9 September 2014

Mounting data concerns

Forthcoming Solvency II reporting regulations are causing an upheaval in the way that investment data is sourced and managed. Solution providers discuss some of the hurdles insurers face and highlight classification codes and look-through as problem areas. Hardeep Dhillon reports.

Solvency II is placing a significant reporting burden on insurance companies and their asset managers by requiring unprecedented levels of transparency and far greater granularity of asset data.

The regulations mean that insurers will have to calculate the market risk of their investments and report this information on a quarterly and annual basis, which they have never had to do before. The market risk module accounts for two-thirds of the solvency capital requirement (SCR), increasing its overall sensitivity to asset-related risk.

Things become more complex when calculating market risk, because firms also have to consider its sub-modules, such as interest-rate, equity, property, spread, currency, concentration and illiquidity risk.

One fundamental challenge for many firms is that market risk has not received much scrutiny as historically insurers have focussed more attention on their liabilities. At present, many firms do not have as well developed IT systems governing their assets in order to deal with the new extensive data required for regulatory reporting.

"Insurers have not previously needed to store much of the data required for Solvency-II-reporting within their asset management systems, which have not been set up yet for the enhanced level of detailed information," notes Andreas Penzel, director of business development of German IT company Steria Mummert ISS, which recently became part of the Sopra Steria Group.

Collecting this information and the significant disclosure requirements of pillar 3 that necessitate more frequent, detailed and faster reporting are adding to the overall burden for insurers and asset managers.

To appease the regulators, insurers will need to complete the D1-D6 asset data forms of the European Insurance and Occupational Pensions Authority's (Eiopa) quantitative reporting templates (QRT), covering around 80 different data fields (see table 1 and assets templates). This all adds up to a large amount of data.

Table 1: D1-D6 asset data templates

"Data requirements for the QRTs are extensive and not just restricted to portfolio information. They also include enterprise, market and reference data, which is all housed on different IT systems either at the insurer, asset manager or external third-party organisations," says Jeremy Skaling, head of product management at data management firm Eagle Investment Systems.

Further challenges arise in attaining the granularity to enable look-through, in other words, obtaining information on the underlying securities within investment funds. Insurers will have to provide look-through on at least 80% of their invested assets.

Companies unable to collect timely detailed data will have to pay a base capital charge of 49% plus symmetric adjustment, according to Clara Yan, director at UBS Delta, a provider of risk and performance analytics.

This could be a cost worth bearing for some, as the level of investment needed to overcome the challenges of collecting, managing and enriching the data could be far more expensive.

Deadlines looming

The clock is ticking for firms to meet the reporting challenge — Eiopa's reporting deadlines commence 20 weeks after year-end 2014 for annual reporting and eight weeks after the third quarter of 2015 for the first quarterly returns.

John Dowdall, managing director of investment product information provider Silverfinch, observes that most of the large European insurers are still working on the D1 investments data report and not many are applying the D4 requirements for investment funds reporting.

"That will cause problems going into 2015," he warns. "Data issues must be sorted out now, as there is not a lot of time to contact asset managers and get the information."

Indeed, firms are only now starting to get to grips with undertaking this monumental task. Dowdall adds that discussions are growing between asset managers and insurers on exchanging data but he observes that a lot more dialogue and work is still required.

John Dowdall, SilverfinchComments from the Association of British Insurers (ABI) paint a similar picture. "The number and detail of conversations have increased with asset managers, third-party administrators, etc. but there is more to be done in most cases," says the ABI, responding to questions from a PRA report (see IAR, 25 July).

The ABI also noted that not many members have completed a gap analysis on data requirements, "but many are well advanced using what information is available."

The response from the Association of Financial Mutuals noted that many of its member firms were more focussed on pillars 1 and 2. "So gap analysis is rudimentary or high level, or has been focused on governance/actuarial issues," commented the association.

The level of preparedness also differs across continental Europe. For instance, some smaller markets, like Poland, are at a very embryonic stage because it is the first time they have been required to report.

Some markets, like Germany and Austria seem to be quite well progressed in the level of preparedness for reporting. "After a slow start into Solvency II asset managers are able to provide most, if not all, the information insurers require for each reporting date in a timely fashion or are currently working on it," says Penzel of Steria. The firm offers a regulatory reporting solution called Solvara that also calculates the standard formula SCR.

Penzel notes that insurance associations and authorities might come up with a solution to the CIC problem after active discussions with their member firms. "They most certainly are discussing algorithms that will help insurers to derive further reporting-data required from the CIC code," he says.

In Denmark, many domestic insurance companies have pressured asset managers to deliver the required data to automate the market risk calculation, says Bo Koldenborg, partner and executive VP of Danish IT company Asseco, whose SII Engine provides SCR calculations and QRT reporting.

"Many claim they can deliver this information. However, from speaking to our insurance clients it is apparent that this is not the case. When it comes to deliver, insurers have found it difficult to obtain information from their asset managers," he says.

Managing data is crucial

Data management will be essential for the SCR calculation and reporting to the supervisors, as the data content used and reported has to be consistently applied.

"This will require a great deal of back-office co-ordination between the insurance company and its fund managers, administrators and data vendors to ensure consistency and completeness," says Chris Johnson, head of product management, market data services, at HSBC Securities Services.

"Data management is central to Solvency II," adds Eagle's Skaling. "Not only must data be gathered and consolidated, it also has to be enriched to comply with the regulations."

Clara Yan, UBS DeltaThe multitude of issues facing insurers and their supplier companies range from identifying the required data and where it can be sourced to verifying that data is complete and accurate, he outlines. Also important is defining and implementing data quality and validation processes that can establish the completeness and traceability of data.

Other bones of contention can be a lack of control and ownership of data and the prevalence of redundant processes and inconsistencies in data between systems.

"Some insurers are finding it difficult to extend their data management function to deal with the new data requirements regarding the short time span until the effective date and the frequent amendments," adds Steria's Penzel.

The process of gathering investment data is difficult because it usually resides with many external parties often in different formats, says Yan. These can also include fund managers, accounting sources, third-party custody and fund administration firms, and vendors of market and reference data.

"In addition," Yan says, "the data will have to meet certain standards and be reported quarterly with an expected delivery date within five to 10 days of each quarter's end."

For those that outsource investment management, there may be some insurance-specific information that external providers may not be able to access.

"The full market risk calculation requires additional liability-specific data such as claims and premiums provisions to calculate interest-rate risk," says Asseco's Koldenborg. "Nonetheless, the security-level information will be sufficient enough to calculate the majority of the market and counterparty default risk."

Classification challenges

Firms are also facing categorisation issues since much of the reference data is new and only applicable to Solvency II. These include Eiopa's complementary identification code (CIC) that maps assets to asset types for risk analysis, the NACE, which is an industrial classification to identify the economic sector of the issuer and the Financial Stability Board's legal entity identifier (LEI), which identifies the parties of financial transactions.

The ABI's feedback to the PRA noted that codes and identifiers are the most problematic issue as not all of the data is available, such as NACE codes on unlisted securities and the LEI.

"The data vendors will not be accountable for any errors. It is the chief executives of insurance companies who need to sign off the CIC codes and make sure they are accurately assigned," Andries Beukes

While, Mark Sherwood, insurance consultant at Rimes Technologies, observes that firms with exposure to private equity and derivatives might find that some of their counterparties do not have an LEI (see IAR, 2 July).

The Investment Management Association's response to the PRA identified that insurers do not have all the information required to allocate the CIC code for some OTC derivatives because these can contain bespoke terms and conditions.

The CIC is a challenge for insurers because many do not always have the expertise and time to classify their assets into the relevant codes, especially for alternative asset classes, says Andries Beukes, director at risk management and technology firm Muller, Beukes, Edvardsen (MBE).

He says that some vendors can classify the assets but warns that insurance companies cannot wholeheartedly rely on vendors for this, as the insurer is ultimately responsible for making sure the CIC codes are correct.

"The data vendors will not be accountable for any errors. It is the chief executives of insurance companies who need to sign off the CIC codes and make sure they are accurately assigned," he says.

Skaling says that Eagle derives the CIC code from the underlying data supplied by a number of sources, such as market vendors, accounting platforms and custodians. Eagle can then create the CIC code based upon the security classification data that each client has available.

"The challenge is in providing standard mapping for all the required data since each client will have a variety of different sources," he adds.

Underlying transparency

There are numerous hurdles to trying to obtain look-through and full transparency of a fund's constituent holdings and their weightings. One challenge facing asset managers and insurers is that a large insurance firm might hold between three and six thousand funds, notes Connor Sloman, head of asset management solutions, EMEA for Morningstar.

Connor Sloman, Morningstar"That is a lot of data to process and it is a very complex data requirement," he says.

In the UK alone, IMA members report portfolio data from 2,500 funds to Morningstar, which the firm then suppresses and permissions based on each fund company's requirements.

Previously, full holdings might only have been disclosed on a quarterly basis. Solvency II requires full holdings to be analysed within the first 10 to 15 business days of each month.

"This is a big change for the industry and one that does require new processes to be put in place," he adds.

Silverfinch, MBE, Asseco, UBS Delta and Morningstar are some of the providers that offer solutions to the look-through problem.

Silverfinch sources underlying fund data from asset managers to provide to insurers. The firm automatically matches targeted fund portfolios to products listed by asset managers whenever a fund is added to an insurer's target list.

MBE's Transkap solution derives required data from a number of external third-party vendors and delivers it directly to the client. The portfolio data permissioning process could overcome any embargo period that data suppliers may have that restricts the provision of timely portfolio data. 

"Under the data permissioning process, vendors can lift the embargo period for specific clients, subject to approval from fund managers, and that enables us to provide the freshest data," says Beukes.

"It is understandable that some do not want to reveal trading positions or asset allocation strategies but full look-through can only work if the fund manager actually releases the information," Bo Koldenborg

Asseco SII Engine calculates, based on security-level detail, full look-through on funds and funds of funds and can integrate assets managers security-level data input in order to do automatic calculation of market risk and counterparty risk and to produce the list of assets QRT for the insurance clients.

UBS Delta and Morningstar launched a solution in April to deliver look-through functionality on funds as well as the SCR calculations.

"For an insurer or asset manager to build up that level of line-by-line reporting capabilities for external funds as well as any associated technology can be quite costly to implement from an operational perspective," notes Yan at UBS Delta.

Andreas Penzel, SteriaMBE's look-through analysis of the one funds of funds from a single source highlighted a number of issues. One was missing data, some of which could be supplemented by using secondary data sources. However, this would have to be undertaken manually, on a painstakingly labour-intensive line-by-line basis, says Beukes.

There were also gaps in credit ratings and ultimate parent codes.  In addition, the sub-funds had different portfolio dates from the parent fund.  All of these issues would impact the accuracy, completeness and relevance of the data to be used for calculation and reporting purposes.

Beukes adds that using assumptions could complete the data set but this could be at the cost of accuracy. "Our system allows you to plug the missing data gaps. It also identifies which fund managers are not providing up-to-date information. Insurers can then approach these firms to ask for the freshest data for calculating market risk," he says.

However some fund managers may be reluctant to give out real-time information about the assets in their funds because they do not want to compromise commercial confidentiality.

"It is understandable that some do not want to reveal trading positions or asset allocation strategies but full look-through can only work if the fund manager actually releases the information," says Asseco's Koldenborg.

Some argue that providing data required by insurance clients could become a competitive advantage for asset managers. Silverfinch's Dowdall notes that some asset managers are taking proactive steps to address their clients' issues, with a view to gaining market share and increasing assets under management.

Steria's Penzel observes a number of European firms offering information that is supplementary to pillars 1 and 3. "Some are also thinking of going beyond this and providing additional information for pillar 2," he says.

"Data management is central to Solvency II. Not only must data be gathered and consolidated, it also has to be enriched to comply with the regulations," Jeremy Skaling

Asseco's Koldenborg concurs, stating that insurers are quite annoyed about having to do manual work in regards to the SII market risk data input from the asset managers, and are starting to judge asset managers not only on their ability to generate returns on investments but also on their ability to deliver data. 

"We are still yet to see an insurance company deselect an asset manager, but it is something Scandinavian insurers, in particular, are currently discussing," he says. "Other insurers across Europe should also re-consider their best strategic partners for regulatory reporting."

Click table for larger image.