Latest Research

---

Economic Variable Model Available on CAS Web Site
by Phil Heckman, Past Chairperson, Committee on Theory of Risk

An economic time series model is now available in the Research Section of the CAS Web Site. The model was created by CAS members Steve D'Arcy and Rick Gorvett of the University of Illinois, and Kevin Ahlgrim of Illinois State University, under the joint sponsorship of the CAS Committee on Theory of Risk and the CAS Dynamic Risk Modeling Committee and the Society of Actuaries. This model generates multiple financial scenarios, incorporating interrelations among interest rates, inflation, and other financial and economic variables. The model has applications for dynamic financial analysis (DFA), cash flow testing, investment analysis, and financial planning.

The model is written in Microsoft Excel using Visual Basic and the Palisade Corp.'s @Risk add-on. Users can run the program with the base parameters already installed, or change these parameters to reflect the judgment of the user. The time series it can simulate are

• Long Term Interest Rates and Yield Curve
• Short Term Interest Rates and Yield Curve
• Inflation Rates
• Stock Market Price Levels—Large Capitalization Stocks
• Stock Market Price Levels—Small Capitalization Stocks
• Equity Dividend Yields
• Real Estate Price Levels
• Unemployment Rate

The model documents include an Excel workbook containing the model itself and another with a large number of economic scenarios generated by the model for those users who do not have the @Risk add-in. Also posted on the Web site is an extensive report and bibliography with thorough discussion of the research team's modeling decisions and a comparison of the output of the model with historical values. Before undertaking the modeling itself, the research team conducted an extensive literature review. The report includes discussion of key contributions to the literature.

The original impetus for the model came from Chuck Emma, then chair of the Dynamic Risk Modeling Committee (formerly DFA Committee). Emma saw the need for a basic simulation model, for economic time series commonly used in DFA, which would incorporate the correlations and dependences among these series and provide a modeling platform that reliably produces reasonable and realistic economic scenarios. In my capacity as the chair of the Committee on Theory of Risk, I led oversight of the project jointly with Steve Segal, director of research for the SOA.

The model is intended to support a basic level of professional quality in DFA modeling and should be of interest to any practitioner, even those with more sophisticated models at their disposal. Nonpractitioners will find the report well written and an efficient way to become informed on economic modeling issues.

The model is publicly available, and all interested parties are welcome to download and try out the model, perhaps in a live DFA analysis. The authors presented their work at the recent Enterprise Risk Management Symposium, which was held in Chicago and cosponsored by CAS, SOA, the Professional Risk Managers' International Association, and Georgia State University. The authors invite your questions and comments.