models in Solvency II
D.A. van den As MSc
Master’s Thesis to obtain the master degree in Private Law, Commercial Practise track
University of Amsterdam Faculty of Law
Amsterdam Graduate School of Law
Author D.A. van den As MSc
Student nr. 10004690
Email duco.vdas@gmail.com
Date 19 August 2016
Supervisor prof. mr. dr. E.P.M. Joosen Second reader mr. K.W.H. Broekhuizen Supervisor Aegon mr. A.J.A.D. van den Hurk
In this thesis an insight is developed into the level playing field of the application pro-cedure for internal models used for the quantification of insurers solvency requirements, which is a crucial part of Solvency II. This thesis starts with an introduction into Sol-vency II after which we discuss the level playing field of the internal model application procedure for the six largest EU member states. National interpretations of the ap-proval criteria underlying the internal model are then revealed by a discussion on the long-term guarantee measures introduced in Omnibus II and their impact on internal models and the level playing field.
The patchwork of legislation that was in place before the introduction of Solvency II, has been replaced by more harmonized rules that still contain important flaws. The consistent assessment of assumptions underlying the internal model is important and will be most challenging for domestic insurance groups and solo insurers as they are not directly related to the supervision of EIOPA through the College of Supervisors. The hurried development of the long-term guarantee measures caused legislation to be introduced of which the precise implications for the level playing field were not properly measured. Nevertheless, the long-term guarantee measures introduced national discre-tions that are probably the most important reason that Solvency II was introduced in 2016. Achieving a level playing field for the application procedure of internal models will remain difficult without the introduction of a centralised structure in which the as-sessment of assumptions underlying internal models takes place. Benchmarking studies carried out by EIOPA could prove to be fundamental in achieving this objective.
Keywords Solvency II, Internal models, Level playing field, Omnibus II, Legislative effi-ciency, (Dynamic) Volatility adjustment, Matching adjustment, Transitional measures
This section contains the list of abbreviations that will be used throughout this the-sis. Abbreviations in the second part of this glossary will be marked bold at the first occurrence.
Directive 73/239/EEC First Council Directive 73/239/EEC of 24 July 1973 on the coordination of laws, regulations and administrative provisions relating to the
taking-up and pursuit of the business of direct insurance other than life assurance (OJ L 228, 16.8.1973, p. 3-19)
Directive 79/267/EEC First Council Directive 79/267/EEC of 5 March 1979 on the coordination of laws, regulations and administrative provisions relating to the taking up and pursuit of the business of direct life assurance (OJ L 63, 13.3.1979, p. 1-18)
Directive 2002/13/EC Directive 2002/13/EC of the European Parliament and of the Council of 5 March 2002 amending Council Directive 73/239/EEC as regards the solvency margin requirements for non-life insurance undertakings (OJ L 77,
20.3.2002, p. 17-22)
Directive 2002/83/EC Directive 2002/83/EC of the European Parliament and of the Council of 5 November 2002 concerning life assurance (OJ L 345, 19.12.2002, p. 1-51)
Lisbon Treaty Treaty of Lisbon amending the Treaty on European Union and the Treaty establishing the European Community, signed at Lisbon, 13 December 2007 (OJ C 306, 17.12.2007, p. 1-271)
Solvency II Directive 2009/138/EC of The European Parliament and The Council of 25 November 2009 on the taking-up and pursuit of the business of Insurance and Reinsurance (OJ L 335, 17.12.2009, p. 1-155)
Regulation (EU) 1094/2010 Regulation (EU) No 1094/2010 of the European Parliament and of the Council of 24 November 2010 establishing a European
Supervisory Authority (European Insurance and Occupational Pensions Authority), amending Decision No 716/2009/EC and repealing Commission Decision
2009/79/EC (OJ L 331, 15.12.2010, p. 48-83)
Directive 2012/23/EU Directive 2012/23/EU of the European Parliament and of the Council of 12 September 2012 amending Directive 2009/138/EC (Solvency II) as regards the date for its transposition and the date of its application, and the date of repeal of certain Directives (OJ L 249, 14.9.2012, p. 1-2)
Directive 2013/58/EU Directive 2013/58/EU of the European Parliament and of the Council of 11 December 2013 amending Directive 2009/138/EC (Solvency II) as regards the date for its transposition and the date of its application, and the date of repeal of certain Directives (Solvency I) (OJ L 341, 18.12.2013, p. 1–3)
Omnibus II Directive 2014/51/EU of the European Parliament and of the Council of 16 April 2014 amending Directives 2003/71/EC and 2009/138/EC and
Regulations (EC) No 1060/2009, (EU) No 1094/2010 and (EU) No 1095/2010 in respect of the powers of the European Supervisory Authority (European Insurance and Occupational Pensions Authority) and the European Supervisory Authority (European Securities and Markets Authority) (OJ L 153, 22.5.2014, p. 1-61)
Delegated Regulation Commission Delegated Regulation (EU) 2015/35 of 10 October 2014 supplementing Directive 2009/138/EC of the European Parliament and of the Council on the taking-up and pursuit of the business of Insurance and Reinsurance (Solvency II) (OJ L 12, 17.1.2015, p. 1-797)
ITS on internal models Commission Implementing Regulation (EU) 2015/460 of 19 March 2015 laying down implementing technical standards with regard to the procedure concerning the approval of an internal model in accordance with Directive 2009/138/EC of the European Parliament and of the Council (OJ L 76, 20.3.2015, p. 13-18)
ACPR L’Autorit´e de contrˆole prudentiel et de r´esolution
AP Application Package
BaFin Bundesanstalt f¨ur Finanzdienstleistungsaufsicht
CAP Common Application Package
CoEIM Centre of Expertise in Internal Models
CoA Contents of Application
CoS College of Supervisors
DNB De Nederlandsche Bank
EBA European Banking Authority
EC European Commission
EEA European Economic Area
ERB Eigen Risico Beoordeling
ESA European Supervisory Authority
ESFS European System of Financial Supervision
ESMA European Securities and Markets Authority
ESRB European Systemic Risk Board
EU European Union
IMC Internal Models Committee
IRSG EIOPA’s Insurance and Reinsurance Stakeholder Group
ISM Informal Supervisory Meeting
ITS Implementing Technical Standards
IVASS Istituto per la Vigilanza sulle Assicurazioni
LTG Long-Term Guarantee
LTGA Long-Term Guarantee Assessment
MA Matching Adjustment
MCR Minimum Capital Requirement
NCA National Competent Authority
ORSA Own Risk and Self Assessment
PRA Prudential Regulatory Authority
QIS Quantitative Impact Study
RTS Regulatory Technical Standards
SAT Self Assessment Template
SCR Solvency Capital Requirement
SFCR Solvency and Financial Condition Report
SOT Supervisory Oversight Team
UFR Ultimate Forward Rate
VA Volatility Adjustment
1 Introduction 1 2 Solvency II 5 2.1 Development . . . 6 2.2 Legislative efficiency . . . 8 2.3 Framework . . . 10 2.3.1 Quantitative requirements . . . 10
2.3.2 Governance and risk management . . . 12
2.3.3 Disclosure and transparency . . . 13
2.4 Solvency 1.5 . . . 13
3 Internal models 16 3.1 Insurers objectives with internal models . . . 17
3.1.1 Risk sensitivity . . . 17 3.1.2 Capital . . . 19 3.1.3 Supervision . . . 20 3.2 Applications . . . 21 3.3 Supervision of EIOPA . . . 22 3.3.1 Insurance groups . . . 22 3.3.2 Oversight . . . 24 3.3.3 Internal models . . . 25 3.4 Methodology . . . 26
3.4.1 Self assessment template . . . 27
3.4.2 Application package . . . 28
3.4.3 Harmonized procedure . . . 29
3.5 Underlying assumptions . . . 30
3.5.1 Sovereign risk . . . 30
3.5.2 Third country equivalence . . . 33
3.5.3 Benchmarking studies . . . 33
4.1.1 Volatility adjustment . . . 37
4.1.2 Matching adjustment . . . 39
4.1.3 Transitional measures . . . 40
4.2 Interaction . . . 41
4.3 Disclosure . . . 42
4.4 Solvency ratio sensitivity . . . 44
4.5 The way forward . . . 46
1 Introduction
Solvency II is the new insurance legislation framework that modernises supervision, im-proves consumer protection and increases international competitiveness of EU insurers. Solvency II introduces harmonized legislation for insurers that consists of directives that require implementation by member states and regulations that are directly applicable in member states. The implementation requires sophisticated legislative procedures and raises the question whether the national application of Solvency II legislation can guar-antee the level playing field between EEA insurers. In this thesis an insight is developed into the level playing field of the application procedure for internal models used for the quantification of insurers solvency requirements, which is a crucial part of Solvency II. In October 2008 the EC appointed the De Larosi`ere (2009) committee to advise the EC on the measures necessary to strengthen the supervision on the EUs financial markets. The weaknesses exposed by the financial crises urged the need for a compre-hensive approach towards financial supervision on all sectors in the financial industry. The goal of the EC was to create proposals that strengthen the supervision on all financial sectors to obtain an integrated, efficient and sustainable approach towards the supervision of the financial sector. Following the advice of the De Larosi`ere (2009) committee the ESRB was introduced on macro prudential level that maps risks that could affect financial stability and that identifies systemic risks. On micro prudential level the legislative quality was improved by reorganising the functioning of the EBA, EIOPA and the ESMA that form the ESAs. These ESAs focus on the banking sector, the insurance and pensions plan sector and the securities and financial markets respec-tively. They are closely entwined with the NCAs and form together with the ESRB the ESFS.
Solvency II was adopted in 2009 and was scheduled to enter into force in November 2012. Parallel to the development of Solvency II the Lisbon Treaty was signed which increased the effectiveness of the EU legislative procedures. The structure of the new ESFS as well as the new procedures introduced in the Lisbon Treaty were not incor-porated in Solvency II. The Omnibus II directive that addressed these shortcomings was not adopted until 2014 due to ongoing legislative negotiations. The trilogue par-ties (EC, the Council and EP) could not agree on the measures to value long-term insurance products and therefore EIOPA was given the assignment to conduct an as-sessment on the valuation methods for long-term insurance products. From the LTGA (EIOPA, 2013e), the importance of this problem became apparent. Certain aspects of Solvency II promoted pro-cyclicality which means that in times of financial downturn insurers would suffer disproportionally. Insurers tend to stabilize systemic risk (EIOPA (2013e, p. 41) and Bobtcheff et al. (2016)) which means that pro-cyclical effects of the
framework could result in more systemic risk which increases the risk of a collapse of the financial system as a whole. EIOPA introduced LTG measures that should prevent pro-cyclical investment behaviour and should protect against short-term financial mar-ket volatility. The development of these LTG measures would not be completed before the scheduled date of January 2014 when Solvency II would enter into force. Sufficient preparation time was crucial and therefore a second postponement of the application date of Solvency II to January 2016 was adopted in Directive 2013/58/EU.
The delays in the starting date of Solvency II have resulted in some of the member states to begin working on solvency supervision regimes derived from Solvency II and thus created a predecessor of Solvency II. Solvency II was not final at the moment interim regulation was made by certain member states and therefore the problem of uniformity arose. This problem is enlarged by the fact that Solvency II is structured as a directive that requires implementation in national law by member states. Disparities between the final Solvency II text and the national transposed laws may therefore arise. A question to be dealt with is therefore what type of legal act lead to the most optimal legislation implementation for Solvency II.
An important part of the Solvency II framework are the new methodologies intro-duced for calculation of capital adequacy for insurers. The required solvency capital should accurately reflect the risks an insurer is exposed to in order to shelter the in-surer against undue market volatility. Solvency II allows the required solvency capital to be calculated using a standard approach or using an internal model. The standard approach does what the name suggests and comprises of a generalised framework con-sisting of modules to calculate the Solvency Capital Requirement (“SCR”). The insurer should map the risks that he is exposed to and choose the appropriate modules that most accurately describe these risks. When modules of the standard approach do not suffice to accurately describe their risks, they must determine their required solvency capital using a partial or full internal model. This internal model encourages the insurer to think about their risks and should therefore lead to better risk sensitivity. Besides the mandatory requirement for an internal model, another reason to apply internal models voluntarily could be a potentially lower required solvency capital than in the standard approach because required but unsuitable modules can be omitted.
The standard approach is the same for all European insurers which implies that a European-wide implementation will likely result in a fair degree of uniformity of capital requirements between insurers assuming comparable businesses. Insurers implementing (partial) internal models must obtain approval from the NCA in the sense that their risks are adequately quantified. The criteria on which this approval is based are not fully covered in the Solvency II framework and lead to national interpretations. The application procedure of internal models is therefore not harmonized in the EEA and may create inequality between insurers using internal models.
be calculated using the LTG measures introduced in Omnibus II. The application of the LTG measures is interesting because they have a significant impact on the required solvency capital of the insurer (CRO Forum, 2014). Member states are also allowed to require approval in advance for the application of some of the LTG measures which could imply a more lenient review when approval is given in hindsight. The use of the LTG measures, when incorporated in the internal model of insurers, therefore effectively result in different treatments depending on the member state in which the insurer is located. In addition to the question on the overall level playing field of the internal model application procedure, the implementation of the rules concerning these LTG measures will indicate the level playing field of a specific component in internal models. The main question that we answer in this thesis is whether an insurer using an internal model in the Netherlands has different capital requirements than an equivalent insurer in another member state. For this purpose, we evaluate the implementation process of Solvency II and conduct a comparative law study into the internal model application procedure of the six largest EU member states, being the Netherlands, the United Kingdom, Germany, Italy, France and Spain (Insurance Europe, 2016). Moreover, we review the application criteria of the LTG measures to see to what extend the level playing field is achieved with these instruments when used in internal models. In answering these questions, we also review the use of directives for EU legislation and whether the delay of Solvency II caused additional implementing challenges.
A study on the subject of the level playing field in Solvency II is important because differences in the approach by NCAs may lead to inequality of the outcome of solvency calculations by insurers. Due to the nature of internal models, a NCA with lower ap-proval standards could reduce regulatory costs significantly thereby affecting the level playing field. Consistent application of Solvency II legislation on internal models ensures a comparable level of protection to all policyholders and prevents regulatory arbitrage in the internal market. Our main question should therefore be answered negatively as regards the existence of a level playing field within the internal model application procedure in Solvency II.
We expect that differences exist between approaches followed by NCAs due to the nature of internal models and the fact that Solvency II is implemented in national law with a directive. A directive must first be implemented in national law and provides for minimum or maximum requirements that the national law must satisfy. Although Solvency II is mostly maximum harmonisation legislation (EIOPA, 2015, p. 11), the remainder is not and may result in regulatory arbitrage in the sense that NCAs affect competition on the internal market with a lenient approach towards national insurers. We commence with an introduction in Solvency II and the EU legislation procedure. Then we direct our focus towards internal models in Solvency II and provide an analysis of the LTG measures introduced in Omnibus II and the effect they have on internal models and the level playing field. This thesis is completed with a conclusion regarding
the level playing of Solvency II with respect to the subjects we discussed.
In this thesis Solvency II legislation is featured with a strong economic and quan-titative component. This is due to the economic nature of Solvency II and the fact that Solvency II was only implemented recently such that no body of jurisprudence law exists. We assume that the reader already acquired basic knowledge about the insur-ance industry. This includes the division of the insurinsur-ance industry between life-insurers, non-life insurers and reinsurers, where we will primarily focus on life-insurers. The dis-cussion on the LTG measures requires basic knowledge of an insurer’s balance sheet. This knowledge includes the fact that most insurers have liabilities with a longer ma-turity than the mama-turity of assets. Consequently, increasing (decreasing) interest rates cause the value of liabilities to decrease (increase) faster than the decrease (increase) in the value of assets, such that more (less) capital is available. The objective of insurers is to align future cash flows from their liabilities as closely as possible to the payments that they receive from their assets. For example, if the insurer sold a product that requires a payment due in ten years’ time, then the insurer aims to buy an asset with a cash flow with the same maturity. Finally, it is important to know that the yield of a bond indicates the amount of return that an investor can earn.
2 Solvency II
Solvency II is the newest European insurance legislation framework and replaces 14 non-life and non-life Solvency I directives (EC, 2015). Directive 2002/13/EC was the most recent Solvency I non-life directive and amended the 1973 First Council Directive 73/239/EEC. Directive 2002/83/EC was the most recent Solvency I life directive and repealed the 1979 First Council Directive 79/267/EEC. The rationale for harmonized European insurance legislation is to develop a Single market of insurance services and to maintain high standards of consumer protection throughout the EU. The Solvency I directives that were in effect for 40 years could no longer achieve these objectives as they were outdated and took certain risks not into account (EC, 2007, p. 6). An example is the absence of capital requirements for market risk, which is the risk of a loss resulting from adverse financial market movements. The absence, until the introduction of rules in Solvency II, of capital requirements for market risk resulted in inadequate supervision laws as the riskiness of a financial asset should determine the amount of capital the insurer holds.
Solvency II aims to provide a comprehensive approach towards regulation of the European insurance industry and introduces risk-based solvency requirements that are more risk-sensitive and sophisticated than Solvency I. The solvency requirements are estimated such that they better fit the risks that an insurer is exposed to. In addition, a prospective focus is added to the management of risks in the sense that insurers must identify and measure possible future developments. Disclosure requirements are expanded which enables the NCA to identify insurers heading for difficulties as well as increasing competition and enabling market participants to evaluate the insurer’s solvency position better. The framework will also facilitate easier management of in-surance groups, such that they are treated like a single economic entity. When solvency requirements are calculated on group level, the cooperation between NCAs responsi-ble for the supervision of insurance groups operating on a cross border basis should increase.
The main objectives of Solvency II are therefore to harmonize the regulatory frame-works for insurers across the EU and the EEA; to ensure financial health of insurers in volatile times; and to protect policyholders and the financial stability as a whole (EC, 2007). The framework requires insurers to take all relevant risks into account which will be reflected in the required solvency capital and eventually into the design of their products. The framework can be seen as a “solvency thermometer” that introduces an early-warning mechanism to enable NCAs to intervene before a serious possibility of insolvency arises. More specifically, in the case that the available solvency capital be-comes insufficient, NCAs will require the insurer to restore their financial situation. If
the financial situation further deteriorates, supervisory intervention will be intensified. This system should therefore identify threats before they can harm policyholders.
In this chapter we commence with the path leading up to the introduction of Sol-vency II. With this knowledge we are able to evaluate the efficiency of the legislative procedure with which Solvency II is developed. We continue with the pillars to achieve Solvency II’s objectives, where our main source of input is a publication by the EC (2007) on Solvency II’s structure. Finally we look into Solvency 1.5 that was created by certain member states as a response to multiple postponements in the starting date of Solvency II.
2.1
Development
The official start of Solvency II was given by the EC in 2004 with a framework con-sultation (EC, 2004). The objective was to create a single market of insurance services that improves consumer protection, that modernises supervision and increases the in-ternational competitiveness of European insurers and reinsurers. The system should feature maximum harmonisation without additional country specific requirements and a structure consistent across financial sectors. This implies that the approach and rules used in the banking field (Basel 2) should be resembled in the new solvency system. Following this framework consultation, the predecessor of EIOPA conducted the first two Quantitative Impact Studies (“QISs”) in 2006 to gather data for the development of the technical methodology on solvency requirements.
After the framework consultation, the Solvency II proposal of the EC was sent to the EP and the Council in 2007 but was only was only adopted in 2009 by the EP due to negotiations between the EC, EP and the Council. At the same time of the work on Solvency II, the Lisbon Treaty of 2007 introduced several measures that increased the efficiency of the legislative procedures in the EU. The Treaty reduced the number of legislative acts to 5 categories. Based on article 288 of the TFEU, European institutions may now adopt 5 types of legal acts, the regulation, the directive, the decision, the recommendation and the opinion. The first three are binding legal acts, whereas the last two are non-binding legal acts. Furthermore the EP and the Council may now delegate the power to adopt non-legislative acts concerning non-essential elements of legislative acts. These delegated acts allow the EP and the Council to focus on a policy direction without entering into a technical debate. The final change in the Lisbon Treaty was the improvement of the EC’s powers to adopt implementing acts. Normally the responsibility to implement binding EU acts lies with EU countries. In certain cases there is emphasis on the uniformity with which binding legal acts must be implemented. The EC is therefore given the option to adopt implementing acts (article 291 TFEU) that provide for uniform conditions for the implementation of legislation.
The Solvency II framework did not incorporate the Lisbon Treaty principles and therefore had to be updated. In January 2011, the EC published a proposal to adjust insurance legislation to ensure that the new ESAs can work effectively. The Internal Market and Services Commissioner Michel Barnier said:
”The financial crisis in Europe exposed weaknesses in the supervision of financial mar-kets, which the new EU financial supervisory structure intends to correct. Today’s pro-posal is an important building block to ensure that the new supervisory bodies will run smoothly. By giving the new supervisors a clearly defined mandate and bringing existing legislation in line with that mandate, the Commission further delivers on the promise of creating more solid and stable markets and mitigating future crises.”.
In this proposal the areas in which the ESAs can exercise their powers is set, which includes drafting technical standards to increase the uniformity of supervision and to contribute to the single rule book. This single rule book should lower compliance costs and should prevent regulatory arbitrage as legislation is applied uniformly.
After this proposal was sent to the EP and the Council, EIOPA released the final QIS in 2011 for the calibration of Solvency II’s implementing measures. This study showed that Solvency II was becoming increasingly complex which was not the idea of the original 2004 Solvency II framework consultation. Furthermore, the mark-to-market valuation of long-term insurance products resulted in significant volatility in the value of own funds. The calibration of solvency requirements depends greatly on the fact whether an insurer enters into a short-term or long-term insurance contract. A long-term insurance contract requires long-term investing whereas a short-term contract requires short-term investing. Solvency II provided insufficient information on this important topic and therefore Directive 2012/23/EU postponed Solvency II to 1 January 2014.
Following this postponement, EIOPA was given the mandate to conduct a technical assessment on long-term insurance products to find out how long-term insurance prod-ucts should be treated when short-term market movements occur. In June 2013, this LTGA (EIOPA, 2013e) was published and resulted in a Volatility Adjustment (“VA”), a Matching Adjustment (“MA”) and transitional measures that could be used in the valuation of long-term insurance products. Due to the short implementation time frame to update the EC’s proposal, Directive 2013/58/EU was adopted which now postponed Solvency II to 1 January 2016. After the legislation for these adjustments was incorpo-rated in the EC’s proposal, Omnibus II was adopted in 2014 which amended Solvency II such that the EC can adopt delegated and implementing acts, the ESAs could function efficiently and that included the LTGA measures to value long-term insurance products. The introduction of Omnibus II changed the legislative procedure. Implementing measures that were in place before Omnibus II are replaced by delegated and imple-menting acts. The delegated acts cover the Solvency II Directive requirements in more detail for individual insurers and make up the core of the single prudential rulebook
for insurers. On the basis of article 290 TFEU and 291 TFEU, the EP and the Council can delegate power to the EC to adopt Regulatory Technical Standards (“RTS”) (arti-cle 10 of Regulation (EU) 1094/2010) and Implementing Technical Standards (“ITS”) (article 15 of Regulation (EU) 1094/2010) that become legally binding upon approval by the EC. The RTS are standards for the consistent harmonisation of rules in EU legislative acts, whereas ITS are purely technical and give the conditions of application of the delegated acts. The EC may only amend RTS in extraordinary cases because EIOPA is in closest contact with the insurance sector. When RTS or ITS do not cover certain areas, EIOPA can issue guidelines and recommendations on the application of Union law. The reason for this legislation procedure is the technical expertise needed to develop these standards. EIOPA is closest to the insurance sector and is therefore given the assignment to draft RTS and ITS.
Among other regulations adopted in the context of Solvency II, the Delegated Reg-ulation of 2015 is the result from the new legislative structure introduced in Omnibus II and contains detailed requirements for individual insurers in Solvency II that make up an important part of the single rule book.
2.2
Legislative efficiency
Before the financial crisis new legislation was constructed using the Lamfalussy leg-islative procedure. This procedure consisted of 4 levels that each focused on a certain implementation stage. Level 1 is the highest level and comprises of a formal proposal of the EC for a directive or regulation. After adoption by the EC, this proposal is send to the EP and the Council for consultation. When an agreement is reached the EC starts working on level 2 legislation. The EC will seek advice on the implementing measures and the EP must be kept fully informed on the progress. When sufficient im-plementing measures are created, the EC adopts the measures and the legislation will continue to level 3 where guidelines and common standards are generated to ensure consistent implementation and application throughout the EEA. The final level 4 of this procedure controls the implementation process of directives and the application of regulations. In extreme cases the EC may take legal action against member states that insufficiently satisfied requirements, such as timely implementation of a directive in national legislation.
The report of the De Larosi`ere (2009, par. 103) committee pointed out that a lack of harmonized European prudential law was an important reason for the severity of the financial crisis as insurers amplified the financial cycle. The national transpositions of directives led to distortions in competition and regulatory arbitrage that both must be avoided at all times. Solvency II must therefore be complemented with lessons from the crisis in order to remedy the fragmentation of prudential law in the EEA.
Earlier in this chapter we saw that following their recommendation, the EC proposed the creation of the ESFS with the ESA’s and the ESRB. These bodies together with the changes from the Lisbon Treaty amounted to a shift in the regulatory procedure in the EEA. The level 2 implementing measures that the EC composed before Omnibus II was implemented, were replaced by a new structure where the EC was given the power to adopt delegated acts and implementing acts. The implementing acts may now be drafted by EIOPA and become legally binding upon approval of the EC.
Year
2000 2005 2010 2015
Number of legislative acts
0 5 10 15 20 25 Regulations Directives
(a) Insurance companies
Year
2000 2005 2010 2015
Number of legislative acts
0 5 10 15 20 25 Regulations Directives (b) Financial institutions
Figure 2.1: Number of directives and regulations in the Insurance Company and Financial institution EUR-Lex categories over the last 15 years.
Figure 2.1 shows the number of directives and regulations for the insurance sector and the financial sector over the last 15 years. Directives in the insurance sector category include Solvency II in 2009, Omnibus II in 2014 and other directives related to insurance such as the insurance of shipowners directive in 2009 and the civil liability insurance directive in 2009. The regulations in the insurance sector category include the Delegated Regulation and the ITS on internal models in 2015. In the financial institutions category we can see the four banking capital requirements directives in 2000, 2009, 2010 and 2013 and the capital requirements regulation in 2013. We can see that in the last years in both categories more regulations were adopted than directives, although this effect is much stronger in the insurance sector category.
The shift in the EU legislative procedure is therefore confirmed by the increased number of regulations compared to directives. This shift can be seen as an indication that Omnibus II with the new legislative procedure was a turning point and that the Lamfalussy legislative procedure insufficiently achieved the maximum harmonisation objective of Solvency II. The current situation is therefore an environment where regu-lations form the most important part of the EU legislative procedure and it is therefore the most effective tool for the EC to achieve Solvency II’s goals.
2.3
Framework
Three pillars aim to accomplish the objectives of Solvency II. The first pillar specifies minimum amounts of capital that insurers are required to hold in order to survive volatile times. The second pillar ensures a solid governance and risk management system and the third pillar ensures harmonized disclosure standards and transparency.
2.3.1
Quantitative requirements
The first pillar focuses on the capital requirement that an insurer is required to hold in order to survive volatile times. The insurer will first have to calculate the total amount needed to cover expected future claims from policyholders. This amount is called the “technical provision” and should equal the amount another insurer would pay in order to take over the insurer’s obligations to policyholders. In addition to the technical provisions the insurer should have sufficient funds to cover the SCR to meet obligations calculated with 99.5% certainty covering obligations due within a 1 year’s time frame. The SCR is calculated using the standard approach or calculated using an internal model verified by the NCA.
The standard approach consists of modules that represent risks that an insurer is exposed to, see figure 2.2. The idea of the standard approach is to map the risk exposures of the insurer and subsequently to choose the appropriate risk modules to calculate the SCR. An example that illustrates the modular structure of the standard approach is a fully reinsured insurer. In such case all claims that are be made against this insurer are fully insured with another insurer. This insurer will therefore have no exposure to the market risk module as changes in for instance the interest rates will not affect him, but he will have high exposure to the default risk module as this is the risk that the counterparty may default on his obligations.
Solvency Capital Requirement
Market Health Default Life Non-life
Interest rate Spread Concentration
Modules
Sub-modules
Figure 2.2: The modular structure of the standard approach in Solvency II. Not all (sub-) modules are included to preserve clarity.
Another insurer that is not fully re-insured, can also have exposure to the market risk module. Important sub-modules from the market risk module in the context of this thesis are risks arising from changes in interest rates, changes in spreads and changes in concentration. Interest rates are not only used as a percentage that borrowers charge each other for lending capital, but the interest rate is also used to calculate the technical provision of liabilities. The interest rate risk sub-module quantifies the impact of a shock applied to interest rates on the value of assets and liabilities. The spread risk sub-module is another important sub-module in the market risk module and takes the risk related to changes in credit spreads into account. The credit spread is the difference between the yield on government or corporate bonds over the interest rate. This spread is an indicator for the riskiness of bonds such that a high spread resembles a risk premium that the investor obtains by investing in that particular bond. The spread risk sub-module quantifies the impact of a shock applied to credit spreads on the value of assets. The lack of diversification in the number of counterparties in the assets portfolio is quantified in the concentration risk sub-module. This sub-module is calculated by assigning a threshold to asset classes such that high exposures to one particular asset class results in higher capital requirements for the concentration risk sub-module. The methodology of the SCR in internal models will be discussed in the next chapter.
When an insurer fails to meet the SCR, a recovery plan must be sent to the NCA within two months, containing measures to meet requirements within six months (article 138(2) and 138(3) Solvency II). Possible measures include raising additional capital to cover the SCR or mitigating risks that result in a lower SCR. Apart from the SCR, a Minimum Capital Requirement (“MCR”) is calculated from the SCR that must be present at all times. If an insurer fails to meet this obligation he must compose a short-term recovery plan within one month for the NCA, containing measures to meet the MCR within three months (article 139 Solvency II).
The exact amount of solvency capital an insurer will hold in practice is difficult to estimate. The amount of solvency capital will likely be higher than the SCR as the insurer will not want to send a recovery plan to the NCA every time when small market fluctuations occur. The amount of excess capital over the SCR will however not be too high as these funds may not be invested in higher yielding investments. In figure 2.3 we compare the solvency ratio year-end 2015 of a selection of large EU insurance groups, where the SCR method is added between brackets if a (partial) internal model was used. The solvency ratio indicates the extend to which the insurer is able to cover the SCR with available capital. A solvency ratio of 100% indicates that the insurer has precisely enough capital to cover the SCR.
Solvency ratio
50% 100% 150% 200% 250% 300% 350% 400%
Poste Vita (IT) Munich Re (DE, IM) NN (NL, PIM) Talanx (DE, IM) Axa (FR, IM) Generali (IT, PIM) Prudential (UK, IM) CNP (FR, PIM) Allianz (DE, IM) Aviva (UK, IM) Legal & General (UK, IM) Intesa Sanpaolo (IT) Vivat (NL) Aegon (NL, PIM) Delta Lloyd (NL)
Figure 2.3: Solvency ratios of large insurers from year-end 2015 annual reports, data on Spain was not available.
The average solvency ratio is 211% and we see that the solvency ratio is with no exception always substantially higher than the SCR, which begs the question why this is done. The Dutch insurers Delta Lloyd, Aegon and Vivat show relatively low solvency ratios around 150% and German and Italian insurers show relatively high solvency ratios around 190%. There is even an Italian insurer with a solvency ratio of over 400%.
The level of capital that insurers aim to hold in excess of 100% will depend on a number of factors. One factor is that the level of capital in the first pillar is calculated with the assumption that sufficient capital to cover the SCR should be available with 99.5% certainty. Given this level of certainty insurers can choose their optimal level of capital such that their targeted credit rating is reflected in the available capital to cover losses. For example, when the 99.5% certainty level equates to a credit rating of BBB, and the insurer is targeting a credit rating of AA to facilitate easier lending, then an additional amount of capital over the SCR is necessary. This reality goes beyond the original thought of Solvency II that the ratio should act as solvency thermometer which triggers gradual intervention of the NCA once the SCR or MCR is breached.
2.3.2
Governance and risk management
The second pillar ensures solid governance and risk management systems. The internal processes of the insurer should be structured such that risks are identified, assessed and handled in a timely manner. Furthermore the prudent person principle must be em-ployed which aims to ensure well-thought conduct by the insurer in making investment decisions. These elements are complemented by the ORSA that considers the impact of risks that are not explicitly considered in the calculation of the SCR. Despite the
fact that some risks are not explicitly considered in the SCR, the ORSA does not imply that an insurer must develop a (partial) internal model. The ORSA aims to include a more prospective focus with possible risks, the consequences of these risks for the financial position and ways to avert these risks. The ORSA is therefore not designed as an additional capital requirement, but the NCA can still apply an additional capital requirement if the capital requirements from the ORSA and the SCR do no coincide (recital 27 Solvency II).
2.3.3
Disclosure and transparency
The third and final pillar is structured to ensure harmonized disclosure standards and transparency. Solvency II requires insurers to report information to a far greater extent than previously was the case. Information under Solvency II is published in four com-ponents, the first component is the reporting of the solvency and financial position, the second component the regulatory report including quantitative information, the third component consists of a report containing pre-defined scenarios and the influence of these scenarios on the financial position and the fourth component contains the pol-icy concerning the disclosure of information and reporting to the regulator. Important information about the financial health of the insurer is reported to the public, which enables the NCA to better identify insurers heading for difficulties. In turn market participants will be able to evaluate the insurer’s solvency position better which will increase the competition between insurers in favour of consumers.
2.4
Solvency 1.5
In the years leading up to Solvency II, it became apparent that the risks faced by insurers were not adequately covered by Solvency I and therefore the EC started working on Solvency II in 2004. The development of Solvency II had multiple delays that were mainly caused by the negotiations on Omnibus II. At the time of the negotiations the timetable in which Solvency II would be implemented was unclear (EIOPA, 2012) and the fear arose that a smooth transition towards Solvency II could only be achieved with a realistic timetable. The need to ensure sound prudential supervision together with the postponements caused member states to reform Solvency I parallel with the development of Solvency II. In the Netherlands this period of transition was paraphrased as Solvency 1.5. These reforms were not coordinated on a European level which led to an unharmonised framework of regulatory requirements in the EEA. To prevent further dispersal of regulatory requirements and to prepare insurers for Solvency II, EIOPA released an opinion (EIOPA, 2012) on the key areas that must be prepared before the introduction of Solvency II. Following this opinion EIOPA announced guidelines for a preparatory phase at the beginning of 2013 that contained important elements
of Solvency II that must be in place by January 2014. This competence is based on article 16 of Regulation (EU) 1094/2010, which gives EIOPA the possibility to issue guidelines that aim to increase uniformity of NCAs in the implementation stage of Solvency II. NCAs are obliged to comply to the best of their ability with each category of these guidelines. To ensure transparency and to strengthen compliance by NCAs, EIOPA can publish the reasons for NCAs’ non-compliance. These guidelines are geared towards NCAs and insurers and are divided into four categories. The pre-application of internal models, reporting, requirements for business organisation and risk management and a forward-looking assessment of own risks.
To evaluate the stance of NCAs towards these guidelines, EIOPA held a public consultation with NCAs regarding all categories (EIOPA (2013a) for the system of governance, EIOPA (2013b) for the forward looking assessment of own risks, EIOPA (2013c) for reporting and EIOPA (2013d) for the pre-application of internal models). When the NCA responded that he will comply or intends to comply, he must incorporate the guideline into his supervisory framework. The results of the Solvency 1.5 public consultation are given in table 2.1 where six NCAs are compared.
Internal Models Reporting Country 70 provisions 39 provisions
Netherlands C C (1), IC (38) United Kingdom C (65), IC (5) C (29), IC (9), NC (1) Germany IC IC Italy IC C (1), IC (38) France C C (1), IC (37), NC (1) Spain C IC
System of Governance Forward Looking Assessment 52 provisions 25 provisions Netherlands C C United Kingdom IC (42), NA (10) C(19), IC (6) Germany C (6), IC (46) C (3), IC (22) Italy C (24), IC (28) IC France C (5), NC (47) C (1), IC (14), NC (10) Spain C (11), IC (41) IC
Table 2.1: EIOPA preparatory phase guideline compliance.
In this table the total number of provisions is given for each guideline and the compliance is subsequently specified with regard to each provision. NCAs have 4 choices, they can comply (C), they can intend to comply (IC), they can choose not to comply (NC) or they can decide it is not applicable (NA) to them. If parentheses are added, variation in compliance exists and the number of provisions is given for that compliance choice. C(6) therefore indicates that the NCA complies with 6 provisions of the guideline and filed a different choice for the remaining provisions.
The Dutch regulator decided to almost fully comply with EIOPA’s guidelines (DNB, 2013) in March 2013. EIOPA’s guidelines on the internal model pre-application were im-plemented, which are mainly used as guidance for the development of an internal model. The governance requirements based on EIOPA’s guideline were also implemented as far as they were sufficiently clear. In anticipation of the ORSA, the Dutch regulator imple-mented an ERB which is based on the ORSA and the forward looking assessment of own risk guideline but is less thorough. With respect to reporting, the Dutch regulator implemented EIOPA’s respective guideline, which implies the insurer must provide an annual report over the past year as well as two quarterly reports over that year. With regard to certain provisions in this guideline the regulator intended to comply due to ongoing discussions on the required amendments. As of 1 January 2014 the WFT also contained an arrangement where insurers opting to pay dividends, must first apply for approval from the NCA when their financial position is not satisfactory at the time the dividend is paid out. The Dutch authorities therefore have applied the option to restrain a dividend pay out which caused problems in those years with the level playing field and makes it harder to compete with other European insurers which will not benefit policyholders (Verbond van Verzekeraars, 2013).
The PRA followed not long after the Dutch authorities and implemented EIOPA’s guidelines in December 2013 (PRA, 2013). With some provisions the PRA intended to comply or did not intend to comply. The latter approach was followed in cases where the absence of the final technical specifications and the generality of certain provisions prevented the adoption of the interim rules anticipating the final regime. The guideline compliance of France shows more mixed results. The ACPR cannot comply with 58 provisions due to a national rule that prohibits formal guideline implementation before the transposition of Solvency II in 2015. The ACPR proposed however to report the preparation and progress of insurers through surveys to EIOPA.
Earlier in this chapter we noted that Omnibus II was a turning point in the Solvency II legislative procedure and that regulations became the most effective tool to achieve Solvency II’s goals. In this section we have seen that Solvency 1.5, that prepared insurers for the introduction of Solvency II, failed to harmonize the regulatory frameworks within Europe which allowed problems with the level playing field to arise.
3 Internal models
Solvency II introduces new methodologies for the calculation of capital adequacy for insurers. Insurers should map the risks that they are exposed to and evaluate whether the modules of the standard approach suffice to accurately describe these risks. In case risks are not adequately described, the risk sensitivity of the standard approach can be improved by using a (partial) internal model (article 112 Solvency II). A partial internal model only addresses risks not or not appropriately accounted for in the modules of the standard approach, whereas a full internal model requires the insurer to calculate all risks internally such that all modules of the standard approach are replaced in its entirety. Note that partial internal models are used most frequently and that seldom a full internal models is developed. These approaches allow insurers to align their solvency positions more closely with their own assessments on required regulatory capital.
The internal model is subject to prior approval by the NCA to provide policyholders with an equivalent level of protection as under the standard approach. Developing an internal model is not always a voluntary option for insurers. The NCA may insist on the application for an internal model if the products offered by the insurer differ too much such that they are not aligned with the assumptions underlying the standard approach (article 119 Solvency II). This is the case for variable annuity products that are paid for during a certain period and convert to cash flows dependent on investment returns at, for instance, the retirement age. Products dependent on investment returns are not appropriately accounted for in the standard approach EIOPA (2011b, par. 13) and consequently a (partial) internal model will be required by the NCA.
Preferably the dialogue with the NCA starts before the formal internal model ap-plication is made (recital 5 ITS on internal models). NCAs should therefore put pre-application processes in place to form a view on the insurer’s readiness for submission of an internal model application and its ability to meet the ongoing requirements (EIOPA, 2014d, Guideline 1). Once the application is made, articles 112(4) to 112(6) Solvency II give the NCA the authority to assess the internal model. Approval is given when the NCA is satisfied that the systems to identify, measure, monitor, manage and report risk are adequate, that at least the six tests1 from articles 120 to 125 Solvency II are fulfilled
and when using a partial internal model, that the additional requirements from article 113 Solvency II are met. The timeline of this procedure consists of four phases with different durations. The completeness assessment phase takes one month, the maximum duration of the assessment phase of the application itself is six months, the optional remediation phase of EIOPA can take one month and the decision letter phase can also take one month. Once the internal model is approved by the NCA, it is not possible for
1(i) Use test, (ii) Statistical quality standards, (iii) Calibration standards, (iv) Profit and loss
attribution, (v) Validation standards, (vi) Documentation standards. Later on these tests will be discussed in detail.
the insurer to revert to the standard approach in whole or in part (article 117 Solvency II). Only when the internal model ceases to be compliant with articles 120 to 125 Sol-vency II and remediation of this deviation failed within a reasonable time frame, can the NCA require the insurer to apply the standard approach (article 118(2) Solvency II) for the calculation of solvency requirements.
First we discuss why an insurer would apply for an internal model, after which we compare the number of internal model applications and evaluate the structure at EIOPA aimed at resolving level playing field issues. We then discuss the methodology of the approval process, which allows us to evaluate the level playing field of the internal model application procedure.
3.1
Insurers objectives with internal models
We divided the reasons for an insurer to apply for an internal model into three cate-gories. These categories are (i) risk sensitivity, (ii) capital and (iii) supervision and will be discussed in the next sections.
3.1.1
Risk sensitivity
Developing an internal model starts with mapping the risks that the insurer is exposed to. The insurer then determines what level of risk sensitivity in the SCR methodol-ogy is adequate to describe this risk profile. Risk sensitivity is therefore dependent on the complexity of the insurer such that the standard approach is adequate for a low complexity insurer, whereas the internal model is adequate for a more complex insurer. When two insurers differ in size but not in complexity, the same SCR methodology can still be applied. Therefore, risk sensitivity involves the ability to proportionally quantify the complexity of the insurer.
Increased complexity arises when the risk profile deviates from the assumptions underlying the standard approach. Deviations from the standard approach occur when either certain risks are not accounted for or when risk modules are less granular than would be needed to accurately describe the complexity of the insurer. Inflation risk is an example of a risk that is not fully covered in the standard approach and deals with changes in (the volatility of) inflation rates. Inflation risk is sometimes referred to in modules of the standard approach to the extent that a future inflation rate of 1% per annum should be assumed. The importance of this module depends on the type of products an insurer has sold. If products with guaranteed future indexation were sold, inflation risk would be present and the SCR should be quantified appropriately with an internal model. Another risk is reputational risk, which is the risk that the trustworthiness of an insurer is affected and in turn results in a loss of revenue. Apart from the difficulties of quantifying this risk, large insurers should still consider the
impact of this risk. Furthermore, the standard approach was designed for solo insurers and not specifically to insurance groups (EIOPA, 2014e, p. 10). Some risks exist because of their position within an insurance group and are therefore not taken into account in the standard approach. Despite these gaps in the standard approach, the availability and granularity of modules is quite complete, with the exception of products with specific risk characteristics (Doff, 2016).
Considering the potential mismatch between the chosen SCR methodology and the risk profile, the standard approach can still be applied when the risk profile differs not much and when this deviation is adequately accounted for in the ORSA. The consid-eration for a relative high level of risk sensitivity has of course the benefit of a better fit with the actual risk profile which leads to better risk management, but on the other hand requires a significant effort both to develop and to implement into the model. Solvency II is structured such that the workload imposed by mandatory provisions of law is proportional to the complexity of the insurer (article 29(4) Solvency II). Simpli-fied methodologies should be applied to insurers characterised by low complexity risk profiles, who in addition also have insufficient resources to conduct the full Solvency II workload. Slightly more complex insurers should apply the standard approach and, if appropriate, in addition determine “Undertaking Specific Parameters” such that addi-tional risk sensitivity is added (EIOPA, 2015, p. 2). When adjusting standard approach parameters is insufficient, the risk sensitivity can be increased by using a partial or full internal model. Although this principle of proportionality is formally incorporated in Solvency II legislation, some weak points exist as the EC requested EIOPA in July 2016 to review the application of the principle of proportionality for smaller insurers (EC, 2016, p. 2). Especially the cost arising from the need to obtain external credit ratings from credit rating agencies is not proportional to low complexity insurers that need those credit ratings to perform SCR calculations. The proportionality principle of the SCR methodology in Solvency II is depicted in figure 3.1.
Complexity Proportionality Standard approach with simplifications Standard approach Standard ap-proach with USPs
Partial internal model Internal model
When the chosen SCR methodology and the risk profile deviate significantly from the assumptions underlying the standard approach, a capital add-on can be imposed by the NCA. Capital add-ons may be imposed in exceptional circumstances in both the standard approach as well as the internal model as measures of last resort (article 37(1) and recital 27 Solvency II). In the standardised approach a capital add-on is possible when the mandatory use of an internal model by the NCA through article 119 Solvency II is either inappropriate or when an internal model is being developed. When the SCR methodology is based on an internal model, a capital add-on is still possible when the risk profile deviates significantly from the assumptions underlying the internal model and remediation of deficiencies as they are observed by the NCA failed within a reasonable time frame. Once the deficiencies that lead to the imposition of the capital add-on are addressed, the capital add-on can be removed during the yearly review of the NCA (article 37(4) Solvency II). Inadequate use of the standard or internal model SCR methodology may therefore result in additional capital requirements, which makes risk sensitivity an important consideration.
3.1.2
Capital
Development and application of an internal model requires the application of significant costly resources that make the prospect of capital relief an important consideration. An insurer may expect that the better fit of the internal model to the risk profile lowers the SCR because required but unsuitable modules from the standard approach can be set aside as they are not relevant. NCAs may expect an increased SCR due to complexity of insurers and products not accounted for in the standard approach such as inflation-linked products. This misalignment of interests must be handled with care as the NCA will be wary if the SCR comes out significantly lower than the SCR in the standard approach. A report by Morgan Stanley and Oliver Wyman (2012, p. 47) that evaluated the group internal model approval process, found that approval is becoming increasingly difficult when the application of an internal model results in a significantly lower SCR than what the standard approach would produce. This is especially true with regard to the categories sovereign risk and policyholder behaviour, where the NCAs may insist on assumptions that lead to a higher SCR than the standard approach. Although the information on which this statement is based is unclear, it may provide an incentive for insurers to alter their internal model from the most accurate reflection of their own view on risk. Hence, the capital consideration lies between the risk and capital considerations because a trade-off between less accurate risk management and higher capital requirements exist.
In March 2011, EIOPA conducted the fifth QIS (EIOPA, 2011a) and also compared the level of capital calculated with the application of internal models to the capital cal-culated when applying the standard approach. A total of 234 EU insurers were asked
to calculate the SCR for both the standard approach as well as the internal model. An important remark about these internal models is that they were not yet approved by the NCAs which makes the results not quite comparable to each other. Despite this fact, the results showed that in thirteen of the nineteen participating countries most insurers had a lower internal model SCR, whereas in the other six countries most insurers showed a higher internal model SCR. The average ratio of capital between internal models and standard approach was 91%, whereas the weighted ratio was 99% which indicates that large insurers deviated less (when applying internal models) from the standard approach capital requirement. Analysis showed that the higher internal model results of insurance groups were caused by low diversification and a risk profile characterised by large exposures to one type of risk. When assessing the level of capital in an internal model on a country by country basis, this QIS indicated a clear distinc-tion between countries where most insurers obtained higher internal model results and countries where most insurers obtained lower internal model results. Identifying the specific countries in which this occurred is however impossible due to the confidential nature of this QIS.
3.1.3
Supervision
Insurers using internal models for SCR calculations are subject to supervision by the NCA in both the application and maintenance stage of the internal model. Contrary to the standard approach, evidence of compliance with Solvency II and the Delegated Regulation must be submitted to the NCA in order to get approval. Consequently, multiple NCAs are developing own internal model approval processes and criteria. In case of adverse market conditions, these processes and criteria will be tested in real life circumstances and may expose NCAs to reputational risk. The prospect that failure of the internal model will put blame on the NCA, will likely expose insurers to additional supervision in both the application stage as well as the maintenance stage during the lifespan of the internal model.
Another important aspect of supervision on insurers are changing supervisors at NCAs. Supervisors at NCAs evaluate whether the evidence provided by insurers is suf-ficient to meet all requirements for the application of an internal model. When these supervisors change, expertise needed to comprehend the specifics of the internal model methodology of insurers may be lost. This could give rise to a brain drain that stands in the way of maintaining the supervision of adequacy of internal models over time. Assumptions accounted for during the application stage of the internal model are sub-ject to the scrutiny of new supervisors that are not obliged to align their opinion on assumptions with those of their predecessors. Changing supervision for insurers using the standard approach will be less important as the standard approach is coordinated from the EU with less discretions for NCAs.
3.2
Applications
The reasons to apply for an internal model were discussed in the previous section and outline some important decisions that the insurer must make. Legislation on internal models evolved extensively in the years leading up to Solvency II, which could have changed the reasons for an insurer to apply for an internal model. To relate this changing mindset to the difficulty of getting approval from the NCA, we have retrieved data from press releases by the NCAs on the number of internal model applications. In figure 3.2 the number of internal models applications for the Netherlands, the United Kingdom, Germany, Italy and France is depicted in the years 2014 and 2016. The year 2016 indicates the number of internal model applications that have received approval from the respective NCA.
Year
2012 2013 2014 2015 2016
Number of internal model applications
0 1 2 3 4 5 6 7 8 9 10 NL (a) Netherlands Country NL UK DE IT FR
Number of internal model applications
0 5 10 15 20 25 30 35 40 45 50 2014 2016
(b) Member state comparison
Figure 3.2: Internal model applications from the Solvency II Wire website, data on Spain was not available).
All member states show a decrease in the number of internal model applications over time, although the strength of this effect differs between member states. Interestingly, the difference between the number of internal model pre-applications and the number of internal model approvals was small in Germany. In the UK, France and Italy this difference was higher and a substantial number of internal model applicants did not obtain approval. This was also the case in the Netherlands were data from more years was available and where a consistent downward trend is also visible. Although the decrease in the number of internal model applications on itself is not conclusive evidence for the strictness of NCAs, it does provide an insight into the demand for and stance towards internal models by NCAs. The high approval rate of Germany could imply splendid preparation, clear communication and a more tolerant approach from the NCA. The lower approval rates of the Netherlands, UK, Italy and France could indicate the opposite of the previous explanation. This indicates that as the legislation on internal
models evolved, the reasons to apply an internal model have changed more in some member states than in others.
3.3
Supervision of EIOPA
The activities of EIOPA can be divided into regulatory and supervisory tasks. During the development of Solvency II, EIOPA was mainly concerned with the regulatory task to support development on all levels of the Solvency II legislative procedure. Now most Solvency II legislation is finalised, more emphasis is needed on her supervisory task to ensure consistent application of Solvency II legislation across the EU.
In the next sections we discuss the role that EIOPA has in the supervision on insurance groups and the role that EIOPA has in keeping oversight. This provides the necessary information to discuss the structure currently present at EIOPA that is aimed at resolving level playing fields related issues. We can then return a verdict on the level playing field of the internal model application procedure.
3.3.1
Insurance groups
Group supervision is carried out on insurers with multiple entities that form a group (article 213 and 214 Solvency II). The key issue for the NCAs charged with the supervi-sion on these insurance groups, is to identify the level at which group supervisupervi-sion must be carried out. The level at which group supervision is carried out depends primarily on the location where the insurance group is headed.
Insurance groups headed in the EEA are subject to group supervision at the ul-timate EEA parent, although group supervision is also possible at lower levels under the ultimate EEA parent. Group supervision on insurance groups headed outside the EEA depends on the equivalence of the local system of group supervision (article 260 Solvency II). Note that multiple definitions of equivalence exist and that we solely look at the equivalence of group supervision. The EC is given the competence to adopt a delegated decision (articles 288 and 290 TFEU) on the equivalence assessment of the principles underlying the local group supervision with the principles underlying Sol-vency II. When the EC assessed the local group supervision as being equivalent to Solvency II, local group supervision is carried out and no additional requirements are imposed by Solvency II. If local group supervision is assessed as not being equivalent to Solvency II, then group supervision is carried out on the ultimate parent outside the EEA or alternatively on the ultimate EEA parent. As of now, the EC only confirmed full equivalence of the Swiss and Bermuda supervisory regimes. Group supervision regimes in the USA, Canada, Australia, Brazil, Mexico and Japan are provisionally equivalent (article 227(5) Solvency II). Provisional equivalence means that not all criteria are met and that equivalence is only valid for a limited period of time after which the period
may be extended. The assessment of foreign systems of supervision requires a signifi-cant amount of time and therefore Omnibus II introduced provisional equivalence as a practical measure to not further delay the introduction of Solvency II.
Group supervision inside the EEA is a joint effort between the NCAs from the member states in which the group and her subsidiaries reside. For each insurance group a College of Supervisors (“CoS”) is set up to ensure cooperation and to facilitate an exchange of information between NCAs. The group supervisor provides the solo supervisors with information on the group and the solo supervisors provide the group supervisor with the financial developments and risks in the solo entities. In most cases the group supervisor is the NCA from the member state in which the insurer that is heading the group is located. The CoS allows a shared view on the risks and future strengths of the insurance group to be developed among NCAs. From 2011 onwards, EIOPA participates in most CoS to set targets, to organise training days and to share practical solutions such as the assessment of the group ORSA (EIOPA, 2016b, p. 1). Every year EIOPA devises an action plan with areas that need attention. In the action plan for 2016 the appropriateness and consistency of the ORSA and the group SCR calculations across the group form the most interesting topics.
The pillars of supervision that apply to solo insurers, generally also apply to insur-ance groups. A group SCR must be determined from the first pillar, a group ORSA from the second pillar and the reporting requirements from the third pillar must also be met. With respect to the calculation of the group SCR, a group internal model may be developed (recital 102 Solvency II). In the group internal model the aim is at how the total model works, at the appropriateness of the standard approach or internal models at lower levels in the group and at the local implementation of group models. In addi-tion to the internal model applicaaddi-tion at local level, the internal model applicaaddi-tion at the group level must include additional information. The most important additions are a list of entities included, the legal and organisational structure, an estimate of the SCR calculated with the group internal model, an estimate of the SCR calculated with the standard approach and an explanation for the difference between them (articles 343(5) and 347(6)(a) Delegated Regulation). When insurance groups are using the Common Application Package from EIOPA to structure their application, they are allowed to submit only one application containing both the solo and the group information. Later on, this package will be discussed in more detail. Another difference is the assessment of the group internal model as the group supervisor will have the lead and consult with the supervisors of all subsidiaries included in the group internal model.
In total, EIOPA participates in the CoS of 100 insurance groups across the EEA. It is important to emphasize that there is a distinction between international insurance groups and domestic insurance groups and solo insurers. Furthermore, not all insurance groups use internal models and therefore not all insurance groups are of interest to use. In table 3.1 we provide an overview of the number of international insurance groups
in which EIOPA participates in the CoS and the number of approved internal models as of January 2016. The international insurance groups are classified according to the location where the insurance group is headed.
Comparison between
international insurance groups and internal models
International insurance groups in Approved internal models Member state which EIOPA participates in the CoS as of January 2016
Netherlands 4 3 United Kingdom 15 19 Germany 17 6 Italy 7 1 France 18 10 Spain 2
-Table 3.1: Comparison between the number of international insurance groups with a CoS and the number of internal model applications (Source: EIOPA).
In total, 63 insurance groups operate internationally and are headed in one of the six countries that we specified. Earlier, in the previous chapter, we found that in total 39 internal models were approved as of January 2016. When assessing the application procedure of internal models it is therefore important to differentiate between insurance groups that operate internationally and insurers that mainly operate domestically. In the first case approval of the group internal model is based on a process involving multiple NCAs that take part in the CoS, whereas approval of the internal model for a domestic insurance groups or solo insurer is solely based on the assessment of the NCA of the country in which the insurer is located. According to the annual report of EIOPA on the functioning of the CoS, the NCAs reached a joint decision on the approval of an internal model in 10 cases (EIOPA, 2016b, p. 3). This provides interesting grounds for reconciliation as the other 29 internal models that were approved, were solely assessed by NCAs on national basis.
3.3.2
Oversight
In addition to the input and information that is shared with NCAs, EIOPA wants to give NCAs feedback on their practices. For this purpose, EIOPA created a Supervisory Oversight Team (“SOT”) in 2014 that is aimed at developing and maintaining relations with the NCAs. The SOT provides NCAs with feedback on their practices on the way regulations are interpreted and implemented in practice by NCAs in their day-to-day supervision. Furthermore, feedback on the way that NCAs cooperate with one another, on the process and outcomes of risk analysis conducted by NCAs and on the output of benchmarking studies across NCAs is given.
