Global capital standards

The International Association of Insurance Supervisors (IAIS) has been working for a number of years on the first global, risk-sensitive solvency regime for insurance groups. BaFin is collaborating on the project and is working hard to achieve compatibility with the European Solvency II¹ framework.

In summer 2017, the IAIS published ICS 1.0, the first version of the Insurance Capital Standard², for extended field testing³. This global capital standard is intended to apply to all large insurance groups with international operations (Internationally Active Insurance Groups – IAIGs). The field test is intended to try out different options for the various elements of the ICS, including discounting and the criteria for own funds. The findings of the test will be fed into the ongoing development of the ICS from version 1.0 to version 2.0, which should then no longer include any options.

In the next step, the IAIS has begun to analyse the results of the field test which started in May 2017, with the participation of German undertakings. Following intensive discussions in which BaFin was also involved, the IAIS announced in November how the process of transitioning ICS 2.0 from the development phase to the implementation phase from 2020 will function.⁴

Next steps

From 2020 onward, the intention is that, during an initial five-year monitoring period, ICS 2.0 will be reported using only market-adjusted valuation. The results that will then be available will provide further information to be used in assessing the solvency of IAIGs, but will not be used as a basis to trigger supervisory action during the monitoring period. The data obtained in this way will also be able to contribute to improved comparability in the supervisory colleges. After completion of the monitoring phase, ICS 2.0 is planned to be introduced as a prescribed capital requirement (PCR) worldwide according to the wishes of the IAIS and its members.

BaFin voices support for internal models

An important concern for BaFin is that, in addition to the standardised method which all undertakings will be required to apply during the monitoring phase, internal models should also be authorised as an equal basis for the purpose of calculating the ICS. During the monitoring period, internal models are intended to be used only on a supplementary basis. It is planned to review their use as a permitted alternative to the standardised method at the latest by the completion of the monitoring phase. The IAIS is currently working on defining the criteria for the application of internal models. This process is scheduled to be completed by the next round of the field test in 2018. Other IAIS members prefer a method based on adjusted accounting standards (GAAP+)⁵, which will also have to satisfy a review after five years. The agreed plan also envisages bringing the different concepts together without delay and defining them in detail, as well as reducing the options in the field test.

One of the issues concerns the methodology used to determine the yield curve for the purpose of discounting the liabilities. At the same time, agreement is needed on a comprehensive method of calculating the risk-free yield curve. The IAIS members must also agree on criteria for allocating the available own funds to different tiers in order that a solvency ratio can be determined. Other unresolved points are the calculation of individual elements of the capital requirements and the calculation of the risk margin.

Identification of G-SIIs und ABA

The Financial Stability Board (FSB) did not designate any global systemically important insurers (G-SIIs) in 2017.⁶ However, the supervisory measures for G-SIIs remain in force for the nine insurance groups identified in 2016 as before. The FSB stressed that new discussions about designation were planned for 2018, including in the light of progress made on work towards an approach based on the undertakings' different activities (activities based approach – ABA).

Direct and indirect systemic risk

The ABA for the purpose of assessing systemic risk in the insurance industry is currently being developed as a supplement to the approach focused on undertakings. BaFin is involved in the development of the ABA. In the undertakings-based approach, the central question is whether the (uncontrolled) insolvency of an undertaking could have a direct negative impact on the financial system (domino effect).⁷ The ABA approaches systemic risk from a different point of view. In this case, the critical issue is whether particular activities, undertaken collectively by a number of actors, have the potential to affect the financial markets negatively (tsunami effect). In BaFin's view, this distinction between direct and indirect systemic risk is very important, and both could be combined in a hybrid approach.⁸ A hybrid procedure of this kind, which BaFin supports, brings together both perspectives, resulting in a comprehensive monitoring approach – and therefore also a reduction in systemic risk.

The IAIS stepped up work on the ABA in 2017 and published a first consultation paper⁹ in December. In the paper, the IAIS sets out a potential framework and a possible scope for an ABA. The framework is based on the IAIS's work to date on systemic risk. In addition to identifying activities posing a potential systemic risk – work that is being carried on in 2018 –, the IAIS intends to match the available supervisory tools against the identified risks of these activities. If this process reveals gaps in the IAIS toolbox, those gaps are planned to be closed.

Group of undertakings covered

While only about 50 undertakings were included in the data query for the purpose of the previous identification of G-SIIs, this group should be expanded for an ABA, in the opinion of BaFin. Only then will it be possible to evaluate any possible negative effects of particular activities. BaFin is convinced that the principle of proportionality embedded in Solvency II should also be observed for the purposes of an ABA. This applied both to microeconomic and macroeconomic risks.

Footnotes: