent to $725 million per annum, while reflects a broader trend in the flow of other
LMICs and UMICs accounted for around 40% external private capital flows, such as foreign
each (Benn et al., 2017).72 These findings chime direct investment (FDI), which flows to countries
with the findings of the joint MDB surveys on with more favourable investment climates
the direct mobilisation of private finance. The and is commensurate with the relative size of
first MDB report (World Bank, 2017b) estimated their economies compared with LICs, as seen
that MDBs directly mobilised $1 billion in LICs in Figure 5. Blended finance is better suited to
($5.9 billion in total). The most recent MDB MICs, as the majority of DFIs have traditionally
survey (World Bank, 2018a) estimates that MDBs operated on a commercial basis just below
and the EDFI directly mobilised $1.6 billion in market, with very little DFI investment classified

Figure 5 External financial flows: income-group comparison

LIC LMIC UMIC HIC

100 100

80 80
Share of global total (%)

Share of global total (%)

60 60

40 40

20 20

0 0
Mobilised ODA Remittances FDI inflows Population Population living GDP
private finance in poverty

Source: Mobilised private finance data (2012–2015) from Benn et al. (2017); ODA (2015) data from OECD CRS;
Remittance data (2016) from World Bank (2017c); FDI inflows data (2016) from the United Nations Conference on Trade
and Development (UNCTAD) (FDI inward and outward flows and stock, annual, 1970–2016); population data (2016)
from the World Bank World Development Indicators (SP.POP.TOTL); poverty data (2013) from World Bank (2017a); gross
domestic product (GDP) data from the World Bank World Development Indicators (NY_GDP_MKTP_KD). All data accessed
November 2017.

72 A substantial share of blended finance is invested on a regional basis, rather than in a specific country. This is particularly
the case where investments are made in intermediated investment funds, which on-lend or invest directly in companies.
This type of investment cannot easily be categorised by income classification. In the 2016 OECD survey, 9% of the total
finance mobilised was classified as regional, 45% of it is through intermediated investment funds. As regional and global
funds may also provide financing in LICs, the real share of blended finance is probably slightly higher than the OECD
survey suggests.

41
as ODA (as it did not meet ODA concessionality Figure 7 compares each actor’s share of the
thresholds). As expected, we also see greater total private finance mobilised across all actors
dependence on ODA in the external financing and their share of finance mobilised in LICs
mix in LICs (Figure 5). only. MIGA, France, the Private Infrastructure
Development Group (PIDG) and Norway stand
6.2 Who is mobilising private out as playing an outsized role in LICs compared
finance in LICs? with their overall importance in the blended-
finance landscape. For example, while France
Based on OECD data,73 the largest actors accounts for only 3% of total private finance
mobilising private finance generally are also, mobilised, it accounts for 20% of the total
to a large extent, the largest actors mobilising finance mobilised in LICs. The opposite is true
private finance in LICs. In most cases, this owes for the EIB, EBRD, UK, IADB and ADB, which
more to the sheer size of their blended-finance are large blended finance actors. For some, such
operations than a particular focus on LICs. For as the EIB, EBRD, ADB and IADB, this is to be
example, the US is the third-largest mobiliser of expected, because of their geographical focus.
finance in LICs, but only 3% of its total blended The UK seems to play a smaller role in LICs than
finance goes to LICs. France, in contrast, is its overall level of importance in blended finance
only the ninth-largest blended-finance actor in would imply.
overall terms, but 20% of its blended finance is According to the 2016 OECD survey (Benn
mobilised in LICs, making it the second-largest et al., 2017), MDBs and RDBs account for
actor in those countries (Figure 6). around two thirds of all private finance mobilised

Figure 6 Private finance mobilised through blended finance by institution, 2012–2015

Low-income Non-low-income Share of blended finance to LICs

18 45%
16 40%

Share of the institutions' private

14 35%
Private finance mobilised

finance mobilised in LICs

12 30%
10 25%
$ billions

8 20%
6 15%
4 10%
2 5%
0 0%
s

um
s

Po y
d
GA

d
UK
IFC

EIB

De n
the y

itz l
A

k
ted e

DB

Sw uga
ate

an
nd
Ne wa

an

lan
e

ar
iUS anc

PID

ID

ed
MI

lgi
Af

nm

rm
rla

l
D/
St

Fin

er
rt
No

Sw
Un Fr

Be
IBR

Ge

Note: Institutions with no blended finance in LICs are not included in these figures (EBRD, IADB, ADB, the Credit Guarantee
and Investment Facility (CGIF), the Development Bank of Latin America (CAF), Japan, Spain, Canada, the Global Energy
Efficiency and Renewable Energy Fund (GEEREF), Luxembourg, Austria). US mobilisation is 86% from OPIC and 14%
from the United States Agency for International Development (USAID).. French mobilisation is 60% is from AFD and 40% is
from Proparco. EIB excludes activities in EU and European Free Trade Area (EFTA) countries.
Source: OECD semi-disaggregated dataset (OECD, 2017).

73 The agency-level data in the OECD dataset (OECD, 2017) do not disaggregate by recipient country.

42
through blended finance. This finance is usually are investment companies and/or funds, which
generated by their private-sector arms, with are not regulated in the same way as banks and
bilateral institutions mobilising the remaining they are not financed by the capital markets.
third. Interestingly, when we look at which This means they can have a relatively higher risk
institutions are mobilising private finance in appetite than some MDBs and RDBs. France
LICs, bilateral DFIs appear to be playing a more and the US mobilise more private finance in LICs
important role, as they account for a combined than the IFC. Also, France, the US, Norway and
50% of private finance mobilised. DFIs are the Netherlands mobilise more blended finance
different institutions to MDBs and RDBs; many than the ADB, which suggests, given its capital,

Figure 7 Share of total private finance mobilised, overall and in LICs, 2012–2015
25% LICs Total

20%

15%

10%

5%

0%

Luxembourg
Portugal

Austria
AfDB
IBRD/IDA

CGIF

CAF

Finland
ADB

Denmark
US

EBRD
UK

IADB Group

Belgium

Spain
France

Norway
PIDG
Germany

Switzerland

Japan
Sweden
EIB
MIGA
IFC

Canada
Netherlands

GEEREF

Source: Authors’ calculations based on the OECD semi-disaggregated dataset (OECD, 2017).

Figure 8 Private finance mobilised in LICs, 2012–2015

Low-income Share of blended finance to LICs
0.7 45%

40%
0.6
$ billions

Share of the institutions' private

35%
finance mobilised in LICs

0.5
30%
0.4 25%

0.3 20%

15%
0.2
10%
0.1
5%
0.0 0%
the y

Sw gal
A

k
ce

um

Ge d

y
s

DB
GA

d
US

UK
IFC

EIB
IBR G

en
a

ar

an
nd

lan
ID

lan
PID

rw
an

ed

rtu
Af
MI

nm

lgi
D/

rm
rla

Fin

er
No
Fr

Sw

Be

Po
De

itz
Ne

Source: Authors’ calculations based on the OECD semi-disaggregated dataset (OECD, 2017).

43
geographical focus74 and mandate, that the AfDB selection process and Annex 2 outlines the
should play a larger role than these bilateral methodology we used to identify blended-finance
actors (Figure 8). commitments. Data were available for 10 of
We would also expect the volume of private our 11 selected institutions.76 Together, 7.1%
finance mobilised to increase in the coming of their total own-account commitment to
years as several new initiatives bed down, most mobilising private finance was directed to LICs
notably the new $2.5 billion IDA18 IFC-MIGA between 2013 and 2016 – amounting to an
PSW and the new IFC general capital increase annual average total institutional commitment
of $5.5 billion. The PSW will be a key tool of $1.6 billion. A total of 313 blended-finance
supporting the IFC’s new 3.0 strategy, along projects in LICs were identified in the 2013–2016
with its general capital increase and MIGA’s period. The analysis that follows uses the ODI
2018–2020 strategy to scale up efforts in the dataset on institutional commitments to analyse
poorest countries.75 blended finance in LICs in more detail.
Based on this OECD overview of the largest Figure 9 shows the average annual
blended-finance actors in LICs, we selected commitments by MDBs and DFIs to mobilise
11 institutions for a more detailed drill-down. private finance in LICs. Compared with the
Combined, these institutions mobilised 77% relative mobilisation rankings of the OECD
of total private finance mobilised overall and mobilisation survey (Figure 8), the IFC makes the
85% of total private finance mobilised in LICs largest annual commitment to mobilise private
between 2012 and 2015, according to the (2016) finance, while OPIC, MIGA and the French
OECD survey. Annex 1 contains our institutional agencies (AFD and Proparco) made significantly

Figure 9 Average annual MDB and DFI commitments to mobilise private finance in LICs, 2013–2016
800

700

600
Commitments, $ millions

500

400

300

200

100

0
IFC OPIC MIGA IDA EIB Norfund CDC AFD Proparco ADB

Note: Average for years with available data. For ADB, AFD and Proparco, data are only available for 2016. ADB value is
$5 million. Data were not available for AfDB.
Source: Authors’ calculations based on the ODI dataset.

74 Of the 34 countries classified as LICs as of July 2018, 27 are in sub-Saharan Africa.

75 For example, it is envisaged that by 2030 15% to 20% of the IFC’s commitments will be in LIC-IDA and IDA fragile and
conflict-affected countries (IFC, 2018).

76 Data were not available for the AfDB.

44
lower annual commitments. At first sight, this during 2013 to 2017, suggesting that they were
suggests that several bilateral DFIs may have at the higher end of the LIC income spectrum.
higher leverage ratios in LICs than the IFC. Senegal, in contrast, is the only country to have
been reclassified downwards from LMIC to LIC
6.3 In which LICs does blended during the period (in July 2016). If we look at
finance mobilise private finance? total flows to countries classified as LICs today,
Senegal would be the largest destination country
Between 2013 and 2017, according to our ($0.8 billion). However, it does not make the list
dataset, 73% of institutional commitments to in Figure 10, because most of the commitments
blended finance in LICs went to sub-Saharan going to it were made when it was still an LMIC.
Africa, with the share increasing over time as Because the largest destination countries
non-sub-Saharan African LICs transitioned to transitioned to LMIC status, the share of total
LMIC status and the remaining LICs became MDB and DFI commitments to blended finance
increasingly concentrated in Africa. As of July going to LICs decreased from 2013 to 2016.
2018, our data show that 27 of the world’s 34 Figure 11 shows the share of total institutional
LICs were in sub-Saharan Africa. commitments going to countries classified as LICs
The top 10 LIC blended-finance destination at the time of the commitment compared with
countries accounted for 82% of total blended- the share of the total commitment to countries
finance commitments in 2013 to 2017, classified as LICs at the start of the period.77
according to our calculations. Kenya, before The concentration of destination countries
it was reclassified as an LMIC in 2015, was varies by MDB and DFI, while membership
by far the largest recipient of blended finance affects the financing destinations of the RDBs.
among LICs (Figure 10). Indeed, several of the While the IFC has activities in 23 LICs, for
main destination countries (Kenya, Myanmar, example, the ADB only has one (Cambodia).
Bangladesh and Cambodia) were reclassified On average, the institutions have activities in

Figure 10 MDB and DFI commitments to mobilise private investment in LICs by destination, 2013–2017

1.6

1.4

1.2
Commitments, $ billion

1.0

0.8

0.6

0.4

0.2

0.0
.
*

h*

da

ia

ue

nia

ts
ep
ya

ar

dia

ine
iop

ien
es

an

biq
nm

.R

a
n

bo

nz

Gu
lad
Ke

cip
Ug

em
Et
a

Ta
My

za
ng

re
Ca

,D
Mo
Ba

er
o

oth
ng
Co

20

Note: Measures commitments by MDBs and DFIs to countries only for the years they were classified as LICs. *Kenya,
Myanmar, Bangladesh and Cambodia were reclassified as LMICs during the period.
Source: Authors’ calculations based on the ODI dataset.

77 Note that this is not how income classification is handled in the OECD mobilisation survey. See Annex 2 for more information.

45
Figure 11 MDB and DFI commitments to mobilising private finance in LICs (selected MDBs and DFIs),
2013–2016
Countries that were LICs in 2013 LICs at the time of commitment
14%
Share of total blended finance commitments

12%

10%

8%

6%

4%

2%

0%
2013 2014 2015 2016

Source: Authors’ calculations based on the ODI dataset.

Table 3 Largest blended-finance recipient countries by MDB and DFI

AFD ADB CDC EIB IDA IFC MIGA OPIC Proparco

Tanzania Cambodia Sierra Kenya Kenya Ethiopia Bangladesh Kenya Benin
Leone
Uganda Malawi Senegal Uganda Myanmar Congo, Myanmar Burkina
Dem. Rep. Faso
Madagascar Bangladesh Malawi Sierra Bangladesh Kenya Cambodia Tanzania
Leone
Source: ODI Dataset, 2013–2017

nine LICs, but there is a large degree of overlap. 6.4 In which LIC sectors is
Kenya features as a top three destination country blended finance mobilising private
for four institutions, while Bangladesh registers
in the top three for three institutions. Table 3 lists
investment?
the top three blended-finance destinations for The sectoral distribution of MDB and DFI
each of our selected institutions. commitments to mobilise private finance in
LICs largely reflects the trend at global level,
with commitments largely concentrated in
infrastructure, banking and finance,78 and
the productive sectors. The remaining sectors
account for a very small share. Very little
MDB and DFI blended finance goes to the

78 As previously noted, much of the investment that is categorised as flowing to the banking and finance sector is usually
lent to local financial institutions, which then lend on to local end-borrowers (mainly SMEs) in a range of sectors, so the
sectoral split is not captured.

46
Figure 12 Commitments to mobilise private finance by sector (selected MDBs and DFIs), 2013–2017

LIC LMIC UMIC

60%

50%
Share of total commitments

40%
by income class

30%

20%

10%

0%
Banking and Extractive sectors Infrastructure Other/Multisector Productive Social sectors Unknown
finance sectors

Source: Authors’ calculations based on the ODI dataset.

social sectors, possibly due to the economic at the IFC (Figure 13). Notably, Norfund and the
fundamentals of these sectors, or the fact that two French institutions are the outliers in terms
there is limited opportunity to invest at scale in of sectoral focus – both Proparco and AFD have
the social sectors in a way that is efficient for had a stronger focus on banking and finance than
structuring blended-finance transactions. on infrastructure in LICs.
There are, however, three noteworthy Individual MDBs and DFIs also show variations
differences in the sectoral allocation of in their sectoral commitments to LICs versus
commitments between country income groups. MICs. As Figure 13 suggests, infrastructure is
MDBs and DFIs allocate more capital to mobilise larger in LICs than in MICs for all the MDBs and
private finance in infrastructure in LICs than in DFIs, although the scale of the difference varies;
MICs, where the focus is on the banking and IDA commitments in LICs are three times more
finance sector. Interestingly, commitments to likely to be in infrastructure, while at AFD, the
mobilise private finance in the social sectors are difference is only three percentage points. In the
mainly made in MICs. Only three DFIs made banking and finance sector, the trend is less clear.
commitments in the social sector in LICs between Most MDBs and DFIs are more likely to make
2013 and 2017 (CDC Group, 2%; IFC, 1%; commitments to mobilise finance in this sector
OPIC 5%). in MICs than in LICs, but the opposite holds for
Almost half of the blended-finance AFD and Proparco.
commitments to mobilise private finance in
LICs by our selected MDBs and DFIs are to 6.5 Which instruments are used to
infrastructure, making it by far the largest sector mobilise private finance in LICs?
overall in LICs (Figure 12).
This aggregate picture, however, masks Loans are the most commonly used instrument for
variations among the MDBs and DFIs. While mobilising private finance in all country income
infrastructure is the largest LIC sector for six groups, but they are even more common in LICs.
out of the nine selected MDBs and DFIs with The opposite is true for lines of credit, which are
sufficient data, the share of infrastructure ranges more commonly used in MICs, but hardly used at
from 93% of commitments at the IDA to 32% all in LICs. This is partly due to the composition

47
Figure 13 Blended-finance commitments by sector and MDB and DFI, 2013–2017
Infrastructure Banking and finance Productive sectors Extractive sectors Social sectors Other/Multisector
100%
Share of blended finance commitments (%)

90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
LIC MIC LIC MIC LIC MIC LIC MIC LIC MIC LIC MIC LIC MIC LIC MIC LIC MIC

AFD CDC EIB IDA IFC MIGA OPIC Proparco Norfund

Source: Authors’ calculations based on the ODI dataset.

Note: ADB data not available by sector; AfDB data not available.

Figure 14 Commitments to mobilise private finance by instrument (selected MDBs and DFIs), 2013–2017
LIC LMIC UMIC

60%

50%
Share of blended-finance commitment

40%
by instrument

30%

20%

10%

0%
Equity Fund Guarantee Insurance Line of credit Loan Risk management

Source: Authors’ calculations based on the ODI dataset.

Note: ADB data not available by sector; AfDB data not available

of the institutions studied; lines of credit are only use of equity in all income categories; aside
used by the EIB,79 for example, and it tends to from the IFC, it seems to be the preserve of the
use them more often in MICs. We also see limited bilateral DFIs. There is very little variation in the

79 In our dataset, only the EIB used lines of credit. Although France and Norway reported lines of credit to the OECD 2016
survey (Benn et al., 2017), they were not identified as such in the project descriptions in their public databases. The IFC
did not report the use of lines of credit to the OECD survey. The AfDB was excluded, as data were not available, and
ADB data were not included, as a breakdown by instrument was not available.

48
remaining instruments between income groups: that holds across all country income groups,
guarantees, direct equity, funds, insurance and while 65%–70% of IFC commitments are in
risk-management tools are all commonly used, loan form, with almost no variation between
regardless of country income category. country income groups (Figure 15). Norfund
Loans are by far the preferred method of stands out as using a large share of equity
mobilising private finance in LICs; 56% of in all country income groups, albeit slightly
commitments were made using loans (Figure 14) less in LICs. MIGA and the IDA only use one
between 2013 and 2017. The remaining instrument – guarantees – in all of their blended-
MDB and DFI commitments took the form of finance operations.
guarantees (22%), and a very small amount At AFD, CDC Group and the EIB, however,
was committed using direct equity, insurance, different instruments are used for different
lines of credit and risk management (each less income groups – group B (Figure 16). At AFD,
than 10%). As mentioned in section 6.1, funds guarantees are more commonly used in LICs,
are usually (but not always) regional or global, with loans accounting for a larger share of their
meaning they are not classified as investing in portfolios in MICs. Meanwhile, at CDC Group
LICs specifically, though in practice, some of and the EIB, loans are used less commonly in
their investment probably goes to LICs. This MICs; at CDC, loans give way to direct equity
implies that the significance of funds in LICs is investment in richer income groups, while at the
being underestimated in our dataset. EIB, lines of credit are preferred.
As mentioned, any variation in the use of loans For six of the nine MDBs and DFIs for which
and lines of credit between income groups is data are available, there appears to be limited
largely the result of just a few MDBs and DFIs. variation in the instruments used to mobilise
For most of the MDBs and DFIs (which we call private finance across country income groupings
group A), there is limited variation between (Figure 15). At first glance, this may imply a
country income groups. More than 80% of the rather limited ‘toolkit’ and little tailoring of
commitments to mobilise private finance by instruments to country circumstances. In some
OPIC and Proparco are made using loans, and cases, however, this is down to the mandate of

Figure 15 Limited variation in instruments use, by income classification (MDB and DFI group A)

Loan Equity Fund Guarantee Insurance Other

100%
90%
Share of blended-finance commitments

80%
70%
60%

50%
40%

30%
20%
10%

0%
LIC LMIC UMIC Regional LIC LMIC UMIC Regional LIC LMIC UMIC LIC LMIC UMIC Regional
IFC OPIC
OPIC PROPARCO Norfund

Note: 100% of MIGA- and IDA-only blended-finance operations use guarantees, so there are six institutions with
limited variation.
Source: Authors’ calculations based on the ODI dataset.

49
Figure 16 Variation in instrument use by income classification (MDB and DFI group B)

Loan Equity Fund Guarantee Line of credit Other

100%
90%
Share of blended-finance commitments

80%
70%

60%
50%
40%

30%
20%

10%
0%
LIC LMIC UMIC Regional LIC LMIC UMIC Regional LIC LMIC UMIC Regional
AFD CDC EIB

Source: Authors’ calculations based on the ODI dataset

Figure 17 Instruments used for private-sector mobilisation in LICs, by sector

Loan Guarantee Equity Risk management Fund Line of credit Insurance

100%

90%

80%

70%
Share of blended-finance
commitments (%)

60%

50%

40%

30%

20%

10%

0%
Banking and finance Extractive sectors Infrastructure Other/Multisector Productive sectors Social sectors

Source: Authors’ calculations based on the ODI dataset, 2013–2017.

the institution, as in the case of MIGA, which 6.6 Instrument use by sector in LICs
only issues political risk insurance and credit
enhancement. It can also be due to operational Although the use of loans dominates all sectors,
restrictions imposed by regulation and/or there are minor trends worthy of attention.
statute, which may prohibit the use of certain Infrastructure stands out as the sector with the
instruments, as in the case of OPIC, which is not lowest share of loan commitments and the highest
authorised to make equity investments. share of commitments via guarantees (Figure 17).

50
The large share of infrastructural guarantees project size in LICs in 2016 was $14 million,
is, to a large extent, a reflection of the heavy the average size was $32 million in LMICs and
involvement of MIGA and the IDA, both of which $74 million in UMICs. However, these averages
only use guarantees to mobilise private finance. are inflated by a small number of particularly
They are also large institutions, accounting for large projects. The respective median project sizes
almost 20% of total commitments in LICs. At the are just $1 million, $4 million and $37 million.
other end of the spectrum, commitments to the Furthermore, behind the averages, we find large
social sectors see the heaviest use of both loans differences between institutions. While project
and equity. Investments in funds are often not sizes tend to be much smaller in LICs, there are
classified by sector, as the funds themselves are notable differences between average project sizes
usually multi-sector. Nor do we know the ultimate from institution to institution in LICs. While
sector destination of blending classified as going to the AFD average project size in LICs was only
the banking and finance sector. $900,000 in 2016, for example, MIGA projects
in LICs average $95 million.80
6.7 Blended-finance project size Breaking down LIC projects further, we also
in LICs find large variations in project size between
sectors and instruments. Projects in the extractive
A commonly cited challenge to expanding blended and infrastructure sectors are much larger, on
finance in LICs is the high transaction costs average, than other projects, while projects
associated with small project size. Large-ticket in banking and finance, and the social and
investments are thin on the ground in LICs, which productive sectors are smaller than the average
arguably limits investment attractiveness, for LIC project (Figure 18). Risk-management
example, from an institutional-investor perspective. projects, meanwhile, are by far the largest
Our dataset shows that projects are much projects,81 while equity and fund investments
smaller in LICs than MICs. While the average tend to be smaller.

Figure 18 Average project size by sector in LICs

70

60
Average project size, $ million

50

40

30

20

10

0
Banking and finance Extractive sectors Infrastructure Productive sectors Social sectors

Source: Authors’ calculations based on the ODI dataset, 2013–2017.

80 This, however, is based on only two projects.

81 This is based on only seven projects.

51
7 The need for a more
tailored approach for LICs
Key findings
•• There is limited use of concessional finance to blend in LICs, compared with UMICs and HICs.
•• The potential of blended finance in LICs is constrained by several factors, most notably poor
investment climates, a lack of investable opportunities, limited use of subordinate instruments
and the limited risk appetites of MDBs and DFIs.
•• Private commercial finance will not flow freely to countries, especially LICs where the local
investment climate is challenging, where markets are not functioning and where the risk-
adjusted rate of return is uncompetitive. Blended finance can tip the balance, but it will not
work if the fundamentals are not in place. The emphasis on blending should not eclipse the
need for grant financing to strengthen local investment climates.
•• There is a scarcity of private investors willing to take on early-stage and pioneer risks in LICs, but
also limited use of subordinate instruments by MDBs and DFIs, suggesting that blended finance
may not be meeting the risk-mitigation needs of the private sector in challenging markets.
•• Conservative MDB and DFI financing models and the returns required on blended
concessional finance are constraining risk appetite and the ability to engage in LICs. Donors
should revisit the rates of return they require on concessional resources and the ‘hurdle rates’
of bilateral DFIs.
•• MDBs should develop a more uniform gauge of capital adequacy that elucidates the scope for
risk. They should also incorporate callable share capital into their capital-adequacy models.
•• Since 2014, global private investment in infrastructure has been declining and is now
virtually negligible in LICs. Blended finance is not currently crowding in private investment
in infrastructure in LICs. There is a need for more blended concessional finance for project
preparation, early-stage project financing and to develop more innovative risk-sharing
instruments.
•• MDBs and DFIs need to adopt a tailored approach to blended finance in LICs.
•• This will be crucial to ensure that increased investment of ODA in blended finance delivers for
LICs.

need to help pioneer and create markets and/

7.1 A limited use of concessional or overcome the most pervasive market failures.
finance to blend in LICs However, data available to date suggest a much
more concentrated use of concessional finance for
As explained in section 4.1 we would, in theory, blending in MICs than in LICs. The DFI Working
expect to see concessional finance playing Group on Blended Concessional Finance reports
a particularly important role in blending to that 21% of the concessional finance used to
mobilise private finance in LICs, given the acute blend in 2017 ($243.2 million) was used to

52
mobilise finance in LICs, compared with 51% 7.2 Poor investment climate and a
($592.5 million) in LMICs, 22% ($252.4 lack of investable opportunities
million) in UMICs and 6% ($70.2 million) in
HICs82 (DFI Working Group, 2018).
in LICs
This is not to say that MICs do not have a Much has been written on the impediments
valid claim on the use of concessional finance to private investment in developing countries,
for blending; indeed, the official financing mix especially LICs, but this is not the focus of this
and the terms thereof evolve along the income report. Put simply, it is far easier to do business
spectrum. We would expect to see blended in more ‘developed’ developing countries, where
concessional finance feature strongly in LMICs the institutional, legal, regulatory and policy
struggling to access finance on reasonable terms frameworks are stronger, the local capital markets
and where there is a need to pioneer and create are more developed and the risk-adjusted rate of
markets. However, we are questioning whether return is more competitive, than in LICs, where
more blended concessional finance should be the contrary is often true. Private commercial
used in LICs, for example, rather than in UMICs finance will not flow freely to countries, especially
and HICs, where near market-term finance LICs, where the local investment climate is
(funded by MDB and DFI own-account resources challenging, where markets are not functioning
and not concessional resources) is much more and where the risk-adjusted rate of return
likely to be appropriate and only small amounts is uncompetitive. These weak fundamentals
of subsidy might be required to bring the risk- adversely affect the ability of blended finance to
adjusted rate of return in line with the market. mobilise private finance in LICs.
As donors seek to scale up their investment The importance of the investment climate
of ODA in blended finance, it is necessary to and these fundamentals can be seen in the
understand what factors are driving this more relationship between private finance mobilised
limited use of concessional blended finance in and country credit ratings. In total, we calculated
LICs. Our research identifies three key drivers, 96.3% of the total private finance mobilised
which are interacting with each other: a weak between 2012 and 2015 flowed to countries
enabling environment and a lack of investable with a credit rating (Figure 19).83 A country’s
opportunities in LICs; the limited use of credit rating appears to be an important factor
subordinate instruments by MDBs and DFIs; influencing the destination of private finance
and the limited risk appetites of MDBs and mobilised through blended finance. This makes
DFIs. We discuss each of these drivers in turn. sense, as private commercial finance is unlikely
To illustrate how blended finance in its current to flow to sub-investment-grade investment
form is not successfully crowding in private opportunities. To illustrate this point, the average
investment in LICs, we discuss the collapse in credit rating for many African countries and the
private investment in LIC infrastructure in LICs. private sector84 is sub-investment grade; private
capital, especially that of institutional investors,
is unlikely to flow at scale to such high-risk
markets. As 27 of the 34 countries classified as
LICs as of July 2018 are in sub-Saharan Africa,

82 The DFI working group notes that the concessional finance used to blend in HICs was mostly for infrastructure projects
and climate projects.

83 We conducted further tests on correlation, looking at fragility status, Doing Business index score, least developed country
(LDC) status and natural-resource dependence. None of these showed any correlation. We also looked at the correlation
to financial depth and the Human Assets Index (www.ferdi.fr/en/indicator/human-assets-index), noting slight correlations,
but these indicators have a positive correlation with GDP per capita. See Annex 3 for scatter plots on financial depth and
the Human Assets Index.

84 Based on the convention that private-sector borrowers are rated one to three notches below the sovereign rating.

53
this may explain why so little private finance finance in MICs. It is more suited to markets
is mobilised in LICs. Only one country in sub- where it can bring the risk-adjusted rate of return
Saharan Africa is investment grade (Botswana, in line with the market. It can tip the scale where
rated A). there are investments with close-to-commercial
This should encourage policy-makers and risk-adjusted rates of return, but it will not work
donors to think more about the sequencing of in markets where the fundamental economics are
support in LICs to mobilise private finance, by not right, and the investment is not financially
placing more emphasis on strengthening local feasible. We estimate an overall leverage ratio of
investment climates85 and adopting a more 1:0.37 in LICs, which implies that that blended
holistic, coordinated and long-term approach to finance is actually picking up the bulk (around
help build markets rather than ad hoc blended- 73%) of the total cost of investment. This reflects
finance investment. The current emphasis on the lack of commercial or close-to-commercial
blending (at investment level) should not eclipse investment opportunities in LICs.
the need for grants to strengthen the investment Furthermore, the task of mobilising private
climates of LICs, to support country-led finance, especially in LICs, is made even more
programmes of policy reform, local capital- difficult by the current bleak external financing
market development and capacity-building. That environment. For example, the OECD reports
is a long game and there will be no quick wins. that, in 2016–2017, FDI in developing countries
Weak fundamentals adversely affect the fell by 30% to $750 billion and that project
risk-adjusted rate of return for investors and, finance fell by 30% in the first quarter of 2018
hence, the availability of investable commercial (OECD, 2018h).
opportunities. As previously noted, we should In this context, and as MDBs and DFIs are
not be surprised by the fact that blended finance tasked with mobilising more private finance in
(broadly defined) is mainly mobilising private LICs, there is a risk that MDBs and DFIs will

Figure 19 Mobilised private finance flows to countries with a credit rating

A-AAA B-BBB CCC No rating/default

4.0

3.5
Private finance mobilised, 2012–2015

3.0

2.5
$ millions

2.0

1.5

1.0

0.5

0.0
0 2,000 4,000 6,000 8,000 1,0000 12,000 14,000
GNI per capita

Source: Benn et al., 2017; Standard & Poor’s (S&P) Global Ratings, as of October 2017; gross national income (GNI) per
capita data (2016 or most recent) from World Bank World Development Indicators (NY.GNP.PCAP.CD).

85 Local investment climates also depend on other forms of public investment, such as health and education, which affect
the availability of human capital.

54
chase fewer and fewer projects and invest in more insurance and risk-management tools are used
marginal investment opportunities with higher just as commonly, regardless of the income
costs to the public purse (in other words, a higher status of the destination country. For six of the
subsidy to attract private investors). As overall nine institutions with data available, the use of
risk is not mitigated in many blended-finance instruments was fairly consistent across country
structures, but rather transferred from the private income groups.
to the public sector, there will be a point at which At first glance, and in the absence of more
the provision of such large subsidies to the private detailed data to understand the capital structure
sector does not represent value for money. of the investments financed, these findings may
suggest that the blended-finance approach
7.3 Limited tailoring of the employed is not specifically tailored to meeting
blended-finance toolkit in LICs the risk-mitigation needs of private-sector
investors in LICs. This may limit the potential
The rationale underpinning blended finance of blended finance to mobilise private finance in
suggests that the level of subsidy funded by ODF LICs. This issue suggests that further exploration
will rise, the higher the level of market failure, and discussion is warranted about whether there
and that this subsidy will fall over time as markets should be greater use of subordinate instruments
develop. This suggests that a significant role will and grants to mobilise private finance in LICs
be played by blended finance, especially blended and, considering the small ticket size of projects
concessional finance, in helping to pioneer and in LICs, whether blended-finance solutions can
create new markets, foster innovation and invest be better constructed for smaller investments.
at the earliest stages of projects, when risk levels
are at their highest and when private investors 7.4 Limited blended private
need a greater risk mitigation. investment in infrastructure in LICs
Given the scarcity of private investors willing
to take on these risks, we would expect to see In the international discourse, blended finance
MDBs and DFIs using blended concessional has been assigned a particularly critical role
finance to take on these early-stage or pioneer in mobilising private commercial investment
risks using instruments such as subordinated to finance the infrastructure gap in developing
debt, equity, risk-sharing facilities, guarantees countries, estimated at $1.5 trillion per annum
and grants. Data from the most recent DFI by the G20. However, recent ODI research
Working Group on Blended Concessional concludes that while there have been successful
Finance, however, shows these instruments innovative initiatives by MDBs and DFIs to
playing a lesser role and a focus on the use of mobilise private infrastructure investment
senior debt86 (44% of concessional commitments including the use of blended finance, this
by volume and 48% of concessional project has fallen sharply from an annual average
count) in 2017 (DFI Working Group, 2018). of $37 billion between 2008 to 2014 to just
MDBs and DFIs hardly use grants at all. Data $13 billion between 2015 and 2017 and is not
show that grants accounted for only 1% of mobilising at anything like the scale required
concessional commitment by volume and 1% of (Tyson, 2018).
concessional project count in 2017 (ibid.). The research finds that although LICs have the
Our analysis (section 6.5) also finds that debt greatest need for infrastructure investment, only
finance is the most common instrument used to 2% of total private infrastructure investment
mobilise private finance and that its use is more went to LICs between 2008 and 2017. Although
prevalent in LICs. The use of other instruments in line with their relative share of global GDP,
shows very little variation between country this is low. The research also highlights the
income groups – guarantees, direct equity, funds, collapse in private investment in infrastructure

86 In the absence of more data, senior debt investment suggests that MDBs and DFIs are likely to be investing pari passu (i.e.
same ranking as private lenders in the capital structure with equal rights of payment).

55
in LICs from just under $3 billion in 2012 to expectations of the potential of blended finance
$300 million in 2016, which blended-finance to fund infrastructure investment, especially
efforts have not been able to arrest. Illustrative in LICs. It is unrealistic to expect institutional
of the constraints on mobilising private investors to risk trillions in markets where the
infrastructure investment in LICs is the case average credit rating is sub-investment grade.
of the risk-mitigation facility of the IDA18 As recent ODI research concludes, at best MDBs
PSW. A $1 billion allocation was made to the and DFIs can
risk-mitigation facility under the IDA18 PSW,
designed to mobilise private investment in nudge markets in the right direction,
infrastructure in the poorest and most fragile of while the true step-change in investment
countries, yet no commitments were made during patterns will only happen as a result of
the first 15 months of the PSW’s operation. The deeper forces over which MDBs have
mid-term review of the IDA PSW forecasts that limited influence. ‘Billions to trillions’ is
the full $1 billion allocation is unlikely to be a catchy phrase, but when it comes to
fully committed by the end of the IDA18 period the ability of MDBs to directly crowd
(World Bank, 2018c). institutional investors into [emerging
This raises questions about whether blended and developing country] infrastructure,
finance, in its current form, can crowd in ‘billions to tens of billions’ is more
private infrastructure investment and at the realistic. (Humphrey, 2018a: 7)
scale required. The recent DFI report on
blended concessional finance shows senior debt As many infrastructure projects in LICs lack
being predominantly used to mobilise private secure streams of positive cash flow, there will
infrastructure investment, which accounts for still be a critical role for standalone international
72% of blended concessional investment by and domestic public investment in infrastructure
MDBs and DFIs. This, in turn, suggests that the for many years to come (ibid.).
MDBs and DFIs are investing pari passu with
the private sector. Again, we see very limited use 7.5 The limited headroom and risk
of subordinate instruments such as subordinate appetites of MDBs, DFIs and donors
debt (12%), equity (9%) and risk-sharing (7%),
to mobilise private infrastructure investment The ability of MDBs and DFIs to engage in
(DFI Working Group, 2018). Considering the and mobilise more private finance in LICs will
constraints on the mobilisation of private finance depend, in part, on their risk appetite, which is
for infrastructure investment, there is a clear constrained by the need to maintain a AAA credit
need to use more blended concessional finance to rating for the MDBs and for DFIs the need to
finance project preparation, early-stage project remain profitable and financially sustainable.87
financing and to develop more innovative risk- As MDBs and DFIs are increasingly tasked
sharing instruments to help mitigate political with investing more of their own-account
and macroeconomic risk where there is a valid resources in riskier sectors and countries, they
case for subsidising private investment in will need to be empowered to take on more risk.
infrastructure. The answer so far has been for donors and third
Even with more innovative use of blended parties to provide concessional capital, which
concessional finance at the earliest stages the MDBs and DFIs use to blend with their
to develop a pipeline of bankable projects, own-account resources (blended concessional
policy-makers will still need to recalibrate their finance), enabling them to shift a part of the

87 The major MDBs’ financing model is a powerful one. Small amounts of paid-in donor share capital can leverage large
sums of private capital from the international capital markets for investment in the SDGs, but it is dependent on the AAA
rating of their bonds.

56
risk off balance sheet and invest beyond their appear to play a more important role in
traditional risk appetites.88 To date, however, mobilising private finance in LICs than MDBs
this approach does not seem to have resulted in and RDBs (section 6.2). For example, France
a material shift in the overall risk appetite of the and the US mobilise more private finance in
MDB and DFI system, as implied by the limited LICs than the IFC. France, the US, Norway and
use of this blended concessional finance and the the Netherlands mobilise more private finance
relatively limited use of subordinate instruments than the AfDB, which, given its mandate and
to mobilise private finance in LICs. geographical focus, would suggest that the AfDB
Although it is early days,89 and should play a larger role.
notwithstanding the impact of the lack of At the multilateral level, it is not immediately
investable opportunities in LICs, more could obvious how conservative the MDB financial
be done by donors and shareholders to enable models are in terms of capital adequacy. Some
MDBs and DFIs to increase their risk appetite. argue that that financial models of several MDBs
For example, many of the concessional funds and RDBs are overly conservative and that there
and facilities provided by donors that MDBs and is room to expand risk appetite and investment
some DFIs use to blend require a positive rate of in LICs. All of our selected MDBs and RDBs
return, or is returnable capital, which restricts the have very strong capital-adequacy positions,
level of risk this capital can carry. Donors could defined by S&P as risk-adjusted capital ratios of
start by reviewing the rates of return they require more than 15%, up to 23% (Figure 20).
on these facilities to see whether they could MDBs’ and RDBs’ scope to take on more risk
provide more funding in pure grant form. without affecting their AAA rating is a matter
At the bilateral DFI level, there may be more of debate, complicated by the different and
room for manoeuvre among those not ‘slave’ to frequently changing rating methodologies of the
the markets, as shareholders could reassess DFI three main credit-rating agencies (S&P, Fitch and
profitability targets. The UK’s DFID, for example, Moody’s), some of which are not transparent in
has lowered CDC Group’s required rate of their assessment or are overly conservative. This
return from 3.5% to remaining profitable at the makes it hard for MDBs and their shareholders
institutional level. Other bilateral shareholders to understand their headroom according to
could follow its lead. There are other ideas on the credit-rating methodologies compared with their
table, too, such as the creation of special-purpose own models of capital adequacy, which naturally
vehicles (SPVs) that focus on providing high-risk limits their operational capacity.
capital (such as early-stage finance and high-risk Even so, some argue that if MDBs and RDBs
project tranches) (Lee, 2018) and the IADB’s (other than the IFC)90 included a portion of
proposal to explore the creation of dedicated callable capital in their own internal capital-
off-balance-sheet facilities to help overcome adequacy models, lending capacity could be
institutional capital-adequacy constraints substantially increased (Humphrey, 2018b).
hindering risk appetite and the ability to expand Relevant for LICs (as the majority are in sub-
the use of risk-mitigation instruments (Pereira Saharan Africa), Humphrey sees potential for
dos Santos and Kearney, 2018). the AfDB to increase lending by $14.1 billion
Interestingly, we find that in contrast to the (almost double its lending portfolio). These
overall landscape, bilateral DFIs and actors illustrative calculations imply that there is room

88 In our review of 11 institutions, we noted that this approach was used more commonly by MDBs, multilateral DFIs
and RDBs. It was a less common approach for the bilateral DFIs to blend external concessional finance with their
own‑account resource.

89 A number of large blended concessional initiatives (e.g. the IDA PSW) and several of the more ambitious MDB and DFI
strategies, which place more emphasis on investment in LICs, are new (e.g. the new IFC 3.0 strategy and the new CDC
2017–2021 strategy), so it will take time for the results to filter through.

90 The IFC has no callable capital, so S&P issues a standalone credit profile.

57
Figure 20 MDB and RDB risk-adjusted capital-adequacy ratios, 2014–2016

2014 2015 2016

40%

35%

30%

25%
Extremely strong
20%

15% Very strong

10% Strong

5%

0%
IFC EIB ADB AfDB

Source: S&P Global Ratings (2017).

Note: ADB increase is due to the merger of the Asian Development Fund into ADB ordinary capital resources on 1 January 2017.

for MDBs and RDBs to increase their risk not an easy problem to solve, but, borrowing
appetite without affecting their AAA rating.91 from Humphrey (2018b), two important starting
There is much debate about MDBs’ lack of a options include the development of a more
uniform approach to assessing their capital uniform approach to assessing capital adequacy
adequacy, the approach of the ratings agencies and the incorporation of a portion of callable
and their impact on MDB operations. This is capital into capital-adequacy calculations.

91 These illustrative calculations assume that the risk profile of the portfolio remains the same, but increased concentration in LICs
would reduce this potential headroom. By how much would need to be explored with further research. Given the size of the
headroom in relation to their portfolios, however, it is reasonable to assume that there is room to increase risk appetite.

58
8 Conclusions and
recommendations
The operational reality and blended-finance SDG financing gap need to be recalibrated: a
experience to date suggest that there is an urgent ‘billions to billions’ framing is a more realistic
need to recalibrate the policy debate and temper operational reality.
policy-makers’ expectations about the potential 2. Better understanding of the poverty and
of blended finance to plug the SDG financing development impacts of blended finance and
gap in developing countries. There is a significant its true costs is required to assess value for
disconnect between policy rhetoric (‘billions to money and ensure effective policy-making
trillions’) and the operational reality of blended and allocation of ODA.
finance (‘billions to billions’). This will help inform 3. The potential of blended finance in LICs is
policy decisions on the allocation of ODA. constrained by several factors, most notably
While blended finance presents a global the countries’ poor investment climate, lack
opportunity to maximise the catalytic effect of of investable opportunities, lack of tailored
ODA and mobilise significant sums of additional approach and limited risk appetites of the
private commercial finance, we would argue that MDBs and DFIs.
it also comes with attendant risks, which may 4. The big push to invest more ODA in blended
have unintended consequences for providers of finance risks diverting ODA away from its
ODA. First, the big push to invest more ODA core agenda of helping eradicate poverty in
in blended finance may result in the further the poorest countries.
concentration of MIC investments in MDB 5. Effective policy-making is hindered by
and DFI portfolios. Second, there is a risk that the lack of a common official blended-
mobilisation targets will shift the emphasis finance framework and very poor data
away from prioritising development impact and availability, which hampers transparency and
ensuring financial additionality. Third, it may accountability and undermines public trust in
well become more expensive to mobilise more this approach.
private finance in LICs, as MDBs and DFIs chase
fewer ‘investable’ projects and invest in more This report offers a suite of recommendations
marginal investment opportunities. with a view to lowering expectations for blended
MDBs and DFIs will need to collectively adopt finance to more appropriate levels and improving
a clearer, tailored approach to blended finance in its use:
LICs. This will be crucial to ensure that increased
investment of ODA in blended finance, especially 1. If blended finance is to be scaled up, MDBs
blended concessional finance, delivers for LICs. and DFIs will need to get better at using
This report finds that: blended finance to mobilise private finance,
while managing the higher level of risk this
1. On average, $1 of MDB and DFI investment implies. If blended finance is to be scaled up
mobilises $0.75 of private finance. This falls rapidly, overall leverage ratios will need to
to $0.37 in LICs. Unrealistic expectations increase significantly. Fundamental systemic
of blended finance’s potential to plug the changes will be required, which will mean

59
 taking on far riskier projects. Such changes transparent than providing a direct subsidy
might include: to the private sector. For example, MDBs,
a. making more use of concessional finance DFIs and donors could make greater use of
and subordinate instruments to meet the grant finance to fund efforts to strengthen
risk-mitigation needs of the private sector local investment climates, focused on
in LICs supporting country-led programmes of policy
b. the use of more concessional finance to reform, local financial-sector development
fund project preparation, early-stage and capacity-building in LICs. As blended
project finance and to develop more concessional finance has not targeted well the
innovative risk-sharing instruments poorest countries, and is not used equally in
c. revisiting required rates of return on all sectors, donors need to manage the risk
concessional resources used in blending that increased investment of ODA in blended
and the ‘hurdle rates’ of bilateral DFIs, finance could divert ODA from the countries
in other words, accepting higher levels of and sectors that need it most.
financial risk 3. There is an urgent need for better data and
d. MDBs developing a more uniform transparency. Efforts should be made to align
approach to assessing capital adequacy, and harmonise the OECD and MDB blended-
which would allow a more transparent finance frameworks. All institutions should
understanding of the scope to take on more publish disaggregated project-level data.
risk and incorporate callable share capital The OECD DAC urgently needs to resolve
into their capital-adequacy models. outstanding issues on the treatment of private-
2. Donors need to think carefully about the sector instruments in the modernisation of
allocation of ODA and the risks and trade- ODA, and efforts should be made to publish
offs of investing ODA through blended the ‘grant equivalent’ of blended-finance
finance. There may well be other public transactions to the OECD CRS. This should be
policy interventions that could support the disclosed publicly at some semi-aggregated level
achievement of the same stated development to overcome the commercial confidentiality
objective, which are more effective and concerns of MDBs, DFIs and donors.

60
References

Benn, J., Sangaré, C. and Hos, T. (2017) Amounts mobilised from the private sector by official
development finance interventions – guarantees, syndicated loans, shares in collective investment
vehicles, direct investment in companies and credit lines. OECD Development Co-operation
Working Paper 36. Paris: Organisation for Economic Co-operation and Development (www.
oecd-ilibrary.org/docserver/8135abde-en.pdf?expires=1552399091&id =id&accname=guest&
checksum=9EE9516B020DB0A19F064C5FBDE47Bd4).
Blended Finance Taskforce (2018a) Mobilising private capital for the SDGs at scale.
Programme of Action. London: Blended Finance Taskforce (static1.squarespace.com/
static/5acdc066c258b4bd2d15050b/t/5bd1f24deef1a1396c5267fe/1540485730232/BFT_Action_
Programme_Oct2018.pdf)
Blended Finance Taskforce (2018b) Better finance, better world. Consultation paper of the
Blended Finance Taskforce in consultation with the Business & Sustainable Development
Commission and SYSTEMIQ. London: Blended Finance Taskforce (www.blendedfinance.earth/
better-finance-better-world)
Carter, P. (2015) Why subsidise the private sector? What donors are trying to achieve, and what
success looks like. ODI Report. London: Overseas Development Institute (www.odi.org/sites/odi.
org.uk/files/odi-assets/publications-opinion-files/9948.pdf)
Carter, P. (2018) ‘The pitfalls of leverage targets’. CDG Blog. Washington, DC: Center for Global
Development, 9 February (www.cgdev.org/blog/pitfalls-leverage-targets)
Carter, P., Van De Sijpe, N. and Calel, R. (2018) The elusive quest for additionality. Working Paper
495. Washington, DC: Center for Global Development (www.cgdev.org/sites/default/files/elusive-
quest-additionality.pdf)
Development Initiatives (2016) The role of blended finance in the 2030 Agenda: setting out an
analytical approach. Discussion paper. Bristol, UK: Development Initiatives (devinit.org/post/
the-role-of-blended-finance-in-the-2030-agenda-setting-out-our-analytical-approach/#)
DFI Working Group (2017) DFI Working Group on blended concessional finance for private sector
projects – summary report. A joint report by the AfDB, ADB, AIIB, EBRD, EDFI, EIB, IADB Group, ICD
and IFC. Washington, DC: International Finance Corporation (www.ifc.org/wps/wcm/connect/30635fde-
1c38-42af-97b9-2304e962fc85/DFI+Blended+Concessional+Finance+for+Private+Sector+Operations_
Summary+R....pdf?MOD=AJPERES)
DFI Working Group (2018) DFI Working Group on blended concessional finance for private sector
projects. A joint report by the AfDB, ADB, AIIB, EBRD, EDFI, EIB, IADB Group, the Islamic
Corporation for the Development of the Private Sector (ICD) and IFC, October 2018 Update.
Washington, DC: International Finance Corporation (www.adb.org/sites/default/files/institutional-
document/457741/dfi-blended-concessional-fiance-report.pdf)
EDFI – Association of European Development Finance Institutions (2018) ‘Members facts & figures’.
Brussels: European Development Finance Institutions (www.edfi.eu/members/facts-figures/)
G7 (2018) ‘Co-chairs’ summary: G7 joint development and finance ministers’. Summary
of the meeting of G7 development and finance ministers in Whistler, Canada, 1 June.
Charlevoix, Canada: Canadian G7 Presidency (g7.gc.ca/wp-content/uploads/2018/06/
CO-CHAIRS’-SUMMARY-G7-JOINT-DEVELOPMENT-AND-FINANCE-MINISTERS_EN.pdf)
G20 Germany (2017) ‘G20 Hamburg Action Plan’. Annex to G20 leaders’ declaration, 8 July.
Hamburg: German G20 Presidency (www.g20germany.de/Content/DE/_Anlagen/G7_G20/2017-
g20-hamburg-action-plan-en___blob=publicationFile&v=4.pdf)

61
Gulrajani, N. (2017) ‘Bilateral donors and the age of the national interest: what prospects
for challenge by development agencies?’ World Development 96: 375–389 (www.odi.org/
publications/10815-bilateral-donors-and-age-national-interest-what-prospects-challenge-
development-agencies)
Humphrey, C. (2018a) Channelling private investment to infrastructure: what can multilateral
development banks realistically do? ODI Working Paper 534. London: Overseas Development
Institute (www.odi.org/sites/odi.org.uk/files/resource-documents/12176.pdf)
Humphrey, C. (2018b) The role of credit rating agencies in shaping multilateral finance; recent
developments and policy options. Policy paper commissioned by the Intergovernmental Group of
24. Washington, DC: G–24. (www.g24.org/wp-content/uploads/2018/04/The-Role-of-Credit-Rating-
Agencies-in-Shaping-Multilateral-Finance.pdf)
IFC – International Finance Corporation (n.d.) ‘Global toolbox to advance private sector
development’. Webpage. Washington, DC: International Finance Corporation (www.ifc.
org/wps/wcm/connect/publications_ext_content/ifc_external_publication_site/publications/
MDBs-Global-Toolbox)
IFC (2013) DFI guidance for using investment concessional finance in private sector operations.
Private Sector Roundtable Note. Washington, DC: International Finance Corporation (www.
ifc.org/wps/wcm/connect/2f1575b9-9420-4b84-a440-43b5aeae4cf9/2013+DFI+Guidance+_
Concessional+Finance_Private+Sector+Operations.pdf?MOD=AJPERES)
IFC (2018) ‘IFC strategy and capital increase’. Webpage. Washington, DC: IFC (www.ifc.org/wps/wcm/
connect/8e5d9b9f-f70e-414a-8edb-330ade432421/Capital+Increase+-+Presentation+with+public+in
formation.pdf?MOD=AJPERES)
Lee, N. (2018) More mobilizing, less lending: a pragmatic proposal for MDBs. CGD Brief, April.
Washington, DC: Center for Global Development (www.cgdev.org/expert/nancy-lee).
Manuel, M., Desai, H., Samman, E. and Evans, M. (2018) Financing the end of extreme poverty. ODI
Report. London: Overseas Development Institute (www.odi.org/sites/odi.org.uk/files/resource-
documents/12411.pdf)
OECD – Organisation for Economic Co-operation and Development (2017) A new DAC: innovations
for the 2030 Agenda. OECD Development Assistance Committee High Level Communiqué, 31
October. Paris: – Organisation for Economic Co-operation and Development (www.oecd.org/dac/
DAC-HLM-2017-Communique.pdf)
OECD (2018a) Making blended finance work for the Sustainable Development
Goals. Paris: OECD Publishing (read.oecd-ilibrary.org/development/
making-blended-finance-work-for-the-sustainable-development-goals_9789264288768-en#page1)
OECD (2018b) OECD DAC blended finance principles for unlocking commercial finance for the
Sustainable Development Goals. Paris: Organisation for Economic Co-operation and Development
(www.oecd.org/dac/financing-sustainable-development/development-finance-topics/OECD-Blended-
Finance-Principles.pdf)
OECD (2018c) Mobilisation of private finance and support for the 2030 agenda: briefing on
efforts to harmonise OECD and MDB measurement methodologies. Development Co-operation
Directorate Briefing Paper for the OECD DAC. Paris: Organisation for Economic Co-operation
and Development (www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=DCD/
DAC(2018)25&docLanguage=En)
OECD (2018d) Methodologies to measure amounts mobilised from the private sector. OECD DAC
document. Paris: Organisation for Economic Co-operation and Development (www.oecd.org/dac/
financing-sustainable-development/development-finance-standards/OECD-DAC-Methodologies-
to-measure-amounts-mobilised-from-the-private-sector-by-official-development-finance-
interventions-2016.pdf)

62
OECD (2018e) ‘Provisional reporting arrangements for private sector instruments’. OECD
DAC announcement, 5 December 2018. Paris: Organisation for Economic Co-operation and
Development (www.oecd.org/dac/private-sector-instruments.htm)
OECD (2018f) United Kingdom mid-term review. Development Co-operation Directorate report.
Paris: Organisation for Economic Co-operation and Development (www.oecd.org/dac/peer-reviews/
UK-mid-term-review.pdf)
OECD (2018g) ‘Development aid stable in 2017 with more sent to poorest countries’. OECD
Newsroom. 9 April. Paris: Organisation for Economic Co-operation and Development (www.oecd.
org/newsroom/development-aid-stable-in-2017-with-more-sent-to-poorest-countries.htm)
OECD (2018h) Global outlook on financing for sustainable development 2019: time
to face the challenge. Paris: OECD Publishing (read.oecd-ilibrary.org/development/
global-outlook-on-financing-for-sustainable-development-2019_9789264307995-en#page1)
Pereira, J. (2015) Leveraging aid: a literature review on the additionality of using ODA to leverage
private investments. London: UK Aid Network
Pereira dos Santos, P. and Kearney, M.C. (2018) Multilateral development banks’ risk
mitigation instruments for infrastructure investment. IDB Technical Note 1358’.
Washington, DC: Inter-American Development Bank (publications.iadb.org/en/publication/
multilateral-development-banks-risk-mitigation-instruments-infrastructure-investment)
S&P Global Ratings (2017) Supranationals. Special Edition Report. Stockholm: S&P Global (www.
spratings.com/documents/20184/908545/US_SR_Event_Webcast_MultilateralNov17_Article2.pdf/
fc892865-5302-4dd6-93ca-8469c8752818)
Tyson, J.E. (2018) Private infrastructure financing in developing countries: five challenges, five
solutions. ODI Working Paper 536. London: Overseas Development Institute (www.odi.org/sites/
odi.org.uk/files/resource-documents/12366.pdf)
UNCTAD – United Nations Conference on Trade and Development (2014) World investment report
2014: investing in the SDGs. New York and Geneva: United Nations Conference on Trade and
Development (unctad.org/en/PublicationsLibrary/wir2014_en.pdf)
UNDESA – United Nations Department of Economic and Social Affairs (2015) Addis Ababa action
agenda of the Third International Conference on Financing for Development. New York: United
Nations (sustainabledevelopment.un.org/index.php?page=view&type=400&nr=2051&menu=35)
United Nations (2015) Paris agreement. New York NY: United Nations (unfccc.int/sites/default/files/
english_paris_agreement.pdf)
Warner, A.M. (2013) A framework for efficient government investment. IMF Working Paper
13/58. Washington, DC: International Monetary Fund (www.imf.org/en/Publications/WP/
Issues/2016/12/31/A-Framework-for-Efficient-Government-Investment-40362)
World Bank (2015) From billions to trillions: transforming development finance. Document prepared
jointly by the AfDB, ADB, EBRD, EIB, IADB, IMF and World Bank Group for the 18 April
Development Committee meeting. Washington, DC: World Bank (siteresources.worldbank.org/
DEVCOMMINT/Documentation/23659446/DC2015-0002(E)FinancingforDevelopment.pdf)
World Bank (2017a) ‘Poverty measurement data’. Electronic dataset. Washington, DC: World Bank
(iresearch.worldbank.org/PovcalNet/povDuplicateWB.aspx)
World Bank (2017b) Mobilization of private finance by multilateral development banks – 2016 joint
report. Joint report by the AfDB, ADB, AIIB, EBRD, EIB, IADB, the Inter-American Investment
Corporation (IIC), the Islamic Corporation for the Development of the Private Sector (ICD), IFC,
the IsDB, MIGA, the New Development Bank and the World Bank. Washington, DC: World Bank
Group (documents.worldbank.org/curated/en/860721492635844277/pdf/114433-REVISED-11p-
MDB-Joint-Report-Mobilization-Jul-21.pdf)
World Bank (2017c) ‘Annual remittance data’. Electronic dataset. Washington, DC:
World Bank) (www.worldbank.org/en/topic/migrationremittancesdiasporaissues/brief/
migration-remittances-data.)

63
World Bank (2018a) Mobilization of private finance by multilateral development banks and
development finance institutions 2017. Joint report by AfDB, ADB, AIIB, EBRD, EIB, ICD, IADB
and IDB Invest, IFC, IsDB, MIGA, the New Development Bank and the World Bank. Washington,
DC: World Bank Group (documents.worldbank.org/curated/en/691741532976526194/
Mobilization-of-private-finance-by-multilateral-development-banks-and-development-finance-
institutions-2017)
World Bank (2018b) MDB methodology for private investment mobilisation – reference guide.
Washington, DC: World Bank Group (documents.worldbank.org/curated/en/813091529416636675/
pdf/WP-PUBLIC-DocumentsPrivInvestMob-Draft-Ref-Guide-Master-June2018-v3.pdf)
World Bank (2018c) IDA18 IFC-MIGA private sector window (PSW): IDA18 mid-term review
Washington DC: World Bank Group (documents.worldbank.org/curated/en/157801542813052758/
IDA18-Mid-Term-Review-IDA18-IFC-MIGA-Private-Sector-Window)
World Economic Forum (2016) Insights from blended finance investment vehicles and facilities.
Geneva: World Economic Forum (www3.weforum.org/docs/WEF_Blended_Finance_Insights_
Investments_Vehicles_Facilities_report_2016.pdf)

64
Annex 1 Institutional
selection process
In selecting institutions for analysis, our aim was to include the largest actors in blended finance
overall, as well as the most important actors in a LIC context. We also wanted to compare differences
between (a) MDBs, (b) RDBs and (c) bilateral actors. The selection process was as follows:92,93

1. The three largest institutions in the blended-finance world in each category

The top nine institutions fit perfectly into the three groups:
a. MDBs: MIGA, IFC and IBRD/IDA
b. RDBs: EIB, EBRD and IADB
c. Bilateral: US, UK, France
However, at the time of selection in November 2017, the EBRD had no recipient countries classified
as LICs, while IADB only had one (Haiti). These banks will naturally have very small or non-existent
flows to LICs, which makes them less relevant for our study. We thus replaced them with the next
largest institutions, the ADB (with two LICs) and the AfDB (with 27 LICs).

Table A1 Largest blended-finance actors

Actor Total blended finance ($) (2012–2015) Category

US 15,984,962 Bilateral
EIB 13,007,690 RDB
MIGA 11,850,566 MDB
IFC 9,891,875 MDB
EBRD 4,487,264 RDB
UK 3,672,242 Bilateral
IBRD/IDA 3,546,052 MDB
IADB Group 2,943,397 RDB
France 2,766,117 Bilateral
Source: see footnote 92.

92 All data analysis for the selection process is based on the 2016 OECD mobilisation survey (Benn et al., 2017), as it has
the most comprehensive available disaggregated data on blended finance.

93 While the public version of the OECD mobilisation survey disaggregates the dataset by agency, it does not disaggregate
it by recipient country, so we cannot analyse flows to LICs. However, an internal dataset the OECD shared with the
authors does disaggregate by recipient, though not by agency, only by donor. For example, it has data on the US, but it is
not possible to see the split between OPIC and USAID. We used the internal dataset, as we were interested in the recipient
country. If we had focused on agencies at this stage, Germany’s Kreditanstalt für Wiederaufbau (KfW) and Denmark’s
Investment Fund for Developing Countries (IFU) would have been larger than France’s AFD and Proparco.

65
2. The three largest blended-finance institutions in LICs
The top three institutions in each category are largely the same as the above list. The only addition is
Norway. The Netherlands is the fourth-largest bilateral actor in LICs and does not make the cut, even
though it is a larger mobiliser in LICs than the regional banks. Note that only two regional banks mobilise
in LICs. The ADB does not have any mobilisation in LICs, although there are LICs among its recipients.

Table A2 Largest blended-finance actors in LICs

Actor Total BF (2012-2015) Category

MIGA 661,926 MDB
France 546,304 Bilateral
US 502,210 Bilateral
IFC 326,948 MDB
PIDG 191,608 N/A
IBRD/IDA 171,000 MDB
Norway 107,721 Bilateral
Netherlands 78,058 Bilateral
EIB 73,101 RDB

AfDB 62,850 RDB

Source: see footnote 92.

3. Thus, the final list is:

a. MDBs: MIGA, IFC and IDA94
b. RDBs: EIB, ADB and AfDB
c. Bilaterals:95 US (OPIC), UK (CDC), France (AFD/Proparco) and Norway (Norfund)

94 Because of the different country focus of the IDA and IBRD, we assume that IDA is the largest donor in LICs.

95 Using the public dataset, we can analyse the relative sizes of agencies in total (although not by recipient). In France, AFD
and Proparco are of similar importance, while in the other countries, bilateral DFIs are by far the biggest mobilisers. We
thus selected only the DFIs (OPIC and CDC Group).

66
Annex 2 Data-collection
methodology

Introduction
Comparable and good-quality granular data are not readily available to enable a nuanced analysis
and understanding of how much ODF is invested in blended finance and what private finance this
investment is mobilising, disaggregated by instrument, sector and country. Consequently, we have had
to build our own database to try to paint a granular picture.
Our objective was to collect data on the amount of ODF invested to mobilise private finance by our
11 selected institutions. Our key assumption was that not all MDB and DFI activities were aimed at
mobilising private finance, so using the full commitment of an MDB or DFI overestimated the funds
used for blended finance.
For the purposes of our quantitative analysis of the landscape and in light of the data available, we
focused on the strategic use of official development finance96 to mobilise additional97 private finance
for development purposes. In practice, this meant we measured institutions’ funding for projects that
aimed to directly mobilise private finance, proxied by the institution’s commitments.
To get a detailed picture of the institutions’ commitments, we aimed to collect as much of the
disaggregated data as possible. Specifically, the data had to be detailed enough to map which
countries received the financing, in which sectors, using which instruments. To obtain this level of
disaggregation, we used publicly available project-level databases provided by the institutions, where
available (Table A3), and screened individual projects to determine whether or not they aimed to
directly mobilise private finance (see below for methodology) to build our ODI blended-finance
commitment data.
Due to data availability issues, and to ensure good data coverage and coherence while ensuring
sufficient coverage for trend analysis, we limited the datasets to 2013–2017. For many institutions,
public disclosure of project information was lagged for confidentiality reasons, which meant that more
recent projects might not have been disclosed. Consequently, the picture for 2017 was not complete
at the time of compilation. Thus, we excluded 2017 from those parts of the analysis that dealt with
annual averages. For AFD, Proparco and the ADB, we only had 2016 data.
In an attempt to verify the completeness of the individual institutions’ databases, we compared total
annual commitments in the project databases with those in the institutions’ annual reports. Figure A1
compares the total annual sums from our database with the numbers from annual reports. Although
the numbers are not identical, they are roughly in the same ball park, suggesting the projects covered
largely reflect total operations. This means we can be reasonably confident in the disaggregated
database we have built.

96 This is essentially the official finance subset of the OECD definition on the input side and includes concessional
and non-concessional finance (i.e. MDB and DFI own-account resources). Our work excludes analysis of private
development finance.

97 An assessment of additionality is beyond the scope of this research.

67
Table A1 Project-level databases, by institution

Institution Database
ADB ADB has a separate dataset on projects involving commercial co-financing (data.adb.org/dataset/
projects-involving-commercial-cofinancing-2016/resource/cccac7be-de31-4990-8361-
9745d438f7de). Although the data are not at project level, we were able to obtain country-level totals.
AFD, Proparco The agencies shared with us an internal non-public dataset where blended-finance projects were
already identified.
AfDB We were not able to obtain sufficient project-level data to conduct an analysis.
CDC IATI dataset (iatiregistry.org/dataset/cdc-201217)
EIB EIB website (www.eib.org/projects/loan/list/index.htm)
IDA World Bank website (datacatalog.worldbank.org/dataset/world-bank-projects-operations)
IFC World Bank website (finances.worldbank.org/Projects/IFC-Investment-Services-Projects/efin-cagm)
MIGA World Bank website (datacatalog.worldbank.org/dataset/miga-project-portfolio)
Norfund Norfund shared with us a confidential internal dataset with flow data at project level.
OPIC We used the Center for Global Development ‘scraped’ OPIC portfolio dataset (www.cgdev.org/media/
opic-scraped-portfolio-dataset), which captures flow data by project.

Figure A1 Comparison of total annual commitments, project-level datasets versus annual reports

Annual report sums Dataset sums

50

45

40

35

30
$ billions

25

20

15

10

0
AFD (EUR, CDC (GBP, EIB (EUR, IDA (USD, IFC (USD, MIGA (USD, Norfund (NOK, OPIC (USD, Proparco (EUR,
2016) 2013-2016) 2013-2016) FY13-FY17) FY13-FY16) 2013-2016) 2013-2017) 2013-2016) 2016)

Source: Individual institutions’ annual reports and the institutions’ public project-level databases.

Table A2 Exchange rates

Currency 2013 2014 2015 2016 2017

EUR 1.377614 1.211023 1.086653 1.052255 1.199861
GBP 1.657413 1.558642 1.476337 1.23321 1.350291
NOK 0.164767 0.133856 0.112942 0.115646 0.121898
Note: 1 unit of currency = $ value.

68
 In practice, because the raw data vary in coverage, typology (in terms of instrument, sector, region,
etc.) and definition, they are not fully comparable. However, they represent a good proximation. All
findings based on the collated dataset should be treated as estimates, rather than exact results.

Standardising datasets
Because of varying reporting standards, the institutions’ datasets are not always directly comparable.
To make them comparable, we had to standardise the instruments, sectors and values in US dollars.
Values were calculated to US dollars from their original currency, based on the exchange rate on
31 December in the year of commitment (Table A4).

Instruments

•• Any investment in a fund was classified as ‘funds’, regardless of whether the instrument used was
debt, equity or otherwise.
•• Projects classified as risk management (at IFC) and risk participation (at CDC) were grouped into
the ‘risk management’ instrument.
•• For projects where multiple instruments were used, we applied the instrument classification used
by the institution in its database.

Sectors

•• We customised sector classification to translate institutional classifications into a common system

for comparison and analysis. This was approximate (lowest common denominator) to allow
automatic translation from institutions’ groupings, without manual identification (Table A5).

Identifying blended-finance commitments

Identification was carried out at project level. For each project we identified whether the commitment
was used to mobilise private finance or not. To create comparable results across institutions, and to
ensure reproducibility, we used a standardised methodology for all institutions.
We made the following assumptions when assessing whether individual projects were identifiable as
blended finance:

General

1. We assumed all projects intending to mobilise private finance were blended finance. Without
detailed information on investments, we were unable to determine whether projects did mobilise
any private finance. We did not attempt to assess additionality, as this was beyond the scope of this
exercise.
2. Only non-sovereign operations can be blended finance. We excluded all sovereign operations, as we
assumed that these did not involve private financiers.

69
 Table A5 Consolidated sector classifications

New AFD and CDC EIB IDA IFC MIGA Norfund OPIC
classification Proparco*
Banking and 24030-25020 Financials – commercial Credit lines Banking institutions Financial institutions Banking Banking Financial
finance banks services
Financials – consumer Other non-bank financial Capital markets Financial
finance institutions services
Financials –- diversified Financial Microfinance
financials markets
Financials – microfinance Financial
– institutions services
Financials – microfinance Leasing
– SME banks
Extractive 32210-32268 Oil and gas Oil, gas and mining Mining Extractives
sectors
Other energy and Oil and gas
extractives

70
Infrastructure 21010-23640 Information and Composite Energy transmission and Infrastructure Infrastructure Energy ICT
communications infrastructure distribution
technology (ICT)
– telecoms
Infrastructure Energy Non-renewable energy Telecoms, media, and Power Infrastructure
tech
Solid waste Other ICTs Solid waste
management
Telecoms Other transport Telecoms
Transport Power Transport
Urban development Railways Water
Water, sewerage Renewable energy – Water and
hydro/solar/ wind wastewater
Rural and inter-urban
roads
 New AFD and CDC EIB IDA IFC MIGA Norfund OPIC
classification Proparco*
Other/ Everything else Financials – Funds Investment General
Multisector funds funds
other Humanitarian
assistance
Productive 31110-32182 Agribusiness Agriculture, Fisheries Agribusiness and Agribusiness Agriculture Agriculture
sectors and food fisheries, forestry forestry
32310-33210 Consumer – general Industry Forestry Manufacturing Chemicals Food and Hospitality and
agribusiness tourism
Consumer services Services Other agriculture Tourism, retail Manufacturing Manufacturing Industrial
and property
Industrials Manufacturing Other Real estate
and services industrial
partnerships
Real estate Services Other services Retail

71
Tourism Tourism Services
Social 11110-16064 Education – Education Central government Health and education Education
sectors core education
Healthcare – Health Health Healthcare
healthcare providers and
services
Healthcare – Other public admin
life sciences tools and
services
Public admin, transport
Note: * The AFD/Proparco dataset uses CRS purpose code classifications. Codes in the table refer to a range of purpose codes. Sectoral classification is not available for the ADB.
These have been classified as ‘unknown’ in the dataset.
Instruments

3. All projects involving direct equity in private companies were treated as blended finance – we assumed
that the motivation behind direct investment was to attract further capital from private investors.
4. All projects involving guarantees and insurance (to the private sector) were treated as blended
finance – we assumed that the intention of the guarantee/insurance was to unlock a private
investment that would otherwise not have been made.98
5. Projects involving investments in funds were treated as blended finance if the fund also involved other
private investors. If the fund was only financed by DFIs or other public sources, we did not classify
the project as blended finance. In cases where no information on the fund’s investors was found, we
assumed it included private investors (as there tends to be more information on DFI-only funds).
6. Projects involving lines of credit were treated as blended finance if the financial intermediary had
to cover some of the cost of the sub-loans, take on some of the risk or provide some additional
financing on top of the credit line. In practice, in our data, the EIB was the only institution to use this
instrument and it does not disclose in its project descriptions the extent of the sub-loans it covered.99
We thus treat all the projects as blended finance, as long as they go to a private-sector intermediary.
7. For projects involving loans, identification was more complicated. First, if the project description
explicitly stated the loan was in the form of subordinated debt, B-loan or similar, we classified
the project as blended finance. Where this information was not available, we compared the
commitment of the institution to the total project cost. If the total project cost was higher than the
commitment, unless there was any specific information suggesting the co-financer was not a private
investor, we classified the project as blended finance. Note that this applied even if the co-financer
was the investee/lending company. For CDC Group and Norfund, for which no information was
available for total project cost, we assumed loans were blended finance.

Table A6 illustrates the assessment process through project examples, highlighting the differences.
In total, 72% of the projects were classified as mobilising.100 The majority of projects not identified
as blended finance were EIB and AFD sovereign projects, excluded on the basis of assumption 2,
above. Among the remaining projects not identified as blended finance, most were IFC projects. These
were often projects involving direct debt financing to a financial intermediary. Figure A2 shows the
breakdown of mobilising and non-mobilising projects. The share is slightly lower for LICs only (66%).

98 Note that some MIGA projects provide guarantees to public entities, such as DFIs. These are not marked as blended
finance, on the basis of assumption 2 (public-sector recipient).

99 Although France and Norway reported the use of lines of credit in the OECD survey, we were not able to identify these
from the project descriptions in the project databases provided to us.

100 This section excludes IDA sovereign lending.

72
Table A6 Examples from the blended-finance assessment process

Instrument Institution Project & amount of commitment Classified Why?

recognised in ODI database as blended
[Project ID/name] finance?
Equity IFC $22 million equity investment in Cimenterie de Yes All direct equity projects are classified
Lukala, a private cement manufacturer in the as blended finance
Democratic Republic of the Congo
[ID: 36898]
Guarantee MIGA $2.2 million guarantee to the Burundian coffee Yes All guarantee projects are classified as
operating company Budeca SA blended finance
[ID: 13502]
Funds CDC $17 million investment in the EuroMena III No The EuroMena III investment fund is
investment fund only financed by DFIs (CDC, IFC, DEG,
[ID: GB-COH-03877777-F312701] Proparco), and no private investors were
identified
CDC $11 million investment in the Catalyst II Yes The fund is financed also by other
investment fund private investors
[ID: GB-COH-03877777-F315001]
Line of EIB €70 million credit line to Development Bank No Financial intermediary (Development
credit of Ethiopia for on-lending to SMEs Bank of Ethiopia) is not a private-sector
[Name: LEASING AND LENDING FOR SMES] entity; the signatory to the loan is the
Federal Democratic Republic of Ethiopia
EIB €20 million credit line to AccessBank Yes Financial intermediary (AccessBank) is a
Azerbaijan for on-lending to SMEs private bank
[Name: ACCESSBANK AZERBAIJAN LOAN FOR
SMES II]
IFC $20 million A loan<?> to private Armenian bank No Total project cost ($20 million) is
Inecobank to support its acquisition of another covered by the IFC – no other sources of
bank [ID: 37196] finance are involved
Loan IFC $100 million A loan to Argentinian company Yes Total project cost ($600 million) is not
Telecom Personal to support deployment of covered by the IFC. Project description
nationwide telecom network mentions mobilisation from B Lenders/
[ID: 36171] parallel loans of up to $400 million

73
Figure A2 Breakdown of projects by mobilising status
Unknown Non-mobilising Mobilising

60

50
Sum of commitments, $ billions

40

30

20

10

0
ADB AFD CDC EIB IDA IFC MIGA Norfund OPIC Proparco
Note: ‘Unknown’ implies insufficient evidence was available in project documents to assess status.

Income classifications
World Bank income classifications (LIC, LMIC, UMIC, HIC) were used to classify countries. We used
individual classifications for each project, based on the classification of the recipient country on the
date of the commitment.101 Because some countries were reclassified during the period under analysis,
two projects to the same country (but at different times) can have different income classifications. This
is in contrast to the OECD mobilisation survey, for example, which uses a single classification per
country – this can cause numbers to diverge.102

101 World Bank income classifications are adjusted every July (the start of the World Bank Financial Year (FY)), based on its
GNI per capita the preceding year. A country has a certain income classification from July to June. For example, Kenya
was reclassified as an LMIC in July 2015 (at the beginning of FY2016), before which it was classified as an LIC. For an
investment in Kenya in August 2015 (FY2016), we classified the project as an LMIC, but for an investment in June 2015
(FY2015), we classified it as an LIC. For some institutions, information on date was not included, but year was. In these
cases, we treated the commitment as having been made on 1 January of that year.

102 Although we could see that the OECD applied a single income classification to a country throughout the period, it was
not clear what its methodology was; some countries were classified according to their income classification at the start of
the survey period and some at the end of the survey period.

74
Annex 3 Scatter plots
Figure A3 shows a slight positive correlation between financial depth and private finance mobilised. To
some extent, this is explained by the slight positive correlation between financial depth and GDP per capita,
but also the fact that around one third of the private finance mobilised is in the banking and finance sector.
Figure A4 shows a slight positive relationship between the Human Assets Index103 and the amount of
private finance mobilised, probably because the Index is very closely correlated with GNI per capita.

Figure A3 Relationship between private-finance mobilisation and financial depth

4.0
Private finance mobilised 2012–2015

3.5

3.0
2.5
$ billions

2.0
1.5

1.0
R² = 0.0553
0.5

0.0
0 20 40 60 80 100 120 140 160 180
Domestic credit to private sector (% of GDP)

Source: Benn et al. (2017), World Bank Development Indicators (FS.AST.PRVT.GD.ZS).

Figure A4 Relationship between private-finance mobilisation and the Human Assets Index

4.0

3.5
Private finance mobilised 2012–2015

3.0

2.5
$ billions

2.0

1.5

1.0

0.5 R ² = 0.0188
0.0
0 10 20 30 40 50 60 70 80 90 100
Domestic credit to private sector (% of GDP)

103 www.ferdi.fr/en/indicator/human-assets-index

75
 Evidence.
Ideas.
Change.

ODI is an independent, global think tank,

working for a sustainable and peaceful
world in which every person thrives. We
harness the power of evidence and ideas
through research and partnership to
confront challenges, develop solutions, and
create change.

ODI
203 Blackfriars Road
London SE1 8NJ

+44 (0)20 7922 0300

info@odi.org

odi.org
odi.org/facebook
odi.org/twitter