Appendix I. Survey on the Use, Compilation, and Dissemination of Macroprudential Indicators
- International Monetary Fund
- Published Date:
- April 2006
1. The Survey on the Use, Compilation, and Dissemination of Macroprudential Indicators was conducted by the IMF in 2000. It was an important step in the IMF’s program to develop a common set of FSIs.1
2. The objective of the survey was to obtain information on national needs and practices related to FSIs to (1) gauge the usefulness of specific indicators, (2) assess compilation and dissemination practices to help identify international best practices where possible, (3) evaluate whether the SDDS or other vehicles would be appropriate to encourage the public dissemination of FSIs, and (4) explore the analytical frameworks used by member countries in macroprudential analysis.
3. The survey had two parts. The first part, the User Questionnaire, gathered information from financial supervisors, financial policymakers, and the private sector on the usefulness of the FSIs and methods of macroprudential analysis. The second part, the Compilation and Dissemination Questionnaire, inquired about national practices in compiling and disseminating FSIs.
4. The FSIs included in the survey largely focused on information about depository corporations (banks) but included some key information on their corporate and household counterparties. This focus was determined in light of the importance of banking institutions and the greater availability of information for banks compared with other types of institutions.
5. Central banks in each economy received the survey, with a request that they coordinate its distribution, completion, and return to the IMF. They were asked to distribute the survey within their economy to whichever parties they judged could best provide representative information on needs and practices relating to FSIs, such as the supervisory agency, the central government, and private sector participants.
6. A total of 122 responses (74 percent of those receiving the survey), covering 142 countries and other jurisdictions, was received. The first part of the survey (User Questionnaire) was completed by all 122 respondents, while 93 respondents completed the second part (the Compilation and Dissemination Questionnaire). The high response rate to the survey is an indication of the importance being attached worldwide to issues relating to macroprudential analysis and the possible role of FSIs in such analysis. This view is bolstered by the evident effort made by respondents to thoroughly answer the survey and provide detailed comments.
The Most Useful FSIs
7. Respondents judged all major categories of FSIs to be broadly useful. Indicators of capital adequacy, asset quality (lending institutions), and profitability were deemed the most useful, followed by indicators of liquidity and sensitivity to market risk. Users in industrial countries in particular deemed the liquidity and sensitivity to market risk indicators less useful than the others. Several respondents from industrial countries commented that the liquidity and sensitivity to market risk indicators were sophisticated and possibly difficult to construct with precision.
8. With the highest score possible being 4, Table A1.1 presents the 13 FSIs with an average usefulness score of 3.5 or over. These FSIs include central elements of bank soundness. Two of them—the Basel capital adequacy ratio and one of its components—relate to the capital base, which serves as a buffer to withstand shocks; four of them measure profitability, which serves to sustain the capital base. The remaining FSIs assessed to be most useful relate to the quality of banks’ assets—as covered by data on nonperforming loans, the distribution of assets, and asset liquidity. This list is the basis of the core indicators provided in Chapter 1.
|FSI #||FSI||All Countries||Industrial Countries||Emerging Countries||Developing Countries|
|1.1||Basel Capital Adequacy Ratio||3.8||3.7||3.9||3.6|
|1.1a||Ratio of Basel Tier 1 capital to risk-weighted assets||3.6||3.6||3.6||3.5|
|2.4||Distribution of loans, by sector||3.6||3.5||3.6||3.5|
|2.5||Distribution of credit extended, by sector||3.5||3.3||3.6||3.6|
|2.8||Ratio of total large loans to own funds||3.5||3.2||3.6||3.6|
|2.9||Ratio of gross nonperforming loans to total assets||3.9||3.9||3.9||3.8|
|2.10||Ratio of gross nonperforming loans net of provisions to total assets||3.8||3.8||3.8||3.8|
|3.2||Ratio of profits to period-average assets (ROA)||3.6||3.5||3.8||3.6|
|3.3||Ratio of profits to period-average equity (ROE)||3.6||3.5||3.8||3.6|
|3.4||Ratio of net interest income to total income||3.5||3.3||3.6||3.6|
|3.8||Spread between reference lending and deposit rates||3.5||3.4||3.6||3.5|
|4.3||Ratio of liquid assets to total assets||3.5||3.2||3.6||3.5|
|4.4||Ratio of liquid assets to liquid liabilities||3.6||3.2||3.7||3.7|
9. Table A1.2 presents the FSIs with an average usefulness score of 3.0 to 3.4. These FSIs form the basis of the list of encouraged indicators set out in Chapter 1. They cover some of the elements of capital adequacy, the distribution of bank credit by risk weight category and by country, the financial condition of the corporate and household sectors, some of the elements of operating income and expenses of banks, the maturity and duration of assets and liabilities, and other market risks.
|FSI #||FSI||All Countries||Industrial Countries||Emerging Countries||Developing Countries|
|1.1b||Ratio of Basel Tier 1 + 2 capital to risk-weighted assets||3.4||3.2||3.6||3.4|
|1.1c||Ratio of Basel Tier 1 + 2 + 3 capital to risk-weighted assets||3.0||2.9||3.1||3.1|
|1.2||Distribution of capital adequacy ratios (number of institutions within specified capital adequacy ratio ranges)||3.3||3.3||3.4||3.1|
|1.3||Leverage ratio (ratio of total on-balance-sheet assets to own funds)||3.2||2.9||3.3||3.3|
|2.1||Distribution of on-balance-sheet assets, by Basel risk weight category||3.4||3.2||3.5||3.4|
|2.4a||Loans for investment in commercial real estate||3.2||3.3||3.3||3.1|
|2.4b||Loans for investment in residential real estate||3.2||3.3||3.2||3.2|
|2.6||Distribution of credit extended, by country or region||3.1||3.2||3.2||2.8|
|2.7||Ratio of credit to related entities to total credit||3.4||3.0||3.6||3.5|
|2.11||Ratio of corporate debt to own funds (“debt-equity ratio”)||3.4||3.4||3.5||3.3|
|2.12||Ratio of corporate profits to equity||3.3||3.1||3.4||3.2|
|2.13||Ratio of corporate debt-service costs to total corporate income||3.2||3.2||3.4||3.0|
|2.14||Corporate net foreign currency exposure||3.2||3.2||3.4||2.9|
|2.15||Ratio of household total debt to GDP||3.0||3.2||3.0||2.8|
|3.5||Ratio of trading and foreign exchange gains/losses to total income||3.3||3.2||3.4||3.3|
|3.6||Ratio of operating costs to net interest income||3.4||3.0||3.6||3.6|
|3.7||Ratio of staff costs to operating costs||3.2||2.8||3.4||3.4|
|4.5||Average maturity of assets||3.4||3.0||3.4||3.6|
|4.6||Average maturity of liabilities||3.4||3.0||3.4||3.6|
|4.10||Ratio of customer deposits to total (noninterbank) loans||3.2||2.9||3.3||3.3|
|5.1||Ratio of gross foreign currency assets to own funds||3.1||2.7||3.2||3.2|
|5.2||Ratio of net foreign currency position to own funds||3.4||3.1||3.6||3.5|
|5.3||Average interest rate repricing period for assets||3.0||2.8||3.3||3.0|
|5.4||Average interest rate repricing period for liabilities||3.0||2.8||3.2||3.0|
|5.5||Duration of assets||3.2||3.0||3.4||3.0|
|5.6||Duration of liabilities||3.2||3.0||3.3||3.0|
|5.8||Ratio of net equity position to own funds||3.0||2.8||3.0||3.1|
10. The User Questionnaire also asked respondents to identify FSIs they considered useful but that were not covered in the survey. The most frequently identified useful additional FSIs were asset prices. Among the asset prices suggested were the prices of real estate, both commercial and residential, and equity prices, including the stock prices of the depository corporations subsector relative to the overall stock price index and stock prices disaggregated by industry. Moreover, to prevent the masking of relevant information through the aggregation process and to help in the identification of outliers, clustering of problem cases, or tiering in markets, there were calls for more information on the distribution or dispersion of observations. Several respondents identified the ratio of gross nonperforming loans to total loans as useful, in lieu of the FSI in the survey that used total assets as the denominator.
Importance of Nondepository Financial Institutions
11. About 80 percent of the respondents reported that information on nondepository financial institutions, markets, and activities was important to the overall analysis of financial sector soundness. On nondepository financial institutions,2 the majority of the respondents were most interested in information on insurance corporations and pension funds, followed by information on other financial intermediaries. Many of these institutions were viewed by respondents as playing an important role in financial intermediation and possibly in contagion. Several respondents mentioned the importance of specialized financial intermediaries, such as venture capital funds for advanced economies, and microcredit institutions and development banks or funds for developing countries. Some respondents noted the importance of information on financial conglomerates, especially those that included insurance companies.
12. On financial markets, about 90 percent of those responding on the issue indicated that data on the securities markets (public and private debt and equity markets) were important.3 A few thought that information on foreign exchange markets (16 percent) and derivatives markets (6 percent) was also important.
13. Several respondents noted that borrower information (indebtedness and asset-liability mismatches) was useful, as it provided some indication on emerging credit quality trends and risks in the corporate, household, or foreign sectors. Some respondents said that they paid particular attention to large corporations, while a few others mentioned the importance of monitoring other financial activity, such as the functioning of payment, settlement, and clearing systems. In addition, some respondents emphasized that qualitative information—such as the thoroughness of supervision and the transparency of financial policies—was important to the overall assessment of financial sector stability.
Disaggregation of “Depository Corporations” into Subsectors
14. Almost 60 percent of the respondents thought that more disaggregated information on depository corporations was needed, particularly breakdowns by ownership, function, exposure to risk (for example, geographical, asset type, borrower type), and size. A few respondents felt that disaggregated data that highlighted distributions among banks or allowed for peer group analysis were also useful. One respondent felt that the disaggregation of banks’ data should be as fine as possible to enable distinctive activity patterns to be identified. Several respondents, however, stressed that the type of disaggregation would depend on the issue being analyzed.
15. Almost 30 percent of all respondents (about half of those who felt that more disaggregation was useful) mentioned that they analyzed or would like to analyze institutions by ownership characteristics (for example, domestic versus foreign, private versus state-owned, and publicly held stock versus privately held equity). Of these respondents, almost all stated that a breakdown between domestic and foreign institutions was useful, with some emphasizing that the domestic/foreign distinction was important because foreign institutions might operate under different regulatory and supervisory regimes. At the same time, one-fourth of the respondents stated that a breakdown between private and state-owned institutions was important.
16. About 20 percent of the respondents said that disaggregation by function or exposure was useful. The functions most often mentioned were commercial banking, universal banking, and specialized banking (especially mortgage lending and, to a lesser extent, development lending). About 80 percent of the respondents interested in disaggregation by exposure indicated that they would like information on internationally active banks. Sixteen percent wanted disaggregated information on offshore banks, while another 16 percent wanted information on banks disaggregated by their geographical market.
Systemically Important Institutions
17. Almost 60 percent of the respondents reported doing some evaluation of systemically important institutions. Supervisors tended to be more concerned about such institutions—two-thirds of them reported that they evaluated the condition of these institutions, as opposed to less than half of market participants and about half of the government policy or research analysts.
18. Most respondents reported using a measure of size (of assets and/or deposits) to ascertain the importance of an institution. Sometimes size was coupled with other criteria, for instance, exposure to certain risks (such as foreign exchange risk), complexity of transactions, or complexity of ownership structure. However, some respondents mentioned only risk exposure, or used legal or prudential definitions, while others evaluated all institutions by sector or a particular category. This had as a result that all institutions within a particular classification (for example, problem banks, deposit-taking institutions, institutions with insured deposits, commercial banks, international banks) ended up sometimes being considered systemically important. This was often the case in countries with small, developing, or concentrated markets.
19. Many respondents said that the techniques used to evaluate the condition of systemically important institutions were similar to those used to evaluate other institutions. Most mentioned using the CAMELS framework or ratio analysis. Among the variables stressed by the respondents as important in their evaluations were interbank activity, liquidity, large exposures, foreign exchange exposure, consolidated positions for institutions that are part of a financial group, and risk management practices (including assessment by internal models).
20. Many respondents reported that specific norms, benchmarks, or thresholds were not used in macroprudential analysis. While some of them were considering using norms and benchmarks in the future, others preferred using comparisons with peer group countries to establish relative rankings.
21. Among those who reported using norms and benchmarks for FSIs, some highlighted their critical role in guiding interpretation of the indicators. For this purpose, benchmarks were constructed in a number of ways, including (1) historical averages, (2) bank supervisors’ prudential thresholds applied at the aggregate level, (3) trigger points, (4) cross-country comparisons, and (5) criteria constructed from econometric studies.
22. Overall, about half of all respondents reported that they made use of business survey results—qualitative or quantitative measures of business expectations and potential leading indicators of instability—to supplement macroprudential analysis.
Compilation and Dissemination Practice4
23. Country practices on the compilation and dissemination of FSIs and their components were mixed. With only a few exceptions, compilation of FSIs themselves was quite limited, and dissemination of FSIs—especially outside the industrial countries—was scanty. However, compilation and dissemination of components of FSIs were more extensive.
24. The average number of FSIs compiled and disseminated by industrial countries, emerging countries, and developing countries is shown in Figure A1.1. Industrial countries compiled and disseminated the largest number of FSIs, and emerging countries compiled and disseminated the second largest number of FSIs. Industrial and emerging countries compile on average more than half of the indicators specified in the survey.
Figure A1.1.Average Number of FSIs Compiled and Disseminated, by Type of Economy
Note: Comparison of the number of FSIs compiled and disseminated indicates that around 60 percent of compiled FSIs were disseminated; this percentage was broadly the same for each type of economy. Divergence between the number of FSIs compiled and disseminated indicated that the private sector had access to a narrower range of FSIs than is available to national authorities. It also indicates that there is scope for increasing the number of publicly available indicators of financial sector soundness in all types of economies.
25. For almost all FSIs, users in countries subscribing to the SDDS rated the usefulness of FSIs nearly identically with users in industrial and emerging countries. Although subscribers’ performance in the dissemination of components of FSIs was somewhat better than nonsubscribers’ performance, the overall results were broadly similar to those for the total population of respondents—that is, SDDS subscribers had somewhat limited compilation and dissemination of FSIs but relatively extensive compilation and dissemination of the component data series used to compile the FSIs.
26. The survey also inquired about country practices regarding the periodicity of compilation and dissemination as well as users’ needs in those areas. The periodicity of dissemination of FSIs varied considerably among the different categories of FSIs. No general pattern could be ascertained, and the number of responses was too low to draw valid conclusions.
27. The Compilation and Dissemination Questionnaire asked a series of quantitative and open-ended questions about accounting and statistical issues to assess the state of existing practices, possibly identify best practices that might be used as a basis for development of international standards, and help identify strategies for improving the comparability of FSIs.
28. The responses highlighted a diversity of national practices and revealed many reasons why FSIs might not be comparable across economies:
Different, and often complex, standards exist for recognition of substandard claims and provisioning.
National definitions of regulatory capital differ: for instance, as regards deductions and components of each tier of capital. Moreover, numerous countries indicated that they had not approved the inclusion of Tier 3 capital within the base.
Consolidation practices for foreign branches and subsidiaries differ (see section below). Within each country, some FSIs use global (cross-border) consolidations drawn from supervisory data, while other FSIs use domestic (national residence basis) consolidations drawn from statistical sources. Overall, however, some degree of international conformity exists in consolidation practices because of the rather widespread use of domestic consolidation.
Valuation practices for financial instruments differ (see section below). Key issues include the limited use of market valuations for debt securities and shares, and the diverse practices for on-balance-sheet recognition of derivatives, repurchase agreements, and securities lending.
Different rules exist across countries for revaluing foreign currency positions. Although there appears to be convergence in industrial countries toward use of market exchange rates in revaluing foreign-currency-denominated positions, continued use of official rates in a number of emerging and developing countries might hinder the comparability of FSIs.
29. The list of issues above indicated that practices were diverse and that cross-country comparison of FSIs was challenging.
30. The survey sought information on country practices for consolidating information on foreign branches and subsidiaries of financial institutions into single accounting statements or statistical reports. A key issue was whether data were compiled on a domestic or global consolidation basis.
31. Strong differences in practices by type of economy were found. Respondents in developing countries adhered overwhelmingly to a national residence basis for most FSIs. This possibly reflects the fact that banks with headquarters in developing countries may often have few or no nonresident branches or subsidiaries. It might also reflect limited supervisory infrastructures in developing countries that may not always effectively monitor and supervise nonresident operations. To some extent, adherence to domestic consolidation was also reported by respondents in emerging countries. In industrial countries, supervisors used global consolidation most often but also reported that data using both approaches to consolidation were available for numerous FSIs.5
32. Differences in practices by category of FSI were also found. These differences often reflected whether the primary source data are supervisory or statistical in nature. A summary of the practices by category of FSI is provided below.
Capital adequacy. In industrial countries and emerging countries, data were primarily from supervisory sources and generally on a global consolidation basis, although data using both consolidation approaches were often available. In a number of emerging countries and many developing countries, only data on a national residence basis were used. In terms of worldwide totals, the two approaches were used about equally, and in the case of some FSIs up to one-fourth of respondents used both. A small number of countries reported nonstandard consolidations in their data, such as including nonresident branches but not nonresident subsidiaries.
Asset quality (lending institutions). FSIs derived from monetary statistics were overwhelmingly on a national residence basis. FSIs derived from supervisory sources were most often on a global consolidation basis, but in many cases they were available on a national residence basis or on both bases.
Asset quality (borrowing institutions). FSIs were almost exclusively on a national residence basis because the underlying data were drawn from national macroeconomic statistical series.
Profitability and competitiveness. Data were most often on a national residence basis or were available on both bases. However, a number of countries had data only on a global basis. Within the profitability category, nonstandard consolidations were used by a number of countries.
Liquidity. Consolidation on a national residence basis was most common, but the FSIs on liquid assets and average maturities of assets and liabilities were often consolidated on a global basis. Global consolidation is not relevant for some of the liquidity FSIs that refer solely to national conditions.
Sensitivity to market risks. Consolidation on a national residence basis was most common. Consolidation on a cross-border basis was used to some extent in supervisory data in industrial and emerging countries.
33. For deposits and loans, historical valuations were most commonly used—in supervisory data, in at least three-fourths of all responses, and in statistical data, in about 9 out of 10 cases. In contrast, for securities (other than shares), as well as for shares and other equity, no valuation method clearly predominated, although use of market values was more common than other valuation approaches. For financial derivatives, market valuations were used most often, with supervisors also reporting fairly common use of “other” valuation methods, such as hedge valuations. Historical valuations predominated in miscellaneous receivables and payables and in nonfinancial assets, but use of the other types of valuations was not uncommon.
34. On the translation of the value of foreign-currency-denominated instruments into domestic currency equivalents, end-of-period exchange rates were used most often for all types of financial instruments. A large minority of emerging countries and developing countries reported that they used official exchange rates. Foreign currency positions were revalued most often at the rate applying on the balance sheet closing date. However, revaluations of foreign currency positions at other frequencies were not uncommon for securities (other than shares), shares and other equities, and financial derivatives.
35. The majority of respondents preferred the use of ratios and growth rates in presenting their FSIs. However, many respondents also felt that the preferred mode of presentation depended on the particular FSI in question and the type of analysis being conducted. For example, for sectoral aggregates, it was useful to have weighted averages as well as simple averages, accompanied by the frequency distribution of institutions according to the range of values of the indicators.
36. Some respondents noted that measures of dispersion (standard deviations, histograms, Gini indices, and so forth) could be particularly useful in presenting FSIs because they allowed the analyst to identify, among other things, outliers, trends in concentration, and tiering in markets, which could be relevant for the analysis of financial stability.
A more detailed discussion of the survey and its results is provided in Sundararajan and others (2002).
Defined, in line with the MFSM (IMF, 2000a), as insurance corporations and pension funds, other financial intermediaries, and financial auxiliaries.
The types of data mentioned included trading volumes, bid-ask spreads, and credit spreads.
For a detailed discussion of compilation and dissemination practices, see Slack (2003).
The availability of FSI data on both bases could have some important advantages. For example, one respondent noted, “The survey does not address the main statistical aspect, which is reconciliation between the home and host country approach, which will be viable if both supervisory and macroeconomic statistical data sources are used.”