IMF-FSB Users Conference WANTED datasets during the financial markets turmoil - Securities and Derivatives Yuko Kawai, Bank of Japan 8 July, 2009 Agenda Market observations during the turmoil What we wanted to know Then-available datasets and what we had missed 2 Part I Market observations during the market turmoil Before the crisis “Shadow Banking” sector risk has loomed up Banks $ $$$$ $$$ $$ Final Borrowers Final Investors Brokers “Shadow Banking” $$ $$$$ Funds $$ 4 Before the crisis We did not have the full knowledge about “them” What are they? How Banks LARGE are they? What share of money $$$$ flow between final investors (and who are they?) and$$ Final Borrowersfinal borrowers isBrokers intermediated by this sector? $ $$$ Final Investors “Shadow Banking” $$ $$$$ Funds $$ 5 During the crisis Money flows dwindled, loss/defaults surged everywhere Banks Loss $ Defaults Final Borrowers $$ Defaults Brokers Final Investors Defaults “Shadow Banking” $ Funds $ Defaults 6 During the crisis Through market data, we tried to identify…. Banks How resilient are they? How are the financing/capital raising conditions? Loss $ Defaults Final Borrowers How levered were they? By what instrument? $$ Defaults Brokers How much is the current/maximum loss? $ Final Investors Defaults “Shadow Banking” Funds Defaults Who holds what risk? How much? $ 7 Other considerations Cross-border money flow/arbitrage Heightened correlation of price movements among different asset classes and market locations Maturity mismatch (short-term financing vs. long-term asset holding) Hyper-leverage through re-securitization Re-intermediation of risk upon the drawdown of liquidity facility provided by banks to off-balance sheet balance sheets 8 Part II What we wanted to know 1. Magnitude of “De-leveraging” Detect which particular market/product suffers dysfunction. ✓ Check the price movements. Estimate the current and possible maximum influence for final borrowers and investors of troubled products. ✓ Check the composition of investors and ultimate borrowers. Evaluate the systemic implication. ✓ Check the transaction volume <issuance, secondary>/outstanding amounts of such market/products. 10 Spillover of risks / Recovery progress Identify the risk transmission mechanism from the troubled ones to other markets/products. ✓ Check who the cross-over dealers/investors are. ✓ Find the cross-market trading strategies. Evaluate the impact of the central bank’s unconventional methods in the money market operation over the concerned products/market. Estimate the level or risk appetite, availability of money liquidity. ✓ Check the concerned market conditions (price, transaction volume, volatility and dispersion of the prices). 11 Implications to financial stability and the real economy Estimate the impacts of turmoil on major bank/brokers’ capital. ✓ Check the outstanding volume and potential loss exposure of troubled bank assets. ✓ Check the market conditions for bank equity issuance. Analyze the recovery of borrowings by household/corporate sector. ✓ Check the issuance volume and price of various bonds and loans. 12 Part III Then-available datasets and what we had missed Examples of “flow” products Price Flow Stock Buyers and Sellers Risk holders Listed securities Exchange Exchange Exchange (Exchange) (Exchange, FoF) Corporate Bond Vendor, Broker (clearing?) Vendor, FoF N/A (FoF) CDS Vendor, Broker DTCC ISDA, BIS rough estimate by BBA N/A GSE securities Vendor, Broker ? Vendor, FoF, Issuer N/A FoF “vendor” includes news/data vendors such as Bloomberg, MarkIt, LipperTass, Datastream, Dealogic ( ) means “not always available, depending on the product” 14 “Shadow Banking” products – tough ones RMBS Price Flow Stock Buyers and Sellers Risk holders (Broker) Primary (Vendor, Broker) (Rating Agency, Vendor) N/A N/A N/A N/A CMBS Secondary – N/A SIV N/A (Broker) (Rating Agency) N/A N/A ABCP Fed N/A Fed N/A N/A 15 “shadow banking” products, additional data required For first-level securitization: Private RMBS, CLO, CMBS, ABS … Outstanding amounts by underlying loan quality (subprime, Alt-A, Jumbo), or by rating Rating transition Second-level securitization: ABCP conduits, SIV, ARS,… Outstanding amounts by underlying asset class Liquidity (commitment line) providers, level of drawdown, trigger of drawdown 16 Market Liquidity – difficult to identify Proxy by Offer-bid, on- vs. off-the-run spreads, transaction volume, volatility, or daily high-low spreads 17 Cross market interaction - even tougher “Combination” of correlated products must be detected. By statistical analysis By collecting market-common trading strategies “Cross-over market participants” must be identified. Through regulatory bodies’ monitoring of exposures of regulated entities with global activities ? 18 Observations “Price” data is relatively easy to obtain, from exchanges, broker/dealers (through vendors or directly), or central banks’ survey. Standardized index-type derivatives were often used as the benchmark to individual cash products. Availability of “Volume” data depends. Data for “Issuance (primary flow)” and “Outstanding” can be obtained from rating agencies, vendors (Thomson Reuters, Datastream). Issuance may also be obtained from arrangers. Data of “Transaction volume (secondary flow)” of unlisted products is scarcely available. Information of “Investors”, “Buyers” and “Sellers” are almost nonexistent, except for products covered by FoF and/or ad-hoc survey. 19 What did we miss ? (1) Market data coverage is limited even for TRADITIONAL products. e.g., who are the major sellers of listed equities? Developments of hard-to-recognize products and trading infrastructure make the situation worse. e.g., dark pool, off-balance-sheet derivatives, VIEs Newly developed products/markets, if not too customized, mostly have price/issuance volume datasets while holders’ information is very limited. Investor (Holder) information and shadow banking sector information are hard to obtain. Information obtained through regulatory monitoring canNOT be freely shared. Furthermore, even monitoring information is imperfect as not all the risk holders are monitored by regulators, or they can be CROSSBOARDER. 20 What did we miss ? (2) “Genuine” risk exposure is difficult to identify. Notional amount does not necessarily reflect the risk amount. Mark-to-market value, or even the potential exposure calculated for risk management purpose under some stress scenario, did not produce useful information given the massively excessive liquidity environment. 21 Tentative Conclusion It’s impossible to obtain perfect datasets for every single product without a significant time lag. Comprehensive statistics/survey are time consuming and incur heavy costs, while information may be obsolete when published. Such statistics may not include new developments as they have to ask the same questions for the purpose of continuity, and therefore may not work as a forward-looking risk detector during the turmoil. ➵ Some sort of coordinated efforts to compile data and qualitative assessments gathered through the regulatory monitoring across the globe may help. ➵ Sharing the list of “data sources available to the public (or quasipublic)” will greatly help. ➵ Rating agencies, clearing houses, and central counterparties may be able to offer data with more details. ➵ To gather anecdotal information ahead of data collection will help, especially when changes are so quick. 22