## wpiea2022212-print-pdf ## Source details **Canonical URL:** [wpiea2022212-print-pdf](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022212-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022212-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022212-print-pdf.pdf.json) --- ### Introduction and Rationale - Financial risks outside traditional banking can rapidly spread and disrupt the real economy, as evidenced by the Global Financial Crisis (GFC). - Pre-GFC growth in credit intermediation via shadow banking received limited attention in part due to lack of detailed statistics on size and interconnectedness. - Post-GFC initiatives increased data collection on nonbank financial intermediaries (NBFI), including: - G-20 Data Gaps Initiative (DGI-I) and DGI-2. - FSB’s annual global NBFI macro data collection, with focus on OFIs and a narrow NBFI measure for entities posing bank-like financial stability risks (maturity/liquidity transformation, leverage transformation, leverage or imperfect credit risk transfer). ### Scope and Objectives of the Paper - Reviews current NBFI taxonomy within Other Financial Corporations (OFC) and selected financial instruments across Financial and Macroeconomic Statistics Manuals (FMSMs). - Identifies areas for taxonomy updates, clarification, streamlining, and consistency. - Suggests expanded coverage to capture potential financial stability risks. - Compares proposed breakdowns with SNA and Balance of Payments Manual updates and proposes a common core taxonomy. - Focuses on conceptual and methodological guidance updates; excludes detailed cross-border discussions for a follow-up version. ### Data Gaps Initiative (DGI) — Goals and Progress - DGI origins and structure: - Launched after the GFC to identify data gaps (complex instruments, off-balance sheet entities, cross-border linkages) and recommend strengthened data collection for financial stability analysis. - DGI-1 produced 20 key recommendations; DGI-2 launched to complete DGI-1 and implement new recommendations (derivatives, CDIS). - Status at completion of DGI-2 (2022): - About half of recommendations fully completed; significant progress on the rest. - Recommendations not fully completed by end of DGI-2 (verbatim): - Recommendation 5) Shadow banking. Most G-20 countries made limited progress on collecting and disseminating securities financing transaction data (loans securitized with securities, including repos). - Recommendation 10) IIP. Gaps relate to other financial corporations (OFCs) and currency composition. - Recommendation 11) IBS. Several countries do not report consolidated banking statistics, which measure banks’ country risk exposures on a worldwide consolidated basis. - Recommendation 14) Cross border NBFI data. Only thirteen G-20 countries reported NBFI data, and this is done via IMF’s Standardized Report Forms for OFCs (SRF-4SR), ECB, and one other method, raising comparability questions. ### Guidance Notes and Recording Issues (FITT guidance notes) - Key recommended additions and clarifications: - Split MMFs into constant and variable NAV; separate insurance into life and non-life; separate pension funds into defined benefit and defined contribution. - Detailed subcategories added for non-MMF investment funds (12), OFIs (6), and captive financial institutions and money lenders (4) (reflected in Appendix IV). - BPM additions: MMFs, non-MMF investment funds, insurance corporations, pension funds, OFIs (including CCPs), captive financial institutions and money lenders, financial auxiliaries. - "Loans" in updated SNA/BPM to include an "of which" line for repos; in ESS nonfinancial corporations separated from households and NPISHs. - Crypto assets: - Examines classification of crypto assets/currencies; key issue whether crypto assets without a corresponding liability should be classified as non-financial assets or financial assets. - Fintech: - Existing standards allow recording fintech activities without introducing a new "fintech" sector; proposes an "of which fintech" category within existing classifications. - Derivatives and CDSs: - Recommend breaking down derivatives by market risk category (6), instrument (7), and trading venue/clearing status (3). - Continue to classify CDSs as options; recommend derivatives breakdowns so credit derivatives can be separately identified. - Reverse Transactions (RTs): - Maintain current recording in BPM6 and 2008 SNA; clarify recording of short RT positions, interest/dividends on borrowed securities, and borrower counterparties. ### State of Play: Gaps, Inconsistencies, and Need for a Common Taxonomy - Cross-manual shortfalls: - FSB templates require more detail (economic funds) than the nine financial corporation subsectors in 2008 SNA; reporters find SNA taxonomy insufficient. - Planned SNA/BPM enhancements (targeted in 2025 update) will address some issues (e.g., equity, bond, hedge fund detail) but gaps remain (e.g., government MMFs statistics requested by FSB). - Inconsistencies and ambiguities exist across manuals (example: classification of REITs and real estate funds is complex and ambiguous across 2008 SNA, FPFS handbook, BPM6). - Need and recommendation: - Update and expand guidance in SNA and BPM and ensure consistency across all official manuals (including GFS and MFSM). - Provide a common core taxonomy across official manuals, augmented with domain-specific material. - Analytical implications: - Lack of clarity reduces international comparability for FSB collections and can introduce measurement errors in macro indicators (saving, investment, equity in sectoral accounts), potentially misleading policy guidance. ### Process and Timelines - BPM6 and 2008 SNA update process launched in 2020 with target publication date of 2025. - Appendices I–V and figures/boxes provide taxonomy comparisons, recommended updates, proposed breakdowns, and crypto categorization discussion. ### Appendix I — Current Taxonomy Assessment (summary findings) - Cross-check across SNA, MFSM, HSS, FSI, GFS, BPM: examined 31 sector/instrument categories for a total of 186 elements, with coding outcomes: - 24 green - 28 blue - 44 yellow - 7 red - 83 missing - Seven additional NBFI subsectors proposed in Appendix III are currently not covered in any manuals and excluded from this review: constant NAV MMFs, variable NAV MMFs, government MMFs, prime MMFs, equity funds, fixed income funds, principal trading firms. - MFSMCG covered the most selected elements; GFS covered the least. - Color-code definitions: - Green: includes all core elements of modern taxonomy. - Blue: includes core elements plus additional detail. - Yellow: would benefit from clarification. - Red: inconsistent across manuals. - Gray: lacks definitional detail. - Priority inconsistencies: taxonomy of closed-end funds, ETFs, REITs; distinction between foreign currency swap and cross-currency interest-rate swap. - Need for additional granularity to capture stability-relevant features (institutional investors, asset managers, market intermediaries, FMIs). ### Two Illustrative NBFI Examples (Box 3) - NBFI liabilities and financial cycle: - Noncore funding expanded faster than core funding pre-2008; noncore-funded parts contracted most during the crisis. - Pre-2008 risk build-up elements: Repos; Federal funds; MMFs; Uninsured deposits. - Data collection proposals for runnable liabilities: - Collect additional breakdowns for size, composition, and maturity of runnable liabilities to enhance policymakers’ view of NBFI roles and emerging risks. - Non-bank Payment Service Providers (NBPSPs): - Distinguish NBPSPs by whether they (1) directly offer storage of value or (2) rely on third parties for storage. - Examples and classification notes: classic MSBs (e.g., Western Union) vs modern MSBs (e.g., Venmo, Alipay, M-Pesa); many U.S. NBPSPs regulated as MSBs. - Data capture depends on regulation (whether assets must be stored with depository corporations); coverage varies across jurisdictions and may be subnational. - Stability concerns: NBPSP liabilities are typically uninsured; maturity mismatches may arise; value-storage providers should be captured in financial statistics. - Recommended statistical actions: create basic taxonomy of NBPSPs by funding sources and regulatory regimes; distinguish payment processors that store value from passthrough processors; capture NBPSP lending and monetary liabilities where relevant. - Illustrative classification examples from decision tree: Alipay, M-pesa, Venmo classified as payment providers using own balance sheet and active intermediary; Zelle, Swish as passthrough/processor. - Challenger banks and neobanks: - Challenger banks: regulated digital-only banks (examples: Marcus, Discover Banks, Varo Bank). - Neobanks: non-regulated fintech firms offering bank substitutes (examples: Chime, Current, Revolut in some contexts); classification depends on funding, regulation, and partnerships. - Crypto Assets Without a Liability (CAWLs): - Consensus: classify digital assets as financial assets if a corresponding liability exists; no consensus on CAWLs. - 2008 SNA/BPM6 principle: financial assets must have a corresponding liability (exception: gold bullion held as official reserves). - Proposal: classify CAWLs and financial gold within a new "Valuables" category in the Financial Account (see Appendix V). - Stablecoins: - Types: Centralized vs Decentralized; Collateralized (Custodial, Synthetics) vs Un(der)collateralized (CBDC, Undercollateralized). - Observations: custodial stablecoins grew quickly; Figure notes top stablecoins by market capitalization (in billions); centralized stablecoin growth attracted attention. - Technical/operational risks: examples of failures — Bean attacker drained $182 million (April 17, 2022); Terra failure illustrated algorithmic stablecoin instability. - Classification depends on backing assets and redemption conditions; stablecoins that promise immediate redemption at par may resemble depository institutions; non-redeemable stablecoins may resemble ETFs or constant NAV MMFs. - CBDC: - BIS definition: digital money, denominated in national unit, representing a direct liability of the central bank (BIS 2021). - Variants: Wholesale CBDCs (restricted to financial intermediaries); Retail CBDCs (for day-to-day payments). - Design choices: account-based or token-based. - Distinction from narrow banks: liability of a private bank (even if backed by reserves) is not CBDC in many analyses; sometimes called "synthetic CBDCs." - Potential benefits: improved financial inclusion, competition, monetary policy tools, faster/secure settlement finality. - Policy question: whether retail payment reforms will reduce urgency for CBDC adoption. ### Central Bank Retail Payment Systems (Box 4) — CODI and Pix comparison - Key differences between Codi (Mexico, Sep 2019) and Pix (Brazil): - Pix mandates broad participation and mobile-app initiation/receipt functionalities; Codi does not. - Codi data stored at each financial institution; Pix data stored at central bank. - Codi slower to enable government transfers; Codi participation limited to SPEI members. - Usage and metrics (preserved exactly): - Codi usage as of April 2021: monthly transactions numbering in the hundreds of thousands and volume around 150 million Mexican pesos. - Pix adoption and volumes: - By end-February 2022 (15 months after launch), 114 million individuals, or 67 percent of the Brazilian adult population, had either made or received a Pix transaction. - 9.1 million companies have signed up – fully 60 percent of firms with a relationship in the national financial system. - Volume: initially around R$ 30 billion; over R$ 600 billion by February 2022 (about twice the amount of currency in circulation). - Transaction composition: over three quarters of transactions are between two individuals; business to business transactions account for about a third of the total volume. - Pix operational features: - Near instantaneous transfers (within a few seconds) at any time, low/zero cost for individuals, cellphone-based transfers, fraud detection and reversal, no minimum size, optional participant username privacy, transactional accounts linked to demand/savings/prepayment accounts, ability to obtain cash including at ATMs. - Cost comparisons (merchant/fee metrics preserved exactly): - Pix costs an average of 0.22 percent of a transaction’s value for merchants. - Debit cards in Brazil cost slightly above 1percent. - Credit cards in Brazil reach 2.2 percent. - Comparative credit card benchmarks: 1.7 percent in the United States, 1.5 percent in Canada and 0.3 percent in the European Union (BIS 2022). - Implications for CBDC and classification: - Retail payment reforms (PIX, CODI, FedNow) may reduce urgency for CBDC; CBDC classification would likely be straightforward as a central bank liability. - Synthetic CBDCs (two-tier with commercial bank involvement) may be classified like demand deposits ("inside" money) depending on design. ### Main Recommendations (IV. Main Recommendations, verbatim structure retained) - Update the international statistical standards’ classification systems. Provide cross-manual, common, and updated core definitions for each NBFI subsector, starting with the proposed taxonomy shown in Appendix III. - Expand the FMSM classification systems to include NBFI financial institutions and activities currently not covered. - Incorporate additional discussion of NBFI’s connections to financial risks in SNA and MFS manuals; tie taxonomy to funding risk, excessive leverage, and credit risk (Annex III). - The international statistical community needs an agile approach to update manuals to keep pace with financial innovation. - Continue to inventory the coverage of relevant NBFI instruments, particularly those associated with sectorization, valuation, cross-border risks, and those suggested through reader feedback. - Statistical frameworks and Statistical Business Registers must capture non-traditional providers that compete with traditional banks and report on them timely. - Decentralized, peer-to-peer crypto assets need monitoring to the extent possible. - Reporting guidance for crypto assets and stablecoins needs establishment; suggest creation of a new Valuables category within the Financial Account to include self-redeeming assets like Bitcoin and financial gold. - The activity of fintech payment providers should be reported and monitored; when they store value within their organization, their assets should be reported and monitored nationally. - Lending by fintechs should be reported and monitored like bank lending. - Funding structures for fintech lenders should be monitored for leverage and funding risks. ### Appendix II — Recommended Taxonomy Updates (selected substantive proposals) - MMFs (S123): subdivide into Constant Net Asset Value MMFs; Variable Net Asset Value MMFs; Of which: Government; Of which: Prime. Risk flags specified per subtype (maturity mismatch, collateral, counterparty, concentration, foreign exchange risk). - Non-MMFs (S124): subdivide open-end (S124a) and closed-end (S124b) funds with subcategories: Real estate funds (Of which: Mortgage REITs; Equity REITs), Equity funds, Bond funds, Mixed/balanced funds, Hedge funds, Fund of funds, Private equity funds (Of which: Venture capital), Exchange-traded funds. - ETFs risk flags: model/price/ineffective hedges; underlying market liquidity; stresses on authorized participants. - REITs: distinguish Equity REITs and Mortgage REITs; note equity REITs may be classified in nonfinancial corporate sector in some manuals. - Insurance (S128) and Pension Funds (S129): include defined-benefit and defined-contribution distinctions and associated risk flags. - Market Intermediaries/OFIs (S125): identify securities dealers, principal trading firms, CCPs, and FMIs with associated risk profiles. - Financial Instruments (F.4, F.7): recommend instrument-level breakdowns and trading/clearing status: Exchange traded; Over-the-counter (cleared); Over-the-counter (not cleared). ### Appendix V — Categorization of Crypto Assets (key proposals) - Distinguish inside vs outside money; define redeemability classification for financial assets: self-redeeming vs redeemable into other financial instruments. - Application: - Some crypto assets (e.g., bitcoin) catalogued as self-redeeming (no corresponding liability). - Stablecoin liabilities that are redeemable into other financial instruments classified as redeemable. - Statistical recommendation: create a new Valuables category within the Financial Account to classify self-redeeming crypto assets (e.g., Bitcoin) and financial gold. - Decision tree logic preserved: classify assets first as medium of exchange or not; then as self-redeemable or redeemable; apply to crypto and monetary assets accordingly. ### Concluding Remarks and Next Steps - Need for granular data: financial stability focuses on tail risks; granular data required to detect buildup of vulnerabilities. - Purpose: update guidance for NBFI and financial innovation data collection; current guidance needs updating to reflect rapid restructuring and innovation. - Scope and future work: - National authorities should prioritize granularity relevant to their risk profiles. - Sectorization, valuation, cross-border, and other risk topics deferred to future versions. - Further deliberation needed on instrument detail (investor leverage, from-whom-to-whom detail). - Develop Table examples to encourage collection of agreed data. - Future work may consider securitization vehicles and climate/biodiversity financial stability risk elaboration. *Source: International Monetary Fund — wpiea2022212-print-pdf (References and Appendices content).* ### References _________________________________________________________________________ 58 ### wpiea2022212-print-pdf - References _________________________________________________________________________ 58 ### I. Introduction and Rationale - Financial risks outside traditional banking can rapidly spread and disrupt the real economy, as evidenced by the Global Financial Crisis (GFC). - Prior to the GFC, rapid growth in credit intermediation via shadow banking (e.g., securitization) received relatively little attention due in part to a lack of adequately detailed statistics on size and interconnectedness of emerging risk exposures. - Post-GFC initiatives by authorities, international organizations (including the IMF), and the G-20 sought to strengthen data collection on nonbank financial intermediaries (NBFI), including the G-20 Data Gaps Initiative (DGI-I) and the FSB’s annual global NBFI macro data collection. - The FSB tracks broader NBFI aggregates and also: - Other Financial Intermediaries (OFIs) - A narrow measure of NBFIs defined as “entities that may pose bank-like financial stability risks (i.e., credit intermediation that involves maturity/liquidity transformation, leverage transformation, leverage or imperfect credit risk transfer).” ### II. Scope of the Paper and Proposed Objectives - The paper: - Reviews the current NBFI taxonomy within Other Financial Corporations (OFC) and selected financial instruments across Financial and Macroeconomic Statistics Manuals (FMSMs). - Identifies areas for taxonomy updates, clarification, streamlining, and consistency. - Suggests expanded coverage of the NBFI sector to capture potential financial stability risks. - Compares proposed breakdowns with those planned in the SNA and Balance of Payments Manual updates and proposes a common core taxonomy. - The paper focuses on conceptual and methodological guidance updates and excludes detailed cross-border discussions to be addressed in a follow-up version. ### III. The Data Gaps Initiative (DGI): Goals and Achievements (Box 1) - DGI origins: - Launched in the wake of the GFC by the G20WG to identify data gaps (complex instruments, off-balance sheet entities, cross-border linkages) that hindered policy responses, and to recommend ways to strengthen data collection for financial stability analysis. - DGI structure: - DGI-1 produced 20 key recommendations covering monitoring risk in the financial sector, tracking international network connections, and expanding sectoral and other financial and economic datasets. - DGI-2 launched to complete DGI-1 recommendations and to implement new recommendations on derivatives and the CDIS. - Status by completion of DGI-2: - At completion of DGI-2 in 2022, about half of the recommendations had been fully completed and significant progress had been made on the rest. - Recommendations not fully completed by end of DGI-2 (listed exactly as in source): - Recommendation 5) Shadow banking. Most G-20 countries made limited progress on collecting and disseminating securities financing transaction data (loans securitized with securities, including repos). - Recommendation 10) IIP. Gaps relate to other financial corporations (OFCs) and currency composition. - Recommendation 11) IBS. Several countries do not report consolidated banking statistics, which measure banks’ country risk exposures on a worldwide consolidated basis. - Recommendation 14) Cross border NBFI data. Only thirteen G-20 countries reported NBFI data, and this is done via IMF’s Standardized Report Forms for OFCs (SRF-4SR), ECB, and one other method, raising comparability questions. ### IV. Guidance Notes Related to NBFI and Fintech Recording (Box 2) - Overview of FITT guidance notes and key recommendations: - “Capturing Nonbank Financial Intermediation” and “More Disaggregated Institutional Sector and Financial Instrument Breakdowns”: - Recommend additional sectoral and instrument breakdowns in SNA and BPM to better capture NBFI. - Proposed additions: split MMFs into constant and variable NAV; separate insurance into life and non-life corporations; separate pension funds into defined benefit and defined contribution plans. - Detailed subcategories added for non-MMF investment funds (12), OFIs (6), and captive financial institutions and money lenders (4). (These are reflected in Appendix IV.) - BPM additions include MMFs, non-MMF investment funds, insurance corporations, pension funds, OFIs (including central clearing counterparties), and captive financial institutions and money lenders, and financial auxiliaries. - “Loans” in the updated SNA and BPM would include an “of which” line for repos; in external sector statistics (ESS) nonfinancial corporations would be separated from households and NPISHs. - “The Recording of Crypto Assets in Macroeconomic Statistics”: - Examines classification of crypto assets/currencies. - Key issues: whether crypto assets without a corresponding liability should be classified as non-financial assets or as financial assets. - “Impact of Fintech on Macroeconomic Statistics”: - Concludes that existing international statistical standards allow proper treatment and recording of fintech companies and fintech-related activities in most cases without introducing a new “fintech” sector. - Proposes an “of which fintech” category can be included within existing activity and sector classifications for countries with significant fintech activities. - “Financial Derivatives”: - Recommends breaking down derivatives by market risk category (6), instrument (7), and trading venue and clearing status (3). (Market risk and instrument breakdowns are included in Appendix IV.) - Clarifications and reemphasis of existing guidance. - “Credit Default Swaps” (CDSs): - Recommends continuing to classify CDSs as options instead of forward contracts. - Recommends breakdowns of financial derivatives by risk category so credit derivatives can be separately identified. - “Reverse Transactions” (RTs): - Recommends maintaining current recording of RTs in BPM6 and 2008 SNA. - Recommends clarifications (recording of short RT positions, interest and dividends on borrowed securities, identify borrower counterparties, etc.) and discussing separate presentations of RTs in the updated SNA and BPM. ### V. State of Play: Gaps, Inconsistencies, and Need for a Common Taxonomy - Role of manuals: - Financial and macroeconomic statistical manuals provide conceptual and methodological guidance for producing NBFI statistics used by FSB and others. - Key gap issues: - The detail needed by FSB templates (moving from OFIs to economic funds) exceeds the nine financial corporation subsectors in the 2008 SNA; reporters find SNA taxonomy insufficient. - Planned SNA enhancements (targeted in the 2025 update of BPM6 and 2008 SNA) will address some issues (e.g., detail on equity, bond, and hedge funds) but gaps will remain (e.g., the FSB request for government MMFs statistics). - Inconsistencies and ambiguities in guidance exist across manuals; example: classification of REITs and real estate funds is complex and ambiguous. - 2008 SNA: real estate activities by trusts or companies are classified under non-financial corporate sector; real estate funds are classified under non-MMF investment fund sector. - Financial Production, Flows and Stocks SNA handbook distinguishes real estate funds, equity REITs, and mortgage REITs but also describes equity and mortgage REITs as types of real estate funds, creating ambiguity. - BPM6 mentions REITs should be included in the investment fund subsector but does not distinguish equity vs mortgage REITs or other real estate funds. - Need and recommendation: - Update and expand guidance in SNA and BPM and ensure consistency across all official manuals (including GFS and MFSM). - Aim to provide a common core taxonomy across all official manuals, augmented with domain-specific material as relevant. - Analytical implications: - Lack of clarity can reduce international comparability for FSB collections and introduce measurement errors in macro indicators (saving, investment, equity in sectoral accounts), potentially misleading policy guidance. ### VI. Process and Timelines Referenced - The latest update process for BPM6 and 2008 SNA was launched in 2020 with a target publication date of 2025. - Appendices and figures included in the paper (as listed in the source) provide detailed taxonomy comparisons, recommended updates, and proposed breakdowns: - Appendices: I–V (I: Current NBFI Taxonomy; II: Recommended Taxonomy Updates; III: Recommended NBFI Subsectors, Potential Sources of Financial Stability Risk and Taxonomy; IV: Proposed NBFI Breakdown; V: On the Categorization of Crypto-Assets in Official Statistics). - Figures and Boxes illustrating taxonomy gaps, ETFs, funding, runnable liabilities, NBPSP taxonomy, financial asset decision tree, stablecoins, etc., are included in the paper. *International Monetary Fund — wpiea2022212-print-pdf (References section content provided above).* ### Appendix I lists the current taxonomy of all the main NBFI categories—institutional investors and asset managers; market ### wpiea2022212-print-pdf - Appendix I lists the current taxonomy of all the main NBFI categories—institutional investors and asset managers; market ### Current taxonomy assessment and cross-manual review - Appendix I cross-checks the taxonomy of NBFI categories across the manuals: SNA, MFSM, HSS, FSI, GFS, BPM. - For this review, the authors examined 31 different sector or instrument categories across the 6 manuals for a total of 186 elements (of which 24 green, 28 blue, 44 yellow, 7 red, and 83 missing). - Seven additional NBFI subsectors proposed in Appendix III are currently not covered in any of the manuals and have therefore been excluded from this review: constant net asset value (NAV) MMFs, variable NAV MMFs, government MMFs, prime MMFs, equity funds, fixed income funds, and principal trading firms. - The MFSMCG covered the most of the selected elements while the GFS covered the least. ### Color-coded findings on definitional alignment - Green: Includes all core elements of modern taxonomy. - Blue: Includes all core elements of modern taxonomy plus additional detail that may be removed if not domain-relevant (example: HSS description of open-end funds lists the most popular types). - Yellow: Current taxonomy would benefit from clarification (example: MFSMCG says that the benefits received by participants in defined contribution (DC) pension plans are “based on the participant’s contributions to the pension fund and the investment performance of the fund” (3.200); this would benefit from clarifying that employers may also contribute to a participant’s DC plan). - Red: Current taxonomy is inconsistent across manuals (example: BPM description of closed end funds indicates that they are also known as “exchange-traded funds” (ETFs), while most ETFs are open-end funds). - Gray: Lacks definitional detail; maintaining common definitions across manuals may aid compilers’ understanding and data collection. ### Identified inconsistencies and priority areas - Instances of inconsistency are rare but indicate priority areas: taxonomy of closed-end funds, exchange-traded funds, REITs; distinction between a foreign currency swap and a cross-currency interest-rate swap. - The review highlights that some manuals omit definitional detail for items that other manuals treat in depth (examples: open-end and closed-end funds in FSI and GFS). ### The need for additional granularity (rationale and examples) - Rationale: - To provide a clearer window into financial stability risks, additional granularity for institutional investors and asset managers, market intermediaries, and financial market infrastructures is proposed. - Since the GFC, several NBFI episodes have been central to market distress (example: March 2020 dash-for-cash with large redemptions in prime MMFs). - NBFIs can be important sources of financial stability risk via channels such as leverage, funding reversals, procyclicality of credit, maturity/liquidity/funding mismatches, excessive leverage, and counterparty risks. - Example — Money Market Funds (MMFs): - MMFs have become key intermediaries for short-term sovereign securities and corporate debt; liquidity transformation exposes them to runs. - Current and proposed SNA and BPM MMF breakdowns would not allow tracking detailed MMF developments. - The paper suggests compilers consider capturing data on government and prime MMFs separately for both stable and variable NAV, beyond the currently recommended split of constant NAV MMFs and variable NAV MMFs. - Measures to mitigate runs since the GFC (e.g., allowing stable NAVs only for MMFs investing in government securities) have been inadequate; initiatives under consideration include additional redemption gates, liquidity fees, and swing prices. - Example — Exchange Traded Funds (ETFs): - ETFs have grown significantly since the late 90s and hold close to $ 10 trillion in assets. - ETFs issue shares designed to track passive investments and are redeemable only by Authorized Participants (APs) who arbitrage price deviations versus NAV. - Liquidity of underlying asset markets and liquidity constraints on APs have led to share prices deviating from underlying asset fair value during stress (example: corporate bond ETFs trading at steep discounts to underlying asset values in March 2020). - ETFGI reports the global ETFs industry gathered US $105.88 billion of net inflows in February 2022. ### Recommendations for taxonomy updates and additional granularity - Review and update NBFI taxonomy across the suite of manuals to enhance clarity and consistency to reduce confusion and improve availability and quality of statistics. - Provide additional granularity beyond the planned SNA and BPM update in conceptual guidance for NBFI data collection for macro-financial surveillance. - Guidance should be tied to the type of systemic risk posed by the sector/financial market; selection criteria for taxonomy and granularity updates should prioritize systemic-risk relevance (examples: MMFs for liquidity mismatch risk; ETFs for potential liquidity and valuation risk transmission). - Suggested additional breakdowns to aid surveillance (summarized and elaborated in appendices): - Additional breakdowns in MMFs. - Additional breakdowns in non-MMFs. - Additional breakdowns in REITs. - Additional breakdowns in hedge funds. - Additional breakdowns in derivatives. - Acknowledge that not all suggested breakdowns will be relevant to all jurisdictions, but each proposed breakdown is tied to specific risks. ### Appendices and complementary materials referenced - Appendix III: details potential financial stability risk triggers associated with suggested NBFI subsector breakdowns and presents suggested taxonomy for each proposed subsector. - Appendix IV: summarizes the list of additional NBFI breakdowns proposed in this paper, relative to the new proposed financial sector breakdowns in the context of the SNA and BPM updates. - Note: Additional NBFI subsectors and instruments that could be considered later include warehoused asset loans, insured versus uninsured deposits by investor, securitization funds, and interval investment funds. *Source: IMF Working Paper (authors’ compilation from SNA, MFSM, HSS, FSI, GFS, BPM materials).* ### Box 3. NBFI Data for Financial Stability: Two Examples ### Box 3. NBFI Data for Financial Stability: Two Examples ### NBFI liabilities and the financial cycle - Existing monetary and financial sector data can track the composition of NBFI liabilities to provide insights into the stage of the financial cycle and the buildup of risk. - Prior to 2008, noncore funding of the financial sector increased at a much faster pace than core funding. - The noncore funded part of the financial sector suffered by far the greatest contraction when the crisis occurred. ### Granular market elements and the 2008 crisis - Specific elements that increased sharply prior to 2008 and represented a buildup of risk: - Repos - Federal funds - Money Market Funds (MMFs) - Uninsured deposits (to a lesser extent) - These riskier markets experienced greater contractions after the crisis in 2008. - Mechanisms of impact: - Repos: affected in part because of higher margin requirements; financial intermediaries (brokers and dealers) that relied on continuous rollover of repo for funding (including Lehman Brothers) were severely impacted. - MMFs: runs occurred because small movements in safe assets called into question their ability to honor strict constant NAV obligations. ### Data collection and statistical enhancements for runnable liabilities - The paper proposes additional breakdowns to improve collection of information on: - Size of runnable liabilities - Composition of runnable liabilities - Maturity of funding of runnable liabilities - Expected benefits: - Enhanced understanding of the increasingly important role NBFIs play in the financial sector - Better view for policymakers of emerging risks and developments to be addressed ### Non-bank Payment Service Providers (NBPSPs): taxonomy and classification - NBPSPs distinction for NBFI statistics: whether the NBPSP 1) directly offers storage of value; or 2) relies on third parties for this storage. - Examples and regulatory notes: - In the U.S., many NBPSPs are regulated as Money Services Businesses (MSBs)—money transmitters. - Classic-type MSBs (e.g., Western Union): issue only specific claims (redeemable only by a specific person and at a specific place). - Modern-type MSBs (e.g., mobile network operators such as Venmo, Alipay, M-Pesa): provide a general-purpose liability similar to a demand deposit. - Firms named: WeChat pay, Alipay, Venmo, Revolut, M-Pesa. - Data capture issues: - Whether NBPSP assets are captured in monetary statistics depends on: - Whether regulation requires NBPSPs to store assets with depository corporations (central banks and commercial banks). - Compiler care in capturing electronic money balances as part of monetary aggregates. - Most countries require NBPSPs that offer storage of value to deposit holdings at a regulated depository institution, so they are captured in broad money aggregates—though transaction capture varies by jurisdiction and regulation. - Heterogeneous regulation can lead to limited coverage or coverage only at subnational levels (e.g., U.S. statistics on service volumes may be available only at the state level). - Financial stability concerns: - Nonbank regulation does not require NBPSPs to maintain a minimum stock of liquid assets to redeem liabilities on demand—maturity mismatches may arise. - Liabilities of MSBs and other NBPSPs are not insured, posing run risk. - Concerns are attenuated for NBPSPs that provide only front-end communication services and do not store value. - Recommended statistical action: - Statistical manuals should create a basic taxonomy of NBPSPs based on funding sources and regulatory regimes. - For payment processors, distinguish whether the NBPSP stores value on its balance sheet or provides access to another institution’s monetary liabilities. - For NBPSP lenders, determine whether lending is from own balance sheet, via issuance of asset-backed securities, or as passthrough (“processors”). - NBPSPs that lend directly or issue monetary liabilities should be captured in financial statistics. - Illustrative classification outcomes (from the hypothetical decision tree): - NBPSP payment provider using own balance sheet and active intermediary: Alipay, M-pesa, Venmo. - NBPSP payment provider as passthrough/processor: Zelle, Swish. - NBPSP lender using own balance sheet and active intermediary: Revolut (in some countries), SoFi. - NBPSP lender as passthrough/processor: Revolut (in some countries). ### Technological challenges to incumbent banks: challenger banks and neobanks - Definitions: - Challenger banks: regulated banks providing digital-only, branchless services (examples: Marcus, Discover Banks, Varo Bank). - Neobanks: non-regulated fintech firms offering close substitutes to bank services; often partner with banks (examples: Chime, Current rely on partner banks; Revolut in the U.K.). - Statistical and classification issues: - Challenger banks report statistics like other banks—classification straightforward. - Neobanks may be classified as non-financial corporations in the embryonic stage but may require reclassification as activities mature. - Mature neobanks may be difficult to classify depending on funding sources and regulation (e.g., those funded through securitization schemes without banking licenses). - Neobanks that partner with incumbent banks may have activity reported in partner banks’ statistics. ### Classification criteria for digital money and CAWLs (Crypto Assets Without a Liability) - Consensus and open questions: - Statistics community consensus: digital assets should be classified as financial assets provided there exists a corresponding liability. - No consensus on treatment of CAWLs even if suitable as or designed to act as a medium of exchange. - 2008 SNA and BPM6 principle: - Financial assets, by definition, must have a corresponding liability—gold bullion held as official reserves is the only exception. - Inside vs outside money and self-redeeming assets: - Inside money: has corresponding liability and disappears upon full consolidation across balance sheets (examples: checking accounts, e-money such as Alipay or M-pesa). - Outside money/self-redeeming: does not disappear upon consolidation (examples: currency, central bank reserves, monetary and non-monetary gold used for financial purposes); proposed to include CAWLs (for example, Bitcoin). - The paper proposes classifying CAWLs and financial gold within a new “Valuables” category in the Financial Account (see Appendix V reference). - Additional classification dimensions that may emerge: - Whether a jurisdiction deems a cryptocurrency legal tender. - Whether a cryptocurrency is programmable. - Whether the cryptocurrency is “native” (not tied to a product or service). - Whether the cryptocurrency builds on a native currency (DApps, NFTs). - Whether the asset arises from ICOs or similar funding products. ### Stablecoins: types, risks, and classification criteria - Types (bird’s eye view, Figure 7): - Centralized vs Decentralized - Collateralized (Custodial, Synthetics) vs Un(der)collateralized (CBDC, Undercollateralized) - Observations: - Custodial stablecoins have grown quickly (with a recent slowdown in growth rates); Figure 8 shows top stablecoins by market capitalization (in billions). - Although they represent a small portion of overall capital markets in aggregate, growth has attracted attention from intermediaries and policymakers. - Technical and operational risks: - Many centralized stablecoins operate as software on decentralized blockchains—subject to technical and operational risk (Narula (2021) referenced). - Examples of failures: - Bean (Beanstalk) attacker drained $182 million (April 17, 2022). - Terra failure illustrated algorithmic stablecoin instability. - Classification determinants and examples: - Key factors: type of backing assets and conditions for withdrawal/redemption. - Stablecoins that promise redemption at par upon request and offer immediate redemption rights may resemble depository institutions. - Stablecoins not redeemable may resemble ETFs or constant NAV MMFs if they have procedures to adjust token supply. - Because backing assets and redemption conditions are often unclear, stablecoins may not resemble traditional financial institutions. - Hypothetical classification decision tree (Figure 9) examples: - If access to central bank account = Yes, portfolio = 100 percent in bank reserves, immediate redeemability = Yes => Classify as DTn* (narrow bank). - If No access to CB account, portfolio held at deposit takers (DTs), immediate redeemability = Yes => Classify as DTs. - If reserves are cash and cash-equivalents, immediate redeemability = No => Classify as MMF. - If securities to track specific index and not redeemable => Classify as ETF. - Note: DTn stands for narrow bank. ### Central Bank Digital Currency (CBDC) - BIS definition: a form of digital money, denominated in the national unit of account, representing a direct liability of the central bank (BIS 2021). - Two widely recognized variants: - Wholesale CBDCs: restricted to financial intermediaries for interbank transfers and wholesale transactions. - Retail CBDCs: for day-to-day payments and transfers by the wider economy. - Design choices: - CBDCs can be account-based or token-based. - Key distinctions and related concepts: - CBDCs are issued by a monetary authority, unlike stablecoins issued by the private sector. - Narrow banks (accounts backed fully by central bank reserves) are not considered CBDCs by all analysts because the liability is of a private bank—not the central bank; such narrow bank deposits are sometimes referred to as “synthetic CBDCs.” - Potential benefits discussed in literature: - Improved financial inclusion - Improved competition in the banking system - Better tools for implementation of monetary policy - Faster and more secure finality of payment settlement - Policy question: - Whether ongoing reforms to retail payment systems will be sufficient to achieve faster and more secure finality of payment settlement, reducing the urgency for CBDC adoption. *Source: IMF Working Paper — Financial Innovation and Statistical Methodological Guidance (Box 3 content).* ### Box 4. Central Bank Retail Payment Systems ### Box 4. Central Bank Retail Payment Systems ### Comparison of recent retail payment reforms: CODI and Pix - Mexico introduced Codi in September 2019; it shares many similarities with Brazil’s Pix but differs in several key aspects. - Critical differences highlighted: - Pix mandates that large financial institutions participate and requires that all participating PSPs provide customers with the functionalities needed to initiate and receive instant payments in their mobile applications; Codi does not take these universal-access steps. - Codi data are stored at each financial institution; Pix data are stored at the central bank. - Codi was slower to introduce the capacity to make transfers to and from the government. - Participation in Codi is more restrictive: allowing participation only by financial institutions that are members of SPEI, the Bank of Mexico's real-time gross settlement (RTGS) system. - Comparative usage and reach: - Codi usage in Mexico: monthly transactions numbering in the hundreds of thousands and volume of transactions amounting to around 150 million Mexican pesos as of April 2021. - Pix usage in Brazil (rapid adoption): - By end-February 2022 (15 months after launch), 114 million individuals, or 67 percent of the Brazilian adult population, had either made or received a Pix transaction. - 9.1 million companies have signed up – fully 60 percent of firms with a relationship in the national financial system. - Volume of transactions: initially around R$ 30 billion; over R$ 600 billion by February 2022 (about twice the amount of currency in circulation). - Transaction composition: over three quarters of transactions are between two individuals; business to business transactions account for about a third of the total volume of transactions. - Transactions with the government are possible, but the number and volume are negligible. - Demographics: Pix is used by people of all ages, with only a somewhat larger though gradually declining share of younger users. - Anecdotal usage extends to very small businesses and even panhandlers (Bloomberg, 2021). ### Pix features, costs, and operational details - Operational features: - Near instantaneous transfers (within a few seconds) at any time of day—including when offline. - Zero cost for individuals (other than in a small set of circumstances) and low cost for businesses. - Transactions can be made using only cellphones; alternatives include bank branches and ATMs. - Security mechanisms detect fraud and can quickly reverse transactions when detected; transactions occur over a protected network. - No minimum transaction size limit; participants may establish upper limits per payee, day, or month. - Participants may join Pix with usernames that don’t reveal personal information; usernames are linked to accounts that are not anonymous (not revealed to every counterparty). - Account and cash access: - Pix transactions are made to and from “transactional accounts”, which include demand deposit, savings, or prepayment accounts operated by a bank or other financial institution. - Pix can also be used to obtain cash, including at ATMs. - Pix participants do not extend credit, and there are no balances associated with a Pix account; only with the accounts that Pix is linked to. - Cost comparisons (merchant/fee metrics preserved exactly as presented): - Pix costs an average of 0.22 percent of a transaction’s value for merchants. - Debit cards in Brazil cost slightly above 1percent. - Credit cards in Brazil reach 2.2 percent. - Comparative credit card fee benchmarks: 1.7 percent in the United States, 1.5 percent in Canada and 0.3 percent in the European Union (BIS 2022). ### Implications for CBDC and classification in official statistics - Context: - Ongoing reforms to current retail payments systems (PIX, CODI, and the upcoming FedNow) aim to address speed and finality of payments; these reforms are ongoing and limited to specific countries. - A number of countries are actively studying the adoption of retail CBDCs; guidance for the classification of CBDC is needed. - Classification options and implications: - A CBDC would presumably be classified as a liability of the central bank or monetary authority in the same manner as traditional currency (the likelihood that it is nonredeemable notwithstanding) making it straightforward to classify in official statistics. - Synthetic CBDCs (two-tier systems where commercial banks are involved in issuance): - Banks would act as intermediaries issuing CBDC that would represent a claim on the central bank. - It could also be that the synthetic CBDC could represent a claim on the issuing commercial banks, backed by deposits at the central bank. - In the latter case, CBDC could be classified like a demand deposit, as “inside” money. - Either classification would be a relatively straightforward application of current statistical methodology. ### Main recommendations (IV. Main Recommendations) - Update the international statistical standards’ classification systems. Provide cross-manual, common, and updated core definitions for each NBFI subsector, starting with the proposed taxonomy shown in Appendix III. The review of NBFI subsectors and selected financial instruments, assembled in Appendix I and illustrated in Appendix II, reveal both overlaps and inconsistencies in definitions across the STA manuals. A set of streamlined, common core definitions will aid statisticians in the assembly of necessary counterparty data and help ensure cross-domain compatibility. These core definitions should be supplemented with content appropriate to the domain. - Expand the FMSM classification systems to include NBFI financial institutions and activities currently not covered. As shown by the red text Appendix IV, there are several NBFI subsectors relevant to financial surveillance that are not discussed in the current suite of FMSMs nor in the proposed updated standards. Providing methodological guidance on these —their definitions, primary activities, and links to financial stability risks—will aid compilers in gaining support and traction for the collection of such data in jurisdictions where they are relevant, and support international comparability when data are published. - Incorporate additional discussion of NBFI’s connections to financial risks in SNA and MFS manuals. For example, NBFIs that allow immediate and at constant NAV withdrawals, such as by the authorized participants of exchange-traded funds, face higher funding risk—even when their liabilities are backed by short term assets, such as government fixed income securities of short maturity. Annex III outlines how this paper’s recommended NBFI breakdowns are associated with funding risk, excessive leverage, and credit risk. While there is a robust discussion of financial risks, particularly relating to depository corporations, in the Financial Soundness Indicators 2019 manual, compilers in other domains would also benefit from a better understanding of data needs for monitoring NBFI subsectors. - The international statistical community needs to take a more agile approach in the update of statistical manuals to ensure methodological guidance keeps pace with financial innovation. They will need to be able to quickly establish and adopt new international methodological guidance to keep pace with developments in NBFIs and Fintech. - Continue to inventory the coverage of relevant NBFI instruments, particularly those associated with sectorization, valuation, cross-border risks (not fully assessed in this paper) and those suggested through reader feedback. Identified instruments should be considered and documented in order to provide a more comprehensive view of potential risk exposures. - Statistical frameworks and Statistical Business Registers must seek to capture non-traditional providers that compete with traditional banks and report on them in a timely manner. The increasing digitalization of technology has not only broadened access to financial services but has allowed some of these non-traditional providers to operate without banking supervision and regulation, potentially leading to systemic risks to the financial system. - Decentralized, peer-to-peer crypto assets need to be monitored to the extent possible. They are rapidly evolving, are becoming entrenched in the financial system, and a lack of information will complicate the work of policymakers, investors, and other stakeholders. - Reporting guidance for crypto assets and stablecoins needs to be established. Self-redeeming private crypto assets are intended to act as financial assets and should be classified as such while recognizing their novel characteristics. The paper suggests the creation of a new Valuables category within the Financial Account to include self-redeeming assets, like Bitcoin and financial gold. The wide variety of rapidly growing stablecoins pose conceptual challenges to statisticians that should be addressed by the international statistical community under the aegis of Advisory Expert Group on National Accounts (AEG) and the IMF’s Balance of Payments Committee (BOPCOM). - The activity of fintech payment providers should be reported and monitored for purposes of both micro- and macro- financial stability. In cases where they store value within their organization, their assets should be reported and monitored on a national basis, consistent with the reporting required of banks. - Similarly, lending by fintechs should be reported and monitored in the same way as lending by banks. Fintech lending is an important category of loan origination in many markets, and, in some jurisdictions, the dominant sector for some loan types, such as for home mortgages in the U.S. - Funding structures for fintech lenders should be monitored for both leverage and funding risks. The risk of sudden decreases in funding availability is a key micro-prudential risk and may be associated with broader systemic risks. ### Concluding remarks and next steps - Need for granular data: “Financial stability ... is more about the tail of the probability distribution than the central probability” (Cunliffe, 2017). Granular data are required to recognize the buildup of potential vulnerabilities through time. - Purpose of the paper: update guidance for the collection of NBFI and financial innovation data; current official statistical guidance needs updating to reflect rapid restructuring and innovation in the financial sector. - Scope and future work: - Not all suggested NBFI granularity will apply to every economy; national authorities must prioritize collection efforts around areas of greatest risk. - Sectorization, valuation, cross-border, and other relevant risk topics have been deferred for future versions. - Further deliberation is needed on instrument detail required to monitor financial risks (e.g., degree of investors’ leverage, from-whom-to-whom detail). - Developing Table examples, such as those in the FSI Guide, would encourage collection of agreed data. - Future work may consider financial vehicle corporations engaged in securitization transactions for their role in dispersing credit risk. - Additional future work: elaborate on financial stability risks posed by climate change and biodiversity loss, and establish consensus around Annex III taxonomy and appropriate granularity. *Source: Box 4. Central Bank Retail Payment Systems — IMF Working Paper content unit wpiea2022212-print-pdf* ### Appendix I. Current NBFI Taxonomy in Official Statistics ### Appendix I. Current NBFI Taxonomy in Official Statistics ### Institutional Investors and Asset Managers - Money market funds (S123) - SNA: "Money market funds (MMFs) are collective investment schemes that raise funds by issuing shares or units to the public. The proceeds are invested primarily in money market instruments, MMF shares or units, transferable debt instruments with a residual maturity of not more than one year, bank deposits and instruments that pursue a rate of return that approaches the interest rates of money market instruments. MMF shares can be transferred by cheque or other means of direct third-party payment. Because of the nature of the instruments the schemes invest in, their shares or units may be regarded as close substitutes for deposits." (4.107) - MFS: Same core definition; adds classification criteria: "(1) a certain degree of capital certainty (store of nominal value); and (2) the possibility to withdraw funds immediately or on short notice." Also notes third-party payment facilities and limitations. (3.144-145) - HSS: Similar description; references 2008 SNA paragraph 4.107. (4.20) - FSI: Repeats MFS criteria: "(1) a certain degree of capital certainty (reliable store of value); and (2) the possibility to withdraw funds immediately or on short notice." (2.45) - GFS: Emphasizes investment in short-term money market securities and note on inclusion in broad money and separate subsector recording. (7.175) - BPM: Mirrors SNA; recommends showing MMFs as an extra subsector when included in monetary aggregates. (4.73) - Non-money market investment funds (S124) - SNA: "Non-MMF investment funds are collective investment schemes that raise funds by issuing shares or units to the public. The proceeds are invested predominantly in financial assets, other than short-term assets, and in nonfinancial assets (usually real estate). Investment fund shares or units are generally not close substitutes for deposits. They are not transferable by means of cheque or direct third-party payments." (4.108) - MFS: Adds that non-MMF funds invest predominantly in long-term financial assets and nonfinancial assets; "Shares or units issued by non-MMF investment funds are not close substitutes for deposits" and "do not meet the definition of broad money and are not included therein." (3.149) - HSS: Mirrors SNA. (4.21) - FSI: Defines non-MMF funds as collective investment schemes excluding MMFs; lists denominations (mutual funds, investment pools, investment trusts, unit trusts, institutions for collective investment). (2.46) - GFS: Defines investment funds and lists asset types they invest in (debt securities, equity, commodity-linked investments, real estate, shares in other investment funds, structured assets). (7.174; 7.176) - BPM: Notes investment funds may be open or closed ended, various legal forms, and excludes pension funds (part of insurance companies and pension funds subsector). (4.74) - Open-ended investment funds (of which: open end) - SNA: "Open-ended investment funds can issue and redeem shares on a continuous basis, or at certain defined (short-term) time intervals..." (FPFS 2014, 4.256-257) (FPFS citation retained) - MFS: "Open-ended investment funds issue and redeem shares on a continuous basis..." (3.150) - HSS: Describes open-ended funds, cites index funds and MMFs as popular types, references NAV behavior and CFI standard definition in footnote. (3.58) (footnote 6) - BPM: "Open-ended funds or open funds are those whose shares or units are, at the request of the holders, repurchased or redeemed directly or indirectly out of the undertaking’s assets." (4.74) - Closed-ended investment funds (of which: closed end) - SNA: Closed-ended funds open for subscription only at launch; thereafter shares traded on secondary market; may allow periodic repurchases at NAV-related prices. (FPFS 2014, 4.262) - MFS: Same description. (3.150) - HSS: Notes fixed number of shares, sometimes specified maturity (e.g., five to seven years); shares rarely issued post-launch and not usually redeemable until liquidation. (3.58) - BPM: "Closed-ended, closed, or exchange-traded funds are those with a fixed share capital, where investors entering or leaving the fund must buy or sell existing shares." (4.74) - Index funds - MFS: "Index-tracking funds that mirror the performance of a specific group of shares." (3.151) - HSS: "Index funds track the performance of particular stock or bond indices." (3.58) - Fund of funds - SNA: "Funds of funds are investment funds investing in other funds... holding a portfolio of other investment fund shares or units..." (FPFS 2014, Box 4.7) - MFS: "Funds of funds... hold a portfolio of other investment funds shares rather than investing directly." (3.151) - BPM: Notes investment funds that invest in other funds ("funds of funds"). (4.74) - Hedge funds - SNA: Detailed list of distinguishing features (unconstrained, high minimum investments, light regulation, wider asset ranges, long/short positions, objective to secure positive return, manager co-investment, management and performance fees, limited accredited investors, may invest in other hedge funds). Assessment requires prospectus and similar documents; data collection is difficult due to light regulation. (FPFS 2014, Box 4.7) - MFS: Concise: "Hedge funds Involving high minimum investments and light regulation, which invest in financial derivatives, take long and short positions in securities, and may sell over-the-counter derivative contracts." (3.151) - BPM: "Hedge funds are a kind of investment fund... typically involving high minimum investments, light regulation, and a wide range of investment strategies." (4.75) - Private equity funds - SNA: "Private equity funds are closed-end funds... typically limited partnerships with a fixed term of ten years (often with annual extensions). At inception, institutional investors make an unfunded commitment... drawn over the term of the fund." (FPFS 2014, 4.262, 7.42) - MFS: "Private equity funds (PEFs) ... invest in unlisted equity... usually constituted as closed-ended investment funds or as limited partnerships... included within non-MMF investment funds." (3.152) - Venture capital funds (of which: venture capital funds) - SNA: Venture capital fund definition: invests in corporations too risky for standard capital markets or banks; "Most venture capital funds have a fixed life of 10 years..." (FPFS 2014, Box 4.7) - MFS: "Venture capital funds are a subcategory of private equity funds that invest in start-up companies." (3.152) - Exchange-traded funds (ETFs) - SNA: "Exchange-traded funds (ETFs) are types of index-tracking funds... market price of an ETF is typically within 1 per cent of its NAV" and describes leveraged/inverse ETFs and arbitrage mechanism. (FPFS 2014, Box 4.7) - MFS: "A subset of index funds that are priced continuously throughout the trading day, and therefore trade like a stock." (3.151) - HSS: "Exchange-traded funds are a subset of index funds that can be traded on an exchange during the day, just like common stocks." (3.58, footnote 5) - BPM: Notes closed-ended/ exchange-traded funds definitions under closed-ended. (4.74) - Real estate investment trusts (REITs) - SNA: Describes REITs as corporations that own/operate real estate, lists equity REITs and mortgage REITs, dividend reinvestment plans, and notes REITs may not be considered funds in some countries. (FPFS 2014, Box 4.7) - MFS: "Real estate investment funds invest in debt and equity securities of companies that purchase real estate... Equity REITs are not considered financial intermediaries as they specialize in owning and managing real estate. Thus, they are not part of the FCs sector." (3.151) (Footnote 18) - BPM: "Real estate investment trusts are included [in non-MMF investment funds]." (4.74) - Insurance Corporations (S128) - SNA: "Insurance corporations consist of incorporated, mutual and other entities whose principal function is to provide life, accident, sickness, fire or other forms of insurance... Captive insurance is included... Deposit insurers, issuers of deposit guarantees and other issuers of standardized guarantees that are separate entities and act like insurers by charging premiums and have reserves, are classified as insurance corporations." (4.115) - MFS: Expanded functional description including captive insurance, reinsurance, life and non-life insurance operations, captive subsidiaries, and reinsurance purchasing behavior. (3.190-193) - HSS: Mirrors SNA. (4.29) - FSI: Detailed description of insurance activities, product types (term insurance, non-unit linked, unit-linked), composite insurers, and reinsurance mechanics. (2.47-2.50) - GFS: Defines life and nonlife insurance activities. (A4.69-70) - BPM: Mirrors SNA and lists deposit insurers, deposit protection schemes, and standardized guarantees. (4.88) - Pension Funds (S129) - SNA: Defines employer/union and government arrangements; subsector consists only of social insurance pension schemes that are institutional units separate from creators; excludes non-autonomous schemes, pay-as-you-go, and employer arrangements where reserves are included with the employer. (FPFS 2014, 2.38) - MFS: "Pension funds subsector consists of autonomous pension funds... have their own separate sets of pension-fund assets and liabilities..." (3.196) - HSS: Mirrors SNA; references BPM6 paragraph 4.90. (4.30) - FSI: Mirrors MFS description. (2.51) - GFS: Notes classification of autonomous employment-related pension funds and treatment as public financial corporation if under government control; mentions defined-benefit and defined-contribution forms (see paragraphs A2.47–A2.53). - BPM: Describes pension liabilities, organizational forms, funded/unfunded and subsector inclusion criteria. (4.89) - Defined benefit (of which) - SNA: "A defined benefit scheme is one where the benefits payable to an employee on retirement are determined by the use of a formula..." (FPFS 2014, 3.352) - MFS: "Future retirement benefits are determined by an actuarial formula related to participants’ length of service and salaries, expected retirement ages, mortality rates, etc." (3.200) - FSI: "Under a defined benefit plan, the future retirement benefits are determined by specific factors such as the participants’ length of service and salaries and age at retirement." (2.54) - GFS: Mirrors SNA description. (A2.54) - Defined contribution (of which) - SNA: "A defined contribution scheme is one where the benefits payable to an employee on retirement are defined exclusively in terms of the level of the fund built up..." (FPFS 2014, 3.352) - MFS: "Benefits to be received by a participant are based on the participant’s contributions to the pension fund and the investment performance of the fund." (3.200) - FSI: "Under a defined contribution plan, the benefits to be received by a participant are based on contributions to the pension fund and the investment performance of the fund..." (2.54) - GFS: Mirrors SNA and notes "For a defined contribution pension scheme, a pension fund is always deemed to exist." (A2.55) - Sovereign Wealth Funds (part of S127) - MFS: "Sovereign wealth funds (SWFs) are created and owned by the general government to hold, manage, or administer assets to achieve financial objectives... commonly established out of balance of payments surpluses, official foreign currency operations, privatization proceeds, fiscal surpluses, and/or receipts resulting from natural resources or commodity exports." (3.187) - BPM: Similar definition and list of common funding sources. (6.93) ### Market Intermediaries - Security and derivative dealers (part of S125) - SNA: Notes recording treatment when debt securities are marketed through underwriters/dealers; differences treated as service payments by issuers. (FPFS 2014, 5.97) - MFS: "Underwriters and dealers specialize in securities market activities... Only underwriters and dealers that act as financial intermediaries are classified in this category. Securities brokers and other units that arrange trades... but do not purchase and hold securities on their own account are classified as financial auxiliaries." (3.159) - Central clearing counterparties (part of S125) - SNA: "These organizations provide clearing and settlement transactions in securities and derivatives... The central clearing counterparties involve themselves in the transaction and mitigate counterparty risk." (FPFS 2014, 2.27) - MFS: Describes CCPs' clearing/settlement, tri-party repos inclusion, offsetting obligations, and classification as FCs but not money issuers; "CCPs should, therefore, be classified as other financial intermediaries except ICPFs." (3.157) - BPM: Mirrors SNA description. (4.77) ### Financial Market Infrastructures - Securities markets, exchanges and electronic platforms (part of S126) - MFS: "Public exchanges, securities markets, and clearing houses provide facilities in which commodities, debt and equity securities, and financial derivatives are transacted and/or settled." Lists criteria for identifying exchanges and includes associated entities (securities depository companies, accounting and clearing offices, etc.). (3.167) ### Financial Instruments - Repurchase agreements, securities lending with cash collateral, and margin lending (part of F.4) - SNA: Defines securities lending, repurchase agreements, buy/sell-backs, and notes economic effect as secured loan; legal title passes but economic ownership retained; margin calls in cash under a repo are classified as loans. (FPFS 2014, Box 4.4) - MFS: Repo defined; view from securities provider and reverse repo perspective; repo similar to collateralized loan. (4.71-72) - HSS: Treats repos and securities lending as collateralized loans; excludes them from debt securities statistics. (3.40) - FSI: Recommends recording repurchase agreements as loans collateralized by securities; securities remain on balance sheet of cash taker; repos resembling standard deposits should be classified as deposits; if securities acquired under a repo are sold to third parties, security taker records a liability equal to the current market value (short position). (5.44) - GFS: Economic nature of repo is collateralized loan/deposit; underlying securities remain on security provider's balance sheet despite legal ownership change. (7.159) - BPM: Repos, securities lending with cash collateral, and sale-buybacks are "different terms for arrangements with the same economic effect as a securities repurchase agreement—all involve the provision of securities as collateral for a loan or deposit." (5.52) - Derivatives (F.7) and subtypes - SNA: "Financial derivatives are financial instruments that are linked to a specific financial instrument or indicator or commodity..." (FPFS 2014, 4.323) - MFS: Defines financial derivatives, underlying reference prices, and purposes (risk management, hedging, arbitrage, speculation). (4.156-157) - HSS/FSI/GFS/BPM: Consistent definitions emphasizing linkage to underlying items, no principal advanced, valuation at market prices, and separate treatment from underlying items. (HSS: 2.7; FSI: 5.55; GFS: 7.204; BPM: 5.80) - Options - SNA: Definition of options (call and put) and American/European exercise. (FPFS 2014, 4.327) - MFS: Purchaser acquires right to buy/sell; purchaser pays premium. (4.176) - FSI/BPM: Similar descriptions; buyer pays premium and acquires right but not obligation. (5.61; 5.85) - Forwards and related instruments (i.e., futures) - SNA: "A forward contract is an unconditional financial contract that represents an obligation for settlement on a specified date." (11.120) - MFS: Defines forward-type contracts (forwards, futures, swaps), notes futures are exchange-traded and standardized, forwards OTC. (4.162-164) - FSI/BPM: Similar descriptions; futures traded on organized exchanges with margining requirements. (FSI: 5.57-58; BPM: 5.88-89) - Swaps (including interest rate swaps, foreign currency swaps, cross-currency interest-rate swaps, equity swaps) - SNA/MFS/FSI/BPM: Swap contracts involve exchange of cash flows based on reference prices; forward-type contracts include currency swaps, interest rate swaps, cross-currency interest rate swaps, equity swaps; obligations may arise at different times and interim payments recorded as transactions in the derivative contract. (SNA: 11.121; MFS: 4.169; FSI: 5.57; BPM: 5.91) - Interest rate swap: "An interest rate swap contract involves an exchange of cash flows related to interest payments, or receipts, on a notional amount of principal, which is never exchanged..." (SNA; MFS: 4.170) - Foreign currency swap: "A foreign exchange swap is a spot sale/purchase of currencies and a simultaneous forward purchase/sale of the same currencies." (SNA; MFS: 4.171; BPM: guidance on distinguishing derivative transaction and underlying currency transactions) (11.121; 5.92) - Cross-currency interest-rate swap: "A cross-currency interest rate swap... involves an exchange of cash flows related to interest payments and an exchange of principal amounts at an agreed exchange rate at the end of the contract." (SNA; MFS: 4.172) - Forward rate agreement (FRA) - SNA: FRA definition and settlement by net cash payments; buyer receives payment if prevailing rate exceeds agreed rate, etc. (11.121) - MFS: Adds equivalence to a swap agreement and payment math including notional principal. (4.167) - Credit derivatives - SNA: "Credit derivatives are financial derivatives whose primary purpose is to trade credit risk... take the form of both forward-type and option-type contracts..." (11.123) - MFS: Explains forms (total return swaps, CDSs), mechanics of total return swaps and CDSs, and master agreements/collateral and margining. (4.180-183) - BPM: Mirrors MFS description and emphasizes contrast with market-risk derivatives. (5.93) *IMF Working Paper — Appendix I. Current NBFI Taxonomy in Official Statistics* ### Appendix II. Guidance for Taxonomy Updates ### Appendix II. Guidance for Taxonomy Updates ### Assessment of Current Taxonomy Coverage - Includes all core elements of modern taxonomy; would benefit from updates. - Includes all core elements of modern taxonomy plus additional detail that may be removed. - Current taxonomy would benefit from clarification. - Current taxonomy is inconsistent across manuals. - Lacks definitional detail. ### Identified Gaps and Clarification Needs (as mapped against existing manuals) - Non Bank Financial Intermediaries coverage varies across: - System of National Accounts 2008 - MFS Manual and Compilation Guide 2016 - Handbook on Security Statistics 2015 - FSI Compilation Guide 2019 - Government Financial Statistics Manual 2014 - Balance of Payments Manual Sixth Edition - Specific subsector and instrument delineations shown as inconsistent or needing additional definitional detail (see taxonomy codes used throughout appendices: S123, S124, S124a, S124b, S125, S126, S127, S128, S129; instrument codes F.4, F.7). ### Recommended Taxonomy Updates for NBFI Subsectors (summary of substantive taxonomy proposals) - Money Market Funds (MMFs) (S123) - Subdivision into: - Constant Net Asset Value MMFs - Variable Net Asset Value MMFs - Of which: Government - Of which: Prime - Definitions and risk flags: - Constant NAV: maturity mismatch, collateral management, cliff and run effects if NAV falls below, or is expected to fall below, critical level; counterparty, sector, concentration risk; foreign exchange rate risk. - Variable NAV: maturity mismatch, collateral management, excessive management of NAV if coarsely measured; counterparty, sector, concentration. - Government MMFs: maturity mismatch, collateral management; counterparty risk in reverse repo positions. - Prime MMFs: maturity mismatch, collateral management; counterparty, sector risk. Prime securities include certificates of deposit (CD) and high-grade commercial paper (CP). - Non-Money Market Funds (non-MMFs) (S124) - Subdivision into: - Open end funds (S124a) - Real estate funds (Of which: Mortgage REITs; Equity REITs) - Equity funds - Bond (fixed income) funds - Mixed or balanced funds - Hedge funds - Other open-end funds (Index funds; Fund of funds) - Closed end funds (S124b) - Real estate funds (Of which: Mortgage REITs; Equity REITs) - Equity funds - Bond (fixed income) funds - Mixed or balanced funds - Hedge funds - Other closed-end funds (Funds of funds; Private equity funds; Of which: Venture capital) - Private equity funds (Of which: Venture capital) - Exchange-traded funds - Risk characteristics by fund type: - Open end funds: model, price, ineffective hedges; maturity mismatch, collateral management, short positions; counterparty, sector risk. - Closed end funds: model, price, ineffective hedges; maturity mismatch, collateral management, short positions; counterparty, sector risk. - Equity funds: model, price, ineffective hedges; maturity mismatch, collateral management, short positions; counterparty, sector risk; foreign exchange rate risk. - Fixed income funds: model, price, ineffective hedges; maturity mismatch, collateral management, short positions; counterparty, sector risk. - Index funds: model, price, ineffective hedges; liquidity of markets in which indexed securities trade. - Fund of funds: model, price, ineffective hedges; liquidity of underlying markets. - Hedge funds: leverage, maturity mismatch, short positions, collateral management; leverage lending, credit, sector, OTC contracts; price, model risk, ineffective hedges; sovereign risk from positions; foreign exchange rate risk. - Private equity and venture capital: model, price, ineffective hedges; maturity mismatch if investors not subject to lock-up period; quality of investment concerns (see closed end non-MMFs). - Exchange-traded funds (ETFs) - Risk flags: model, price, ineffective hedges; underlying market liquidity; credit or liquidity stresses in authorized participants. - Structural features: intraday trading, redemption-in-kind by selected intermediaries, market price trades near NAV, authorized participant redemption mechanisms. - Real Estate Investment Trusts (REITs) - Two types: - Equity REITs: generate rental income through owning/managing real estate properties. - Mortgage REITs: invest in financial assets backed by real estate, including mortgages and mortgage-backed securities. - Risk flags: counterparty and collateral value. - Insurance Corporations (S128) - Entities providing life, accident, sickness, fire, theft or other insurance, and reinsurance services; may provide investment funds to policy holders and retail investors. - Risk flags: model, price, ineffective hedges; counterparty, sector; leverage, maturity mismatch, short positions, collateral management. - Pension Funds (S129) - Defined benefit pension funds - Risk flags: leverage, maturity mismatch, short positions, collateral management; counterparty, sector, OTC contracts; model, price, ineffective hedges; sovereign risk from positions. - Future retirement benefits determined by actuarial formulas; contributions invested in public-market and private-market assets. - Defined contribution pension funds - Risk flags: leverage, maturity mismatch, short positions, collateral management; counterparty, sector, OTC contracts; model, price, ineffective hedges; sovereign risk from positions. - Benefits based on contributions and investment performance; no employer guarantee. - Sovereign Wealth Funds - Vehicles managed by state-affiliated entities, often focused on long-term illiquid assets. - Risk flags: leverage, maturity mismatch, short positions, collateral management; counterparty; model, price, ineffective hedges; sovereign risk from positions. - Market Intermediaries and OFIs (S125, others) - Securities dealers: leverage, maturity mismatch, short positions, collateral management; counterparty, leverage lending, sector, OTC contracts; model, price, ineffective hedges; sovereign risk from positions. - Principal trading firms: same risk profile as securities dealers; high-frequency trading characteristics noted. - Central clearing counterparties: settlement risk, model risk bearing on margin requirements; counterparty, settlement risk. - Financial market infrastructures (S126): securities markets, exchanges and electronic platforms — settlement risk; counterparty risk. ### Financial Instruments and Associated Risks - Secured financing transactions (repurchase agreements, reverse repos, security lending and borrowing, margin lending) - Primary risks: settlement risk (including failure to deliver security timely), counterparty risk, model risk. - Derivatives (F.7) and employee stock options - General description: instruments linked to a specific financial instrument, indicator, or commodity; used to trade specific financial risks (interest rate, foreign exchange, equity, commodity, credit). - Risk flags: counterparty, leverage lending, sector; model, price, ineffective hedges. - Instrument-level breakdown and associated risk flags: - Options: counterparty, leverage lending, sector; model, price, ineffective hedges. Definition: right but not obligation to buy (call) or sell (put) at predetermined strike price within specified time span. - Forwards and related instruments (i.e., futures): counterparty, leverage lending, sector; model, price, ineffective hedges. Forwards represent unconditional obligations for settlement on a specified date; futures often settled by cash payment and valued separately from underlying. - Swaps: counterparty, leverage lending, sector; model, price, ineffective hedges. Includes: - Interest rate swap contracts: fixed-rate payments swapped for floating-rate payments. - Foreign currency swaps: payments in one currency swapped for payments in another currency. - Cross-currency interest-rate swaps: fixed-rate payments in one currency swapped for floating-rate payments in another currency. - Equity swaps: one party’s payment based on stock price/index performance; other party’s payment based on fixed/floating rate or other stock/index. - Forward rate agreements (FRA): over-the-counter obligation applying predetermined interest rate to notional principal over future period. - Credit derivatives: primary purpose is to trade credit risk. - Trading venue and clearing status distinctions recommended: - Exchange traded - Over-the-counter (cleared) - Over-the-counter (not cleared) ### Structural and Classification Recommendations - Adopt finer subsector breakdowns to capture potential sources of financial stability risk, including distinctions by fund type (constant NAV vs. variable NAV), open-end vs. closed-end, and instrument-level granularities (e.g., interest rate swap contracts, FRA). - Highlight where existing statistical frameworks (SNA, MFS, GFS, BPM, Handbook on Security Statistics, FSI guides) are inconsistent or lack sufficient definitional detail, motivating updates aligned with the proposed taxonomy codes (S123–S129; F.4; F.7). - Emphasize inclusion of REITs contextually for assessing interconnectedness, noting that equity REITs are part of the nonfinancial corporate sector but included for a holistic view of real estate–financial intermediary interconnections. *Appendix II. Guidance for Taxonomy Updates — IMF Working Papers Financial Innovation and Statistical Methodological Guidance* ### Appendix V. On the Categorization of Crypto ### Appendix V. On the Categorization of Crypto Assets in Official Statistics ### Distinction between inside and outside money - Economists distinguish inside and outside money: "Inside money is inside the economy in the sense that each unit is someone’s asset and someone else’s liability. That is, inside money disappears if there is sufficient consolidation across the balance sheets of agents in the economy. Outside money, in contrast, does not disappear when balance sheets are consolidated."1 - Lagos definition emphasizing zero net supply in the private sector: "Outside money is money that is either of a fiat nature (unbacked) or backed by some asset that is not in zero net supply within the private sector of the economy...Inside money is an asset representing, or backed by, any form of private credit that circulates as a medium of exchange. Since it is one private agent’s liability and at the same time some other agent’s asset, inside money is in zero net supply within the private sector."2 ### Treatment in official statistics and resulting issues - Official statistics treat both notes and reserves of central banks as "liabilities backed" by assets held by the central bank, overlooking the inside/outside and private/public money distinction. - This accounting convention creates difficulties in categorizing some self-redeeming digital assets (crypto assets) that lack corresponding liabilities. - Self-redeeming digital currencies are described as assets that "do not yield utility as might ownership of a Picasso and is not an input into the production of other things as is farmland, a factory" (not "valuables" in SNA terminology). ### Proposal: classify financial assets by redeemability - Recommendation: characterize financial assets as either redeemable into some other financial instrument or as self-redeeming. - Self-redeeming financial assets (examples given): financial gold (gold used for financial purposes), currency, and central bank reserves. - Redeemable financial assets (examples given): commercial bank deposits and other near-monies such as money market mutual fund shares. - Rationale: commercial banks can redeem deposit liabilities by providing currency or deposits at other banks, and they hold assets to "back" deposit liabilities; central bank redemption differs because redemption by the central bank returns currency or reserves—another central bank liability—rather than some asset owned by the central bank. - Equity is noted as not self-redeeming: "Equity is not a self-redeeming asset as in bankruptcy, the equity may be 'redeemed' by allocation of the assets to be liquidated or conserved, ownership in a new enterprise, or proceeds from the sales of assets."4 ### Implications for crypto assets - Under the redeemability classification: - Some crypto assets, such as bitcoin, would be catalogued as self-redeeming (no corresponding liability, no "backing"). - Stablecoin liabilities that are redeemable into other financial instruments would be classified as redeemable (akin to commercial bank deposits). - Central bank currency and reserves are both "backed" by central bank assets and supported by legal tender status and institutional support; Bitcoin has no "backing." - This approach aligns crypto assets like Bitcoin with financial gold as a financial asset without corresponding liability, but expands the category to include currency and central bank reserves. ### Statistical recommendation: new Valuables category - Primary recommendation: create a new Valuables category within the Financial Account and classify self-redeeming crypto assets, like Bitcoin, and financial gold within this new Valuables account. - Alternative, more radical recommendation: classify all outside money into this new Valuables account (noted as harder to implement). ### Decision tree summary (classification logic) - All financial assets are first classified as either a medium of exchange or not. - Along the medium-of-exchange branch, assets are then classified as either self-redeemable or not. - Monetary assets such as currency (self-redeemable) vs. demand deposits in commercial banks (redeemable) align with economic notions of outside and inside money, respectively. - The redeemability distinction is applied to crypto assets: bitcoin = medium of exchange + self-redeemable; stablecoin tokens that offer redemptions = medium of exchange + redeemable. - General Accounts include fixed-income securities (non-medium-of-exchange branch). *Source: Appendix V. On the Categorization of Crypto Assets in Official Statistics (Espinosa-Vega and McAndrews).* --- _Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022212-print-pdf.pdf_