This blog is authored by our Head of Development Economics, Paddy Carter. These are personal views and do not necessarily represent the views of British International Investment.
Guarantees are an extremely useful tool, when risk is the barrier to investment. In abstract, investors care about risk-adjusted returns so risk should not present a problem when sufficiently compensated by returns. Nothing would be gained by sacrificing returns to reduce risk, leaving risk-adjusted returns unchanged. In practice, the level of risk can be what prevents investment even if expected returns seem commensurate, especially for banks and other regulated institutions where the capacity to take risk is rationed.[1]
But there are also arguments that in general it would be more effective for development finance institutions to guarantee loans rather than make loans. This article considers whether and how guarantees could be more effective than loans by reducing the cost of capital and mobilising private investment.
Reducing the cost of capital
When borrowers and guarantors assess and price risks in the same way, guarantees do not change the cost of capital. If you start with a risky loan where lenders would want 10% and change that to a 5% low-risk loan and 5% fee for a credit guarantee, you have got nowhere.
Guarantees sometimes cover an entire loan, but often public guarantors will offer only a partial guarantee so that the private sector still bears some risk.[2] For example, a private lender might agree to make a $10m loan if half of it was guaranteed. That sounds like we are getting somewhere – the public has only needed to assume the risk on $5m but $10m of financing has been supplied. That could help an MDB get more projects financed compared to extending the full $10m loan itself, making better use of its given capital base. [3]
But from a cost of capital perspective if the private lender would now want to charge 7.5% on its partially guaranteed $10m loan and the fee on the $5m guarantee would equate to 2.5% of the $10m, we still haven’t got anywhere.[4] We are still just moving risks around.[5]
A staple of financing is that risks should be allocated to those best placed to bear them, and if possible, mitigate them. Different risks can be split out and allocated to most suitable parties.[6] MDBs are ideally placed to insure against political risks, for example, because they can do things to reduce them.
If lenders want 10% because they are risk averse, whereas a commercially run but public provider of credit guarantees can afford to be more risk neutral, a guarantee could turn that $10m loan at 10% into something with a lower overall cost of capital, by charging a fee that is lower than the risk premium priced by the commercial lender.[7] The publicly-owned guarantor might also price risk differently than the commercial lender because it is better at appraising and managing investment risks than the private lender.[8] Hold that thought.
However, while a partial guarantee looks more effective from a mobilization perspective, if public institutions do have a superior ability to assess and manage risk then mobilisation works against cost of capital minimisation. If the public development finance institution can see that a 3% fee (or an 8% interest rate) would be risk-reflective, then mobilising a private investor who still wants 10% on the uncovered portion of their loan is going to result in a higher overall cost of capital than if the public sector just financed the whole thing with a loan at 8%.[9]
Why are dollars left on the sidewalk?
The case for guarantees is often made on the basis that private investors are not merely risk averse; they are misinformed. They don’t have data, and they don’t like uncertainty. Perceived risks are higher than actual risks.[10]
We have been telling the story that the risks in EMDE investing are lower than perceived yet the returns are high, for a long time. We keep pointing to dollars left on the sidewalk (or pounds left on the pavement), but nobody is picking them up. I had the pleasure of listening to Professor Stiglitz at a workshop organised by the Finance for Development Lab, where he said the bald facts are that private markets are not very good at their job and there is a gigantic opportunity that the public sector could seize.
I am not entirely comfortable with this story. There are things that lenders can do when borrowers are struggling to repay that incur costs, including some rescheduling of repayments, but which do not result in a declared default, and private investors may suspect that MDBs’ data showing low default rates, as evident in the GEMS database, is not the whole story.[11] That’s a point that could be cleared up with greater transparency.
I am also not entirely comfortable with this story because if there is money lying around on the pedestrian walkway, I’d like to know why anyone hasn’t picked it up. We all know markets are imperfect and there are “frictions” that prevent opportunities from being taken, but are these frictions sufficient to prevent the global financial industry, with all its layers of intermediation, from getting capital into well-publicised investments with significantly higher risk-adjusted returns?[12]
One potential explanation for the persistent existence of these apparently unexploited opportunities could be unobserved transaction costs. Benchmark risk-adjusted returns are often for intrinsically low transaction cost instruments, such as listed corporate bonds. If what investors really care about are returns net of costs, and these EMDE investments come with higher transaction costs, that could explain why large global asset managers are leaving those relatively high risk-adjusted return investment opportunities unexploited. If that is the explanation, then Professor Stiglitz might be wrong about commercial investors being bad at their jobs, but could he still be right about there being a gigantic opportunity for the public sector to exploit?
Guarantees as subsidies
Some of the best-known guarantee providers, such as GuarantCo, price their guarantees on commercial principles, meaning that they charge fees that reflect their assessment of the risks, and which should generate some profit. But we might also decide that providing guarantees without charging commensurate fees is a good idea. That would really get the cost of capital down.
In principle there is nothing wrong with subsidising private sector investments, if the resulting social returns are high enough (and the subsidy is necessary to achieve them and no greater than needed). This, though, raises the question of why, if you want to subsidise the private sector to undertake an investment where the risk-adjusted returns are too low for them to be willing to invest, a guarantee is more effective than the alternatives (grants in various guises, concessional loans or equity).[13] The answer to that question could be that guarantees are more effective when there are “frictions” to do with risk, and the problem is not only inadequate risk-adjusted returns.
Let’s return to the idea that MDBs and other experienced public sector guarantors might be better at appraising and managing the risks of EMDE investments than the commercial investors who are less familiar with these markets, and who would hence want to charge high interest rates. That would explain how guarantees can be offered with risk-reflective fees that meaningfully reduce the cost of capital.[14] What if that advantage in risk appraisal and risk management is less a question of the expertise of the individuals involved, a more a question of being willing to bear the higher costs of more “hands on” investment processes? If so, public guarantors could be seen as offering a small subsidy, even if fees are risk-reflective, because a commercial guarantor with the same costly expertise would want to charge a higher fee to recoup those costs.
We might conclude that guarantees are attractive because they allow the public sector to confer a small subsidy that changes private sector willingness to invest, at little cost. If a public sector guarantor is seen as a revolving fund, then the cost to the public of providing guarantees and charging risk-reflective fees consists of capitalising the vehicle and then tolerating lower-than-commercial net profits from providing guarantees, which shows up as having less money to recycle into future transactions.[15]
Why do we think guarantees are so effective, anyway?
This is a guess, but I would point towards OECD analysis that shows guarantees have been responsible for the largest volumes of reported mobilization, among all instruments. Here is what worries me about that. Suppose we live in a world where the big mobilisation numbers come from big infrastructure projects, the world is going to do a certain number of big infrastructure projects each year, and in many cases the optimal way of financing those projects is to combine (relatively) low-risk loans with guarantees, rather than try to find suppliers of (relatively) high risk loans. That would result in a situation where guarantees report the big mobilization numbers. It would not tell us that if we try to scale up the supply of guarantees, we will see more big infrastructure projects and more big mobilization numbers. We cannot move from information about which instruments report the most mobilization to knowing where the greatest marginal gains are from increasing the supply of different instruments. That is a question about what’s stopping potential investments from becoming reality.
If we think that guarantees can reduce the cost of capital, the question becomes how responsive demand for capital is to that price reduction – what is the “price elasticity” of demand? If investments are stalling not because risk-adjusted returns are too low but because lenders do not want to bear the level of risk (“frictions” again), then guarantees could unlock a larger volume of investment than the price elasticity may suggest.
Conclusion
The purpose of this blog has been to explore the question of how guarantees can reduce the cost of capital and mobilise private investment. If we are prepared to assume that commercial investors are simply mistaken about risks but public guarantors are not, then guarantees can reduce the cost of capital at no cost to the public. If we believe that commercial investors are not making mistakes, as such, but do not wish to incur the costs of assessing and managing risks, then guarantees can reduce the cost of capital but at some small cost to the public. Professor Stiglitz still could be right that the relatively high returns and low risks in EMDE investing are an enormous opportunity for public development finance, because that small cost still represents a good deal for the public.
There are certainly places where the supply of guarantees is inadequate and where more resources combined with innovations in their application could achieve great things. Development finance needs an array of different instruments in different contexts.[16] PIDG, for example, seems to have found a gap in the market and is setting up onshore entities that offer local currency guarantees to domestic infrastructure investors – and guarantees can be used to accelerate local capital market development more generally. If your objective is to get new markets off the ground and introduce new investors to market segments where they would otherwise fear to tread, guarantees are often the right instrument. They are also useful tools to help financial institutions expand their activities beyond their internal risk appetites and try new things like adaptation finance.
Finally, we should not forget that some of the arguments for the greater use of guarantees are not about their impact on the cost of capital, but about how they allow DFIs to “do more with less” – if, for some reasons, a DFI has the capital to bear risk but not the liquidity to finance projects itself.
[1] Solvency II sets minimum credit rating requirements for insurance; Basel specifies how much capital must be held against lending long term into sub-investment grade countries.
[2] This can be desirable when the objective for private investors is to get comfortable with the risks in a new market segment. If a guarantee is sufficiently strong that a lender fully inherits the AAA rating of the guarantor, they have little incentive to learn about the underlying risks and the transaction may fail to build new markets. Risk sharing is also desirable if you take that principled position that risk-bearing is the essence of finance, so that when private investors are only providing risk-free liquidity, they are not really providing finance. However, partial guarantees have drawbacks. Credit ratings are driven by probability of default and partial guarantees are treated as affecting recovery upon default. Some rating agencies ignore recovery and those that don’t will only offer a maximum of around 2 notch uplift (say from B- to B+).
[3] At this point the argument could take a diversion into a discussion of how banks ought to provision against losses from loans and guarantees. That topic is left for another time.
[4] In more detail, assume a notionally risk-free rate of 5% and a risk premium on this loan of 5%, which amounts to a 10% interest rate. The private lender wants to charge 10% on the uncovered $5m but only 5% on the guaranteed $5m. The guarantor is charging a 5% fee on the $5m. That works out as 10% overall on the $10m, with 7.5% going to the lender and 2.5% to the guarantor.
[5] By creating assets more suited to investor preferences, guarantees could also reduce the cost of capital in a similar fashion to how structured finance can create a “structuring premium” as explained here.
[6] Once we are dealing with specified risks the better term is ‘insurance’ rather than ‘guarantee’. Insurance comes with exclusions and claims may be denied – a guarantee should pay out on demand whatever the cause, so that investors are paid in line with the underlying debt obligations.
[7] Some people would argue that the public sector ought to price risk in the same way as the market, and not do this. However, if we think that the public sector should fundamentally price risk differently than commercial actors, so that the yardstick by which we judge profit and loss from public finance is not the market opportunity cost but some other discount rate, then it is easy to understand how public finance can bring down the cost of capital. This argument does not only apply to guarantees, but all instruments.
[8] Private investors might bridle at this suggestion, and I am not asserting it generally true. But it might hold in some markets where DFIs have a track record and the private investors in question do not.
[9] The idea that public lenders can price risk differently and undercut commercial lenders raises difficult questions about crowding-out, which will have to wait for another blog.
[10] The results of a modelling exercise, presented at a Chatham House roundtable, concluded that the case for guarantees being superior to public lending is only compelling when allowing for exaggerated risk perceptions.
[11] See the reference to “helping borrowers through temporary distress” here. Another point, which is often made, is that what happens with MDB loans is relevant to the private investors if they are participating in MDB loans, but less indicative of the performance of EMDE lending without MDB participation. But the private sector understands this and can adjust for it.
[12] The term “friction” is shorthand for institutional details that, for example, mean investors are not indifferent between investments on the market risk-return frontier, or are unable to find the capacity to take every profitable investment opportunity they encounter.
[13] Barder and Talbot (2015) suggested guarantees and subsidies are equally efficient. Other authors, such as Fernandez-Arias and Xu, present a theoretical argument for the superiority of guarantees in some circumstances.
[14] It would also explain how MDBs and DFIs could price loans more cheaply than commercial lenders, whilst still pricing risks in the same way as those commercial lenders would – because the assessed risk is different.
[15] This point probably applies to development finance in general. In Risk, return and impact I note that when DFIs say they “price on commercial terms”, that can mean charging the same prices but bearing higher costs.
[16] Guarantees also have a role in sovereign financing and green bonds, but this article is about private finance.
