You bought yield; then the issuer rewrote rules
The screen looks clean: a bond with a coupon that fixes the income gap in the portfolio without taking obvious equity-like risk. The offering deck says “flexible” and “modern” terms, which is easy to gloss over when the spread feels like the point. A quarter later, the first earnings miss hits, and the questions shift from “will they pay?” to “what are they allowed to do?” That’s usually when advisors discover the uncomfortable part: the contract wasn’t just light on restrictions, it was designed to be edited in practice—through baskets, add-backs, and definitions that move with management’s story.
What tends to sting isn’t default; it’s the issuer improving its position before any default is visible. “Covenant-lite” here doesn’t mean no covenants—it means the protections that would normally slow leverage, block asset leakage, or force equal treatment are softened into tests the company can pass by reclassifying items or using permissive capacity baskets. The yield you “bought” was often compensation for that optionality. In a stress path, the issuer can refinance into priming debt, shuffle collateral, or upstream value to shareholders while staying technically compliant, leaving the bondholder with the same coupon and a worse recovery setup.
First filter: what loss you can’t afford
Once you accept that the documents can be “updated” midstream, the first job is less heroic: decide what drawdown you can actually live through in this sleeve. In an income portfolio it’s rarely the absolute default probability that breaks discipline; it’s the path. A bond that can gap from 98 to 82 on an amendment headline creates a different problem than one that drifts to 92 on widening spreads, even if both ultimately pay par. The constraint is usually real—client withdrawals, a quarterly performance bogey, a policy limit on below-investment-grade exposure, or simply the fact that selling into a down tape turns a paper loss into a permanent one.
So the filter is “maximum tolerable loss before the portfolio starts making decisions for you.” Translate that into a recovery assumption and a time assumption. If the business is likely a 40–60 recovery in a restructuring but the covenants allow assets to move away from the issuer, your floor is not 50; it might be 20–30, and the position size has to match that reality. After that, covenant analysis stops being academic: you’re only debating terms on bonds you can afford to hold through the ugly part.
Spot the three moves covenants should block
With sizing anchored to a realistic downside floor, the next step is to watch for three issuer “moves” that usually show up before the credit breaks, and that decent covenants should at least slow down. First is leverage creep without a clean refinancing moment: incremental debt baskets, ratio debt definitions that add back too much, or EBITDA games that let the borrower re-lever while the bond is still trading near par. That matters because the mark-to-market hit tends to land fast, long before any rating action, and it collides with real constraints like client liquidity or a mandate drawdown limit.
Second is asset leakage: selling collateral or key subsidiaries to an unrestricted entity, licensing away IP, or doing sale-leasebacks that turn owned assets into fixed claims senior to you. Third is priority shifting: new liens, priming debt, or guarantees that appear in the wrong place—especially through “permitted liens” language that sounds narrow but isn’t. If the documents don’t make these three moves expensive, the yield isn’t income; it’s payment for optionality you just handed to management.
When ‘restricted payments’ still leak value out
After those three “moves” are on the radar, restricted payments are usually where the document starts to feel generous without looking reckless. The label reads like a hard stop on dividends and sponsor extractions, but the math often turns it into a slowly refilling allowance. A builder basket tied to “cumulative consolidated net income” (plus big addbacks) can grow even while leverage stays high, and “available amount” language can sweep in equity proceeds, asset sale proceeds, or permitted debt incurrence that effectively manufactures new RP capacity. In a portfolio, the friction shows up as timing: the bond can trade fine until the quarter the basket clicks large enough, then you’re forced to decide whether to eat the spread widening or accept a weaker credit.
The leakage doesn’t have to look like a dividend. Watch for payments routed as management fees, monitoring fees, or “permitted payments” to affiliates. More subtle is value leaving through “investments” into unrestricted subsidiaries, where cash and assets move outside the restricted group and your recovery perimeter shrinks. If restricted payments are paired with loose “investments” and “transactions with affiliates” tests, the covenant can still read protective while the collateral base quietly thins.
Debt can rise without ‘debt’ technically rising
The next surprise is how often leverage climbs while the “Debt” line in the covenant stays compliant. It usually starts with instruments that behave like debt in stress but are drafted as something else: operating leases, receivables facilities, supply-chain finance, or vendor notes that sit outside the defined debt buckets. The issuer doesn’t need a big refinancing event to do it; they can layer these on quarter by quarter, then point to the covenant calc and say nothing material changed. Meanwhile, the bond market notices the fixed-charge load rising and reprices the risk faster than any leverage ratio in the documents will admit.
In review, I treat this as a definition problem with a real portfolio cost: the day spreads gap wider, the exit is expensive and the client asks why “leverage is up” if the covenants weren’t tripped. Look at what’s excluded from “Indebtedness,” whether lease-adjusted leverage is ever tested, and how “Permitted Debt” interacts with cash netting. If cash can be netted broadly while new obligations stay gross, the covenant can show flat leverage even as the company becomes harder to refinance.
Security and collateral: protection or illusion
By this point the documents may still look “safe” because the bond is described as secured, sometimes even first-lien. In practice, the constraint is what the collateral package actually covers on day one and what it’s allowed to become later. If key assets sit in subsidiaries that never guarantee, if foreign ops are excluded, or if “excluded assets” carve-outs take the crown jewels off the table, the security label can be more marketing than recovery. The tell in a portfolio is the spread: a bond priced like it’s protected, but structured like it’s only partially inside the fence.
Then comes the slow erosion. Permitted liens can stack on top of you; secured debt can migrate to different collateral; asset sales can swap hard assets for cash that’s quickly upstreamed. Even clean-looking intercreditor terms can still leave you financing a shrinking base. The test is blunt: in a downside case, can you describe what you actually get, how fast, and who gets paid ahead of you.
Turn covenant strength into a comparable verdict
At the point where “secured” is no longer comforting on its own, the practical move is to score the paper the way you already score yield and duration: with a repeatable verdict you can defend in a client review. I keep it simple—three buckets tied to real portfolio pain. (1) Leverage control: how much incremental debt can appear without a true refinancing? (2) Value leakage: how easily can cash, assets, or guarantees leave the restricted group? (3) Priority risk: how quickly can someone get ahead of you through liens, priming, or structurally senior subsidiaries.
Give each bucket a 1–5 and force a written “break point”: one sentence on what would have to happen for recovery to fall from, say, 50 to 25. The constraint is time—this has to work across an ETF sleeve and a handful of individual issues—so the goal isn’t perfect legal nuance. It’s comparability. When two bonds offer 150 bps of yield difference, you can decide if that spread is paying for a weaker score, or if you’re just underwriting disappointment.
