SYSTEMS AND METHODS FOR USING DATA MART TECHNOLOGY TO MORE ACCURATELY PREDICT PROJECTED PREMIUM TRENDS FOR AUTOMOBILE INSURANCE
Publication Date: 2008-Jan-03
The IP.com Prior Art Database
The projected premium trend is a key component of many actuarial analyses. It is also one of the least studied--only a few methods are used in practice today. However, this structural analysis approach, when combined with knowledge of a company’s operational plan, can be a very powerful tool in improving the accuracy of a projected premium trend estimate. As disclosed herein, the data demands for a large scale study can be met by fully utilizing database and data mart technology. Results from such a large scale study in turn allow an insurer to make informed pricing and business decisions.
SYSTEMS AND METHODS FOR USING DATA MART TECHNOLOGY TO MORE ACCURATELY PREDICT
PROJECTED PREMIUM TRENDS FOR AUTOMOBILE INSURANCE
An integral part of the typical actuarial rate level review for automobile insurance is the premium trend analysis. This step in the actuarial process quantifies the current and projected changes in an insurer¶s premium volume that can be attributed to factors other than policy growth and rate changes. The current premium trend is fairly easily determined by analyzing the change over the historical experience period in the average written or earned premium after it has been adjusted to the appropriate rate level. The projected trend estimates premium changes between the time of the review and the proposed effective date of new rates. This projected trend estimate is not as simple to determine because it includes many assumptions and estimates of future unknowns.
The most common approach to estimating projected premium trend is to fit the historical data to a trend line. There are two flaws to this approach. The first is that history is not always the best indicator of the future. The second flaw is that there is no explanatory power to this estimate. In other words, it is difficult to attribute specific causes to support the rationale for the selected projected trend.
In order to improve both the accuracy and the explanative power of a projected premium trend estimate, a collection of several data cubes can be used to both calculate dynamically and display average rating factors across coverage and over several historical months. Accuracy is improved because a more detailed structural analysis is
possible using a collection of data cubes. Explanatory power is also increased and made more specific because assumptions behind changes in specific rating variables can now be attributed to the premium trend selections.
Improvements in Projected Premium Trend Selection
Premium trend analysis is the analysis of distributional changes that impact the overall rate level. Changes in the distribution of deductibles, increased limits, vehicle symbol, and model year are all examples of variables that impact premium trend. The current method of premium trend analysis studies these changes in the aggregate and then projects them into the future. One-time changes, insurer operational changes, or other external factors that are part of the historical data and that, for purposes of accuracy, would best be left out as determinative factors or variables are often erroneously included in a calculation of a projected premium trend forecast. Another consideration is that other changes may be expected to occur in the future but these are not reflected in the current record.
One way more accurately to capture what should be included in a future projection...