In this series, we apply the magnifying glass to how the standard formulae for selected SCR sub-modules were calibrated. We investigate the history behind the calibration, the risks that were excluded from the calibration, and potential shortcomings as a result.
In PART I we covered mortality risk1, PART II considers retrenchment risk. Look out for future articles in this series on other sub-modules.
Currently, retrenchment risk is calculated by permanently increasing retrenchment rates by 50% relative to best estimate assumptions. This calibration was however done at a very high and pragmatic level. It is suggested that insurers with significant exposure to retrenchment risk, should evaluate whether the standard formula gives an adequate estimate of retrenchment risk. This can be done using internal models as part of the ORSA process. Some limitations of the current calibration are that it neither reflects the cyclical nature of retrenchment, nor does it make allowance for catastrophic events.
Background - calibration:
After the first South African Quantitative Impact Study (SA QIS1), a need was identified to have a separate sub-module for Retrenchment risk within the Solvency Capital Requirement (SCR). As a result, a retrenchment rate shock equal to a permanent 50% increase in retrenchment inception rates relative to best estimate assumptions, was included from SA QIS2 onwards.
The calibration for this stress was done at a very high and pragmatic level. This was given the relatively small exposure in the SA insurance industry to Retrenchment risk overall, and the relatively small part of the industry risk capital represented by the retrenchment risk sub-module. At the time of the SA QIS 2 report, retrenchment risk represented less than 5% of the SCR for the insurers who completed the retrenchment risk calculation. In addition, not a lot of statistically credible data on South African retrenchment claims was available at the time of calibration.
Although retrenchment risk is not one of the larger risks across the industry as a whole, it may be significant for some insurers. However, the QIS2 and QIS3 task groups recommended retaining Retrenchment risk as a separate sub-module. It is included in the SCR standard formula at the original calibration of a permanent 50% increase in retrenchment inception rates relative to best estimate assumptions. Importantly, the Solvency Asset and Management (SAM) Steering Committee task group noted that those insurers for whom retrenchment is a significant risk should address this via a partial internal model or the Own Risk and Solvency Assessment (ORSA).
As part of its SAM Phase II subcommittee work, the Actuarial Society of South Africa is currently focusing on retrenchment risk. The existence of a task group dedicated to this, indicates the need to relook the original calibration.
Unpacking the potential shortcomings:
The current calibration is limited in that it does not reflect the cyclical nature of retrenchment and does not allow for catastrophic events (resulting from a concentration of risk). These limitations could result in retrenchment risk being understated within the SCR.
Concentration & Catastrophe
The current SCR for retrenchment risk captures the risk of long-term retrenchment assumptions being too low, but it does not capture a short-term catastrophe (‘CAT’) type event (similar to the CAT risk that is calculated for mortality and morbidity risks). This is particularly relevant to products with shorter contract boundaries.
Allowing for CAT risk is important because retrenchment events tend to be catastrophe-like – every few years retrenchment rates spike for some reason or another. These increases may be concentrated in some areas (geographical or industry) more than others, leaving some insurers more vulnerable than others.
Consider the COVID-19 pandemic as an example. While a lot of attention was on mortality risk, we have seen short-term retrenchment rates increase substantially for a number of our clients. The changing landscape from ESG initiatives may also result in industry specific retrenchments. This suggests the need for a retrenchment CAT risk component in the standard formula, to capture these ad hoc, possibly short term, increases in retrenchment risk going forward.
The current calculation does not reflect the cyclicality of retrenchment risk. Following QIS2, some insurers were of the opinion that the 50% increase is not onerous enough. The view was that the increase might well be higher, but for a shorter period only (i.e. not a permanent increase as is currently being assumed in the retrenchment risk calculation).
Conclusion: The current standard formula may underestimate retrenchment risk. This is because no allowance is made for catastrophe risk or the cyclical nature of retrenchments. These limitations may have substantial short-term impacts. Insurers that are more exposed to retrenchment risks should consider evaluating its retrenchment risk via internal modelling and include a consideration for this in the ORSA.
Reference & further reading: This article uses information from:
SAM steering committee Position Paper 108 - Life SCR - Retrenchment Risk (fsca.co.za)
SAM Report on the results from the 2nd South African Quantitative Impact Study - SA QIS2 (fsca.co.za)
Get in touch:
For assistance with evaluating the appropriateness of the standard formula and impact on your specific business, or for developing alternative internal capital stresses, reach out to us.
1 PART I of the series can be assessed here: Mortality Risk Calibration, Shortcomings & Risks Not Considered.