Cognalysis Reserving System FAQ


Is Cognalysis Reserving System™ spreadsheet-based?

No. The software is a stand-alone Windows application. You can cut and paste from and into spreadsheets, and the analysis files it stores data in are csv files, so it is straightforward to use Cognalysis Reserving System™ in conjunction with spreadsheets.


The benefits of stand-alone software are significant. We added drill down features to take the mystery out of calculations (i.e. double click on a number to see how it was calculated). Specialized graphs focus on key issues in reserving. Features such as selecting factors via drag and drop in loss development factor graphs help to streamline the process.


What are the data import options?

Import data can currently be in the form of a flat file (.csv or .txt) or a Microsoft Access database. Data can be pre-aggregated or transactional in nature. The data layout is flexible, with wizards to guide the user through the process of describing new data layouts to the software. Once a particular layout has been used, it can be automatically recognized by the software for future imports.


What sort of reporting options are there?

The software includes a standard set of reports that the user can select from. These reports can include customized headers and footers for the date, page number, etc. The user can also decide whether or not to include his/her notes on relevant exhibits.


How are reserve ranges determined by Cognalysis Reserving System™?

Cognalysis Reserving System™ measures the variability of actual development from projected development using selected loss development factors and seed loss ratios. It uses this observed variability, including effects of correlation across development periods and accident periods, to develop estimates of future loss payment variability. Correlation between analysis lines is also included.


How are the Bornhuetter-Ferguson seed loss ratios determined by Cognalysis Reserving System™?

The seed loss ratios are ultimately selected by the user. The software does generate defaultvalues based on multi-period averages of loss ratios from link-ratio indications. In addition to being able to select any loss ratios the user feels appropriate, the user can also change the number of periods that are used in calculating these multi-period averages, to be more responsive or less responsive to changes in the loss ratio over time.


What is the Minimum Variance (Optimized) Weighted Average Estimate?

In addition to the standard paid and incurred link ratio and B-F indications, the Cognalysis Reserving System™ provides a weighted average indication that blends these standard estimates together based on the relative predictive accuracy of the different indications, which varies from analysis to analysis. While the user still may select any estimate he or she feels is appropriate, this weighted average indication, as well as the weights used to develop it, can provide valuable information to the actuary.


What is the regression diagnostic, and how is it used?

Most of the functionality within the Cognalysis Reserving System™ is built around standard actuarial techniques, but with additional techniques to aid the actuaries.


The regression diagnostic, on the other hand, represents a more radical departure from the standard techniques. It poses the reserving question as one of estimating future incremental cash flows based on a number of predictive variables — premium, case reserves, historical payments, and the time period. From among the wide range of potential multivariate models, the regression diagnostic selects the one that generates the smallest predictive error for the incremental payment in question. The incremental payment being projected from the standard techniques is then compared to the regression prediction range to see if it is inconsistent. Investigating cases where such inconsistencies are occurring can be useful to suggest alternative models of development to the user, when standard techniques may not be appropriate for a particular analysis.