Sri Lanka: Recovery values updated as IMF approval nears

We recently wrote how Sri Lanka’s restructuring was getting bogged down by China’s insistence that multilateral creditors accept haircuts and its failure to provide “specific and credible” financing assurances that would allow the IMF Executive Board to approve the agreement that was reached in September for a 48-month, US$2.9bn extended fund facility (EFF).

There has since been a breakthrough. China Exim Bank sent an updated letter last week extending its "firm support to Sri Lanka through a debt treatment" and the IMF has said that Sri Lanka has obtained the financing assurances needed to bring the programme to the Board, which is now scheduled for 20 March (as confirmed on the Board calendar). Sri Lanka has reportedly already signed the Letter of Intent, with all signs pointing to Board approval on or directly after 20 March.

In addition to the financing assurances, Sri Lanka has reportedly also completed all the prior actions needed for the Board to approve the EFF, culminating with a 100bps interest rate hike by the Central Bank of Sri Lanka (CBSL) earlier this month and the adoption of a more flexible exchange rate after removing the prior trading bands, which has seen LKR appreciate sharply 366/US$ towards the end of February to 319/US$ on Friday (a nearly 15% rise). It is unclear how durable this rally will be, though, with many commentators calling for a full (or more than full) reversal by year-end.

Although Board approval would be a major milestone in Sri Lanka’s restructuring process, it does not mark the end of the road by any stretch. For example, Zambia was granted Board approval in August, after the IMF adopted a relatively flexible definition of financing assurances, but China has since backtracked by insisting that the debt sustainability analysis (DSA) assumptions are revised and that non-resident holders of domestic debt and multilateral creditors participate in the restructuring, while bondholders have also taken issue with the DSA thresholds. As such, 6.5 months after Board approval, Zambia’s restructuring is still stuck in the mud.

In Sri Lanka, China Exim’s assurances reportedly mirror the earlier letter that was deemed insufficient, barring a statement of the goal of “finalising the specifics of a debt treatment in the coming months". Further, the letter reiterated China’s insistence that multilateral creditors “do their utmost to make contributions” to the restructuring, which could continue to serve as a roadblock, even after Board approval. As such, Sri Lanka could face the same difficulties that Zambia has converting financing assurances into actual debt relief.

Further, Sri Lanka may struggle to keep its reforms on track over the course of the programme amid a difficult political backdrop. Local government polls are scheduled for 25 April, after being delayed by seven weeks due to insufficient funding, while the opposition has spoken out vocally against the IMF-backed austerity measures and a strike has been launched by half a million public and private sector workers to protest the government’s tax hikes, electricity and fuel price increases, and privatisation efforts. We have already highlighted how Sri Lanka’s fiscal plans are unrealistically ambitious, and political resistance will test the government’s resolve.

Still, the finalisation of financing assurances for Sri Lanka would be a major step in the restructuring process, especially given recent concerns that disagreements with China could derail the process. A resolution to the deadlock in Sri Lanka could also be an encouraging breakthrough for the sovereign debt restructuring framework more broadly and could bode well for other countries, like Zambia, where restructurings have been held up for similar reasons, but it remains to be seen if it can be replicated in other cases where China is a major creditor or if disputes will continue to be resolved on ad hoc basis.

Updated recovery analysis

Once Board approval is granted, the staff report and accompanying DSA will be published, which will clarify the key DSA thresholds and provide greater clarity on the amount and type of relief that may be required to bring debt to sustainable levels. According to CBSL Governor Nandalal Weerasinghe, Sri Lanka aims to outline its restructuring strategy in April and engage with commercial creditors with the goal of reaching an agreement by the time of the first EFF review, which Prime Minister Wickremesinghe has said will occur six months after Board approval (implying sem-annual reviews).

In the meantime, the thresholds referenced in India’s letter provide a starting point to update the recovery analysis that we first published last March and updated in July to account for higher exit yields. In its letter, India references a public debt/GDP ratio of 95% by 2032 and gross financing needs (GFN) of 13% of GDP on average from 2027-32, split between 8.5% of GDP domestic and 4.5% of GDP external.

It is not clear what cashflow targets will apply in the 2023-26 period which the EFF spans or what quantum of net present value (NPV) relief the IMF is after but, on the face of it, a 95% debt/GDP threshold does not seem very ambitious and calls into question our previous assumption of a 30% nominal haircut, potentially pointing to some upside to our previous base case recovery value of US$33-40 at 12-14% exit yields (in line with current pricing of US$36.7 for the SRILAN 7.55 03/28/2030s).

Updating our DSA based on the IMF’s growth and inflation assumptions from the October 2022 WEO, the budgeted fiscal targets from 2023-27, and conservatively assuming a 400/US$ exchange rate by year-end and future movements in line with PPP already sees debt declining from 128% of GDP in 2022 (based on the recently released 2022 debt bulletin) to 104% of GDP by 2027. Maintaining the 2.3% of GDP primary surplus through 2032 and assuming that growth will converge towards 5% over the medium term would push debt to 85% of GDP by 2032 without the need for any relief.

Using more conservative assumptions of 3% real growth and the primary surplus reverting to 0.6% of GDP over the medium term (which it achieved in 2018 and was a three-decade high), debt still declines to 99% of GDP by 2032. With “restructurable” debt comprising two-thirds of the debt stock at the end of 2022, this would require a 10-15% nominal haircut on the remaining external debt to bring it below the 95% threshold in 2032. Either way, there does not seem to be a need for large nominal haircuts if the optimistic IMF and government macro and fiscal assumptions hold. We, therefore, revise our base case nominal haircut assumption from 30% to 10-15%.

Beyond the issue of nominal haircuts, our base-case scenario assumed that three news bonds would be issued maturing in 2028, 2033 and 2043 and amortising over the last three years to smooth out the repayment profile, with coupon step-ups from 1.5% to 7.5% in 1.5% increments for the ‘28s and ‘33s and from 2% to 8% for the ‘43s. But, with our ‘28s and ‘33s beginning to amortise during the EFF and the 2027-32 target window, respectively, and coupons stepping up to a hefty 7.5-8% during the 2027-32 window, it is possible that greater cashflow may be needed to satisfy the IMF’s thresholds.

Without more clarity on the near-term cashflow thresholds and data on the debt service profile over the next decade, it is hard to make an informed judgement about what sort of relief will be required. However, Prime Minister Wickremesinghe claims that Sri Lanka has cUS$6bn-7bn of external debt payments every year until 2029, which would be equal to 6.6-7.2% of GDP on average during the 2027-29 period. This implies 2.1-2.7% of cashflow relief would be needed to satisfy the IMF’s 4.5% GFN threshold, before factoring in positive contribution from any potential primary surplus.

Barring better data on Sri Lanka’s debt service profile, we make some simplifying assumptions. Our back-of-the-envelope calculation shows that a 50% coupon cut delivers 1.2% of GDP of cashflow relief over the 2027-32 period, while a 15% nominal haircut would deliver 0.2% of relief, together accounting for more than half of what would be required based on our extremely rough assumption of 2.1-2.7% of GDP of needed relief. Maturity extensions, we think, could easily deliver the rest, assuming bonds are termed out beyond the 2023-32 DSA period.

To keep our analysis tractable, we focus on the SRILAN 7.55 03/28/2030s, which we assume will be restructured into a new 20-year bond maturing in 2043 and amortising semi-annual over the last three years to maturity (eg the same as our previous analysis – this equates to a 13-year maturity extension and could be seen as overly harsh, but we maintain this assumption for consistency and to err on the conservative side). We also adopt a slightly more onerous coupon structure, assuming a 4% coupon from 2023-32 (eg roughly half the original 7.55% coupon on the ‘30s) and an 8% coupon thereafter (eg a slight pickup on the original coupon).

Using this new cashflow profile and assuming no nominal haircut yields gives a bond price of US$41.5-50.2 at a 12-14% exit yield, assuming the restructuring is completed six months after Board approval with the first coupon due on 30 March 2024. Assuming a 10% nominal haircut, the bonds fall to US$37.4-45.2 and, assuming a 15% nominal haircut, they fall to US$35.3-42.7, each of which is an upgrade to our previous US$33-40 base case.

Taking a different perspective, we note that the IMF has called for debt relief of 49% in NPV terms in Zambia at a 5% discount rate (which corresponds to 69% NPV relief at a 12% discount, per our calculations). This is based on a low-income country (LIC) DSA for a country with “low” debt-carrying capacity. If Zambia’s bondholders succeed in changing its carrying capacity to “medium”, we calculate that this would lower the required NPV relief to 24%.

As a market access country (MAC), Sri Lanka’s debt sustainability thresholds and required NPV relief should be less onerous than Zambia’s, all else equal (and they are coming from a similar starting point in terms of debt stock, service and composition, so this assumption should hold). This is already evident in the relatively light 95% debt/GDP threshold and should translate to lower required NPV relief (although the revised MAC DSA framework is relatively untested and Sri Lanka will be a key test case so there is still some uncertainty).

Indeed, this assumption is what underpins our preference for Sri Lanka’s bonds over Zambia’s from a relative value perspective, with the Zambia ‘27s trading at a 6.4pt (17.5%) premium to the Sri Lanka ‘27s despite the likely need for greater debt relief. That said, this is partly justified by accrued interest, with the Zambia ‘27s trading with 26.9pts of accrued interest versus just 12.4pts for the Sri Lanka ‘27s, so the comparison is not so straightforward.

As such, we calculate the impact on Sri Lanka's bonds if the IMF calls for NPV relief of 25%. Applying this to our new bond maturity and cashflow profile (above), we calculate that a 28% nominal haircut would be required to deliver 25% of NPV relief at a 5% discount rate, which would translate to NPV relief of 66% at a 12% exit yield and a recovery value of US$36.1 for the Sri Lanka ‘30s (again, roughly in line with current prices).

For illustrative purposes, using the same approach would yield a recovery value of US$31.3 with NPV relief of 35% at a 5% discount rate (implying a 36% nominal haircut and 71% NPV relief at a 12% exit yield) and US$40.95 with NPV relief set at 15% (implying an 18% nominal haircut and 62% NPV relief at a 12% exit yield). This implies a slightly lower cUS$31-41 recovery value range at a 12% exit yield under what we deem a plausible range of required NPV relief, with a US$36 midpoint.

Putting it all together, our new back-of-the-envelope recovery analysis gives us an updated recovery value range of US$31-50 at a 12-14% exit yield, a wider and higher band than the previous US$33-40 range. However, if haircuts indeed turn out to be negligible, then a 12-14% range for exit yields could arguably be overly optimistic, as an endpoint of 95% for public debt/GDP would leave Sri Lanka highly vulnerable to distress in the future (especially with the budget predicated on unrealistic fiscal assumptions), which makes the upper-end US$50 estimate unlikely.

Conversely, larger nominal haircuts may call for lower exit yields, potentially in the 10-12% range in our more conservative 18-36% nominal haircut scenarios, making the lower range estimate of US$31 unlikely as well. Together, these considerations would narrow our recovery band by lowering the upside and raising the floor, pushing us to adopt a qualitative recovery band of “low to mid 40s”. This still implies some upside relative to current prices, so we maintain our Buy recommendation on the SRILAN 7.55 03/28/2030s at US$36.7 (29.1% YTM) at cob on 10 March on Bloomberg.

We also maintain our Hold recommendation on the SRILAN 5 ⅞ 07/25/2022s at US$40.4, with the premium over the ‘30s rising back to 10% from parity in December. While this is down from highs of 30-45% at the peak of last year’s sell-offs, it is still on the high side, in our view, with more attractive entry points elsewhere on the curve.

Board approval of the IMF programme later this month could provide a boost to Sri Lankan eurobonds but, ultimately, any future changes will be dictated by the DSA once it is published (likely to happen later this month, after Board approval is finalised) and bondholder negotiations. Beyond the near-term horizon, potential difficulties converting financing assurances to actual relief could pose a downside risk to bonds, while potential slippage on reform targets could also threaten IMF and creditor negotiations and cap upside over the longer run by pushing up exit yields.

There are also still numerous Sri Lanka-specific issues to be worked out, including an ongoing suit by a holder of the ‘22s and potentially contentious decisions on how to treat domestic debt, SLDBs, guaranteed debt and short-term financing extended to the CBSL. While we are encouraged by the recent progress, it is too early for complacency and there are still ample opportunities for the restructuring process to be derailed. In the meantime, we await the IMF’s report for more concrete information on the restructuring.

Research

Access & Events

Country Risk Index

Sponsored Research

Macro & Market Data