The $190M settlement Johns Hopkins Health System announced last month is a stark reminder of the importance of a sound, responsive risk transfer program. Investigators found over 1,300 videos and images and plaintiffs’ attorneys estimate more than 8,000 patients could have a claim. Johns Hopkins noted in the communication that the claim would be covered by insurance and their ability to pursue the mission would not be impacted.
The first thing to remember when designing your risk transfer program is why do you purchase this coverage? Is it to protect the balance sheet against variation in known risks? Protect the organization against Black Swan events? Both?
Protecting against variation in known risks deals with the attachment point of insurance – an interesting discussion but one where errors cost the organization a million here or there. I’d like to focus on the latter – protecting against Black Swan events where errors can be the difference in continuing to exist. To be more specific: how do we protect the organization against Gray Swans (known, large risks) and Black Swans (unknown unknowns)?
Protecting an Organization from a Gray Swan
Thinking about these large, unlikely scenarios is difficult, and humans tend to make repeated errors when doing so. First, we place too much emphasis on our own experience and have short term memory. Second, we forget that the future is under no obligation to act like the past. Consider the following logic: “well our largest claim has only been six million dollars and our risk management program is really strong so we don’t need any more limit.” And lastly, we don’t think what has happened to others can happen to us. But if it can happen to Johns Hopkins, why couldn’t it happen to you?
Next, let’s think about the nature of excess insurance and large losses. Think of excess insurance like an option – you pay a small annual fee for the option to transfer a significant loss. Law of large numbers is on the insurers’ side – not yours. The incredible variation in outcomes favors the spread of risk.
Consider the following example: an insured has the ability to pay a minimum premium of $7,500 per million of coverage (this is an example price, not representative of any specific insurer) for $50M of excess coverage above a $100M attachment point. This results in a not small premium of $375,000, but it also avails the insured to $50M of protection. Ignoring the time value of money, it would take the insured 133 years of saving that premium to fully fund a full $50M loss. The insurance company on the other hand, can spread that loss among all of its insureds – 133 of them to fully fund a $50M loss.
We don’t know what the potential losses are – it may be a rogue employee like at Johns Hopkins, a natural disaster that results in cascading errors, or anything else. It doesn’t matter. What matters is we understand the option and the nature of the decision.
Pricing Excess Insurance is Difficult
Insurers love to model risk. But models work best with a closed data set – and models fall apart at the extremes. Small changes in assumptions have outsized impacts on the “tail “of the distribution. Pricing excess insurance – for you and for the insurers is exceedingly difficult. There is just very little data to inform the models. Some also make the error of using the same model for working layer analysis and excess layer analysis. The difference in the nature of losses in these layers could not be more different.
So, if it is difficult to price the excess layers, is it more likely that it is too expensive or too cheap? Well, does it matter? If we can pay $7,500 per million but over a simulation it turns out that we really should only be paying $5,000 per million…we’re overpaying the $50M layer by $125,000. While it’s not desirable to overpay, the organization can take a $125,000 hit.
The price per million of insurance coverage is somewhere between $0 and $1M. Perhaps with profit and overhead of the insurer, you can go over $1M but at that level we’re not talking about insurance anymore. Excess pricing of $7,500 per million is much closer to $0 than $1M, and I think it’s much more likely that $7,500 is too thin than too fat. Again because of the limits – there is a lot of room for upwards correction but not much for downwards.
Insurance company’s ability to move down depends on getting increased spread of risk (number of insureds). This is intuitive – an insurer needs adequate spread of risk to fund the claims of the few. New capital entering the market (increasing supply) actually has the effect of decreasing spread of risk for individual insurers. This is because as the market becomes saturated, each insurer gets fewer insureds and needs more premium per million to appropriately price the risk. But at the same time competition is up, driving prices down. To get adequate spread of risk, insureds need to price competitively to keep or gain market share. Consider also that by the high excess, rare (but not that rare) nature of these claims means the market could take years to show the insurers they’ve been under pricing. Said more simply: competition breeds price competition and then the market doesn’t correct for a decade or more. I think this is where we are in the market cycle.
Buy Higher Limits?
All of this – limited data and over-reliance on modelling, asymmetrical upside, and a hyper competitive property and casualty market lead me to believe that $7,500 per million (or $5,000 or $10,000) is likely a very good deal for insureds. Combine that with the understanding that future losses will be larger than past and we don’t know when or how the next Black Swan will hit – buy the higher limits and then buy some more. At some point enough is enough but that’s a topic for another day.
By: John Littig
John currently serves as the Chief Finance and Underwriting Officer of The Risk Authority – Stanford, as well as the Vice President of Risk Finance of Stanford University Medical Center. John has more than 11 years of experience as a health care risk finance professional with experience creating and operating self-insurance trusts, reciprocal risk retention groups, direct issue captives, as well as domestic and offshore captives. John is responsible for procuring and maintaining all property and casualty coverages for the Stanford University Medical Center.