Banking StandardsLetter from Andrew Bailey, Managing Director, Prudential Business Unit, Financial Services Authority

Thank you for your letter setting out concerns on the implementation of Solvency II (SII). I should say that I am pleased this issue has come to the surface because, while it is not a banking issue, it is a major point that we are dealing with as we move from the FSA to the PRA. I would like to deal with two charges that are levelled at the implementation of SII, namely: first, that it is lost in detail and vastly expensive; and second, drawing on the lessons from HBOS and the implementation of Basel II, that it is dangerously distracting from the application of prudential supervision today. In these points, I would like to set out some background on SII.

SII introduces a new, EU-wide, insurance regulatory regime that replaces previous EU regimes. The stated aim of SII is to strengthen the prudential regulation of the insurance sector and ensure policyholder protection. It does this by setting out strengthened requirements on capital adequacy and risk management. It introduces requirements to value assets and liabilities at market value and risk-based capital requirements based on internal models.

The EU introduced Solvency I as a “minimum harmonizing” directive in 2002, imposing minimum standards but allowing individual member states to impose higher standards if they wished. In practice this has proved a relatively weak constraint and national supervisors within the EU have pursued quite different approaches.

The UK introduced its Individual Capital Assessment (ICAS) regime for insurers in 2004. The key features of this regime are the requirement to value assets and liabilities at close to market value, and a risk-based capital requirement that relies upon some internal models. SII is therefore an evolution of the current UK framework, but it represents a major change for Germany, France and Italy, where support for SII appears to have diminished.

In particular, it is unclear to us that the French authorities will now be able to agree to any directive that we consider prudentially acceptable. Germany in contrast has long been supportive of a relatively prudent agreement on SII but needs a long transitional arrangement (eg 10 years or more) to allow its domestic industry to adjust. A particular issue for German industry is a large “back book” of guarantees, many issued in conjunction with the Government, which have been rendered difficult to support by the low interest-rate environment.

These concerns mean that, as has been the case on many previous EU directives, SII has been substantially delayed. It is, however, politically highly unlikely that the EU would operate long term without harmonised standards for insurance so the issue is really what sort of Directive is eventually agreed. The FSA’s longstanding approach to SII has therefore been to assume that a final version of the Directive will eventually be agreed, and to seek to maximise influence on the policy content.

The process to finalise SII in the EU has ground to a halt in the face of these different national interests. I have three major concerns with what could emerge eventually. First, SII is a “maximum harmonised” directive (as will be large parts of the Basel 3 implementation for banks), so there will be a more limited ability for national supervisory authorities to impose sensible treatments to deal with more idiosyncratic risks. There is a risk that this whole area of maximum harmonisation will be a battleground of the future as the judgmented approach of the PRA comes up against narrow interpretations of EU law.

Second, SII envisages that firms may calculate their risk sensitive capital requirements by using a bespoke internal mode. This is an option that most of the larger firms in the UK are taking up. There are good reasons for this because the UK has two large sections of its insurance industry whose business does not fit well into the standardised approach of SII, namely the London Market including Lloyd’s with its global focus, and the With Profits business of life insurers. That said, as with banks, there is a risk of over-reliance on complex models and of models being used to pare capital requirements. Moreover, there is a risk that SII overloads supervisors with very detailed model approval requirements. Such an approach is not consistent with the PRA’s proposed approach to supervision. To mitigate this risk, we plan to use “early warning indicators” in our supervisory work. These are designed to be simple measures which are exogenous to the model and alert supervisors as to potential threats to solvency from models being used to pare capital requirements, triggering immediate supervisory action. The analogy here is with the Basel leverage backstop. We believe we can implement these Early Warning Indicators in the UK within the SII framework but in any event we would pursue this approach and accept the risk of EU challenge.

Third, SII is designed to set standards for the equivalence (to SII) of insurance supervision in the rest of the world. If the EU determines the insurance regime of a country outside the EU to be not equivalent to SII, EU firms operating in those jurisdictions will be required to hold additional capital, potentially making them uncompetitive in those markets. This is a big issue, and one where we will need if necessary to resist narrow interpretations, bearing in mind the UK’s large presence in the global insurance industry.

Since there is no timetable for the implementation of SII, I have been very concerned about the implications for costs, which have been, frankly, staggering. In my view it is not wise to take an incremental view of the slippage in the timetable and go on spending money on implementation as if the process was moving when in fact it is not the case. As a consequence, we have told firms that we do not expect implementation until at least 2016, and we have to plan as such rather than let the thing slip in small increments. We have scaled back plans for future work to implement SII. We rejected the option of stopping project work altogether; that would not be the preference of either the firms or us. We are all of the view that there are some useful elements of the work done to date, and we want to bring them into effect. The work to date will therefore be used where possible to introduce an enhanced domestic insurance capital adequacy standard.

The scaling back of plans has resulted in a substantial future cost-saving for the industry which, in addition to funding its own internal model development, was being levied to pay for the FSA’s preparatory work to approve those models. In 2009 the FSA Board approved a total spend by the FSA on preparation for the implementation of SII (primarily approving SII models) of between £100 and £150mn. This was to be financed by a levy on the insurance industry. The total expenditure set against the special levy to fund SII is currently just over £63mn. Our re-plan has reduced the estimate of total expenditure to £88mn. This has been warmly received by the Chairmen and CEOs of the major insurance companies.

I have to say that I find the history of the EU process on SII shocking. I have attached a copy of a lecture that I gave last week in which I made this point. It would be a help if Parliament could cast light on a process which has gone on for the best part of ten years, and in which the EU process has assumed that firms and regulators will spend very large amounts of money to prepare for something that carries no promise in terms of when, or in what form, it will be implemented. We have taken action since last summer to deal with this as best we can, and I think the outcome is sensible and pragmatic, but it does not cancel the nature and scale of the issue that the PRA inherits caused by an EU process that makes no allowance for value for money.

Apologies for the length of the answer, but I hope this casts light on the first concern, namely that SII is lost in detail and vastly expensive. The obvious answer is yes to both of these charges. We are putting in place an approach which I hope will mitigate these risks, but it does not I admit deal with the root cause, which lies in the EU process.

The second charge is that SII preparations have crowded out prudential supervision of insurers, the HBOS risk. Ironically, the cost of SII preparation points against this risk, because the FSA in effect set up a second supervision operation while continuing to apply the ICAS regime. We will have to watch this in the future as we seek to economise on spending, but I am content at present that we are alert to this risk, and have scaled the SII work back rather than have it compete with current supervision.

I hope this rather long explanation helps.

14 February 2013

Prepared 19th June 2013