Institutional Investment - July 20, 2007

Interview with Samuel Sender

Samuel Sender, research associate with the EDHEC Risk and Asset Management Research Centre and ALM expert and senior insurance consultant with Finalyse, is the co-author of a major study drawn up by EDHEC entitled The Impact of IFRS and Solvency II on Asset-Liability Management and Asset Management in Insurance Companies. He also co-wrote, with Philippe Foulquier, director of the EDHEC Financial Analysis and Accounting Research Centre, three position papers in response to the consultations and quantitative impact studies carried out by CEIOPS (the Committee of European Insurance and Occupational Pensions Supervisors) in order to prepare the Solvency II directive: QIS 2: Modelling that is at odds with the prudential objectives of Solvency II; CP20: Significant improvements in the Solvency II framework but grave incoherencies remain; and QIS3: meaningful progress towards the implementation of Solvency II, but ground remains to be covered.

Samuel Sender


The European Commission adopted the Solvency II framework directive on July 11th. This directive sets out all the principles for solvency capital requirements and supervision of insurance and reinsurance undertakings, but leaves all implementation measures and detailed specifications for later update. It will now be examined by the European parliament and will be incorporated into EU legislation by amendment of existing directives governing this area. We are hopeful that this will occur by 2009 so as to ensure progress on the standard formula for capital requirements as it is intended to be implemented.

EDHEC, in its reports, has been positive about the Solvency II project as a whole, but has criticised many of the details relating to its implementation. Before we go on to discuss those details, could you tell us why the Solvency II project itself is useful?

Samuel Sender: Solvency II is a major advance from Solvency I. Under Solvency I capital was absolutely not related to risk, so much so that in some cases companies could appear stronger from a regulatory standpoint while they were actually becoming economically bankrupt. Solvency II has the objective of linking regulatory capital to risk capital – a huge improvement from a philosophical perspective.

In today’s world, only financial companies have to hold regulatory capital to protect customers – on the other hand most companies, such as carmakers, are not regulated. Because regulations such as Solvency II are made to protect customers, they must take a modern view on risks. From a quantitative standpoint, it is necessary to hold some capital against risks - a good regulation is one that imposes capital to be as close as best practices would impose, i.e. that a defined level of protection for clients is reached. From a qualitative standpoint, what really protects businesses from failure is not capital but rather good risk management, so that good regulations must first of all be understood as an incentive to apply best risk management practices.

Poor regulations can be understood as inefficient tax systems. When shareholders are required to hold more capital than necessary to cover risk, they suffer an immediate tax that can be approximated by the cost of capital, i.e. in CEIOP’s own calculations, approximately 6% of the amount of extra capital locked in. On the other hand, when institutions hold less capital than necessary, then policyholders lack protection, which is a form of tax. When best practices are not recognised, then both parties may be penalised.

Management may have the option of avoiding regulations that are inconsistent with best practice by performing regulatory arbitrage. In our work, we have given some examples of how a given business could react to inadequate regulations, by showing how regulatory arbitrage could be implemented. This type of practice is common in the finance industry even though not widely publicised.

Though this technique can be used by sophisticated investors to correct poor regulations, most businesses would rather try to align their risk management practices and internal monitoring systems with regulatory requirements. Regulations thus have a mammoth influence on business practices. For this reason, we have found it extremely important to point out how the upcoming Solvency II directive may be improved.

In your position paper on QIS2, you focused on certain aspects of the modelling suggested by the CEIOPS. You suggested that the choices of concepts, measures and calibrations were sometimes not only hazardous, but above all could lead to strategic management decisions being taken by the insurance companies that were totally at odds with the initial objectives of Solvency II. Has the situation improved since then?

Samuel Sender: The philosophy of Solvency II is as good and modern as can be; on the other hand its application is not always as luminous, and what really matters for regulated entities is what they need to apply.

The principle is to apply full market consistency principles when possible. However, some of the proposed measures and calibrations of the standard formula are inconsistent with these very principles.

Though there has been constant improvement since the first drafts, ground remains to be covered.

You indicated in your paper on CP20 that the lack of acknowledgement of dynamic asset management strategies would incite companies to carry out regulatory arbitrage rather than implement best practices in risk management. Could you explain exactly what you mean by that?

Samuel Sender: In this comment, we wanted first to outline the importance of risk management to insurance companies. Capital is needed as a buffer against risks, but more than capital, what protects a company is good risk management. And the very definition of risk management is the capacity to react to a deterioration in the balance sheet before bankruptcy occurs, which is synonymous with dynamic hedging strategies.

As these practices are not recognised at all by CEIOPS (a CPPI fund is considered to have the same risk as a pure equity fund), companies that are applying modern management techniques are subject to extra tax. One way to escape that undue tax is to perform regulatory arbitrage, i.e. to take actions to reduce regulatory capital requirements without changing the risk profile of the company.

One could argue that these strategies could be captured in internal models, but when and how these will be reviewed and accepted is still an open question. In banking the profit and loss is computed on a daily basis and therefore allows the accuracy of the predictions of internal models to be backtested easily. This is not however the case for the insurance sector where equity and P&L are usually available on a quarterly or even yearly basis, so how this will be done is still to be defined.

This is why it is important in the short run that regulated entities be allowed to report risk management as they use it, even in the standard formula approach.

You also have been critical of the option proposed in QIS3 to reduce equity risk significantly by taking into account the duration of the liabilities. Why are you critical of this evolution?

Samuel Sender: The “alternative” approach to equity and property risk reduces the stress scenario for equity risk from 36% for short-tail businesses to 13% for long-tail businesses. We think there is no theoretical backing for this idea, as the volatility of the equity portfolio over one year is naturally independent of the duration of the liability.

This choice will also create distortions between interest rate risk and equity risk, as bonds may be more penalised than equities for long-tail businesses. While equity investments that are backing a 10-year liability will require 13% of capital, investment in a 10-year bond will lead to a capital requirement of 15%1.

It seems as if CEIOPS has tried to lend support to the traditional view that long-term liabilities allow more risk-taking. As far as equity risk is concerned, given that Solvency II relies on a one-year horizon, a one-year measure of volatility must be used rather than a twenty-year one. Long-tail businesses will still be able to hold more equities in their books because they have more expected future profits as well as more expected profit-sharing as a buffer to risk.

This tendency of CEIOPS to incorporate traditional practices into a supposedly radically modern set of regulations probably reflects its desire to satisfy all participants, even though this sometimes means derogating on the very principles it has set for these regulations. The rule-making process has its own weaknesses.

What about the 45% capital charge for hedge funds?

Samuel Sender: A paper by Noël Amenc, Lionel Martellini and Mathieu Vaissié entitled The Benefits and Risks of Alternative Investment Strategies illustrates perfectly that hedge funds as an asset class brings significant diversification benefits to a portfolio.

Hedge fund is a very generic word for a fund that uses sophisticated techniques. These can be used either to reduce risk or maximise returns. In Hedge Fund Performance: A Vintage Year for Hedge Funds?, Véronique Le Sourd (2007) shows that hedge funds as an asset class have a lower Value-at-Risk than the S&P 500 over the past ten years.

Source: V. Le Sourd (2007)

Is CEIOPS afraid of systemic risk – the risk that hedge funds will make the system collapse – and on what grounds? Is it afraid that insurance companies may place a huge undiversified bet on a hedge fund? The nice formula for concentration risk could be easily extended to hedge funds so as to penalise companies that make large “bets” on a single hedge fund, as investment from these companies would certainly not be optimal.

What in your opinion are the major questions that remain to be addressed after the advances proposed in QIS3?

Samuel Sender: Within the standard formula, the three major advances that remain to be made are the following:

  1. Acknowledgement of dynamic strategies, because the current lack of recognition is unfortunately more of an incentive to regulatory arbitrage than to good risk management practices.
  2. Testing for interaction risk (e.g. supplementary redemptions during a bond market crash)
  3. Discount cash-flows to estimate non-life underwriting risk. Rather than computing the volatility of the premium relative to undiscounted cash-flows, the premium must be compared to discounted cash-flows. In theory, only discounted claims matter for pricing purposes, not undiscounted claims.

As far as internal models are concerned, the major issue is to determine how these will be recognised by the regulator. The notion of carte blanche that seems to prevail currently is probably not what can be expected to be the norm in the future.

Finally, are you hopeful at this stage about the overall framework for Solvency II, or do you think insurance companies and other institutional investors should be worried about it coming into effect?

Samuel Sender: When rules change, there must be winners and losers. As diversification benefits are now recognised, undiversified businesses shall be less favoured than well diversified ones. Equally penalised will be the companies that have both little available capital and small regulatory buffers for instance because of high interest rates, low profit sharing and low balance sheet reserves (such as capitalisation reserves, deferred profit sharing reserves like PPE, etc.).

However these businesses can take corrective action before Solvency II is implemented. For instance, one can notice that risk-transfer techniques such as reinsurance and securitisation will be far more recognised under Solvency II than they currently are. Under Solvency I, reinsurance could not diminish capital requirements by more than 50%. Under Solvency II, reinsurance will be fully accounted for in reducing the risk transferred, even if credit risk and concentration risk will rise in case of large exposure to reinsurers. This means that even weak companies have levers to correct their weaknesses. They may use them if they are convinced that there is a high degree of visibility on the agenda and content of the Solvency II project.

Timing is thus important. Too much delay may either make firms postpone their investment in risk management or conversely take the risk of implementing an organisational structure and processes that do not fit with the requirements for Solvency II.

For the industry as a whole, another worry is that the rule-making process may prove too weak to stick to its own yardstick and accepts counter-productive compromises with traditional practices or specific concerns.

Hopefully, as Solvency II is more principle-based than older regulations, companies can start to implement it, even if they are aware that some of the current proposed rules need to be revised.


1 This result can be approximated as follows: a 10-year bond has a duration of 8 and we will make the calculation with an 8-year zero-coupon bond instead. The stress scenario for interest rate risk is a 46% rise in the 8-year zero-coupon rate, which would jump from 4.7% to 6.9%. (1+6.9%)-8/(1+4.7%)-8=0.85, a 15% fall in market price under current market conditions. Computed with the full yield curve, the fall is slightly larger.

About Samuel Sender

Samuel Sender is a Research Associate with the EDHEC Risk and Asset Management Research Centre as well as ALM expert and senior insurance consultant for Finalyse, a consultancy specialised in implementing risk and performance measurement solutions at finance institutions.

Formerly with the Economics, Strategy and Quantitative Research unit at HSBC Asset Management Europe, and then head of Asset-Liability Management for a life insurer, he is a graduate in Statistics and Economics from ENSAE (Ecole Nationale de la Statistique et de l’Administration Economique) in Paris.

URL for this document:

Hyperlinks in this document: