Zambia: the Official Sector Strikes Back
Zambia’s agreement in principle with bondholders was met with reservations by the IMF and the Official Creditor Committee, and is apparently being revised.
There are two main reasons for the official sector to balk at the deal: compliance with IMF targets, related to the DSA and financing gaps, as well as Comparability of Treatment which is assessed by official creditors.
While time will tell which of these two aspects — if not the two — was the main hurdle in the assessment, this development sends the strong message that the official sector is serious about enforcing the rules, even if it means calling out a deal and reopening negotiations with commercial creditors.
This latest plot twist will further delay Zambia’s restructuring process — hopefully by not more than a few weeks — and could also have spillover effects for other restructurings, especially in Common Framework cases.
(TL;DR — I also did a short thread laying out my main thoughts back on the day of the announcement)
IMF targets: blurred line between program reviews
The first test that the agreement in principle (AIP) with bondholders has to pass is the compliance with IMF program parameters which are twofold: meeting DSA targets, and ensuring that the IMF program is fully financed — meaning there is no residual financing gap for each year of the program.
According to Zambia’s October 2022 investor presentation, the binding targets for Zambia are related to external-debt-to-exports and external-debt-service-to-revenue, while bridging the financing gap at that time also required creditors to deliver $8.4 billion of flow relief over the program period.
The IMF’s wording in the FT after expressing their reservations about the AIP seems to indicate that the deal did not fit within the targets mentioned above:
“Further discussions and modifications are needed to bring this initial proposal more fully into line with the requirements of the programme.”
One thing to take into account is that the IMF’s assessment is most likely based on the updated DSA published with the first review in July 2023, which featured significant changes compared to the one published when the program was approved back in September 2022.
Several parameters of the original DSA were sharply criticized by creditors: Zambia’s weak debt carrying capacity (DCC), the application of 40% buffers on already low targets, and non-resident holdings (NRH) of local debt eating up an important share of available cash flows.
While the IMF could not change the DCC or the buffers, it did modify the data related to the NRH and other aspects of the macro-fiscal framework when approving the first review: NRH were revised down from $3.2 billion to $2.6 billion in face value, reducing the stock constraint for other creditors. The IMF also modified the forecasts for the evolution of NRH, judging from BoP data: net portfolio inflows were reduced by c. $2 billion over 2022-2025, as I discussed in an earlier blog. In addition, projections for exports — a key anchor for DSA targets — were improved, for instance by c. $600 million in 2024 and $900 million in 2025 between the 2022 and the 2023 staff reports (data from the BoP tables).
These changes arguably relaxed the DSA and financing gap constraints: for example between the program approval and the first review, the flow relief to be delivered by external creditors over 2022-2025 was reduced from $8.4 billion to $7.7 billion — that is $700 million more for servicing restructured debts in that period.
Looking at the terms of the AIP, a likely point of contention with regards to IMF targets might be the early amortization of the base bond (“bond A”). As we noted with Brad Setser, the amortization structure of this bond looks rather steep: 25% of the $2 billion bond A was expected to mature in 2024 and 2025.
For reference, the October 2022 investor presentation noted that Zambia would have $400 million of available cash flows for all restructured debts (Paris Club, China, bondholders, etc.) in 2024 and 2025 combined (slide 13). The AIP on the other hand implied that bondholders would get north of $700 million for these two years with principal and coupon payments — even if the IMF freed up some cashflows, the gap appears difficult to bridge.
Another issue could be around the design of the state-contingent instrument featured in the AIP (“bond B”). The IMF and official creditors have repeatedly insisted in recent months that any state-contingent instrument designed as part of restructuring negotiations should be in line with IMF program parameters, in all states of the world. It is not totally clear whether it is the case with the triggers for the upside scenario in the AIP — let’s dive in.
The upside scenario is triggered upon the occurrence of any of two events: Zambia's DCC is upgraded from weak to medium by the IMF, or "the 3-year rolling average of the USD exports and the USD equivalent of fiscal revenues (before taking into consideration grants) exceeds the IMF’s projections as laid out in the First Review of the IMF’s Extended Credit Facility Arrangement released in July 2023".
Therefore, a scenario could occur in which Zambia barely outperforms IMF forecasts for exports and revenues in the observation period but the DCC is not upgraded. In that case the upside scenario would materialize — one of the two triggers is sufficient — and the debt repayment schedule would steepen sharply, potentially pushing Zambia over the debt sustainability thresholds for countries with a weak DCC.
As a side note, the reasoning for bondholders to combine the two triggers had some merits in theory: the DCC is a prime candidate for a state-contingent trigger due to the massive threshold effects it induces for IMF debt targets, meaning creditors can leave a lot of money on the table if Zambia’s DCC is agonizingly close to being upgraded. However, the DCC is also a complex indicator taking into account variables like world growth which could fail to capture variations of Zambia’s actual ability to service debt. It also relies on qualitative factors which could increase legal risks when including it in a bond contract. Therefore bondholders chose to combine the DCC with a somewhat more objective trigger related to exports and revenues. Time will tell if this drew the ire of the IMF.
Bottom line, there is a lot of moving pieces: it is hard to say with publicly available data how far off the mark the AIP currently stands, whether it has to do with the amortization schedule or the state-contingent features, and therefore how long renegotiations could last.
Comparability: no more free riding?
Another important test for the AIP is Comparability of Treatment (CoT) as assessed by the Official Creditor Committee (OCC). The G20 Common Framework (CF) communiqué from November 2020 indicated that “A debtor country that signs an MoU with participating creditors will be required to seek from all its other official bilateral creditors and private creditors a treatment at least as favorable as the one agreed in the MoU.”
The communiqué also specified that CoT would be assessed along the three indicators used by the Paris Club (PC): “changes in nominal debt service, debt stock in net present value terms and duration of the treated claims”. Indeed, the PC has historically used a holistic approach when assessing CoT, calculating the contribution of all creditors with the three methodologies and making a final judgement call. There is no set threshold above which a treatment is deemed not comparable — a debt treatment could provide a lower NPV haircut than the PC and compensate with more short-term flow relief and longer duration extension.
As noted by Diego Rivetti, the PC has never withdrawn a debt treatment due to a breach of the CoT provision. Historically, the PC also did not include claw back clauses in its agreements: the main threat for debtors offering sweeter terms to non-PC creditors came from the non-compliance with IMF targets — if the debt treatment is not compliant then the IMF will suspend disbursements.
The CF was bound to reschuffle the cards.
In October, the progress report of the Global Sovereign Debt Roundtable (GSDR) provided further clarity on the assessment of CoT in CF restructurings. It confirmed that the three PC formulas would be used, and clarified that the NPV reduction would be calculated using the present value of old/new debt as numerator/denominator respectively, all discounted at the standard 5% rate used in the LIC DSA.
Most importantly, the GSDR progress report stressed that the OCC would enforce CoT through different mechanisms, evolving from the PC’s historical practices. These mechanisms include claw back clauses — the OCC debt treatment would fall through if a sweeter one is given to other creditors — or the request for the debtor country to commit not to repay any creditor which has not agreed to comparable debt treatment.
While often overlooked in the analyses of the GSDR progress report which came out after the IMF meetings, this paragraph is key. The design of the enforcement mechanisms in the Zambia CF MoU also probably explains the delay between the June deal with the OCC laying out the financial terms, and the signing of the MoU which happened only in October during the IMF meetings.
The lack of enforceability of CoT — or at least the perception thereof — had been an issue for China since the creation of the DSSI and then the CF. It explained part of China’s reluctance to negotiate restructurings in a multilateral setting and accept the same terms as PC creditors, if it believed that bondholders would be better off. For instance Deborah Brautigam and Yufan Huang acutely described the perception of unfairness by China towards bondholders in their paper on the participation of China in the DSSI.
The official sector balking at the first ever bondholder deal concluded for a CF case therefore sends a strong message: the PC is now serious about enforcing CoT. This could as a result increase the trust of China in the overall CF process — let’s see if any positive spillovers materialize on other cases.
Similar to the IMF debt targets, it is hard to say how much of an issue CoT was in the official sector’s assessment of the AIP without the precise cashflows pre- and post-restructurings for all creditors. In order to improve the accountability of the process, it would be a major progress for the PC/OCC to make its CoT assessment public, along the three formulas. This would coincidently reduce a potential perception that the official sector balks at deals it doesn’t like for purely political reasons. Finally, the publication of calculations would gradually provide increased clarity ex ante on the wiggle room that debtors have when negotiating with bondholders.
An important related question is the role of the IMF with regards to CoT. The longstanding policy of the IMF is to not have views on the allocation of losses across creditor classes — hence on CoT — and to focus entirely on the overall debt reduction. However, the IMF has an indirect interest in CoT because the lack of comparability can delay restructuring negotiations and create inter-creditor tensions which in turn hamper the return to debt sustainability.
In addition the IMF has significant technical capabilities within its staff when it comes to debt math. The IMF recently started publishing partial CoT calculations, specifically the respective NPV reductions of the PC and bondholders in Suriname at different discount rates. Building on this example, it would make a lot of sense to leverage the IMF’s expertise and increase the strength and transparency of the CoT assessment, making it a standard disclosure in staff reports during restructuring.
What happens now in Zambia and elsewhere?
The priority in the short term will be for Zambia to adjust the deal and make it compliant with the two tests described above. The wording of the IMF in the media seems to indicate it should be doable in a reasonable timeframe, though it might seek to avoid scaring off the market. In addition, it is unlikely that the IMF would block subsequent disbursements because a deal with bondholders has not been finalized — the announcement of the AIP still shows that negotiations are progressing and the arrears policies are more permissive with commercial creditors than with official ones.
Beyond Zambia, the clarity provided in the GSDR report with regards to CoT applies to all CF cases. This means that recent developments matter a great deal to Ghana and Ethiopia. In Ethiopia specifically, the treatment of the country’s sole Eurobond has been a recurring topic of discussion, and some bondholders have made a preemptive restructuring proposal. The looming threat of the official sector balking at a deal after concluding their own debt treatment a few months later might disincentivize a preemptive liability management exercise which arguably has otherwise strong merits. This illustrates issues around the sequential aspect of the CF and the persisting lack of clarity ex ante for the assessment of CoT — a topic for another blog.
For countries outside of the CF, namely Sri Lanka and Suriname, the impacts of Zambia’s process are less obvious in terms of CoT, although they will still have to ensure strict compliance of any debt deal with IMF parameters. First, the GSDR report underlined how a common understanding of CoT in the official sector is yet to emerge for non-CF cases. For instance, if China pushed for the inclusion of claw back clauses in the Zambia MoU, it is uncertain whether the PC would use the same mechanisms when negotiating without China as is the case in non-CF countries.
Second, non-Common Framework cases are less sequential and it would surely be much more politically costly for official creditors to call out a bond exchange after it has been executed. This might in turn create incentives for all creditors to front run others and announce a deal quickly — something which one can glimpse from China Exim’s restructuring agreement in principle of which nobody seems to have seen the terms.
An important point: the rejection of the Zambia bondholder deal does not spell the end of state-contingent instruments in restructurings — it simply means that their design warrants increased attention to the compliance with IMF targets and CoT. For instance the “Macro-Linked Bonds” proposed in Sri Lanka are linked to the USD GDP which is also the anchor for all debt targets (debt-to-GDP, GFN-to-GDP, and external-debt-service-to-GDP), making it easier to ensure their compliance with these targets. Whether they return the country to a sustainable footing is another question, and the government balked at the bondholders’ proposal for now.
Bottom line, one can only hope that Zambia will soon be out of the woods, and that the transparency and accountability of the CoT assessment will improve going forward. While it seems unlikely that all creditors will agree on a unique quantitative CoT formula — none could fully capture complex debtor-creditor relationships — there are low hanging fruits which could greatly contribute to streamlining restructuring processes.
I am grateful to Bernat Camps, Mathilde Gassies and Stephen Paduano for valuable comments.
If you want to discuss this further, please reach out: tmaret@globalsov.com
You can also follow me on twitter @TheoMaret
Views expressed here are my own, not my employer’s.