Development Finance Institutions and Infrastructure: Findings from a Systematic Review of Evidence for Development Additionality
By leveraging private investment, Development Finance Institutions (DFIs) aim to reduce the infrastructure financing gap in the developing world. This article summarises the findings of a Systematic Review of the evidence on the development impact of DFI support for private participation in infrastructure.
Infrastructure is vital for development but is seriously underprovided in much of the developing world. The financing shortfall is estimated at $48 billion per year in Africa alone (1). With public finance insufficient to fill this gap, recent decades have seen an increasing focus on mobilising private investment. The volume of private finance flowing to infrastructure in developing countries, however, has been well below that anticipated by many following deregulation in the 1990s (2). There has been progress, but not enough: if this situation is to be resolved, new combinations of public and private initiatives will be required.
Development Finance Institutions (DFIs) aim to leverage private investment for projects that are close to commercial viability, have large potential developmental impacts, but are in sectors or countries where commercial banks are reluctant to invest due to perceptions of excessive risk. By investing their own resources in projects, DFIs seek to mitigate these risks and so give private investors the confidence to invest. A number of instruments are employed to achieve this: investment (loans and equity), risk mitigation (for example loan guarantees), advisory services (to governments), and project preparation and development services.
As well as the quantity of investment, however, quality also matters. New facilities that provide low quality services, or which poorer sections of society cannot access or afford, may tick boxes with respect to quantity, but do badly in terms of quality. Accordingly, Development Finance Institutions (DFIs) seek to use their resources to improve outcomes in both areas: to attract more private investment, but also to enhance the development impact that this finance can achieve.
It is within this context that the systematic review upon which this article is based addressed the following questions: What is the evidence of the impact of DFI support for private-participation-in infrastructure (PPI), on economic growth and poverty reduction? What conclusions can be drawn from this evidence to help DFIs better target their investment to maximise their impact on economic growth and poverty reduction?
In approaching these questions we focused on the ‘additionality’ that DFIs might create with respect to growth and poverty. Additionality is defined as impact beyond that which would have occurred without DFI participation. For example:
1. Financial additionality, raises the total quantity of investment, both in terms of DFI resources and the private investment they can leverage as co-investors.
2. Demonstration effects increase the quantity of investment by demonstrating to private investors that the risks are lower than they had perceived, resulting in a significant increase in private investment that is independent of DFI activity.
3. Policy additionality may increase both the quantity and quality of investment by creating an ‘enabling environment’ that is attractive to investors, and influencing the nature of this investment to boost development impacts
4. Design additionality increases the quality of investment by amplifying the development impacts of a project through the use of particular design features.
The review proceeded in two phases. Phase 1 examined publicly available evidence in the literature, including DFIs’ own material. Much relevant evidence, however, is to be found in DFIs’ internal evaluations, which are usually not publicly available due to issues of commercial confidentiality. To address this, Phase 2 reviewed internal evaluations for a group of 5 major DFIs (3). In total, more than 400 documents were reviewed, roughly half in each phase of the review. Throughout, evidence that DFIs do (or do not) create these different forms of additionality was gathered and assessed.
Overall, the evidence gathered in the review suggests that DFIs have positive development impacts. However, the review also identified ways in which they could enhance the developmental outcomes of their investments, and thus increase the impact of donor resources. Our main findings are as follows:
First, the evidence suggested that DFIs do create financial additionality, particularly in less commercially viable sectors. Specifically, we found support for the propositions that DFIs are able to: a) supply long-term finance, which is often essential for infrastructure but frequently unavailable in low-income countries; b) mitigate early-stage project risk, thus leveraging additional finance by improving the attractiveness of deals; and c) provide and leverage finance counter-cyclically.
Second, we found support for the view that DFIs influence project design and the policy context to boost growth. Both in terms of project selection (e.g. projects that remove ‘bottlenecks’ to growth), and during the project design phase, DFIs seek to enhance growth effects. Similarly – though to a lesser extent – some DFIs seek to influence regulatory frameworks with the aim of enhancing growth (e.g. through liberalisation) or by building public sector capacity to pursue private sector development.
Third, we found little to suggest that DFIs influence project design and policy to improve direct poverty impacts. As well as its indirect effects via growth, infrastructure can have direct effects on poverty by,for example, providing affordable access to services that were previously not available, or access to new markets, will have direct poverty impacts. Certain aspects of project design will greatly influence the extent of these direct effects, such as the ability of the poor to physically access services, or their ability to afford fees. We found very little evidence that DFIs actively seek to influence these design features to increase direct poverty effects. There was also limited evidence that DFIs proactively seek to create local employment and stimulate broader economic development through, for example, forging linkages with local suppliers including SMEs.
Importantly, we found a small number of examples of this type of activity, demonstrating that it would be feasible to do more. Indeed, in many of these areas, it was not clear why greater efforts are not habitually made by the DFI sector, as they appear to have scope to do so.
Fourth, it is clear that more emphasis should be placed on project selection. If the goal of DFI activity is to maximise the impact of scarce donor funds, they should select those projects with the potential to create the greatest impacts. While this may seem obvious, it is not generally reflected in current practice. DFIs often appear to accept the projects that come their way and clear a certain hurdle, rather than choosing projects on the basis of a comparison of alternatives and their likely development impact. While this may be justified because of a lack of investable options a better approach might be to devote more efforts to identifying and preparing projects with the greatest potential impacts. In our view, the drawbacks in terms of additional cost and time are outweighed by the greater impacts that could be achieved.
Given that the funds available to DFIs cannot fill the infrastructure financing gap, the demonstration effect they create that increases private financing is, in some ways, their core task. This needs to be put in perspective, however. Our fifth finding is that the ability to achieve such demonstration effects may have real limits. In part, DFIs are able to do what they do because they are DFIs. Political backing allows them to borrow on favourable terms as there is no default risk. Further, DFI-backed loans are less risky than pure commercial loans: borrowers are reluctant to default due the harm this could do to their relationship with the donor government (or institution in the case of the World Bank and Regional Development Banks). Given these factors, DFIs can lend on better terms (e.g. for longer maturities) and hold riskier portfolios than is possible for private actors. It is therefore not always possible for private actors to follow DFIs’ example and make the same investments.
While we found very few projects where DFIs appeared to have had direct poverty effects, it was striking that all projects for which this was the case were part-funded by concessional finance. This suggests that achieving some forms of impact may not be compatible with a purely commercial model. To maximise development impacts, therefore, public agencies will have to perform different roles in different projects. In many cases, this will be to demonstrate commercial attractiveness, thereby mobilising a future increase in private investment. In other projects, however, maximising impacts may require concessional finance to be deployed, implying that public actors will need to be engaged in the sector for the longer term.
Our sixth finding, therefore, is that projects should be assessed and categorised more systematically. This will enable support from public agencies to be structured in the most appropriate and effective way.
As a first step on this road, we developed the following framework for categorising projects:
i. Fully commercially viable, could go ahead without DFI involvement (4).
ii. Commercially viable but DFI ‘political insurance’ essential to mitigate risks sufficiently to assure investors.
iii. Project commercially viable but only if finance is structured in ways that only DFIs will or can do.
iv. Only commercially viable if ‘blended’ model of concessional and commercial finance is used.
v. Not commercially viable, should be publicly funded.
In our view, DFIs should not be engaged in categories (i) and (v) (although we found some cases where they were). For categories (ii) and (iii), financial additionality is a result of the particular qualities of DFIs, particularly their ability to offer ‘political insurance’ and favourable forms of finance. In these circumstances, we suggest that the ‘premium’ paid for this insurance should be a greater commitment to social and environmental standards by the private investor, as well as commitments on local employment and supply chain linkages. Projects in category (iv) are those where, to ensure access and affordability of the poor, the project is only commercially viable if concessional finance is used in conjunction with private investment. Category (iv) projects are therefore fundamentally different from categories (ii) and (iii).
One option would be for DFIs to undertake category (iv) projects using concessional finance to realise development impacts. They may feel, however, that their primary purpose is to create demonstration effects, which would mean they should restrict themselves to categories (ii) and (iii) (5). In this case, other public agencies would need to step in to fill this gap.
In sum, the evidence studied led us to conclude that there is a set of projects that are neither fully commercially viable nor suited to full public funding. Attempting to ‘shoehorn’ these into either camp is likely to lead to sub-optimal development outcomes, which fail to achieve the commercial success needed to create a positive demonstration effect, or to realise the full set of potential development impacts. There is a real opportunity here for public agencies (DFIs or other) to both mobilise private finance and create developmental impacts through the use of blended finance. However, crucially, the projects in this category must be recognised so that support from public agencies can be structured appropriately. Where this does not happen, scarce public funding will not be put to best use.
The full review is available here: http://www.ids.ac.uk/publication/development-finance-institutions-and-infrastructure-a-systematic-review-of-evidence-for-development-additionality
Lily Ryan-Collins is Infrastructure Adviser, UK Department for International Development, and Stephen Spratt is Research Fellow. Institute for Development Studies.
The views expressed in this paper are those of the authors alone, and do not necessarily reflect the UK Government’s position
1. Foster, V. and Briceño-Garcia, C. (Eds.) (2010) Africa’s Infrastructure: A Time for Transformation, Washington D.C.: World Bank.
2. Estache, A. & Fay, M. (2007) Current Debates on Infrastructure Policy. World Bank Policy Research Working Paper No. 4410. Washington D.C.: World Bank.
3. IFC, KFW, CDC, AsDB and FMO. 12 DFIs were approached to participate. Those not named were not prepared to release their internal documents.
4. In the case of category, (i) projects DFI advisory services can still play a valuable role in mobilising finance for projects that are commercially viable without DFI investment.
5. As well as advisory services in the case of category (i).
This article was published in Great Insights Volume 2, Issue 4 (May-June 2013)