1) Stephen Catlin claims insurers are $100-200B underreserved.
2) I show the math can’t get you there.
3) A $30B US shortfall is plausible.
4) Earnings estimates are too high across the board regardless.
First, A Pat on the Back!
Before we discuss last week’s “news” that Stephen Catlin thinks the insurance industry is > $100B underreserved, I have to take my victory lap!
It was only six weeks ago that I suggested buying futures on Bloomberg to be President. At the time, they were 200-1. Now they are 12-1. If you were smart enough to jump on a plane to Ireland to get your bet down, I hope you will do the right thing and leave me a generous tip!
Are Catlin’s Claims Credible?
At a recent Insurance Insider conference, Convex CEO Stephen Catlin dropped a hot take claiming the industry is $100-200B underreserved for casualty losses. This has all the makings of an IANS ALERT! CEOs are aware of the giant reserve hole but have kept the actuaries in the dark so they don’t have to admit it. He even accused the auditors of being in on the plot!
However, before issuing an ALERT, I need to dig a little deeper into this claim. First, Catlin provided no evidence for his allegation (or, if he did, II didn’t write about it). He didn’t share any data to indicate that losses were worse than companies are booking, any proof that auditors are looking the other way, or any math to justify such a bombastic conclusion.
Now, some might say an IANS ALERT has all the same flaws. The difference is I at least have compelling anecdotes that entertain and amuse! Plus, I think most people read them and recognize they sound like the sort of things that have been known to happen in this industry.
I have yet to indicate a magnitude, only a direction, of reserve deficiency. In other words, I am confident things are worse, but I have not said how much worse. I took this approach for good reason. Most people still believed (and, given the reaction to most earnings reports this quarter still do believe) there was no problem.
The consensus was pricing is improving and there was nothing to worry about. My goal was to make people aware that they were likely to be disappointed going forward if they believed that. There was no reason to pile on by throwing out a big number that was more likely to be wrong than right.
Let’s Do Some Math
But now there is at least a mini-consensus building that things are indeed worse, even if most believe it is only at a few companies and not broad based. Catlin’s broadside, if nothing else, highlights that more people are seeing the evidence of pain so it is probably time for me to go beyond direction and start addressing magnitude.
Given that most of the worries are about general liability (GL), commercial auto (CA), and professional liability (PL), I will start there. Medical is also worrisome though it’s a small line and there are still redundancies in workers’ compensation to offset the medical and any other areas of weakness.
The combination of GL + CA + PL is about $100B in annual premium and has about $175B of held reserves. If we assume the redundancy in comp offsets any weakness elsewhere, this means Catlin’s $100-200B deficiency suggests GL + CA + PL is ~ 50-100% understated. That sounds ridiculous!
Now, to be fair I am using US reserves, so if we bumped that to global perhaps it is $300+B of reserves. Still, he’s assuming a 50% deficit.
We can also look at it from a premium standpoint. If I assume loss ratios are understated by 10-15 points for each of 2017-19, that would produce a cumulative deficit of $30-45B in the US. Maybe that would get you to $75B globally (and that’s only if the shortfalls are the same internationally as in the US, which is unlikely). That’s still short of $100B, let alone $200B.
It’s also unlikely each of the last three years is equally bad. Even if 2019 is 15 points off, I would guess ’17 is perhaps 5 points and 18 is 10 points. That would get you to $30B US and perhaps $50B globally.
What If The Late 90s Repeats?
One other check we can do is look at the actual deterioration in results in the late 90s. There are some limitations as the industry data I have are calendar year, not accident year, but we can draw some conclusions.
Commercial auto in ’97 produced an 83% LR. That went up to 93% in ’01, so ten points worse and probably with some 9/11 cat losses in there. The next worst year was an 88% in ’00.
Other liability went from an 83% in ’97 to a 108% in ’02. However, there were significant asbestos charges in that number. I would guess ex-A&E the deterioration was perhaps half that.
In other lines, workers comp did deteriorate 20 points and medical malpractice by over 30 points. I guess if we were to use those numbers over multiple accident years in today’s problem areas we’d get close to Catlin’s estimate. However, that’s quite a stretch and I don’t think it’s a good base case working assumption.
Additionally, I tried to approximate an AY result by looking at the average calendar year LR across the ’97-02 block. For example, the 25 point deterioration in liability over the period is largely adverse development and comes from a number of prior accident years. It would be more helpful to spread that out over the entire block to gauge how much results deteriorated in total from initial expectations rather than attribute it all to the 02 AY.
Other liability had an average LR of 91, so 8 points worse than the starting point. Commercial auto averaged 87 vs 83 at the start. Comp went from 69% in 97 to an 82 average and med mal started at 88 and averaged 108.
This supports my estimate of a possible 15 point deterioration in GL + CA + PL for each of ’17-19 being a reasonable worst case. And again, that gets you nowhere near $100+B of deficiency.
The Goldilocks Analogue
I feel like Goldilocks trying to communicate the proper level of anxiety to have about reserve levels. Originally, nobody seemed to be concerned much at all. The porridge was too cold. Now, Catlin is dramatically overconcerned, aka the porridge is too hot.
The answer likely lies in between. There is a problem. Those who are saying they’re not seeing it or “we’re aware of it, but we don’t see it in our book” should be looked at askance. However, this isn’t likely to be the late 90s redux. Maybe it will get there, but it’s not where we are today.
I think the outcome that is “just right”, aka most likely is reserve releases will go away and accident year picks will rise (yes, even with more pricing being earned through). In some cases, development will flip to adverse and remain there for a few years.
This is a paradigm shift that I don’t think investors are prepared for. In fact, I plan to write more about it in the near future. Catlin doesn’t have to be right for reality to be a big disappointment for investors.