There is a number that ought to be more famous than it is. In 2023 the United Nations Development Programme calculated that African sovereigns pay roughly 74.5 billion dollars more to borrow than their actual financial behaviour warrants. That figure is not a measure of debt, or of poverty, or of mismanagement. It is a measure of perception. It is the price the continent pays, in excess interest and foregone financing, for being seen as more dangerous than it is. We think it is the single most instructive statistic in international finance, and almost nobody outside a handful of development economists can tell you it exists.
We want to use it to argue something the firms we compete with will not say out loud, because saying it gives away the trade. Most of what the world calls risk in frontier and emerging markets is not risk. It is illegibility. The two are routinely confused, they are priced identically, and the difference between them is where a great deal of money quietly sits, waiting for someone who can tell them apart.
Start with the evidence, because the claim sounds like boosterism until you look at it. Moody's spent fourteen years studying more than eight thousand project-finance loans made between 1983 and 2018. The default rate on African infrastructure projects over that period came in at 5.5 per cent, the lowest of any region in the world. When Moody's refreshed the work, the African figure fell to under two per cent. Over the same window Eastern Europe ran at 12.4 per cent and Latin America at 10.1 per cent. North America, the supposed safe harbour, sat at 6.6 per cent, higher than Africa. These are not advocacy numbers produced by a regional development bank with an agenda. They are the loss rates compiled by one of the three agencies whose ratings govern the cost of capital for most of the planet.
Now set them beside what the same continent actually pays. African dollar bonds yield around 9.1 per cent on average. The comparable figure in Latin America is 6.5 per cent, and in emerging Asia 4.7 per cent. So the region with the lowest measured default rate on infrastructure borrows at the highest cost, and the gap is not small. Of fifty-five African countries, only a handful hold an investment-grade rating from the agencies that control more than ninety per cent of the ratings market. The rest are stamped junk, or left unrated entirely, which is worse, and the stamp does its own work regardless of what any individual borrower has ever actually done. The IMF found in 2023 that sub-Saharan governments pay materially higher coupons than Latin American and Asian issuers with the same ratings. Same rating, higher price. The rating, supposedly the objective measure, does not even clear its own bar.
What is being priced here, if not default? Three things, and none of them is risk in the sense the word pretends to mean.
The first is contagion by category. When Zambia defaults, the cost of borrowing rises for Kenya, Ghana and Senegal, countries that did nothing, are structurally unrelated, and in some cases were improving at the moment the penalty landed. The market is not pricing those countries. It is pricing the word Africa, applied as a regional overlay, because looking closer is expensive and the overlay is free. That is not a risk assessment. It is the absence of one, dressed as prudence.
The second is the cost of looking. A pension fund in London or a desk in Frankfurt can underwrite a German Mittelstand borrower in an afternoon, because the information arrives in a familiar format, audited to a familiar standard, in a familiar language, inside a legal system the analyst already understands. The same fund, asked to assess a perfectly sound Ivorian agribusiness, faces a wall of unfamiliarity: different accounting conventions, a different legal tradition, a language barrier, no comparable peers in the model, and no cheap way to verify what it is told. Faced with that wall, the analyst does the rational thing for an individual and the destructive thing for the system. He declines, or he demands a premium fat enough to cover his own ignorance. The premium is real money charged for a problem that lives entirely on the lender's side of the table.
The third is simple distance, the kind no spreadsheet captures. The deals that fail in these markets mostly fail for unglamorous reasons. The wrong introduction, or none. A contract structured for a legal system it will never actually meet. A counterparty who was trustworthy and unknown, which in the eyes of capital is the same as untrustworthy. A timeline set by someone who had never been to the place. None of this is risk in the actuarial sense. It is friction, and friction is not a property of the market. It is a property of the gap between the market and the outsider trying to enter it.
The world has built an enormous and largely unexamined machine for converting unfamiliarity into a number and calling that number risk.
Put the three together and you arrive at the thing we will state plainly, since it is the whole of our view. The conversion is lazy, it is systematic, and it is wrong often enough to be a standing opportunity rather than an occasional one. The UNDP's 74.5 billion dollars is simply the visible, sovereign-level tip of it. The same error runs all the way down, through corporate credit, through trade finance, through private deals too small to rate, repriced upward at every step by people who mistook the cost of their own ignorance for the borrower's probability of failure.
We do not say any of this to be contrarian, and we are emphatically not saying these markets are safe. Some of them are genuinely treacherous, and a firm that cannot tell a real hazard from an imagined one is worse than useless. That is exactly the point. The entire value of operating here rests on the ability to distinguish the two, and that ability cannot be bought as a data feed or hired as a credit model. It is built slowly, through presence, through language, through knowing which counterparty actually pays and which courthouse actually functions and which official actually decides. It is, in the most literal sense, the work of making an illegible market legible to the capital that would otherwise misprice it or flee it.
Most of the financial world has organised itself to avoid that work. It prefers the overlay to the visit, the rating to the relationship, the familiar borrower at a thin margin to the unfamiliar one at a fat one. That preference is rational for any single institution and quietly catastrophic in aggregate, and it is the reason the African infrastructure project that repays more reliably than its North American equivalent still cannot borrow at a comparable rate. The gap will not close on its own. The incentives that produce it are stable and the people who benefit from explaining it away are few.
We have made a different choice, which is to treat illegibility as the asset rather than the obstacle. What looks like risk to a desk that has never left the building looks, to someone who has done the work, like a market that has simply not yet been read correctly. The premium the world charges for not understanding a place is, to the party that does understand it, the entire reason to be there.
The cost of not looking is 74.5 billion dollars a year and counting, on the sovereign ledger alone. We think that is less a tragedy than an address. It tells you precisely where the mispricing lives. The rest is a matter of being willing to go and stand in it.
This is the view of the firm. Figures are drawn from the United Nations Development Programme (2023), Moody's Analytics project-finance default studies, and the International Monetary Fund (2023).