Actuaries may not think of themselves as inventors or as persons who might get a patent in their name. But some recent patents issued for a statistical sampling technique show that “patentability” can apply to novel techniques developed in the actuarial profession.
The consulting firm Towers Watson touted receiving two patents for a statistical modeling technique in a March 15, 2012 press release:
Global professional services company Towers Watson (NYSE, NASDAQ: TW) announced today that the United States Patent and Trademark Office (USPTO) has issued two key patents for its innovative Replicated Stratified Sampling (RSS) financial modeling technique for the insurance industry….
The Towers Watson RSS modeling technique uses statistical sampling to accurately measure changes in risk metrics while producing dramatic reductions in run time. Sampling has long been used in other industries, in applications as familiar as television ratings and consumer polls. But in the insurance industry, where access to the full population of contracts is not a barrier, sampling has not been commonly used in the past.
Generally when an insurer performs certain calculations, it includes data from all its policies. If it has a million policies, that means a lot of processing as various scenarios are considered. Sometimes, the work can take days.
I believed I had a solution to this cumbersome and costly process: create subgroups from the database, sample policies from each, repeat the process several times, then combine the results.
My technique provides results similar to those from studying all policies, and saves time and money. And unlike consumer polling, which requires finding participants, insurance databases contain ready-made samples.
As far as I knew, no one had proposed this model. I developed and tested it for several months before I told my boss, Craig Buck, head of United States life insurance consulting. We agreed that it should be patented, and that Towers Watson could license it to clients.
An actuary might normally think of the development of a new modelling technique as something to publish in a paper of the Society of Actuaries or to present at a seminar, not to include in a confidential patent application. In his blog, actuary Kevin Pledge commented on the Towers Watson patents:
The real issue is not really whether this practice is sufficiently different from existing practice, but why they would want to patent it at all.
Valid actuarial practices are normally written up in peer reviewed journals and modified and improved over time as they are they are adopted by the community. If this technique is different it will be branch off from existing and developing actuarial practice. If your work is valid and your goal is to get credit for this work it is far better to publish it in a journal in form that can be read by a peer group rather than by a lawyer in patent language.
The primary reason for patenting something is usually to stop others from using it or to collect royalty from those who do. It is possible that we will see law suits against insurance companies and smaller consulting firms that are not using software from Towers Watson. Defending such law suits will rely on deciphering the patent description and explaining actuarial practice to a jury. Smaller companies in particular will find it hard to defend against such law suits despite the fact that statistical sampling is an established practice.
The point Pledge raises is a valid policy issue. When should the advancement of knowledge in a science such as statistics or actuarial science be patentable, and when should it be part of the increase in technical knowledge available to all in the field? It’s a policy debate which courts and Congress struggle with from time to time. For now, actuaries should be aware of the possibility that patent law protection might apply to “inventions” they develop for better ways to calculate actuarial outcomes.
Disclaimer: I work at a law firm which represents actuaries and which has a large patent and intellectual property law practice. We are clearly interested in being able to apply for patents for the inventions our clients develop.