From stories in the popular press to dire analytical reports to pronouncements from Federal Reserve Chair Janet Yellen to statements from the OCC, a growing chorus is warning of a bubble in the leveraged loan market. Driven in part by the liquidity that (ironically) has resulted from the Fed’s zero-interest-rate/QE policy, banks and other lenders are said to be seeding the loan market with increasingly aggressive transactions and could end up reaping a bitter harvest of high defaults and stinging credit losses.
Our analysis of how far the pendulum has swung is based on hard data and is informed by conversations with loan market participants. One point on which everyone agrees: structure and terms are looser than they were earlier in the cycle, when liquidity was scarce and the economy was fragile.
Where views diverge, however, is on the question of whether we are witnessing the excesses normally associated with late-cycle booms such as those of the late 1980s (hyper leverage on thin-margin retail businesses), 1999-2000 (blueprint telecom deals), or 2006-2007 (highly leveraged, jumbo take-private LBO trades). Many market participants argue that the answer is no – or at least not yet.
To illustrate this point, the state of play in the leveraged loan market can be viewed along three broad lines:
- Credit: Today’s loan issuers sport higher-than-average leverage multiples, but levels are well inside the peak readings of the past. However, this theme is mitigated in two ways. First, today’s low borrowing costs offset higher leverage by bolstering coverage ratios. Second, few 2014-model deals carry either the heroic leverage associated with the large-scale LBOs of 2007 or the negative out-of-the-blocks cash flow.
- Risk/return: Though loan yields are low on an absolute basis, spreads relative to the base rate are still in the middle of the historical fairway.
- Terms and conditions: Loan T&Cs have evolved in the post-crisis period in a way that offers issuers unprecedented optionality, most notably in the widespread use of incurrence tests and incremental tranches.
We’ll take each of these in turn.
New-issue credit profile
Large LBO deals provide a consistent gauge by which to measure the credit hygiene of the loan market. And, indeed, the average leverage ratio at which these deals have been struck has crept higher in recent quarters from the more conservative levels of 2010-2011.
At the same time, PE firms are reeling in equity contributions:
Still, what is missing from the recent mix is the large tail of highly leveraged, thin-coverage deals that were rampant in 2006/2007. This chart, for instance, shows the share of LBO deals structured at 7x-plus leverage:
In the past, such loans produced an outsized share of defaults when the cycle ultimately turned.
As well, rising leverage levels are being ameliorated by low borrowing costs, which bolster coverage ratios, as this chart illustrates:
More to the point, the super-thin-coverage deals of the 2006/2007 period are not in play. So far this year, just 2% of LBO loans had an initial ratio of EBITDA less capex to cash interest of less than 2.0x, versus 64% in 2007.
Here again, narrower coverage produced the expected spike in defaults.
Of course, some of this cushion could evaporate if interest rates spike. After all, just 14% of the average debt stack for first-half LBOs is in the form of fixed-rate bonds, and provisions that once required borrowers to hedge half of their floating-rate exposure are a relic of a bygone era, participants say. Thus, if LIBOR climbs past the average floor of 1%, issuers could be in for higher interest costs, and loans will reset higher. That said, forward three-month LIBOR doesn’t cross the 1% barrier until January 2016, according to Bloomberg, with expectations for 2% in January 2017 and 3% in April 2019. As these figures imply, the market believes it is unlikely that rising rates will crimp coverage any time soon.
Credit profile of the S&P/LSTA Index universe
By most measures, the credit profile of the loan market remains robust, despite the fact that leverage levels across the S&P/LSTA Index have climbed steady – if slowly – in recent quarters as a result of (1) record recap activity and (2) as noted above, more highly geared new-issue transactions.
Loan issuers’ ability to service debt, for instance, has improved dramatically in recent years as a result of two salutatory trends: (1) muscular EBITDA growth since the recession ended in June 2009 and (2) shrinking financing costs.
The result: fatter cash-flow-coverage multiples across the universe of S&P/LSTA Index issuers that file publicly.
While wider coverage doesn’t rule out defaults, it obviously affords issuers greater wherewithal in the event of a downturn.
Looking at rating dispersion shows a similar pattern. The share of performing S&P/LSTA Index loans outstanding that Standard & Poor’s rates CCC+ or lower climbed to a two-year high of 5.98% by June, from a five-year low of 4.19% in January. In the interim, S&P downgraded a few borrowers’ loans to triple-hook territory, including Caesars Entertainment, Gymboree, Education Management, and Auxilium. Despite the recent run-up, however, the share of CCC paper in the Index remains well inside the historical highs of 10-13%, from 2009.
Yield versus risk
Lining up the loan market’s current risk/return proposition with those of the past is a difficult exercise, given today’s historically low interest rates. In terms of absolute yields, today’s loans clearly offer a low prospective return.
In part, this situation is a product of razor-thin underlying base rates. Loan spreads, by contrast, are more generous. This chart shows the same trend but uses the discounted spread over LIBOR.
Putting this trend in context, BB Index loans offered a discounted spread of L+361, on average, as of July 11, which is 23 bps wide of the historical median level (we use median here because the heights of the dislocation skew the averages). Single-B loans, meanwhile, were at L+473, or one basis point wide.
It gets more interesting when we exclude the benefit of LIBOR floors, which many players see as a way to compensate lenders for a particularly steep yield curve, with three-month LIBOR lingering in the 0.20-0.25% range and 10-year Treasuries recently in a 2.5-2.6% band. Excluding LIBOR floors, the current BB discounted spread is L+296, 15 bps inside the long-term median, with single-Bs at L+391, or 20 bps lower.
Taken together, these metrics show that the market remains in the historical fairway between boom and bust.
Looking at the data another way, the average discounted spread of the S&P/LSTA Index was L+441 as of July 11, which implies an imputed default rate of 2.83%, outside both June’s reading of 1.08% (excluding Energy Future Holdings) and managers’ average June 2015 forecast of 1.76%.
As this chart makes plain, the margin of safety has tightened significantly in recent years in the face of robust market liquidity and consistent economic growth. That said, loans are not priced today as if there will never be another default, as was the case in late 2006/early 2007.
Where the factors discussed above suggest a Goldilocks scenario in the loan market, the erosion in documentation protections is the poster-boy trend of the bubble brigade. Covenant-lite is the most visible example. So far this year, 62% of new institutional loans cleared with only incurrence tests, topping 2013’s prior high of 57%. As a result, a record 56% of S&P/LSTA Index loans are now covenant-lite, up from 46% at year-end 2013.
There is no doubt that managers see the proliferation of covenant-lite technology as an unfortunate outcome of today’s market liquidity. But the reason is not that covenant-lite loans are inherently more risky. In fact, the historical record shows that maintenance tests did not produce either (1) lower default rates or, more to the point, (2) lower losses given default. As reported, Energy Future Holdings’ $19.5 billion loan default, from April, pushed the 2008-to-date default experience of covenant-heavy S&P/LSTA Index loans to 19.0%, from 16.6%. By comparison, the corresponding figure for covenant-lite loans is 10.3%. Drilling down to loans that S&P initially rated single-B, the stats are similar, at 19.8% for loans with maintenance tests and 11.0% for those with incurrence tests. As for recovery rates, the average price at which defaulted incurrence-test-only loans exited bankruptcy was 70.0 cents on the dollar, versus 65.6 for loans with traditional tests.
That’s not to say that toggling from maintenance to incurrence tests is without consequence to lenders. Clearly, traditional financial covenants provide an effective repricing option for lenders when an issuer’s financial performance slides. Take the default spike 2008-2010. During that three-year period, issuers loosened tests on roughly 21% of covenanted loans that were outstanding at year-end 2007. On average, these waivers cost the borrowers in question a fee of 57 bps and 201 bps of incremental spread. Taking this calculation a step further, 86% of Index loans had maintenance tests when 2007 ended. Thus, from 2008-2010, lenders were able to increase the average spread of their legacy portfolio by 36 bps.
Apply those same statistics to today’s Index universe (44% of which is are subject to maintenance covenants), and the spread lift – based on the same pace of waivers for covenanted loans and average spread increase – would be just 19 bps.
Where covenant-lite is viewed more as a damper on future returns than a spur for higher credit losses when the cycle turns, managers do worry about the credit implications of free-and-clear incremental tranches. These provisions, which allow issuers to tap the market for additional loans unfettered by the restrictions of incurrence tests, have become a pervasive feature of newly minted covenant-lite loans over the past 18 months. Some parameters:
An anecdotal scan of recent deals shows that the vast majority of covenant-lite loans have an incremental carve-out under incurrence tests that allow issuers to add another turn, give or take, of fresh loans outside the scope of the covenant. In LCD’s review of the data, 2014 covenant-lite loans allowed free-and-clear tranches equal to 0.96x pro forma EBITDA.
The lenders have some protection, in that with rare exception – Gates Global being a notable recent example – loans are fortified with most-favored-nation provisions that do not sunset and therefore provide pricing protection for the life of the loan. So far this year, the average MFN is 50 bps. Moreover, managers have scored small victories by pushing back against excessively large incremental amounts. Recent examples include Vantiv, which scaled back its free-and-clear limit to $650 million, from $800 million; Phillips Medisize (to $80 million, from $100 million); and Nextgen (to $65 million, from $80 million).
Still, credit vigilantes worry that in an extreme case an issuer could raise an incremental tranche at some high rate to pay a dividend, thereby allowing a PE firm to get some (or all) of its bait back while lenders are left with a less creditworthy debtor vulnerable to default. Another potential downside risk: struggling issuers might use incremental tranches to effect a distressed exchange that converts subordinated debt into a lower-face-value amount of secured loans, but, in the process, dilutes first-lien collateral coverage. Those are just two obvious ways in which free-and-clear tranches could work to the detriment of secured lenders, managers say. Because free-and-clear technology is a recent innovation, however, there may be other pitfalls as yet unseen.
More issuer-friendly documentation in the form of incremental tranches, incurrence tests, and equity cures may well be a direct response to the rising liquidity in the loan asset class, which allows managers to enter and exit positions more easily and therefore may obviate the need for the buttoned-up terms that were necessary when the lender was married to a loan until maturity or repayment. Of course, this is cold comfort to managers who studied credit at the old school and are concerned that today’s less-restrictive terms may result in less-favorable recovery outcomes down the road.
Innings of a baseball game is useful shorthand to describe where the market stands relative to the overall cycle. Some managers, however, prefer to describe the current cycle as a three-game series, with the first game over and the second now underway. To wit:
- Game One (recovery/QE): That part of the cycle, which produced outsized returns for loans and virtually every other risk asset, concluded in 2013 as the Fed began tapering bond purchases and 10-year Treasury yields climbed from a decades-long low of 1.43% in July 2012 to the recent level of 2.5-2.6%.
- Game Two (rate/event risks): The large early-cycle gains of Game One are now in the rearview mirror, but there appears to be little risk of a default spike over the next 12-24 months. Instead, managers say the paramount risks of this part of the cycle are a sudden change in the interest-rate curve or a disruptive outside shock, either of which could send liquidity to the sidelines and cause loan prices to fall.
- Game Three (default/loss spike): This game will be played eventually, as it always is. The timing, however, will likely be several years into the future based on most economic prognostication, which views the chance of a near-term recession as remote – unless, of course, the U.S. economy is rocked by a catastrophic outside shock.
Given this framework, managers say the market is likely to be in the middle game of the cycle for a while longer. For the time being, the biggest risks for investors will be:
- Event risk, in the form of an outside shock that sends investors to the sidelines and causes outflows from loan mutual funds to accelerate, while the CLO window slams shut.
- Rate risk, such as either (1) an unexpected decline in rates or (2) a steeper rate curve that, in either case, could drive money from short-duration products like loans to longer-duration products like bonds, thereby crimping loan demand.
– Steve Miller
Follow Steve on Twitter for an early look at LCD analysis, plus market commentary.