LexisNexis shifts life insurance underwriting into high gear
Compliance isn’t the only pet peeve of insurance agents and advisors. Another is underwriting.
The rating of a life insurance policy, which can make a purchase untenable for the client, is one source of annoyance. Potentially more aggravating is the wait time—the worst cases extend to weeks or months—in procuring physician statements or other health data needed to bind a policy application.
Such delays may soon be a thing of the past. A myriad of “insurtech” and “fintech” players are joining a burgeoning market for solutions that automate underwriting. These include software and services that not only assess the policy risk but also provide workflow, audit, and data analytics capabilities.
A long-established company in this area, Alpharetta, Georgia-based LexisNexis Risk Solutions, is betting that software which speed policy transaction times by dispensing with medical underwriting will prove transformative to the market. The company’s claim to fame, Risk Classifier, aggregates a mountain of information—billions of public records and more than 20,000 data sources—to assess mortality risk for a policy applicant.
The kicker: The tool can use this data to approve preferred policies—no blood work of doctor’s record needed. For advisors and their clients, the benefit can be measured not only in fewer hassles and speedier policy binding but also greater convenience. When paired with a self-service web portal offering other digital capabilities, including product education, a needs analysis, and online application processing, the tool can secure coverage in minutes.
To learn more about the offering, LifeHealthPro interviewed Elliot Wallace, a vice president and general manager for life insurance at LexisNexis. The wide-ranging conversation—conducted in advance of the National Association of Independent Life Brokerage Agencies‘ annual meeting, taking place Nov. 17-19 in Dallas, where the executive will be representing the company—explored topics likely to be of keen interest to life insurance brokerages and affiliated agents.
Among them: data analytics capabilities underpinning the tool; the ability to assess mortality risk using motor vehicle records; and how “accelerating underwriting” may evolve amid other trends buffeting the industry. The following are excerpts:
LHP: What are LexisNexis’ objectives at the NAILBA annual meeting? How might NAILBA’s members benefit from the LexisNexis data?
Wallace: LexisNexis Risk Solutions main objective is to lead discussions around accelerated underwriting and meet with carriers, distributors and brokers to answer their questions about data, analytics and how the distribution broker community can leverage new processes to increase their businesses.
Brokers, distributors and the entire life insurance community are going to benefit from the transformation we’re experiencing now. Consumers have different expectations today. The more data we can use to improve marketing, application fulfillment, underwriting and policy owner services and the more we can align life insurance products with their needs with how consumers want to do business, the more success we can expect.
LHP: How are motor vehicle records, a key component of the LexisNexis tool, predictive of mortality experience?
Wallace (pictured below): Motor vehicle records or MVRs, have “protective value,” that is, they provide insights into an individual’s lifestyle risk. By examining the number and severity of moving violations on an applicant’s MVR, life insurance underwriters can accurately assess the mortality risk associated with a policy applicant.
If you have multiple points in your MVR for driving under the influence, then you’re living a risky lifestyle. You’re also more likely than not to have an underlying health issue that could evolve into a life-threatening condition, such as liver cancer.
So motor vehicle records let life insurers price policies more accurately. We put these MVRs and other data sources into our Life Risk Classifier predictive modeling software for determining a policy applicant’s relative mortality.
This model is key to doing streamlined or accelerated underwriting online. After a prospect submits an application, an insurer can underwrite and bind the policy — generally within 15 minutes. We have carrier clients doing this today.
LHP: Can you elaborate on the application process? How many steps are involved?
Wallace: There are three. The first is identity verification: checking the accuracy of information the prospect provides — name, address, date of birth and so on. In the next step, we pre-fill fields in the online app based on the applicant’s data.Thereafter, our predictive model, a proprietary software algorithm compliant with the Fair Credit Reporting Act, assigns a score to the individual’s “relative mortality.” This value tells us whether he or she represents an above- or below-average risk of dying by a certain age compared to the general population.
The insurer then coverts this score into a risk class — super-preferred, preferred, standard, etc. — all in real time and without time-consuming medical underwriting. You can dispense with blood samples, attending physician statements other health records about the applicant.
LHP: Let’s talk now about clients. Which life insurers are using the LexisNexis solution to speed underwriting? Or is this proprietary information?
Wallace: Most of our partnering carriers don’t want to be named. One I can disclose is the Savings Bank Life Insurance Co. of Massachusetts. SBLI is using our data to underwrite a block of new policies.
LHP: Why not the whole book of business? Is there skepticism as to the accuracy of the data?
Wallace: Let’s call it due diligence. SBLI is looking to validate the data by insuring that future claims experience aligns with the predictive model. In a 2012 benchmarking study of more than 7.4 million motor vehicle records, LexisNexis and RGA Reinsurance Co. found that MVRs can reliably predict an individual’s “all-cause mortality,” or the death rate from any cause.
For example, people with major violations, such as alcohol-related infractions and excess speeding, have all-cause mortality that is 70 percent higher than individuals who don’t. The presence of 6 or more driving violations on an MVR elevates all-cause mortality by 80 percent.
Individuals with a higher number of violations represent the worst risks. For women, the all-cause mortality rate is 100 percent greater among those with these violations than for those who don’t. In short, our predictive model held its own against conventional underwriting. The 2012 study evolved into three-year conversation with carriers to convince them the data is real and actionable.
LHP: But just how accurate? If you were to plot mortality experience on a chart, there would surely be some outliers — policy claims that don’t dovetail with the predictive model — yes?
Wallace: Correct. But we can identify many of these outliers by combining the model with a drug history check. If the policy applicant is taking prescription medications for certain health conditions that garner a low score — making life expectancy harder to forecast — then the carrier may opt for traditional underwriting.
LHP: I imagine also the insurer may prefer medical underwriting for particularly large policies. What face amount ranges are insurers underwriting using the data?
Wallace: They typically start with policies carrying a death benefit of $350,000 or less, then ratchet up the amount gradually. Some of our clients now cap the contract limit at $1 million or $2 million.
LHP: These would seem to be amounts sufficient to cover many middle market households. What resources went into building your predictive model? Who on the LexisNexis staff was involved?
Wallace: We have data scientists on an analytics team. From an IT standpoint, we’re systems-agnostic. And, in fact, we don’t provide software to the carriers. Rather, they send us the data from their policy applications. We then run the model at our site and return back a score they use to assign a risk classification.
No underwriters are involved. Again, the only the cases they see are those falling below a minimum threshold in our scoring system. Scores above the threshold are immediately sent thorough accelerated underwriting and a policy is issued.
LHP: I would think the Risk Classifier would lend itself to websites like Policy Genius, a one-stop shop where prospects can get a quote, access educational resources, complete an application online and be approved for a policy. Is LexisNexis partnering with carriers that are developing such portals?
Wallace: Some of our distribution partners have built what you’re describing. All we’re doing is providing the accelerated underwriting component.
LHP: Let’s move on to another hot underwriting topic, cannabis. Does the LexisNexis data factor in mortality risk for those who smoke marijuana for recreational or medical purposes?
Wallace: Though it could affect the mortality rate, marijuana is not currently factored into our predictive model. We do have a separate solution coming to market that scores mortality risk for tobacco smokers that could be adapted for marijuana users. But this would be separate from the relative mortality scoring we’ve been talking about.
LHP: We had earlier interviewed a famous geneticist, Dr. Craig Venter, whose team at Human Longevity Inc. is creating a comprehensive genetic database using large-scale computing and machine-learning. Do you envision incorporating new genetics discoveries into the Risk Classifier in the future?
Wallace: Great question. I can’t say “near future,” as there are different regulatory hurdles to overcome. But once we do, the genetics data would become another attribute of the solution. So, yes, in the longer-term we do see the potential to feed that type of information into our predictive model.
LHP: How else do you see the LexisNexis product, and the company itself, evolving over the next three or five years?
Wallace: We’ve been active in the insurance industry, including the property and casualty market, for more than 25 years. Our predictive model is just one component. We also market data and services for compliance, credit risk management, fraud, identity and investigative needs.
In terms of insurance underwriting, related products include our electronic inspection reports, which offer fast access to public records information about proposed insureds. The data includes, for example, criminal records history, properties owned and bankruptcies.
For life insurers, we’ll be rolling out new products over the next several years. In 2017, the tobacco smoker model we talked about we’ll be coming online. We also have a lot of marketable wealth data that can do a needs analysis for proposed insureds.
LHP: When do you think accelerated, non-medical underwriting will be the predominant industry practice?
Wallace: For certain face amounts — say, life insurance policies topping $20 million — insurers will likely continue to demand attending physician statements and blood work. Nobody wants to be on the hook for $20 million dollars, right?
But for policies carrying less than $1 million or $2 million in death benefit, carriers will have to adopt predictive modeling in the next several years — be they solutions from us or a competitor — or they’ll be left behind. Their sales will decline as prospective customers turn to other insurers that offer a fast, hassle-free experience.
Secondly, the adopters can look forward to reduced underwriting expenses: Invasive procedures like blood testing or medical exams are more costly than our predictive model. Add to this the convenience now available to consumers — the ability to apply and be approved for a policy in minutes, wherever their location — and it’s clearly a win-win for carriers.
LHP: One last question: How do you see accelerating underwriting benefiting our readers — life insurance agents and financial advisors — as opposed to carriers’ direct-to-consumer distribution channels?
Wallace: The benefit will be as big for your audience as for direct-to-consumer. The value-add producers provide is in helping consumers determine how much life insurance they really need. What they definitely can do without is time wasted on underwriting. With predictive modeling, they can strip that way and just focus on what they do best: the needs analysis, closing the sale and binding the policy right at the office or the client’s kitchen table.