Which Door Into Insurance? A Lawyer’s Map of the 50-State System

Michael Byrne has spent almost 25 years as a lawyer working only in insurance. He does transactional and regulatory work rather than litigation or coverage disputes, and his clients include insurers, brokers, investors and service companies across the ecosystem. The job, as he describes it, is keeping clients out of trouble, minimizing disruption and enforcement risk if they do get into it, and helping them buy, sell and partner.

Gilad first met Michael through an introduction from Hippo’s Assaf, when Gilad’s own embedded insurance startup was looking for a law firm that knew the space. They met face to face at an OnRamp event at Soldier Field, and Michael also serves as a mentor and sponsor to the Israeli insurtech accelerator, partly to help founders understand the US market.

About Michael Byrne

Michael Byrne is a partner at McDermott Will & Emery and co-leader of its InsurTech practice, advising insurers, MGAs, brokers and startups on insurance regulation and transactions.

Why Insurance Is Regulated Like a Public Utility

Michael started with the basics. Insurance is a 50-state system with 50 chief regulators and career staff who stay when commissioners leave. The commissioners themselves are roughly half elected and half appointed, so there is always a political undertone. The regulators’ job is to protect consumers and company solvency, and that is why retailers and other companies that look similar aren’t treated the same: insurance is regulated much like power or telephone companies. A license is needed to sell insurance, to receive commissions and to take on risk.

His read on innovation is positive. Regulators ask whether something helps or hurts the consumer, and they have largely facilitated insurtech or at least not been an obstacle. The 50-state structure also lets states compete: unlike a single national regulator, if one state is hostile there are others, and many states want the jobs and commerce that startups bring. Captives are the clearest example, with Vermont and Utah competing to attract them by writing laws that let people do business the way they want. That doesn’t mean they are easy regulators, only that their laws don’t get in the way.

There are no real sandboxes like the UK’s, but most states are approachable. The questions that matter are what your business model is, what you can offer the state, and where you need to be.

Should You Talk to Regulators First?

Michael’s answer depends on what you are launching. If you need to launch an insurance company, you will have to talk to the domicile regulator eventually, so do it early. That choice is important, because the domicile regulator is the one you will live with for your existence: it regulates transactions within your corporate family and how much capital you must hold. If you are an agent or an MGA that just needs a license and isn’t doing anything unusual beyond doing it digitally, you generally don’t need to raise your hand, and sometimes you don’t want to.

Agency, MGA or Carrier? The Cost and Timeline Ladder

The central question for any founder is how to enter the market, and Michael’s framework is mostly about money and time:

• Agency: a few thousand dollars to start, two to four months to get licensed (a 40-hour course and a few weeks of back and forth with the state). Turning it into a real digital agency is in the hundreds of thousands.
• MGA: faster to market than a carrier because the barrier is a license, though he noted that being an MGA in practice just means doing some underwriting; many founders don’t realize an MGA is regulated as an agent unless it is licensed as one.
• Carrier: tens of millions of dollars at a minimum, since you need capital and surplus, a rating, and people to run governance, compliance and regulatory reporting. Licensing a carrier takes at least six months and often a year or more. Roughly half of states impose seasoning requirements of one to five years of transacting business before they will license you.

Because of that, starting as a carrier at scale is, in his words, almost impossible, and the exceptions prove the rule: Lemonade started with about $500 million and large reinsurers behind it, while Hippo started as an MGA and Metromile started as an agency and later bought a carrier. A shell company is the usual way around seasoning, which keeps shell prices high; finding, diligencing and getting domicile approval takes about six months. He also noted that you can start as a carrier in a single, friendly state with a limited geography, and gave an example of a health insurance startup that did.

Where to start also depends on the model. If you are proving a concept, you don’t need California or New York, since Utah and Idaho have drivers and people who need health insurance too. But many founders want all 50 states right away, because that is where the scale is and what the next investors will look for.

Captives and Skin in the Game

Profit sharing with a carrier doesn’t require a risk entity. If you want to share in the risk, you need an insurer, reinsurer or captive. A captive typically acts as a reinsurer behind an MGA and a fronting carrier, and a rule of thumb is about three dollars of premium for every dollar of surplus. In practice, a fronting carrier reinsures nearly everything to a large reinsurer, and the captive might take, say, five percent.

Michael said reinsurers and insurers are increasingly asking MGAs whether they plan to take any risk, because they want to see that the MGA has skin in the game, and bad business will hurt them too. Gilad noted how quickly reinsurers have adapted, given that they enabled the insurtech wave in the first place.

What Comes Next: The Black Box

Michael doesn’t expect the landscape to look very different in two years, but he expects the focus on big data and algorithms to grow. The concern is the black box: rating and underwriting models that regulators see only the output of, and which they worry might discriminate on race in an unlawful way. With attention on race across the country, the NAIC and states are looking at disparate impact, and an NAIC group is examining accelerated underwriting, so life insurers will soon be asked to report whether they use it.

Regulators are underfunded, underpaid and overworked, and a smart engineer is more likely to go work for Amazon than for a state insurance department. Michael thinks that more fees in exchange for faster filings and better service could help. He also predicts that in some states the presumption will be that a black box discriminates and the carrier will have to prove it doesn’t. New York’s cybersecurity regulation shows how states push each other.

His advice is to assume your technology business will be regulated at some point, and that someone will eventually look at your model. There is also a trust gap: companies are afraid to bring an algorithm to a regulator because opening up may invite endless questions, and regulators assume the industry doesn’t want to share. Michael has worked with an ad hoc industry group to engage the NAIC’s Innovation Task Force, and he worries about an adversarial stance that ends with heavy-handed rules that stop everyone operating legally.

Embedded Insurance, Defined

Gilad closed with embedded insurance, since his own platform was one of the first and had to shut down the previous year. Michael said the whole conversation depends on how you define it. Some people mean insurance literally embedded in a retail product, mandatory or free, and dismiss it. His view is broader: being continuously connected to a customer. A homeowner needs insurance for the mortgage, but the home is so much more: a repair when a window breaks, a water sensor that detects a leak. That is an opportunity to earn trust and build a relationship over decades rather than just cross-selling.

It is, he said, why people talk about a three trillion dollar opportunity, but it takes execution and good partners, and probably a few failures. He pointed out that Lemonade has done well at selling a different vision of insurance, as something that helps people mitigate risk rather than an adversarial relationship.

Advice: Watch Something, and Turn Off Notifications

For his recommendation, Michael named a few shows: Sons of Anarchy (a big commitment), The Serpent, and The Handmaid’s Tale. He also recommended the documentary The Social Dilemma, after which he turned off notifications on nearly all his apps, including LinkedIn.

Key Takeaways

• Insurance is regulated like a public utility in a 50-state system, and the competition between states gives startups choices of where to launch.
• Talk to regulators early if you are launching a carrier, and consider not raising your hand if you are a plain agency or MGA.
• An agency costs thousands and takes two to four months; an MGA is faster than a carrier; a carrier needs tens of millions and at least six months, plus seasoning in many states.
• Starting as a carrier at scale is nearly impossible: Lemonade is the exception, and most successful startups began as MGAs or agencies.
• A shell company can shortcut seasoning in about six months, which keeps shell prices high.
• Reinsurers increasingly want MGAs to retain some risk, often through a captive, as proof of skin in the game.
• Expect more scrutiny of black-box models, accelerated underwriting and disparate impact, and assume your technology business will be regulated.
• Michael defines embedded insurance as staying continuously connected to the customer, not as a free add-on.