Ghana External Debt: Tough targets, creativity required to lift recovery (1)
While only the publication of the Debt Sustainability Analysis (DSA) by the IMF (likely after board approval of the programme) will provide clarity about the ultimate restructuring parameters and goals, recent guidance by Ghana’s authorities indicates a substantial Net Present Value (NPV), reduction of external debt by about 16 per cent of GDP by 2028 will be a central target.
Advertisement
As discussed in Ghana external debt: Time for a haircut, 14 April 2023, we estimate that this will require an NPV reduction of about 55 per cent of commercial debt.
This is because the qualifying external debt for restructuring is only US$20bn, or some 30 per cent of GDP, including c.US$13.2bn of outstanding eurobond debt.
A multidimensional optimisation exercise Bondholders may prefer notional haircuts to large coupon reductions and maturity extensions.
Taking the existing eurobond structure as a starting point, and using the IMF’s five per cent discount rate to calculate the NPV of future debt payments, there are infinite combinations of principal haircuts, coupons and maturity adjustments that can achieve the required NPV reduction.
And any such combination would unavoidably entail some trade-offs for bondholders, as well as for Ghana.
An interesting twist comes from the fact that the IMF uses a five per cent discount rate for NPV calculations in its DSA framework - a significantly lower rate than what the market would likely see as a reasonable exit yield.
From a purely mathematical and mark-to-market-oriented perspective, this should, in our view, make market participants more inclined to accept notional haircuts relative to maturity extensions and grace periods for coupon payments.
That is because any extension of cash flows further out into the future would naturally lead to a harsher NPV reduction under a higher (market-based) discount rate than the NPV debt “savings” this would bring under the IMF framework.
For example, a 100 notional bond accruing interest at a five per cent rate would have an NPV of 100 if discounted at five per cent, no matter if it matures in five, 10 or 30 years.
Hence, extending the maturity of a five per cent coupon bond would deliver no NPV savings under the IMF framework.
But of course, the longer the maturity, the lower the NPV/market price of the bond, if one applies a 10 per cent discount rate. In the case of Ghana, we believe that a successfully concluded debt restructuring could result in ultimate exit yields of 10-12 per cent.
Maturity and coupon adjustments can be critically important for distressed issuers to manage near-term liquidity and smooth the debt-maturity profile.
In the case of Ghana, the concentrated maturities in the 2025-2029 period have long been a cause for concern and are arguably one of the reasons why a pre-emptive debt restructuring had become unavoidable (Figure 1).
And while, in our opinion and as outlined above, there are some aspects that may make mark-to-market oriented commercial creditors more inclined to accept notional haircuts than harsh maturity extensions and coupon reductions, other restructuring situations have shown different types of creditors find notional cuts difficult to accept.
This means that, as highlighted in Ghana External debt: Time for a haircut, 14 April 2023, we believe that a combination of notional haircuts, and coupon and maturity adjustments, is the most likely outcome for the Ghana debt restructuring.
All these considerations mean debt restructurings can resemble multidimensional optimisation exercises with multiple (and sometimes not clearly defined) constraints.
Financial creativity is required to find a solution that is acceptable to most, if not all, of the stakeholders.
In the case of Ghana, we believe that some solutions can be found within the framework of known targets, even though not all parameters and targets of the debt restructuring have been released ahead of the IMF’s publication of its DSA.
Based on our assumptions, these options would likely deliver recovery rates for bondholders in the low 40s.
We provide an illustrative example in which all outstanding bonds (including the 2030s partially guaranteed bond) are restructured into a series of nine new bonds, with an upfront 40 per cent haircut applied to the aggregate outstanding notional.
We assume notional repayments from 2027, but in order to smooth the overall debt-repayment profile in light of heavy local bond redemptions in 2027-2029, we assume payments would gradually increase up to 2034, with a final, longer-dated bond maturity in 2038.
All the post-restructuring bonds in this structure would have at least USD700m of notional (and would hence be eligible for most standard EM indices), while the larger bonds would have a notional in excess of USD1bn, the minimum size for some indices followed by ETFs.
We assume that all bonds carry a coupon of five per cent, starting in 2024. Of course, this aggregate approach raises the question of how bondholders of the 2030s will be compensated for the value of the guarantee.
There may also be different allocations of new bonds for different holders of particular old bonds, for example, the shorter maturity 2023s and 2026s that have single series CACs but do not feature the aggregate CACs of the more recently issued bonds.
We note that in some previous restructurings, such as Ukraine’s 2015 exercise, front-end bonds received a higher share of allocations in shorter-maturity post-restructuring bonds.
While such questions will undoubtedly be part of the negotiations, they should not have much impact on the aggregate recovery values, which we focus on in the following.
In our example, the post-restructuring bond structure would deliver the required NPV savings, support near-term liquidity, smooth the debt profile and result in recovery rates in the low 40s.
The post-restructuring bond structure as described above and illustrated in Figure 2 and Figure 3, would deliver an NPV saving on Ghana’s eurobonds close to the targeted 55 per cent under the IMF’s five per cent discount rate framework.
While the notional haircut of 40 per cent is rather large and exceeds the 30 per cent mentioned in previous comments by the Deputy Finance Minister, we have allowed for some coupon payments from 2025, contrary to the indications that Ghana would seek a grace period for coupon payments for three years.
As noted above, such extensions/grace periods of payments are particularly punitive for mark-to-market oriented bondholders.
Even with small coupon payments, the liquidity savings that Ghana would enjoy over the IMF programme period are still very substantial and while debt payments, in our example, would increase from 2027, overall external payments would remain well below the pre-restructuring peaks and would be lower in each year up to 2030.
As a small trade-off/concession from Ghana, our example would assume that the longest-dated maturities (the current longest-dated maturity is the Ghana 2061 bond) are brought forward, with the 15-year maturity bucket the longest-dated in our example.
Using an 11 per cent exit yield (the mid-point of what we see as the likely range of 10-12 per cent), the new bond structure would deliver recovery rates of 41-42, with the new bonds trading at an average price just below 70 (per 100 new notional).
While the bonds, due to the below-market-yield coupon of five per cent, would still trade in the discount territory, the distance to par is not as large as it would be in a structure with larger maturity extensions and lower coupons.
Such a structure should make Ghana’s eventual return to markets and pricing a new issue based on the post-restructuring curve easier.
A solution with a 30 per cent notional haircut, a 3-year interest grace period and a 1/3 cut to coupons thereafter (one approach discussed by market participants following the Deputy Finance Minister’s comments along these lines in November last year) would also deliver recovery values just above 40, according to our framework and calculations.
From a bondholder perspective, the benefits from the lower notional haircut would be more than offset by the interest grace period.
Coupon payments after the initial grace period would be just marginally higher than the five per cent we have assumed above (around 5.25 per cent, according to our calculations), but the effect on recovery values from this is rather marginal.
However, this would result in marginally lower NPV debt savings for Ghana, with the IMF’s five per cent discount rate framework slightly under-delivering against the 16 per cent of GDP NPV reduction target for external debt, according to our calculations.
It is worth noting that the comments made by the Finance Ministry in November 2022 about the expected restructuring approach and NPV reductions were likely based on assumptions of a successful domestic exchange programme as initially designed.
The fact that the final exchange - and associated cash flow savings - was smaller than initially anticipated and that other domestic debts such as those held by local pension funds, IPP creditors, etc., are yet to be restructured points to the need for deeper NPV adjustments on the external side.
Moreover, under the IMF’s five per cent discount-rate framework, it would mean that maturity extensions actually add to the NPV of Ghana’s debt, rather than help meet the NPV saving targets.
Hence, while a 30 per cent notional haircut may look more appealing from a headline perspective if this comes with a three-year interest grace period, then we believe this would be an inferior solution for both Ghana and market-value-oriented bondholders.
Upside from current market prices, but speed is of the essence Based on our calculations and on the parameters/targets known to date, we believe that recovery values for eurobond holders in the low 40s are achievable and hence see some upside to current trading levels in the low-to-mid 30s (which also underpins our Overweight rating on the credit).
Of course, if negotiations provide some additional leeway around these targets (or underlying assumptions), this would increase the potential upside.
Similarly, any potential addition of a recovery value instrument, with payments contingent on Ghana’s future economic performance, could have a similar positive effect on overall recovery values.
However, as our discussion above also shows, the room for maneuver does not seem large, and given the high opportunity costs of holding defaulted debt in a world of much-increased interest rates, the return prospects for Ghana eurobond holders will ultimately depend on the speed of the restructuring process and bondholders’ ability to realise the potential upside promptly.
Undoubtedly, there are risks to the process, including the likely difficulty for some investors to accept such large haircuts, China’s stance within the G20 common framework 1 (China’s portion of Ghana’s debt load is smaller than in other cases where an agreement with China on equal treatment has remained elusive, but at c.USD1.7bn it is not negligible), the required flexibility of all actors, including the IMF, to find alternative solutions if the process does not move forward at speed within the G20 CF and Ghana’s political cycle ahead of the end-2024 elections.
We also note that the timeline for the IMF board approval of the programme has repeatedly shifted. At the same time, we still think that the starting point for Ghana’s external-debt restructuring is better than in many other cases.
Ghana has already completed its local-debt restructuring (while the scope of local bond inclusion in restructuring efforts is still an ongoing debate in Sri Lanka, for example).
And contrary to Zambia, Eurobonds represent the lion’s share of external debt likely to be in-scope for restructuring, arguably giving bondholders a stronger voice in negotiations.
Not all parameters of the external-debt restructuring are known yet, but authorities’ indications imply a 55 per cent NPV reduction in eurobond debt may be targeted.
With some creativity, we estimate such an ambitious target could be achieved with bond recovery values in the low 40s, c.5-8 points above current market prices.
Key takeways
• Not all parameters are known, but eurobond-debt reduction targets are likely to be very ambitious.
Ghana’s authorities have indicated that to comply with IMF targets, a 16 per cent-of-GDP NPV reduction of external debt by 2028 may need to be achieved in the upcoming restructuring.
We estimate that this would translate into an ambitious c.55 per cent NPV reduction for eurobond debt.
• The difference between the five per cent discount rates used by the IMF and likely market-based exit yields may skew bondholders’ preferences towards larger upfront notional haircuts, relative to larger coupon cuts and maturity extensions.
At the same time, maturity and coupon adjustments will also need to be part of an overall solution to provide near-term liquidity relief and to smoothen the debt profile.
• We provide an illustrative example that we believe broadly achieves the outlined targets and results in a recovery value for bondholders in the low 40s.
This implies some upside from current market prices, which could be enhanced if negotiations provide leeway around the targets, or if any additional recovery value instruments are introduced contingent on Ghana’s future economic performance.
However, a speedy process allowing investors to realize the potential upside is of the essence, and risks to this are non-negligible.
