Evidence submission to the International Development Committee’s enquiry on Investment for Development: the UK’s strategy towards development finance institutions

Charles Kenny

Senior Fellow

Center for Global Development

This evidence submission will focus on the question posed by the Chair in her comments on the enquiry: “whether BII delivers impact and value for the UK taxpayer.” I will argue that while BII (British International Investment) has been a leader amongst development finance institutions (DFIs) in important respects, and has a strong staff and management focused on improving development outcomes, it still suffers from constraints that limit the impact of DFIs as a whole in supporting sustainable development, especially in the poorest countries. Like other DFIs, BII is unable to mobilize significant capital by crowding in private finance to its investments, its efforts to create new markets for private finance have had limited success and its use of subsidized capital has been inefficient. BII has some specific issues around leverage and transparency that might be amenable to reform but these larger problems are inherent to the DFI model, and only likely to become more severe given the growing scale of the DFI sector. Given that, future flows of official development assistance (ODA) are best allocated elsewhere.

BII is an innovative DFI

BII is one of the most innovative DFIs worldwide, with a talented and committed staff and management. That talent and innovation is reflected in investments such as Globeleq, an independent power producer created by CDC in 2002 and now owned 70 percent by BII and 30 percent by Norfund. Globeleq is the largest private developer, owner and operator of independent power plants in sub-Saharan Africa. CDC created Globeleq to overcome what it saw as a significant gap in the market in terms of private electricity generation in developing countries. In contrast to the traditional DFI model of waiting for private firms to approach with proposals, this proactive stance, akin to industrial strategy, is more likely to deliver transformative results. Other examples in the BII portfolio include Gridworks financing equity investments in transmission and distribution networks and MedAccess, which supports rollout of medical products with high cost effectiveness in poorer countries.

More broadly, BII has what appear to be comparatively strong internal controls in place to ensure its investments are additional, crowding in better private investment rather than crowding it out, yet still appears able to find sufficient projects in its target markets (focused on the poorest developing countries) to support.[1] As a strong private sector is critical to poverty reduction and economic growth, that suggests the potential for significant development impact. In that regard, it is worth emphasizing that the social returns on BII investments are internally estimated to compare to ODA-financed public investments.[2]

But DFIs including BII have limited mobilization impact

That said, the DFI model is not delivering on its promise. The transformative impact of DFIs is predicated on them crowding in significant private investment to developing countries directly through mobilization or indirectly through demonstration. For example, the Blended Finance Taskforce for the Global Goals in 2015 suggested an additional $1.1–$1.5 trillion in private sector investment in developing country infrastructure could be mobilized a year if $100 billion in aid and multilateral development bank-backed blended finance was leveraged at a 9:1 ratio. UK government ministers have suggested the central role that BII would have in this ‘billions to trillions’ agenda.[3]

But things have not worked out as planned. The recent Global Financing Pact agreed in Paris dramatically ratcheted down mobilization expectations to 1:1 for multilateral development banks as a whole, but this is still above their actual performance.[4] According to analysis by Samantha Attridge and colleagues at ODI, the leverage ratio of BII has historically been a little below the DFI average at around 75 pence for each pound invested and has recently fallen below 40 pence per pound.[5] Not 9:1 but 0.4:1. Note also, standard mobilization calculations are exercises in asserting additionality rather than assessing it.[6] Some of these projects might have gone ahead absent DFI involvement, and in which case mobilization is zero.[7] (A dated estimate for the World Bank’s DFI, the IFC, is that in only 27 percent of IFC projects was IFC investment necessary for the project to go ahead, but this proportion would likely be higher in poorer countries.)[8]

And they have limited demonstration impact

The Globeleq example discussed above suggests another way that DFIs could have outsize impact: by creating or demonstrating a market that is then further developed by the private sector. But the experience of Globeleq itself suggests how complex that can be. Take the Central Térmica De Temane gas power plant in Mozambique. It is majority owned by Globeleq, itself owned by DFIs. Other investors are Mozambique’s state-owned electricity transmission and distribution company, the IFC (with the support of IDA), the U.S. government through the International Development Finance Corporation, and the OPEC Fund for International Development. While the power produced is affordable, and the project likely has a strong development impact, what the investment demonstrates is that two decades after the creation of Globeleq, even most ‘private’ power projects in low-income countries are overwhelmingly financed by the public sector, and much of that through DFIs. A second example is IFC’s Scaling Solar, which set out to demonstrate the financial viability of private solar power plants, but after five years has only delivered projects with majority involvement of DFI and subsidized government finance.[9]

The ‘demonstration’ impact of DFI investment is limited because DFIs aren’t the private sector. Indeed, they are not too different from the Chinese state-owned enterprises involved in the Belt and Road Initiative. The complaint made about those Chinese enterprises is that they enjoy access to subsidized finance combined with the benefits of being a sovereign creditor while protected with contractual provisions and returns designed for private investors. It is this same combination which allows DFIs to offer financing at below market rates while also ensuring low default risk on the deals they are involved in, a combination unavailable to true private investors.[10]

Any demonstration impact is further curtailed by the fact that DFIs as a group are notably less transparent than other aid-funded institutions. BII itself has a score on Publish What You Fund’s DFI Transparency Index that is half that of the IFC, and ranks below the Asian Infrastructure Investment Bank.[11] Deal terms are kept confidential even where they involve contingent liabilities for governments—take-or-pay conditions in electricity contracts, for example.

BII subsidized finance is poorly targeted

The targeting of DFI support is also ill-matched to generate outsize development impact. Like other DFIs, BII effectively subsidizes investments by backing them where its expected returns are considerably below what the market would demand. This is justifiable if subsidies support an investment with strong development outcomes that actually need a subsidy to be viable. But robust empirical evidence that DFIs meet those tests is thin on the ground, because it is hard to run experiments on large investments.[12] That should leave relying on ‘best practice’ which suggests subsidies should be used to further public policy priorities of the country where they are being deployed, open offer or competitive wherever possible and transparent in every case.[13] Instead, BII (like other DFIs) offers bespoke, noncompetitive, nontransparent subsidies based on its own internal decision-making process.

And while subsidies should make DFI finance broadly attractive, in fact the institutions are limited to a small portion of low-income economies in particular because of their need for investment partners which are reasonably large and meet international accounting and ESG standards. Most of the private sector is off limits to direct investment when the informal economy accounts for more than 50 percent of GDP in Nigeria and Tanzania, for example.[14] Again, in 2010 Ethiopia only had 43 firms which employed more than 500 people.[15] This explains a DFI focus on banking and infrastructure, where firms are usually larger and formal-sector (though note the banking projects tend to focus on on-lending to smaller firms).[16] Infrastructure and banking are dominated by state-owned enterprises, however, especially in low-income countries. Eighty-three percent of infrastructure investment in developing countries as a whole is publicly financed. That rises to 95 percent in sub-Saharan Africa.[17] A number of developing countries, including India, see public banks with a 50 percent share of the market or more.[18]

In other sectors where large-scale investment is crucial to meet the development targets, including health and education, the formal private sector is an even smaller player. Again, DFIs cannot focus their subsidy resources on public sector priorities simply because there are not the private sector partners to support in those areas. And efforts like Globeleq suggest that even a proactive market-creating approach by DFIs may do little to change that in the medium term. We are left with the choice between an approach where ODA backs ‘public-private’ investments that are nearly all public made by donor DFIs demanding quasi-market returns, or where ODA backs public investments by recipient countries using loans that require (lower) quasi-market returns. Given the second is considerably more scalable, transparent and financially sustainable, as well as (recipient) country-owned, it seems like the more attractive option.

DFI subsidies are all too often allocated simply on the basis that there is a project to be financed and a considerable supply of financing that needs to be deployed, then. And a growing number of DFIs with ever more money are chasing these projects, especially in the poorest countries. A 2019 estimate suggested the DFI resources available for investment in low-income countries and lower-middle-income countries from the IFC, the US DFC and European DFIs was forecast to rise from about $67 billion to $146 billion between 2019 and the mid-2020s.[19] That helps explain why so many DFIs pile into the same deals, and significantly increases the risk of DFIs competing with each other on the basis of how large a subsidy they can offer. Providing additional finance to BII will only exacerbate this problem.

BII is also unable to leverage resources by borrowing

BII specifically faces an additional challenge to impact in that it is unable to leverage its capital. As of the 2021 accounts, BII had paid-in capital of £4.7 billion. Under today’s DAC, rules that would all count as ODA. In addition, it had retained earnings from past profitable investments worth £3 billion. These resources enabled a portfolio of outstanding investments worth £5.9 billion and disbursements of £1.3 billion. Effectively, one pound of capital and retained earnings gets you 76 pence of portfolio and 17 pence of disbursements. Put in an additional pound of ODA into BII next year, and that is about the ratio you would expect in terms of portfolio and disbursement expansion over the medium term.

Compare the World Bank Group’s development finance institution, the IFC, which operates more like a traditional bank than BII, in that it leverages its capital by borrowing on markets. That means each pound of capital and retained earnings—or new ODA—goes further. At the IFC, one pound of capital and retained earnings gets you £1.35 of private investment (almost double the level at BII) and 40 pence of disbursements (more than double the level at BII).[20] Note there are two important differences: First, BII primarily invests equity rather than debt while the position is reversed in the IFC, and equity can be used to leverage financing by the company receiving the investment. Second, BII is limited to investing in low- and lower-middle-income countries while the IFC only manages to funnel about half of its investments into those countries,[21] and IFC investments in low-income countries are about 8 percent compared to 12 percent for BII.[22] Nonetheless, that suggests more direct investment in low-income countries from a pound of ODA dedicated to the IFC than to BII.

All of this suggests uses of ODA are likely to generate greater development impact

BII’s capacity for innovation as demonstrated by Globeleq is admirable—if it moved toward supporting client country industrial strategy in creating markets where that was a public policy priority, that innovation could have a greater return. If it were given borrowing authority, BII’s relative standing compared to other DFIs in terms of investments per pound of ODA might improve (hopefully not at the cost of risk appetite). And transparency reforms (particularly aimed at transparency around financing terms) could help it set the pace in an opaque industry. All of that is possible with its existing capital base. Meanwhile, other DFIs could learn from BII’s market-making ventures and attention to additionality. But the DFI community as a whole is struggling to deliver the kind of results that previous capital infusions were predicated upon, and the problem now is not insufficient money but insufficient ways to spend it well. That is a fundamental constraint on BII’s impact.

Compare the potential impact of providing capital or hybrid capital investments to the public sector arms of multilateral development banks—a live topic at the moment. This would create far higher returns in terms of quasi-market rate development finance. One pound of capital and retained earnings gets you £4.32 of portfolio from the IBRD or 54 pence of annual disbursements, five times the investment per dollar of ODA than provided by BII. That public finance can support the education, health and infrastructure investments which are overwhelmingly provided by governments and that are required to meet climate targets and the Sustainable Development Goals. Such loans are a far more flexible, scaleable and efficient use of ODA resources.

It is true that the IBRD doesn’t invest in low-income countries or those at the bottom end of the lower-middle-income scale. That said, the top five countries BII invests in, collectively accounting for 49 percent of the BII portfolio, are all IBRD borrowers, so there is considerable overlap.[23] The UK currently has a 4 percent shareholding at the IBRD. If it had contributed the £4.7 billion it has given to BII as part of a general capital increase to the World Bank instead, it would be directly supporting over £20 billion pounds of IBRD-backed investments, and it would have helped mobilize over £500 billion in extra public lending backed by other shareholders. That is surely a better deal for the UK taxpayer.

And if the choice is between BII and grant-funded investments focused on low-income countries, World Bank IDA contributions are a far more flexible vehicle to meet SDG investment needs, as is FCDO-managed ODA, where 58 percent of spending is in low-income countries, as compared to 12 percent for BII. FCDO might also want to consider the judicious use of quasi-market lending to low- and lower-middle-income sovereign and sub-sovereign entities for greater flexibility.

The UK’s DFI has a role to play in supporting private investment in the world’s poorest countries, but it operates in a sector which already sees too much money chasing too few investments. Additional aid finance is far better spent on BII than diverted to ‘ghost ODA’ spending within the UK, and doubtless has a higher impact than some FCDO-managed spending abroad, but compared to the most efficient alternate uses, it will likely fall short.