AnaCredit - Implementation challenges and lessons for the future stages

Background

AnaCredit (analytical credit dataset) is a three-stage project that was launched by the European Central Bank (ECB). In April 2014, ECB had announced the establishment of a central credit register that would be updated with granular credit (individual loan level), counterparty, and credit risk/exposure data of all of the credit institutions and other financial institutions (FIs) that provide loan within the Eurozone - harmonized across all of the member states. Eurozone states have obligation to participate, while other member states of the European Union could participate voluntarily.

The AnaCredit regulation was finalized by the ECB's Governing Council in 2016. The data collection under AnaCredit for stage 1 is scheduled to start in September 2018. From 30th September 2018, all concerned reporting agents (credit institutions and other FIs that provide loan within jurisdictions in scope) would need to provide standardized, specific, exhaustive, and frequent credit and counterparty related reporting to ECB via their national central banks (NCBs). The scope of stage 1 of AnaCredit reporting entail the credits provided by reporting agents to legal entities (and not natural persons). The three-stage regulation is further expected to come into full effect in 2020.

The purpose of AnaCredit is to support: a) monetary policy analysis & operations, b) risk management, and c) financial stability surveillance. The granular credit data under AnaCredit is expected to be used for broadly monitoring the performance of entire Eurozone credit market. AnaCredit would massively support the macro-level credit analysis tasks of ECB and the NCBs of European System of Central Bank (ESCB).

Implementation challenges

Implementing AnaCredit regulation has, however, proved to be quite challenging for the reporting agents. Following are some of the key challenges faced by FIs.

Data availability: AnaCredit require FIs to report huge amount of data and at high frequency – monthly, quarterly and on-change basis. Concerned FIs need to report around 100 data points for each credit exposure in scope and at a loan-by-loan level. Some of the data points to be reported include instrument (instrument type, non-performing status, currency, payment frequency, interest rate type, etc.), counterparty (LEI code, address, balance sheet total, # employees, etc.), collateral (protection type, location of real-estate collateral, original protection value, etc.), and accounting related data (accumulated impairment amount, encumbrance source, forbearance & renegotiation status, etc.).  

Good number of FIs have however traditionally lacked the availability of many of these required data attributes. For example, data attributes related to legal entity borrowers (such as # of employees, classification of enterprise size, and fair value changes owing to credit risk) have been lacking at most FIs. For many other FIs, data attributes such as SME indicator and LEI code are also unavailable.  Large number of FIs have therefore been having to plan the gathering of counterparty data points (such as # of employees, annual turnover, or balance sheet total) from external sources such as company information bureaus and credit registration offices.

Many FIs' legacy systems and data stores have lacked capability to capture data at the required level of granularity. Even the data that is available within FIs is scattered in myriad siloed databases (including PDF loan contracts and Excel sheets). Consequently, many large FIs have been needing to source data attributes for AnaCredit from over 10s of databases – in some cases even 30! Compiling data from their various siloed source systems –lending, risk management, accounting, reporting etc. - and aggregating and reconciling these at the required level of granularity has been quite challenging for FIs.

Data quality: Unlike the other regulatory reporting mandates that require data at aggregate level (e.g. total loan portfolio value), AnaCredit regulation mandates reporting at a granular level and on loan-by-loan basis. For AnaCredit reporting therefore, FIs don’t have the luxury of making manual adjustments at aggregated value level (to take care of data quality issues). Unfortunately though, the quality of data (e.g. collateral data and client financial indicator) remains questionable in many FIs.  

AnaCredit expects each individual exposure and counterparty data to be accurate, attribute by attribute. Unfortunately however, owing to the sheer volume of data involved, certifying the completeness and accuracy of these data remain a key challenge for FIs. Compounding the challenge for FIs is the fact that to ensure data quality, ECB has defined over 2,000 validation rules for completeness, consistency, and referential integrity. Many NCBs too have added additional validation checks. Most FIs have been facing massive challenges in ensuring compliance with such large number of validation rules. 

IT & process: Many FIs’ existing processes and IT systems lack capability to facilitate data collection at the required level of granularity, and are unable to fulfil the intense data validation and reconciliation requirements of the AnaCredit regulation. AnaCredit also expects reporting on legal entity (as opposed to per business) basis. Many FIs’ existing systems however are capable of reporting only on per business basis.

Enabling robust and flexible processes and IT infrastructure for meeting AnaCredit regulation’s frequent and detailed reporting requirements in a timely and error-proof manner has been quite challenging for FIs. FIs, for example, have had to revamp their client on-boarding process and the associated IT system to ensure all required data are captured. FIs’ data warehouse, statistical/supervisory reporting, and collateral data administration systems have required massive changes for enabling AnaCredit compliance. Many other processes and IT systems too have needed significant changes – including loan origination, loan administration/booking, credit risk management (including modelling, underwriting, monitoring and reporting), client administration, and accounting systems.

Multi-jurisdiction: AnaCredit regulation provides flexibility of national discretion; some of the requirements can be detailed by NCBs for their corresponding jurisdiction – for example requirements related to data collection strategy, mandatory attributes, submission timelines, extension of loans/clients in scope, derogations and facilitations etc. This has resulted in significant requirements variation across countries; thereby substantially increasing the challenges for internationally active FIs. FIs that have businesses in more than one Eurozone country are needing to implement specific AnaCredit requirements for several concerned national discretions. Also, these FIs would need to furnish report to many NCBs, as applicable.

Implementation cost: AnaCredit implementation has proved to be quite expensive for many FIs – especially for the internationally active ones. Complexities related to variety and heterogeneity of products, IT systems, and regions has added to the increased cost for FIs. For many FIs, implementation cost has been upward of €10 million. Data warehouse and data reporting solutions development (or purchase), data sourcing, IT systems enhancements to capture the required data at granular level, and process re-design and testing have been some of the key cost contributors for FIs. 

Lessons for the future

In order to effectively overcome the aforementioned challenges, FIs should take certain corrective actions for stage 2 and 3 implementation of AnaCredit. Following are few recommendations in this regard. FIs can  leverage these recommendations, to the extent possible, for their current AnaCredit stage 1 implementation as well – time permitting, or perhaps as part of FIs’ ongoing improvements plan post the stage 1 go-live date.

Strategic approach: FIs need to have long-term perspective and take a strategic approach for AnaCredit implementation. Given that there are certain similarities of data needs between AnaCredit and few other regulatory reporting requirements, FIs should consider implementing a single robust, integrated, scalable and strategic reporting platform. Such a platform would be capable of fulfilling FIs’ myriad supervisory, statistical and regulatory reporting requirements. It would enable them with granular data drill-down capability, and also fulfil their multi-jurisdictional reporting requirements. The platform would further help increase reporting efficiency, effectiveness and consistency, and avoid redundant activities related to loading/processing of same data numerous times.

FIs should also leverage their AnaCredit implementation efforts towards creating effective bridge between their risk, finance and regulation functions. FIs’ endeavor should be to utilize the rich AnaCredit infrastructure investments for their additional key business imperatives: such as for internal risk management, for gaining valuable business intelligence and insights, and for strategic decision making. As an example, the insights gained from FI’s AnaCredit reports can be leveraged for strategic lending portfolio analysis/management, and for risk-based lending business decision making.

Smart IT, process and data reengineering: FIs need to smartly revamp and optimize their internal processes, data infrastructure and IT systems. They should scale-up and transform their existing credit workflow, processes and systems - origination, collection, monitoring, provisioning, recovery etc. Data warehouse, collateral data administration, and supervisory/ statistical reporting systems also need to be thoroughly reviewed and revamped. Client and loan administration systems, credit risk reporting, and accounting system are other key systems that would need smart reengineering.

FIs should also work towards enabling analytics capability on top of their AnaCredit solution. This will help them gain rich insights, and enable prompt data analysis and quality checks. FIs can also consider leveraging grid and in-memory computing capabilities for concerned processes (e.g. mapping and calculations). This would provide them with the required scalability and performance for processing large volume of data for AnaCredit reporting.

FIs should focus on enabling robust data aggregation and normalization capabilities. They should work towards establishing single golden data source (for their various regulatory, risk, finance etc. reporting). Towards this, they should leverage robust integrated and universal data lake / warehouse - which in turn would have data captured from various source systems. All reporting templates would be populated from this universal data lake / warehouse.

Ensuring consistency amongst individual data attributes is crucial. For this, specific interdependencies and relationships between individual data attributes need to be thoroughly considered. Further, over and above the ECB defined validation rules, FIs can implement additional checks throughout their internal workflow to ensure consistency, completeness and integrity of reports. For example, automatic checks to ensure that the reported amount in the AnaCredit report is not larger than the corresponding balance sheet value; and automated controls for reconciling the AnaCredit data with FinRep, CoRep and other statistical reports.

Robust tools: FIs need to implement robust tools for providing complete data lineage. This is crucial from auditability perspective and for allowing users to drill down data from aggregate AnaCredit report to the underlying details of the contract. FIs could work towards thoroughly specifying the AnaCredit data requirements via standard API into the AnaCredit solution data model.

Robust data quality assurance tools should also be implemented - for example, to allow users to conduct variance analysis at different granularity levels, and to slice and dice the data per various attribute types (instrument, country, industry etc.) FIs should implement smart and flexible dashboards for data quality validation, and for interactive data visualization. These dashboards should enable drill down capabilities - starting with aggregated view to portfolio view and until single loan view. The dashboards could be structured along various dimensions - e.g. organizational unit, recipient groups etc.

Reuse: FIs should optimally reuse their data and reporting infrastructure. This would enhance efficiency and also enable consistency in reporting across multiple regulatory reporting mandates. So for example, most banks, should be able to reuse significant amount of data and their current IT infrastructure related to COREP/FINREP, BCBS 239, IFRS, and other statistical reporting for their AnaCredit implementation. Quite a few banks should also be able to reuse significant volume of their data and IT infrastructure related FATCA, AQR, MiFID/MiFIR and EMIR.

For optimal reuse, it is important that FIs gain a thorough view of their available data, and a complete understanding of IT synergies and overlaps amongst their myriad regulatory reporting requirements.

Conclusion

It is beyond doubt that AnaCredit implementation has been posing significant challenges for the concerned FIs. To overcome these challenges, FIs should carefully consider the overall end state of AnaCredit (not just stage 1), holistically rethink their implementation approach, and take corrective actions, as appropriate.

By doing so, FIs would gain immense business benefits – not only with regards to ensuring AnaCredit compliance, but they would also achieve significant improvements in their overall data management and reporting process efficiency and effectiveness, and the strengthening of their credit risk management and lending processes.