© 2014 KPMG LLP, a Delaware limited liability partnership and the U.S. member firm of the KPMG network of independent member firms affiliated with KPMG International Cooperative (“KPMG International”), a Swiss entity. All rights reserved. 3 World Bank Group Valuation Issues in Today’s Marketplace May 20, 2015 Presentation Topics Contents 1 Fair Value and Associated Adjustments 2 Credit Valuation Adjustments (CVA) / Debit/Debt Valuation Adjustments (DVA) 3 Funding Valuation Adjustments (FVA) 4 XVA 5 Curve Construction 6 MTM Swaps 7 Islamic Finance 8 Catastrophe Bonds World Bank Presentation, May 20, 2015 Fair Value Definition and Various Adjustments 1. Fair Value Definition of Fair Value 2. CVA/DVA FAS 157 defines fair value as: - Definitions and Key Drivers - Calculation Methodologies and Implementation - The price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date. World’s Bank Current Situation 3. FVA Adjustments to Base Mark-to-Market Valuation of Derivatives FAS 157 indicates that fair value should include nonperformance risk. As derivatives are commonly modeled under a risk neutral framework, a Credit Valuation Adjustment (CVA) to risk-neutral MTM becomes necessary to be fair value compliant for uncollateralized or partially collateralized positions. 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance The type of credit valuation adjustment that reflects the counterparty default risk is referred to as Counterparty Valuation Adjustment, also known as CVA. The type of credit valuation adjustment that reflects the entity’s own non-performing risk is referred to as Debt Valuation Adjustment (DVA). 8. Catastrophe Bonds CVA and DVA may be combined into a single bilateral number used to reflect the bilateral credit valuation adjustment that captures both a counterparty’s and entity’s own non-performing risks. Funding Valuation Adjustment (FVA) is a non-credit valuation adjustment. It captures the funding difference between collateralized and uncollateralized derivatives positions. FVA is a relatively new phenomenon tied to the credit crisis and its aftermath regulatory reforms especially the push for central counterparty (CCP) clearing for derivatives and collateralization for non CCP cleared derivatives. World Bank Presentation, May 20, 2015 Fair Value Definition and Various Adjustments 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Calculation Fully Collateralized Derivatives Fully collateralized derivatives either by ISDA agreements or through cleared by CCP assume zero CVA or DVA. Fair value calculation for fully collateralized derivatives is performed using overnight index swap (OIS) discounting. Methodologies and Implementation - World’s Bank Current Partially Collateralized Derivatives Situation Non-zero collateral posting threshold and non-zero minimal collateral posting amount cause partial collateralization of derivatives. CVA/DVA for partially collateralized derivatives is typically performed by assuming CVA/DVA for the collateralized portion of the position to be zero, while calculating CVA/DVA for the uncollateralized portion. 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds If OIS is used for discounting, FVA is often calculated to account for the difference between OIS and the funding cost for uncollateralized portion. World Bank Presentation, May 20, 2015 Credit Valuation Adjustments (CVA) and Debit/Debt Valuation Adjustments (DVA) CVA - Definition and Key Drivers 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Definition CVA represents the risk to an entity that its counterparty will not be able to fulfill its end of a derivative transaction. CVA quantifies expected losses due to its counterparty’s default, and because of this, CVA brings down the value of an entity’s derivative assets. Calculation Methodologies and Implementation Key Drivers 3. FVA Like other expected loss calculations, CVA is an exposure * probability of default * loss given default calculation. CVA increases when each of the following increases for a given position, or portfolio of positions: 4. XVA • Counterparty credit spreads 5. Curve Construction • Loss given default (loss severity) 6. MTM Swaps • Time to maturity 7. Islamic Finance • Expected positive future exposures to the counterparty - World’s Bank Current Situation 8. Catastrophe Bonds We note the following: • PD is usually a positively sloped term structure and therefore long-date expected future exposures have more impact on CVA • Loss given default (LGD) is calculated as 1 – recovery and often treated as a constant. • As a first order approximation, credit spread can be considered as the product of PD * LGD. For an entity, the higher the seniority on the capital structure, the lower the credit spread for the liability is. For example, a secured loan margin is usually lower than an unsecured bond spread. World Bank Presentation, May 20, 2015 Credit Valuation Adjustments (CVA) and Debit/Debt Valuation Adjustments (DVA) DVA - Definition and Key Drivers 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Definition DVA measures the risk to that entity’s counterparties from its own credit and represents the risk to an entity’s counterparties that it will not be able to fulfill its end of a derivative transaction. As a result, DVA brings down the value of an entity’s liabilities. Calculation Methodologies and Implementation Key Drivers 3. FVA Like CVA, DVA is also an exposure * probability of default * loss given default calculation. DVA increases when each of the following increases for a given position, or portfolio of positions: 4. XVA • Own credit spreads 5. Curve Construction • LGD 6. MTM Swaps • Time to maturity 7. Islamic Finance • Expected negative future exposures to the counterparty - World’s Bank Current Situation 8. Catastrophe Bonds World Bank Presentation, May 20, 2015 Credit Valuation Adjustments (CVA) and Debit/Debt Valuation Adjustments (DVA) Calculation Methodologies and Implementation 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Calculation Methodologies and Implementation - World’s Bank Current Situation Calculation Methodologies – Analytical Formulas versus Monte Carlo Simulation In some cases, CVA/DVA for individual positions may be calculated using analytical formulas where there are closed-form solutions to determining expected future exposures. However, there are many derivatives where analytical formulas cannot be used and the projection of future exposures cannot be done so easily. For this reason (among others), many larger banking organizations, including all of the largest money center banks, calculate CVA/DVA using Monte Carlo simulation to project future exposures, and then apply probabilities of default and recovery rate assumptions to calculate expected nonperformance risk. 3. FVA 4. XVA Individual Position-level versus Portfolio-level calculations 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds CVA/DVA may be calculated for individual positions or on a counterparty portfolio-level. Portfolio-level CVA/DVA is often preferred or even required by netting agreements. Portfolio-level CVA/DVA calculations essentially require that a Monte Carlo simulation methodology be used in order to capture all of the correlated movements in market data that drive derivative valuations. As CVA/DVA is not additive, it can be difficult to allocate portfolio-level CVA/DVA to individual positions. The marginal CVA/DVA approach (computing the net CVA/DVA of a portfolio with an without that position) may be most accurate but can be time consuming. World Bank Presentation, May 20, 2015 Credit Valuation Adjustments (CVA) and Debit/Debt Valuation Adjustments (DVA) Calculation Methodologies and Implementation (continued) 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Calculation Methodologies and Implementation - World’s Bank Current Situation 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Netting CVA and DVA Together – What to Expect and What Could Happen Between two parties of roughly equal creditworthiness, netting the two together typically results in a valuation that is smaller in absolute value; assets would become smaller in size, as would liabilities. In some cases, however, calculating and applying CVA/DVA may result in an asset that turns into a liability, or vice versa. While counterintuitive, these situations are more likely to occur when: • There is a large discrepancy in credit spreads between the parties involved • The mark-to-market value before adjustment of the position is relatively small and projected to flip from an asset to liability, or vice versa • There is a long time to maturity A large discrepancy in credit spreads could mean that even if the potential future exposures are projected to remain relatively stable over time, the calculated CVA (or DVA) may outweigh the other due solely to the large difference in default probabilities. The last two reflect the possibility that the exposure itself may change over time, so even if you were to exclude CVA and DVA altogether, the position may be expected to flip to begin with, making an initially unexpected CVA/DVA calculation more understandable. World Bank Presentation, May 20, 2015 Credit Valuation Adjustments (CVA) and Debit/Debt Valuation Adjustments (DVA) World Bank’s Current Situation 1. Fair Value 2. CVA/DVA - Definitions and Key Drivers - Calculation Methodologies and Implementation - World’s Bank Current Situation 3. FVA 4. XVA Implications for World Bank The World Bank Group has ISDA (International Swaps and Derivatives Association) agreements with all of its counterparties that require its counterparties to post collateral to the World Bank in the event a derivative becomes a liability to them (and an asset to the World Bank). These agreements also require that the World Bank post collateral in the event that World Bank’s credit rating falls below AAA, which means that no collateral is being posted on the part of the World Bank right now. What this means is that World Bank is not exposed to the nonperformance risk of its counterparties (CVA is zero) because collateral is being posted. The DVA for World Bank is also zero because its AAA rating means the market thinks the market-perceived probability of World Bank defaulting is close to zero. 5. Curve Construction 6. MTM Swaps OIS or LIBOR Discounting? 7. Islamic Finance Because of the one-way collateralization, neither OIS discounting nor LIBOR discounting is perfect for World Bank’s derivatives valuation. Using OIS for discounting and then applying a liability FVA (LFVA) to capture funding difference for WB’s liability discounting is one of the acceptable approaches. 8. Catastrophe Bonds World Bank Presentation, May 20, 2015 Funding Valuation Adjustments (FVA) Definition and Controversy 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction Definition FVA is a post-crisis phenomenon after the 2008 credit crisis as more derivatives are either collateralized or cleared by CCPs. It is used to reflect the impact of additional funding cost over OIS on the valuation of uncollateralized derivatives positions. FVA becomes significant for broker dealers when they enter transactions in a retail market where a substantial portion of derivatives positions is uncollateralized and then hedge them in a broker dealer market where almost all derivatives transactions are collateralized. 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds How it Works FVA is similar to CVA and DVA in that its magnitude depends on the size of expected uncollateralized future exposures. For that reason, large financial institutions have built their FVA calculation based upon their CVA framework for expected future exposures calculation. Essentially, the only additional critical model input for FVA is the funding spread over OIS. Market Approach vs. Entity Specific Approach FVA using an entity’s own funding is more accurate and easier to be hedged while FVA using a market average funding cost has the backing of the fair value accounting guidelines, which says the fair value of a derivative is not entity specific but the exit price accepted by general market participants. Many large US money center banks have adopted the market approach by estimating funding spread from the market consensus provided by Markit. We note that using the spread of LIBOR over OIS for FVA calculation assumes LIBOR to be the market average funding for uncollateralized positions. World Bank Presentation, May 20, 2015 Funding Valuation Adjustments (FVA) Definition and Controversy (continued) 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps Symmetric vs. Asymmetric Funding Costs Some large banks consider funding costs to be asymmetric with respect to assets and liabilities. The evidence for such an approach is about half of Totem’s participants quote asymmetric FVA. You may also find support for the argument in accounting rules – asset valuation requires marking to exit prices while liability valuation requires assuming a hypothetical transfer to a 3rd party of equal credit standing. Using LIBOR-OIS spread for FVA calculation assumes funding cost is symmetric. 7. Islamic Finance 8. Catastrophe Bonds Netting Considerations Two levels of netting need to be considered – counterparty level and funding desk level. As default of a counterparty will cause termination of existing positions with that counterparty and therefore eliminate the need to fund these positions, expected future exposures for FVA are calculated contingent on the survival of the counterparty, similar to CVA/DVA calculation. In other words, certain aspects of expected future exposure calculation for FVA need to be done at counterparty level. That is the reason why FVA calculation often leverages the infrastructure constructed for CVA/DVA calculation. However, funding is also affected by the netting at the entire funding desk level provided rehypothecation is allowed. That means across counterparty netting is sometimes necessary for FVA calculation. If a symmetric funding cost is taken, e.g. LIBOR-OIS spread is used for both asset and liability FVA calculation, netting at any level is trivial. However, that’s not the case for any asymmetric approaches. World Bank Presentation, May 20, 2015 Catch-All Valuation Adjustments (XVA) XVA – and Basel Steps in 1. Fair Value 2. CVA/DVA 3. FVA Operational Inefficiencies and Creation of the XVA Desk Many organizations have created an XVA Desk to allow an entity to price a derivative in real time with all of these adjustments at once through one centralized desk. 4. XVA 5. Curve Construction 6. MTM Swaps Credit Valuation Adjustment (CVA) Desk Debt/Debit Valuation Adjustment (DVA) Desk Funding Valuation Adjustment (DVA) Desk Various Other Adjustments 7. Islamic Finance 8. Catastrophe Bonds XVA Desk Uncertain Outcome On 3/20/2015, the Basel Committee on Banking Supervision launched a project reviewing FVA at various banks – and so have US bank supervisors. Given that more than 20 banks have begun to implement FVA, with varying methodologies, losses have ranged wildly ($1.5 billion at JPMorgan, 267 Swiss Francs at UBS). While it’s not certain what the scope of this project will entail, market participants do not expect regulators to speak about FVA specifically but may provide guidance on the relationship between FVA and DVA, as the two capture similar risks. World Bank Presentation, May 20, 2015 Multi-Curve Construction Shift from Single to Dual Curve Bootstrapping 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Rise of OIS Discounting and Impact The increase in the number of collateralized derivatives trades after the credit crisis and shift to OIS discounting for those deals forced a change in the manner in which LIBOR rates are quoted. For collateralized positions, whereas LIBOR is still used for cashflow projection, OIS is used for discounting the cash flows. Changes in Curve Construction – Single and Dual Curve Previous curve construction methodologies were “single curve” methodologies, which were not only general market practice, but also very computationally quick. Current curve construction methodologies call for a “dual curve” model, where OIS and LIBOR curves are generated simultaneously. Various steps can be taken as part of this curve construction process, with a tradeoff between accuracy and computational time. In a dual curve methodology, the following steps are generally taken: • Bootstrap OIS curve instruments to generate OIS discount factors • Use LIBOR instruments and OIS discount factors to bootstrap LIBOR curve • Bootstrap OIS again using the generated forward LIBOR curve • Repeat until the desired level of tolerance is reached • Use this forward LIBOR curve to generate other LIBOR tenor and cross-currency basis- adjusted curves Another popular alternative is global bootstrapping, a computational method used to generate both curves at once with the goal of optimizing the fit of both curves to the market data provided though this method is computationally expensive, with other various methodologies used to try to optimize the tradeoff between accuracy and speed. World Bank Presentation, May 20, 2015 Mark-to-Market (MTM) Cross Currency Swaps Variant of Typical Cross Currency Swaps 1. Fair Value 2. CVA/DVA 3. FVA Typical Cross Currency Swaps Most cross currency swaps are traded with a fixed notional. This can cause significant FX risk, as changes in FX rates between the time a swap is entered into and the final notional exchange can cause significant gains and losses and drive most of the value of these types of positions. 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Mark-to-Market (MTM) Cross Currency Swaps This significant FX risk has given rise to a new form of cross currency swaps, referred to MTM swaps or resetting swaps. In these positions, the notional on one side is fixed, while the notional on the other leg varies periodically, typically resetting with each payment period. A typical example would be a EUR/USD cross-currency swap in which the EUR leg is fixed at a pre-specified notional amount (i.e. 100 million EUR) and pays a floating rate (i.e. 3M EURIBOR). The USD leg would have an initial notional amount based on the FX rate at the time the swap was entered into (i.e. 1.25) and pay 3M LIBOR, with its notional amount resetting based on the observed EUR/USD FX rate at the time of each payment. These notional resets would trigger a principal exchange equal to the difference in the new principal amount and the then-current principal amount, but would also reduce the impact of the final principal exchange at maturity as well. Date EUR Notional FX Rate USD Notional USD Notional (Paid)/Received by USD Leg 12/31/2014 EUR 100MM 1.25 USD 125MM 3/31/2015 EUR 100MM 1.20 USD 120MM (5MM) 6/30/2015 EUR 100MM 1.30 USD 130MM 10MM World Bank Presentation, May 20, 2015 Islamic Finance Principles and Products 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Rise of the Industry Islamic banking is a subset of banking activities that is meant to be in compliance with sharia, or Islamic law. While sharia has been codified for some time, the development of Islamic banking practices and principles did not begin to develop until relatively recently; though since its development, growth has been tremendous. Per Ernst & Young, Islamic banking assets have grown at an annual rate of 17.5% between 2009 and 2013, and are expected to grow an average of 19.7% a year through 2018. This growth is not due entirely to a growth in Islamic investors trying to adhere to sharia, but also value investors looking for a good deal. Principles behind Islamic Finance Products Sharia prohibits the following activities: • Usury - payment or charging of interest • Maysir – speculation • Investment in activities deemed sinful under Islamic law (alcohol, pork, gambling) World Bank Presentation, May 20, 2015 Islamic Finance Principles and Products (continued) 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Examples of Products In practice, these prohibitions mean that workarounds are made to achieve compliance with sharia, though they may seem obtuse at first. Because sharia prohibits the payment or charging of interest, Islamic mortgages are structured differently from a typical mortgage. Instead, a bank may buy a home from its seller directly, and then resell the home back to the original intended buyer at a profit, allowing the buyer to make those above-market rate payments, nominally all principal, in installments that would be equal to the payments due under a more typical mortgage. A sukuk, or Islamic bond, does not pay interest per se. Instead, the buyer owns a nominal share of whatever the money raised from issuing the sukuk purchased, and receives money derived from the profit generated from that asset or from rental payments made by the issuer to compensate the buyer for its use. Functionally, however, these payments are typically structured such that they are no different from a fixed or floating rate bond. In addition, the prohibition on speculation generally rules out derivatives, options and futures, because they do not represent an investment in their underlying assets and are considered to be speculative gambles on their movements. As a result, many banks adhering to sharia withstood the credit crisis better than their less pious counterparts because of their relatively smaller derivative exposures. World Bank Presentation, May 20, 2015 Catastrophe Bonds Market and Motivation 1. Fair Value 2. CVA/DVA 3. FVA 4. XVA 5. Curve Construction 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds Definition Catastrophe bonds (also commonly called cat bonds) are a specific type of bond, typically used by insurance companies, to transfer the risk of catastrophic events (hence the name) insured by the issuing companies. As an example, a catastrophe bond may pay a specified coupon to the holder, but if the specific event type named (i.e. hurricane or earthquake) in the bond occurs, the issuer would keep some or all of the principal invested and not be required to repay that to their investors. These bonds are usually issued by special purpose vehicles which are then cash collateralized; therefore, the risk of principal default is not the risk that the issuer will default, but that a catastrophic event will occur during the lifespan of the bond – usually around or less than three years. Motivations for the Broad Insurance Market Cat bonds were created in response to Hurricane Andrew, which took place in 1992. The motivation for the market at large was to share insurance risk among more possible participants, as catastrophe reinsurers began to become reluctant to bear too much reinsurance risk: World Bank Presentation, May 20, 2015 Catastrophe Bonds Market and Motivation (continued) 1. Fair Value Motivations for Individual Issuers and Investors 2. CVA/DVA The motivation for individual issues were numerous as well: 3. FVA • The ability to achieve a fixed price for reinsurance over a potentially longer time frame 4. XVA • Potentially lower cost of reinsurance 5. Curve Construction • Quicker payouts in the event of catastrophic events by being able to keep principal 6. MTM Swaps 7. Islamic Finance 8. Catastrophe Bonds already invested rather than waiting for reinsurance payouts • Off-balance sheet SPV provide benefits with respect to capital requirements and help with issuer credit risk The motivation for investors such as hedge funds, pension funds, asset managers, and other large institutional investors include high coupons (ranging from 3-20% over LIBOR depending upon the risk), and the fact that their value is usually uncorrelated to other more standard investments in fixed income and equities. While this makes hedging these instruments difficult, the added diversification is often considered to outweigh that risk. World Bank Presentation, May 20, 2015