Last week was a roller coaster week for El Salvador bonds after hopes for an IMF programme, which the authorities had been discussing for several weeks, were put in doubt. This followed the government-controlled legislature's surprise removal of top judges and the attorney general last weekend (1-2 May), which provoked international condemnation and the suggestion that the US could use its voting power to block an IMF deal. It led to a sharp sell-off in El Salvador bonds on Monday (3 May), with prices falling by 4pts at the front end (‘23s) and 10pts at the long end (‘52s), according to Bloomberg (mid price basis). However, the bonds rallied later in the week after the authorities reportedly held an investor call on Wednesday (5 May), with prices rising by 5-6pts in the middle of the curve and 7pts at the long end.
The exercise in damage limitation appears to have done the trick, for now at least, assuaging bondholders and reigniting optimism about an IMF agreement. After yields spiked to over 7.0% on the 23s, and longer bonds rose some 100bps, with yields on the 32s rising to 8.1% and the 52s to 9.2%, yields had eased by the end of the week, to 6.1% at the front, 7.4% on the 10-year and 8.5% on the long end. But even after the rebound, the bonds were still generally down some 3-4pts on the week, as investors were left wondering what the US will do.
We see an IMF agreement as still the most likely outcome, with the IMF confirming last week that virtual talks are taking place on a possible programme, especially as the government may become more dependent on multilateral funding given its high funding needs and with limited alternative funding options. This would provide upside to the bonds, putting the convergence trade with Costa Rica back in play.
However, an IMF programme is not a done deal. The path to an IMF programme may not be straightforward. First is the economics. Given the high debt burden (we estimate public debt at 92% of GDP in 2020), the fiscal commitments alone may be demanding, and more demanding than the government is willing to tolerate, although admittedly previous signals from the finance ministry as to its willingness to undertake the necessary adjustment are encouraging. A sizeable primary adjustment will be needed to put debt firmly on a downward trajectory. Second is the politics. Investors may worry that the US may be more obstructive.
Hence there remains plenty of downside too, either in the form of no programme resulting in a muddle through, with the risk of more heterodox policies, or – in the limit – some kind of debt restructuring given high debt, high share of bonded debt and high financing needs (although the problem looks more like a liquidity problem rather than insolvency which might mitigate downside). El Salvador is already rated B3/B-/B- (with negative outlooks from Moody’s and Fitch), which suggests an elevated level of risk.
We assign a Hold to El Salvador bonds, with a yield on the 2032s of 7.4% (z-spread 595bps) as of cob Friday, 7 May on Bloomberg (mid price basis). Our expected values in a scenario weighted analysis, based on plausible restructuring and non-restructuring scenarios, are some 3-4% below current prices across the curve.
The catalyst for last week's events came over the weekend of 1-2 May when the new parliament voted to remove the country’s attorney general and five Supreme Court judges. It was the first sitting of parliament under its new composition since the legislative elections in February, in which President Nayib Bukele’s ruling Nuevas Ideas party gained a super majority winning 56 out of the 84 seats. Although Bukele was elected president in February 2019, his new party had no seats in parliament, which resulted in legislative gridlock.
Already perceived as an authoritarian president, the move was seen as a further attempt by the government to consolidate power and an attack on institutional checks and balances. It attracted criticism from foreign governments, business leaders and NGOs. The US expressed its grave concern. The Biden administration continues to make the case for rule of law as well as adhering to democratic norms, and has pressed President Bukele to restore a strong separation of powers. Investors fretted that the US response could put an IMF deal, which has been a key support for the bonds in the last few months, in doubt. Could the US use its vote to block an IMF programme, and if so, what happens? Without a deal, it is hard to see where El Salvador goes.
The legislative vote runs counter to what had been the conventional wisdom. The prevailing view was that the legislative majority would allow the president finally to pursue reform, after these had stalled in congress since his own election victory. Indeed, the argument used by the government to justify the court moves was that it was the courts which had blocked reform. It will be difficult to know if this is the case until we see reforms. If the US rhetoric cajoles the authorities into normal behaviour (there are no concrete actions as yet from the US, although some have mentioned suspending aid and sanctions), then investors will breathe a sigh of relief. Otherwise, the government could seek an alternative path and pursue less orthodox policies, especially given its limited policy flexibility in the context of a dollarisation (its ability to print money is severely constrained). This might see it rely more on autarky (eg like Argentina), domestic financing, short-term debt or forcing it to other sources of external finance (such as China – interestingly last Thursday, the finance minister said the country is not in talks with China, according to media reports, perhaps to head such concerns off at the pass). Alternatively, the authorities may be forced to restructure.
The investor call was therefore a positive sign and El Salvador bonds reacted favourably. Not only did the authorities reiterate their commitment to the IMF programme, noting that negotiations continue, subsequently confirmed by the IMF at its regular press briefing on 6 May, but in holding a call, it also shows that the authorities seem to care what investors think (ie knowing that the government needs to retain market access as part of its financing plan). This suggests that an IMF agreement is still the most likely outcome, if the US (and other IMF board members) can put aside their concerns, and reach agreement on a new programme.
Status of IMF programme
The authorities’ interest in an IMF programme is well known, and follows IMF support through an emergency disbursement under the Rapid Credit Facility (RCF) in April 2020 of US$389mn (100% of quota). Finance minister Alejandro Zelaya confirmed in early March that programme talks were underway. He said the authorities were after a US$1.3bn three year EFF (which we calculate would be quite big, at some 330% of quota), similar to Costa Rica.
The minister’s comments followed the ruling party’s legislative victory, which may have given the government the political cover it needed to seek an IMF programme, ending a political impasse which had hindered policy implementation. The impasse also undermined the availability of external funding sources, thereby increasing the country’s reliance on short-term domestic borrowing. The comments prompted a bond rally, after the authorities had previously squashed any suggestion of an IMF programme before the legislative elections during the IMF/WB Annual Meetings in October 2020.
The IMF confirmed during its press briefing on 6 May that, at the authorities’ request, staff are currently holding a virtual mission with the authorities to conduct the 2021 Article IV and to discuss the possibility of a Fund programme. But while events this week may have been a setback, the surprise is that it has taken this long to get a programme. Maybe it’s been a matter of waiting for the new congress to take shape.
However, politics might still get in the way. Last week’s events might also reveal cracks, or lack of coordination, between the technocrats in the finance ministry and politicians in government. Indeed, despite the investor call doing the trick and calming markets (the bonds reversing part of the losses earlier this week), the President’s subsequent comments that the constitution allows the assembly to remove judges brought further criticism from the US. This suggests that despite the best efforts of the finance ministry, politics might mean an IMF programme may not be a done deal.
What could the US do? If so minded, we think the US could either seek to impose more structural conditions, or prior actions, to ensure the independence and integrity of the judiciary and anti-corruption measures, reminiscent of Ukraine (which may raise the bar to reach an agreement on a programme and stick to it), or the Biden administration could even choose to be more obstructive if it is that concerned about the situation in El Salvador and what doing nothing may signal to other leaders.
Need for multilateral funding amid high financing needs
Moreover, multilateral financing is an important part of the government’s financing plan as it faces high public sector gross financing needs (GFN) in coming years. An IMF programme would be intended to unlock additional financing from other MDBs, including loans that were approved but not disbursed in 2020, according to Fitch. Without this, there are few alternative, reliable options.
The government estimates financing needs of around US$2bn in 2021. It was hoping for US$450mn from the IMF, along with previous commitments from the CABEI, IADB and World Bank of US$600mn, US$250mn, and US$200mn, respectively, according to media reports in March. Indeed, CABEI approved funding of US$600mn on 28 April (before the removal of the judges).
Meanwhile, Fitch estimates financing needs are higher, at US$2.4bn (9.2% of GDP) this year, based on a projected fiscal deficit of 7.5% of GDP and amortisation of US$543mn (excluding short-term debt and arrears). The next bond maturity isn’t until 2023 however, which gives some breathing space.
An SDR allocation may also help, although at only US$390mn on our estimates (based on a US$650bn general allocation), that’s not going to change things dramatically. El Salvador’s reserves were reported at US$2.5bn in April, according to Bloomberg.
Wide fiscal deficits threaten debt sustainability
Key to any IMF programme will be agreement on a fiscal path that restores debt sustainability, especially given the marked fiscal deterioration caused by the pandemic and El Salvador’s already high debt level.
We calculate the overall fiscal deficit widened to 10.2% of GDP in 2020, from 3.1% in 2019, based on central bank figures for the non financial public sector (NFPS) balance deflated by nominal GDP in the IMF WEO (note that the authorities’ figures for the NFPS balance, including the pension deficit, correspond to the IMF WEO general government measure). The primary balance fell from a surplus of 0.6% in 2019 to a deficit of 5.8% in 2020.
The fiscal outturns are modestly worse than the IMF projected in the WEO, which estimated the overall deficit at 8.1% and the primary deficit at 3.7% in 2020. The WEO projected the overall deficit to narrow to 7.5% in 2021, and stay there in 2022, and average 8% over 2023-2026 on unchanged policies. Under this path, public debt rises to 113% of GDP in 2026, compared to 88% in 2020. Yet, even that cautious fiscal path would need to be reviewed given the weaker starting point; the fiscal deficit was 2ppts wider in 2020, so the IMF’s expected 60bps adjustment in 2021 is moot.
But the deficit is expected to narrow this year. Fitch expects the deficit will narrow to 7.5% of GDP in 2021, due to a cyclical rise in revenues and lower pandemic related spending, and fall further to 4.9% of GDP in 2022.
Public debt rose to 91.9% of GDP in 2020, according to our calculations (using central bank figures for consolidated total public sector debt deflated by nominal GDP as per the IMF WEO). This represented a rise of c19ppts from 2019, exceeding what the IMF had expected at the time of the RCF approval in April 2020. It had projected a 12ppts rise from 70% to 82%. It also compares to the IMF gross government debt figure in the WEO of 88% in 2020 (we observe that the authorities’ debt figures, deflated by WEO GDP, are about 3ppts higher than the WEO over 2017-2020).
Debt sustainability questions
El Salvador’s public debt, despite being high, was assessed as being sustainable under its gradual fiscal adjustment scenario in the IMF’s last DSA contained in the RCF report from April 2020. Public debt was seen as at high risk because of its high level. Risks associated with gross financing needs were seen as benign and the long average maturity of the debt (twelve years) was a mitigating risk factor. GFNs were however still high, projected at 14.1% in 2020 and 10.5% in 2021, and averaging c7.7% over 2022-2023. The IMF’s new MAC DSA framework places more weighting on GFNs.
The IMF’s DSA noted that a primary surplus of 1.5% of GDP was needed to stabilise the debt (we think it would be a little higher now given the rise in debt since then). Moreover, the DSA noted that a primary surplus of 3.5-4.0% was required to reduce the debt burden to 74% by 2025, and 60% by 2030, in line with the government’s own debt targets (from 70.2% in 2019 and 82.2% projected for 2020).
But the situation has deteriorated since then. In our DSA, we calculate that the debt stabilising primary surplus is 2-3% of GDP, given a debt/GDP ratio of 92% and reasonable assumptions for real GDP growth (g) and the real interest rate (r). We assume trend growth is around 2-2.5% as per IMF WEO projections. While the IMF forecast real GDP growth of 4.2% this year, this represents a cyclical recovery after the pandemic-induced -8.6% in 2020. Growth has rarely been this high in recent history, only exceeding 4% once before (4.3% in 2006), and averaging a lowly 2% over the last twenty years. We assume a real interest rate of 5%, based on a nominal effective rate of 6% (derived as interest payments in time t over the debt stock in time t-1 using WEO data) and inflation of 1%.
We calculate that the required primary surplus to reach the government’s debt targets is of the order 5-6% of GDP. This is unlikely to be desirable or achievable, and could point to the need for some kind of debt reprofiling.
To put the size of the fiscal effort needed into context, the country ran a small primary surplus over 2017-2019, averaging about 0.7% of GDP, but averaged a deficit of 1.5% in the preceding 20 years. This makes it hard to expect that the country can durably achieve a surplus of even 2%, let alone something much higher (4% or more), although other countries have done so.
However, according to Fitch, the government has set out plans for a fiscal adjustment worth 4% of GDP over three years in its effort to get an IMF programme. This consists of 2ppts of spending cuts and 2ppts of increased revenue (through new taxes and improved tax administration).
But we don’t have specifics. Finance minister Zelaya has seemingly ruled out a VAT increase although has recognised that fiscal sustainability also needed a comprehensive reform of the retirement system. We estimate the pension deficit is about 1.2% of GDP in 2020 (albeit down from 2.2% in 2017). Clearly, however, tax increases and pension reform can be difficult to implement, as other countries have found out – including Colombia (not in an IMF programme) last week.
High debt and a challenging fiscal adjustment might therefore make restructuring a more likely outcome, especially if agreement on an IMF programme cannot be reached. A more realistic or feasible adjustment might therefore seek to distribute the burden evenly between creditors and taxpayers.
Assuming the goal is to reduce public debt from 92% of GDP (in 2020) to a target of 60% by 2030, that implies a reduction of 32ppts (or 35%). Assuming the burden of the required debt reduction is shared equally between creditors and the fiscal consolidation, that implies a c17.5% upfront nominal haircut and a c3.5% primary surplus, which might be deemed more achievable.
However, notwithstanding some cyclical improvement following an easing in the pandemic, that still implies a primary adjustment of something like 5ppts, which may be optimistic.
The composition of public debt, and within that, public external debt, is another consideration in any restructuring.
Public debt amounted to US$22.6bn in 2020, according to the central bank, and was divided roughly evenly between public external debt (PED) and public domestic debt (PDD), with PED standing at 45.3% of GDP and PDD at 46.6%, based on our calculations. Public external debt stood at US$11.2bn in 2020, according to the central bank, a 12% rise from 2019 (although we think the figure might be a bit higher, at US$12.1bn, if we try to reconcile central bank data with World Bank IDS data). Prior to last year, PED was fairly stable averaging US$9.7bn over 2017-2019.
Our estimates for the creditor composition show that international bonds appear to make up the bulk of public external debt. Bonds amount to US$7.7bn nominal outstanding according to Bloomberg, which would account for about 70% of PED and about a third of total public debt.
Adjusting World Bank IDS data (reported to 2019), which appears to underestimate the bond stock, we find that the bonds are about 63% of PED, still by far the biggest single creditor group, while multilaterals account for 34% of PED. Official bilateral debt is very small, at only US$346mn in 2019 (just 3% of the total).
The debt service burden on the bonds is also quite heavy. While the amortisation profile is fairly favourable, with the next maturity not until 2023 (US$800mn), interest is US$581mn per annum until then. The bonds carry an effective nominal interest rate of 7.6%. Debt service payments on the bonds over 2021-2025 – which more than spans the life of a possible IMF programme – amount to US$4.4bn, comprising US$1.7bn in amortisation (2023 and 2025) and US$2.7bn in interest.
El Salvador took advantage of improved conditions following the reopening of the primary market last April, issuing a US$1bn 30-year bond (2052 maturity) in July 2020, albeit with a high (9.5%) coupon.
The composition of PED might be an important factor for bondholders in the event of any restructuring. The high share of bonded debt would mean they would inevitably need to be bailed into a restructuring. On the other hand, there is very little bilateral debt while multilaterals would resist participation due to their senior (preferred creditor) status.
However, another consideration is that domestic debt is equally significant, amounting to US$11.5bn in 2020, and accounting for half of total public debt and nearly 50% of GDP. Even zeroing the international bonds, public debt would still be 60% of GDP – enough to restore debt sustainability but an unlikely solution. Hence, domestic debt may also need to be part of a restructuring. The government already restructured some domestic pension debt in 2017 after it missed a series of payments for pension certificates (CIP bonds), extending payments from 25 to 30 years, with a grace period and modified interest rates.
Indeed, domestic debt has risen more sharply over the last few years compared to external debt, given the lack of external financing. We calculate PDD has risen by 30% since 2017, while PED has risen by 15%.
Moreover, domestic issuance may be reaching its limit. Fitch notes that the government is near the legal US$1.5bn ceiling in short-term Letes and issued cUS$645mn in one-year Cetes in the local market.
Limited financing options in the local market may be the spur the government needs to do what it takes to get an IMF programme, in order to open up external sources of capital. Alternatively, if politics gets in the way, it might force the government into more heterodox measures. Given the government’s legislative majority, it could presumably also change the legal limits to issue more locally, although this could stretch capacity of the banks, and would be seen negatively by the market and IFIs.
Comparison with Costa Rica
Investors may believe after the investor call, and the renewed commitment to an IMF agreement, that the convergence trade with Costa Rica is back on. El Salvador 32s trade some 200bps wide of Costa Rica 31s (yields of 7.5% vs 5.5%).
That said, even Costa Rica’s new programme is not yet in effect. Despite the IMF board approving Costa Rica’s new programme on 1 March (a US$1.8bn three year EFF), the programme is still waiting the approval of the country’s National Assembly, pending (among other things) agreement on restructuring the public sector wage bill.
Even then, although El Salvador is seeking to emulate Costa Rica in its programme, El Salvador’s fiscal adjustment path may be more demanding than Costa Rica’s. Costa Rica’s public debt was 68% of GDP in 2020, and is projected to rise further in coming years, to a peak of 76% in 2023, as per the IMF country report. The overall fiscal deficit falls over the programme period to 3.4% in 2024 and to 2% over the medium term, compared to 8.7% in 2020. The primary balance shifts from a deficit of 3.9% of GDP in 2020 to near balance by 2022, a surplus of 1% in 2023 and 1.7% in 2024, by the end of the programme. The primary surplus is projected to rise to 2.8% in 2026. That’s still a demanding 5ppts swing over most of the programme, and 7ppts by 2024; the IMF’s DSA noted structural measures worth c5¼% of GDP over 2021-25. Although with public debt at 72% by 2026, that still leaves it higher than in 2020 (and about the same as expected this year), and still 15ppts higher than in 2019. The impact of Covid has a long-lasting impact on public debt dynamics. El Salvador’s starting point may be somewhat weaker, however, with a much higher initial public debt/GDP ratio.
Scenario analysis and expected recoveries
We run a scenario analysis to derive our fair value estimates for El Salvador’s eurobonds. We identify four scenarios – two “no restructuring” scenarios and two “restructuring” scenarios. To illustrate the impact across different bond maturities, but to keep things tractable, we focus on five bonds rather than the whole curve – the 23s, 25s, 27s, 32s and 52s. Our exit yield assumptions are scenario-specific to provide a bit more realism.
Scenario 1 assumes an IMF agreement, under which El Salvador bonds rally across the curve, converging towards their COSTAR equivalents. We assume the spread over COSTAR narrows, but is not eliminated, given El Salvador’s weaker fundamentals, growth outlook, institutions, and lack of policy flexibility due to dollarisation. We assume the spread over COSTAR ranges from 25bps at the front end (2023s) to 50bps in the middle of the curve (27s, 32s) and 70bps at the long end (52s). A 50bps spread on ELSAV 32s over COSTAR 31s is in-line with the average pre-Covid performance. This implies yields on El Salvador's bonds ranging from 3.4-7.5%, providing a lot of upside from current levels, ranging from 4% at the short end to 12% on the longer bonds.
Scenario 2 assumes no IMF programme and more of a muddle through, without restructuring, which continues the country’s reliance on domestic financing and potentially may force it to seek more non-traditional bilateral lending (eg from China), given that under such a scenario, multilaterals may feel they have reached the limits of their lending capacity. We assume yields widen under such a scenario given such a solution may not be seen as sustainable and the potential lack of transparency over bilateral and domestic financing, as well as possible more heterodox policies. We assume yields widen to 11%, which is roughly where they were in autumn 2020 when IMF lending was dismissed. This implies a lot of downside from current prices, ranging from 7% at the short end to c20% on the longer bonds.
Scenario 3 and 4 are restructuring scenarios. The former involves an upfront 17.5% nominal haircut, in line with our DSA, but unchanged coupons and maturities (which is a simplifying assumption). The latter involves a PV haircut, consisting of a 50% coupon haircut and five year maturity extension on the 23s and 25s, but no principal reduction (note, to keep things tractable, we don’t change other payment terms, and recognise that creditors could negotiate higher step up coupons to compensate for PV losses from maturity extension and to invoke intercreditor equity). In terms of PV impact, it is more severe than the modest nominal haircut, providing more liquidity relief (lower GFNs), as it recognises the fairly high debt service payments on bonded debt in coming years. It would reduce the effective coupon from 7.6% to 3.8%, and reduce cumulative debt service over 2021-2025 by about 70% from US$4.4bn currently to just US$1.4bn on our calculations (interest only, because of the maturity extension on the 23s and 25s, and averaging about US$272mn a year compared to US$545mn presently). We assume a 10% exit yield in both scenarios (lower than our usual 12% benchmark).
Under these assumptions, we estimate a recovery value on the ‘32s, for instance, of US$75 in the principal haircut scenario and US$63 in the coupon haircut scenario, expressed per unit of existing principal, at a 10% exit yield, which represents considerable downside from current prices of cUS$106.
We assign a probability of 60% to the IMF programme scenario – that is, an IMF programme is our most likely scenario. We assume as 25% chance to the muddle through scenario. We assign a 15% chance to restructuring, with an equal weight on each restructuring scenario.
On this basis, our estimated fair values are some 3-4% below current prices.
We assign a Hold to El Salvador bonds, with a yield on the 2032s of 7.4% (z-spread 595bp) as of cob Friday 7 May on Bloomberg (mid price basis).