Most risk models are increasingly diverging from the real loss experience of insurers by underestimating the frequency and severity of events. This means cedants are increasingly seeking alternative perspectives, Matt Olsen (pictured left), senior vice president and co-head of catastrophe modelling at Holborn, told APCIA Today.
He explains that Holborn has been mapping the output from various risk models against the actual loss experience of clients for years. It was initially clear that the frequency was often inaccurate. More recently, however, it is clear the severity is too.
“We’ve always known there were issues. Initially, we noticed the frequency of events was woefully understated. Some models would cover maybe only a third of the expected frequency. In the past three years, however, we’re seeing the models also miss on the severity side,” Olsen said.
He said the issue is especially pronounced for so-called secondary perils such as hail, flood, severe convective storms, and wildfires. “Data in king in this game. The risk modellers have branched out but, for certain perils, they just don’t have access to the data they need,” he said.
This means insurers are increasingly seeking other ways to get a handle on the risk. Holborn has developed several tools designed to helps its clients better understand these risks. They are not risk models, just a scientific way of applying different scenarios to a portfolio.
“We’re focused on a client-specific view of risk, not a vendor model.” Matt Olsen, Holborn
“Our aim is to assist our clients gain a more accurate assessment of risk and loss potential,” Olsen said. “We’re focused on a client-specific view of risk, not a vendor model or an industry view of risk. We try to give our clients as many data points as possible to make a well-informed decision and reduce the uncertainty. You are never going eliminate it completely, but we can manage it. The goal is to always learn, always get smarter.”
One insight Holborn offers is devised through the analysis of industry-wide claims data. Olsen said average claims in many sectors are increasing by close to 10 percent a year. Yet most risk models, and some insurers, are factoring in inflation at a much lower figure much lower. “So, claims are woefully understated,” he said.
One specific challenge for insurers operating in the Midwest has been an increase in the number and severity of derechos in recent years. These are long-lasting severe windstorms that can travel great distances—usually in a straight line. They have been responsible for severe and unexpected losses for some insurers.
Olsen says these are very difficult to model but Holborn has helped clients assess the potential risk by pinpointing potential large loss scenarios within client exposures. Again, he stresses, this work is client-centric—not deterministic.
“That is what sets us apart; we are always seeking a client-centric view of frequency and severity. We use vendor models, but we also look at actionable data that we can base decisions on. That requires education for clients, but that is where our analytics team comes in. We need to explain why the need additional perspectives to truly understand their risks,” he explained.
This is important in a market where property-cat reinsurance capacity is becoming scarce, and rates are increasing. Yet as insurers grapple with their risk profile, many require more reinsurance.
Olsen adds that cat activity is going through an elevated frequency. “But this has happened before. You must learn from the data and history. Equally, in a more benign weather period, when the weather’s good, we ensure clients manage their portfolios correctly. It is in these elevated periods that the warts show themselves.”
“I have never seen an insurance company go under because they purchased too much reinsurance.” Frank Harrison, Holborn
Frank Harrison (pictured right), chief executive officer of Holborn, says that some clients are on a steep learning curve as they reassess their exposures. He suggests close attention should be paid to scenarios previously considered infrequent or unlikely to happen—and their likelihood reassessed.
“Cat is the newest four-letter word and aggregate reinsurance policies are not far behind. It’s hard to understand those risks but insurers would be served to consider those worst-case scenarios,” he said.
“1-in-1,000 events seem to be happening a lot. In my lifetime, I have never seen an insurance company go under because they purchased too much reinsurance, but I can think of plenty that went down because they purchased too little.
“A lot more time and devotion are needed to understand the specific risk profile of each client. They’re all different. Some will pay more because they need a lot more. But that change is required now.”
In such a scenario, Harrison stresses, long-term relationships between cedants and reinsurers, facilitated by their broker, are key. Those that have invested in long-term relationships will get better deals. Those that perhaps shopped around for cheaper reinsurance previously, will not.
“It is a question of treating others how you wish to be treated,” Harrison said. “Speaking candidly, not every partner is in it for the long term. The trick is to identify those who are and ensure that everyone is on the same page. Those who have behaved opportunistically in the past may struggle in a market such as this.”
APCIA 2022, Holborn, Risk Modelling, Catastrophe, P&C, Insurance, Reinsurance, Matt Olsen, Frank Harrison, North America