The UK’s Prudential Regulation Authority has published a brief Supervisory Statement, “SS15/16 Solvency II: Monitoring model drift and standard formula SCR reporting for firms with an approved internal model“.

The PRA has long been concerned that internal model quality and output-reliability will gradually fall, if (re)insurers aren’t encouraged to maintain their models with the same rigor and enthusiasm they brought to the model design and build process; and that some (re)insurers will try to “game the system”, in an attempt to inappropriately lower their capital requirements, if the PRA doesn’t watch them carefully enough.  The new Supervisory Statement has been designed with these concerns in mind. In particular, it:

• sets out the PRA’s “expectations“, and “provides further information on the PRA’s approach to monitoring model drift and the reporting of standard formula … SCR information“;
• explains that “the PRA’s approach includes the monitoring of the internal model against certain objective measures of risk” – for example, the standard formula SCR, the pre-corridor MCR, net written premium and best estimate liabilities – and that these “measurements” will be taken (a) when a model is approved; and (b) when it’s re-based after a change in the firm’s risk profile, or a model change that generates a material SCR change as well; and
• (in practical terms, effectively) requires firms to:
  • calculate their standard formula SCR, as part of their risk management, ORSA, and narrative reporting and model validation cycle(s); and
  • (if they’re using a model to calculate their solo SCR), privately report the results to the PRA within 4 weeks’ of the deadline for submitting their annual quantitative reporting templates. (On the plus side, (re)insurers aren’t required to have their standard formula calculation externally audited, although it must still be approved by an appropriate member of the senior management team.)

The PRA has produced and annexed 2 reporting templates to the Supervisory Statement. They’re also available on the regulatory reporting page of the Bank of England’s website,  under the heading “Standard formula SCR reporting templates for firms with an approved internal model“.

There is at least a thread of forbidden “gold-plating” here, although the PRA would almost certainly deny this, by arguing (for example) that:

1. Article 112(7) of the Solvency II Directive gives the PRA the power, “by means of a decision stating the reasons“, to require an internal model firm to provide it with an estimate of the SCR calculated using the standard formula; and
2. Article 118 of the Directive gives the PRA the power to require a (re)insurer to revert to using the standard formula if (a) the firm and its model no longer meet the “use”, “statistical quality”, “calibration standards”, “validation standards”, and other relevant tests; and (b) the firm fails to implement a plan to rectify these faults timeously enough;
3. So internal model firms are effectively obliged to retain their ability to calculate their standard formula SCR; and
4. The PRA is merely pointing out that firms “may” find it helpful to calculate their standard formula SCR as part of their risk management, ORSA, narrative reporting and model validation cycle(s), and it’s merely “expecting” them to privately report the results, rather than making a rule that will always require them to do so.

Talk to professionals in the market and you’ll have a strong sense that, whilst the PRA might not be “gold-plating” in a technical sense, that’s exactly what it’s doing here, and elsewhere.

[View source.]