Regulatory requirements as new challenges
In January 2013, the Basel Committee introduced the new regulation on supervision No. 239 (Basel Committee on Banking Supervision, BCBS No. 239). The lessons of the financial crisis of 2008 led to the disappointing conclusions, that the banking IT-architecture and data models are inadequate and unable to support decision-making processes at the level of board members.
Many banks are unable to quickly identify concentration risks at the group level, by business, subsidiaries, and markets. Several large banks were not able to manage these risks, because they could not adequately aggregate them. This has led to destructive consequences, both for the banks themselves and for global financial markets.
That is why the Basel Committee, which includes representatives of the world’s largest banks and regulators, adopted 14 principles of “Effective Risk Data Aggregation and Risk Reporting” as industry standards. Since 2013 European banking supervisors evaluate and verify data management processes, the IT architecture and the IT infrastructure of banks, as part of the annual audit.
The implementation of IFRS also had a significant impact on data management processes and the IT-architecture. By the end of 2017, banks had to introduce new approaches to the calculation of reserves (International Financial Reporting Standards, IFRS9). This standard affects not only a change in the methodology of classification, recognition and subsequent measurement of financial assets, but also requires banks to change the data model. Banks should, at the level of source systems and data warehouse, be able to record default events, revision of contract terms, change in cash flow characteristics.
The European Banking Authority (EBA) puts forward even stricter disclosure requirements for European banks. Since January 2014, Capital Requirements Directive IV (CRD IV) was launched, which covers two areas of reporting:
— COREP, which is introduced to increase transparency by providing detailed transaction information;
— FINREP, that have been introduced to harmonize the regulatory finance reporting requirements of credit institutions, who provide consolidated IFRS reporting.
The European regulator has moved from reporting in the form of flat reports (Template-based) to reporting based on cubes (Cube-based). This reporting is a set of related tables with detailed transaction attributes and standardized dictionaries. Reporting is based on the EBA Data Point Model (DPM).
Data Governance as a new process
One of the difficult issues, when introducing data warehouse, that I had to face both in Europe and the CIS, is the consolidation of responsibility for data at the level of specific business users. To address these issues of Data Governance, Erste Group Bank AG, for example, has created a separate BIC unit (Business Information Center), which deals exclusively with Data Governance and Data Quality. This unit coordinates of those, who arrange data entry and who own data in the systems (Data Owners), monitors quality indicators (Data Quality Managers), organizes and participates in user testing (UAT testers).
Another function of this unit is the organization and on-boarding of Data Stewards. Data stewards are specialists, who define requirements and solve data quality incidents within their business unit. Data stewards work closely with dedicated methodologists or business experts, who prepare business requirements for reports and analytical applications, as well as for business terms, reference books, mapping and calculations.
Elements of the Data Governance process coordinated by the Business Information Center include:
— Assignment of responsibility over data in the bank (Data Ownership);
— Coordinating the work of data stewards (Data Stewardship);
— Harmonization and support of business terms (Business Glossary);
— Data Quality Control and Quality Indicators (DQI);
— User Testing Organization (UAT).
Many banks have also established a Group Information Data Management (GIDM) division, that closes the following Data Governance components:
— Business analysis (Business Requirements Analysis);
— Modeling and support of the conceptual and logical model of data storage (Banking Data Model);
— Description of calculation algorithms (Calculations) for developers;
— Support for reference books and mapping (References and Business Metadata).
Other necessary components of the Data Governance process are related to the daily operational loading of the storage and the formation of BI reports. The Information Management division is responsible for them.
All three departments are located outside the IT vertical and are subordinate to the top manager, who, in fact, acts as the Chief Data Officer. Thus, Data Governance processes are placed in a separate vertical line, outside of businesses and outside of IT. IT service supports only the technical components of the process:
— Creation and support of IT-infrastructure (server hardware administrators, channels, databases);
— Creation and support of IT-architecture (software architecture, system integration);
— Software administration (ETL, BI, DWH, metadata, dictionaries, etc.);
— Creation and support of data storage processes;
— Creation and support of information security processes (access rights, information protection).
New role of Data Governance officer
A key element of Data Governance is the creation of a unified data architecture in the bank’s IT systems: from front- and back-office systems to data warehouse, analytical applications and reporting systems. A single data architecture is necessary to provide external and internal users (regulators, board, middle managers) with complete, reliable and comparable information. To solve this problem at the board level, it is necessary to approve the bank’s business strategy and business model, and to identify reporting segments, channels, and products.
Based on the business model, the IT architecture is built, as well as the logical and physical data models of the data warehouse are determined. Further, at the level of the entire bank, it is necessary to ensure the consistency of the logical and physical data model of the data warehouse and data models of software applications. If necessary, CDO has to initiate the refinement or replacement of source systems to eliminate data gaps, as well as defined all elements of data governance: the requirements of data warehouse model, metadata and the data quality.
To address these challenges, a role of Chief Data Officer has emerged in large Western banks at the board level. This board member is responsible for defining the bank’s business model, data warehouse model, analytical systems and source systems data models. This issue is outside the area of IT and business units. The role of Chief Data Officer has already become the industry standard in Europe. Some of CIS banks have just recently started to introduce this function into everyday practice.
Addressing data quality issues often requires an escalation of incidents at the board level. Therefore, the corresponding Data Management Committees are created in Western banks headed by the Chief Data Officer. A typical organizational structure of a Western bank data management is presented below.
Organizational structure of data governance
The processes of creating a business model, data model and Data Governance have became the standard in Western banks. In the data warehouse project at Erste Bank, I was struck by the fact, that the entire bank, from board members to ordinary employees, speaks a modeling language, the language of attributes. Moreover, the main goal of Erste Bank for the coming years is not related to improving customer service, it is related to Data Excellence. Bonuses of the bank’s top management are directly linked to data quality indicators (DQI).
What to do banks?
In my opinion, banks in CIS will move towards the creation of similar functions and processes. Moreover, the requirements for Data Governance have already been fixed by international regulators at the level of the industry standard in order to minimize systemic risks in the financial markets. It means, international banks also obliged to face this issues.
However, the main factor will be competition for customers and challenge the new Big Data technologies. Those banking institutions, that are able to organize own processing and data management to effectively analyze customer information at data warehouses, including from social networks, will certainly create a long-term competitive advantage and win the race of leaders.
But apparently, for this they will have to enter improve their own data strategy and processes as well as to enter the IT market in order to establish corporate control and their own rules of the game there. Otherwise, IT-players will sooner or later enter the financial services market.