Tag Archives: lamparelli

Why the Agent Will NOT Be Disrupted

“Google Compare kaput” – Shefi Ben Hutta

A few weeks ago, I published an article here on ITL saying that the insurance industry, in general, would not be “disrupted.” I received both a lot of positive and (politely) negative feedback, including a rebuttal by Nigel Walsh. And then just this week, Google, the single-most-often-pointed-to culprit for the probable insurance disruption, dropped a bombshell: that it is shutting its Google Compare insurance service.

That whisking sound you hear is me taking my victory lap.

All kidding aside, although the urge to take a victory lap is strong, my calmer, rational side realizes that this news does not mean what some might think it means. While my beliefs are that disruption, as has occurred in other industries, will not happen in insurance, Google’s exit from this space is NOT evidence that I am correct. What I believe has transpired is the following:

  1. The insurance business overall is complex. Software cannot eat this elephant whole.
  2. Google underestimated how difficult the business is, especially in the segment Google Compare was fighting for, which is distribution. Getting new customers in insurance is quite challenging. Customers want value in their insurance transactions, which a website and a rater cannot imitate.
  3. Google’s opportunity cost of capital is high, and Google Compare couldn’t meet an acceptable threshold because of its inability to get traction. Brian Sullivan of Risk Information recently said that Google Compare got 10% of the business it forecasted. Ouch!

Those on the disruption side of things promise that, much like the Terminator, Google will “be back.” I actually think that is possible, after some of the issues are ironed out, such as expectations. Once upon a time, I would have been an eager Google Compare customer. So I have no doubt that there is a market for its offering.

But there is a bigger market for insurance customers who want someone else to do the un-thrilling work of getting their insurance in place because those customers either don’t have the expertise or don’t wish to be bothered by the process at all. Consider a recent example in my own timeline.

My current auto and property policies were purchased online several years ago. I didn’t need an agent because I was more than happy to do the work myself to save a few dollars. No longer.

I recently moved across the country, back to the East Coast. The last thing I wanted to do was deal with address and other changes that are required when you move across state lines. I also didn’t want to research all of the licensing and car registration procedures I’d have to go through in the weeks following my move. So I found an agent. Within a couple of days, that burden had been lifted from me. I am less likely to personally do the insurance buying going forward. I would rather be doing something else altogether than researching and buying insurance. The whole experience was well worth the commission paid.

And then there are customers who don’t know much about insurance at all: teen drivers, new homeowners and new parents, to name just a few potential insurance customers where the guidance of a trusted adviser will save a lot of time and future headaches. Can we really expect teen drivers to understand anything more than getting the cheapest policy possible so that they can drive? My newly minted teen driver spent days trying to get her car on the road because she chose the Cockney-accented spokes-lizard insurance, which provided nearly zero support for her real problem, which was the DMV. My response: ”You should have gone to an agent. He would have done all that work for you.” A lesson learned, I hope.

How about a new homeowner trying to get insurance to cover the property and family? An insurance agent will help with issues around replacement values, limits of liability, deductible options and coverage differences between carriers. Can machine learning get to the point where it can replace all of that? Perhaps. But add to this, additional complexities such as how should a family put together auto, property, umbrella and other insurance policies (such as flood, earthquake, jewelry, non-admitted products) together to optimize effectiveness, and I think the technologists looking to disrupt are a long, long way away from being able to effectively deliver the value that an agent/broker is already providing. As the stakes are raised, the human touch will remain invaluable.

This is not to say that the state of the current agency system is acceptable. Agents need to step up their game. Agents have been one of the biggest offenders in not using technology to further their significance. Agents have chiefly been great sales people. They have to be. They are selling an imperfect product whose value is difficult to quantify. In today’s environment, agents need to scale their sales presence outside of the face-to-face transaction toward a digital world. The agent might be able to overcome my objections when we are looking at each other, but, today, I am communicating via digital means, and I can simply ignore the agent. Agents need to use technology to better market to, communicate with an educate customers. They also need to take a page from insurers and use data to understand and quantify risk so that they can recommend the best solutions and not just a policy with the lowest price. Agents are used to providing multiple options to customers; now they need to use data to get an information advantage. Does this mean that agents need to become part underwriter, part adjuster, part actuary while remaining part salesperson to survive? I think so.

For the modern agent or broker, Google Compare was not seen as a serious threat. Top agents know the value they bring and are not easily substituted for with technology. Twenty years from now, the landscape for buying insurance will look very different from today but I wager that, for many of the reasons I have outlined here, the insurance agent will still have a significant role for consumers who value their time and possessions.

No, Insurance Will Not Be Disrupted

I recently had the pleasure of attending the Insurance Disrupted conference in Palo Alto (put on by the Silicon Valley Innovation Center in partnership with Insurance Thought Leadership). This was the single best insurance conference I have ever attended. I was surrounded by hundreds of hopeful, smart, problem-solving professionals from disparate backgrounds and industries all trying to make a difference in insurance without money being the prime motivator.

I was so encouraged by what transpired at the conference, the connections that I made and what I believe would be the promise of a new future that I began to pen this article on my flight home. But something just did not sit right with me as I wrote. Three weeks have gone by, and I am beginning to understand why I felt the way I did; at the end of the day, insurance will NOT be disrupted.

For all the promise of big data, the Internet of Things, autonomous vehicles and peer-to-peer insurance, there was nothing presented at this conference that struck me as disruptive in the way the tech industry is generally thinking of the term today. When technologists think of disruption, they immediately point to Uber and Airbnb, which disrupted the taxi/livery and travel accommodations industries. The taxi industry is literally fighting for its survival. No, that will not be the fate of insurance. Insurance will be a lot more difficult to shake up or disrupt.

Here’s why:

  1. At the core, insurance customers are leasing the potential to access capital. That capital is sitting in predominantly liquid assets. Not real estate, not taxi medallions. How do you make a big pile of money irrelevant?
  2. The modern form of the industry is 300 years old, and the math is pretty solid (that’s why they call it actuarial science). We sell a product whose costs are unknown at the time of purchase. That means scale and immense capital is required to cover worst-case scenarios, which rules out any new business model not having that potential. Peer-to-peer providers just won’t be able to get sufficient scale to efficiently use capital to cover risk. And if they aggressively get scale, then they just become another insurance company, so what’s the point?
  3. Getting a better glimpse into those unknown expenses can create massive competitive advantages. This is where big data and the IoT creators are looking to disrupt, as big data and IoT will generate incredibly large data sets to be used to accurately predict, avoid and mitigate future losses. I have no doubt that these new technologies will make an impact on the industry, but I am less convinced of their disruptive nature. Insurers have already established non-actuarial, big data departments where fraud detections and credit scoring are just a couple of many predictive models being created. IoT devices will slowly be adopted by most insurers as they look to get competitive edges, but the follow-the-leader paradigm of the industry will mean that any edge will disappear quickly, and we will all be running hard just to stay in place. These technologies are impressive. I would classify them as a solid innovations to the industry, but not disruptive. (Disclaimer: I bought a smart battery from Roost.)
  4. Autonomous vehicles represent the one area where some chaos can occur. But notice I use the word “chaos” and not “disruption.” If autonomous vehicles can live up to expectations, then they will be a great service to society, reducing deaths and increasing efficiency. Risk will transfer from a personal lines business to commercial lines, and that could be chaotic for heavy personal lines auto writers such as State Farm and Progressive. But will this be disruptive? Will State Farm or Progressive be fighting for their survival the way that medallion owners in the New York City taxi system are? Again, I doubt it. State Farm is sitting on about $70 billion in surplus capital, and it generally writes at a 100 combined ratio, working the float and cash flow model. I think State Farm and large auto insurers like them will be just fine, and technologies such as autonomous vehicles will be more of an annoyance than an existential threat. And like others, I don’t think autonomous cars are nearly as ready to take over our roads as many seem to think.
  5. For better or worse, state-by-state regulation of insurance is intense and nebulous. Ask Zenefits. The battlefield is already uncertain, and scrutiny by a regulator with political ambitions can kill your disruptive product quickly. Any technology that you think you can create that could potentially benefit the majority of buyers while subsequently raising the price for some other group, alone, would be grounds for a regulator to squash you, as that vocal minority raises their collective voices. In Florida, the state may even create a company to compete against you, writing business at a loss. Insurance regulation might be the ultimate disruption killer.
  6. There was not one presentation on natural catastrophes, which happen to be my area of expertise. How we underwrite, manage and think about natural catastrophe risk has changed quite a bit over the past 20 years. In fact, CAT models have been and may continue to be the most disruptive force in insurance, and yet there is little technology can do to disrupt that area of the industry. I would have been very excited if we had discussions about new business models to help customers with the problems the industry is currently facing with getting adequate flood or earthquake cover to homeowners. If someone had proposed a new product that removed the exclusions of flood and earthquake from the homeowners policy, now, THAT would be disruptive! Alas, nothing on NatCat, and so we will continue to have thousands of homeless families following big storms and earthquakes.

I don’t think insurance will be disrupted, not in the way folks from Silicon Valley are used to doing it. But the future of insurance will look very different than today. Very digital. Streamlined. Less clunky, more efficient. If “disruption” comes to insurance, it is likely going to require the replacement of the current set of leaders with new ones cultured in this digital age and influenced by the successes of technology to make change happen to their business models.

Paul Vandermarck from RMS (a CAT modeling vendor) perhaps summed it up best when he said that no matter how all of this change to the industry plays out, we know of one sure winner: the customer. And that’s how it should be.

It’s Time to Rethink Flood Coverage

“The boat is safer anchored at the port; but that’s not the aim of boats.” — Paulo Coelho

The scenes are now all too familiar. Waters rising, dams breached, cars drifting away, homes and properties inundated with water. As of this writing, 13 people have died in the Carolinas as the “one in a 1,000 years” flood continues to ravage the area. Losses should easily exceed $1 billion.

If all of that was not bad enough, what’s worse is that you and I will be paying for this.

Unfortunately, the song remains the same after all these years:

  1. Property insurance policies exclude flood coverage
  2. Property owners either believe they have coverage or choose not to purchase it
  3. The biblical rains arrive, causing damage, and property owners seek help from the largest wallet available and willing to help…the U.S. government
  4. (Alternatively, and unfortunately, property owners may buy flood coverage, but, because the coverage was mispriced, the National Flood Program will not have the funds to pay the claims and will need to borrow from us taxpayers).

The system is a mess, and my criticism lies directly with the insurance industry. We can solve this problem. These floods are insurable events. We are flush with capital, and each week it seems another technology firm is releasing a flood model to help us manage this risk.

But that sound you hear is crickets. We are not making much progress at all.

The solution cannot be separate, private, flood coverage. That is a nice start but is not the solution, because it’s more of the same, just with a different wallet writing the check.

What we need is to “loosen the exclusion.” Flood needs to become a standard component in the homeowners policy. Just as fire, wind, lightning, theft, vandalism and liability are all standard components of a homeowners insurance package, flood needs to be included as that form of standard coverage.

The advantage to homeowners is true peace of mind.

  • Every homeowner has some ground water risk, and we can eliminate this coverage concern once and for all.
  • We can eliminate policy juggling, with one single policy.
  • A single claims adjuster can determine any losses without needing superhuman insights to know whether water or wind caused the damage.

The enterprising insurer gets to differentiate its personal lines business with a non-correlated premium source. The insurer eliminates the headache of defending flood exclusions and the bad publicity and court judgments around those issues.

Some insurers will be rightly concerned about the increased risks. But isn’t this the business we are in? It may feel safe to exclude coverage, but our role in society is not to exclude coverage. Our role is to find a way to profitably make our capital available for these type of events.

We have all the tools and capital we need to make this happen. Do we have the will?

Why Flood Is the New Fire (Insurance)

With our past few posts on ITL, we have been exploring how insurers can continue to bring more private capacity to U.S. flood (Note: Everything we talk about for U.S. flood is also relevant for Canada flood). We have explored here how technology, data and analytics exist to handle flood in an adequately sophisticated manner, and we have described here the market opportunity that exists. Now, it’s worth a look to explore how a flood program could be introduced, starting from scratch through cherry-picking mischaracterized risks and then to a full, mass-market solution.

What’s a FIRM? It’s not what you think

First, let’s take a quick look at how National Flood Insurance Program (NFIP) rates are determined: the Flood Insurance Rate Maps, or FIRMs. For the NFIP, FIRMs solve two core problems – identifying which properties must have flood insurance and how much to charge for it. The first function is for banks, giving them an easy answer for whether a property to be lent against requires flood insurance – this is what the Special Flood Hazard Area (SFHA) is for. Anything within the SFHA is deemed to be in a 100-year flood zone (basically, A and V zones), and requires flood insurance for a mortgage. The second function sets the pricing and conditions for the NFIP to sell the actual policies. The complexity of solving these two problems should not be underestimated for a country of this size. But it must be remembered that a FIRM is a marketing device and not a risk model.

Considering that FIRMs are a marketing device built on a huge scale, it makes perfect sense that some generalizations needed to be made on the delineation of the various flood zones. The banks needed a general guideline to know when flood insurance was needed, and the NFIP needed rates to be distributed in a way that could result in a broad enough risk pool to generate enough premium to be solvent. While the SFHA has served the banks well enough over the years, the rating of properties has not been so successful. There are plenty of reasons the NFIP is deep in debt (see page 6 of this report); suffice it to say that the rates set by FIRMs do not result in a solvent NFIP.


The fact that the FIRMs are a flawed rating device based on geographical generalizations means there are cherries to be picked. By applying location-based flood risk analytics to properties in the SFHA, a carrier can begin to find where the NFIP has overrated the risk. Using risk assessments based on geospatial analysis (such as measurements to water) and their own data (such as NFIP claims history), a carrier can undercut the NFIP on specific properties where the risk fits their own appetite. Note to cherry-pickers: Ensure you account for the height above ground of the building, because you won’t need elevation certificates for this type of underwriting. So far, cherry-picking has been focused on the SFHA for a couple reasons – homeowners need to have coverage, and the NFIP rates are the highest. There is no reason, though, that cherry picking can’t be done effectively in X zones and beyond.

Mass-market solution

The same data and analytics used for cherry-picking can be used more broadly to create a mass-market solution. By adjusting the dials on the flood risk analytics – and flood risk analytics really should be configurable – you can calibrate to calculate the flood risk at low-risk locations. In other words, flood risk can be parsed into however many bins are needed to underwrite flood risk on any property in the country. With the risk segmented, rates can be defined that can (and should) be applied as a standard peril on all homeowner policies. Flood risk can be underwritten like fire risk.

Insurers have traditionally been confident underwriting fire risk. But consider this: While fire is based on construction type, distance to fire hydrants and distance to fire station, flood risk can be assessed with parameters that can be measured with similar confidence but with greater correlation to a potential loss.

Flood will be the new fire

Insurers have been satisfied to leave flood risk to the Feds, and that was prudent for generations. But technology has evolved, and enterprising carriers can now craft an underwriting strategy to put flood risk on their books. Fire was once considered too high-risk to underwrite consistently, but as confidence grew on how to manage the risk it became a staple product of property insurers. Now, insurers are dipping their toes into flood risk. As others follow, confidence will grow, and flood will become the new fire.

Home Insurers Ignore Opportunity in Flood

Recently, Munich Re announced its plan to step into the U.S. inland flood market to offer a competitive flood coverage endorsement for participating carriers. This is the second notable entry of international capital into an arena dominated by the federal government.

Munich Re is known as a conservative giant of international reinsurance, so it might seem odd that it is joining the National Flood Insurance Program (NFIP) in covering U.S. flood. A quick look at the opportunity shows why the plan makes sense.

U.S. inland flood insurance is an untapped source of non-correlated premium unlike any other in the world. The market is dominated by an incumbent market maker that is in trouble because it offers an inferior product that cannot price risk correctly (this paper nicely summarizes the problems at NFIP). So, here is what the new entrants are seeing:

  1. Contrary to industry beliefs, flood is insurable. The tools are present to accurately segment risk.
  2. Carriers offering flood capacity will differentiate themselves from competitors. This will give them a leg up on the competition in a market that is highly homogeneous. Carriers not offering flood will likely disappear.
  3. The market is massive, with potentially 130 million homes and tens of billions of dollars at stake.

Let’s go into details.

Capital Into a Ripe Market

The U.S. Flood Market

As most readers of Insurance Thought Leadership already know, many carriers have flood on the drawing board right now. The Munich Re announcement was not really a surprise. We all know there will be more announcements coming soon.

Let’s summarize the market reasons for the groundswell of private insurance in U.S. flood.

The most obvious characteristic of the market is the size. For the sake of this post, we’ll just consider homes and homeowner policies. Whether one considers the number of NFIP policies in force as the market size (about 5.4 million policies in 2014), the number of insurable buildings (133 million homes) or something in between, there is clearly a big market. And the NFIP presents itself as the ideal competitor – big, with a mandate not necessarily compatible with business results.

So, there is no doubt that a market exists. Can it be served? Yes, because the risk can be rated and segmented.

Low-Risk Flood Hazard

To be clear: A low-risk-flood property has a profile with losses estimated to be low-frequency and low-severity. In other words: Expected flood events would rarely happen, and not cause much damage if they do. For many readers, joining the words “low-risk” and “flood” together is an oxymoron. We strongly disagree. Common sense and technology can both illustrate how flood risk can be segmented efficiently and effectively into risk categories that include “low.”

Let’s start with common sense. Flood loss occurs because of three possible types of flood: coastal surge, fluvial/river or rain-induced/pluvial (here is more information on the three types of flood). The vast majority of U.S. homeowners are not close enough to coastal or river flooding to have a loss exposure (here is a blog post that explores the distribution of NFIP policies). Thus, the majority of American homeowners are only exposed to excess surface water getting into the home. We’d be willing to wager that most of the ITL readership does not purchase flood insurance, simply because they don’t need it. That is the common-sense way of thinking of low-risk flood exposure.

How does the technology handle this?

There is software available now that can be used to identify low-risk flood locations (as defined by each carrier), supported by the necessary geospatial data and analytics. Historically, this was not the case, but advances in remote sensing and computing capacity (as we explored here) make it entirely reasonable now, with location-based flood risk assessment the norm in several European countries. Distance to water, elevations, localized topographical analyses and flood models can all be used to assess flood risk with a high degree of confidence. In fact, claims are now best used as a handy ingredient in a flood score rather than as a prime indicator of flood risk.

How to Deliver Flood Insurance in the U.S.

Deliver Flood Insurance to What Kind of Market?

Readers must be wondering at the size of market, because we offered two distinctly different possibilities above – is it about 5 million to 10 million possible policies, or 130 million policies? The difference is huge – the difference is between a niche market and a mass market.

The approach taken by flood insurers thus far is for a niche market. The current approach probably has long-term viability in high-risk flood, and the early movers that are now underwriting there are establishing solid market shares, cherry-picking from the NFIP portfolio.

On a large scale, though, the insurance industry’s approach needs to be for a mass market.

Here is a case study describing the mass market opportunity:

  1. The property is in Orange County, CA, where the climate is temperate and dry, almost borderline desert. El Niño might be coming, but that risk can be built in.
  2. Using InsitePro (see image below), you can see that the property is miles and miles away from any coastal areas, rivers or streams. More importantly, the home is elevated against its surroundings, so water flows away from the property, which is deemed low-risk.
  3. The area has no history of flooding, and this particular community has one of the most modern drainage systems in the state.


Screenshot of InsitePro, courtesy of Intermap Technologies. FEMA zones in red and orange

  1. Using Google Maps street view, we can estimate that the property is two to three feet above street level, which adds another layer of safety. Also, this view confirms that the area is essentially flat, so the property is not at the bottom of a bathtub.
  2. And, as with most homes in California, this property has no basement, so if water were to get into the house it would need to keep rising to cause further damage.

To an underwriter, it should be clear that this home has minimal risk from flooding. As a sanity check, she could compare losses from flood for this property (and properties like it in the community) to other hazards such as fire, earthquake, wind, lightning, theft, vandalism or internal water damage. How do they compare? What are the patterns?

For this specific home, the NFIP premium for flood coverage is $430, which provides $250,000 in building limit and $100,000 in contents protection. The price includes the $25 NFIP surcharge.

This is a mind-boggling amount of premium for the risk imposed. Consider that for roughly the same price you can get a full homeowners policy that covers all of these perils: fire, earthquake, wind, lightning, theft AND MORE! It is crazy to equate the risk of flood to the risk of all those standard homeowner perils, combined! We provided this example to show that even without all the mapping and software tools available for pricing, what we can quickly conclude is that the NFIP pricing for these low-risk policies is absurdly high. Whatever the price “should” be for these types of risks, can you see that it MUST be a fraction of the price of a traditional homeowner’s policy? Don’t believe that either? Consider that the Lloyd’s is marketing its low-risk flood policies as “inexpensive,” and brokers tell us privately that many base-level policies will be 50% to 75% less expensive than NFIP equivalents.

The news gets even better. There are tens of millions of houses like this case example, with technology now available to quickly find them. These risks aren’t the exception; these risks can be a market in their own right. Let the mental arithmetic commence!

Summary: Differentiate or Die!

The Unwanted Commodity

Most consumers of personal lines products don’t have the time or the ability to evaluate an insurance policy to determine whether it provides good value. Regrettably, most agents and brokers don’t have the time to help them either. So, when shopping for a product that they hope they will never use and that they are incapable of truly understanding, consumers will focus on the one thing they do understand: price.

Competing on price becomes a race to the bottom (yay! – another soft market) and to death. But there is an opportunity here – carriers that compete on personal lines/homeowner insurance with benefits that are immediately apparent (like value, flexibility, service, conditions and, inevitably, price) have a rare chance to stake out significant new business, or to solidify their own share.

The flood insurance market is real, and it’s big enough for carriers to establish a healthy and competitive environment where service and quality will stand out, along with price. Carriers that would like to avoid dinosaur status can remain relevant and competitive, with no departure from insurance fundamentals – rate a risk, price it and sell it. It’s obvious, right?

Which carriers will be decisive and bold and begin to differentiate by offering flood capacity? Which carriers will evolve to keep pace or even lead the pack into the next generation of homeowner products? More importantly, which of you will lose market share and cease to exist in 10 years because you didn’t know what innovation looks like?