Big Data, AI and the Rate-Making Revolution Can Mean Intellectual Property Trouble for Rating Organizations

The insurance industry has increasingly relied upon the use of big data to enhance its operations and core functionalities, and one way this trend has manifested is in the realms of property and casualty business, where big data is being leveraged using artificial intelligence and other techniques to develop rating factors, methodologies and other rate related functions — this in lieu of decades-old process of relying upon demographic information and actuarial science alone.

While many large insurers develop these tools in-house, there is a burgeoning sector of third-party providers specializing in creating proprietary software, models and rating mechanisms to modernize the underwriting process. In regulatory terms, these insurtech organizations, which assist member carriers by developing and providing loss or expense statistics and rate-making services, differ significantly from state to state and are called by different names.

For instance, California only allows “advisory organizations” under its Insurance Code §1855.2, which may develop and prepare policy or bond forms, or manuals, for use by admitted insurers. However, advisory organizations do not act in a rate-making capacity or file rating plans, although as part of their policy forms and manuals these entities can prepare policy writing rules, rating plans, classification and territory codes and descriptions, prospective loss costs, and rules regarding increased limits factors and classification relativities. Yet in Massachusetts, “rating organizations” pursuant to Massachusetts General Laws Chapter 175A, Section 6, are authorized to file manuals of classifications, rules and rates, rating plans and associated modifications.

Notwithstanding the nomenclature differences, for simplicity we shall refer to them collectively as “rating organizations” — the underlying principle is that they support insurers by collecting and analyzing data and developing information to use in rating. And because they do, these entities register with — and are consequently regulated by — state departments of insurance. Most rating organizations agree that the registration process on its own is not a terribly difficult one, nor is the attendant rule that the work product rating organizations wish to offer their members be filed in the states where carriers wish to use it. But here is the rub: Once the filing is made, and depending upon the relevant laws, state regulation can be very problematic.

The Trade Secret Conundrum

The rating organizations developing these rating tools clearly want to preserve trade secret protections around their proprietary tools and other information that goes into product development. Indeed, the proprietary nature of the data and analytics developed by these rating organizations is a pivotal aspect of their business models, typically guarded as trade secrets to maintain competitive advantage. U.S. laws acknowledge the importance of safeguarding proprietary and trade secret information. Maintaining confidentiality is crucial to preserve the integrity and value of such data, although it can still be sold or licensed without forfeiting its protected status.

Regulators seek complete rate filings that allow them to understand the rates and other information being filed, which often calls for transparency into the rating organizations’ proprietary models and data. In general, rating organizations are willing to disclose sensitive information in regulatory filings that are only viewed by departments of insurance. The challenge arises when this material becomes public by virtue of it being filed, and the degree to which this data remains confidential will vary significantly across states, leading to a complex legal landscape.

In California, for example, filings by rating organizations are available for public inspection. Consequently, a rating organization’s proprietary information might need to be disclosed as part of these filings, which can create real headaches for them by virtue of the department’s historic position that all information submitted for rate applications should be publicly accessible, regardless of its confidential or proprietary status.

At first glance, Massachusetts seems to adopt a different approach. That state requires filing of rates, as well, yet the operative chapter for many types of property/casualty risks (Chapter 175A) declares that nothing in the filing requirement is intended “to prohibit or discourage reasonable competition.” This statement of intent appears to give the Massachusetts Division of Insurance a policy instruction that would favor preservation of proprietary information (read: trade secrets).

In addition, Massachusetts generally recognizes confidentiality over all sectors of industry governed by its laws. That being said, in a 2006 Decision and Order on Massachusetts Property Insurance Underwriting Association Rate Filings, then-Insurance Commissioner Julianne M. Bowler denied a motion to strike a report that included confidential rating information because rate proceedings, like other rate approval hearings, are public, and rate filings, which initiate the public hearing process, must be sufficient to enable any person who might seek to challenge proposed rates to fully evaluate underlying information and methodologies. To this day, the Division of Insurance takes the position that all filings and their contents become a public record at disposition.

Together, the stances taken in jurisdictions such as California and Massachusetts present a conundrum for rating organizations and insurers, balancing between regulatory compliance and the protection of trade secrets. These legal and operational intricacies underscore the evolving challenges of insurance rate development in the age of big data and artificial intelligence (AI).

As the industry continues to navigate these complexities, the tension between transparency and confidentiality remains a critical issue for all stakeholders involved.

Conclusion

The integration of big data into the insurance industry represents a seismic shift towards a more sophisticated and tailored approach to rate-making and underwriting processes. Still, this evolution brings with it a complex interplay between innovation, competition and regulation. As insurance companies and rating organizations lean on proprietary algorithms and data analytics, the need to protect sensitive information conflicts with regulatory demands for transparency and accountability.

While some states do allow for confidentiality of rate filing information, other states, such as California and Massachusetts, highlight the challenges that rating organizations face in navigating legal requirements while striving to maintain the confidentiality of their trade secrets. The balance between facilitating reasonable competition and ensuring consumer protection is delicate and varies significantly across jurisdictions.

As the industry moves forward, it will be crucial for insurers, rating organizations, and regulators to harmonize the objectives of promoting innovation, protecting proprietary information, and ensuring fair and transparent insurance practices. The resolution of these issues will not only shape the future of insurance rate making but also influence the broader trajectory of how big data and AI are utilized in regulatory environments.