Postcodes, credit scores may be 'ethnicity proxy'

An association with postcodes means ethnicity may still be impacting insurance policy prices via “proxy discrimination” – even though it is illegal in many jurisdictions to use protected policyholder characteristics when calculating premiums.

Research from London-based Bayes Business School says insurance pricing models must change if gender and ethnicity discrimination is to stop.

It is urging that a new method be introduced that takes risk predictions for policyholders based on all available characteristics – including protected ones such as gender or ethnicity – and are then “averaged out” from prices.

The researchers say this process “mathematically uncouples” protected characteristics from variables such as postcodes or credit scores and avoids proxy discrimination, enabling “safe” use of these variables for risk discrimination in pricing models.

While this “counterintuitively requires the use of protected characteristics to generate discrimination-free prices,” the researchers say without such information insurers are unable to compensate for proxy discrimination. The artificial intelligence-based method only requires the collection of sensitive information from a small subset of policyholders.

The paper comes as Washington regulators attempt to ban the use of credit scores in insurance pricing models due to potential for proxy racial discrimination. Investigations into possible civil rights violations are underway.

The report examined data from a motor insurance portfolio and found young members of one minority ethnic group would be charged premiums that are higher than if ethnicity had no insurance price impact.

“It is illegal in many jurisdictions to use protected policyholder characteristics when calculating insurance prices, although there remain concerns that such characteristics may still be impacting insurance quotes indirectly,” the paper says.

“Policyholders’ postcodes are commonly used in the calculation of their insurance premium, but this information could be an effective proxy for determining ethnicity.”

Professor Andreas Tsanakas says insurers should demonstrate discrimination is not a material issue in their portfolios and adjust their prices if it is.

“A policy framework for this process is currently lacking, such that insurers do not have clear regulatory signal on how to manage this problem,” he said.

“We have proposed a practical, adoptable method for removing the effects of discrimination from pricing models by removing the proxying of characteristics.

“Proxy discrimination is a real issue, and it should be addressed in pricing but to address this problem insurers need to collect information on protected characteristics, which in turn raises privacy concerns.

“It is important that strict protocols are introduced by regulators about how such information is collected and used, and how this process is explained to policyholders.”

The proposed method for addressing proxy discrimination is detailed here.