5 | Case Studies: Challenges for insurers Summary Where a protected attribute has a distinct relationship with risk, and the insurer charges premiums to customers on the basis of that protected attribute: • For sex, disability, and age, any discrimination will not be unlawful provided the data exemption under SDA, DDA and ADA can be relied upon. • For protected attributes under the RDA, as no data exemption is available for insurers, this is likely to constitute unlawful direct discrimination. Where a protected attribute has a distinct relationship with risk, even if the insurer does not use the protected attribute in price setting (for example, if no insurance exemption applies), it may be at risk of unlawful indirect discrimination, as other correlated rating factors may still infer this relationship within a risk model. However: • This may be mitigated by careful use of the protected attribute in the construction or training of the risk model, where suitable data is available. There are a range of proposed methodologies for this in the academic literature.75 • The insurer may not be able to mitigate against the effect of this protected attribute in price setting where no such data can be reasonably obtained about the protected attribute. The availability of such data is a factor that may be considered by the courts in determining whether the pricing (being the requirement imposed for the purpose of the Discrimination Acts) is reasonable in the circumstances and, hence, whether unlawful indirect discrimination has occurred. This scenario assumes that CIPL has statistical data that shows the riskiness of drivers from Group A and Group B. However, insurers will not necessarily have data clearly showing the risk factors of different groups and it may not be possible, or appropriate, to obtain such data. In such a situation, without relevant data, the insurer may not be aware of any potential discrimination and may not be able to mitigate against it. 5.2 Case Study: Travel Insurance This case study explores common challenges for insurers when defining standard product constructs, terms and conditions which are reliant on data or models. Travel Insurance Pty Ltd (TIPL) is building a pricing model for a new travel insurance product. This product provides emergency health coverage for accident and sickness during travel. 32