Following Executive Board approval of Zambia’s US$1.3bn extended credit facility (ECF) last week, the IMF released the accompanying Staff Report on Tuesday. The report lays out the contours of Zambia’s reform programme and, most importantly for bondholders, includes the debt sustainability analysis (DSA) that will serve as the basis for restructuring negotiations. In the report that follows, we outline the most salient points and the implications for Zambia’s eurobonds.
Based on the DSA, we think there is upside for Zambia’s bonds, with prices currently implying a 50% net present value (NPV) cut versus a proposed 35-45% NPV cut by the IMF/World Bank. Nonetheless, bonds have sold off since the staff report was published, based on headlines citing the requested US$8.4bn of debt relief. While this is 50% of the public external debt stock, such a comparison is erroneous as it is defined as cashflow relief over the programme period and we estimate it is roughly equivalent to the value of 'restructurable' amortisations over the duration of the ECF. We argue this implies maturity extensions alone may provide sufficient cashflow relief.
Further, the DSA represents a starting point and its underlying assumptions and targets will be subject to negotiation with creditors. As such, we don’t think the treatment will be any worse, and a strong and well-organised bondholder committee may be able to extract better terms. On balance, we think this is more likely to result in lower NPV haircuts than currently outlined rather than higher. On this basis, we upgrade our recommendation on Zambian eurobonds to Buy from Hold at US$57 (23% YTM) at cob on 8 September on Bloomberg for the ZAMBIN 8.97 07/30/2027s.
The ECF spans the 2022-25 period, with associated macro assumptions corresponding to that period and debt targets spanning through 2031. Real GDP growth is projected to rise from 3% in 2022 to 4.5-5% over the medium term, which is in line with the 2010-19 average and comfortably below the 6.8% rate that prevailed from 2000-09. Meanwhile, inflation is projected to fall within the Bank of Zambia’s (BoZ) 6-8% target band by H1 23 and remain there over the medium term. At first glance, these assumptions seem reasonable.
The current account is projected to remain in surplus, but will fall from 7.6% of GDP in 2021 to 1.4% in 2022 and 0.3% in 2023 as imports recover, before rising gradually to 2.7% of GDP by 2025 on rising copper exports (growing 6.6% yoy each year from 2022-25). Copper prices are projected to average cUS$9,300/ton from 2023-25 in line with the April 2022 WEO, above the current price of cUS$7,800 and futures pricing of cUS$7,700 over that period, implying some downside risk to projected copper exports.
The key anchor of the ECF is the primary balance (commitment basis), which is projected to consolidate from a 6% of GDP deficit in 2021 to a 0.7% of GDP surplus in 2022 and a 3.2% of GDP surplus in 2025, a massive 9.2pp adjustment relative to the 2021 outturn (or 10.7pp excluding one-off revenue items) and 13.3pp relative to the 2020 deficit of 10.1% of GDP (or 14.1pp excluding one-off revenue items).
This will undoubtedly be a heavy fiscal lift, with only four low-income countries achieving such a large three-year adjustment under an IMF programme since 1990, according to the DSA’s realism tool. Zambia itself has only achieved a primary surplus >3% of GDP twice since 2000, and was only able to sustain a surplus on a multi-year basis from 2004-08, ranging from 0-0.75% of GDP once the exceptionally large 18.5% of GDP surplus in 2006 is stripped out.
While there is thus significant downside risk to the IMF’s fiscal forecasts, there is much low-hanging fruit in the budget, with a rampant increase in inefficient investment pushing capex to 8.6% +/- 0.8% of GDP from 2018-20 and regressive and wasteful fuel and agriculture subsidies reaching 5.7% of GDP in 2021. The consolidation will mainly be achieved by reducing these line items, resulting in 4pp and 4.7pp of consolidation from 2022-25, respectively, plus an additional 3.4pp of new revenue measures.
As such, while the budgeted consolidation is likely an upper bound of what is achievable, it may be possible with significant commitment by the authorities. Indeed, data for 2022 shows the budget performing broadly in line with its target, with the primary balance running 0.4pp behind target on a commitment basis through May 2022. With nearly three-quarters of the consolidation frontloaded in 2022, much of the heavy lifting has already been done, notwithstanding some minor slippage.
The programme will also focus on clearance of domestic arrears, with plans to reduce the stock to zero by 2026 from a massive 15.8% of GDP at the end of 2021 (comprising 13.4% of expenditure arrears and a 2.4% backlog of VAT refunds). While arrears clearance is essential to improve fiscal credibility and boost private sector growth, the gradual pace of clearance means that arrears will continue to weigh on the private sector in the years to come and could threaten the IMF’s growth targets.
Based on the fiscal plans above, public sector debt (including domestic arrears and guaranteed and non-guaranteed state-owned enterprise – SOE – debt) is projected to decline from 155% of GDP (o/w 103% of GDP external on a residency basis) in 2020 and 133% of GDP (79% external) in 2021 to 107% of GDP (76% external) in 2025 and 67% of GDP (51% external) by 2032. This is before the impact of the restructuring is factored in, indicating a significant policy contribution to Zambia’s debt reduction plans.
The debt figures offer no major surprises, with Zambia’s debt position well-telegraphed by the Ministry of Finance. However, the IMF adopts a slightly broader definition of Zambia’s debt stock for purposes of the DSA than both the authorities and most other EM, encompassing central government debt, expenditure arrears, government-guaranteed debt, the nonguaranteed debt of fiscally important SOEs (namely ZESCO) and the arrears of that same SOE.
The latter two categories are the key additions, and their inclusion will be re-assessed over the course of the programme. This expands the 2021 public debt stock to 133% of GDP versus the authorities’ less inclusive definition of 126%, thus increasing the amount of relief required to meet Zambia’s DSA thresholds, but the broad definition of Zambia’s debt stock is appropriate, in our view, and will help maximise the likelihood that debt sustainability is restored after the restructuring.
To support the goals of the ECF, the IMF is seeking “a debt restructuring to ensure the program is fully financed and to achieve the objective of bringing the risk of debt distress to ‘moderate’ over the medium term". These goals will define the contours of Zambia’s debt restructuring, with relief calibrated to 1) fill Zambia’s external funding gap consistent with a given reserve accumulation target and 2) bring Zambia’s debt below the relevant DSA thresholds with enough of a buffer to absorb shocks.
On the former, the ECF aims to boost reserves to 5.5 months of prospective imports (US$6.2bn) by 2025, up from 3.2 months of import (US$3.2bn) at the end of 2021. The IMF projects an external funding gap of US$11bn from 2022-25 that will need to be filled to reach that target, which will be filled with US$2.6bn of official financing (split roughly evenly between the IMF and World Bank) and US$8.4bn of relief from the external debt restructuring over the programme period. This provides a baseline on the amount of “cashflow relief” (eg maturity extensions on principal payments and/or grace periods on interest payments and/or coupon cuts) required from Zambia’s external creditors.
On the latter, debt sustainability thresholds are generally well-defined for low-income countries and vary based on their 'debt carrying capacity' (see below). Zambia is classified as having a 'weak' carrying capacity, so would normally be subject to those thresholds. However, given an ongoing GDP rebasing exercise and associated measurement issues (which will likely result in an upward revision to Zambia’s nominal GDP, but not until end-23 at the earliest), the IMF instead chooses to adopt the external debt service-to-revenue and PV of external debt-to-exports ratios as the key indicators for its DSA.
Focusing on the debt service indicator, the IMF stipulates a reduction of the external debt service/revenue ratio to 14% by 2025 and maintenance at that level on average from 2026-31 to reduce the risk of distress to “moderate”. This implies some additional cashflow relief beyond the US$8.4bn that was already pencilled in from 2022-25, with the ratio projected to fall from 61% in 2022 to 44% in 2025, 45% in 2026, 30.5% in 2027 and 17% by 2032. External debt service would need to be cut by around two-thirds in 2026, one-half in 2027 and one-sixth in 2032 to reach the 14% debt service/revenue threshold, all else equal.
Turning to the PV of external debt/exports ratio, we get a much clearer sense of the total quantum of NPV relief required. The IMF says that in order to bring Zambia’s risk of debt distress to “moderate” over the medium term, “the restructuring would need to bring the PV of external debt-to-exports ratio to a level consistent with ‘substantial space’ to absorb shocks by 2027”, which it defines as 84% for countries like Zambia with a weak debt carrying capacity.
In 2027, the pre-restructuring ratio is projected to stand at 129% of GDP versus the 84% target, implying the need for 35% of PV relief on all external debt. Focusing on the “restructurable” debt stock (eg excluding super-senior multilateral debt), which the IMF estimates at 84% of the total external debt stock, required NPV relief rises to 41.5%.
Notably, the amount of relief is highly sensitive to the target year. Achieving the 84% threshold by 2025 instead of 2027 would, for example, require a c45% NPV cut on all external debt instead of 35%, while reaching the threshold by 2032 would only require a c15% haircut.
Further, World Bank President David Malpass said in a statement on Wednesday that an NPV haircut of 45% would be consistent with the joint World Bank-IMF DSA, pointing to a slightly more aggressive interpretation of the DSA than our read-through seems to imply.
Nonetheless, the DSA, taken together with Malpass’s comments, provides a starting point of 35-45% for the amount of NPV relief desired by the IMF to reduce the risk of debt distress to “moderate” over the medium term.
Areas for further discussion
While the DSA is a starting point for restructuring talks, its assumptions are up for debate and creditors are likely to raise several points of contention during the forthcoming discussions that will need to be ironed out before a final agreement can be reached, although how much of the programme can be changed now is debatable.
The discount rate used to calculate the NPV reduction by creditors is likely to be one of two key points of contention. If the government or official creditors insist on using the IMF’s preferred 5% discount rate to calculate the NPV value of debt relief, it will vastly understate the value of NPV relief offered by bondholders, for whom the market rate is most relevant. Conversely, official creditors may dispute the use of a market rate to value the NPV of bondholder relief, arguing that creditors with a lower discount rate will be disadvantaged by the need for greater nominal relief to achieve the same NPV target.
While the discrepancy between the discount rates used across creditor classes is an issue in every restructuring and is reconcilable, it is likely to be particularly acute for Zambia’s restructuring given the sharp rise in global interest rates (especially for HY EM debt). Indeed, the Fed funds rate is projected to rise to nearly 4% by year-end and the yield on the Bloomberg EM Sovereign HY index is above 11%, up from c7.5% at end-2021, c6% at end-2020, and c7.25% in the decade pre-Covid. Barring the post-GFC and pos-Covid spikes, the gap between market and IMF discount rates has never been higher.
The IMF’s chosen PV of external debt-to-exports threshold in Zambia’s DSA is likely to be a second key point of contention, as the targeted 84% is far below the 140% threshold that applies to low-income countries with weak debt carrying capacity (and even further below the 180% threshold for those with a “medium” carrying capacity, which Zambia would likely be upgraded to by 2027 if the reform programme goes according to plan). At a projected 129% in 2027, this ratio would already fall within those less ambitious thresholds without any sort of restructuring.
That said, the IMF report says that the threshold has been calibrated “to a level consistent with ‘substantial space’ to absorb shocks by 2027.” Indeed, given the fairly ambitious macro assumptions, it would be inappropriate not to leave some space for Zambia to absorb shocks while staying below the threshold. Further, Zambia may argue that the choice of exports as a denominator, which are projected to be north of 47% of GDP by 2027 on the back of rising copper exports, already flatters the debt outlook and lowers the amount of relief required from external creditors.
While these are likely to be the focus areas for bondholders, every aspect of the DSA will need to be digested and discussed during the negotiations. The perimeter of the restructuring will be another such area. While the IMF says that “restructurable” external debt is 84% of the total, bondholders may push for a greater proportion to be included in the restructuring (eg non-IMF and World Bank multilateral debt, which is c70% of total multilateral debt), thus reducing the amount of NPV relief required by bilateral and private creditors to meet the DSA’s targets.
Conversely, China may argue against the inclusion of non-guaranteed SOE debt (comprising US$139m owed by ZESCO) and the classification of the commercial component of guaranteed SOE as bilateral debt (with a footnote in the report highlighting that “the commercial central government and guaranteed SOE external debt backed by guarantee from official export-credit agencies” will be counted as bilateral debt). This classification ensures that this debt is restructured bilaterally instead of being left until later in the process, but contravenes the preference of China to classify the CDB as a 'commercial' creditor.
The macro assumptions will also need to be dissected. While the main assumptions (growth, inflation, etc.) seem to strike a 'conservatively optimistic' balance, there is likely room for discussion on the exchange rate assumptions. ZMW is projected to depreciate by c5% annually in nominal terms from 17.88/US$ in 2022 to 21/US$ by 2025, implying a broadly unchanged REER with inflation at 6-8%. However, recent ZMW appreciation has pushed the exchange rate to 15.5/US$ currently, implying a stronger starting point and potentially reducing the amount of relief required even further.
Looking further down the line, the downside risks to fiscal consolidation plans (see above) could cause yields to rise if slippage occurs. The IMF also acknowledges serious discrepancies in the balance of payments data (which we’ve previously highlighted), which are currently being studied and could negatively impact reserve accumulation over the course of the programme if current account projections need to be downwardly revised. While this will only impact recovery value if creditors do not see the reform programme as credible and increase the exit yield accordingly, it could erode the value of the bonds over time.
There could also be some pushback from creditors based on the relatively small amount of financing offered by the IMF, which totals just 100% of quota and c1-1.5% of GDP over the course of the programme. This compares to cumulative support to Egypt, which has reached over 700% of quota and is expected to be topped up even further with a new programme, and leaves bilateral and private external creditors to plug a larger portion of the funding gap with debt relief.
While IMF and World Bank commitments are likely set in stone over the course of the programme, there could be scope to reduce the amount of relief required in the post-programme period by adopting more generous assumptions on external financing inflows and rollovers and/or assuming a timely resumption of market access. While the IMF says the prospect of renewed market access is low over the medium term, market access could quickly be restored if Zambia sticks to its ECF targets (we estimate that on average it takes 2-3 years for a country to regain market access after coming out of default), so the IMF could relax this assumption as a vote of confidence in its own programme design.
The decision to exclude domestic debt from the restructuring is likely to be uncontroversial, as Zambia has long telegraphed its intentions on this front. The IMF says that given the large proportion of banking sector asserts (around one-third) comprised domestic government debt, “any restructuring of this debt could trigger significant financial instability, potentially requiring public resources to support the sector". Bondholders are unlikely to push back against this conclusion, as non-residents hold US$3.25bn of domestic government debt (28% of the total and more than the US$3bn face value of eurobonds).
That said, the IMF’s conclusion on this front could set an interesting precedent for Sri Lanka, whose domestic debt stock comprises a similar portion of banking sector assets but where non-residents do not hold any of the stock, likely making Sri Lanka’s desire to exclude domestic debt from the restructuring a key point of contention for creditors who will argue that debt cannot be put on a sustainable path without including domestic debt in the restructuring or, conversely, through unacceptably large relief from external creditors if it is excluded.
In Zambia’s case, the classification of non-resident holdings of domestic debt, instead, could be a potential area for discussion. Per its internal policies, the IMF classifies external debt by residency, meaning the US$3.25bn of domestic debt held by non-residents (equal to 12.8% of GDP) would meaningfully improve the DSA if reclassified on a currency rather than residency basis. However, this line of argument is unlikely to yield much fruit, and we think bondholders would be better off focusing their efforts elsewhere.
The cashflow relief requested from 2022-25 equals US$2.1bn annually, which is c90% of the US$2.3bn of annual external amortisations due over the programme period (which, notably, includes a US$750m eurobond amortisation this year, US$1bn in 2024, and the first of three US$417m payments on the ‘27s, before past-due interest (PDI) is accounted for).
Assuming at least 90% of the external debt amortisations due over that period are on “restructurable” debt (which is conservative – multilateral debt makes up just 5-8% of external debt service in 2022-23, per IMF data), then the desired cashflow relief can be achieved simply by extending maturities beyond the ECF period, implying there may not be a need to resort to coupon cuts.
However, given the elevated external debt service/revenue ratio from 2026-31, we think some additional cashflow relief is likely to be required (and although the DSA isn’t clear on this, there may be some resistance to a restructuring that merely pushes debt service into the post-programme period). For bondholders, this is likely to take the form of a light coupon reduction, as the weighted average interest rate on Zambia’s eurobond stock is relatively hefty 7.9% (5.375% on the ‘22s, 8.5% on the ‘24s and 8.97% on the ‘27s).
If maturity extensions and coupon cuts do not deliver enough relief to meet the proposed 35-45% NPV reduction, the remaining NPV relief can be offered via nominal haircuts (eg principal reduction). Different creditors will likely prefer different combinations of cashflow relief and principal reductions based on their own preferences and discount rates. China, for example, has historically expressed strong resistance to principal reductions, preferring instead to term out debt service payments.
Bondholders, on the other hand, could be more open to nominal haircuts if it allows for less aggressive coupon cuts and maturity extensions. Indeed, a preference for a 'normal' bond with reasonably remunerative and steady coupons versus a more esoteric structure with no or low initial coupons that step-up over time is likely a key relic from the widely ridiculed 2020 Argentina restructuring (where creditors faced limited nominal haircuts in exchange for a very low coupon step-up, effectively backloading payments, only to see the bonds collapse post-restructuring to cUS$25 currently).
Rising global yields may also make nominal haircuts more palatable, as the NPV reduction from smaller principal payments at maturity becomes lower relative to the NPV reduction of lower coupon payments earlier in the life of the bond as interest rates rise. Still, bondholders may ask to receive some sort of value recovery right (like a GDP or copper warrant) to be induced to provide nominal haircuts, giving them the opportunity to capture some of the upside if the reform programme goes well (especially if they think the IMF’s growth and copper production targets are overly conservative). That said, this would likely complicate and elongate the restructuring process compared to the 'cleaner' approach of issuing a relatively standard new bond.
In any case, taking the range of 35-45% for NPV haircuts implied by the DSA and Malpass’s comments allows us to back out bond prices without making any assumptions on bond structure. Taking the ‘27s as an example, the NPV of the bonds at a 12% discount rate is US$113.6 once the 22.4pts of PDI are factored in. A 35-45% NPV cut would imply a bond price of US$62.5-73.8 at a 12% discount rate, well above the US$57 price at cob on 8 September. Put differently, at the current price, the ‘27s break even at a more aggressive 50% NPV reduction, implying some upside.
While we are agnostic on bond structure (see above for considerations), for illustrative purposes we can back out a bond structure that achieves this degree of NPV relief. For simplicity, we assume the restructuring is completed in October 2023 in line with the IMF’s proposed timeline (which seems overly conservative – more on this below) and that all three of Zambia’s bonds are rolled into a single new instrument with a cUS$3.8bn face value (assuming all PDI is capitalised over that period – which seems likely given their inclusion in the DSA envelope but would be a significant downside risk if not) that matures in 10 years and amortises in equal annual instalments over the final three years (the same structure as the existing ‘27s) with a 6% coupon rate (around one-quarter below the current weighted average of 7.9%). In practice, there may be a preference for two or more bonds to establish a curve, although creating too many smaller issues risks reducing their liquidity.
Given this payment structure at a 12% exit yield, the principal would have to receive a c20% nominal haircut to achieve a 45% NPV reduction and c5% to achieve a 35% NPV reduction. In aggregate, we think this scenario represents a reasonably palatable bond structure for both the government and bondholders.
With bonds currently implying NPV haircuts of c50% versus the 35-45% implied in the DSA, and a projected recovery value in the low 60s to low 70s versus a current price of US$57 for the ‘27s, we think the risks are skewed to the upside for Zambia’s eurobonds. Indeed, we are a bit perplexed as to why bonds dropped by 2.4-3.8pts across the curve on 7 September on the back of the DSA release, which outlined a notably less severe restructuring than we had previously anticipated (though bonds did claw back 0.7-1.3pts of losses on 8 September, partly reversing drop the day before).
This apparent paradox could be explained by a reading of the headlines, many of which are misleading. While US$8.4bn of debt relief seems large relative to the US$16.8bn stock of public external debt, representing 50% of the outstanding stock, in reality, this only refers to the amount of cashflow relief required over the programme period and implies that maturity extensions may be sufficient without the need for any additional coupon cuts. This should be seen as good news for bondholders, and referencing the US$8.4bn request without this context makes for overly ominous headlines, in our view.
Furthermore, the debt reduction targets are only a starting point and will be subject to further negotiation as bondholders digest the DSA and begin to push for some input into the process (although in reality, it is not clear how much they will be able to change). After sitting on the sidelines while the DSA was formulated, bondholders will argue that they cannot be expected to have a predetermined restructuring imposed on them as a fait accompli, and a strong and well-coordinated committee may very well be able to improve the terms as currently outlined.
As such, recovery value risks seem to be skewed overwhelmingly to the upside, including the possibility of less ambitious debt targets and timelines and/or updated macro and exchange rate assumptions, among others. On the downside, the key risk is that the calculated NPV relief provided by bondholders is understated if it is valued at a below-market rate, thus requiring even greater NPV relief than currently outlined to satisfy the condition of comparable treatment, with an additional tail risk if PDI is not fully capitalised (though this is unlikely). These will all likely be key points of debate during the restructuring negotiations.
Conversely, the risks to bond prices after the restructuring are likely tilted to the downside given the risk of fiscal slippage over the course of the programme. As long as the programme is viewed as credible at the moment of restructuring then this will not impact recovery value via higher exit yields, but it could erode the value of bonds over time if the key fiscal and macro indicators begin to fall off target. The risk of slippage from ambitious targets could prompt some bondholders to offload their holdings post-restructuring to lock in any gains before this risk has the chance to materialise (although we are putting the cart ahead of the horse here).
Overall, the DSA represents the starting point for restructuring negotiations, and there could still be a long road ahead if the potential points of contention outlined above materialise into disagreements and delays. The target of finalising MOUs with official creditors by the time of the first programme review (April 2023) and with private creditors by the second review (October 2023), however, seems overly pessimistic, especially if these issues can be ironed out quickly. Conversely, the published timeline could be read as an indication that the IMF expects negotiations to be difficult and protracted.
Overall, we view the publication of the DSA as a key milestone for Zambia and its creditors, and a more constructive starting point from the bondholder perspective than we had previously anticipated. On that basis, we upgrade our recommendation on Zambian eurobonds to Buy from Hold at US$57 (23% YTM) at cob on 8 September on Bloomberg for the ZAMBIN 8.97 07/30/2027s.
As we have previously highlighted (see here and here), Zambia’s restructuring also has major implications for future restructurings. As a test case for the Common Framework, the IMF and G20 have a strong incentive for the process to conclude in a smooth and timely manner, with the handling of Zambia’s restructuring setting the precedent for future restructurings under the Common Framework.
And as a major borrower from China, the treatment of Chinese debt (in terms of what is included in the perimeter and how equal treatment is calculated and enforced, among others) will set a precedent for other restructurings by borrowers, like Sri Lanka and Ethiopia currently, with large Chinese debt exposure.
While we will refrain from rehashing our previous analyses (see links above for a more comprehensive summary), suffice it to say that sovereign creditors and debtors alike will be keenly watching Zambia for the signals it sends and precedents it sets in a world of elevated risk of sovereign debt crises.
Zambia: Restructuring timeline faces risk of delays, February 2022
Zambia: IMF agreement is crucial first step, December 2021
Zambia provides more clarity on debt stock, October 2021
Zambia: Extent of hidden Chinese debt revealed, September 2021
Zambia: Local debt attractive but eurobonds are stretched, September 2021
Zambia is treading down the wrong path, January 2021