When Loans Go Rogue – How Investors Read the Risk in Securitisations

If you’ve ever wondered what keeps structured credit analysts awake at night, it isn’t the copious cups of coffee they’ve downed that day – it’s two deceptively dull numbers that decide whether a deal sinks or swims.

The first is default frequency – how often the borrowers behind a securitisation are expected to miss their payments. The second is loss severity – how much value disappears when they do. Together, they form the heartbeat of every securitisation model, quietly dictating the pecking order of who gets paid, who doesn’t, and who ends up explaining themselves to the credit committee.

Get those numbers wrong, and what looked like a calm pool of loans can turn into a splashy mess. Get them right, and you’re the one who saw the storm approaching while everyone else was still admiring the prospectus.

So, what do default frequency and loss severity actually tell us about risk? And how do investors work out when loans are about to misbehave? Time to meet the first troublemaker: default frequency.

Default Frequency – The art of predicting who won’t pay

Default frequency, or probability of default if you’re feeling formal, is essentially a posh way of saying: how likely is it that borrowers will stop paying?

Analysts start with history. They look at how similar loans have behaved in the past. If UK prime mortgages have typically shown a 1% annual default rate, that’s your rough guide. But history has a habit of going off script. Lending standards shift, interest rates jump, economies wobble – and suddenly yesterday’s “safe” borrower looks a bit less dependable.

So, investors adjust. They look at:

• Borrower behaviour: credit scores, income stability, employment type.
• Loan features: interest-only or repayment, high or low loan-to-value, fixed or floating rate.
• Economic weather: inflation, job losses, and the inevitable “cost-of-living crunch” headlines.

And because misery loves company, there’s then the problem of defaults arriving together. In good times, they trickle in. In bad times, they come all at once – like buses, or unwanted calendar invites. A jump in unemployment or a dip in house prices can turn a few late payments into a proper traffic jam.  

And even the word default can mean different things. Some deals call a payment 90 days overdue a default; others wait until insolvency papers appear. That tiny definition in the documentation decides when alarms go off, cashflows are diverted, and analysts reach for something stronger than coffee.

Default frequency, then, isn’t fortune-telling. It’s about judging when borrowers might start to wobble – and how quickly that wobble could spread through the whole deal.

Loss Severity – How Bad is Bad?

If default frequency tells you how often borrowers trip up, loss severity answers the next question: how much it hurts when they do.

In simple terms, it’s the percentage of a loan that disappears after all the chasing, selling, and legal wrangling is done. For a mortgage, that might depend on how much the house sells for when it finally changes hands. For a credit card, there’s usually not much to recover beyond an apologetic letter.

Here’s where things get interesting. Two deals can have the same number of defaults but completely different outcomes.

• If a borrower misses payments on a well-secured mortgage, the property might cover most of what’s owed – low severity.
• If the loan was made against something that’s lost its shine (a falling housing market, a struggling business), there may be little left to recover – high severity.

Unsecured lending usually sits at the painful end of the scale. Once a borrower defaults, there’s rarely anything to claw back. Corporate debt lands somewhere in between: senior lenders may get a decent slice of their money back; those further down the ladder often get little more than a polite reminder of the risks they signed up for.

Legal systems play a starring role too. In some countries, repossessions are quick and predictable. In others, they take so long you could grow old waiting for the court date. Then add in lawyers’ fees, maintenance costs, and the odd tax bill, and the gap between what’s owed and what’s recovered widens fast.

So, investors don’t just ask “will they pay?” but also “what’s left if they don’t?” They’ll look at:

1. Collateral value – what the security is worth now, and what it might be worth if market confidence takes a turn.
2. Recovery process – how long, expensive and painful it is to get the money back.
3. Legal backdrop – how friendly the local rules are to creditors.

Loss severity won’t win any awards for excitement, but it’s what separates a bad month from a bad deal. Because when recoveries fall short, even the best-built structure can start to creak.

Putting it Together – Expected Loss and the Safety Cushion

Once you’ve got your default frequency and loss severity, you can finally start answering the question every investor actually cares about: how much pain are we in for?

Multiply the two and you get expected loss. It’s the simplest formula in structured finance and the one that decides whether your securitisation is a careful bit of engineering or a slow-motion headache.

Expected loss = probability of default x loss given default x exposure.

It sounds like the sort of formula scrawled on a whiteboard during a very long meeting, but it’s the one that decides how much protection the deal needs before investors start losing money.

That’s where credit enhancement steps in – the structural safety padding designed to keep the top investors dry while the lower tranches take the splash. Sometimes that means a slice of capital sitting beneath them, sometimes a reserve fund, sometimes a helpful guarantee from someone with deeper pockets. Whatever the shape, the mission’s the same: stop the safe end of the deal from joining the chaos.

Here’s where it gets tricky.

• A deal full of reliable, well-secured loans can get away with a thinner cushion – defaults might happen, but losses are usually small and predictable.
• A deal backed by riskier loans might only see the odd default, but when one hits, it hits hard. That means thicker protection, higher costs, and a lot more nervous glances around the table.

Then come the rating agencies, armed with their spreadsheets and scenarios, testing every “what if” they can think of – what if rates rise, what if unemployment jumps, what if everything goes wrong at once? They’re not trying to guess the future, but to see how much chaos the deal can handle before it breaks.

For investors, this is the real trick: turning those two little numbers into a structure that can survive a storm. Get it right and you’ve built a structure that keeps its head when the market loses theirs. Get it wrong and you’ve built an expensive game of Jenga.

Beyond the Numbers – What Investors Really Watch

Default frequency and loss severity aren’t the whole story. Once the models are built and the ratios neatly lined up, investors start looking at the parts no spreadsheet can capture – the things that make one deal run smoothly and another fall apart faster than a cheap umbrella.

First up, the asset pool.How diverse is it? A pool made up of thousands of small, unconnected loans tends to tick along nicely – a few borrowers stumble, most keep paying, and life goes on. But fill it with similar loans from the same region or sector, and a single economic wobble can send the whole lot tumbling. Diversification rarely takes the glory, but it’s usually the reason no one ends up explaining themselves to the regulators.

Then there’s the structure.
Every securitisation has its own plumbing – waterfalls, triggers and payment priorities deciding who gets paid first. Some are elegant, some resemble a plate of spaghetti. The details matter: a cleverly placed trigger can save a deal; a clumsy one can trip it up just when flexibility is needed most.

Next, the people behind the curtain.
Servicers, swap providers, trustees – the quiet custodians keeping the machine running. If one of them falters, everyone feels it. A solid structure can still be undone by a weak servicer, so investors look closely at track records, replacement rights, and who’s contractually obliged to tidy up the mess.

And finally, the legal landscape.
The best collateral in the world won’t help if you can’t enforce it. Insolvency law, enforcement rights, and local quirks can turn a simple recovery into a courtroom marathon. Cross-border deals add another layer of fun –  and more lawyers.

The spreadsheet shows what a deal should do; these moving parts decide what it actually does. It’s the difference between the recipe and the meal: the ingredients matter, but so does the cook.

How the pros do it

There’s no single way to analyse credit risk in a securitisation – just a handful of methods, each convinced it’s the clever one.

Some investors go bottom-up, testing every loan until their spreadsheets look like modern art. Others go top-down, starting with the economy and working backwards to see how much stress the pool can take.

Rating agencies usually do both: history for guidance, pessimism for safety. Nobody ever got fired for being cautious in a credit committee.

The best investors add a final layer of judgement – looking at who made the loans, who’s collecting them, and whether the reporting feels professional or a cry for help.

Because ultimately, two deals can look identical on paper but behave very differently in practice. The numbers matter, but the people behind them decide how the story ends.

The Last Word

For all its formulas and forecasts, credit analysis comes down to one thing: working out how much uncertainty you can live with.

Default frequency and loss severity give you the map, but not the weather. They’ll tell you where the bumps are – not whether the driver stays awake.

The irony, of course, is that when everything goes to plan, nobody notices. Because in this corner of finance, uneventful is the highest form of praise.