Changes in reporting, data collection and decision support systems. 1

Transcription

1 Marko Kranjec Changes in reporting, data collection and decision support systems. 1 (The case of Slovenia) 1. Introduction This presentation describes the statistical reporting system used by the Bank of Slovenia in its performance of statistical, supervisory and monetary policy functions, and explains the reforms during the run-up to the membership of Slovenia in the European Union, and the subsequent adoption of the Euro. Central bank s experience could be useful for other countries that will ultimately adopt the Euro. The legally defined functions of any central bank need proper information support, regardless of whether the central bank runs its independent monetary policy or contributes to the monetary policy of the monetary union. Nowadays central banks usually take care of the monetary and financial stability, payment systems, and, often, also of the banking and/or wider financial supervision. The functional and institutional arrangements of these functions in practice vary among countries, even within the EMU. Even though the functions of the central banks in general do not change substantially, their activities nevertheless have to be restructured and adapted to external changes such as technical development, financial innovations or financial and economic crisis. In such cases the information support can not remain unchanged, what certainly affects the statistical reporting. This review introduces a specific case of development of the statistical reporting for the central bank functions in the contemporary environment and offers some consequential general considerations. 2. A need for change in statistical reporting 1 I am grateful to Mr. Janez Fabijan and Mr. Dusan Murn for their assistance in the preparation of this presentation. 1

2 Accession negotiations for EU membership, conducted in year 2000, demonstrated the need that Banka Slovenije (hence: BS), as a central bank of future member country, harmonises statistical data collection and reporting with the requirements of the ECB. Traditional reporting of credit institutions to the central bank had to be adapted, changed where needed, and harmonised with statistical requirements and definitions in the ECB regulations and guidelines 2. It soon became obvious that the existing reporting scheme could not be adjusted or partly reformed, so that the main question became in what direction to develop a new reporting system. 2.1 Initial position and requirements For development of a new reporting system it has been most important that BS retains a tradition and positive experience with a common balance sheet reporting of credit institutions data, which were used for multiple purposes, notably for prudential supervision, and statistical and analytical purposes. Another important element, decisive for the design of the new system, was that the previous reporting scheme was based on accounting data of good quality and used by reporting agents (banks) for their business decisions. BS was aware of this fact, and consequently wished to retain this connection with business accounting recordings also in the new reporting scheme. However, the limited structure of the common report system has been assessed as an obstacle for the harmonisation of financial statistics with EU requirements. In spite of relatively detailed balance sheet data that had been collected until 2001 for many purposes, it became evident that its limited structure was not suitable for systematic expansion of the common report, and has required the compilation and submission of many partial additional reports. Most important, it was not possible to expand and augment the details of the most interesting and policy relevant reporting items, such as loans and deposits; neither was it possible to introduce more detailed breakdowns. Another problem was that changes in reporting schemes were rather complicated, time consuming and costly for reporting as well as for receiving 2 Regulation ECB/2001/13, Regulation ECB/2001/18, Guideline ECB/2002/5, Guideline ECB/2003/2 2

3 entities (BS). It was therefore decided that the new solution should allow more flexibility in its dimensions and easier subsequent changes. Therefore a solution had to be found for (a) a common multipurpose report, (b) a report which would be accounting-based, and (c) a report that would have flexible structure. 3. Development The decision to start a project to develop a new reporting scheme was adopted by the BS in From the very beginning reporting entities were heavily involved in order to find mutually satisfactory solutions that would be simultaneously effective and cost efficient. New reporting framework of common data collection should serve different purposes, particularly the need for timely and accurate information for supervisory and monetary policy needs. The reporting entities have been frequently consulted during the development stage and have actively supported the most important decisions. There were two major types of problems that have affected the course of the project: the need for conceptual reconciliation; combining conceptually different reporting requirements for e.g. for statistical and supervisory needs can be achieved either through alignment of different concepts that would satisfy all users of reports or through more detailed granulation; to achieve reconciliation, an extensive process of coordinationamong different services was needed, Timing; changes in supervisory and statistical requirements occurred through time without consideration of previous changes in "counterparty" reporting (e.g. statistical requirements were defined before supervisory changes such as IAS, FINREP, COREP or Basel II very finalized). 3.1 Main characteristics of common reporting system Thorough preliminary analyses showed that the matrix reporting scheme was the most appropriate solution, given the initial constraints regarding data needs and technical possibilities of reporters. The agreement between BS and reporting entities on this question was the most important milestone of the whole project. 3

4 In the matrix system all related data needs (such as those linked to a balance sheet) are: organised in a common reporting scheme for more purposes, and use the combination of reporting items (rows = economic categories) and standard common code lists (columns = required attributes), and are systematically envisaged in multidimensional space but presented in simplified way as 2-dimensional matrices. In such a system, the logical output (in technical sense) of the reporters would be a flat table with data records extracted directly from data warehouse system. The reporting items are (in case of the balance sheet) defined as the balance sheet items depending on the multiple purposes of reporting, such as, for example, statistical and/or supervisory. Multipurpose reporting demands reconciliation between all included data needs. This means, in practice, the reconciliation among supervisory/ financial stability, statistical and analytical requirements, where the statistical requirements are further diversified (for the purposes of monetary and financial statistics, external statistics and financial accounts). Special attention has been devoted to accounting base of reporting scheme of the balance sheet report. The chart of accounts or general ledger for credit institutions serves as a backbone of the matrix reporting system. Due to specific circumstances at the beginning of matrix reporting, a new reporting scheme has been merely connected with the existing accounting report. Standard code list is a breakdown of a specific attribute (such as for example country of residency of the counterparty or original maturity) for one or for several purposes. It is important that breakdowns for different purposes (defined by attributes) are not mixed together in one code list, but separated in different ones. If different purposes (such as for example for statistical and supervisory purposes) are combined in one code list, the breakdowns are normally more detailed (granulised). Breakdowns of the same kind for different purposes have to be commonly defined, which also requires reconciliation efforts. The experience furthermore shows that exceptions are to be avoided in reporting, as they complicate matters. 4

5 Number of codes in the standard code list is, in principle, not limited and nor is the number of reporting items. This gives to the system the desired flexibility and is, consequently, at least technically, easy to introduce subsequent changes in reporting. A drawback of a multipurpose reporting system is that no cutting-off-the-tail or sampling reporting is possible because the supervisory requirements demand the data of all licensed reporting entities. 3 The balance sheet related data (fact tables), transactions and revaluations, and other data (e.g. interest rates) which are collected in a common report are all reported with a single, i.e. monthly frequency which makes reporting and especially potential revisions much easier, compared to the combination of the monthly and quarterly requests. While reception and control of the matrix reports are organised in the statistics department, the data are also used for other purposes. Therefore automatic built-in controls at the reception phase assure that the data can be used for all functional needs of the central bank, particularly for supervision and analyses. The matrix reporting system started with the narrow reporting scheme and was upgraded later by its wider version. Gradual upgrading was fully supported by reporting agents as it enabled them to develop technical solutions at appropriate pace and moderate costs. 4. Project implementation Work on a project started in March Project s objective was to define conceptual and technical framework to ensure balance sheet reporting for the needs of monetary and banking statistics and other needs of BS. Joint balance sheet reporting scheme explicitly aimed at integration of the statistical requirements of ECB/2001/13, ECB/2001/18 (and later 2003/2) with the needs of banking supervision, the needs of monetary policy at the time of own currency, and other needs (e.g. financial stability). 3 These possibilities are offered in the statistical regulations on balance sheet items (ECB/2001/13) and interest rates (ECB/2001/18). 5

6 During project implementation phase the political decision was taken that Slovenia would join the euro area as soon as possible. This required very rapid harmonisation of the measurement of all macroeconomic aggregates, particularly those linked to the criteria for adopting the Euro. With the aim of ensuring price stability, the central bank began implementing a monetary policy with interest rates as the main policy instrument. This required the harmonisation of statistics on interest rates in accordance with Regulation ECB/2001/18. Bank of Slovenia prepared sample reporting scheme for eight largest banks (MFI-s) as a predecessor to the comprehensive reporting of all MFIs. The introduction of the new concept of matrix reporting was followed by necessary harmonisation processes to support monetary policy decision making. Simultaneously, information provided by reporters also included the needs of the supervisory function. The degree of detail was fixed at the lowest possible level, usually taken from contract documentation and from relations of banks with individual clients. Such a breakdown of data also facilitated the compilation of other useful information for various purposes. The cooperation with the reporting entities has been organised through the Banking Association, since the reporting entities are in this case monetary financial institutions. The number of monetary financial institutions (statistical sector S.122), which varies between 20 and 30 has enabled frequent meetings with the representatives of all reporting entities. After these initial steps, BS codified (by Decree and the Guidelines) the reporting requirements. Matrix reporting started in 2004 with the following milestones: Narrow phase of new reporting has started with 3 monthly matrix reports in: o November 2004 (balance sheet items), o December 2004 (revaluation items) and in o May 2005 (interest rates). Reporting entities were originally all other MFI s, i.e. sector S.122 in ESA. Conceptually the matrix has comprised ECB s harmonised requirements in the field of monetary and financial statistics, many other statistical requirements, as well as some supervisory and financial stability needs. Complete reconciliation with the supervisory requirements has been initially postponed, but the matrix has been connected to basic supervisory balance sheet statement. 6

7 Further changes to original reports: o In November 2005, three partial reports on interest rates and one report on interest rates risk were incorporated in the new system, o In May 2006 the system was upgraded by the introduction of the IAS compliant data. The postponed wider reconciled reporting of MFI's has been gradually prepared and finalised in autumn Thorough preparations of revised reporting instructions have taken two years of cooperation between statisticians, supervisors, analysts as well as banking accountants. These modifications that are in line with the ECB s requirements for revision, were introduced in July 2010, and matrix profit and loss items will be implemented one year later. The revised reporting comprises unconsolidated balance sheet data, off-balance sheet balances, profit and loss accounts and interest rates of credit institutions. New requirements introduced in 2010 include: Revisions due to the ECB s regulations on balance sheet and interest rates statistics, Supervisory requirements on credit institutions including non-consolidated balance sheet - core FINREP, Statistics on Financial accounts, Further needs of external statistics. The existing three matrices (see p. 6) were reduced to only two (one on balance and offbalance sheet items with the interest rates and the other on flows), as it has been anticipated in the beginning. In practical terms the matrices have been significantly broadened by deeper breakdown of reporting items and additional code lists. Successful introduction of the revised matrices have enabled to abolish some existing separate reports. Despite significant increase in integration of supervisory and statistical needs the flexibility of new data collecting system resulted in consistencies of time series with many granularities in addition (e.g. amount of each type of guaranties for certain loan to individual customer). Plans for a third phase of the project are underway in which remaining attributes of credit register (from the balance sheet reports) will be included, which will enable to abolish another separate report. 7

8 Encouraged by the success of the multipurpose matrix reporting of monetary financial institutions, the idea has been taken up to extend this method also to reporting of nonmonetary financial institutions. First opportunity presented itself with the requirements of regulation ECB/2007/8 on statistics of investment funds. Securities Market Agency (ATVP), supervisor of capital markets in Slovenia, recognised the advantages of matrix reporting and supported its introduction in reporting entities, which accepted the proposal. The joint reporting began in January 2009, containing requirements of regulation ECB/2007/8 and regulation ECB/2001/13 for Money Market Funds, as well as providing statistics on financial accounts and other requirements of BS (especially for monitoring financial stability), and supervisory requirements of the ATVP. The reporting is not only based on knowhow of BS, but also - to the extent possible uses the same solutions (such as code lists) as in reporting of MFI s, thus linking both reports into a unique system. After a year of successful reporting, it appears that all parties involved are very much satisfied with the new system. In the meantime BS has already begun to cooperate with the Insurance Supervision Agency to expand multipurpose reporting to insurance companies and to pension funds (supervised by ATPV). Preparations are under way to define the reporting needs and to find technical solutions. Described multipurpose national collection of supervisory and statistical data is based on the simple idea that the provision and preparation of information should be done only once and with a minimal reporting burden, while maintaining high quality of statistics. This is in compliance with the principles endorsed by the Governing Council of the ECB in December 2007 within the long-term vision of Euro system s statistics 5. General considerations What are the main benefits of joint reports for statistical and supervisory requirements? i) The reporting entities prepare reports for statistical and supervisory purposes from the same set of data (for example balance sheet), that can be broken down in different ways, 8

9 ii) iii) iv) It is possible to achieve higher consistency of the decision support information for different levels (e.g. micro vs. macro) or for different functions (monetary policy, supervision, etc.), Data in a joint report are compiled and transmitted only once, thus reducing reporters time and costs, Potential common benefit could be a reconciled report on the EU level for the cross border organised reporting entities. For this reason the work of the JEGR group is extremely important and should be strengthened in the future. It is therefore not surprising that the representatives of the EU banking industry strongly support those efforts. 4 v) Another benefit of a single reporting of different data sets is considerable simplification and alleviation of reception and control of data, both for the reporters and receivers of data. In the jointly defined reporting system the relations between conceptually different data requirements are defined in a uniform way for all reporters. This is not the case in separated reports, where relation between different reports is not defined. This feature on one side increases the reliability and homogeneity of data and on the other side enriches their applicability. The reconciliation of conceptually different data requirements leads, as a rule, through lowest common denominator to more detailed breakdowns. Matrix is a suitable flexible tool to define detailed multidimensional data reports. It is technically demanding, and efficient management of such reports is only possible with the use of relational databases, data warehouses and electronic communication means. These technologies are nowadays common, and also already widely utilized by financial institutions, especially by those, whose complex businesses deal with numerous customers and products. It is therefore only reasonable to extend the use of current technical developments to the reporting systems where the number of reporting cells should not be any more an important constraint and cost. 4 This was unambiguously shown at the JEGR seminar with the (banking) industry on 28th of April

10 Matrix reporting influences development of relational databases and data warehouses both at reporting and at receiver's entities. It should not be forgotten that data collected for systematic multipurpose reporting is useful for reporting entities own management needs and encourages further development of their data warehouses for their specific needs. It enables them to detect, through drill-downs, potential errors and inconsistencies, and is thus a perfect tool to support the management of operational, liquidity and credit risk. Reformed reporting system also demands organisational rationalisation of data management at all parties involved in the reporting process, as there are less streamlined reports to be dealt with. The experience shows that it is easier to report and to collect fewer but more detailed reports of related data rather than more reports with less detailed data. The project has also clearly shown that organisationally separated functions of statistics, supervision, financial stability and research can draw on a common multipurpose database for all their needs. This is an important integrative element for most important central banking functions. 6. Conclusions As we are emerging from the most severe financial and economic crisis of the modern era, lessons can be drawn also for the financial statistics function. In the beginning of the crisis, regulators and supervisors lacked some pertinent and important statistical information linked to modern financial instruments, derivatives and exposures. In such circumstances it is of paramount importance that new information is provided as quickly as possible and with minimum disruption of established reporting systems. Efficient performance of policy objectives for monetary, financial stability and supervisory functions linked to financial institutions can only be achieved with timely and relevant information. This is exactly the case with the multipurpose reporting of financial institutions for statistical, supervisory and financial stability needs, which is described on previous pages. Such a reporting system has many comparative advantages on the micro and the macro level. It is detailed and flexible enough to fulfil data needs of supervisors, central banks and controllers of financial stability, such as, for example, the future European Systemic Risk Board. 10

11 Since there is, as yet, no such flexible and efficient reporting system in place at the European level, it seems important to support and prioritize the development in this direction. For this reason the work on reconciliation of different data requirements should be recognized as one of the most important projects in the data collection ever started in Europe in financial area. The work started on the initiative of the ECB by the JEGR group in 2008, should therefore be continued, broadened and strengthened. References: J. M. González-Páramo (2008) "Concluding Remarks" on a Fourth ECB Conference on Statistics: A Strategic Vision for Statistics Challenges for the Next 10 Years. Han, J. and Kamber, M (2001), Data Mining: Concepts and Techniques; Simon Fraser University, Morgan Kaufman, San Francisco. Banca d'italia (1992), I bilanci delle banche: schemi e regole di compilazione; Annex Matrice dei Conti, Banca d'italia, Roma. 11