The Dangers Lurking In Excess Casualty

Most of you reading this are likely aware of the pressure on casualty loss trends due to rising social inflation and aggressive litigation tactics.

However, unless you work in casualty excess or reinsurance, you may not be aware of how destructive rising casualty inflation can be.

There is a significant misunderstanding out there about how loss inflation can affect insurer’s results. Most people think if loss trend doubles, then if you double you’re pricing, you have kept pace.

This is incorrect. In primary business, raising rates equal to claims inflation actually improves your combined ratio (see below)!

However, for excess business (and XOL reinsurance), raising rates commensurate with inflation can lead to very painful results.

Let me knock out the primary case first. The simple explanation here is expense ratio. Non-commission expenses are unaffected by pricing decisions. So, while your loss ratio remains stable, your ER improves (assuming your price actions don’t lower PIF).

To stay at breakeven, you would only need to raise the price 82% for a 100% increase in loss trend.

	Premium	$ Loss	$ Commission	$ G&A	Total Expense
Base Case	100	70 (70%)	15 (15%)	15 (15%)	100 (100%)
2X Base	200	140 (70%)	30 (15%)	15 (7.5%)	185 (92.5%)
82% Price	182	140 (77%)	27 (15%)	15 (8%)	182 (100%)

Excess Casualty

However, when we switch this to excess casualty, something very different happens.

But before addressing the math, let me make a quick analogy. Excess casualty (or cat reinsurance for that matter) is much like corporate debt.

When times are good, the benefits accrue to the equity owners (and primary insurers through lower loss ratios). When times get rocky, equity owners and primary insurers take the pain while the excess insurers and debtholders remain unharmed.

However, when things really get ugly, the debt suffers large losses. This also applies to the excess insurers. They can get overwhelmed by the leverage.

So, just like cat reinsurers don’t care about an increase in tropical storms, excess casualty writers don’t care if loss trend deviates, say 10%, from pricing.

But if trend is up 50-100% cumulatively, this is like a cat 3 hurricane turning into a cat 5. There will be devastating losses.

Alright, let’s dive into the math…

Base Case: $10M XS $25M @3% ROL

Note, for all these examples, I’m going to try to keep things as simple as possible. I know my “loss distributions” don’t look like a real loss curve. I’m not suggesting the frequency is accurate, nor that the ROLs represent market.

I merely tried to come up with a simple enough example for illustration purposes that produced a reasonable loss ratio. Hopefully, that preempts any complaints about assumptions.

So, as noted above, I’m assuming an insurer is writing a $10M layer above a $25M primary retention and charging 3%. For simplicity, I am assuming they have a portfolio of 100 clients, so $1B of limit at risk and earning $30M of premium.

I also assume 90% of clients are loss free while ten have claims. However, some of those claims will be within the $25M retention. My “loss distribution” is below:

Insured Loss	$10M	$10M	$10M	$15M	$15M	$15M	$20M	$25M	$35M	$50M
Treaty Loss	$0	$0	$0	$0	$0	$0	$0	$0	$10M	$10M

So two of one hundred clients produce a claim on the treaty, both for full losses. This results in $20M of claims relative to $30M of premium for a 67% LR. Sounds reasonable, I hope.

What happens when we introduce inflation to this treaty?

Scenario 2: Losses & ROL Double

Let’s assume that over a period of years claims inflation doubles, so what would have been a $15M loss (with no claim to the excess) is now a $30M loss (resulting in a $5M excess claim).

Let’s also assume insurers responded to this news with higher prices and the ROL doubled from 3% to 6%.

You might think the loss ratio would remain at 67% since we doubled pricing in response to doubling inflation. Let’s see if that’s correct…

Insured Loss	$20M	$20M	$20M	$30M	$30M	$30M	$40M	$50M	$70M	$100M
Treaty Loss	$0	$0	$0	$5M	$5M	$5M	$10M	$10M	$10M	$10M

Our premium doubles from $30M to $60M. This is helpful since the max loss per account can still only be $10M.

However, where before only two of ten losses reached the excess layer, now seven do! This results in excess claims of $55M and drives our LR all the way up to 92%!

This is quite a stunning result! As a primary, doubling rates for double the claims keeps your loss ratio steady, but not in excess.

So what is an excess insurer supposed to do? The most common response you’ll hear is raise their attachment points.

While this makes logical sense, does it really help much? Let’s find out.

Scenario 3: Move to $15M XS $35M

The next layer up is likely 15X35. So the excess carrier has moved further away from the loss ($35M vs. $25M), but now has more risk per event ($15M vs. $10M).

For consistency, I’ll reproduce our pre-inflation scenario first. I’m going to assume the ROL is 2.5% instead of 3% since you’re further away from the loss.

Insured Loss	$10M	$10M	$10M	$15M	$15M	$15M	$20M	$25M	$35M	$50M
Treaty Loss	$0	$0	$0	$0	$0	$0	$0	$0	$0	$15M

Note what’s changed – only one claim hits the treaty now instead of two. This makes sense since you’re higher up in the tower. Total losses drop from $20M to $15M.

The premium falls from our original $30 to $25M at the lower ROL. The new LR is 60% (15/25), so close to our base case 67%, but a little lower to account for the volatility risk of higher layers.

Now, let’s replicate the inflation scenario. I’ll assume the ROL went from 2.5% to 4% rather than doubling. We can debate this if you like, but there tends to be less pricing power at the higher layers so I think this is realistic.

Insured Loss	$20M	$20M	$20M	$30M	$30M	$30M	$40M	$50M	$70M	$100M
Treaty Loss	$0	$0	$0	$0	$0	$0	$5M	$15M	$15M	$15M

The inflation results in three full losses and one partial. Relative to $40M in premium, this produces a 125% LR, more than double the pre-inflation case! Even if I let the ROL double to 5%, this would still be a 100% LR.

But we’re not done yet. Let’s do one more example.

Scenario 4: Move to $20M XS $50M

One critique you can offer of the third scenario is the insurer didn’t move far enough up the tower. If inflation doubled, then they should double their attachment point.

Fair enough. Let’s look at a case where the excess carrier moves from 10X25 to 20X50. How does that change the outcome?

I’m going to use a 1.5% ROL here, so half of the original 3% since the attachment is twice as high (arguably it should be lower). That results in $15M of premium.

Insured Loss	$10M	$10M	$10M	$15M	$15M	$15M	$20M	$25M	$35M	$50M
Treaty Loss	$0	$0	$0	$0	$0	$0	$0	$0	$0	$0

Losses in this case drop to zero! Now, obviously that isn’t the expected outcome. There is tail clipped from my simulation. However, when you attach this high, there will certainly be a number of years where the loss ratio is zero.

Remember, due to the lower ROL, it only takes one full loss to produce a LR over 100 (20/15 = 133), so if it helps to frame it, assume every other year one limit loss is expected.

Now, what happens when we once again double everything? It’s not pretty!

Insured Loss	$20M	$20M	$20M	$30M	$30M	$30M	$40M	$50M	$70M	$100M
Treaty Loss	$0	$0	$0	$0	$0	$0	$0	$0	$20M	$20M

We get two full limit losses. Premiums have doubled to $30M (3% ROL). This yields a 133% LR!

Once again, moving up the tower, even at twice the ROL, produces horrific loss ratios.

What have we proven here? Moving up the tower to escape loss inflation can easily backfire. If a company is telling you this is their strategy, you should express ample skepticism.

Summarizing the Scenarios

The final table is a summary of all the scenarios. You can see that in every scenario that doubling price in response to doubling of claims inflation is inadequate.

Instead, it takes tripling or even quadrupling the price to maintain the original loss ratio (per the last column). Finally, the results get worse as you move to higher excess because of the leverage impact.

Excess Layer	Base LR	2X LR	Price Needed
10×25	67%	92%	2.75X
15×35	60%	100%	3.33X
20×50	0%*	133%	>4X

So what do these examples teach us?

First, it is dangerous to write excess when loss inflation is high, even if you are raising pricing meaningfully.

Second, moving up the tower is often a mistake that leads to lower premium and higher loss ratios.

Third, just like a cat book, you can have good results for a long time before things turn south suddenly.

Note in the examples how many claims were just shy of attaching at lower rates of inflation. Once inflation takes off, a lot more claims reach the treaty resulting in the feared “frequency of severity“.

This latter feature makes it very easy to get reserves very wrong. Excess players can be in denial much longer than primaries. But they are also likely to have far larger reserve charges when they hit.

Fourth, this analysis only focuses on one accident year. Remember, inflation doesn’t double in a year. It takes five to ten years.

That means an insurer will have a string of underpriced excess vintages. If the insurer followed the “raise rates with estimated loss inflation” path for five years or more, they will have compounded the error and have an even larger future reserve true-up.

Hopefully, this has been instructive for those who don’t specialize in excess. Be very careful trusting the reserves of excess casualty writers in the current inflationary environment, especially those who write at the top of towers. The same advice applies to excess of loss casualty reinsurers as well.