Age is an important rating factor in such products, since older people are generally more at risk of health emergencies than younger people. TIPL is considering two separate product constructs: • One option is a simple product sold via a ‘white label’ provider. A customer inputs their age, destination and travel duration to receive a product priced on these attributes, allowing a ‘personalised’ price to reflect a customer’s risk. • Another option is for an ‘embedded’ product where insurance is offered alongside other travel products, such as flights and accommodation. In this case, there are no explicit questions about age or trip duration. With no access to detailed information, customers receive the same price and as such products will include inherent crosssubsidies between younger and older groups. The insurer aims to use the insurance exemptions under the ADA. Their data confirms that health risks increase as age increases, but data becomes sparse and unreliable above the age of 90. TIPL is unable to reasonably obtain additional data for people above the age of 90. For simplicity, this illustrative case study is focused on age and the ADA alone. However, from a practical perspective, elderly people may also be more likely than the general population to have health conditions of relevance to travel insurance underwriting and so the DDA may also be relevant to consider for travel insurers in this situation. White label provider As the data is sparse above age 90, the following options are considered for pricing: A. A smooth extrapolation following the general exponential trend for people under 90, and ignoring the sparse, unreliable data for people over 90. It is argued that the extrapolation is justified given that age generally affects health in this manner. B. A smoothed approach using a different extrapolation method. Whilst it has a similar statistical accuracy to Option A, it results in significantly higher premiums for people over 90. C. A flat rate for people over 90. This group already pays the highest rate, and there is limited evidence in the data alone to justify continuing to increase prices over age 90. It is also noted that a greater proportion of people over 90 cannot travel because of their health. Unless an exemption applies, charging higher premiums to people over 90 may constitute direct discrimination on the basis of age. Given that TIPL does not have, and cannot reasonably obtain, actuarial or statistical data on which it is reasonable to rely, the insurer would not be able to rely on the data exemption under the ADA. Guidance Resource: Artificial intelligence and discrimination in insurance pricing and underwriting • 2022 • 33