impacted by the use of different models. For example, for a pass through, use of a normal model would show shorter effective durations than would a lognormal model, by about 0.5 year.

As related to the above discussion on volatility assumptions, it is important to understand that different interest rate models use different forms of volatility assumptions. Whereas, when using the Black-Karasinski model, one might use a relative volatility assumption like 30% and mean reversion of 3%; in using the Hull-White Model, one would use an absolute volatility assumption of 139 basis points and 3.8% mean reversion. Consistency across assumptions and analyses is another issue to be considered when reviewing systems and reports.

These are just two examples of assumptions that have been impacted significantly with changes in the market environment. In any given system used for investment or risk management, there are numerous assumptions that must be made and set to generate specific analyses. If those assumptions are not reviewed and updated on a periodic basis, it becomes a matter of that well-known acronym, GIGO = Garbage In, Garbage Out.

This is just one component of what should be a broader oversight initiative/responsibility - periodic reviews of investment technology and its applicability to investment objectives. It is recommended that systems and any related analyses and reports generated be evaluated for currency and relevance on some regular basis.