1.Challenges specific to internal risk modelling
1.1 Modelling is resource intensive for both insurance groups and supervisors
An industry representative said that one challenge is the resource intensity; this is a multi year work in progress, and many, many people are involved in the groups. There is a very interactive relation with the supervisors and the main challenge really is the resource intensity.
There are two points in relation to the issues that they are facing: first, it is often said that internal models are complex. This is true, but this is because the business is complex and the balance sheet is complex. Therefore, internal models are not bringing in any further unjustified complexity that is already foreseen in Solvency II. There is also the issue of the balance sheet approach: the modelling of insurance liabilities is a complex thing, but they do not add further complexity. The second point is that the question is sometimes asked whether the internal models are black boxes. They are not; they have a very in depth dialogue with a supervisor, and both parties are looking into this ‘box’.
A regulator said that, from a supervisory point of view, the main challenge is the workload, not only for the approval process, but also for the ongoing process. This is a huge opportunity for supervisors as well, because it is the only way to achieve a full understanding of what a group is doing. So far, internal model companies are the best supervised companies, although there is always room for improvement. This means having to have very qualified and educated staff. They have to be on a similar level to the industry, and that is a challenge, but they have been able to achieve this. That creates pressure to maintain this level; this is crucial, and a challenge for the whole supervisory world.
An insurance regulator said that internal models are better than the standard formula in certain specific situations, such as for large and complex groups. In France, they adapted the supervisory organisation; this is something that insurance companies feel, as well, and welcomes the European work that has been done in the colleges of supervisors. This is a mix of supervisory culture, where good practices were identified. There is still some process to be made and more experience needed, but overall, the validation of internal models has definitely been a success. That is a very important message to hear from the supervisor.
1.2. Modelling is costly if multiple layers of validation and multiple regulators
An industry representative said that, regarding challenges, it has been a ‘long and winding road’ to reach this point, and an expensive road. There are a couple of points as to why it was more expensive than they originally wanted it to be. That is to do with multiple layers of validation and multiple regulators, in their specific case, looking at the same thing. It is already a challenge to bring the internal view and a single regulator onto the same page, but when dealing with multiple regulators, this is becoming almost impossible. This is one of or the main reasons why the speaker’s institution took some subsidiaries out of the internal model process, because they cannot have one model governed by many different regulators.
2. Key success factors to maintain trust
2.1 Internal risk-assessment models should be (remain) as accurate as possible
An industry representative had been ‘a bit astonished’ by the discussions in some of the panels, such as whether EIOPA or the European Commission should tweak some factors in the standard formula to foster investment in certain asset classes. Ultimately, this is not what the regulatory model should exist for. The regulator does not exist for purposes of macroeconomic steering. If there are companies with internal models, this debate will not happen because the internal model will do the right thing and steer in the right way, and there is no need to bring the regulators into the macroeconomic debate.
An industry representative said that the riskiness of the different businesses needs to be understood, along with the value that each different business brings to the table, in order to come up with an economic valuation of the balance sheet. Understanding internal models is also required of how to steer asset liability management, and how to use reinsurance to optimise the risk profile and the capital structure of a company.
Regarding the success factors to use as the basis of Pillar 1 capital, there is a very thorough and reliable approval process. This is not only the approval process, because that is for Solvency II and is done with, but there is also the ongoing monitoring. That will be a challenge, even given the big success in the Solvency II countries; the monitoring and maintenance of these models will be an ongoing challenge.
2.2. Three issues that can lead to a deterioration of trust on accuracy.
An industry representative insisted saying that the key challenge is trust, and how to know that what is being said and seen is actually what is being experienced with this internal model. There are three issues that can lead to a deterioration of trust. The first is complexity: the world is very complex, and the internal model models events in the real world, including real life dependencies. This is complex, and cannot be denied. The second challenge to trust is comparability; there is a concern on the supervisory side and the investor side about the comparability of the resulting capital requirement of the internal model. Comparability would be very useful, but it should be asked what the most accurate way to achieve comparability is. The solvency capital requirement should be estimated correctly, although the standard formula does not mean it is correct. To estimate the solvency capital requirement correctly means nothing other than a 99.5% valuated risk of the basic owned funds one year ahead.
Another challenge to trust is conflict of interest. Certainly, between supervisors, the management and investors, there are different interests when it comes to the size of the solvency capital requirement. It is better to have a transparent conflict of interest where the supervisory interaction dialogue is intact; this is what supervision is ultimately all about. That challenge is dealt with the use test, and the use test means nothing other than the internal model is entirely used in the risk management system and for the decision taking processes. This is key, and is the most important thing. Management will obviously care about getting this right, because no one supports incorrect risk taking.
3. Main benefits and achievements in the area of risk modelling
3.1. Improved risk sensitivity allows more risk analysis, better management and reduces pro-cyclicality
An industry representative said that on the benefits, they recalled three reasons not to defend internal models, but just to explain why they are in the regulation and why they exist. The first point is that they are better than standard models, because they are tailor made to the risk that the company is facing. This is something that is very valuable, and the arguments for it are quite well known. If there are different roles in different markets, the standard formula often has a single loss ratio and a loss ratio can be modelled where they are. Internal models can also be used to do more analysis. This is really what is foreseen in Solvency II, which is the risk of the whole balance sheet under consideration.
The second point is that internal models are really a true management tool, and are more and more used in this way within a company. If major decisions are faced in terms of M&A, in terms of acquisition merger business changes, and if many changes are faced on the assimilation of shocks and changes in the environment, the internal model can be used to govern and steer the company. That is something that the standard model does not provide, and that is a second and very significant benefit.
The third benefit has more to do with the macro perspective with financial stability, in that it helps that companies do not act all alike in financial markets. The biggest enemy to financial stability is pro cyclicality: i.e. everybody acting in the same fashion. This is a major risk. Internal models allow for some diversification among the companies, and therefore avoid them all reacting in the same fashion over the cycle. Whenever there is homogeneity, such as the reference model, the reference portfolio and the volatility adjuster, there is a risk that all players behave the same way in the markets. This is certainly not good for financial stability, and this is another important reason as to why internal models are helpful.
An industry representative agreed and said that many statements focus on the use of internal models to Pillar 1 capital: the regulatory capital requirement. The question is obviously broader: it is, regarding internal models, what the challenges are and what the success factors are. The most important thing is that internal models are a key risk management tool. They are required by insurance companies and any financial institution to steer and manage the business.
A regulator said that volatility is clearly an issue in Solvency II. The increased fluctuations in regulatory requirements are true for a standard formula and for internal models, but internal models can reflect this much better. Conscious management of capital is needed, and that can be managed by an internal model even more precisely than it can be done with a standard formula.
3.2. It is necessary to combine bespoke modelling and sufficient comparability and transparency of risk models
The riskiness of the different businesses, the asset and liability management, how to use reinsurance to optimise the risk profile and the capital structure of a company, and other such internal processes, in one way or another, will all need to be reflected in an internal model. It is difficult to come up with a good single internal model that can be used. However, if this is taken in the context of determining Pillar 1 capital requirement, the picture then changes, because the level playing field then becomes the overarching expectation. Competitors all want to compete on the basis of a level playing field. There may then need to be comparability, and maybe even benchmarking, to understand they are not basing Pillar 1 capital on internal experience, but on industry experience and so on.
There is an additional layer of challenge if an internal model is then employed the basis of Pillar 1 capital. However, success factors, in the broad context of internal models, means using them in the day to day business in key strategic decisions, and making them usable by underwriters, asset managers and so on.
An industry representative said that, for a representative of a global reinsurance company, the most important thing for their internal model is having an integrated internal model, which means that their internal model is not only used for the regulatory capital assessment but entirely to determine the economic capital, which is required to support the risks on the book and also steer the risk taking and behaviour. The internal model is therefore embedded in a business cycle, which is the key difference to the banking side. This needs to be borne in mind when talking about internal models on the insurance side.
Whatever model is required, at the end of the day it is most important that the stakeholders, the supervisors, the investors, the management itself and the board of directors of insurance companies understand how the risk profile or the business model is reflected in the model is being looked at. When looking at the model, they should ask whether the purpose of the model is understood, whether the methodology applied to the model calculation is known, whether the model’s limitations are known, and how diversification is taken into account.
3.3. According to the countries with experience, internal modelling is successful
An industry representative stated that at this point in time, it is probably worth taking stock and emphasising that internal models can work, and have been proven to work, in the regulatory context, both in Switzerland – where there is long experience of this – and also in the European Union. This can work, and this should not be forgotten. It is a huge success, and not just on the regulatory side: something has been achieved. Management decisions create the same outcome in the internal view on the risk and capital needs, and on the regulatory view on risk and capital needs.
If operating under the standard formula – which certain subsidiaries in the EU are doing – and steering according to two different directions, one from the standard formula and one from the internal model, there is confusion. If there is a consistent message in terms of what capital needed for a certain business, from an internal and regulatory perspective, the right thing is being done.
The investor community should not be forgotten when talking about internal models. It has taken a while to achieve this, but in the meantime, investors have accepted that steering is according to the internal model and have learned to live with it. The numbers have become more volatile than they have been used to, but ultimately, this is an issue of the volatility of the economic environment.
Regarding multiple validation, there is a robust internal governance for introducing the model. They have independent internal validation, an external auditor and the regulator; there is a need to carefully think through whether there is one step too many in this chain. The independent internal validation that Solvency II introduced is a success; it generally works, and much more reliance should be put on that independent, in house validation for the use of internal models, and more thought given to how to delay the approval and validation process.
4. The ability to achieve an effective comparability and level playing field and define appropriate minimum standards, still raises scepticism among supervisors
4.1. Supervisors at the global level are not totally convinced about comparability and level playing field
A regulator explained that the IAIS had more than 50 pages of Insurance Core Principles (ICP), 17 of which deal with solvency and the internal model. These ICP are the foundation of any regulatory standards, and the speaker’s institution has a very detailed explanation of internal model. One paragraph, or one or two sentences, is extremely important. By its very nature, a standardised approach may not be able to fully and appropriately reflect the risk profile of each individual insurer. Therefore, where appropriate, a supervisor should allow the use of a more tailored approach subject to approval; i.e. the internal model.
Currently, in the context of ICS global capital standards Pillar 1, what will be debated from June onwards at the global level is about what is appropriate: first of all, whether an internal model is going to be used, and to define where it is appropriate. In other words, the regulators at the global level are not totally convinced or sure that they are going to begin the internal model ‘journey’.
It has been said that it works, particularly on the industry side and also the regulatory side from the European context, as a tailor made, manageable and diversified approach that creates more trust. However, as had been mentioned, the key challenge is to persuade all the regulators around the world that they have to debate on comparability and the level playing field. That is critically important. In this context, not only Europe but also in the US and Asia, people are not all familiar with the internal model approach. They are very sceptical, because of comparability.
Furthermore, supervisors’ role is to set the minimum standard, and the regulator’s role is to set the minimum rule. Therefore, the balance that needs to be struck is to assess correctly, accurately and individually, but to also set the minimum rule. In the banking sector, regarding the so-called Basel IV, which is very simple, they have not agreed because they do not agree a minimum floor. They are debating a minimum floor of 80% or 70%. This is very simple, but it is a huge discussion, because they have to set very clear minimum standards comparable to everyone’s level playing field.
4.2. Defining how to progress from using models for business decision toward using them for regulatory purposes
In this respect one person clarified that internal models are essentially more relevant to the insurance sector than the banking sector. In insurance, liabilities are risk models; they are the basis of the business. They are a reality, used to set technical provisions for many years. However, the essence of this integrated model is much more about risk management, to be used for business decisions, pricing, and all those elements from a strategic perspective.
The question is how to progress from this to using it for regulatory purposes, and what is needed from the regulatory framework to ensure that they are used for regulatory purposes and do not prevent the use for risk management. This element of comparability, inherent conflict of interest, model drift and temptation to go down every year to optimise capital, as has been seen on the banking side, is the big challenge. Ultimately, the worst thing that could result from this conflict is that the base, which was used for risk management purposes, will not be there anymore. This would be ‘a disaster’.
4.3. Making all supervisors more familiar with internal models, and improving comparability and trust, are key success factors
An industry representative said that there is not an easy answer, but there may be one element that is relevant. It seems that companies, after many years of work, are comfortable and the supervisors, after many years of validation, are comfortable too. That is a very important starting point. The question that then needs to be answered is what can be done on comparability. First, there is a need to recognise that there is no suspicion that internal models are meant to reduce capital; that they are meant to optimise. Everyone is in the optimisation business to reflect risks correctly, but there is nobody who believes there is a race to the bottom. That is also very important.
Regarding comparability, one issue could be trust amongst supervisors, so that if a supervisor validates the internal model, which often involves colleges, there is already an international aspect. If this is validated, the other supervisors in the community can trust that the supervisor in charge has been throughout the whole process. It is much more important to get assurance on that point, rather than move towards uniformity in the model itself.
A regulator said that their institution is involved in 12 internal models on a cross-border basis. They are now 15 years into the new regime, which remains nevertheless a young regime. A lot is being learned, not only through the interaction with various groups but also in the colleges, about different views about how to reach to a more harmonised and coherent approach, which is very important.
Europe needs to strongly distinguish its own territory from the rest of the world; even in Europe, it needs to overcome some scepticism, but is making good progress. EIOPA has recently launched an internal working group about how to compare parameters, which is very fruitful work. There is a need to distinguish this from the global context, and transparency has not yet been achieved in relation to the reasoning and advantages of internal models. EU regulators did not spend enough energy on promoting their internal models.
5.1. Avoiding uniformisation is essential
A regulator said that internal models need to remain in a situation where the model overall reflects the risks under consideration and is a true reflection of these in the company in question, rather than going point by point and asking whether to use the same parameter across the models. This would also create a race to the bottom and, in a sense, another standard model which is the smallest common denominator of the internal models. They would then lose all of the benefits that have been discussed and agreed upon. There is nothing more important than having trust among the supervisors and an open dialogue, so that they feel comfortable without having to go through uniformisation of the internal models themselves.
If they want to reap the benefits from a financial stability perspective, they should not move to uniformisation; uniformisation is one issue, and another is the hugely differing approaches. Supervisors also need to do work on the ongoing monitoring, because trust is very important, but there inevitably remain different views on certain things. A process will be needed to cater for that.
5.2. A standardised approach or a standard formula, will not create comparability
An industry representative said that their institution has roughly 20 companies that are operating on the standard formula. They are not comparable with respect to the outcome; people should not be under the illusion that a standardised approach, or a standard formula, will create comparability. The comparability has to be on the outcome, but if a model is applied to very different companies with very different risk profiles that is not tailored to this, there will not be comparability.
The speaker’s institution has the benefit of being able to look at their businesses on the internal model tailored to their risk profile and on the standard formula on the European entities. They achieve comparability by looking at them through the internal model, and know that they do not achieve comparability by looking at them through the standard formula. It should be clear that ‘standard formula equals comparability’ is wrong.
5.3. Reliability requires monitoring rather than comparing
An industry representative said that the debate about insurance should not compare insurance with the banking side, or take the examples of the Basel committee introducing floors. They should take the differences into consideration. Secondly, looking at the ICS today and the concept that is in place, comparability will never be achieved given the way that this is set up. Europe should not use the debate around internal models to achieve some kind of comparability.
On the benchmarking side, the speaker’s institution employs the standard formula, as well as the internal model for a European supervisor. Interestingly, the European supervisor also asked the speaker’s institution to stop doing this, because it is not meaningful in any way; there is no way to meaningfully explain the differences. Something that is very useful, which the speaker’s institution fully support, is the internal model ongoing appropriateness indicators that must be in place. It is crucial to know how internal models develop over time; even in the company itself, this must be done, and the speaker’s institution has tried to do this. There are different methods for doing so. Based on the speaker’s institution’s experience, the most promising method is to look at the relative changes in the output, and compare these to the relative changes in the input. If there is no clear explanation for the observations, then the underlying purposes need to be investigated.
The EIOPA has been developing this in recent years. In terms of the shock between the benchmarking and developing internal model indicators in the sense of understanding the way the devolution of the model goes, both of these ideas have merits. The important point is to understand how the different elements in the model compare at the end of the day. There can be a very good explanation as to why different companies with different business models have different risk charges for certain types of risks, but this needs to be understood. This is a process that will originate from the supervisor side, and cannot be given by benchmarking, but benchmarking can play a role.
6. In an environment where insurers are increasingly under pressure to deliver cash to shareholders and optimise their balance sheet, strong governance practices are required
6.1. Possible conflicts of interest within insurance companies exist, which require sound governance practices
An industry representative said that, regarding the terms of the model drift optimisation, they might be falling into the same trap that the banking industry was a couple of years ago, constantly trying to ‘squeeze the last penny out of the internal model’: there might be a need to take a step back. Banks were never really about internal models; they were about internal parametrisation of a standard model. That is very different, because that means that what banks are operating on was different from a regulatory needs perspective, compared to what was used internally. This is exactly what makes this conflict of interest very transparent. If a regulatory standard is optimised against, that is something other than optimising against an internal standard. This conflict of interest exists already within the company, between the first and second line of defence. They have learned to live with this, and have built up strong governance practices to live with this conflict of interest, bring it out into the open, have the discussion and then come to an ultimate conclusion.
This is a very open and transparent way of solving these conflicts of interests for their internal model, and it is a solvable problem, as long as transparency exists. A standard formula, or the application of a standard formula, creates the same conflict of interest, but without the internal governance of the first and second line of defence. The risk of ‘gaming the system’ is even higher when someone is forced into a standard formula that can be gamed.
An industry representative said that regarding how the industry or regulators ensure that companies use their internal model not to optimise as part of a race to the bottom, setting standards and benchmarking are important to some extent, but it ultimately needs to be a supervisory instrument. There needs to be recognition of how strong and successful Solvency II was. The use test and the own risk insolvency assessment are now key instruments for the regulators to hold insurers to account, to use their internal models in an appropriate fashion, to understand, and to prove to themselves, their board and, ultimately, to the regulator that they manage the capital base in a forward looking fashion with a sustainable solvency basis.
6.2. Achieving a reliable valuation of the balance sheet of each company is another challenge to be addressed
An industry representative said that a central question is how to avoid this race to the bottom in an environment where insurers are increasingly under pressure to deliver cash to shareholders and optimise their balance sheet. This is a business reality. There is a strong precedent from Solvency II, and Europe has let its assumptions in this regard be influenced somewhat by Solvency II. One thing is entirely standardised within Solvency II: the valuation of the balance sheet. There is no internal model for valuation of the balance sheet. This is a challenge, and there are probably differences between the valuations, but all the things being discussed – the ultimate forward rate, the matching premium, and the more challenging areas such as the valuation of deferred taxes, the loss absorbing capacity, and policy holder participation – are highly complex. However, the design is that they should be standardised as the valuation of the technical provisions.
This forms a major element of the thinking about an internal assessment of the volatility of the speaker’s institution’s balance sheet that is, within the Solvency II framework, highly standardised. This would be a big challenge for international standards, because that is simply not true. IFRS has struggled to come up with strong valuation principles on an accounting basis. Europe needs to be honest with itself that within Solvency II, there is a great drive towards standardisation of what is a very tricky area of the valuation of the balance sheet.
7. Conditions for defining a common language within supervisory colleges
7.1. To achieve trust, the parties need to speak a ‘common language’
A regulator said that their mandate is to create global comparability. Trust is at the heart of all global coordination and global consistency. In a supervisory college, in the European context as well as the global context, there is a need to establish trust to collaborate and work together. This issue has been debated many times, but the conclusion is that in order to achieve trust, the parties need to speak a ‘common language’. They understand each other and share the same view, but have some common elements that can be communicated with each other.
The question, however, is how to create this common language in the context of this capital tier 1 world, and the answer is not entirely clear. It had been mentioned that they should not compare insurance with banking; the valuation issue had also been mentioned.
7.2. Bank supervisors are trying to define common benchmarks to address possible excess resulting from modelling freedom
If the regulator understands correctly, in the banking world, in order to understand each other better, they tried to set some benchmarks at a more granular level. This is the general framework; ultimate weight is a good example of this. A very concrete benchmark is an aspect of valuation, but there are some who wait to interpret other aspects of valuation. The ongoing indicator approach could achieve a concrete, comparable approach, whereby there is a clearer definition of minimum standards in a very neutral, objective manner, aspiring to some sort of flexibility.
It was very clear from the beginning of the banking crisis that the ways in which different banks treated risk weighted assets were completely different. This was the root of the discussion about the input and output floors for banks in Basel, since this differentiation built up to a level at which trust was lost. Following this, there was the return to leverage ratios and related things, in order to reintroduce more comparability.
There is an inherent difference: on the banking side, this is much more of a parameterisation exercise, devising internal parameters for a more standard model, which is completely different from internal models in the insurance sector. However, Investors often make this point; they say that they have even less trust in the comparability of numbers coming from the insurance side, because the level of freedom that is given to insurers in internal models is much bigger than in the banking sector.
7.3. The main objective should be to enable any stake holder to understand the scope, features and limitations of each internal model: the role of supervisors, the management and investors has to be defined in this respect
An industry representative said that it could be asked whether the individual approach to the risk weighted assets really makes sense on the banking side, if the risks they are addressing are compared. On the insurance side, insurance risks are entirely different between companies, and this is another one of the key differences. The issue of freedom for internal models relates back to the fact that a stakeholder needs to understand what the model is, what it is not covering, and what the limitations of scope are. That responsibility cannot be taken away from investors, and as a company, the speaker’s institution needs to be able to explain this to investors. Understanding of this is growing.
Clarity is needed about the role of the supervisor, and in this discussion around comparability, it is not clear what is driving the call for comparability on the supervisory side. First and foremost, this is the responsibility of the insurance company itself; the board of directors needs to take the accountability for that.
7.4. The main objective is to enable the supervisor to achieve their responsibility, which is to protect the policyholders and ensure that there is sufficient capital in the company
A regulator agreed that this is primarily the responsibility of the companies, and with regimes like Solvency II, this is becoming more and more clear. Ultimately, the protection of policyholders is the role of the supervisor, not of the regulator. This is the balance that needs to be struck. The issue of a thorough understanding of the model by the different stakeholders had been mentioned; this is very important, and constitutes a ‘journey’. Analysts and investors suggest that there is not sufficient information disclosed about the models themselves to make them comparable.
An industry representative said that whenever there are requests from analysts, there is a discussion about them, but the speaker’s institution do not receive an overwhelming number of requests from analysts to go inside the internal model. On the comparability question, the more that this is discussed, the more unclear the aim becomes. Internal models are permitted, are used, and the supervisory responsibility is to protect the policyholders and ensure that there is sufficient capital in the company. It is also unclear what this comparability is, although it could be said that the comparability comes through the 99.5% rule at the bedrock of Solvency II. The industry representative felt that this debate might be going down the wrong track.
Finally, although there has been discussion of a race to the bottom, people should keep in mind that most companies are far above the 100% solvency ratio; they are closer to 200%. Even with investors, the speaker’s institution has been very public about a range of 170% to 230%, and they are moving this range. It is not correct to accuse companies of being involved in a race to the bottom.
The regulator replied that they do not think this is an issue at the present time; what the industry needs, collectively, is to ensure that this does not take place. On the banking side, a race to the bottom destroyed internal models, and that should not be the way that this issue progresses.
8. Conditions to achieving the essential objective which is to restore trust on internal models
Regarding comparability, this also means similar portfolios of investments backing similar types of business having completely different types of capital requirements and risk charges. From a supervisory perspective, that is something that needs to be dealt with; a good explanation and processes are required that, at the end of the day, make this much more consistent.
8.1. Trust requires consistency over time, using models primarily to achieve the best sustainable returns, and having effective transparency regarding the sensitivities
An industry representative said that there needs to be more consideration of how the trust of investors in the internal models can be restored; investors are a very important stakeholder group. Three things have helped in this respect; the first is consistency over time. The speaker’s institution has been publishing internal models for more than 10 years now, and the investors can follow this over time. What the speaker’s institution has demonstrated is not ‘randomly jumping from one quarter to the next ‘. Secondly, there is significant interest in making sure that the speaker’s institution is using this in order to optimise the return on the capital that they use; that it is not just a regulatory exercise, but that it is used for allocating the capital to achieve the best returns.
What has helped a lot in terms of restoring investors’ trust was discussing the sensitivities of the model, and how model results would change if market parameters were to change, but also if the speaker’s institution were to change certain assumptions. They do have different assumptions, and they need to be transparent about that across firms. Helping investors achieve comparability by showing sensitivities has helped the speaker’s institution.
8.2. A more central role for a European supervisor and the colleges of supervisors are certainly positive
An attendee said that insurance does not need to follow the banking path to reassuring stakeholders that the model is properly used. To increase trust, there should be a stronger and more central role for a European supervisor, which accounts for all the differences that exist. He asked whether, if a central role could be given to EIOPA in this regard, the industry representatives would be happy.
In this respect a participant stated that EIOPA will be very attentive to what is being published this year on internal models, and specifically on sensitivities. This might help investors to achieve an understanding of the risks that the different companies and groups have.
Am industry representative noted the work that has been done in Europe in the colleges of supervisors, in which supervisor cultures mix, and good practices were identified had been welcomed. This is also a voice that should be heard.
8.3. Supervisors need to be close to individual companies and take into account their size
An industry representative said that the speaker’s institution respects any way in which the supervisors organise themselves, or the regulators work together with the supervisors. However, it is vital that when this appropriate approach and indicators about internal models are discussed, regulators and supervisors seek remaining close enough to the individual insurance company. If they do not understand its business model and risk profile, then this will not succeed.
Another industry representative said that it needs to be appreciated that internal models are a large company topic and that companies using internal model are the best supervised. However, these are not wholly reflective of the industry. The industry consists of many medium sized and smaller companies, and there is a need to be careful how that is dealt with. However, it is good to have a consistent regulatory structure within the Solvency II countries.