A recent article in Risk Magazine states that insurance companies are struggling to move to OIS discounting for their books and records for the following reasons:

1. There's too much variability in the discounting methodologies being used across different counterparties and different asset classes to switch to OIS.
2. Switching to a new discounting methodology will cause P&Ls to jump, thereby prompting questions, so some insurers prefer to stick with the pre-2008/9 status quo methodology.
3. Switching  to a new discounting methodology will require major changes to back-office processing systems, which will come at a significant cost.

Deconstructing these reasons one by one: (1) The insurers have a valid point. Multi-currency CSAs, and some general market inertia to switch to OIS discounting for anything other than IR derivatives, do cause valuations discrepancies between counterparties. However, as the mandatory central clearing of vanilla IR derivatives continues, and CSA currency silos gain wider adoption for the remaining bilateral derivatives portfolios, many of these valuations discrepancies will be ironed out. (2) The impact of a sudden P&L jump can be considerably softened by running historical back tests of one's portfolio NPV to compare LIBOR discounting with OIS discounting. The results of these back-tests can then be used to reasonably estimate the impact of switching to a new discounting methodology. This information can be used to diffuse any questions around a sudden jump in portfolio MV before-the-fact. (3) If today's knotty valuation problems are resolved in the manner explained in point (1), then firms will have no choice but to update their back-office systems and processes to reflect the prevailing market reality. Ben L.