Bad European Data Validation Rules Give Banks Reporting Jitters

It’s hard enough for bank operations and financial reporting specialists to fill out all the data points and formats required by regulators in filing mandatory documentation evidencing financial soundness.

But when a major banking regulatory watchdog — the European Banking Authority — issues 290 incorrect data validation rules, the whole reporting process can go awry and lead to a lot of unnecessary angst.

“Data submitted in accordance with implementing the technical standards should not be validated against the published set of incorrect rules,” said the EBA in a press statement issued on April 16. The 290 incorrect rules represent about a quarter of the total number of validation rules necessary for so-called COREP and liquidity reports, due at the end of June.

Those rules explain logical relationships between data points and help ensure the data provided is accurate and the data fields are correctly completed. Part of new regulatory requirements designed to evaluate a bank’s financial soundness and potential systemic risk, the COREP (short for common reporting) and liquidity reports provide granular and uniformly structured details on market, credit, operational and liquidity risk including large exposures and leverage ratios. The reports must be formatted using Extensible Business Reporting Language (XBRL) taxonomy.

In the case of credit risk alone, the number of data points required are more than double those required by the UK’s Financial Conduct Authority, according to UK risk specialists. “There is a significant increase in the amount of information required and changes in the way data must be formatted from previous reporting regimes,” says Tony Sarkis, director of account management for Europe, the Middle East and Africa for Axiom SL, a regulatory reporting software firm. “Combined, there will be a lot more work for risk, financial reporting, operations and IT staffers.”

If these professional specialties didn’t work closely before, they will now, so they can gather and validate information from a multitude of internal risk, portfolio accounting, and security masterfiles. Reconciling inconsistencies could prove to be the most challenging task, as the data formats and content are often proprietary to each application.

With a good chunk of the EBA’s data validation rules now proven wrong, risk, finance, operations and data specialists at some of the world’s largest banks must redo some or all of the work they put into filling out and formatting  the reports. None of the ten European middle and back office operations specialists contacted by FinOps Report would predict just how much extra time they will spend, but they are clearly annoyed the EBA didn’t catch the errors in data validation rules before they were issued. “Multiple departments now have to go over each of the preliminary reports, correct any erroneous data, and retag the information all over again,” bemoans one operations expert at a UK bank.

The good news: the COREP and liquidity reports aren’t  due until the end of June, a month or so after the original timetable. It is unclear whether the postponement is related to the invalid data validation rules, but the timing it too coincidental to be ignored. Had the EBA not delayed the deadlines it is possible that some financial firms still trying to implement reporting software would have been in deep trouble. They would not have enough time to identify, test, and implement software within the same calendar month the reporting was due.

Now comes the bad news: in acknowledging it issued the incorrect data validation rules for banks to use when completing the reports, the EBA says it will issue new data validation rules before June 30. It just won’t disclose when, and even so there may be far more than the already-identified 290 mistaken data validation rules. That leaves banks — and their software providers — in a lurch. They have to wait until the eleventh hour to redo their reports, and possibly to even know what they have to redo. “Axiom leverages the EBA’s taxonomy and current validation rules and as such are in the same boat as the wider market,” says Sarkis. “As with all changes, we would expect to do a full regression test before any release to the market.”

Just why the EBA didn’t catch the data validation rules earlier is unclear. Naturally, no one wants to publicly criticize a regulatory body, but financial reporting software specialists and banks insist that the EBA should have tested the rules with real data, before releasing them to the market. The EBA did have the technical rules — a draft data point model and draft XBRL taxonomy, including the data validation rules — available last July after public hearings and comments. The early timeframe, it now tells FinOps, was designed to allow “the maximum available time” for banks to implement the new reporting framework.

That sounds great, except for one major glitch, which the EBA insists was unavoidable.  “The rules were not tested with real data, since they were developed in parallel with the underlying regulatory framework and real data was not available,” says the regulatory authority in a statement issued to FinOps. “Banks had to focus first on compiling the data from their systems to calculate capital requirements and only after that on the reporting. Therefore, there have been more observations [of errors] when banks advanced their implementation.”  The statement went on to say that the EBA expects more questions to be raised after the first reports are submitted and “stands ready to review the first set of rules.”

Presuming they have gathered and standardized all the necessary data to file the regulatory reports, banks can either do the data validation in house or pick a third party reporting specialist to do so. And hope it gets the job done right. The bank is on the hook for any reporting and tagging errors so it will presumably have given itself sufficient time to confirm to the XBRL taxonomy. If the US is any example, European financial firms have plenty of reason to worry. In the three three-year phase-in of XBRL financial reporting required by the Securities and Exchange Commission, there were plenty of technical headaches for filing firms and reportedly plenty of errors.

The EBA’s data validation mistakes show just what can go wrong when adequate planning doesn’t take place. The EBA might give banks a break by not fining them for mistakes in the first round of reports, but they won’t be forgiving forever. The new requirements require adequate technology, data models and governance policies. Banks will need to rise to the occasion quickly. When it comes to highly programmed reporting, half-baked planning won’t cut it.