The Stretch Fund



When the world adopted the SDGs, policymakers knew that aid alone would never meet the financing needs. They embraced the “billions to trillions” vision, believing that an abundance of commercially viable SDG-related investments was ready and waiting for trillions in profitable private investment—if only development finance institutions (DFIs) and others could clear away the obstacles that stand between the investments and private investors. Reality looks different. Finding bankable projects, especially in low-income countries, is hard for commercial investors and hard for DFIs, essentially for the same reasons. The DFI view of risk, tolerance for risk, risk management approaches, and goals for risk-adjusted returns are not greatly different from those of commercial institutions.

The result is a critical gap in the architecture for development finance. Grantmakers fund non-financially sustainable but high-development-impact activities. And commercial impact investors, including DFIs, seek market returns along with impact. We need a new public-private actor in between, one that has a very different risk tolerance and financial objective, and an emphasis on development impact. We propose the Stretch Fund to partner with DFIs in ways that stretch their range of investments, spectrum of clients, scope of investible markets, and scale and mobilization of finance.


When the world agreed to the Sustainable Development Goals (SDGs) in 2015, they came with an enormous price tag. It quickly became clear that meeting the ambitious SDG targets would require a shift in aid strategy. Aid itself could not fill finance gaps. A larger share of the billions of dollars in official development assistance would have to be used to catalyze trillions in new financing, notably from private sources.

The “billions to trillions” vision rested on the assumption that an abundance of commercially viable SDG-related investments would be ready and waiting for profitable private investment. This vision naturally trained the spotlight on development finance institutions (DFIs), both multilateral and bilateral, as the necessary intermediaries, whose mission is to clear away the market failures that stand between the projects and private investors. Market failures in this context could mean private sector misperceptions of risk-adjusted returns. Or, more likely, they could be real risks that DFIs could help share but also mitigate through support for innovations in business models or technologies, closing information and skill gaps, building market infrastructure, and developing local capital markets.

Reality looks very different. Especially in low-income countries, DFIs find it hard to develop projects that will pass their own credit and investment committees. Despite their broad toolkits, DFIs struggle like their commercial counterparts to overcome market failures. In fact, the DFI view of risk, tolerance for risk, and risk management approaches are not greatly different from those of commercial investors. And both seek market risk-adjusted returns—understandable given the interests of DFI shareholders in maintaining profitability and institutional ratings. But these same shareholders criticize DFIs for meager results in development impact and mobilizing private investment.

Stakeholders are calling on DFIs to play a much larger role in SDG finance.

Stakeholders are calling on DFIs to play a much larger role in SDG finance. But they are also asking a more fundamental question: Given widespread aspirations for more development impact from publicly funded DFIs, should DFIs target the same risk-adjusted commercial returns as private banks and investors?

One answer to that question could be fundamental changes in DFI financial models to redefine their financial performance goals. Instead, we propose an approach that adds an off-balance sheet partner to help DFIs boost their impact while effectively managing increased risk.

The Missing Piece

We would argue that DFI intermediation and mobilization are not working effectively or at sufficient scale because a critical piece is missing from the architecture for development finance, which affects both public and private actors. At present, there are grant makers (public and private) that fund non-financially sustainable activities with high development impact, and there are impact investors (public and private) that mostly seek risk-adjusted market returns along with impact. But there is a gap in between. What is missing are investors that target investments at scale with sub-market risk-adjusted returns and high development impact. (Mission-related investments by philanthropic investors are still very small in volume.)

This gap in financial actors is reflected in three persistent gaps in capital markets that neither DFIs nor private investors typically fill:

  1. early stage finance for firms and infrastructure;
  2. the highest risk project tranches (referred to as the top of the capital stack); and
  3. local currency products and services.

Early stage finance for firms is needed after funding from friends, family, and/or angel investors, when the firm has a revenue track record and may be at or approaching the break-even point, and external growth capital is essential to move towards sustainable profitability. This is a universal problem for young firms, not a sectoral problem. The gap affects small and medium-sized enterprises (SMEs) in general, but it is perhaps especially binding for women-owned firms and for first-mover firms (those introducing new technologies or business models, including inclusive or green models) where there are large uncertainties on both the demand and the supply side. For infrastructure, finance is needed in the preoperational construction stage, before revenue streams begin and risks are concentrated. For SMEs engaged in infrastructure, such as off-grid energy and water solutions, the gaps are compounded.

Financing high-risk tranches requires a set of subordinated tools that DFIs now deploy very sparingly—guarantees, various forms of equity, and junior debt. Loans, especially senior loans, still dominate the operations of multilateral DFIs. For 2017, loans accounted for an average of 85 percent of the annual commitments of the five major multilateral DFIs (the African Development Bank, the Asian Development Bank, the European Bank for Reconstruction and Development, IDB Invest, and the International Finance Corporation). Equity (mostly later stage equity) averaged 8 percent, and guarantees 9 percent.[1]

Regarding the third gap, the local currency share of DFI operations remains very small, as do investments in local corporate bonds and hedging and insurance products.

A New Public-Private Actor

Filling these gaps requires an entity capable of taking on more risk and deploying a different mix of tools for managing these risks. Its overriding role would be to stretch the capital of existing DFIs in two ways: (1) expanding the spectrum of investments in which DFIs can participate; and (2) taking on high-risk tranches to open up more DFI investment opportunities. Hence, our proposed name: the “Stretch Fund.”

What would the Stretch Fund look like? As described below, it would look very different from existing DFIs and their donor trust fund partners.

Focus on gaps. The Stretch Fund’s investment strategy would target the capital market gaps described above—not countries, regions, or sectors. As capital market development tends to increase with country income levels, it follows that low-income and lower-middle-income countries would be the principal investment destinations.

Range of financial instruments. In contrast to the dominance of senior lending in existing DFI portfolios, the Stretch Fund would deploy mostly subordinated products: equity (including early stage equity and quasi-equity), guarantees and first loss protection, and subordinated debt, including in local currency.

Human capital. Stretch Fund investment officers and managers would specialize in deploying these instruments, which require different skills than standard senior lending in dollars or euros. The differences in skills and in portfolio management objectives, in fact, are a principal reason that DFIs find it so difficult to shift significantly away from their existing instrument mix.

Deal origination. Most of the investment origination would come from DFIs interested in partnering with the Stretch Fund in order to access its risk-tolerant capital. That would allow the Stretch Fund to capitalize on the reach and on-the-ground presence of existing DFIs in building deal flow. This partnership approach with multiple DFIs (bilateral and multilateral) and with project development funds would be essential to keep the Stretch Fund’s overhead manageable. DFIs could compete for access to the Stretch Fund to promote a race to the top in investment impact. Or high-impact investments at scale proposed collectively by multiple DFIs could receive preferential access.

But, unlike most donor trust and guarantee funds, the Stretch Fund would also have the mandate to originate a limited number of projects on the riskier, more innovative end of the spectrum. This is important for reaching investees that DFIs would normally not touch and for expanding the boundaries of market-making impact. Rather than operating solely as passive project takers, staff would be encouraged and empowered to develop some emblematic investments with exceptional impact and demonstration effects.

Financial goal. The Stretch Fund would be financially sustainable. It would preserve capital at the portfolio level, cover its administrative costs, and not require regular replenishments. Its risk-adjusted returns at the portfolio level would be below-market or zero. Any returns would be retained and added to capital for expanding operations.

The Stretch Fund would look very different from existing DFIs and their donor trust fund partners.

Mobilization. To play a meaningful role in mobilizing private finance for development, growth, and the SDGs, the Stretch Fund would catalyze multiple dollars of finance from other investors for every dollar it commits. It would invest by targeting an unfunded risk tranche or catalytic part of the capital stack, rather than as a stand-alone investor. The aim would be to mobilize both private investment and investment from DFIs. The composition of private versus DFI co-investment would vary based on country and transaction risk, as well as impact.

Permanent capital vehicle. The Stretch Fund would be a permanent capital vehicle with capital contributions in the form of equity. This would enable the fund to play the patient investor role that is so often missing and so critical to unlocking commercial finance.

Public-private fund. The Stretch Fund would be a pooled investment vehicle, serving as a platform for aggregating capital. It would combine capital from public and private investors that are like-minded in prioritizing development impact over returns, such as governments and risk-tolerant philanthropic and foundation investors. The combination would be mutually beneficial. Governments bring experience and emphasis on accountability; transparency; and environmental, social, and governance standards. Private actors bring strengths in innovation, market adaptation, and efficiency. All shareholders in the fund would participate in governance commensurate with their equity share.

Impact. The emphasis on enhanced development impact requires the Stretch Fund to have a strong results measurement framework and equally strong reporting system. Consistent with the investment strategy and rationale, results measurement would need to assess impact in making and building markets, as well as impact on specific beneficiaries (individuals, firms, and farms). It would make sense for the Stretch Fund to adopt an existing best practice measurement framework, such as the IFC’s Anticipated Impact Measurement and Monitoring system (AIMM),[2] rather than expend the time and resources necessary to invent its own framework from scratch.

Challenges and Risks

The Stretch Fund differs in significant ways from existing structures, and there are undeniable uncertainties in assessing its chances for success. Among the most important is financial sustainability: Can a structure with such an investment strategy preserve its capital at the portfolio level? The analysis below directly addresses that question using the track records of the closest comparators we could find and erring on the side of conservative assumptions with respect to risk and returns.

Can a structure with such an investment strategy preserve its capital at the portfolio level?

Equally important, but often neglected, is the question of human capital and institutional culture. The Stretch Fund’s core would be centered around subordinated products. And it would have the advantage of building this focus and skill base into its institutional DNA, rather than as an add-on to an institution dominated by a senior lending culture. Staff performance incentives would be aligned with the impact mission and a tolerance for a portfolio that includes financial underperformers. Fostering the proper culture, recruiting staff with unique skillsets, and aligning incentives not widely present in the current landscape would present significant challenges.

One other major risk is donor fatigue and skepticism regarding new entities. This fatigue comes despite donor interest in using their aid resources more catalytically. Raising capital for the Stretch Fund would not be easy. And creating an entirely new fund or institution would be a heavy lift in the current donor environment.

The good news is that the amount of capital needed for this purpose would be relatively small. The Stretch Fund’s purpose is to be as catalytic as possible—focusing not on maximizing the volume of its own commitments but rather on maximizing space for others to invest. The position it would occupy in the capital stack would limit the volume of capital needed. The analysis below proposes a size large enough for efficiency and global scope, but small with respect to the individual capital contributions needed from shareholders.

It would be desirable to build on an existing fund that is most similar in terms of strategy and attributes to create the Stretch Fund. This would give this effort a head start in human capital, a project pipeline, shareholder confidence, and leadership.

Stretch Fund Size, Structure, and Instruments

To help us shape the specifics of the financial structure and attributes of the Stretch Fund, we looked at comparable investment vehicles designed for development purposes, interviewed professionals managing funds with similar characteristics, and assessed the scale necessary to achieve the Stretch Fund’s envisioned impact. Based on the evidence provided from comparable vehicles, as well as the distinctive features that we wanted to build into the Stretch Fund, we constructed a financial model that allowed us to test a variety of parameter settings with respect to size, instrument mix, investment terms, capital structure, and so on to design a fund that would combine financial sustainability with high risk tolerance and high development impact. (In cases where we were provided the relevant data on a confidential basis, we have not revealed the identities of the comparable vehicles.)

Size of the Stretch Fund and its Investments

The Stretch Fund would be a pooled investment vehicle. Such vehicles are generally defined as funds or facilities in which public and/or private investors interested in development impact, and sometimes commercial investors, pool their capital into a collective investment entity.

There are a considerable number of pooled investment vehicles currently operating. Their advantages for achieving scale and sharing risk give them growing appeal as mechanisms for development finance, as highlighted in separate reports by the World Economic Forum in 2016[3] and Convergence in 2018.[4] The Organisation for Economic Co-operation and Development (OECD) published a flagship book[5] on blended finance in 2018, which reported on the results of a survey conducted in partnership with the Association of bilateral European Development Finance Institutions of facilities and funds created since 2000. In this context, the OECD defines “facilities” as vehicles comprised only of concessional capital (public and private) designed to blend with commercial finance as part of the investment process; it defines “funds” as a blend of donor and commercial finance in the vehicle itself. In aggregate, the survey found the average pooled vehicle managed $400 million in assets. Considering just facilities, the pooling mechanism more closely aligned to the Stretch Fund, the average size increased to $471 million.

Discussions with professionals at individual funds most similar in function and scope to the Stretch Fund provided another valuable window into determining the right vehicle size. Interviews indicated $100 million is the absolute minimum viable amount for a vehicle of this nature when considering the various factors related to financial sustainability, efficiency, and cost per dollar invested. An initial size of $200 million was regarded as still conservative, but consistent with the investment pipeline expected by one comparable fund currently in the process of fundraising. The initial capital could be augmented by successive capital infusions as the Stretch Fund develops a track record.

Investment size and expected deal flow are an important consideration given the distinct features and purpose of the Stretch Fund. The transaction sizes of investment targets are intentionally smaller than many private investors, DFIs, and multilateral development banks (MDBs) would even consider. The Stretch Fund would fill a critical financing gap in smaller, higher-risk investment propositions that can have an amplified development impact. Therefore, the Stretch Fund would focus on transactions under $10 million, with a substantial share of activity involving deals in the $2 to $4 million range.

Deal flow would be relatively heavy compared to vehicles of a similar size. Assuming an average deal size of $3 million, it is envisioned that staff would execute 24 new deals per year on average, and the Stretch Fund would manage a portfolio of approximately 167 active investments and guarantees once it has reached its capacity. This would likely lead to higher administrative and management costs than similarly sized facilities. Interviews with fund managers suggest that a vehicle of this nature would incur an overhead rate of approximately 2.5 to 3.0 percent of assets under management. Our model conservatively assumes overhead costs will be above this range at 3.5 percent, considering the potential for higher transaction costs related to the Stretch Fund’s investment goals.

Combining all these factors leads us to propose a fund on the larger side, aligned with the average pooled facility size reported by the OECD. This would help achieve a global and diversified distribution of assets (especially important in managing high levels of risk), enable a broader array of products in the investment mix, and achieve an investment capacity large enough to co-invest with the gamut of DFIs and MDBs globally.

Further, it is important to be realistic about the willingness of donors to provide ongoing infusions of concessional finance. A higher initial fundraising effort would avoid the need to obtain a second capital infusion soon after the first.

Accordingly, we propose that the targeted size of the Stretch Fund be about $500 million, with capital infusions of varying sizes from perhaps 10 donor governments and philanthropic organizations.

Instrument Mix

The Stretch Fund would provide a range of subordinated investment instruments necessary to accommodate the diverse product needs of its investees. These would include equity, quasi-equity, subordinated debt, subordinated loans and loans in high-risk contexts, and guarantees. Further, it would not be limited to US dollar and euro investments: a sizeable portion of its investments and guarantees would be made in local currency. The associated terms and conditions (i.e., pricing) and instrument mix of the portfolio would be managed to yield sufficient returns to offset losses and cover overhead, promote firm growth and graduation from dependence on subsidies, crowd in commercial capital, and preserve its capital.

The envisioned instrument mix is based on financial modeling as well as interviews with practitioners who work principally in debt, equity, and guarantees in the size, currency mix, and risk categories targeted by the Stretch Fund. Their strong development mandate meant they were taking outsized risk for the level of income they expected in return. In general, we found debt had the best risk-adjusted financial performance followed by guarantees and then equity.


Market data from developed markets show that equity investors are compensated with higher average returns for the increased risks they face. However, the study of comparable funds with a development focus indicated equity does not have a strong record when evaluating risk adjusted returns alone. The research found that early stage equity investments in many developing markets are slow to unwind and often yield inferior annualized financial returns when compared to debt. For example, one of the funds studied with over two decades of experience recovered $0.89 for every $1 invested in equity on average, with an average exit of 10 years after the initial investment. Interviews of professionals at another fund with an equity portfolio most comparable to the one envisioned for the Stretch Fund achieved an average exit multiple of 1.6x with an average duration of eight years. However, once including losses incurred by equity deals failing to unwind and changes in exchange rates, the multiple dropped to 1.3x. From a return perspective, this amounted to a 3.3 percent compound annual growth rate. DFIs pursuing later stage equity investments do better, earning more than twice that amount on average.[6]

Equity investments would mostly be in the form of investments in early stage or high-impact funds, rather than in individual companies, as the Stretch Fund would not have the on-the-ground presence needed for individual company due diligence. The Stretch Fund would not make grants, but it would provide, where warranted, more concessional forms of equity similar to reimbursable grants, recovering in some cases only a portion of its initial investment when exiting. This portion of equity investments would dampen the overall performance of equity investments in the portfolio. As a result, our model conservatively assumes that the equity multiple would be moderately lower than the most comparable fund and significantly lower than DFIs, amounting to a 1.1x exit multiple with an average exit of eight years earning 1.2 percent per annum on average.


The research clearly indicates that debt has the highest net returns per unit of investment in this context. This is true despite the fact that the funds analyzed deployed subordinated debt, lent to entities DFIs would generally reject as not creditworthy, and did some of their lending in local currency. Furthermore, debt provides a predictable and steady income stream important to the sustainability of the Stretch Fund, as well as a critical capital source for many firms and projects.

Interestingly, funds comparable to the Stretch Fund showed lending results that were positive despite the high risk. Average returns earned from their portfolios ranged from around 6.5 to 12 percent—in all cases below the prevailing local rates for loans or bonds for these risk categories. The average loss rates for the portfolios ranged from 5 to 10 percent, which included losses related to depreciation of the local currency portion of the portfolio relative to the US dollar, the British pound, or the euro. The average maturities for the portfolios ranged from four to five years yielding a relatively strong return for the fund managers when compared to the concessional nature of the rates and the risk associated with the investments.

Our model incorporates the experience of practitioners when determining the average yield, loss rate, and maturity of the debt portfolio. Ultimately, we took a relatively conservative approach, assuming an average yield of 9 percent and a 10 percent loss rate to account for the concessional nature of the Stretch Fund, risk profile of its investments, and high proportion of local currency investments. We assumed a slightly higher average maturity of six years, reflecting the Stretch Fund’s goal of offering more patient debt capital than the comparators we studied.


Several appealing attributes make guarantees an essential part of the instrument mix. Reserves are held at a fraction of the guaranteed amount while premiums are based on the total amount guaranteed. Further, reserves can be invested in highly rated liquid securities, yielding additional income. Most importantly, guarantees have a strong track record of being an effective mobilizer of private capital.

The Stretch Fund would provide a range of subordinated investment instruments necessary to accommodate the diverse product needs of its investees.

Several guarantee programs were studied to determine the portfolio allocation to guarantees and model its investment returns. The Development Credit Authority (DCA) at the U.S. Agency for International Development (USAID) proved to be the most instructive given its high level of transparency related to size, pricing, annual loss rates, and reserves. Its typical guarantee had an average size of $9 million and maturity of seven years. The primary product is a 50 percent portfolio guarantee whereby any principal losses that qualified under the terms of the guarantee would be shared equally with the local banks or investors. On average, the agency set aside 8 percent of the guaranteed amount to be utilized in the event of claims by any active guaranteed transaction. The professionals at DCA described it as a product with a 24.3 to 1 mobilization ratio—for every dollar it allocates, it mobilizes $24.30. Fees are assessed at two levels: origination and utilization. The origination fee is a one-time cost at closing ranging from 1 to 2 percent of the total guaranteed amount. The utilization fee is an annual premium ranging from 1 to 2 percent assessed on the guarantee amount being utilized. DCA reports an historical cumulative default rate of 2.94 percent. Its more recent performance shows it paid claims totaling close to $6 million in 2018 and $8 million in 2019 compared to active guarantees of approximately $300 million. This is equivalent to an annual loss rate of 2 percent in 2018 and 2.67 percent in 2019 on the active portfolio. Overall, DCA breaks even; however, this is largely due to concessional pricing and an inability to earn income from reserves it sets aside.

SIDA, Sweden’s agency for development cooperation, also operates a guarantee program. We studied its recent portfolio review, which highlights a broad array of guarantee types—portfolio, balance sheet, project finance, and volume—covering credit or political risks often in partnership with other development agencies, multilaterals, funds, and foundations. It has paid claims of SEK 8 million (approximately $820,000) from the beginning of 2015 through the end of 2018 versus SEK 7.1 billion (approximately $728 million) in total guarantee volume resulting in an overall loss rate of just 0.1 percent.[7] However, more detailed information on risk tolerance, pricing, and reserve practices is not publicly available.

Another comparable guarantee program is part of the European Fund for Sustainable Development (EFSD). In July 2018 the European Union allocated €800 million to the EFSD for guarantees with an anticipated mobilization of €8-9 billion in public and private investments indicating roughly a 10:1 mobilization ratio.[8] The guarantee program run by the European Investment Fund employs a more varied mix of guarantees including first loss, second loss, and shared first loss largely targeted at SMEs. Coverage ranges from 34 to 81 percent of the principal amount and lasts from 10 to 15 years. Their fees are calculated relative to the risk and designed to be self-sustaining in aggregate. Their fee structure also has an up-front element and an ongoing annual amount.[9] However, figures such as average fees, losses, and reserve rates are not publicly available. Even if they were, these would prove less instructive as the guarantees are for European SMEs.

We assumed in our modeling, conservatively, that guarantees would have similar pricing, reserve requirements, and expected losses as DCA, given its higher level of disclosure and a risk appetite more in line with the envisioned Stretch Fund. As a result, the model assumes 8 percent of the guaranteed amount would be held in reserves, seven-year average maturity, and 1.5 percent origination fee. Slight adjustments were made to the loss rate, ongoing pricing, and reserve assumptions. First, the annual loss rate was assumed to be 2.5 percent to reflect a higher risk profile. Second, the model assumes annual premiums of 2.5 percent to balance expected losses and ensure capital preservation. Third, reserves would be invested in a mix of highly rated liquid fixed income securities and US Treasuries yielding 3 percent on average. Finally, the allocation to guarantees was influenced by the fact that in a very rare event, losses could exceed reserves. As a result, care was taken to ensure the aggregate guaranteed amount would never exceed the total value of the Stretch Fund’s assets.

Proposed Portfolio

In addition to the product-level assumptions described above, we made the following general assumptions when modeling and developing the product mix:

  1. The capital structure of the vehicle would be entirely equity without any dividend payments.
  2. The portfolio mix would be periodically rebalanced as income is earned and losses are incurred.
  3. While the Stretch Fund would be created with the flexibility to use a variety of products, only debt, equity, and guarantees were modeled.
  4. The portfolio breakdown was fine-tuned to ensure there was adequate capacity of each product—debt, equity, and guarantees—to best serve the purpose of the Stretch Fund and that the returns were adequate to cover overhead costs and buffer unexpected losses.

In the end, the confluence of research, modeling, and desired impact leads us to propose an instrument mix of roughly 55 percent in equity, 35 percent in debt, and 10 percent in guarantees.

Capital Structure and Portfolio Metrics

Capital preservation—not producing market-level returns—would be the financial objective of the Stretch Fund. This entails a mindset acutely focused on development impact balanced by the need to ensure that investment returns are sufficient to cover losses and overhead costs. The critical benefit of this approach is that it empowers the investment officers of the Stretch Fund to fill demonstrable investment gaps in order to unlock the flow of commercial capital, rather than attempt to occupy the same space as commercial investors. But this risk tolerance has important implications for Stretch Fund capitalization.

Pooled investment vehicles have various capital structures. They can be capitalized entirely with equity or have a debt investment ranging from a very small proportion to a vehicle almost entirely financed by debt. They can also employ different classes of equity and/or debt arranged in a subordinated fashion.

Capital preservation, not producing market-level returns, would be the financial objective of the Stretch Fund.

One comparator fund uses a tranched debt structure made up of a non-interest bearing junior tranche funded by donors, a low-interest mezzanine tranche funded by foundations or impact investors who would benefit from some level of guarantee by a bilateral development agency, and a market rate senior tranche provided by a DFI. The benefits of this structure are clear—it expands the investment pool and has a very low weighted average cost of capital. Conversely, it is not as permanent as equity, imposes a cost of capital, and retains refinancing risks.

Another possibility is a structured fund with low levels of debt and different classes of shares issued to equity investors. Both the Green Growth Fund[10] and Microfinance Enhancement Facility (MEF)[11] employ this capital structure. For example, MEF issues notes to private investors who have a substantial cushion provided by equity investors to absorb losses. The equity has three classes of shares and an additional junior targeted share class. This structure further expands the investor pool, increasing the capital provided by private investors and DFIs as well as lessening the demands for regular debt payments. However, the majority of investors are expecting commercial returns, limiting the fund’s risk tolerance.

We have concluded an equity-only structure would be the most permanent in nature and allow the Stretch Fund to take greater risks, as well as diminish the stress found in debt funding models that have regular interest payments and risks associated with rolling over notes at maturity. The equity-only structure is also better aligned with the capital preservation goal of the Stretch Fund given that portfolio losses would be replenished more quickly by returns generated by the portfolio than in a predominately debt structure. This would further enhance the Stretch Fund’s ability to make a greater share of equity and quasi-equity investments than it would in a different capital structure.

Moreover, the Stretch Fund could avoid a hasty rush to invest capital given that it would not have the need to generate early returns to meet semi-annual interest payments. This would allow staff to be patient to choose the most appropriate investments with maximum mobilization effect and perform more thoughtful due diligence. It further allows fund managers to take a long view on investments and weather periodic losses without the fear of defaulting on a debt payment or requiring cash reserves (excluding guarantee reserves) as a buffer. Good examples of this type of arrangement include facilities established by the Netherlands and managed by its development bank FMO such as MASSIFF[12] and Building Prospects[13] (formerly known as the Infrastructure Development Fund or IDF).

The primary drawback to an all-equity, no-dividend structure is its limited appeal to investors that want some revenue stream. But our sense is that, for the potential investors to be targeted, revenue streams are less important than development impact, mobilization performance, and avoidance of frequent requests for more capital.

The Stretch Fund would be designed to be as transparent as possible, communicating its impact at the core of its reporting. As a result, using accepted DFI methodologies, the Stretch Fund would report on:

  1. impact on investee beneficiaries;
  2. positive spillover effects on markets and on other actors;
  3. transaction/deal-level direct mobilization. This measure would separate mobilization by investor type, financial product, and whether the additional funds were provided in local or foreign currency.
  4. long-term mobilization. The Stretch Fund would track future financing to measure the catalytic effect of its foundational investments.

The financial performance of the Stretch Fund and its portfolio of investments would also be measured and published ensuring that:

  1. income and losses are adequately balanced to fund overhead costs each year;
  2. capital preservation is maintained each year with gains being treated as capital infusions and significant losses in value requiring adjustments in risk appetite to realign new investment;
  3. portfolio mix is properly allocated to maximize impact; the instrument mix would be evaluated every three years to assess consistency with Stretch Fund objectives; and
  4. sufficient shareholder capital is invested in highly rated liquid fixed income securities and investment funds properly aligned with the risks associated with active guarantees.

Each of the measures would be assessed and measured by the management team of the Stretch Fund and audited by an independent body.

The Stretch Fund would maintain a web portal and post annual results on financial performance, portfolio size, investment activity, and impact.

Mobilization Goals

Measuring mobilization is difficult given the multivariable uncontrolled setting of development finance and the challenge in determining the extent to which public funds actually catalyzed the flow of private finance that would not otherwise have happened. Further, it is difficult to tell if finance was simply redirected from other important sectors, creating a financing gap or higher costs as a result.

Data in recent studies indicate only limited success at mobilizing private finance or attracting institutional investors. A study by Convergence in 2018[14] identified over $100 billion in blended finance transactions from 2005 through mid-2018 and found only 11 percent of the total derived from commercial banks, asset managers, and insurance companies. Another 2018 report by the Blended Finance Taskforce[15] found that MDB private sector operations were only able to muster a 1.5:1 mobilization ratio (private to public) with examples of DFIs ranging from 0.6:1 to 2.6:1. Moreover, a 2019 report by the Overseas Development Institute (ODI)[16] showed that every $1 of MDB and DFI resources invested mobilized private finance of just $0.37 in low-income countries, $1.06 in lower-middle-income countries, and $0.65 in upper-middle-income countries. It also found blended finance concentrated in middle-income countries in infrastructure, banking, and financial services, and very little going to social sectors such as health, education, and social protection.

It would be hard to make the case to donors for funding this new entity unless the Stretch Fund does a better job mobilizing private finance. Yet, the transaction and other costs of riskier investments tend to drive down mobilization ratios. Fortunately, by focusing on subordinated, catalytic instruments like equity and guarantees, higher mobilization is built into the Stretch Fund model.

In order to find a realistic benchmark for improving on current DFI performance, we interviewed professionals at a fund making debt and equity investments with similar risk characteristics in developing countries with mobilization as one of its core goals. On average, its portfolio of debt and equity averaged a 4:1 mobilization ratio without a significant difference between debt and equity instruments.

As noted earlier, guarantees have a much higher mobilization effect. Research on guarantees indicated a mobilization ratio ranging from 10:1 to 24.3:1. The Private Infrastructure Development Group (PIDG) provides a useful, broadly comparable datapoint. PIDG’s 2017 annual review[17] reported it achieved a 17:1 ratio of private sector funding to public investment, largely from guarantees.

Ultimately, we assumed the Stretch Fund would likely achieve a similar 4:1 mobilization ratio for debt and equity as the comparable fund and a 15:1 mobilization ratio for guarantees, reflecting a more conservative expectation than some of the comparators we found in our research. Applying the expected portfolio mix, it is anticipated that the Stretch Fund would achieve a 5:1 mobilization ratio—far exceeding the current results achieved by DFIs and MDBs, yet grounded in the experience of entities with comparable development aims and instrument mixes.

Beyond Business as Usual

It has now been four years since the world adopted the Sustainable Development Goals and the Addis Ababa Action Agenda. The vision of “billions to trillions” is not materializing. It is time to ask ourselves if more of the same will change this finance trajectory.

The Stretch Fund would help stretch the range of investments, the spectrum of clients, the scope of markets, and the scale of finance provided by the DFIs.

While the proposed Stretch Fund would be just one mechanism among many needed to achieve the SDGs, it presents an opportunity to change the dynamic of low mobilization and limited development impact that constrain the contribution of DFIs. Its investments would help bridge very real gaps that block the flow of private finance. It would help stretch the range of investments, the spectrum of clients, the scope of markets, and the scale of finance provided by DFIs. Rather than focusing only on periodic capital increases for existing DFIs, shareholders should consider putting some of their capital into this new approach. Private foundations and other philanthropic investors interested in shaping a new vehicle for much greater scale and impact than they can achieve individually should do the same.

See the PDF for financial model assumptions and an example of the kind of investment that the Stretch Fund would make.


[1] Lee, N. and Sami, A., “Still Lending (Mostly) After All These Years,” Center for Global Development, Blog, February 2019,

[3] “Insights from Blended Finance Investment Vehicles & Facilities,” OECD and World Economic Forum, January 2016,

[4] “The State of Blended Finance,” Convergence, 2018,

[5] “Making Blended Finance Work for the Sustainable Development Goals,” OECD, January 2018,

[6] “Comparative Study of Equity Investing in Development Finance Institutions (DFIs),” Inter-American Development Bank, March 2017, The study found average annual returns on equity of 7 to 10 percent for four DFIs (the Inter-American Investment Corporation; the International Finance Corporation; the European Bank for Reconstruction and Development; and FMO, the Dutch development bank).

[7] “Guarantee Portfolio per 31 December 2018,” SIDA, May 2019.

[9] Chatzouz, M., Gereben, À., Lang, F. and Torfs, W., “Credit Guarantee Schemes for SME Lending in Western Europe,” European Investment Fund and European Investment Bank, EIF Research & Market Analysis, Working Paper 2017/42, June 2017,

[14] Convergence, “The State of Blended Finance 2018”

[15] “Better Finance Better World,” Blended Finance Taskforce, 2018,

[16] Attridge, S. and Engen, L., “Blended Finance in the Poorest Countries: The Need for a Better Approach,” ODI, April 2019,


Rights & Permissions

You may use and disseminate CGD’s publications under these conditions.