D E V E LO P M E N T KNOWLEDGE AND LEARNING Inaugural RAMP Survey on the Reserve Management Practices of Central Banks Results and Observations The Reserves Advisory and Management Program (RAMP) DEVELOPMENT KNOWLEDGE AND LEARNING Inaugural RAMP Survey on the Reserve Management Practices of Central Banks Results and Observations The Reserves Advisory and Management Program (RAMP) © 2019 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 22 21 20 19 Books in this series are published to communicate the results of Bank research, analysis, and operational experience with the least possible delay. The extent of language editing varies from book to book. This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2019. “Inaugural RAMP Survey on the Reserve Management Practices of Central Banks: Results and Observations.” Development Knowledge and Learning. Reserves Advisory and Management Program, World Bank, Washington, DC. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an offi- cial World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content ­contained within the work. The World Bank therefore does not warrant that the use of any third-party- owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. SKU: K880263 Cover photo: © MarianVejcik / Getty Images. Used with permission; further permission required for reuse. Cover design: Debra Naylor / Naylor Design Inc. Contents Acknowledgments  v Abbreviations  vii CHAPTER 1: Introduction  1 Notes  3 CHAPTER 2: Key Findings  5 Note  7 CHAPTER 3: Results and Observations   9 Governance  9 Strategic asset allocation   18 Portfolio management  27 Risk management  31 Performance reporting and transparency   34 Notes  35 CHAPTER 4: Concluding Commentary  37 Appendix A   39 Bibliography  43 Figures 1.1 Global foreign exchange reserves, 1998–2018   1 1.2 Global foreign exchange reserves as a percentage of global trade, 1998–2017  2 3.1 Investment policy decision-making process   10 3.2 SAA decision-making process   11 3.3 Organizational structure for reserve management responsibilities   11 3.4 Motives for holding foreign exchange reserves   12 3.5 Motives for holding foreign exchange reserves by country-income group   12 3.6 Reserve adequacy measures   13 3.7 Reserve adequacy assessment by country-income group   14 3.8 Numeraire  15 3.9 Numeraire choice by country-income category   15 3.10 Investment principles  16  iii iv | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks 3.11 Central banks by country-income group that consider return generation highly relevant  17 3.12 Investment horizon of investment tranche   17 3.13 Use of tranching by country-income group   18 3.14 Percentage of respondents that may hold each currency as part of their foreign exchange  19 3.15 Factors shaping currency composition   20 3.16 Distribution of all respondents’ allocations to individual currencies   20 3.17 Distribution of allocations to currencies for those central banks with exposure  21 3.18 Net number of central banks that plan to increase or decrease an allocation to specific currencies within 2 years   22 3.19 Eligible asset classes   23 3.20 Higher risk financial instruments that are eligible asset classes   24 3.21 Distribution of all respondents’ allocations to individual asset classes   24 3.22 Distribution of the allocation to individual asset classes for respondents with exposure  25 3.23 Comparison of allocation to traditional asset classes and months of import coverage of foreign exchange reserves   26 3.24 Comparison of allocation to traditional asset classes and months of import coverage of foreign exchange reserves of emerging market respondents   26 3.25 Portfolio management styles   27 3.26 Reasons for hiring external managers   28 3.27 Foreign exchange reserve allocations to external managers   29 3.28 Use of derivatives   29 3.29 Incorporation of ESG factors into the investment framework   30 3.30 ESG and equity investing   30 3.31 Minimum credit ratings by asset class   31 3.32 Credit assessment methodologies   32 3.33 Credit assessment methodologies for respondents with corporate credit exposure  32 3.34 Credit analysis techniques of respondents with minimum credit ratings of BBB- for at least one asset class   32 3.35 Metrics for measuring market risk   33 3.36 Market risk measurement metrics for respondents that use active or enhanced indexing styles  33 3.37 Types of stress scenarios employed   34 3.38 Highest frequency of reporting to the board   35 3.39 Mandatory or voluntary disclosure of reserve management performance   35 A.1 Respondents’ foreign exchange reserve allocation to traditional asset classes and level of short-term external debt coverage   39 A.2 Emerging market respondents’ foreign exchange reserve allocation to traditional asset classes and level of short-term external debt coverage   40 A.3 Respondents’ foreign exchange reserve allocation to traditional asset classes and reserve cover of ARA metric level   40 A.4 Emerging market respondents’ foreign exchange reserve allocation to traditional asset classes and reserves cover of ARA metric level   41 A.5 Use of external managers by country-income group   41 A.6 Performance attribution models   42 A.7 ESG and investing in corporate bonds   42 Table 1.1 Survey participants’ reserve levels and adequacy metrics   3 Acknowledgments This report shares results and observations arising from data that the Reserves Advisory and Management Program (RAMP) collected from 99 central banks from around the world as part of its inaugural survey of these institutions’ reserve management practices. Although published under an institutional rubric, it is the product of the work of many individuals. The authors of the report, listed alphabetically, are Keyvan H. Alekasir, Nikoloz Anasashvili, Matias Antonio, Philip Dongsoo Hong, Daniela M. H. Klingebiel, Bernard Murira, Guilherme Henrique Pereira Alves, Marco Antonio Ruiz Gil, Alexander E. Slater, and Daisuke Takahashi. In writing this report, the authors benefited from valuable contributions from a host of others and are grateful for their support. An initial draft received com- ment from the following World Bank colleagues: George Bentley, Eric Bouye, Steen Byskov, Otilia Iulia Ciotau, Roberto A. De Beaufort Camargo, Erik Feyen, Bjoern Geir From, Natan Goldberger Rico, Attila Juhasz, Azzedine Lazizi, Wendy Mendes, Stephane Piot, Mahmut Rustem Sen, and James Seward. The following colleagues at the Bank of International Settlements also provided feedback: Pierre Cardon, Stijn Claessens, Mike McMorrow, Vahe Sahakyan, and Michela Scatigna. The authors would also like to thank the staff of the 99 central banks who answered the survey’s questions. Without their input, none of this would have been possible.  v Abbreviations ABS asset-backed securities ARA Assessing Reserve Adequacy AUD Australian dollar CAD Canadian dollar CHF Swiss franc CVaR conditional value at risk EM emerging market ESG environmental, social, and governance EUR euro GBP British pound sterling GDP gross domestic product IMF International Monetary Fund JPY Japanese yen KRW Korean (Rep. of ) won MBS mortgage-backed securities NOK Norwegian krone NZD New Zealand dollar RAMP Reserves Advisory and Management Program RMB renminbi SAA strategic asset allocation SDR special drawing rights SEK Swedish krona SGD Singapore dollar SSA supranational, sovereign, and agency USD United States dollar VaR value at risk ZAR South African rand  vii 1 Introduction Over the past two decades, global foreign exchange reserves have increased more than six-fold, reaching US$12.6 trillion as of March 31, 2018.1 Aggregate reserves peaked in 2014 and decreased moderately thereafter for approximately 2 years—mostly due to a decline in emerging market economies’ reserves that arose in part from a fall in commodity prices—before rebounding (see figure 1.1). Reserve accumulation has also exceeded the expansion in trade over the last 20 years (see figure 1.2). This historically unprecedented growth in foreign exchange reserves accel- erated the evolution of central banks’ asset management practices (Eichengreen and Flandreau 2014). Many of these institutions expanded their focus beyond the traditional investment objectives of liquidity and safety to include income generation. In doing so, they faced additional challenges arising from the extraor- dinary policy responses to the unprecedented economic conditions following the financial crisis. FIGURE 1.1 Global foreign exchange reserves, 1998–2018 14 US$12.6 12 Reserves (trillion) 10 8 6 US$5.4 US$4.1 4 2 US$3.2 0 98 99 00 01 02 03 04 05 06 07 08 20 7 09 10 11 12 13 14 15 16 18 1 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Year Regions China Emerging and developing economies (excl. China) Advanced economies World Source: IMF, International Financial Statistics.  1 2 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 1.2 Global foreign exchange reserves as a percentage of global trade, 1998–2017 60 55 Percent of global trade 50 40 28 30 23 18 20 18 10 8 14 2 0 98 99 00 01 02 03 04 05 17 06 07 08 09 10 11 12 13 14 15 16 19 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Year Regions China Emerging and developing economies (excl. China) Advanced economies World Sources: World Bank, World Development Indicators; IMF, International Financial Statistics. The World Bank Treasury’s Reserves Advisory and Management Program (RAMP) concluded its inaugural survey on central banks’ reserve management practices in the spring of 2018. The goals were to take stock of and develop a more complete understanding of these institutions’ reserve management poli- cies and practices globally. The survey’s objectives were (1) to construct a picture of reserve management activities across multiple regions; and (2) to provide an opportunity for central banks to benchmark their actions and perspectives against peer institutions. The survey addressed key areas of public asset management. Its content covered (1) governance and policy; (2) strategic asset allocation; (3) portfolio man- ­ agement; (4) risk management; and (5) performance reporting and transparency. The survey posed 36 questions across these areas, some of which requested ­additional information depending on the participants’ answers. Some queries gave a prescribed set of potential responses; others requested specific data. The results comprise input from 99 central banks and reflect an overall response rate of approximately 80 percent. Respondents represent countries with different income levels and from multiple regions (see table 1.1).2 Their amounts of foreign exchange assets and levels of reserve adequacy cover a wide range.3 Although most participants provided substantial amounts of informa- tion, some did not answer every question. When presenting data, this report identifies the number of institutions responding to the relevant question (or each section of a question when necessary), either in the main text or in corresponding charts and tables. ­ Data is presented in an aggregate and unattributed format to maintain respon- dents’ anonymity.4 Observations on this information arise from assessments through various lenses, including country-income group, measures of reserve adequacy, and monetary policy and exchange rate regimes. Where this analytical process identified patterns, the report shares these findings. The remainder of this report is organized in three parts. Section 2 highlights its key findings. Section 3 describes the survey’s results and offers observations on patterns in the data. Finally, section 4 discusses potential policy implica- tions arising from the responses and analysis. Introduction | 3 TABLE 1.1  Survey participants’ reserve levels and adequacy metrics NUMBER MEDIAN NATIONAL RESERVE LEVELS RESERVE ADEQUACY METRICS OF GDP PER MEDIAN TOTAL MEDIAN OF MEDIAN MONTHS MEDIAN OF TOTAL REGION CENTRAL CAPITA RESERVES TOTAL RESERVES OF IMPORT RESERVES TO SHORT- BANKS (USD) (000,000 USD) TO GDP COVERAGE TERM DEBT OBLIGATIONS Americas and Caribbean 16 9,934 11,314 0.16 5.1 2.1 East Asia and Pacific 17 5,902 80,692 0.25 5.9 1.6 Europe and Central Asia 35 14,117 25,191 0.19 3.4 0.9 Middle East and North Africa 7 4,828 25,106 0.30 9.1 2.8 South Asia 5 1,861 7,268 0.16 8.4 3.5 Sub-Saharan Africa 19 1,708 2,353 0.15 4.0 2.4 COUNTRY-INCOME GROUPa High income (reserve) 16 16,069 33,159 0.27 5.8 1.6 High income (non-reserve) 16 51,265 75,901 0.06 2.5 0.1 Upper middle income 30 7,491 12,501 0.22 5.5 2.7 Lower middle income 28 2,739 5,901 0.15 5.1 1.6 Low income 9 662 2,108 0.13 4.0 19.1 Source: World Bank, World Development Indicators (October 2018). Note: GDP = gross domestic product. a. Some charts in this report present data on a country-income group basis. Percentages in these charts are based on the total number of survey respondents in a given country-income group, rather than the number of institutions in a country-income group that provided data on the specific topic. Therefore, percentages in these charts will not sum to 100 if the number of respondents to the question were less than the overall survey sample of 99 institutions. NOTES 1. This report defines “foreign exchange reserves” as the pool of non-domestic currency denominated assets a central bank or monetary authority holds for meeting a defined range of objectives. The reserve management entity is responsible for the investment of this wealth and curation of associated risks. At times, this report may use terms such as “reserve assets,” “reserve holdings,” or “foreign currency reserves” or even “reserves” in referencing this pool of wealth. 2. This report uses the World Bank’s customized country-income group categories based on GNI per capita calculated using the World Bank Atlas Method. It separates countries into “low income” (less than $1,005); “lower middle income” ($1,006–$3,955); “upper middle income” ($3,956–$12,235); and “high income” groups (more than $12,235) (World Bank Data Team 2017). Because of the report’s subject matter, it further divides the “high income” category into “high income reserve” and “high income non-reserve” batches. The former encompasses only those that print currencies most often held as foreign exchange reserves by other central banks. The latter encompasses all others in the high-income country category. 3. There are various ways of measuring the adequacy of central banks’ levels of foreign exchange reserves, including coverage of imports and short-term debt obligations. Unless otherwise specified, this report uses the term to reference a central bank’s possession of sufficient levels of reserve assets to execute its mandate and achieve its objectives. 4. RAMP staff believed that confidentiality would facilitate central banks’ participation and candid and comprehensive responses given the sensitive nature of their operations. 2 Key Findings Over the past 20 years, managers of foreign exchange reserves have had to respond to two major market developments—a substantial increase in the amount of these assets globally and the extraordinary policy responses to the unprecedented macroeconomic and investment environment after the global financial crisis. The survey’s key findings suggest that, despite these factors, most central banks continue to employ a traditional reserve management approach. Their investments remain concentrated in high-quality fixed-income assets and the minimum credit rating for these holdings remains conservative. At the same time, the data suggest that important changes are underway as a material number of central banks reported more diversified portfolios with exposure to non-traditional asset classes. A third of respondents hold corporate credit, most of which is investment grade, and almost one in five own mort- gage-backed securities (MBS) or equities, although mostly in limited allocations. Our analysis of this information did not find a relationship between respondents’ measures of reserve adequacy and the size of their exposure to non-traditional asset classes. The data exhibit considerable cross-country differences in the way central banks manage their reserves and, in some circumstances, our analysis suggests these differences correlate with respondents’ country-income groups. The key findings on governance and policy are as follows: 1. Central banks use a diverse set of arrangements to guide and implement their reserve management activities. They divide these responsibilities among various institutional bodies and use distinct approaches to execute mandates like invest- ment policy development and construction of a strategic asset allocation (SAA). 2. Self-insurance against external shocks is the primary motive for holding foreign exchange reserves that central banks most frequently consider highly relevant. They also deem conducting foreign exchange policy and servicing external debt of similar importance albeit less often. Saving for intergenerational equity does not appear to be a major concern even with the substantial increase in global hold- ings of foreign exchange assets.  5 6 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks 3. Most central banks measure reserve adequacy in one way or another. They most frequently use the import coverage method followed by short-term external debt ratio, broad money ratio, and the IMF’s Assessing Reserve Adequacy (ARA) metric.1 4. The United States dollar (USD) and domestic currency are the most frequently used numeraire currencies. Almost all central banks in the middle- and low-­ income country groups use the U.S. dollar as numeraire. In contrast, a substan- tial majority in high-income reserve countries use domestic currency. 5. Almost all central banks consider safety and liquidity as highly relevant priorities. More than a third also highlight returns/income generation as an important motive and almost all the others identify it as somewhat relevant to their reserve management strategies. The key findings on SAA are as follows: 1. Most institutions use tranching, typically separating reserves into liquidity and investment portfolios. Most adopt an investment horizon of more than 1 year for foreign exchange assets allocated to portfolios focused on income generation/ returns. They employ various metrics to express their risk tolerance. 2. Assets denominated in the U.S. dollar and euro (EUR) are the dominant formats for investments. Multiple currencies that central banks consider as eligible investment denominations do not comprise substantial amounts of their portfo- lios. One in ten respondents indicated that, in the near term, they will establish an allocation to financial instruments valued in renminbi (RMB). These institu- tions identified allocations to U.S. dollar and renminbi assets as most likely to increase over this period. 3. The dominant asset classes for portfolio composition are government bonds, bank deposits and money market instruments. Over half of central banks are authorized to purchase non-traditional investments, such as MBS and corporate bonds. One third of institutions own corporate bonds and almost one in five have exposure to MBS and equities. The key findings on portfolio management are as follows: 1. Three quarters of respondents show a high willingness to take on active risk, using either an enhanced indexation or active style. 2. Most institutions use external managers to implement part of their SAA across a range of investment styles. Almost all of them consider the possibilities of knowledge transfer, capacity building and higher returns as highly relevant to their work with these agents. A majority allocate less than 10 percent of their reserves to third party managers. The key findings for risk management are as follows: 1. Respondents reported a relatively conservative approach to fixed income invest- ment as measured by minimum credit rating requirements. All but one do not allow investments below a rating of “BBB-” for government or corporate bonds. Approximately two-thirds use “A-” as the minimum credit rating for sovereigns. Credit ratings are central banks’ main credit risk assessment mechanism. Still, 60 percent indicated they use other methodologies to measure and manage the likelihood of default on the obligations in their reserve portfolios. 2. Most central banks manage market risk using duration and currency limits. About a third also determine limits based on probabilistic risk measures such as VaR (value at risk), CVaR (conditional value at risk), and Tracking Error. Approximately two-thirds also incorporate stress testing into their risk management activities. Key Findings | 7 The key findings for reporting and transparency are as follows: 1. More than half of central banks disclose information (voluntarily or mandatorily) on currency composition, asset classes, investment universe, and reserve manage- ment performance. A majority of institutions do not provide information on institutional regulation, risk metrics, benchmarks, external managers and investment horizon. NOTE 1. The IMF’s ARA metric is a composite of multiple factors relevant to reserve adequacy, including short-term debt, broad money, and imports (International Monetary Fund 2018). 3 Results and Observations This section describes the survey’s results and shares observations where the authors identified notable patterns. Its subsections track the five main areas of public sector asset management that were the foci of the survey’s questions: (1) governance and policy; (2) strategic asset allocation; (3) portfolio management; (4) risk management; and (5) performance reporting and transparency. GOVERNANCE Governance and organizational structure Governance in reserve management refers to the institutional arrangements and processes for policy development and investment of foreign exchange assets. An effective framework ensures clear delegation and separation of responsibilities and establishes the policymaking structure, pathways of accountability, and checks and balances associated with preserving and generating returns from reserves. It defines who makes decisions and who is responsible for them, as well as how they are made, and reflects country-specific institutional, social, and reg- ulatory considerations (de Abreau Faria and Ermes Streit 2016). One model for implementing effective governance is a “three-tier” structure comprising (1) a board, (2) investment committee, and (3) operational units. Under this framework, the board typically sets the policy parameters for reserve management and establishes the investment objectives and horizon, risk toler- ance, tranching criteria, and SAA. It formalizes these decisions through its approval of an investment policy and delegates oversight of reserve management to the investment committee. This body is responsible for setting and approving the investment guidelines, which operationalize the investment policy, includ- ing the SAA. Operational units are responsible for implementing the board’s pol- icy decisions and these guidelines. Some central banks use a two-tier approach without a separate investment committee. Under this framework, the board may be comprised of members with substantial technical skills and experience. In contrast, where a board’s  9 10 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks members represent stakeholders from a broader cross section of society, a formal investment committee may be needed to ensure access to the necessary exper- tise. These different frameworks suggest that a specific structure is less import- ant to effective governance than ensuring that decision makers have sufficient judgment and knowledge to execute their responsibilities. The survey results show that central banks use diverse arrangements to guide and implement their reserve management activities. Respondents divide these governance responsibilities among various institutional bodies in different ways. They also use distinct approaches to execute mandates like investment policy development and SAA. While almost all respondents (92 percent) indicated that a board approves the investment policy, their responses show that numerous entities are involved in its proposal and review (see figure 3.1). With respect to proposing the invest- ment policy, most (86 percent) answered that the operational department responsible for managing foreign exchange reserves proposes the policy and almost half (49 percent) indicated that the risk department also plays a role. When it comes to reviewing the investment policy, a little less than half of respondents (40 percent) use an investment committee while some indicated that the risk management department (24 percent), audit committee (19 ­percent) and/or risk committee (15 percent) also help to discharge this function. Most central banks (73 percent) responded that the board approves their SAA, while (28 percent) answered that an investment committee owns this responsibil- ity (see figure 3.2). The SAA—a central bank’s neutral asset allocation given its risk tolerance—is one of the most critical aspects of effective reserve management. Empirical evidence suggests that an SAA is the key driver of long-term investment success (Ibbotson and Kaplan 2000).1 A central bank’s objectives, risk tolerance and investment horizon all shape this model allocation. The frequent involvement of the board may show central banks’ keen understanding of its importance. The main risk arising from assigning SAA approval to an investment commit- tee is that the board may not understand the source of negative portfolio perfor- mance should market volatility cause returns that do not meet the organization’s long-term objectives. However, if the board does not have the financial expertise to understand the technical aspects of the SAA, delegating its approval to an FIGURE 3.1 Investment policy decision-making process Propose Review Approve No active role Reserves management department 86% 6% 4% Risk management department 49% 24% 16% Investment committee 27% 40% 19% 9% Board 1% 1% 92% 3% Governor 2% 13% 58% 14% Risk committee 3% 15% 7% 42% Audit committee 19% 53% Other 5% 3% 2% 7% N = 97 a Source: RAMP Survey on the Reserve Management Practices of Central Banks. a. “N” denotes the total number of institutions responding to the question, unless otherwise indicated. Results and Observations | 11 FIGURE 3.2 SAA decision-making process Propose Review Approve No active role Reserves management department 79% 8% 2% 5% Risk management department 46% 26% 1% 16% Investment committee 23% 34% 28% 8% Board 1% 6% 73% 13% Governor 1% 15% 54% 16% Risk committee 5% 13% 5% 43% Audit committee 18% 52% Other 7% 2% 6% N = 96 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.3 Organizational structure for reserve management responsibilities Reserves management and open market 63% - Yes 38% - No N = 88 operations in the same department Reserves management’s front, middle, and back 46% - Yes 54% - No N = 90 offices in the same department Reserves management's front and middle offices in the 36% - Yes 64% - No N = 86 same department, back office in a separate department Reserves management risk team is located in the 32% - Yes 68% - No N = 81 same department as enterprise risk management N = 98 Source: RAMP Survey on the Reserve Management Practices of Central Banks. investment committee may be appropriate. In this case, it is critical the commit- tee ensures the board understands the implied risk and return characteristics of the approved SAA. Survey results show that central banks also adopt diverse arrangements for the day-to-day management of foreign exchange reserves. These operations involve units that (1) initiate and execute trades and manage portfolios; (2) mea- sure and report on risk and performance; and (3) settle portfolio trades. These are often referred to as the respective responsibilities of the front office, middle office and back office, which suggests a strict division of responsibilities among units that does not exist in practice. Almost half of respondents indicated that they do not house these operations in separate departments and, in fact, combine front, middle, and back office functions in one unit (see figure 3.3). Motives for holding reserves Central banks have various motives for holding foreign exchange reserves. These include (1) self-insurance against external shocks; (2) conducting foreign exchange policy; (3) servicing external debt or other obligations; and (4) sup- porting monetary policy operations. These aims tend to shape components of 12 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks institutional investment policy and operations, including reserve adequacy determinations, investment objectives, currency composition, investment hori- zon, risk tolerance, and numeraire. Survey results show that self-insurance against external shocks is the primary motive for holding reserves, with most respondents (84 percent) considering it a highly relevant objective (see figure 3.4). Many also considered conducting for- eign exchange policy (66 percent) and servicing external debts or other obliga- tions (55 percent) as highly relevant motivations.2 Saving for intergenerational equity did not appear to be a major reason for central banks’ reserve management even with the substantial increase in global foreign exchange reserves. Very few (9 percent) identified it as a highly relevant objective.3 Respondents’ motives for holding foreign exchange reserves differ across country-income categories.4 All institutions in low-income and lower middle-­ income countries reported that these assets are highly relevant to insuring against external shocks and servicing external debt or other obligations (see figure 3.5). This finding is consistent with a view that these countries may be ­ FIGURE 3.4 Motives for holding foreign exchange reserves Motives Highly relevant Somewhat relevant Not relevant Provide self-insurance against potential external shocks 84% 10% 1% Conduct foreign exchange policy 66% 20% 12% Service external debt or obligations 55% 21% 16% Support monetary policy operations 27% 46% 17% Ensure savings for intergenerational equity 9% 26% 52% Other 4% 5% 5% N = 99 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.5 Motives for holding foreign exchange reserves by country-income group 100 100 100 100 Percent of country-income group 87 81 82 75 78 80 70 61 60 50 50 44 38 40 40 32 22 17 20 13 13 13 13 13 6 7 0 Provide self-insurence Conduct Service external Support Ensure savings for Other aganist potential foreign exchange debt or monetary policy intergenerational external shocks policy obligations operations equity Highly relevant N = 99 High-income (reserve) High-income (non-reserve) Upper middle-income Lower middle-income Low-income Source: RAMP Survey on the Reserve Management Practices of Central Banks. Results and Observations | 13 more vulnerable to contagion and, in certain periods, may have greater difficulty accessing capital markets. Most high-income non-reserve country institutions (81 percent) also consider self-insurance a critical objective. In contrast, less than half (44 percent) of central banks in high-income reserve countries deem it important, an outcome consistent with their option to print reserve currency in periods of market distress. Choice of reserve adequacy metric Like a central bank’s motives for holding reserves, the metrics it uses to evaluate the adequacy of these assets shape the course of its day-to-day operations. One measure may indicate its reserves are adequate while another may not. Much is at stake with this outcome, including the institution’s understanding of its ­ capacity to conduct currency interventions and diversify its reserve portfolios to include non-traditional asset classes. There are various methodologies that institutions can use to assess reserve adequacy. The most commonly adopted tend to be those related to trade and capital flows, such as import coverage and short-term external debt ratio. Beyond these basic assessments, more sophisticated methodologies incorporate combi- nations of these and other indicators (International Monetary Fund 2011, 2013). Survey results show that most respondents (88 percent) measure reserve ade- quacy in one way or another.5 There were 12 central banks that indicated they did not make this assessment, including seven institutions from middle-income countries. Respondents who assess reserve adequacy most frequently reported using the import coverage method (78 percent) followed by short-term external debt ratio (48 percent), broad money ratio (36 percent), and the IMF’s ARA metric (36  percent) (see figure 3.6). Of the 85 central banks who provided this FIGURE 3.6 Reserve adequacy measures 80 66 34 central banks use 70 three or more different measures Percent of total respondents 60 50 41 40 31 31 30 20 20 16 16 10 0 e tio tio ) er s s 13 ric el ag th ra ra od 20 et er O bt ey m m v or co de on n s 1 ve tio rt 01 m rm po er na (2 d te s Im oa re bi A t- Br m AR or al Co im Sh F IM pt O N = 85 Source: RAMP Survey on the Reserve Management Practices of Central Banks. 14 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.7 Reserve adequacy assessment by country-income group 100 100 83 82 Percent of country-income group 80 60 57 56 53 53 50 46 38 38 38 38 40 33 33 33 29 29 19 20 13 13 11 1313 6 6 6 7 7 4 0 Import Short-term Broad money IMF ARA metric Combination Optimal reserve Other coverage debt ratio ratio (2011 or 2013) metricsa models Measures N = 85 High-income (reserve) High-income (non-reserve) Upper middle-income Lower middle-income Low-income Source: RAMP Survey on the Reserve Management Practices of Central Banks. a. Expanded Greenspan-Guidotti rule or Wijnholds and Kapteyn. information, 62 percent reported using at least two metrics and 40 percent employed at least three. Central banks in low-income countries reported primarily using two method- ologies to measure reserve adequacy—import coverage and short-term debt ratio (see figure 3.7). All respondents in this group use import coverage, while some (33 percent) use short-term external debt ratio.6 These choices contrast with the practices of respondents in high-income reserve currency countries, who reported using these metrics infrequently if at all. This difference may be explained by borrowers in a reserve currency country having greater access to loans from international investors in their own domestic currency. Choice of numeraire The choice of numeraire, like the selection of reserve adequacy metrics, tends to frame how a central bank understands its reserve management operations. A numeraire is the specific currency or basket of currencies used to measure investment performance. Empirical evidence shows that different numeraires can produce different risks and returns for a portfolio. Therefore, the numeraire tends to influence a reserve manager’s optimal currency and asset allocation strategy (Papaioannou, Portes, and Siourounis 2006). Traditionally, a central bank’s choice of numeraire reflects one or more prior- ities. For example, an institution that seeks to execute an active foreign exchange policy would benefit from using as its numeraire the currency that it will deploy during an intervention. Another may select a basket of currencies as numeraire weighted according to the composition of trade and/or debt flows. Meanwhile, a central bank concerned with the impact on its balance sheet of fluctuations in the value of reserve assets may choose domestic currency as its numeraire.7 One seeking to accomplish multiple objectives may even utilize more than one numeraire for its reserve portfolios or may report its investment results in mul- tiple currencies.8 Results and Observations | 15 Survey results show that the U.S. dollar (62 percent) followed by domestic cur- rency (41 percent) are respondents’ most frequently used numeraire (see f ­ igure 3.8 below). This may represent a substantial change in practice since the global finan- cial crisis during which many central banks faced dollar liquidity challenges.9 Only 6 percent of central banks reported the use of a basket of currencies as a numeraire. It is possible that some respondents who indicated reporting in the U.S. dollar also track asset values in currency baskets weighted according to trade flows. Survey data also show almost all respondents in the low- and lower ­middle-income country groups (89 percent and 82 percent, respectively) use the U.S. ­dollar as numeraire (see figure 3.9 below). In contrast, a substantial majority of central banks in high-income reserve countries (81 percent) use domestic currency as their numeraire. ­ FIGURE 3.8 Numeraire Currencies USD 62% Domestic currency 41% EUR 13% Currency basket 6% Other currencies 5% SDR 3% 25 central banks selected two or more of the above currency options N = 98 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: Euro-area central banks that use the euro as numeraire are reported in the “domestic currency” basket only. Non-euro-area central banks that use the euro as numeraire are reported in the “euro” basket. SDR = special drawing rights. FIGURE 3.9 Numeraire choice by country-income category 100 89 82 Percent of country-income group 81 80 67 60 44 44 40 30 32 27 22 19 19 19 20 11 10 6 6 6 6 7 3 0 USD Domestic EUR Currency Other SDR currency basket currencies Currencies N = 98 High-income (reserve) High-income (non-reserve) Upper middle-income Lower middle-income Low-income Source: RAMP Survey on the Reserve Management Practices of Central Banks. 16 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks Investment principles of reserve management A central bank engages in reserve management to maximize the likelihood that it will have sufficient liquid foreign exchange assets to achieve its pol- icy objectives. Typically, its investment activities seek to strike a balance among three priorities—liquidity; safety/capital preservation; and returns/ income generation. These aims are complementary when interest rates are high enough to generate satisfactory positive returns for conservative strat- egies. However, in low or negative interest rate environments, they become less so because safe and liquid asset classes generate small or even below zero returns. Almost all respondents identified safety (97 percent) and liquidity (95 per- cent) as highly relevant investment principles (see figure 3.10 below). This is consistent with traditional approaches to reserve management that emphasize these priorities as fundamental. More than a third of respondents (37 percent) also identified returns/income generation as an important objective and almost all the others (60 percent) iden- tified it as somewhat relevant to their reserve management strategies. This focus on returns is less typical for central banks and may be a product of the prolonged period of unprecedented low interest rates for reserve currency assets in the aftermath of the global financial crisis. The focus on return generation appears to be the most pronounced for central banks in the high-income non-reserve country-income category. Overall, approximately a third of respondents within each country-income group identified investment growth as a highly relevant investment objec- tive. High-income non-reserve country central banks (56 percent) most ­frequently prioritized portfolio gains10 and almost half of low-income county respondents (44 ­percent) indicated they were highly focused on this priority. Our analysis did not find a relationship between the sufficiency of institu- tions’ import coverage and their consideration of return generation as highly ­relevant (see figure 3.11 below).11 Investment horizon and risk tolerance Investment horizon and risk tolerance are important policy parameters that influence how a central bank achieves its investment objectives. The former refers to the period over which an institution evaluates risk and performance. The latter defines a reserve manager’s overall appetite for investment risk and is FIGURE 3.10 Investment principles Principles Highly relevant Somewhat relevant Not relevant Safety (preservation of capital) 97% 2% Liquidity 95% 3% 1% Returns (income generation) 37% 60% 2% Other 1% 4% N = 99 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Results and Observations | 17 FIGURE 3.11 Central banks by country-income group that consider return generation highly relevant 31 0.21 4.7 Highly relevant 56 0.26 5.5 33 0.26 7.4 32 0.19 4.4 44 0.21 6.4 0 20 40 60 0 0.1 0.2 0.3 0 2 4 6 8 Percent of country-income group Average reserves-to-GDP ratio Average import cover in number of months N = 37 High-income (reserve) High-income (non-reserve) Upper middle-income Lower middle-income Low-income Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: GDP = gross domestic product. FIGURE 3.12 Investment horizon of investment tranche Investment horizon (years) 0–<1 2 1 28 2 6 3 33 4–5 19 6–10 8 >10 4 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 N = 90 Percent of total Source: RAMP Survey on the Reserve Management Practices of Central Banks. determined according to its ability to withstand asset volatility over the period in which it evaluates risk-adjusted performance. Survey results show that most respondents (70 percent) have adopted an investment horizon of more than 1 year for foreign exchange reserves allocated to portfolios focused on income generation/returns (see figure 3.12). All things being equal, extending this risk and performance measurement period beyond 1 year allows a central bank to adopt an asset allocation that incorporates more risk and, as result, has a greater opportunity to generate income. Institutions in high-income countries reported having longer investment horizons than coun- terparts in middle-income countries. A central bank may use various metrics to help set and express its risk toler- ance. These include probability of negative returns, expected shortfall, and value at risk (VaR). Using probability of negative returns allows an institution to clearly communicate its capital preservation priorities whereas other risk measures are less easy to translate into public discourse. Survey results show that, for the liquidity tranche, 56 percent of respondents use probability of negative returns, 43 percent VaR, and 41 percent expected shortfall. For the investment tranche, the use of the three metrics is similarly frequent (between 41 percent and 47 percent). 18 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks STRATEGIC ASSET ALLOCATION As discussed above, empirical evidence shows that the SAA is the primary driver of a reserve manager’s investment performance. Building an SAA involves multiple steps that aim to translate investment policy into an asset allocation that achieves a central bank’s investment objectives over the appli- cable investment horizon. As a first step, an institution decides whether to use tranching as a tool to build its SAA. Next, it identifies eligible currencies and asset classes. Only investments denominated in these currencies and matching these financial instruments may be included in its reserve portfolio. The use of tranching One common tool for constructing an SAA is “tranching.” In this approach, a central bank segregates foreign exchange reserves into discrete sub-portfolios. The structure and relative size of each of these “tranches” is based on an assessment of liquidity needs across various time horizons and reserve ade- quacy scenarios. Each segregated account is characterized by a distinct objec- tive, risk profile, set of eligible asset classes, currency composition and investment horizon (International Monetary Fund 2015).12 Most respondents (80 percent) reported using tranching. Less than a third of high-income reserve country central banks used this approach as part of their SAA (see figure 3.13). Institutions in countries with non-reserve currencies of all income levels were far more likely to use the practice (between 75 percent and 96 percent). FIGURE 3.13 Use of tranching by country-income group 100 96 87 89 Percent of country-income group 80 75 60 40 31 20 0 e) ) e e e ve m m m rv er co co co se es in -in -in (re -r e- le w on e dl d Lo m (n id id co rm m e -in m er pe co h w ig Up -in Lo H h ig H N = 78 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Results and Observations | 19 Eligible currencies and actual currency composition of reserves Almost all respondents identified as eligible currencies the U.S. dollar percent) and the euro (88 percent) (see figure 3.14). Many indicated (98  ­ that they could invest in assets denominated in the British pound sterling (68 ­percent) and the Japanese yen (55 percent) while almost half (49 ­ percent) reported that the renminbi is part of their currency composition. Its inclusion in the special drawing rights (SDR) basket in October 2016 may drive this result. The renminbi eligibility data are notable because they suggest that, in terms of asset-denomination preference, the currency is on par with or may even have surpassed others such as the Australian ­ dollar and the Canadian ­dollar. Analyzing the data by country-­ income category or level of reserve adequacy does not yield materially ­significant patterns. Macroeconomic considerations and portfolio management concerns drive the actual currency composition of reserve portfolios.13 The former comprise the structure and denomination of external debt, intervention needs, and asset and liability management. The latter consist of the diversification of currency risk and the returns, availability, and liquidity of assets denominated in differ- ent currencies. Intervention requirements and payment of external debt claims tend to be more relevant to a liquidity tranche. Meanwhile diversification of currency risk and pursuit of higher returns play a more important role in shap- ing the investment tranche. The survey results are consistent with this under- standing (see figure 3.15). The U.S. dollar and the euro are the dominant currency denominations for central banks’ reserve assets.14 Figure 3.16 shows the distribution range of the currency composition of all respondents’ foreign exchange holdings, including those with a zero allocation. For each currency, it displays the range of institu- tions’ reported shares and quartiles, as well as the median and average. FIGURE 3.14 Percentage of respondents that may hold each currency as part of their foreign exchange reserves 100 98 88 80 68 60 55 Percent 49 48 44 40 23 21 20 16 16 15 14 11 11 0 USD EUR GBP JPY RMB AUD CAD CHF Other NOK NZD SEK ZAR KRW SGD N = 94 Source: RAMP Survey on the Reserve Management Practices of Central Banks. 20 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.15 Factors shaping currency composition Factors Liquidity tranche Investment tranche Intervention needs/result 72% 15% Payment of external debt claims 72% 25% Asset liability management 67% 50% Diversification of currency risk 38% 67% Size and liquidity of the currency market 28% 38% Pursuit of higher returns 11% 82% Other factors 7% 6% N = 77 N = 76 N = 72 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.16 Distribution of all respondents’ allocations to individual currencies 100 90 80 Distribution of allocations (percent) 70 60 50 40 30 20 10 0 USD EUR GBP AUD RMB JPY CAD NZD Other Number of central 2 20 38 53 58 59 62 75 44 banks with 0% allocation N = 81 2nd quartile 3rd quartile Average Source: RAMP Survey on the Reserve Management Practices of Central Banks. The median for the proportion of assets invested in U.S. dollar-denominated securities is 68 percent and for euro-denominated assets it is 9 percent. average proportion for euro-denominated holdings (25 percent) is signifi- The ­ cantly higher than the median, suggesting that a few institutions’ euro assets skew the average higher. Despite recent changes in the relative sizes of large economies, these denominated results are consistent with historical data showing U.S. dollar-­ assets’ decades-long dominance of reserve portfolios (International Results and Observations | 21 Monetary Fund 2019). Most trade and capital flows still take place using the U.S. dollar or the euro, which may explain their predominance in reserve positions, even though, over the long term, the US and euro-area economies’ shares of global gross domestic product have been in decline.15 These fac- tors may also partially explain respondents’ comparatively large holdings of euro-denominated assets relative to those denominated in non-U.S. dollar currencies, even though many euro-denominated financial instruments have negative yields. Survey data indicates that multiple currencies that central banks c onsider eligible from an investment perspective nonetheless do not ­ comprise substantial amounts of their portfolios. Figure 3.17 comprises i nformation gleaned from the reported currency composition of respon- ­ dents’ ­foreign exchange reserves and shows only data for institutions that indicated exposure to a specific legal tender. (It therefore does not reflect the impact of respondents who reported a zero allocation.) For each cur- rency, it reports the number of central banks that had an allocation and displays the range of reported shares and quartiles, as well as the median and average. The divergence between eligible currencies and actual holdings is most visi- ble for the Australian dollar, the Canadian dollar, the British pound sterling, the Japanese yen and the renminbi. All have eligibility frequencies of between 44 percent and 68 percent. However, none has a median share of exposure above 5 percent. These results suggest that any shift away from the U.S. dollar or the  euro as dominant reserve currencies is, at present, not substantial. FIGURE 3.17 Distribution of allocations to currencies for those central banks with exposure 100 80 Distribution of allocations (percent) 60 40 20 0 USD EUR GBP AUD RMB JPY CAD NZD Other Non-zero 79 61 43 28 23 22 19 6 37 allocations N = 81 2nd quartile 3rd quartile Average Source: RAMP Survey on the Reserve Management Practices of Central Banks. 22 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.18 Net number of central banks that plan to increase or decrease an allocation to specific currencies within 2 years 15 12 11 10 Net count 5 0 –2 –5 –5 USD RMB GBP EUR N = 68 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: The “net count” for a specific currency is calculated by subtracting the number of central banks that reported an intention to decrease their allocation from the number of institutions that intend to increase their allocation. Central banks that reported “no intended changes” to a given currency allocation have no impact on the net count. Their average and median levels for actual allocations remain high while shares of reserve assets denominated in other currencies are low. The Japanese yen aside, the relative size or liquidity of these moneys’ capital markets may drive their small share.16 Challenges associated with accessing renminbi-denominated assets and capital flow management measures in China may also contribute to the currency’s reported allocations. There is some evidence suggesting that, in the near term, central bank allocations to the renminbi are likely to rise. Respondents from 68 institu- tions provided data on expected changes to their actual currency composi- tion over the next 2 years. Many of those with plans to change their allocations reported a likelihood of increasing shares of assets denominated in the U.S. dollar and the renminbi while some anticipated reducing expo- sure to those denominated in the British pound sterling and the euro figure 3.18). (see ­ Eligible asset classes and actual asset class composition of reserves Central banks have broad authorization to purchase traditional reserve man- agement investments. Respondents most frequently cited as eligible asset classes financial instruments generally considered to be highly liquid and low- risk—government bonds (96 percent); sovereigns, supranational and agency securities (85 percent); and money market instruments (79 percent) (see ­ f igure 3.19). This data is consistent with the investment aim of capital pres- ervation and liquidity, which 95 percent of institutions identified as a highly relevant policy objective. A high proportion of respondents (80 ­ percent) also indicated that they were able to hold bank deposits. While safe and convertible Results and Observations | 23 FIGURE 3.19 Eligible asset classes Government bonds (88) 96 SSA bonds (78) 85 Bank deposits (74) 80 Money market insturments (73) 79 Riskier asset classes (54) 59 Gold (47) 51 Other (13) 14 0 20 40 60 80 100 N = 92 Percent Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: The category of “riskier asset classes” comprises corporate bonds, emerging market bonds, covered bonds, mortgage-backed securities, equity, asset backed securities and emerging market equity. SSA = supranational, sovereign, and agency. in the very short-term, when held with commercial banks for terms longer than overnight, these assets take on counterparty risk and have less liquidity than government bonds. More than half of respondents (59 percent) reported having the ability to invest in “riskier asset classes,” suggesting a willingness to increase the diver- sification of their reserve holdings in pursuit of higher returns. Within this category, the most frequently cited financial instruments were corporate bonds (56 percent), emerging market bonds (44 percent) and covered bonds (39 percent) (see figure 3.20). Even as some institutions show a willingness to diversify their reserve port- folios, respondents most frequently identified traditional asset classes as their dominant investment choice (see figure 3.21). Figure 3.21 below shows the distri- bution range of the asset allocation for respondents’ foreign exchange reserves, including those with a zero allocation. For each category of financial instru- ments, it displays the range of institutions’ reported shares and quartiles, as well as the median and average. The most frequently held securities are government bonds and suprana- tional, sovereign, and agency (SSA) securities, followed by bank deposits and money market instruments—holdings broadly consistent with the ­ p olicy objective of safety and liquidity. The median allocation to these investments is 40 percent, 6 percent, 11 percent, and 7 percent respectively, including those institutions that do not have exposure to these asset classes. In contrast, the median allocation to riskier asset classes such as corporate bonds, emerging markets bonds, and equity is 0 percent, suggesting that 24 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.20 Higher risk financial instruments that are eligible asset classes Corporate bonds (30) 33 EM bonds (24) 26 Covered bonds (21) 23 MBS (16) 17 Equity (16) 17 ABS (9) 10 EM equity (5) 5 0 5 10 15 20 25 30 35 N = 92 Percent Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: ABS = asset-backed securities; EM = emerging market; MBS = mortgage-backed securities. FIGURE 3.21 Distribution of all respondents’ allocations to individual asset classes 90 85 80 75 Asset allocations as of September 2017 (percent) 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 Gov. Bank Money SSA Equity Other Corporate Gold EM bonds Covered ABS MBS EM equity bonds deposits Mkt. Inst. bonds bonds bonds Number of central 3 13 21 9 43 40 36 30 43 43 50 44 51 banks with 0% allocation N = 51 2nd quartile 3rd quartile Average Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: ABS = asset-backed securities; EM = emerging market; MBS = mortgage-backed securities; SSA = supranational, sovereign, and agency. Results and Observations | 25 FIGURE 3.22 Distribution of the allocation to individual asset classes for respondents with exposure 90 85 80 75 Asset allocations as of September 2017 (percent) 70 65 60 55 50 45 40 35 30 25 20 15 10 5 0 Gov. Bank Money SSA Equity Other Corporate Gold EM bonds Covered ABS MBS bonds deposits Mkt. Inst. bonds bonds bonds Non-zero 48 38 30 42 8 11 15 21 8 8 1 7 allocations N = 51 2nd quartile 3rd quartile Average Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: ABS = asset-backed securities; EM = emerging market; MBS = mortgage-backed securities; SSA = supranational, sovereign, and agency. these financial instruments are still of limited or no importance for many central banks. Central banks that have investments in riskier asset classes tend to limit these holdings to small shares of their reserve portfolios. Figure 3.22 below is com- prised of information gleaned from the reported asset allocations of respon- dents’ foreign exchange reserves and shows only data for the institutions that indicated exposure to a specific asset class. (It therefore does not reflect the impact of central banks who reported a zero allocation.) For each financial instrument, it displays the range of institutions’ reported shares and quartiles, as well as the median and average. Of the 51 central banks that provided allocation data, less than one third (29 percent) hold corporate bonds and less than a fifth (16 percent) hold equities. For both asset classes, the median allocation is below 10 percent. Survey data does not show a material relationship between respondents’ allocations to traditional asset classes and measures of reserve adequacy (see figures 3.23 and 3.24 and figure A.1–A.4 in appendix A). Analysis did not sug- gest an obvious pattern between reserve levels and reported holdings of bank deposits, government bonds, money market instruments, gold, and sovereign, supranational and agency bonds. This conclusion holds when using months of import coverage as a measure of sufficiency of foreign exchange reserves or other metrics, and also when the data analyzed comes only from respondents in emerging markets. 26 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.23 Comparison of allocation to traditional asset classes and months of import coverage of foreign exchange reserves 18 16 14 Import cover (months) 12 10 8 6 4 2 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 50 Source: World Bank and RAMP Survey on the Reserve Management Practices of Central Banks. Note: Traditional asset classes comprise bank deposits, government bonds, money market instruments, gold, and supranational, sovereign, and agency bonds. FIGURE 3.24 Comparison of allocation to traditional asset classes and months of import coverage of foreign exchange reserves of emerging market respondents 18 16 14 Import cover (months) 12 10 8 6 4 2 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 36 Source: World Bank and RAMP Survey on the Reserve Management Practices of Central Banks. Note: Traditional asset classes comprise bank deposits, government bonds, money market instruments, gold, and supranational, sovereign, and agency bonds. Results and Observations | 27 PORTFOLIO MANAGEMENT Portfolio management style A reserve manager has the option of using different investment styles to imple- ment its SAA. A central bank’s risk tolerance and capacity for risk and portfolio management may have an impact on its manager’s ultimate choice. Styles differ in several ways, including whether they are passive or active. The former rep- licate the risk and return characteristics of a specific benchmark; the latter allow for discretionary departures from these standards within defined risk limits in pursuit of higher returns than the benchmark. Compared to a ­ passive approach, active management is more resource- and skill-intensive, demand- ing more expertise and support in both portfolio and risk management. As a result, it may be beyond the capacity of some institutions to pursue. Central banks that do not use a benchmark to guide allocations tend to invest in time deposits and money market instruments or fixed income assets that they hold to maturity. Most central banks adopt portfolio management styles that use active approaches (see figure 3.25 below). Three-quarters of respondents indicated implementing their SAA using an active style (49 percent) or enhanced index- ation (26 percent). The remaining use passive management (10 percent), a buy- and-hold strategy (7 percent) or invest in only time deposit and money market instruments (6 percent). External management of reserve assets One option a central bank can use to address the skill and resource demands associated with active management is to employ an external manager to FIGURE 3.25 Portfolio management styles Time deposits and money market instruments, Other, 2% 6% Buy and hold, 7% Passive (indexing), 10% Active, 49% Enhanced indexing, 26% N = 98 Source: RAMP Survey on the Reserve Management Practices of Central Banks. 28 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks implement a part of its SAA. Almost three quarters (72 percent) of the 94 respon- dents providing data on the use of third-party service providers reported using organizations outside their institutions for investment management services. These results show central banks in low-income countries (78 percent) and high-income non-reserve countries (88 percent) doing so far more frequently than counterparts in high-income reserve countries (25 ­ percent) (see figure A.5 in appendix A). Other reasons may also motivate a central bank to use an external manager. For example, since these service providers often have significant expertise, their support can create opportunities to share knowledge and build internal capacity. In addition, there is a perception that they may be able to deliver higher returns. At minimum, these third parties can serve as a standard against which to mea- sure the results of internally-managed portfolios. Generating higher returns, however, does not appear to be the most important factor. Of the respondents who use external managers, 87 percent cited as highly relevant to their decision the possibilities of knowledge trans- fer and capacity building (see figure 3.26). A slightly smaller share (78 ­percent) cited performance considerations as an important driver of their choice. Even though engaging external managers is common, central banks generally apportion them only a small fraction of their assets. Of the 70 respondents who indicated using third parties as investment advisors, a majority (56 percent) tasked them with overseeing less than 10 percent of their portfolios (see ­ figure 3.27). Less than one fifth (17 percent) enlisted them to manage more than 30 percent. Use of derivatives Survey results indicate that central banks deploy derivatives extensively (see figure 3.28). They appear to use them primarily for hedging purposes and, to a ­ lesser extent, for active management. The 74 respondents who provided data on this issue most frequently employ FX forwards, interest rate futures/bond futures, and FX swaps (61 percent, 59 percent and 54 percent respectively). Other types of derivatives are not yet widely utilized. FIGURE 3.26 Reasons for hiring external managers Factors Highly relevant Somewhat relevant Not relevant Capacity building/Knowledge transfer 87% (69) 9% (7) 3% (2) Enhancement of return/Performance 78% (62) 16% (13) 1% (1) Internal constraints (e.g. management skills) 42% (33) 39% (31) 13% (10) Increased access to resource-intensive investment strategies 35% (28) 48% (38) 8% (6) Benchmark to internal staff 25% (20) 43% (34) 23% (18) Other 4% (3) 1% (1) 3% (2) N = 79 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Results and Observations | 29 FIGURE 3.27 Foreign exchange reserve allocations to external managers 35 23 30 Percent of respondents 25 16 20 12 15 10 6 5 4 4 5 0 0–<5% 5–10% 10–15% 15–20% 20–25% 25–30% >30% N = 70 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.28 Use of derivatives Derivatives FX forwards 61% Interest rate/bond futures 59% FX swaps 54% FX options 16% Interest rate swaps 16% Currency futures 8% Gold swaps 7% Other 7% Interest rate options 5% Currency options 4% Credit default swaps 4% Gold options 3% Asset swaps 1% None of the above 19% N = 74 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: FX = foreign exchange. Consideration of environmental, social, and governance (ESG) factors ESG factors are tools that can help an asset manager assess environmental, social and governance aspects of an asset issuers’ operations. They aim to eval- uate how a borrower or company is run and the impact of its business practices on society. Examples include a corporation’s carbon footprint, employment practices and governance framework. Use of these factors is an evolving 30 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks investment approach that is currently most developed in its application to public equities and corporate bonds. ­ Although central banks have some awareness of ESG factors, these tools do not play a significant role in their investment frameworks. Only 11 percent of respondents confirmed taking them into account (see figure 3.29). Meanwhile 68 percent indicated they play no role in their current approach to reserve management. Respondents with equity exposures are more likely to use ESG factors as part of their investment framework (see figure 3.30). This result does not hold for those central banks investing in corporate bonds (see figure A.7 in appendix A). FIGURE 3.29 Incorporation of ESG factors into the investment framework Yes, 11% No, 46% Yes, but not explicitly, 20% No, but there have been No, but may consider in the discussions, next two years, 18% 4% N = 98 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.30 ESG and equity investing 50 47 44 38 40 30 Percent 22 21 20 13 10 6 5 4 0 Yes No Yes, No, No, but may but not but there consider in explicitly have been the next discussions 2 years N = 92 Equity No equity Source: RAMP Survey on the Reserve Management Practices of Central Banks. Results and Observations | 31 RISK MANAGEMENT A central bank faces a variety of risks in its reserve management operations. Two of the most important are credit-related hazards and market-related issues. The former arise from the possibility of loss due to an obligor’s deteriorating credit quality, most often in the form of a rating downgrade or default. This danger includes counterparty concerns, making it especially relevant to corporate bond investors and holders of term deposits with commercial banks. The latter stem from the possibility that the price of an asset will decline due to market factors. A comprehensive risk management framework helps an institution identify and assess the magnitude of these threats and maintain them within limits consistent with its tolerance. Credit risk management Most central banks appear to have a relatively conservative approach to fixed income investing based on their minimum credit rating thresholds (see figure 3.31). Almost all respondents (98 percent) indicated that they could not ­ invest in sovereign or corporate debt that is less than investment grade. They most frequently cited the rating group of “A+/A/A” as the acceptable minimum for both government bonds (44 percent) and corporate bonds (31 percent). However, similar ratings across different asset classes do not signal the same level of credit risk. For example, corporate debt with the same credit rating as sovereign debt has a higher likelihood of default.17 Since non-public debt tends to have a higher risk profile than public obligations, it may be appropriate to use a different risk tolerance and additional expertise when measuring and managing its credit risk. A reserve manager can use numerous methodologies to develop an under- standing of the credit risk in its portfolio. While almost all respondents reported using credit ratings (96 percent), many (60 percent) indicated using more than one metric to measure and manage the likelihood of default on the obligations in their reserve holdings (see figure 3.32). FIGURE 3.31 Minimum credit ratings by asset class Government Supranational/ Corporate MBS/ABS Ratings bonds agency bonds bonds AAA 3% (3) 5% (4) 8% (3) 26% (9) AA+/AA/AA- 20% (18) 36% (31) 23% (9) 24% (8) A+/A/A- 44% (40) 38% (32) 31% (12) 24% (8) BBB+/BBB/BBB- 29% (26) 21% (18) 33% (13) 15% (5) BB+ and below 1% (1) 3% (1) Other 2% (2) 3% (1) 12% (4) Total 100% (90) 100% (85) 100% (39) 100% (34) N = 91 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: ABS = asset-backed securities; MBS = mortgage-backed securities. 32 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.32 Credit assessment methodologies Credit rating from rating agency 96% (94) Market indicators (bond spread, credit default swap spread, equity price, etc.) 43% (42) Internal scoring (rating) model 37% (36) Quantitative models (e.g., Moody's Expected Default Frequency and others) 16% (16) Other 7% (7) 30 central banks use three or more methodologies N = 98 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.33 Credit assessment methodologies for respondents with corporate credit exposure Credit rating from rating agency 100% (30) Market indicators (bond spread, credit default swap spread, equity price, etc.) 37% (11) Internal scoring (rating) model 20% (6) Quantitative models (e.g., Moody's Expected Default Frequency and others) 10% (3) Other 7% (2) 14 central banks that invest in corporate bonds use only credit ratings from rating agencies N = 30 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.34 Credit analysis techniques of respondents with minimum credit ratings of BBB- for at least one asset class Credit rating from rating agency 93% (28) Market indicators (bond spread, credit default swap spread, equity price, etc.) 23% (7) Internal scoring (rating) model 27% (8) Quantitative models (e.g., Moody's Expected Default Frequency and others) 13% (4) Other 7% (2) N = 30 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Of the 30 respondents that invest in corporate bonds, almost half (47 percent) or 14 rely only on credit ratings from external agencies (see f ­ igure 3.33). At least 6 of these 14 central banks are authorized to invest in corporates with ratings as low as BBB+/BBB/BBB−. Most respondents with authorization to invest in BBB- rated instruments for at least one asset class use credit ratings to assess credit risk (see ­figure 3.34). Within the group of 30 institutions that invest in corporate bonds, the use of other credit assessment methodologies, such as market indicators, internal scoring models and quantitative models, is less frequent than it is across the complete sample. Results and Observations | 33 Market risk management A reserve manager can use numerous methodologies to develop an under- standing of the market risk in its portfolio. Most respondents indicated using duration limits (87 percent) and currency limits (81 percent) (see figure 3.35). The use of probabilistic risk measures such as Tracking Error (47 percent) and CVaR/VaR (34 percent) are less common. Respondents do not appear to adjust their market risk measurement prac- tices according to their investment styles. The subset of central banks that use active or enhanced indexing styles use the different risk metrics with close to the same frequency as those that do not use active management strategies (compare figure 3.35 with figure 3.36). This suggests that some central banks may be able to gain a deeper under- duration and cur- standing of the market risk in their portfolios. Specifically, ­ rency limits are frequently deployed by fixed income investors to measure market risk and are relatively easy to calculate. However, these metrics only capture one risk factor (namely changes in interest rates or exchange rates) and are difficult to compare with other risks in a portfolio. When an asset pool is exposed to other sources of risk, such as credit risk, optionality, and/ or equity, probabilistic measures are better at measuring market risk because they also consider the correlation between different risk factors. These mea- sures require advanced analy­ t ics, which may explain why they are not as widely adopted as duration and currency limits. FIGURE 3.35 Metrics for measuring market risk Duration limit 87% (83) Currency limit 81% (77) Tracking error limit 47% (45) CVaR/VaR limit 34% (32) Other 4% (4) 49 central banks use three or more limits N = 95 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: CVaR = conditional value at risk; VaR = value at risk. FIGURE 3.36 Market risk measurement metrics for respondents that use active or enhanced indexing styles Duration limit 88% (64) Currency limit 82% (60) Tracking error limit 52% (38) CVaR/VaR limit 37% (27) Other 5% (4) N = 73 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: CVaR = conditional value at risk; VaR = value at risk. 34 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE 3.37 Types of stress scenarios employed Custom designed scenarios 56% (35) Custom designed scenarios with frequent updates 17% (11) Risk system default scenarios (pre-canned scenarios) 16% (10) Both risk system default scenarios and custom designed scenarios 8% (5) Other 3% (2) N = 73 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Stress test and scenario analysis A reserve manager can use stress testing and scenario analysis to enhance risk management because these tools address issues that cannot be captured by tra- ditional risk models and metrics. A stress test, for example, is not bounded by recent or historical market data calibration and return distribution assumptions. This flexibility allows for it to be tailored to various market scenarios, making it a particularly effective tool for uncovering portfolio risk during a period of distress. Approximately two-thirds of respondents reported using stress testing. When analyzed by country-income group, the data show that 80 percent of high-­ income (reserve) and 83 percent of upper middle-income country central banks apply stress tests, whereas 89 percent of low-income country respondents do not. For those institutions that apply stress tests, the majority use custom designed scenarios (56 percent), followed by custom designed scenarios with frequent updates (17 percent) and risk system default scenarios (16 percent) (see figure 3.37). PERFORMANCE REPORTING AND TRANSPARENCY Internal reporting on portfolio risks and returns helps a central bank evaluate the soundness of its investment policy and promote accountability. This pro- cess enables a reserve manager to test the reasonableness of assumptions underlying its choice of benchmarks, eligible asset classes and investments. Through performance attribution, it also allows an institution to identify the key decisions driving results (Bailey, Richards and Tierney 2018).18 A board can also use reporting data to verify that managers are implementing their invest- ment programs effectively and prudently. Where this evidence suggests weak- nesses either in the investment policy or its implementation, changes can be made to assist the institution in its efforts to achieve its long-term objectives. Most respondents (78 percent) indicated that they generate these statistics either on a monthly or a quarterly basis (see figure 3.38). Two-thirds conduct performance attribution and half of these do so using a factor-based model (see figure A.6 in appendix A). Sharing performance-related data externally can bolster confidence in public asset managers, buttressing a central bank’s credibility, legitimacy, and independence. National regulations sometimes set minimum levels of transpar- ­ ency by laying out specific categories of information that must be disclosed. In some cases, institutions choose to produce information beyond that required by law. Results and Observations | 35 FIGURE 3.38 Highest frequency of reporting to the board Quarterly 46% Monthly 32% Annually 8% Daily 6% Other 5% Weekly 2% Total 100% N = 97 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE 3.39 Mandatory or voluntary disclosure of reserve management performance Disclosures Yes No Currency composition 65% 34% Asset classes 63% 35% Performance 53% 42% Investment universe 51% 45% Institutional regulation (e.g., investment guidelines) 35% 60% Risk metrics (e.g., duration, tracking error and VaR) 33% 59% External managers 28% 59% 28% Benchmarks 64% 27% Investment horizon 66% N = 97 Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: VaR = value at risk. Survey results show that respondents take different approaches to producing information on reserve management activities for outside audiences. More than half disclose their currency composition (65 percent) and approved asset classes (63 percent) (figure 3.39). A majority publish their performance statistics (53 ­percent) and investment universe (51 percent). In contrast, only about a third release information on their investment guidelines (35 percent) and benchmarks (28 percent). NOTES  1. The SAA is often referred to as the “neutral” long-term asset allocation because it does not reflect short-term views on the trajectory of any asset class or currency (Cardon and Joachim 2004).  2. There were 99 central banks that responded to this question. Currency regime data is avail- able for 87 of the countries in which they reside. Of these 87, 46 have free-floating regimes (International Monetary Fund 2018). An analysis that assessed these responses according to foreign exchange regime type and monetary policy factors did not find any material correlation. 36 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks  3. An analysis did not show a meaningful relationship between central banks that identified this motive as highly relevant and their reserve-to-GDP ratio or country-income category.  4. An analysis of this data using various reserve adequacy metrics did not produce notewor- thy findings.  5. All but one of the survey’s 99 participants answered questions related to reserve adequacy measurement. The 88 percent is based on the number of respondents to the question not the total number of survey participants.  6. The short-term external debt ratio methodology aims to ensure sufficient foreign exchange liquidity to cover external debt in the event of a sudden loss of access to international cap- ital markets.  7. An institution’s accounting policy also has an impact on the way currency fluctuations affect the value of foreign exchange reserves.  8. For example, a central bank may divide its reserve holdings into separate tranches, each configured to achieve a different investment objective. The tranches may use different numeraires consistent with their specific aim.  9. This result is substantially different from a Bank of International Settlements’ survey con- ducted in 2008, which found that less than a third of respondents used the U.S. dollar as currency numeraire (Borio et al. 2008). This apparent change in reserve management prac- tice may be a lesson learned from the crisis. 10. Neither respondents’ reserve-to-GDP ratio nor ARA level appear to explain the focus on returns/income generation. This outcome does not change when the analysis is done using only institutions within the same country-income group. 11. Short-term debt-to-GDP ratio was also used to test this hypothesis and no relationship was found. 12. Tranching allows the asset allocation of foreign exchange reserves to be dedicated to a specific investment objective. Separating assets into these discrete categories enables a reserve manager to confirm the allocation’s risk-adjusted basis is appropriate to its objec- tive. Where a reserve manager uses multiple tranches, the different asset allocations com- bine to form a single overall portfolio. It is therefore critical to account for correlations across the tranches to achieve an optimal and efficient overall allocation. 13. Some research has found that the currency composition of reserve portfolios depends mainly on trade flows, financial flows, and currency pegs (Eichengreen and Mathieson 2000). In some cases, the currency used to measure portfolio risk and return may also drive currency composition (McCauley 2008; Borio et al. 2008). 14. Overall, 81 central banks provided this information. They provided the actual weightings of their currency composition as of September 30, 2017. 15. According to the Society for Worldwide Interbank Financial Telecommunications (SWIFT), U.S. dollar- and euro-denominated transactions comprised over 80 percent of cross-border payments in 2017 (SWIFT 2018). Research suggests that the drivers of cur- rency composition are based on long-term or mid-term macroeconomic perspectives and country specific considerations (Dooley, Lizondo, and Mathieson 1989; Eichengreen and Mathieson 2000; McCauley and Chan 2014). These tend to change only gradually, which may also explain the reason for the U.S. dollar and the euro’s continued dominance of for- eign exchange reserves. 16. Respondents’ reported reserve shares by currency as of September 31, 2017, is similar to IMF data on the Currency Composition of Official Foreign Exchange (COFER). The IMF shows a 64 percent share for the U.S. dollar, a 19.7 percent share for the euro, a 4.4 percent share for the British pound sterling, a 4.2 percent share for the Japanese yen, a 2 percent share for the Canadian dollar, a 1.8 percent share for the Australian dollar, and a 1.1 percent share for the renminbi as of September 31, 2017 (International Monetary Fund 2019). 17. According to Fitch Ratings, one-year global corporate finance default rates were positive between 1990 and 2014 (0.11 percent, 0.03 percent, 0.07 percent, and 0.17 percent for AAA-, AA-, A-, and BBB-rated debt, respectively), while they were 0 percent for sovereigns over a similar period (from 1995 to 2014) (Fitch Ratings 2015). 18. “Performance attribution provides an informed look at the past. It identifies the sources of difference-from-benchmark returns (differential returns) and their impacts on an account’s performance.” Fitch Ratings 2015. 4 Concluding Commentary Over the past 20 years, managers of foreign exchange reserves have had to respond to two major market developments—a substantial increase in the amount of these assets globally and the extraordinary policy responses to the unprecedented macroeconomic and investment environment after the global financial crisis. The survey’s key findings suggest that, despite these factors, most central banks continue to employ a traditional reserve management approach. Their investments remain concentrated in high-quality fixed-income assets and the minimum credit rating for these holdings remains conservative. At the same time, the data suggest that important changes are underway as a material number of central banks reported more diversified portfolios with exposure to non-traditional asset classes. A third of respondents hold corporate credit, most of which is investment grade, and almost one in five own mortgage-backed securities or equities, although mostly in limited allocations. ­ Our analysis of this information did not find a relationship between respon- dents’ reserve adequacy and the size of their exposure to non-traditional asset classes. The data exhibit considerable cross-country differences in the way central banks manage their reserves and, in some circumstances, our analysis ­ suggests these differences correlate with respondents’ country-income groups. These trends suggest at least three areas of opportunity for central banks to enhance their reserve management operations—first and foremost, they high- light the need for further development of strong governance frameworks con- sistent with existing institutional arrangements. Here the goal is to adhere to a principal-based governance structure that achieves a clear delegation of responsibilities and a separation of functions, and establishes adequate over- sight. As central banks invest in more asset classes, a strengthened governance framework should generate investment decisions that are more likely to yield outcomes that are consistent with an institution’s risk tolerance. Second, risk management should become more robust in response to the increase in actively managed portfolios and the greater diversification in cur- rency and asset allocations. Central banks that are increasing their credit expo- sure are encouraged to revisit their process of credit risk management by  37 38 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks improving their analysis of bond issuers, counterparties, and collateral. In addi- tion, market risk methodologies, such as stress tests and scenario analyses, need to be at the center of this framework to complement the limitations of traditional risk models and to account for potential tail risks. Third, enhanced transparency may be necessary to retain and enhance cen- tral bank legitimacy because larger and more diversified foreign exchange reserves mean their operations take on more risk. Central banks need to find the proper balance between providing enough information to relevant stakeholders and maintaining an efficient investment operation. Country-specific consider- ations, like the mix of institutional and political factors, can influence the degree of disclosure. However, more and better public disclosure is likely to increase the effectiveness of reserve management and the credibility of a central bank, as its stakeholders and the public develop a better understanding of its investment objectives, policy and track record. Appendix A FIGURE A.1 Respondents’ foreign exchange reserve allocation to traditional asset classes and level of short-term external debt coverage 20 Ratio of reserves to short-term external debt 18 16 14 12 10 8 6 4 2 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 33 Source: RAMP Survey on the Reserve Management Practices of Central Banks, IMF, and CEIC.  39 40 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE A.2 Emerging market respondents’ foreign exchange reserve allocation to traditional asset classes and level of short-term external debt coverage 20 Ratio of reserves to short-term external debt 18 16 14 12 10 8 6 4 2 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 21 Source: RAMP Survey on the Reserve Management Practices of Central Banks, IMF, and CEIC. FIGURE A.3 Respondents’ foreign exchange reserve allocation to traditional asset classes and reserve cover of ARA metric level 2.5 2.0 Reserves cover of ARA metric 1.5 1.0 0.5 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 24 Source: RAMP Survey on the Reserve Management Practices of Central Banks and IMF. Note: ARA = Assessing Reserve Adequacy. Appendix A | 41 FIGURE A.4 Emerging market respondents’ foreign exchange reserve allocation to traditional asset classes and reserves cover of ARA metric level 2.5 2.0 Reserves cover of ARA metric 1.5 1.0 0.5 0 100 90 80 70 60 50 40 30 20 10 0 Allocation to traditional asset classes (percent) High allocation Low allocation N = 19 Source: RAMP Survey on the Reserve Management Practices of Central Banks and IMF. Note: ARA = Assessing Reserve Adequacy. FIGURE A.5 Use of external managers by country-income group High-income High-income Upper Lower Low-income reserve (non-reserve) middle-income middle-income No 69% 6% 30% 11% 22% Yes 25% 88% 60% 89% 78% N = 94 Source: RAMP Survey on the Reserve Management Practices of Central Banks. 42 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks FIGURE A.6 Performance attribution models 60 53 50 40 Percent 30 19 19 20 10 8 0 Factor-based Sector-based Hybrid Other model model model N = 62 Source: RAMP Survey on the Reserve Management Practices of Central Banks. FIGURE A.7 ESG and investing in corporate bonds 60 53 50 40 33 Percent 30 27 20 19 20 16 13 10 7 8 3 0 No No, but No, but may Yes, Yes there have been consider in the but not discussions next 2 years explicitly N = 92 Corporate bonds No corporates Source: RAMP Survey on the Reserve Management Practices of Central Banks. Note: ESG = environmental, social, and governance. Bibliography Bailey, J., T. Richards, and D. Tierney. 2018. Evaluating Portfolio Performance. CFA Institute. Borio, C., G. Galati, and A. Heath. 2008. “FX Reserve Management: Trends and Challenges.” BIS Papers 38. https://www.bis.org/publ/bppdf/bispap38.pdf. Cardon, P., and C. Joachim. 2004. “Strategic Asset Allocation for Foreign Exchange Reserves.” In Risk Management for Central Bank Foreign Reserves, edited by E. C. Bank, pp. 13–28. Frankfurt: European Central Bank. https://www.ecb.europa.eu/pub/pdf/other/risk​ managecbreserves2004en.pdf. de Abreau Faria, F., and R. Ermes Streit. 2016. “Governance in Central Banks: A Comparative Study of the Governance Practices Used by the Central Banks of Brazil, Canada and England.” Revista de Administração Pública 765–793. http://www.scielo.br/pdf/rap/v50n5​ /en_0034-7612-rap-50-05-00765.pdf. Dooley, M., S. Lizondo, and D. Mathieson. 1989. The Currency Composition of Foreign Exchange Reserves. International Monetary Fund. Eichengreen, B., and M. Flandreau. 2014. A Century and a Half of Central Banks, Interantional Reserves and International Currencies. https://www.norges-bank.no/contentassets​ /3fba8b3a3432407d929ae9218db1ffc4/04_eichengreen_and_flandreau2014.pdf. Eichengreen, B., and D. Mathieson. 2000. The Currency Composition of International Reserves: Retrospect and Prospect. Washington, DC: International Monetary Fund. Fitch Ratings. 2015. 2014 Transition and Default Studies. Fitch Ratings. https://www.fitchratings​ .com/site/dam/criteria/Transition-Default-Data.xlsx. Ghosh, A. J. 2010. “Shifting Motives: Explaining the Buildup in Official Reserves in Emerging Markets since the 1980s.” IMF Working Paper 12/34. Ibbotson, R. G., and P. D. Kaplan. 2000. “Does Asset Allocation Policy Explain 40, 90 or 100 Percent of Performance?” Financial Analysts Journal 56(1), 26–33. International Monetary Fund. 2001. Guidelines for Foreign Exchange Reserve Management. International Monetary Fund. https://www.imf.org/external/np/mae/ferm/eng/. ———. 2011. Assessing Reserve Adequacy. Washington, DC: International Monetary Fund. https:// www.imf.org/external/np/pp/eng/2011/021411b.pdf. ———. 2013. Assessing Reserve Adequacy—Further Considerations. Washington, DC: International Monetary Fund. Retrieved from https://www.imf.org/external/np/pp/eng​ /2013/111313d.pdf. ———. 2014. Revised Guidelines for Foreign Exchange Reserve Management. Washington, DC: International Monetary Fund. ———. 2015. Assessing Reserve Adequacy—Specific Proposals. Washington, DC: International Monetary Fund. https://www.imf.org/external/np/pp/eng/2014/121914.pdf.  43 44 | Inaugural RAMP Survey on the Reserve Management Practices of Central Banks ———. 2016. Data Query . AREAER Online. https://www.elibrary-areaer.imf.org/Pages​ /ChapterQuery.aspx. ———. 2018. Annual Report on Exchange Arrangements and Exchange Restrictions. Washington, DC: International Monetary Fund. ———. 2018. Assessing Reserve Adequacy. International Monetary Fund. https://www.imf.org​ /external/datamapper/ARA/index.html. ———. 2019. Currency Composition of Official Foreign Exchange. International Monetary Fund. http://data.imf.org/?sk=e6a5f467-c14b-4aa8-9f6d-5a09ec4e62a4 Jones, B. 2018. Central Bank Reserve Management and Internatioanl Financial Stability—Some Post-Crisis Reflections. IMF Working Paper. https://www.imf.org/~/media/Files​ /Publications/WP/2018/wp1831.ashx. Lybek, T. and J. Morris. 2004. Central Bank Governance: A Survey of Boards and Management. IMF Working Paper. https://www.imf.org/external/pubs/ft/wp/2004/wp04226.pdf. McCauley, R. 2008. “Choosing the Currency Numeraire in Managing Foreign Exchange Reserves.” In RBS Reserve Management Trends, edited by R. Pringle & N. Carver. London: Central Banking Publications. McCauley, R., and T. Chan. 2014. “Currency Movements Drive Reserve Composition.” BIS Quarterly Review (Dec. 14): 23–36. Mwase, N. 2012. “How Much Should I Hold? Reserve Adequacy in Emerging Markets and Small Islands.” IMF Working Paper WP/12/205. International Monetary Fund, Washington, DC. Papaioannou, E., R. Portes, and G. Siourounis. 2006. Optimal Currency Shares in International Reserves: The Impact of the Euro and the Prospects for the Dollar. Cambridge: National Bureau of Economic Research. http://www.nber.org/papers/w12333.pdf. Stewart, S. 2013. Manager Selection. Research Foundation of CFA Institute. Sunner, D. 2017. “Trends in Global Foreign Currency Reserves.” Reserve Bank of Australia Bulletin (September Quarter 2017). SWIFT. 2018. RMB Tracker. https://www.swift.com/our-solutions/compliance-and-shared​ -­services/business-intelligence/renminbi/rmb-tracker. World Bank Data Team. 2017. “The World Bank Data Blog.” New Country Classifications by Income Level: 2017–2018. https://blogs.worldbank.org/opendata/new-country-classifications​ -income-level-2017-2018. ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, we leverage electronic publishing options and print- on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. We follow the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine–free (PCF), or enhanced elemental chlorine–free (EECF) processes. ­ More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility. In the Spring of 2018, the World Bank Treasury’s Reserves Advisory and Management Program (RAMP) concluded its inaugural survey on central banks’ reserve management practices. RAMP sought to assess whether there had been a significant evolution in this activity over the last two decades given: 1. The substantial increase in global foreign exchange reserves over that time period, and 2. The extraordinary policy responses to the unprecedented macroeconomic and investment environment during and after the global financial crisis. The survey results show that most central banks continue to employ a traditional approach: their reserve holdings are concentrated in high-quality fixed-income assets and the minimum credit rating for their investments remains conservative. At the same time, the data suggest important changes are under way, as a material number of central banks reported more diversified portfolios with exposure to nontraditional asset classes: a third of respondents hold corporate credit, most of which is investment grade, and almost one in five own mortgage-backed securities or equities, although mostly in limited allocations. Analysis of this information did not find a relationship between respondents’ measures of reserve adequacy and the size of their exposure to nontraditional asset classes. The data exhibit considerable cross-country differences in the way central banks manage their reserves and, in some circumstances, the analysis suggests that these differences correlate with respondents’ country income groups. SKU K880263