I still remember the first time someone walked me through a product that promised a fat coupon, a top rating, and almost no chance of going wrong. The room was calm. The slides were glossy. The models made extreme outcomes look like folklore. Twenty years on, I keep hearing a version of that same pitch, only the wrappers have changed. Liquidity is still easy. Spreads are still tight. Returns on plain assets feel thin. And once again, people are asking how to manufacture extra yield instead of asking whether the extra yield is worth the hidden bite.
When Confidence Looks Like Safety
The Global Financial Crisis did not arrive as a single lightning bolt. It arrived after years of comfort. Credit had been well behaved. Cash was cheap. Sophisticated structures were treated as proof of progress rather than proof of appetite. That is the part I find hardest to shake. The danger was not only leverage. It was the way benign years quietly narrowed the list of outcomes investors were willing to take seriously.
A structure launched two decades ago still sits in my head as a case study. It was designed to look like something for nothing: a high rating and a meaningful pickup over cash. The mechanics increased exposure as credit markets weakened. On paper, that sounded clever. In practice, it meant the product leaned into stress at the exact moment stress became expensive. One version tied to financial names came out with a top rating in early 2007 and had failed before the year was over. That is not a trivia fact. It is a reminder that ratings, models, and recent calm can travel together and still miss the point.
The problem was not that investors ignored risk. It was that years of quiet markets made some risks feel too awkward to price.
Today the ingredients look uncomfortably familiar. Funding is available. Investment grade credit does not pay a lot for the uncertainty it still carries. Expected returns on many public assets feel compressed. Leverage is finding new doors. A fresh wave of product design is pulling in money from people who simply want more income than the underlying market wants to give them. I am not arguing that 2008 is about to repeat itself beat for beat. I am arguing that the GFC lessons were about human framing as much as they were about bank capital.
The Product That Taught A Brutal Lesson
Call the old structure what it was: a constant proportion debt obligation dressed up as innovation. The pitch leaned on stability. Spreads had been orderly for a long stretch. Liquidity looked endless. Probability models treated violent widening as a rounding error. If you only looked at the recent past, the story almost made sense. If you asked what the product needed in order to survive, the story fell apart.
Success depended on spreads staying well behaved. The valuation framework effectively treated serious widening as too rare to matter. That is a quiet form of self-deception. Unlikely is not the same as impossible. Inconvenient is not the same as irrelevant. When markets finally moved, the gap between those two ideas became a pile of losses. I have found that investors rarely fail because they cannot name a risk. They fail because they assign it a probability so small that the position size never reflects it.
There is a sentence I keep coming back to. Many people did not start with, “What return do I need for this risk?” They started with, “This return is not enough. How do I juice it?” That inversion still runs through a lot of product conversations. Returns are visible. They sit on a term sheet. Risks hide in convexity, in funding, in the moment when everyone wants out through the same door.
Why Tight Spreads Feel Comfortable Until They Do Not
Look at investment grade yield the way a skeptical allocator should. Part of it is the government curve. Part of it is the extra slice paid for credit. When that extra slice gets skinny, you are not being paid much to own corporate risk. You can still own it. Plenty of portfolios should own it. But you should stop pretending the skinny slice is a gift from a market that has defeated the cycle.
I have sat in meetings where tight spreads were treated as evidence of quality rather than evidence of crowding. That is a habit. After a long stretch without a messy default wave in the parts of the market people care about most, the imagination shrinks. Models trained on that stretch become more confident. Committees become more impatient with cash. The search for yield stops being a tactic and starts being a personality.
- Abundant liquidity makes poor terms easier to accept.
- Compressed expected returns make leverage look like a solution.
- Recent stability makes tail events feel theoretical.
- Product innovation gives the search for yield a respectable wrapper.
None of those points is new. That is the point. The cycle does not need a brand-new sin. It only needs the old sin to feel modern.
The Question Investors Keep Asking Backwards
There is a commercially painful truth here. You can see rich valuations and still get punished for acting early. Markets can stay expensive longer than a cautious mandate can tolerate. Clients notice underperformance before they notice prudence. So people hold on. Then the repricing arrives in a rush, and suddenly everyone claims they always knew the emperor had no clothes.
That pattern is not a cognitive glitch in the textbook sense. People often recognize the risk. They just cannot afford, politically or commercially, to look cautious while the tape is still kind. I have watched that tension eat good analysis alive. The file note says “asymmetric.” The portfolio says “still in.” Then the bid disappears and the note looks prophetic at the worst possible time.
Returns seduce because they are printed in large type. Risks wait for stress to introduce themselves.
If you want a practical filter, flip the pitch. Do not ask how to lift the coupon. Ask whether the coupon is adequate once you admit that spreads can gap, liquidity can thin, and crowded trades can become funding trades. If the answer is no, the product is not clever. It is a demand for extra return with a delayed invoice.
Covenants, Carry, And The Slow Death Of Protections
High yield markets have already shown the movie. When money is plentiful, lenders give away the clauses that exist to protect them. Maintenance tests get softer. Restricted payments get easier. Collateral packages get messier. It feels fine while defaults sleep. Then a default cycle shows up and everyone remembers why those clauses were invented in the first place.
The yen carry trade is another clean example of enthusiasm turning into dependence. For years the trade was praised as a sensible harvest of rate gaps. Leverage built. Positioning crowded. The risk looked manageable right up until it did not. The unwind was violent not because the original idea was exotic, but because too many people needed the same exit at the same time.
Today the wrappers include leveraged funds, single-name vehicles, and leveraged single-name vehicles. Different plumbing. Same instinct. Use structure and borrowed exposure to squeeze extra performance from assets that no longer offer much on their own. I do not think every one of those products is a landmine. I do think the instinct behind them deserves more suspicion than it is getting.
Banks Look Stronger. Transmission Still Runs Through People.
We are told, often and loudly, that the financial system is sturdier than it was in 2008. In important ways that is true. Banks hold more capital. Balance sheets are cleaner. Some of the specific pipes that failed last time have been rebuilt. I will not pretend those reforms were theater. They mattered.
But investors have a habit of staring at the pipe they fixed and ignoring the pipe they did not. Risk still travels through the owners of the risk. If a leveraged holding falls and more collateral is demanded, people rarely sell the thing that already collapsed. They sell what they can sell. Assets that have not broken become the cash machine. Distress spreads because portfolios are linked, not because every security shares the same legal structure.
That is why “the banks are safer” can be true and still incomplete. Safer banks do not automatically mean calmer asset owners. A pension fund, an insurer, a family office, a levered exchange product, a hedge fund running a crowded factor book: any of those can become the transmission channel. In my experience, the market underprices that human plumbing until a week when everything correlates for no elegant reason.
| What improved | What still travels | Why it matters |
| Bank capital and cleaner books | Investor leverage outside banks | Losses can still force selling elsewhere |
| Clearer rules in parts of credit | Soft covenants when money is easy | Recoveries look worse after the fact |
| Better awareness of 2008 pipes | New products chasing the same yield | The wrapper changes, the need does not |
| More talk about liquidity risk | Confidence during tight spread regimes | Tails stay underweighted in models |
Innovation Is Not The Villain. Incentives Are.
I get tired of the lazy version of this argument, the one that treats every new structure as a moral failure. Markets invent tools. Some tools help hedge. Some tools help match liabilities. Some tools simply help a salesperson leave a meeting with a mandate. The difference is rarely visible in the acronym. It is visible in the incentive.
When underlying yields feel “relatively compelling” but still not compelling enough for the target return, product design fills the gap. That can be honest. It can also be a way to postpone an adult conversation about lowering the target. I have seen both. The second version is more common when committees are measured against peers who stayed fully invested in last year’s winners.
Perhaps the most interesting aspect is how quickly a strategy becomes a habit. First it is a sleeve. Then it is a core allocation. Then it is the thing the whole book needs in order to hit the number. At that point you no longer own a trade. You depend on a regime. Regimes change without sending a calendar invite.
What “Unlikely” Really Means In Credit
Credit spreads can stay tight for a long time. That sentence is not a forecast. It is a warning against false precision. A long quiet stretch does not prove the cycle has been retired. It proves that the last set of shocks was absorbed, postponed, or hidden in places the index does not emphasize.
A former political leader once talked as if the economic cycle itself had been tamed. Events disagreed. I hear a softer version of that boast in credit conversations now. Not always in those words. More in the shrug that says defaults will stay contained, refinancing will remain open, and any wobble will be met with enough liquidity to keep spreads from mattering. Maybe. I would like a better reason than “it has been fine.”
From historically tight valuations, betting against widening is not a heroic stance. It is a common one. Common stances get crowded. Crowded stances become fragile when the first forced seller shows up. You do not need a full-blown banking crisis for that fragility to hurt. You only need a change in the price of liquidity and a lot of people who sized positions as if liquidity were a public utility.
A More Useful Set Of Questions
Instead of hunting for the exact spark that widens spreads, accept that they can widen. Then work backwards. What does your book do if they do? Which holdings become cash sources? Which products need more collateral? Which clients call first? Those questions are dull. They are also the ones that survive contact with a bad month.
- Separate the visible coupon from the path of losses under stress.
- Ask whether leverage is solving a return problem or hiding one.
- Map who owns the risk, not only which bank underwrote it.
- Treat tight spreads as a valuation fact, not a character reference.
- Keep some dry powder that does not depend on a friendly bid.
Notice I did not say “go to cash and wait for the apocalypse.” That is a different religion. Plenty of long-term capital should own credit through cycles. The issue is price, structure, and the story you tell yourself about why this time the tail is a museum piece.
How Models Quietly Edit Reality
Models are not villains either. They are compressions. They take a messy world and make it discussable in a committee. The trouble starts when the compression gets confused with the world. If severe widening is assigned a tiny probability because the sample period was polite, the model is not measuring risk. It is measuring manners.
I have found that the most dangerous slide in a deck is the one that shows a smooth distribution with a thin left tail. Everyone nods. Nobody asks whether the tail was thin because the world is kind or because the data window was short. A product that dies when spreads jump should not be blessed by a framework that barely allows spreads to jump. That is not analysis. That is circular comfort.
Stress tests help only if they are allowed to be impolite. A two-standard-deviation wiggle that still leaves the rating intact is a brochure. A test that breaks the funding assumption, the rollover assumption, and the “someone will bid this” assumption is closer to adult supervision. You will not get every number right. You can at least stop pretending the ugly paths are too rude to print.
Leveraged Vehicles And The Illusion Of Precision
Leveraged exchange products and single-name geared vehicles are easy to mock and easy to use. They give a clean number. They reset. They look like tools. On a quiet Tuesday they behave. On a loud Thursday they can turn a market move into a funding event. The holder who wanted a little extra juice discovers that the juice came with a daily mechanical appetite for buying high and selling low.
That does not make every user reckless. Some people hedge with them. Some people trade them with a stopwatch and no illusions. The broader issue is the migration from tool to habit across a market that already feels underpaid for vanilla risk. When many participants use similar gadgets to reach for the same missing return, the gadgets stop being independent. They become part of the weather.
I keep a simple rule on my desk. If I cannot explain the loss path in one spoken paragraph, I do not understand the product well enough to size it. Fancy names do not count as an explanation. Neither does a rating. Neither does last year’s Sharpe.
Liquidity Is A Story Until It Is A Price
Abundant liquidity is a mood as much as a statistic. It shows up as tight bid-ask, easy refinancing, and a willingness to own complicated paper overnight. It feels like a feature of the system. Then one week it becomes a price, and the price is worse than the model’s “illiquidity premium” ever admitted.
Credit investors learn this the hard way because credit is not a video game with an infinite order book. When you need to raise cash, you sell what still has a bid. That can be your best name. That can be the government bond you promised yourself you would never touch. That can be the equity sleeve that had nothing to do with the original problem. Correlation is often just the sound of many people opening the same drawer.
A simple stress sketch: Tight spreads + easy funding = comfort Comfort + extra leverage = larger book Larger book + sudden bid gap = forced sales Forced sales + shared owners = wider contagion
That sketch is crude. Fine. Crude sketches prevent elegant disasters. I would rather sound unsophisticated in a memo than sophisticated in a postmortem.
What The Last Crisis Actually Taught, If We Are Honest
The useful lessons were never only about one product family. They were about humility under compounding confidence. They were about the difference between a risk you can name and a risk you can survive. They were about the way good times edit the imagination.
- Ratings are opinions with marketing value, not shields.
- Stability is a sample, not a law of nature.
- Leverage converts a tolerable drawdown into a process problem.
- Innovation can hide an old reach for yield behind a new label.
- Owners transmit shock even when issuers look fine.
Did we learn them? In parts of the banking system, yes. In the way capital still hunts for extra return when vanilla credit pays little, I am less impressed. Learning is not a commemorative plaque. Learning is the willingness to leave some yield on the table while everyone else is still collecting it.
A Personal Filter I Use On New Structures
When a new idea lands on my desk, I run a short, slightly unkind checklist. It is not academic. It is the product of sitting through too many meetings that aged poorly.
- What has to stay true for this to keep its rating or its coupon story?
- Who is on the other side, and can they disappear for a month?
- If the position falls, what else in the book becomes the ATM?
- Is the extra return payment for complexity, for illiquidity, or for hope?
- Would I still like this if I could not mark it for a quarter?
If the answers get foggy, I walk. Fog is information. Product teams hate that sentence. Portfolio survivors tend to like it.
The Credit Cycle Is Not A Puzzle You Solve Once
Every generation wants to believe it has professionalized the cycle out of existence. Better data. Better capital rules. Better central-bank vocabulary. Better risk systems. Some of that is real progress. None of it repeals the fact that credit is a claim on cash flows that can disappoint, funded by people who can get scared at the same time.
Are today’s investors sure the credit cycle has been defeated? A lot of pricing behaves as if the answer is close to yes. I am not in that camp. I also do not think certainty in the other direction is wisdom. The grown-up stance is duller. Own credit when you are paid. Be picky about structure when you are not. Refuse to treat a quiet decade as a permanent climate.
Gordon Brown’s old boast about ending boom and bust is an easy punchline now. The more useful version is inward facing. Which of our current boasts will look silly in a file review five years from now? My candidate is the idea that abundant liquidity plus tight spreads plus clever wrappers equals a market that has finally grown up.
Markets are often most exposed when the room feels smart, funded, and slightly bored.
What To Do With This Without Turning Into A Prophet
You do not need a crash call to use this lens. You need process. Revisit position sizes that only work if spreads stay polite. Look at documents you skimmed during the easy years. Ask credit teams to show the ugly path, not only the base case. Reduce the number of strategies that require the same funding weather. Keep a list of assets you would actually want if the bid for risk-free and near-risk-free paper got sloppy.
For income-focused investors, the temptation is strongest. A little extra yield feels like responsibility, especially when liabilities do not shrink to match a thinner opportunity set. That is understandable. It is also how covenants disappear and how levered sleeves sneak into “conservative” books. If you need income, say so. Then pay for it with duration, credit, or liquidity in a way you can explain to a skeptical trustee. Do not pay for it with a structure you only understand on the way up.
For growth-oriented investors, the issue shows up as crowded factor trades and levered expressions of a view that was already consensus. Same disease, different ticker. If the view is good, it may not need a multiplier. If the view is only good with a multiplier, the view may be an admission that the underlying price is no longer attractive.
The Uncomfortable Ending
Have we really learned the lessons of the last systemic scare? We learned some of the plumbing. We learned the slogans. We learned how to say “this time the banks are better.” We are less convincing when the conversation turns to tight credit, rising leverage outside the old center of the system, and a fresh appetite for products that turn a modest opportunity into a louder one.
I do not know the date of the next ugly widening. Nobody honest does. I do know that betting against it from rich levels is not a free lunch, even if the lunch has been free for a while. The last cycle punished people who confused a calm sample with a closed case. This cycle will punish a version of the same confusion if investors keep asking how to lift the return instead of whether the return is enough.
Spend less time hunting the precise catalyst. Spend more time admitting that spreads can move, that owners can transmit pain, and that innovation is often just hunger in a nicer suit. That is not pessimism. It is the only version of the GFC lessons that still earns its keep when the room is confident again.
