Capital in an MDB, a benchmark for blended finance?

Or, as Chat GPT would put it: "MDBs: Where Low Risk Meets Low Yield (But Hey, at Least You're Not Losing Money)".

International development finance is scarce are reportedly insufficient. Scarcity implies choice. Choice implies opportunity cost. A blended finance practitioner, it’s your bread-and-butter dilemma. Is that a good use of taxpayer’s money? Could I get more impact for lesser public cost? And perhaps more importantly, how do I come up with process(es) that do an effective job in translating those grand principles into actionable and sound investment decisions.

In the investment business, beating the market means beating the benchmark (“my portfolio is beating the S&P 500”). What would be a good benchmark for blended finance? Ideally, you would need a standardized blended finance vehicle, different investments under that same structure, historical performance, and comparable datasets. Look no further than our good old MDBs, especially those that focus on the private sector. They’ve been around for a while. They have comparable structures and business models. Finally, they publish a decent amount of data that you can rely upon.

From the prospective of a donor looking to invest in blended finance, I would argue that capitalizing an MDB meets all the different definitions of blended finance (I’ll elaborate further on).

Of the ratio of impact to public cost, this post will focus on the denominator. What’s the cost to the taxpayer to put money in an MDB? I recall during my induction at IDB Invest our then CFO explaining that MDBs typically seek “capital preservation”. In layman term’s, it means doing better than inflation.

Let’s delve into it. Together with my colleague Mengxue, we looked at ADB, AfDB and IFC (US$ equivalent balance sheet, availability of historical data, etc.). MDBs are not in the dividends business, but they sometimes transfer retained earnings to trust funds or separate institutions from the same group. Also, capitalization (i.e. paid-in capital) tends to occur over a multi years period, with a commensurate increase in shares. We therefore focused on the compounded annual growth of capital per shares, with several adjustments to zero-out extraordinary transfers, etc. I’ll save you the details of the calculations, if you want them, message me, it’s all based on publicly available information. The outcome of the calculation is summarized below. We’re right about a couple of percentage points above inflation. Neat!

Inflation measures the increase in price of goods and services. If an investment performs like inflation, it implies that no money is lost (hence the term capital preservation used previously). In other words, I can buy the same burger today than 10 years ago, even if nominally the price of burgers increased (it did!). Thus, a way of interpreting the above results is that when you are putting equity in an MDB, you are not losing any money.

But is that necessarily a good proxy of public cost? Shouldn’t we factor-in shareholders cost of funding (the cost to government to borrow on the markets)? We ran the numbers for one shareholding country (with a strong credit rating). We assumed 1 unit of local currency invested at the beginning of our reference period converted in US$. We then factored-in the increase in capital per share and converted it back to the local currency at the end of the period. Finally, we discounted it based on the country’s compounded historical 1-year treasury. We could have gone for a longer issuance, but that would have taken a trip to our Bloomberg terminal on a separate floor and trying to figure out the many passwords. Plus relying on the yearly treasuries gives more granularity to the result. Here’s what we get:

Looking at it from a shareholder’s cost of funding, it appears that investing in an MDBs remains good value for money from a financial standpoint. Obviously, this is largely tied to the cost of funding of each shareholding country. For those that have higher funding costs and or more challenging access to hard currencies, the cost would be higher. In the case of recipient countries however, this would be compensated by accessing MDBs resources.

Are we done yet? Not quite. From an opportunity cost standpoint, MDBs shareholders could theoretically decide to put their money in a commercial bank, which, for argument’s sake we’ll assume have a similar portfolio than MDBs. Returns on equity for banks are around 8%-10% (thanks Chat GPT, that sounds about right). We again take a whole lot of shortcuts by comparing returns on equity to capital per share CAGR, and we see that returns on commercial banks are about 4-5x times those of MDBs. There are at least two counterarguments for such an approach. First, investing in an MDB is arguably more comparable to investing in a highly rated fixed income instrument than equity in a commercial bank (low risk, low and stable yield, no upside). Second, should we even be comparing MDBs to commercial banks to begin with, because their activities, shareholding structures, governances, etc. are so different? In reality, the underlying question we are trying to answer is whether MDBs shareholders are transferring concessionality to the MDBs. Thus, a simple way of looking at it is to compare our return proxy to the institutions’ bond issuances. Those are AAA-rated obligations of the same underlying and are theoretically less risky than an equity participation, hence should feature lower returns. I focused on the 5-year bullet benchmark size (> US$ 500 million) that is the closest to year end 2010. I took the G-spread (spread to US Treasuries in this case) as of Dec-31st 2010, then added it to the interpolated US Treasury rate at the same date for a 9-year tenor (to think I used to do that as a living). This time I could not avoid the trip to The Bloomberg, and after several failed attempts with the login-in, et voila:

This is getting interesting, isn’t it? The bond “pays” more than the equity, or in the case of IFC, they are about the same. This sort of makes sense when you think about it. On the one hand MDBs seek inflation adjusted capital preservation. On the other hand, a AAA US$ bond is going to be behave similarly than US Treasuries, which in turn mirror interest rates and inflation. It suggests that in seeking capital preservation, MDBs shareholders are transferring concessionality to the institutions. How that concessionality is ultimately used is another debate. It could be that all else been equal, it’s covering the incremental costs of impact measurement, environmental, social and governance safeguards, that are characteristic of MDBs. Whatever the explanation, it does lend to the idea that there is some level of subsidy (whether implicit or explicit) embedded in MDBs’ operations (more on the topic here, here and here). ?

All the above is approximate, and I wouldn’t dare drawing any conclusions (I selected the data range that gave the most consistent results, some ranges looked messier, so it should be taken with more than a pinch of salt). Be that as it may, I believe it suggests a couple of avenues that worth looking into it:

First, looking at an MDB as a standalone blended finance seems to hold across the various definitions of blended finance. Private capital is always involved (as a minimum, through corporate bond issuances) and there are signs of concessionality.

Second, if MDBs can indeed be considered blended finance vehicles, this suggests a natural efficiency benchmark for blended finance interventions. If the public cost of a blended finance intervention is higher, then one should expect more impact (or an outcome that could not be achieved through an MDB). Conversely, if impact is the same, then public cost should be lower. The crucial point is to base this test on the “public cost” equivalent of the intervention and not its face value.??

Third, there are different ways of looking at public cost, all of which seem to have merits. This touches the topic of concessionality and suggests there are different ways (likely all correct) of looking at it: cost to the donor, opportunity costs, market driven implied subsidy, etc.

Hopefully this is food for thought for further investigation.

Fancy more on the topic, here is a compilation of my articles, essays, etc. and team's blogposts on blended finance.

Acknowledgement

Thank you to Mengxue Tang for her work on assessing MDBs balance sheets and to Paddy Carter for the exchange that led to the idea that MDBs could possibly be used as efficiency benchmark for blended finance. This post also draws on the work and lessons learned of IDB Invest’s Blended Finance team ( Eduardo Gutierrez , Elee Muslin , Joan Miquel Carrillo , Mengxue Tang , Pilar Carvajo Lucena and Sofía Ahualli ).

Disclaimer

I am writing those articles from the practitioner’s point of view, recognizing that I am primarily relying on empirical (at best) or anecdotical (at worst) evidence. In blunter terms, I may very well end up being wrong. Blended finance is still at a nascent stage and any honest debate on the efficient use of concessional finance for development will contribute to strengthening the practice. Views are my own, and English is not my native language, so pardon my mistakes. Legalese version click here:?link.