The Balance Sheet Approach and its Applications at the Fund
June 30, 2003
A Balance Sheet Approach to Financial Crisis
December 1, 2002
Data Provision to the Fund for Surveillance Purposes
April 26, 2002
External Debt Statistics Guide for Compilers and Users
June 25, 2003
Original Sin-Balance Sheet Crises, and the Roles of International Lending
December 1, 2002
Integrating the Balance Sheet Approach into Fund Operations
Prepared by the Policy Development and Review Department
(In consultation with other departments)
February 23, 2004
1. This note provides Directors with a summary of the work that is underway to integrate the balance sheet approach (BSA) into the Fund's operations. During the informal Board discussion last July, Directors concluded that the BSA provides a useful analytical framework to detect currency mismatches and other balance sheet weaknesses of an economy as a whole and of its main sectors, which could help the Fund to strengthen its vulnerability analysis and policy advice.1 While it was noted that elements of the BSA are already being used as a complement to traditional flow-based analysis, Directors encouraged staff to develop a strategy to further advance the BSA in Fund operations.
2. In light of the resource constraints, the strategy focuses on four areas that are likely to be most productive: (i) analytical work; (ii) country applications; (iii) Fund operations; and (iv) outreach. Furthermore, the work has followed the risk-oriented approach suggested by some Directors: It is mainly geared towards emerging market countries, where balance-sheet-related weaknesses, in particular currency mismatches, present the most immediate vulnerability concerns. As other Directors noted, however, it is also useful to apply the BSA in mature market countries; indeed, Fund surveillance in these members is increasingly drawing on the BSA. Regarding low-income members, the cost of adopting the fully fledged balance sheet approach may well outweigh its benefits, as these countries are generally not exposed to the sort of balance sheet risks (e.g., rollover risk, corporate exposure) that emerging markets face.
3. The Board will discuss a paper on liquidity management after the spring meetings. The paper, prepared by PDR, makes proposals on reserve adequacy and debt management that are derived from a balance sheet framework. It provides an analytically and empirically based explanation of how external vulnerability can result from combined maturity and foreign currency exposures within sectoral balance sheets. Based on this finding, the paper outlines policy principles and operational tools to improve the management of debt and official foreign reserves.
4. A forthcoming Board paper on debt-related vulnerabilities in emerging market economies will use cross-country analysis to demonstrate how the structure of debt, and balance sheet mismatches more broadly, can contribute to financial crises. This project, undertaken by PDR, takes a broad look at the universe of emerging markets with high public debt levels and draws some general conclusions for Fund policy. A more detailed analysis of some country cases is intended to illustrate how sectoral balance sheet mismatches can help explain both the reasons for, and the severity of, recent financial crises. The analysis will thereby also serve as a didactic device, showing how the BSA might be usefully applied even with relatively limited data resources.
5. Research is underway to explore potential innovations aimed at reducing vulnerabilities on government balance sheets. RES is preparing a paper that analyzes possible measures that may result in lower exposure to the risk of government debt crises. Such measures include the development of domestic debt markets, the use of debt instruments indexed to key macroeconomic indicators, and the design of explicit seniority structures in debt issued by the government. The paper is scheduled for an informal Board seminar in May.
6. Work on the development of Financial Soundness Indicators (FSIs) and the strengthening of FSAPs' analytical tools continues. MFD's last Board paper and background paper on FSIs and describes how FSIs can be used to monitor the vulnerabilities of the financial sector, including how to gauge the impact of weaknesses in other sectors' balance sheets on the resilience of the financial sector. Following the last FSAP-related Board papers that reviewed the experience with stress testing in FSAPs, the focus of further work in MFD is now on developing practical guidance for conducting such stress tests.
7. The commercially available Moody's MfRisk model is being tested in parts of the Fund. A paper by ICM aims at explaining and applying the MfRisk model. The model uses a so-called macrofinance approach that applies contingent claim analysis—a concept originally developed in the corporate finance literature—to model country risk. ICM's current work should give staff an indication of the extent to which the model can be useful for the Fund in identifying balance sheet vulnerabilities. The paper is intended for presentation at a conference on early warning systems in March 2004, sponsored by the World Bank.
8. At the same time, new strategies for early warning systems are being developed that explicitly build on the BSA. For example, a forthcoming Fund working paper provides a set of early warning systems that focus on balance sheet measures in the corporate and the financial sector, such as debt structure, leverage, liquidity, and profitability, as well as some indexes of corporate governance. By introducing microstructural indicators, such models can complement the traditional warning systems in capturing external vulnerabilities.
9. Finally, BSA-related issues in general have received considerable attention in research undertaken by staff across departments. In recent years, research projects initiated by individual staff or coordinated by departments have addressed a variety of topics that relate to the BSA. These include, for example, papers on the implications of corporate balance sheet structures for financial stability, the effects of liability dollarization on banks' balance sheets and the role of official reserves, the measuring of off-balance-sheet leverage, and the role of international lending in balance sheet crises.
10. The BSA already serves as a conceptual framework for Fund surveillance in mature markets. For example, within the last year the Article IV consultations for Australia, Ireland, the United Kingdom and the United States focused on potential changes in real estate values and the implications for mortgage lending and household debt. The international linkages of the bank and insurance sectors have been the subject of selected issues papers for Germany and Spain. In the case of Austria, currency mismatches—rapidly expanding foreign exchange denominated loans to households—have been subject of staff scrutiny. These studies have all looked at specific sectors and the data provided by the authorities have generally been adequate. A fully-fletched intersectoral balance sheet analysis is very data intensive, although some members (such as the United Kingdom) are making headways in this area.
11. For emerging market countries, cross-country issues related to the current level and structure of public debt could be presented to the Board, possibly as part of a regular WEMD session. This would respond directly to a request in the September IMFC communiqué for the Fund to focus on currency mismatches, since these would naturally be at the center of the presentation.
12. We aim at including more cross-sectoral balance sheet analysis of vulnerabilities in future Article IV consultation reports for selected emerging market countries. So far, there are only two cases (Thailand and Peru) where such an analysis was included in a Board paper and two cases (Turkey and Lebanon) where country teams are well advanced in developing balance sheet frameworks that cover several sectors. A few other countries have been identified as candidates for balance sheet analysis, based on data availability, interest expressed by the authorities, and their vulnerability to mismatches. Preliminary work has started, in some cases with technical assistance from STA.
13. In some cases, the analysis of individual sectoral balance sheets may also be valuable. While the BSA is only fully exploited if cross-sectoral linkages are covered and the model closes, examining important individual sectoral balance sheets—in particular that of the financial sector—can be useful to detect weaknesses that have the potential to spill over into other sectors. Examples are:
14. Recent experience with applying the approach, however, also highlights the paramount importance of good data. Data initiatives over the recent years have led to considerable improvements in the availability and timeliness of data in general. Nevertheless, for many countries balance sheet information beyond what is readily available from existing sources must be gathered to make a complete intersectoral analysis possible. Establishing sound data bases will take much time and effort by country authorities and staff. A fairly long gestation period is therefore likely before BSA-based analysis can evolve from its current experimental stage to an exercise that can more routinely be integrated into staff's country work.
15. Especially the analysis of the corporate sector faces many practical limitations. The available data bases often only include publicly listed corporations. Although these are typically the ones that have access to international capital markets, and are therefore exposed to the associated risks, they still present only a subgroup among the corporations, and for this reason may not adequately reflect the complex vulnerabilities of this heterogeneous sector. Yet, while this heterogeneity often requires industry-specific analysis, it also makes such a differentiated analysis more difficult. For example, the conceptually important disaggregation of the sector into producers of tradable and nontradable goods has proved difficult to make in practice, even where aggregate data are available. Finally, weak and widely differing accounting standards often pose a further constraint.
16. The forthcoming joint PDR-STA Board paper on data provision to the Fund discusses in some depth the data requirements for the BSA, the actual use of such data by staff to date, and related statistical and operational challenges from a user's perspective. It will at the same time assess how the BSA initiative relates to the debt sustainability framework (DSA) and the compilation of financial soundness indicators (FSIs). This paper will help to form an integrated view of the Fund's data needs for both surveillance and program-related work, through exploring how existing data sources can be used more intensively and data availability under the Fund's ongoing data initiatives can be improved.
17. Work is underway to further improve external debt statistics from the creditor and the debtor side. STA, in the context of the work program of the Fund-chaired Inter-Agency Task Force on Finance Statistics, is involved in a World Bank project that develops a standardized, quarterly, cross-country time-series database of external debt statistics for SDDS-subscribing countries, drawing on data from the debtor side. Countries are being consulted, and the database should be made operational during 2004. While its primary focus is on the stock of external debt, the inclusion of data on forward debt-service schedules and the currency composition of debt is also being considered. At the same time, work continues on revamping and extending the coverage of the existing joint BIS-IMF-OECD-World Bank external debt statistics which disseminate data obtained primarily from creditor and market sources. The BSA underscores the importance of also improving domestic debt statistics, which are often more deficient than those on external debt.
18. The guidance note on the coverage of financial sector issues in Article IV consultations will necessarily touch on the BSA. The note, prepared jointly by PDR, MFD and STA will clarify the objective, scope and focus of financial sector surveillance, as well as its operational aspects. The proposed methodology follows the same macroprudential analysis framework employed in FSAPs, in which the assessment of financial institutions' exposure to various types of balance sheet risks—credit risk, market risk (which includes currency and interest rate risk) and liquidity risk—plays a central role.
19. In this context, the coordinated compilation exercise for Financial Sector Indicators (FSIs) is a concrete action that can support the BSA. Based on the FSI Compilation Guide that has been prepared by STA in cooperation with MFD and is currently being updated, selected countries are being encouraged to make a concentrated effort to compile FSIs. The data involved in compiling FSIs, in particular those for the key nonfinancial sectors covered by the encouraged set of FSIs, will also be of great use for any other BSA-related work.
20. The new Government Finance Statistics Manual 2001(GFSM 2001) provides a basis for the systematic analysis of the government balance sheet. The formal integration of stocks and flows, and the incorporation of a government balance sheet are among the defining features of the GFSM 2001. In this context, a forthcoming FAD paper on Public Investment and Fiscal Policy reviews the implications of adopting the balance sheet approach to fiscal policy. Another paper is being prepared by FAD, PDR and STA—in consultation with other departments—that discusses the implications of the GFSM 2001 for country work, and additional work is underway to provide specific operational guidance for GFSM 2001 implementation.
21. Insights of the BSA will also be taken into account in a revised guidance note on setting debt limits in Fund-supported programs. The paper will be brought to the Board after the paper on debt structures mentioned above. It would reinforce the importance of an approach to setting debt limits that puts greater emphasis on improving the currency and maturity structure of public debt, and more explicitly takes into account contingent liabilities.
22. As the approach is developed within the Fund, outreach—kept to a modest level to date—will become increasingly important to help broaden the understanding and the usage of the approach. In October, PDR gave seminars on the BSA at the European Central Bank and the Austrian National Bank. Discussions concentrated on its possible use for analysis of EU accession and other emerging market countries. While these institutions are very interested in the BSA as a concept, they are not currently implementing the approach systematically, at least not outside the Euro zone. Further dissemination of the BSA that reaches beyond academic and institutional audiences, could be achieved through articles in Finance & Development (building on the overview article in December 2002), the IMF Survey, and some specialized but non-technical journals (e.g. Central Banking).
23. In its financial programming courses for country officials, INS staff is beginning to cover sectoral balance sheet exposures and their implications. Once the development of the balance sheet methodology is sufficiently advanced, INS envisages arranging internal and external training seminars on the topic.
24. During the present experimental stage, the resource costs of the BSA will be met by redeployment within the existing resource envelope. On the level of country teams the limited experience to date suggests that completing a selected issues paper on the BSA requires 0.25-0.5 staff years, depending on the complexity of the case and data availability. In some cases demands on staff may be higher, not only to compile the necessary data, but also to scrutinize adequately the results of such an exercise and analyze them. This means that staff resources within mission teams will have to be redirected from other surveillance issues. The budgetary costs of maintaining regular subscriptions to commercially available databases (mainly for corporate data) for a small number of authorized users currently amount to some US$50,000 per year and are already included in the joint library's budget allocation. The additional budgetary implications for STA are discussed in the forthcoming paper on data provisioning.
25. Resource costs would, however, rise substantially if the approach were to be made a regular element of Fund surveillance. Country work during the experimental phase is necessarily concentrating on members where data are relatively easy to obtain and analyze. Should the BSA approach be adopted to a larger group of countries, the costs to the Fund—both in terms of staff resources (on mission teams and for technical assistance) and data subscriptions—would increase significantly. For example, applying a fully-fledged BSA analysis to 20 high-risk emerging markets alone could amount to at least 10 staff years (US$1.7 million) in area departments and a similar need for TA resources. In the current zero-real-growth budgetary environment this would require a fundamental redeployment of resources earmarked for surveillance.
1 For a description of the BSA see WP/02/210.