Consumer protection in the age of big data.

Author:Helveston, Max N.
Position::II. The Data Revolution Will Change Insurance Markets, for Better and for Worse C. Advanced Insurance Analytics' Threat to Traditional Market Goals through Conclusion, with footnotes, p. 887-917
  1. Advanced Insurance Analytics' Threat to Traditional Market Goals

    The significant benefits that would be gained by advances in insurers' predictive abilities and actuarial fairness, however, would not be without costs. As described earlier, any commercial entity's collection and use of data increases certain risks. (134) Drastic expansions in insurers' use of data would not only raise these generic problems, but would pose unique threats to public interests.

    While increased actuarial fairness would improve several metrics used to evaluate insurance markets (e.g., subscription rates, rate fairness, insurer stability), these are not the only qualities that healthy markets are expected to possess. A market that performs exceptionally well on the factors improved by actuarial fairness could still be deemed toxic if, for instance, it systematically disadvantaged racial minorities. It is possible to identify several values society expects insurance markets to respect and that would be imperiled by the industry's expanded use of data analytics.

    This Subpart provides a brief description of each of these values and how they are affected by insurers with advanced data capabilities. In the course of doing so, it shows why an insurance market that is fully committed to actuarial fairness and loss reduction would not only fail to advance these values, but would actively work against them. This sets the stage for Part III's discussion of how regulation could help ensure that markets possess an optimal balance of these societal interests.

    1. Personal Liberty and Autonomy Norms

      One set of values that the public expects markets to recognize is the preservation of personal liberties and autonomy. As a general premise, modern society has attempted to prevent individuals' freedom from being limited in ways that it deems to be unfair, exploitative, or coercive. Much of contract and consumer law seeks to regulate market conduct in ways that prevent these types of abuses. (135)

      Given that all contractual agreements place constraints on autonomy, it cannot be the case that limiting autonomy is an inherently negative act. Rather, society has identified specific types of restrictions as being too unfair or too exploitative. (136) It is only these restraints that markets are expected to respect.

      Commercial deals can impermissibly limit individuals' autonomy in two situations. Most commonly, this occurs when a powerful party forces a consumer to agree to contractual terms that unfairly limit her freedom, either by compelling, banning, or incentivizing certain conduct. (137) There are countless examples of legislatures, agencies, and courts creating rules that make specific contractual terms unenforceable. (138) Less often recognized is the concern for consumer autonomy in situations where a powerful party engages in pre-contractual behaviors that impermissibly affect an individual's behaviors. (139) Due to the fact that price discrimination is uncommon in consumer goods markets, the state has rarely needed to police consumer markets for this type of behavior. (140) Perhaps the best examples of regulation of pre-contractual behaviors are laws that ban insurers from considering certain types of information when determining whether to insure individuals. For instance, the Affordable Care Act's community rating rules prevent insurers from refusing to insure diabetics, nullifying any behavior-modifying incentives that would be created by such a practice. (141)

      Insurers' means for constraining the autonomy of private actors are relatively straightforward. Whenever an insurer adopts a rule that governs its price-setting, underwriting, or other customer-related processes, it has taken an action that may force a private actor to behave in a certain way. (142) For example, an insurer might influence policyholders' behaviors by deciding that it will no longer offer commercial property coverage for warehouses that lack sprinkler systems. Whether a private party's behavior is affected by an insurer's policy will depend on a variety of factors, such as the availability of alternative providers with different policies, the degree to which the private party needs coverage, etc. If such rules attain a sufficient level of ubiquity across insurers, however, consumers will have no choice but to comply with the rules' requirements or forego coverage. (143)

      It is widely accepted that insurance companies' practices have large impacts on consumer behaviors. Insurers compel policyholders to take certain actions by fiat (e.g., we will only insure you if you install smoke alarms) and economic incentives (e.g., we will give you lower rates if you install a telematics device in your car and drive a certain way). Individuals' autonomy is also constrained by the conditions that insurers place on coverage. Insurance policies are somewhat unique among consumer contracts in the number of continuing obligations that they impose on policyholders. In homeowners' policies, for instance, the standard policy contains requirements concerning claim reporting, (144) cooperation with the insurer regarding claim investigation and claim-related litigation, (145) payment of premiums, (146) cancellation and non-renewal of the policy, (147) and duties regarding post-loss repairs to the property. (148) Further, the limitations placed on covered losses can influence policyholders' behaviors--they can affect individuals' maintenance and occupancy decisions, (149) discourage construction projects, (150) and alter personal behaviors. (151)

      Insurer-imposed regulation of consumer conduct is not an inherently bad thing. The industry has a long history of premising policyholders' coverage on their installation and maintenance of safety equipment or on the performance of other risk-reducing actions. (152) For instance, insurers have required business owners to install sprinkler systems in their buildings, retain security personnel, and refrain from performing specified types of business activities as conditions of coverage. (153) While such demands directly regulate applicant and policyholder behaviors, they do so in a way that most individuals recognize as legitimate.

      But not all of insurers' conduct-regulating measures have been as innocuous or well received by the public. Perhaps the best-known example of an objectionable exclusionary rule would be pre-ACA health insurers' refusal to offer coverage to individuals with certain chronic or severe diseases at any price. (154) Another would be certain companies' refusal to issue homeowner's insurance to households with pets that it deems to be high-risk (e.g., pit bulls). (155)

      Even if one believes that insurers have not yet overstepped their bounds in constraining consumer autonomy, this could easily change in the Big Data era. As will be discussed in Part III, the current regulatory system gives insurers large amounts of discretion over their core business practices--for example, how they set rates, what types of coverage they offer, etc. (156) Once insurers have unlimited access to data and analytics, it is reasonable to believe that they will begin to abuse this discretion and act in ways that threaten consumer autonomy. Why? First, such insurers will have much more information about what acts, characteristics, and qualities correlate with risk. As described earlier, market forces will drive them to use this information to discriminate among consumers or risk losing their low-risk customers to competitors. (157) Second, technological advances will enable them to monitor policyholder behaviors at increasingly lower costs. (158)

      The data revolution will encourage insurers to take actions that will influence individuals' decisions. Insurers can affect individuals directly (e.g., requiring applicants to take (or refrain from taking) certain actions as a prerequisite for coverage (159)) as well as indirectly (e.g., refusing to provide coverage to households with certain types of pets (160) or setting incredibly high rates on homeowner's insurance in certain neighborhoods). (161) Using this power to compel some types of behavior, such as compliance with safety codes, is not objectionable. Advances in data science, however, will create incentives for insurers to reach far beyond this point. For instance, it is already the case that insurers issuing homeowner's policies are considering whether they should mandate (or incentivize through discounts) the installation of certain forms of monitoring equipment in insured homes. (162) It is easy to imagine that insurers in other lines are exploring similar policies. (163)

      Insurers compelling private conduct will have significant repercussions for consumers' personal liberties. As the scope of conduct that is analyzed by insurers grows, the domain of conduct that can be regarded as personal (i.e., decisions that individuals are entitled to make free of coercion by the state or other external actors) will shrink. If the datafication of the world becomes as extensive as some have projected, then a sword of Damocles might loom over many personal decisions. One's expressions of political views, whom one befriends or networks with, and other actions that have been regarded as wholly within the personal sphere could become affected by insurers' views of these activities.

      The concern that Big Data analytics will have a coercive effect on personal decisions extends to areas of individuals' lives that have a long history of being protected from outside influences. Religious affiliation or membership in political or social groups provide the most salient examples of such domains. (164) Assume data analysis indicates that members of a certain societal subset (e.g., supporters of the Tea Party movement) have significantly higher than average auto risk profiles, while members of another group (e.g., Roman Catholics) have significantly lower than average risk profiles. Allowing price discrimination on the basis of political group...

To continue reading