April 2014

University Programs for Academic/Practitioner Collaboration—Part One

By Jay Vadiveloo and Ian Duncan

At the recent Society of Actuaries Annual Meeting in San Diego, I was part of a panel that discussed various models currently in place to foster academic/industry collaboration. I participated in my role as director of the Goldenson Center for Actuarial Research at the University of Connecticut (UConn). The Goldenson Center was formally established in 2009 with a one million dollar gift from Janet & Mark L. Goldenson. The donors were very clear about the purpose of this gift and it is reflected in the mission statement of the Goldenson Center: to undertake applied research projects to serve the needs of industry in the region. But it is a big step between coming up with a mission statement and actually implementing it. I believe over the past five years, the Goldenson Center has managed to remain faithful to its mission and is a good example of successful academic/industry collaboration. In this article, I hope to share what it has taken to accomplish this mission and what I believe are key stumbling blocks to avoid, which could jeopardize this effort.

First, some facts about the Goldenson Center structure. The annual income from the endowment serves as the operating budget for the Goldenson Center and it is primarily used to support students (and faculty, if necessary) on research projects. As director of the Goldenson Center, I have a faculty position at UConn with teaching responsibilities for one class each semester, as well as the responsibility for supervising Ph.D. students in actuarial science. However, my main responsibility is generating and managing projects for the Goldenson Center.

This structure may be similar to other university models. However, where the Goldenson Center is very different, is in the active participation of Towers Watson, a global professional services company. Towers Watson’s involvement with the Goldenson Center is reflected in several ways:

  • Towers Watson funds my position as director of the Goldenson Center and I also serve as a senior global research consultant for Towers Watson
  • Towers Watson provides client and research projects for students to work on, as well as supervisory and project management support for Goldenson Center projects
  • Towers Watson provides students working on client projects with proprietary actuarial software and training at no charge to the Goldenson Center

I believe Towers Watson’s active support of the Goldenson Center is one of the main reasons for its success. The funding of the director’s position by Towers Watson allows all earnings from the endowment to fully support students. In addition, the project management and modeling assistance on client research projects provide credibility and a level of professionalism to the research work done at the Goldenson Center.

There are other reasons for the success of the Goldenson Center model:

  • The Goldenson Center, since its inception, has established an Advisory Board of industry representatives representing most of the major insurance companies in the region. The Advisory Board plays a key role in ensuring that the Goldenson Center stays true to its mission, and companies represented on the Advisory Board are the main source of projects for the Goldenson Center. The Advisory Board meets once a year to hear presentations by students on projects they have worked on in the prior year, and to suggest new projects for the coming year.
  • With Towers Watson’s oversight and supervision, the Goldenson Center has achieved a strong track record of projects being delivered on time and meeting the needs of clients, as well as maintaining high standards of academic rigor and professional excellence
  • As director, I have connected with industry to market the Goldenson Center and generate an ongoing stream of new projects, as well as take ultimate responsibility for the successful execution of these projects. This requires time and commitment and is necessary to ensure that the Goldenson Center continues to be supported by industry

It is important to explore further the Goldenson Center’s philosophy of applied research, which has shaped both the choice of research projects undertaken by the Goldenson Center as well as how the research work is done. The research projects are not generated from UConn and then adapted to suit the needs of industry. The projects come directly from industry and are formulated during one-on-one meetings between the director and Advisory Board members. The project deliverables are clearly laid out, as well as costs, and timelines, with the intent that completed research projects can be immediately applied to benefit the client. Many research projects naturally have some academic or theoretical underpinnings and while that is a nice feature to have, it is not a criterion for selecting a research project.

The one advantage to industry of doing collaborative research with academia is the ability to focus on both short-term and long-term applied research needs. While most Goldenson Center projects are designed to be completed within the academic year for obvious practical reasons, we have undertaken projects that have extended beyond a year. These projects involve a changing batch of student resources, and it is important that knowledge and models are seamlessly transferred between different batches of students to ensure continuity in the projects. Many longer term Goldenson Center research projects evolve into Ph.D. research topics for doctoral students in actuarial science and they help to maintain the continuity of these projects.

One potential stumbling block that could jeopardize any academic-industry collaboration is the whole issue of data security and confidentiality. It is important that this issue is addressed right from the start when discussing client projects. At the Goldenson Center, we have several ways to ensure the confidentiality of client data:

  1. All client data resides in a dedicated virtual server established at UConn for the Goldenson Center. Only students selected for the project are allowed access to the data using strict password protocols, and the data cannot be loaded to a student’s personal computer or any external devices. A similar arrangement is available using Towers Watson’s servers.
  2. All students working on client data have to sign a non-disclosure agreement
  3. Any personal information which is not needed for the project is removed by the client before being provided to the Goldenson Center
  4. Any additional research work using client data after the project is completed requires sign-off by the client

While the Goldenson Center to date has not violated any data confidentiality requirements, I am well aware that it takes just one inadvertent mishap to completely destroy any future academic-industry collaboration. From UConn’s point of view, I have to ensure that Goldenson Center projects do not violate any student visa restrictions that international students have to observe. While this is largely a UConn issue, any negative publicity arising from Immigration & Naturalization Services (INS) violations would impact the strong and trusting relationship the Goldenson Center shares with its board members. As a rule, on any issue which is not straightforward, it is best to bring it up to the attention of experts at your university and the client and get it resolved before the project commences.

So what are some typical Goldenson Center projects that we have undertaken over the past five years? The first two research projects are long-term in nature while the last two sets of projects have generally been completed within a semester.

  • Replicated Stratified Sampling (RSS) modeling algorithm which uses statistical sampling techniques to dramatically speed up run times for complex actuarial modeling. The RSS technique is the subject of the research of a recent Ph.D. actuarial student at UConn.
  • The National Retirement Sustainability Index (NRSI) is an annual retirement preparedness index for the United States which captures both economic and non-economic factors impacting retirement readiness. The non-economic factors include state of health at retirement, level of adaptability at retirement, level of job satisfaction and impact on age at retirement, and impact of pre and post retirement financial planning
  • A series of Goldenson Center funded risk management projects for small businesses. This has led to a text on this subject that will be published by the Society of Actuaries in 2014. I will be the Editor of this text with articles contributed by academia and risk management professionals
  • Life predictive modeling projects for long term care, long term disability and for modeling the mortality risk for pension liabilities.

While the Goldenson Center model has been successful, it is not necessary for every successful academic-industry collaborative effort to follow this design. However, any successful model has to incorporate at least the following:

  • A faculty member dedicated to managing this effort either on a full time or least 50% basis with clear and measurable deliverables
  • An actively participating Advisory Board of industry representatives
  • Some form of committed annual funding, preferably by a consortium of companies, to support students on any exploratory applied research projects. This is in addition to fee-for-service client projects involving student and faculty resources
  • Commitment to the highest standards of excellence on any project including meeting client deadlines and data confidentiality needs.

No successful and sustainable academic-industry collaboration for actuarial science research can be created without a commitment to this effort. Business departments at universities have historically been very successful in doing this. However, this is a new experience for actuarial science programs which are often housed in the department of mathematics. Traditionally, mathematics departments do not collaborate with industry. Hence, it may be helpful to speak to other departments at the university or possibly partner with them when embarking on this initiative. While it takes effort and dedication, I believe that academic-industry collaboration, if managed to meet the needs of industry, is critical for the growth and development of the actuarial profession. Plus most importantly, students who have real life experience working on projects with industry while they are in school, become far more marketable and valuable to industry. For me personally, based on feedback from students, clients and Advisory Board members, managing the Goldenson Center has been a truly rewarding and fulfilling experience.

For more information on the Goldenson Center, please visit www.goldensoncenter.uconn.edu.

Jay Vadiveloo, Ph.D, FSA, MAAA, CFA, is professor-in-residence and director, Janet & Mark L. Goldenson Center for Actuarial Research at the University of Connecticut. He is also a senior research consultant at Towers Watson in Weatogue, Conn. He can be reached at jay.vadiveloo@towerswatson.com.

Ian Duncan FSA, FIA, FCIA, MAAA, is adjunct professor of actuarial statistics in the department of statistics & applied probability at the University of California-Santa Barbara in Santa Barbara, Calif. and vice president, clinical outcomes & reporting at the Walgreen Co. in Deerfield, Ill. He can be reached at duncan@pstat.ucsb.edu.