As insurers submit their first Solvency II reports this month, Christopher Cundy asks what supervisors will do with the billions of data points now at their fingertips
It is an astonishing statistic: each year, because of Solvency II, UK insurers will collectively submit some 650 million items of information to the regulator. Scale that figure up across Europe and the total reaches well into the billions.
As insurers battle to get their first quarterly reports delivered this month, it begs the question: what are supervisors going to do with all that information?
The volume of data coming into supervisors is unparalleled. "The new regime has not only raised the bar for undertakings but for national competent authorities (NCAs) as well," comments Frank Grund, director of insurance supervision at Germany's regulator BaFin.
The UK's Prudential Regulation Authority (PRA) estimates that the volume is five times the quantity collected under the banking regulation CRD IV - although that's partly because the PRA supervises more insurers than banks.
"We expect to perform more assessments and analyses, including more comparisons between undertakings." Frank Grund, BaFin
Most supervisors have had to invest heavily to develop or adapt systems that can cope with the flow and format (data needs to be submitted in extensible business resource language, XBRL).
Jan Parner, deputy general director at Danish regulator Finanstilsynet, says its reporting platform was developed with the central bank. "Originally it was built for Solvency II, but before it came into life it dealt with the core submissions of the CRD IV directive. It means we already have the tools to receive the XBRL data, put it in the data warehouse and perform analysis on it."
For the PRA, a sum in the region of £18m ($26m) has been spent on IT infrastructure to create the Bank of England Electronic Data Submission (Beeds) portal.
But building the mechanics is only part of the story. The next chapter involves using the inundation of information to provide value for both supervisors and insurers.
PRA's special data team
The PRA's data analytics team of 18 was established in December, corralling staff from other PRA departments with actuarial, supervisory, economics and finance experience. The team does not have direct contact with insurers, but provides four functions: working with the IT team to ensure data integrity; producing tools to help supervisors; performing analysis on firms; and providing sector-level analysis.
As data flows into the system, it undergoes validation and plausibility checks. Spotting mistakes may be relatively straightforward, but there is another important step in ensuring consistency: checking whether firms have interpreted the data items in the same way, something that can only be assessed once the first submissions come in.
The team will also be building toolkits to help supervisors interrogate firms and provide information to other parts of the Bank. The three main tools in development are one to help understand the different basis between the old and new regulatory regime; another that supervisors will use to understand firms' quarterly and annual reports; and a peer comparison tool.
Developing new tools
BaFin's Grund notes how the market-consistent valuation and the risk-sensitive capital requirements introduced by Solvency II make insurers' financial position more susceptible to changes in external factors.
"Consequently, not only undertakings but also NCAs need to be more alert to incipient negative trends and developments," he says. "We expect to perform more assessments and analyses, including more comparisons between undertakings, with the quantitative data we receive, identifying new key risk indicators and early warning signs to react early and fast if necessary."
"There is so much data on Solvency II, we cannot use it all on day one." Jan Parner, Finanstilsynet
However, all three regulators agree that developing analytical tools and indicators will take some time. "There is so much data on Solvency II, we cannot use it all on day one," says Parner. "The priority is to be able to do the same thing we do today with new data. The second step is to make use of data that we did not receive before."
The PRA has adopted an agile approach to IT development, so new tools and analytics can be brought into play relatively quickly. Most of the current work is undertaken in Excel and VBA, along with software such as PowerBI, PowerPivot and Tableau that enable the handling of large datasets.
Asset data
The biggest revolution in Solvency II is the level of detail to which assets are reported, which goes way beyond any previous experience.
BaFin's Grund notes that the abolition of quantitative investment limits and the introduction of the prudent person principle means insurers have greater responsibility for their investment activities. Solvency II also happens to have arrived at a time in the economic cycle when yields are low and regulators are watching for insurers taking excessive investment risk in a bid to maintain returns.
"To assess whether undertakings comply with the qualitative investment requirements at all times, NCAs need to be able to drill down into the asset portfolio of individual undertakings, benchmark the asset allocation and to identify shifts to more risky asset classes over time," he says.
The Bank of England has previously identified pro-cyclicality of investment as a possible risk and the new details revealed in Solvency II data will allow the PRA to monitor and model that risk.
Parner says the billions of Solvency II data points should help avoid the situation seen in the banking sector where liquidity became a major issue during the financial crisis, but nobody had good data on liquidity.
"We have not identified any data points where the same would happen for insurers. But we have to develop more supervisory tools to track the asset side, and we have to do some monitoring on how own funds are structured," he adds.
Solvency II will not allow supervisors to get the whole picture, however. Currently there is an issue with coding of unlisted assets, where there are no industry-standard datasets that recognise what the assets are. Asking firms to provide this information is expected to accelerate the development of data standards, but it will take time to work through.
Extra data
Beyond the official Solvency II data request, most EU countries require insurers to submit extra data on national-specific templates (NSTs). The UK has 13 NSTs, although two are only for Lloyd's syndicates, gathering detail on cash generation, for example.
There is also an opportunity to bring in external data to perform additional cross-checks and analysis – for example, on whether a shift in portfolio composition is reflective of the wider market.
And as the industry is increasingly aware, Solvency II data templates will not remain static and new data requests will be made to meet whatever objective the supervisor has.
Micro to macro
One of the most powerful outcomes from the data exercise is the ability to monitor macroprudential risks with company-level activity.
Grund explains that Solvency II data will give supervisors better opportunities to calculate sensitivities and analyse the impact of economic stress scenarios.
"Supervisory authorities should take this chance to strengthen their macroprudential view," he says. "Given the current low interest rate environment, it will become increasingly important for supervisory authorities to monitor macroprudential trends. BaFin will carefully select important themes, where it will engage in in-depth discussions with the undertakings. BaFin has set up a series of standard reports to fulfil its role in this regard and will refine them as it gains further experience with Solvency II."
"Given the current low interest rate environment, it will become increasingly important for supervisory authorities to monitor macroprudential trends."
Cooperation with industry
Another question on the lips of insurers is will all this work bring any benefits for them?
The PRA's industrial liaison group on pillar 3 data, which has hitherto focused on the process for submitting data, is to morph in an effort to collaborate on the use of data. The goal is for the industry to start using data for its own insights, although decisions to share the PRA's data sets – even in aggregated and anonymised formats – will be not be taken lightly.
Supervisors will also be collaborating with each other and via the European Insurance and Occupational Pensions Authority (Eiopa) to share best practice. Eiopa has launched a workstream on the possible use of a Solvency II data for additional analysis and indicators to support supervision, and the data will feed into other aspects such as benchmarking studies on internal models and studies on consistency between supervisors (IERM, 21 April, Eiopa builds its toolkit for Solvency II convergence)
Big data has always held the promise of big insights and in the case of insurance supervisors this seems particularly true. Insurers should expect more questions from supervisors and they would do well to keep an eye on the regulatory metrics being developed. The new data will also allow better peer comparison than ever before, so they should be prepared to explain if their business is found to be an outlier on certain metrics.
Channels:Solvency II