Pricing fairness: tackling big data and COVID-19 insurance discrimination

Research highlights the need for regulators, consumer advocate groups and industry associations to be involved in determining what insurers can – and cannot – discriminate against.

Man in shirt and tie shows document to young couple sitting at a desk

Insurers often discriminate and charge more from some people for the same product. Photo: Shutterstock

Unfair discrimination has been a crucial topic for the insurance industry over the past few decades.

The insurance industry is based on discrimination, and discrimination issues continue to evolve in part due to insurers' extensive use of big data, says UNSW Business School's Dr Fei Huang, Senior Lecturer and Honours Program Coordinator in the School of Risk & Actuarial Studies

"Discrimination issues regarding privacy and use of algorithms are taking on increased importance, especially as insurers' extensive use of data and computational abilities evolve," she says.

Dr Huang explores this in her paper: The Discriminating (Pricing) Actuary, alongside co-author Edward W. (Jed) Frees, Emeritus Professor in the Risk and Insurance Department at the University of Wisconsin–Madison and Professor in Actuarial Studies & Statistics at the Australian National University.

In the paper, the authors present different arguments to identify different situations when insurance discrimination is fair and ethical and when it is unfair and morally indefensible.

"For example, auto insurers often charge younger (and presumably riskier) drivers more than older (presumably safer) drivers, but do not make a distinction between brown-haired and red-haired drivers (presumably because the two groups are equally risky).

"So, discrimination based on age is done routinely, whereas discrimination based on hair colour is not," explains Dr Huang.

The treatment of 'related variables' is a key factor here, and raises potential issues. The use of variables related to a prohibited variable constitutes indirect discrimination. Although they do not have the usual characteristics of an unfair variable, they negatively affect society.

A classic example of this is redlining – drawing red lines on a map to indicate areas insurers will not serve, areas typically containing high proportions of minorities or people of lower socioeconomic status. 

"In this example, it was prohibited to use race as a rating variable and yet, through the use of a geographic proxy (such as an area with a concentration of minorities), at one time insurers were able to indirectly discriminate against minorities," says Prof. Frees.

They conclude courts, legislatures, and other stakeholders including regulators, consumer advocates, and industry associations should all play a role in determining what insurers can and cannot discriminate against.

Privacy and proxies

Big data presents two key issues: privacy and the use of proxies.

In terms of privacy, detailed information is sometimes provided voluntarily by individuals to insurers and is not treated as sensitive. This includes information from global position systems (GPS) that we put in our cars, comparable devices for our homes (the Internet of Things), devices we wear to improve our health etc., explains Dr Huang.

Insurers may also use other information that is not provided directly by individuals for their own commercial purposes. For example, privacy issues are raised any time a carrier classifies risks on intimate, personal information, like HIV status, marital status, sexual orientation, or genetic information, she says. 

Proxy discrimination occurs when a surrogate (proxy) is used in place of a prohibited trait, such as gender, race or nationality. This proxy is a facially neutral trait, like the size of an automobile's engine being used as a proxy for gender.

In the world of big data, complex algorithms are being developed using thousands of traits. Proxy discrimination may be unintentional; moreover, Dr Huang says an insurer may not even be aware it is engaging in discriminatory behaviour due to the opaqueness of machine-driven algorithms.

The impact of COVID-19

As with other parts of the global economy, the COVID-19 pandemic has rocked the insurance industry.

The lines of business most affected on the commercial side include workers' compensation, business interruption insurance, cyber liability, general insurance liability and event cancellation, while health and travel insurance have been most affected on the personal side, says Dr Huang.

But insurance legislation is being introduced to prohibit discrimination based on the diagnosis of this disease.

“For example, on 14 April 2020, the Australian Competition and Consumer Commission granted interim authority to the Financial Services Council and its members to ensure frontline healthcare workers are not excluded from coverage due to exposure to COVID-19.

"That means life insurers cannot use the exposure to COVID-19 as a factor for pricing or applying risk exclusions to any new policy,” says Dr Huang.

Such legislation has several implications.

"For example, in absence of this legal restriction, rates may well increase for grocery store workers, due to their exposure and increased suspicion of a diagnosis of COVID-19. Is this in the best interest of society?” Dr Huang asks.

"For a pandemic, the weight of evidence suggests societal concerns dominate and that a prohibition based on diagnosis, real or suspected, of COVID-19, is warranted."

Ensuring pricing fairness

Understanding the principles of ethical discrimination is vital for actuaries and other financial analysts. Actuaries are heavily involved in setting insurance prices; they are also often influential in determining the scope of contractual insurance coverages and who the company insures, both initially and at renewal, say the authors.

Indeed, the research provides a framework to help actuaries present financial cost recommendations in a meaningful way by summarising different perspectives.

To tackle discrimination in the industry, Dr Huang says actuaries can make significant contributions to these discussions by quantifying policy alternatives' financial impact.

Policies describing what insurers cannot discriminate against (having ever contracted COVID-19, for example) will also increase consumer confidence in the insurance system more broadly.

"Our position is not that actuaries should dictate whether or not information should be unrestricted, partially restricted, or prohibited.

"Rather, choices regarding insurance prohibitions involve policy choices should also involve legal and economic scholars, as well as government representatives and advocates for the industry and consumers," say the authors.