During the last several years, many new evidence and data sources have become available to the life insurance industry.
This has enabled both the development of more sophisticated simplified issue (SI) products and expanded qualifications in accelerated underwriting (AU) programs. The ongoing evolution of these increasingly important underwriting methods has raised a critical question: Are AU and SI converging? To answer this nuanced and multifaceted query, life insurers need to start by asking five other questions.
Question 1. Who is applying?
When fishing, what determines the type of fish being caught? Is it the pond or the fisher? In insurance, a similar question can be asked regarding the effect of the target market and distribution channel on the applicant pool.
The target market is like the pond and is often defined by core ages and coverage amounts. While a product may target younger people for lower coverage levels, it is important to consider the edges of the offerings as well. Ask: Do the limits make this product a market outlier? Will it be cross shopped against a product with more rigorous underwriting and substantially lower premiums?
Factors beyond age and coverage amounts may be more indicative of the mortality level or anti-selection risk. These include elements such as membership in an affinity group or motivation for buying (e.g., becoming a new parent or homeowner). Socioeconomic factors can also have significant impact as identified in recently published research from the Society of Actuaries.
Lastly, the application of tools, such as propensity-to-buy models, can influence the target market as well through both underwriting and product marketing. Examining the impact of all tools can inform the underwriting process and attract individuals at risk levels that correspond with product pricing.
Beyond the pond is the fisher’s activity – the distribution channel. One key factor is the presence or absence of agents. The types of applicants entering the sales funnel via an agent can differ significantly from those entering though a direct-to-consumer (DTC) channel – which is more akin to actively fishing vs. simply dropping a net into the water. In addition, a lack of agents may limit applicant coaching, but it also removes field underwriting, which could lead to individuals applying for a product not well-suited to their risk profile. In a DTC setting, technology is expected to play a key role in both limiting non-disclosure and screening applicants. The effectiveness of digital platforms can vary greatly.
Whether working with agents or through DTC approaches, distribution incentives are a crucial consideration. Regardless of channel, there may be similar incentives to maximize sales volume. The alignment of incentives is a key factor in establishing a mutual goal of writing sustainable business.
Question 2. What information is used to make decisions?
New evidence is changing the underwriting process for both SI and AU business. Traditionally, AU employed more evidence and more robust evidence than SI products. AU utilizes an application with a full Part 2 medical information, compared to the short-form applications more commonly used in SI, and makes greater use of new data sources and modeling techniques.
The distinction between the two processes is beginning to narrow. SI products are now adopting tools that have historically been more closely associated with AU, including more robust applications and new scoring tools. On the AU side, insurers are increasingly reducing requirements for full evidence, such as lab panels and paramedical exams, by using alternative evidence sources including medical billing data and historical clinical lab results.
Regardless of the tools in use, it is critical to ensure they are working together whether in an AU or SI environment. For example, it may be beneficial to pair a propensity-to-buy model with a propensity-to-qualify model to address the potential correlation between buying appetite and risk level. Taking time early in the process to ensure all tools are working together as intended can help reduce unintended results later.
Question 3. What happens to those who don’t qualify?
The differences between the AU and SI application processes create a substantial gap in qualification rates. An applicant disqualified by the AU process would traditionally be required to undergo an exam and lab testing. In contrast, those who do not qualify for a SI product are typically declined. Paired with different risk levels – usually a standard or better risk for AU or risks up to Table Rating 3-4 for SI – these process differences give rise to significantly different qualification rates. That leads to the question: What factors may be narrowing this gap today?
AU is evolving to introduce a “middle lane” process, allowing cases that cannot be accelerated in the traditional sense to be approved without full age and policy amount requirements. Additional review of these cases could be as simple as an underwriter reviewing information a bit too complex for an automated engine or considering supplemental evidence such as historical clinical lab results, medical billing records, or even an attending physician statement (APS).
The rise of new data sources is also spurring development in SI underwriting, enabling more granular risk segmentation in the absence of traditional lab results. While many SI premiums historically varied by gender and smoker status alone, products can now offer preferred classes. Such options open the door to price points for some SI risk classes that are closer to some fully underwritten (FUW) product risk classes.
Question 4. How do you determine the baseline mortality assumption?
The answers to the three previous questions can all affect mortality performance. Perhaps the most obvious impact stems from the applied evidence. But what about the market segment? Does that really matter? Historical mortality experience demonstrates that business with very similar underwriting requirements sold to different populations may result in different levels of mortality. This is true even when blood testing is performed on every applicant. These effects could be magnified with a weaker underwriting filter.
Historical experience also shows a significant gap between FUW and SI mortality, with many SI programs experiencing mortality nearly double that of an otherwise similar FUW product. Along the mortality spectrum, AU is expected to be much closer to FUW – within 25% in most cases – with some programs far less than that.
As SI programs gain sophistication to presumably bring down mortality experience and AU programs expand to include more applicants and potentially increase mortality experience, does this mean they are converging? In most cases, even if AU and SI mortality move closer together, they still may not meet any time soon because the gap remains quite wide.
Additionally, these products tend to attract different buyers, which may be a further headwind to mortality convergence. For new products that seem to blur the lines between SI and AU, considering the mortality assumption from multiple angles may be informative. For example, how do assumptions differ if loading up FUW compared to discounting SI?
Question 5. How do we measure success?
Measuring progress against goals has been another point of differentiation between AU and SI. AU programs have featured robust monitoring programs, such as random holdouts and post-issue audits, designed to provide performance indications much more quickly than credible claims experience.
SI programs, on the other hand, traditionally have not had this type of monitoring in place, waiting instead for mortality experience to emerge. That noted, some concepts employed in monitoring AU business could translate effectively to SI products, particularly if policy limits continue rising. For instance, what might be revealed when conducting post-issue APS reviews for SI policies in the highest face amount band?
Other metrics can be informative as well, even if their impact on mortality is secondary or tertiary. Placement rates may offer insight into the degree of cross-shopping among prospective buyers. Lapse rates, particularly in early durations, may provide early evidence of potential anti-selective behavior. The same scoring tools used in underwriting can offer a glimpse into the makeup of market space by analyzing score distributions for all applications, not just approved policies.
Even more traditional distribution analysis can be useful, with trends on attributes such as age and face amounts providing perspective on whether the product is reaching its intended audience. Similarly, if surging sales are observed at the product’s fringes (i.e., the very highest ages and face amounts), what might that indicate? Distribution channel analysis also may open the door to enhanced opportunities by identifying patterns of favorable risk characteristics and behaviors (e.g., high rates of disclosure) and minimizing loss from anti-selective behaviors such as non-disclosure.
In summary, a confluence of developments in recent years are enabling more AU applications to be approved without traditional evidence and increasingly granular SI pricing. While starting from different points along the mortality spectrum, AU approvals are increasing and SI refinement is producing lower mortality rates and correspondingly lower premiums.
Are these paradigms converging? Yes, but in varying degrees. The five preceding questions offer a means to evaluate elements that may influence the answer in one direction or another for different carriers. Will SI and AU ever meet in the middle? Maybe. The industry may be getting closer, but it will take more time.
Learn more about trends impacting the insurance industry from Taylor Pickett and Ryan Holt at the SOA Life Meeting August 24-26 in Chicago. Taylor will be speaking at the session "Point, Counterpoint – Experts. debate hot topics in underwriting," and Ryan will be presenting at the session "Fluidity of Fluidless Underwriting: Time to drink the Koolaid?"