Conway Williams, Head of Private Markets at Prescient Investment Management
This year’s global bond sell-off is usually discussed in terms of what it does to bond portfolios. I want to look at a different consequence: what governments do when borrowing becomes expensive, and what that means for the investors they turn to.
Long-dated yields in the US, the UK, Germany and Japan are at their highest levels in more than a decade. Inflation has played a part, but heavy government borrowing is a large part of the story. There is a great deal of government debt to absorb and more is on the way.
Expensive funding does not have to mean governments stop building. One response is to move the building off the balance sheet: offer guarantees, sign long-term contracts and invite private capital to fund what the state would otherwise fund directly. South Africa has relied on this approach for years, most visibly through the Renewable Energy Independent Power Producer Procurement Programme (REIPPPP) since 2011. If borrowing costs stay high, governments elsewhere may return to similar tools, as the UK did with its Private Finance Initiative from 1992 to 2018. The question for every allocator is whether they are being paid for the project and sovereign risks these structures carry and for giving up liquidity.
A constraint South Africa knows well
In its April 2026 Fiscal Monitor the IMF put global public debt at nearly 94% of GDP and projected it to reach 100% by 2029. Rising interest bills put long-term investment under pressure.
South Africa starts from a different place. The fiscus has largely stabilised. In the 2026 Budget Review National Treasury projected gross debt to peak at 78.9% of GDP in 2025/26 and then decline. S&P upgraded the country in November 2025, and Fitch followed in June 2026. That progress is real but it has not created room to spend. Debt service was budgeted at R420.6 billion for 2025/26, about 21% of revenue. Meanwhile the transmission plan alone calls for about 14,500km of new lines by 2034 at a cost of roughly R440 billion. That gap between what the country needs and what the fiscus can carry is where private capital comes in.
How South Africa learnt to build off balance sheet
Under REIPPPP, private sponsors and lenders funded construction, and Eskom bought the power under long-term agreements. Government stood behind Eskom’s obligation to pay through an implementation agreement with each project. According to the IPP Office’s latest published figures, by June 2024 renewable projects that had reached financial close had attracted R272.5 billion of investment, more than four-fifths of it from domestic investors.
The next phase builds on that model. National Treasury is establishing a Credit Guarantee Vehicle, designed as an independently governed insurer that will guarantee payments to infrastructure projects and cover them if contracts are terminated early. It is set to start with about $500 million of capital, $350 million of it from the World Bank, and aims to mobilise up to $10 billion of investment over time. Its first guarantees are aimed at independent transmission projects. Separately, amendments to Regulation 28 that took effect in January 2023 allow retirement funds to hold up to 45% in infrastructure across asset classes. The plumbing for local savings to fund the build is theoretically in place.
What investors are really buying
In these structures private capital takes on construction and operating risk. The state, or a vehicle it seeds, keeps much of the tail risk: the rare but severe outcomes. National Treasury put its exposure to signed independent power projects at R244.3 billion as at 31 March 2026, carried as a contingent liability.
A senior loan to a renewable energy project with a state-backed offtake agreement is two exposures in one. The first is project risk: construction, performance and the grid. The second is the state standing behind the offtaker. Investors who treat it purely as private credit understate the second. Investors who treat it as a government bond proxy ignore the first. A fair price pays for both, plus a premium for not being able to sell.
Measuring against the right yardstick
It is tempting to read high government bond yields as a hurdle every private asset must beat. The benchmark South African ten-year government bond yielded about 9% in early October 2026, a real return of roughly 4.6% against August inflation of 4.4%. That looks like a hard number to argue with. But the comparison mixes up two different decisions.
The first is whether to hold real infrastructure assets at all. That choice is made when an investor selects a mandate, and it is the right place to weigh the liquid alternative. A government bond cannot offer direct exposure to real assets, contracted cash flows from those assets or a hand in building what the country needs. Investors who want those things are not choosing between a solar loan and a bond. They have already chosen.
The second decision is whether a particular loan belongs in that mandate. Here the government bond is the wrong yardstick. Measuring a twenty-year senior project loan against a listed instrument of similar duration compares it with an exposure the investor has not asked for. The questions that matter are whether the loan clears the mandate’s real-return objective after the risks identified, and whether it is priced in line with comparable assets: senior against senior, construction against construction, on a like-for-like real basis. For floating-rate debt, rising rates lift the headline yield without making the deal any better. Converting the margin into a real spread is the proper test. Because private assets are valued using models rather than traded prices, their returns look smoother, and less risky, than they really are.
What could go wrong
The model has limits that investors should price in. Guarantees are contingent liabilities, so if the fiscus came under pressure again, government bonds and state-supported infrastructure debt would likely weaken together. And retirement savings directed towards national priorities still have to earn their place on merit. If pricing discipline slips, the risk is an informal version of prescribed assets, which serves neither members nor the projects. Both are reasons to demand a fuller premium.
The answer for allocators
So what do governments do when borrowing becomes expensive? On South Africa’s evidence they can keep building by changing who pays. Investors here should expect more guarantees, more long-term contracts and more invitations to private capital. For South Africa this is a developmental necessity. For long-term savers it is a real opportunity.
What should those investors expect to be paid? Enough for each risk they carry, measured against the right yardstick. High sovereign yields rightly raise the bar when an investor chooses a mandate. They should not become the test for each asset within it. That test is the mandate’s objective and comparable private assets.
ENDS
