Life insurance underwriting has traditionally been perceived as an "art" due to the subjective judgment involved in the decision-making process. Not surprisingly, this has led to sporadic concerns about arbitrary decisions lacking scientific evidence and validated information. However, with the advent of big data, insurance companies now have access to a vast array of data, and the methodologies underpinning data analytics are maturing. As a result, the underwriting process is increasingly becoming data-driven, making decisions more science-based.

In evaluating risks, the underwriting process often needs to consider three important aspects: the design of the insurance application health questionnaire, the assessment of the non-medical limit, and the development of a underwriting manual. 

Design of the Insurance Health Application Questionnaire

Simplified Issuance Offer (SIO) allows for quick approval of the underwriting process without a medical exam, based on limited and simplified health questions. This approach is gaining acceptance in many Asian markets but can lead to underwriting substandard risks at standard rates. To price the risk accurately, underwriters typically add a “SIO loading”, which is challenging to quantify and assess. One solution is to evaluate and assign a value to each health question in the application form. The assessment of the value of a health question can be split into two parts, based on the applicants’ health declarations. [1]   

1. Impaired applicants[2] who honestly declare their health conditions: The assessment is mainly done at the underwriting stage. We recommend calculating a value (Disclosure Value) based on the prevalence of the disease (Disease Prevalence), an estimate of the share of honest disclosure of impairment (Disclosure Ratio), and an average extra mortality/morbidity loading on the disease (Underwriting Rating).

2. Impaired applicants who fail to declare their health conditions properly: The assessment is mainly reflected at the claims stage. For these applicants, we recommend calculating a value (Non-Disclosure Value) based on the prevalence of the disease (Disease Prevalence), an estimate of the share of non-disclosure of impairment (Non-Disclosure Ratio), an average extra mortality/morbidity loading on the disease (Underwriting Rating), and the projected share of non-disclosure claims being detected (Claim Detection Ratio).

The total value of a health question is the sum of these two parts (i.e. Disclosure Value and Non-Disclosure Value). Underwriters can estimate the value of all health questions by analysing available data and making assumptions for each impairment. Given the limited question in an SIO, underwriters can add a certain anti-selection loading to reflect the additional risk of missing health questions. For a better estimation of risks, an automatic calculation tool for SIO questions can be designed. This practice can facilitate the underwriters’ task and enable them to calculate the value of each health question more accurately and efficiently.

Assessment of the non-medical Limit

Medical exams are crucial for underwriters to control risk. Introducing simplified underwriting process without a medical exam thus represents a paradigm shift, often times necessitating the quantification of non-medical limit. This involves determining a reasonable non-medical limit and evaluating the pricing loading for high non-medical limit.

1. Determine a Reasonable Non-medical Limit (Breakeven Point Sum Assured)

Most insurers determine their medical table through surveys and peer references. Some use the Cost-benefit Analysis method – where the Cost of Medical Test = Breakeven Point Sum Assured * Saved Present Value per Unit Sum Assured. This is the traditional way to assess non-medical limit based on data analysis. In the Cost-benefit Analysis method, the most important cost (lapse cost) changes dynamically with the calculated breakeven point, making it difficult to estimate accurately. It is often observed that the calculated breakeven point is significantly lower than the average market level, causing the assessment results to be hard to use. To overcome this problem, an alternative method can be considered which is the Profit Maximisation method – by comparing the Profit of Medical Policies and Profit of Non-medical Policies (i.e. waiving the medical test) to derive the business growth rate for the studied portfolio. This method facilitates the calculation of the non-medical limit by replacing the challenging estimation of lapse cost with easily assessable parameters, making the non-medical limit more reasonable and explainable.

2. Assess Pricing Loading for High Non-medical Limit

Sometimes, the non-medical limit must exceed the reasonable breakeven point for business reasons. This means a portion of sub-standard applicants may be accepted at standard rates. To mitigate underpricing risk, a non-medical limit pricing loading is recommended to be assessed through the following steps.

The first step is the calculation of the "Primary Loading" which is defined as Primary Loading = (1-Disclosure Ratio) * Substandard Ratio * Underwriting Rating. The next step is to add other factors to the Primary Loading to derive the Final Loading. These other factors can be the policy distribution by age and sum assured, product net amount at risk (NAAR) etc.

Development of an Underwriting Manual

Drafting the underwriting manual relies primarily on two sources: medical literature and insurance experience data. However, due to insufficient data and system issues (such as a lack of optimisation of impairment codes or digitalisation of system data), medical literature often serves as the primary source for many insurers and reinsurers. This can lead to pitfalls since the studied population in medical research may differ from the insured population. As insurers accumulate more data and continue to optimise their data systems, their data platforms is expected to play a more important role in developing the underwriting manual.

On the other hand, insurers can consider leveraging health data from digital devices. With the widespread use of smartphones and wearables, analysing non-traditional health data can lead to the development of new or non-traditional underwriting factors and their respective manuals. With consent from customers to obtain and use their health data, these factors can be used individually or in combination with traditional factors to provide a faster and better risk assessment, eventually enhancing the customer experience.

In an era of rapid digitisation, underwriters should strengthen their capability in data analytics and quantifying non-traditional underwriting risks. These capabilities will support more accurate underwriting assessments and case evaluations. Ultimately, these skills will allow product development which better fits customers’ needs, a win-win situation for both insurers and their customers. A gradual shift from "Art" to "Science" should be the future direction of the underwriting profession.

Glossary

Breakeven Point Sum Assured: the sum assured limit at which the costs associated with medical tests equal to the benefits saved by medical tests

Claim Detection Ratio: the ratio of the number of non-disclosure claims detected to the total number of claims assessed

Disclosure Ratio: the proportion of applicants who disclose pre-existing health conditions or issues compared to the total number of applicants

Disease Prevalence: the proportion of individuals in a population who have a specific disease or condition at a particular time or over a specified period

Non-disclosure Ratio: the proportion of applicants who don’t disclose pre-existing health conditions or issues compared to the total number of applicants

Profit of Medical Policies: financial gain earned from policies with medical test

Profit of Non-medical Policies: financial gain earned from policies without medical test

Saved Present Value per Unit Sum Assured: discounting the long-term mortality / morbidity saving from medical test to the present value per unit sum assured

Substandard Ratio: the proportion of substandard cases detected by a medical test to the total number of cases with the medical test.

Sum Assured: the amount that the insurer agrees to pay to the policyholder's beneficiaries upon the insured person's death or other insurance.

Underwriting Rating: the additional premium charged by an insurance company to cover the increased risk associated with insuring an individual