Specialty property’s permanent shift reshapes programs and E&S

Specialty property insurance has moved from the margins to the mainstream. What was once viewed as a channel for distressed risks has become a structural pillar of the property and casualty market.

“The starting point of specialty,” said Ari Chester (pictured), head of specialty at Argo Group, “is either channel, meaning E&S or MGA program channel, or it’s a product niche in the market.”

In his framework, specialty is defined either by distribution or by technical depth. Sometimes it requires expertise. Sometimes it reflects a shortage of admitted capacity. Often, it is both.

“AI and data center is a good example. It’s both a capacity need, and, to do it, you need more expertise. Same with renewables,” he said. “It’s not just a dearth of capacity from markets, and it’s also real expertise that is needed on the ground to do it and to do it well.”

That dual identity – channel and specialization – has reshaped how brokers and carriers approach property risk.

Over the past five years, non-admitted and program business underwent what Chester described as a “seismic shift.”

“E&S five years ago, as a portion of all property P&C in the US, has gone from7-12%,” he said. In commercial lines, he placed that figure closer to 20%. In certain statutory lines such as fire, quake and private flood, “it’s actually now 40% of all of the market is E&S.”

While those figures were shared as part of Argo’s internal analysis and not independently verified in this interview, they illustrate a broader shift. E&S has moved from a peripheral outlet to a permanent placement strategy.

“It went from being more of an alternative or peripheral or distressed part of the market to more of a permanent staple for placement in the market,” he said.

At the same time, the MGA and program segment doubled from roughly $60 billion to $120 billion in premium, according to Chester. Growth was fueled by hybrid fronting innovation, greater risk syndication, improved data and alignment, and capacity gaps during the hard market.

With scale came discipline. “In today’s world, if you’re a poor-performing MGA, you will likely get a bad reputation and lose capacity pretty quickly,” he said. “That’s a little bit different than what I saw even 10 years ago.”

Still, he cautioned against assuming the same growth trajectory ahead. “If I had to summarize the macro outlook, it’s permanent, but it’s also plateauing,” he said. “There’s no room for continued explosive growth, three to five years out… it’s hitting a ceiling.”

As programs mature, execution risk becomes more visible. Chester grouped the challenges into three areas: balance, execution and volatility measurement.

“Balance means, to be viable in the market, you need enough scale… both spread of risk and the breadth and the depth to absorb a few losses,” he said. Premium-to-limit and premium-to-PML ratios matter. “You can’t be small and dabble; you have to be big enough to have spread.”

Execution often proves harder than business plans suggest. “Insurance is easy to build, it’s easy to grow; you just have to have a better price. But that’s a very dangerous way to grow,” he said. “Whatever the target subject premium is two to three years out, I just assume it will be half of that or a third of it.”

Volatility remains the third pressure point. In property underwriting, reliance on modeled PML or MFL can create a false sense of security. “If you think you’re writing to a 20% PML and it turns out to be a 40%–50% loss… that’s a problem,” he said.

Incumbents with scale and historical data hold an advantage. “Once you do reach scale and balance, and you have a decade of experience under your belt, it helps with the measurement of volatility,” he said.

Climate risk has further reshaped specialty property. “A secondary peril is really no longer secondary,” Chester said, citing wildfire, flood and convective storm.

Program resilience, he argued, rests on awareness, measurement and product alignment.

“Just because something is not really in the model now, it doesn’t mean it’s not a risk,” he said, referencing winter storm and civil unrest as examples of under-modeled exposures.

Measurement has shifted from broad geographic assumptions to location-level underwriting. The industry moved “from more of a zip-code mentality… to location-level underwriting,” he said.

Wildfire assessment now incorporates vegetation, topography and environmental data, while portfolio accumulation management is increasingly real-time.

Product alignment is equally critical. Coverage structures are recalibrated through sub-limits, percentage deductibles and, in some cases, shifting from replacement cost to ACV is just one of many examples.

“It’s not really narrow,” he said. “It’s right-sized, and because it’s right-sized, it’s more sustainable and more fair… the capacity will be more sticky and more permanent.”

Technology has evolved from industry buzzword to underwriting tool.

“It’s gone… from a lot of hype and hyperbole to a more tangible and tactical application of AI and LLMs,” Chester said.

He identified three areas of impact. First is knowledge management, where submission material could be ingested and summarized “in minutes or seconds, instead of hours.”

Second is AI-enabled inspection using drones, satellite imagery and smartphone-based tools. “It’s a very powerful, tactical application of AI that really wasn’t at all widespread… five to seven years ago,” he said.

Third is what he described as “deep, specific, spontaneous insight” from large language models. In one example, a warehouse in an X500 flood zone appeared manageable based on elevation and structure. A deeper query revealed a recent history of 25-30ft flooding at that address.

“It’s not perfect,” he said. “But it’s a very, very powerful tool… mostly not adopted yet in a widespread way.”