Compliance with the Technical Actuarial Standards (TAS), the standards framework set for actuaries by the Financial Reporting Council, is compulsory for those members of the Institute and Faculty of Actuaries engaged in traditional actuarial work in the UK.
This means that actuaries carrying out Solvency II-compliant Capital Requirement calculations are subject to the requirements around (amongst others) Modelling, Data and Reporting as dictated by the standards, and are ultimately responsible for any judgement calls made on the data or application of the standard.
But, as I outlined in a previous article, An Asset to the Insurance Industry, just because the standards apply to insurance companies, it does not mean that other stakeholders in the industry do not have a role to play in helping to meet these standards.
Asset managers in particular can be a valuable resource when it comes to ensuring that the data used in Solvency II calculations is compliant with the Technical Actuarial Standards on Data (TAS D) and Reporting (TAS R).
TAS D applies in respect of all data used in preparing actuarial information which is used in reporting. The principles in TAS R cover the need for a description of the data used, its source, limitations and associated uncertainties.
Over the course of the past few years, the implementation of TRANSKAP, the MBE risk management and capital planning solution, has revealed that the collection, validation and processing of good quality look-through data is more complex and time-consuming than could previously have been imagined.
When accessing fund data, up to 16 iterations of look-through have been needed in order to access the required asset-line level of granularity tucked away under multiple layers of parent funds. TRANSKAP has revealed that, on average, a parent fund contained five unique child funds, which translates into five times the effort required to look through compared to what would originally have been anticipated.
In one specific example, 4 000 parent funds invested in by one insurer broke down into more than 23 000 child funds. Each of these funds required further drilling-down, resulting in almost 9 million non-unique asset lines, and for each of these assets a number of data fields (needed for the Solvency II Quantitative Reporting Templates) had to be furnished.
To add to the complexity presented by the sheer volume of asset data to be processed, it quickly came to light that all the fields required could not plausibly be provided by individual asset managers, despite many insurers believing that this would be the case.
Instances have occurred where up to five data providers, including specialist data vendors, have been needed in order to source and validate the full dataset.
And these were the hurdles faced before even starting to think about the Capital Requirement calculations themselves. So, given this context, what are some of the standards stipulated by TAS D and TAS R where asset managers can make the most valuable impact?
Technical Actuarial Standards on Data
Data Requirements: Whenever work is undertaken an assessment shall be made of the data required in order to deliver the actuarial information needed by the user.
This is a case where communication between insurers and asset managers is crucial. If asset managers have a clear understanding of the data required, and provide that to insurers (excluding everything that is unnecessary), both parties can eliminate a lot of waste from their processes.
This will also prevent insurers from having to make multiple requests to asset managers as they discover that not all data has been provided as expected.
Data Definitions: The definitions of all items of data shall be documented.
Asset data fields are often supplied to insurers with vague or misleading labelling and scant documentation. Asset managers best understand the data that they are providing to insurance companies, and are so best placed to define it. TAS D even states that “definitions may be documented in many ways, including documentation provided by those supplying the data…”
By providing concise, yet complete data definitions, in language consistent with that understood by insurers, asset managers will help insurers both comply with the regulation, and also understand the data they receive more thoroughly.
Validation: A set of checks shall be constructed and performed in order to determine the extent to which, taken overall, the data is sufficiently accurate, relevant and complete for users to rely on the resulting actuarial information… The checks that have been performed shall be documented.
The extent of the checks is a matter of judgement and depends on factors such as the source of the data and the extent and nature of checks carried out by other parties.
If asset managers provide data that insurance companies can be confident has been sufficiently checked and tested, Solvency II actuaries can spend less time checking data, more time modelling and analysing it.
Examples of possible data validation checks that asset managers can carry out include comparisons with data used for a previous period, and checking that data values lie within reasonable limits.
Incomplete or inaccurate data: When data that is required is materially incomplete or inadequate, an assessment shall be made to determine whether the reliability of the data can be improved by adjusting or supplementing it…The treatment of, or action taken for, incomplete or inaccurate data shall be documented.
In order to assist actuaries in their assessment of how to remedy incomplete or inaccurate data, asset managers can conduct an assessment of their own data before providing it to the insurers.
Even if they are not able to plug all of the data gaps, asset managers documenting these gaps will save actuaries a lot of time in their assessment of how to improve the raw data.
Transparency – Data: An aggregate report shall: a) describe any data or any other information used; and b) state the source of the data or other information.
If there is any material uncertainty over the accuracy of the data, an aggregate report shall: a) describe the uncertainty; and b) explain any approach taken to the uncertainty in the calculations or in the results.
As suggested above, asset managers can greatly assist in providing useful descriptions of any data supplied, and also inform insurers of any material uncertainty they uncover in their own assessment of the data.
If, for example, the asset manager is aware of a timing error in the market value fields of certain assets, they could point out this very specific information to the insurer upon receipt of the data. This will save the time of the insurer (in trying to understand the error), the asset manager (in having to answer queries which could have been pre-empted) and it will mitigate the risk of the insurer not discovering the error at all and potentially reporting incorrect results.
There are many ways in which asset managers can use their unique access to and insights into data to add value to insurers’ capital calculations and regulatory compliance. By being aware of the standards to which their clients are subject, asset managers can focus their energy on delivering data of adequate quality, so that the actuaries can get on with the modelling.