In the financial industry, we have seen an increase in data use, especially in the area of credit risk analysis. The first wave came with the introduction of Basel 2 around 2004, followed by Solvency II, where institutions have to collect and store data for the calibration of credit risk models. The second wave arrived with the introduction of standardized data definitions across jurisdictions: Loan Level Data Initiative of the ECB (2013) and Bank of England (2012), EU-wide stress test & AQR by EBA (2014), data collections by Global Credit Data (2004) and the planned European-wide credit register, AnaCredit (2016).We are in the middle of the third wave where banks have realized that good quality data could be used for better service, higher client satisfaction.
Implications of New Data Standards
Standardized data will make it much easier for market participants and supervisors to make informed judgments on credit risk instead of being dependent on market price comparison or credit ratings – they will be able to benchmark portfolios and financial institutions across entire markets. At the same time originators can benchmark their own observations or give their model a more solid basis; this is especially useful for so-called low default portfolios.
Standardizations will make data quality between originators comparable, putting more peer pressure on institutions to improve their data quality. The ECB, the supervisor in Europe, has announced they will become more data driven and the European Commission has defined new requirements on data quality for securitisation to support Simple, Transparent and Standardized (STS) securitizations. This means that financial institutions want to make data quality manageable, measurable and their employees accountable. At the same time there are many benefits for the institutions themselves – reliable, consistent management information, decrease in costs and timely risk information through immediate translation of new information into the models. They will also benefit from higher staff morale and motivation levels as there will be less data issues to resolve.
Our Solutions For Data Quality
From incorporation, OSIS has been very active in providing data quality solutions to financial institutions. We provide services to originators and investors on evaluation, validation and monitoring of data quality at loan level through our LoanQualifier™ product. We apply automatic checks on data formats, uniqueness of loan identifiers, consistencies within loans, between loans in the same pools and consecutive loan reports. We also provide warnings on strange patterns at portfolio level. We do over 1,000 checks on mortgage loans and over 1,200 on SME’s. The information we derive is translated into pool quality reports, data quality dashboards, data quality scores – we store each error in an analytical database. Originators can use the LoanQualifier™ dashboard to comply with the data quality requirements of the ECB, Bank of England, the new proposed EU regulation on transparency for ABS and the Simple-Transparent-Standardised (STS) certificate.
Data Quality Scoring
Each quarter we run over a billion checks on all the mortgage loans at the European DataWarehouse and 150 million on SME loans. The results are published and are freely available on this website under data quality scoring. The quality scores are available at transaction, originator and country level and it is possible to see which transactions meet the 95% confidence score to be STS compliant.
Each detected error on mortgages or SME loans in the European DataWarehouse is recorded in a database. Therefore the investor or any other user can analyse transactions with and without errors and compare the results.
AnaCredit: EU-wide Credit Register
The data requirements in the Loan Level Data Initiative (requirement ECB for repo funding), Anacredit (requirement ECB) and the FRY-14Q (requirement FED) are very similar and cover many data fields per loan. Therefore they can be used for multiple application like benchmarking, modelling and capital market transactions. Originators can harmonise their internal data quality processes by complying with several regulations at the same time, saving resources, money and time by using these requirements as their internal standard. At the moment only the data from the Loan Level Data Initiative are publicly available. We encourage institutions to pool and to share the Anacredit and FRY-14Q data as well in order to provide a better understanding of bank loan credit risk internally, for their stakeholders and investors.