HSBC Holdings Regulatory News (HSBA)

Trading assets

1 Loans and advances to banks and customers include reverse repos, settlement accounts, stock borrowing and other amounts.

Trading securities valued at fair value1

1 Included within these figures are debt securities issued by banks and other financial institutions of US$22,399m (2013: US$22,989m), of which US$2,949m (2013: US$3,973m) are guaranteed by various governments.

2 Includes securities that are supported by an explicit guarantee issued by the US Government.

3 Excludes asset-backed securities included under US Treasury and US Government agencies.

Trading securities listed on a recognised exchange and unlisted

1 Included within listed investments are US$5,956m (2013: US$3,836m) of securities listed in Hong Kong.

2 Unlisted treasury and other eligible bills primarily comprise treasury bills not listed on an exchange but for which there is a liquid market.

Fair values are subject to a control framework designed to ensure that they are either determined or validated by a function independent of the risk‑taker.

For all financial instruments where fair values are determined by reference to externally quoted prices or observable pricing inputs to models, independent price determination or validation is utilised. In inactive markets HSBC will source alternative market information to validate the financial instrument's fair value, with greater weight given to information that is considered to be more relevant and reliable. The factors that are considered in this regard are, inter alia:

· the extent to which prices may be expected to represent genuine traded or tradeable prices;

· the degree of similarity between financial instruments;

· the degree of consistency between different sources;

· the process followed by the pricing provider to derive the data;

· the elapsed time between the date to which the market data relates and the balance sheet date; and

· the manner in which the data was sourced.

For fair values determined using valuation models, the control framework may include, as applicable, development or validation by independent support functions of (i) the logic within valuation models; (ii) the inputs to those models; (iii) any adjustments required outside the valuation models; and (iv) where possible, model outputs. Valuation models are subject to a process of due diligence and calibration before becoming operational and are calibrated against external market data on an ongoing basis.

Changes in fair value are generally subject to a profit and loss analysis process. This process disaggregates changes in fair value into three high level categories; (i) portfolio changes, such as new transactions or maturing transactions, (ii) market movements, such as changes in foreign exchange rates or equity prices, and (iii) other, such as changes in fair value adjustments (see further below).

The majority of financial instruments measured at fair value are in GB&M. GB&M's fair value governance structure is illustrated below as an example:

In certain circumstances, HSBC records its own debt in issue at fair value, based on quoted prices in an active market for the specific instrument concerned, where available. An example of this is where own debt in issue is hedged with interest rate derivatives. When quoted market prices are unavailable, the own debt in issue is valued using valuation techniques, the inputs for which are either based upon quoted prices in an inactive market for the instrument, or are estimated by comparison with quoted prices in an active market for similar instruments. In both cases, the fair value includes the effect of applying the credit spread which is appropriate to HSBC's liabilities. The change in fair value of issued debt securities attributable to the Group's own credit spread is computed as follows: for each security at each reporting date, an externally verifiable price is obtained or a price is derived using credit spreads for similar securities for the same issuer. Then, using discounted cash flow, each security is valued using a Libor-based discount curve. The difference in the valuations is attributable to the Group's own credit spread. This methodology is applied consistently across all securities.

Structured notes issued and certain other hybrid instrument liabilities are included within trading liabilities and are measured at fair value. The credit spread applied to these instruments is derived from the spreads at which HSBC issues structured notes.

Gains and losses arising from changes in the credit spread of liabilities issued by HSBC reverse over the contractual life of the debt, provided that the debt is not repaid at a premium or a discount.

Fair values of financial assets and liabilities are determined according to the following hierarchy:

· Level 1 - valuation technique using quoted market price:financial instruments with quoted prices for identical instruments in active markets that HSBC can access at the measurement date.

· Level 2 - valuation technique using observable inputs:financial instruments with quoted prices for similar instruments in active markets or quoted prices for identical or similar instruments in inactive markets and financial instruments valued using models where all significant inputs are observable.

· Level 3 - valuation technique with significant unobservable inputs: financial instruments valued using valuation techniques where one or more significant inputs are unobservable.

The following table sets out the financial instruments by fair value hierarchy.

Financial instruments carried at fair value and bases of valuation

The increase in Level 2 derivative balances reflects the overall increase in derivative balances and is discussed in Note 16. There were no other significant movements during 2014.

Transfers between Level 1 and Level 2 fair values

Transfers between levels of the fair value hierarchy are deemed to occur at the end of each semi-annual reporting period. Transfers from Level 1 to Level 2 mainly reflect the reclassification of settlement balances and cash collateral following reassessment of the application of levelling criteria to these balances.

Fair value adjustments are adopted when HSBC considers that there are additional factors that would be considered by a market participant which are not incorporated within the valuation model. HSBC classifies fair value adjustments as either 'risk-related' or 'model-related'. The majority of these adjustments relate to GB&M.

Movements in the level of fair value adjustments do not necessarily result in the recognition of profits or losses within the income statement. For example, as models are enhanced, fair value adjustments may no longer be required. Similarly, fair value adjustments will decrease when the related positions are unwound, but this may not result in profit or loss.

Global Banking and Markets fair value adjustments

The largest change in recurring fair value adjustments was a decline of US$403m in respect of the credit valuation adjustment, as a result of both reduced derivative counterparty exposures and general narrowing of credit default swap ('CDS') spreads. Narrowing HSBC credit default swap spreads similarly contributed to a reduction in the debit valuation adjustment ('DVA') of US$346m.

Funding fair value adjustment ('FFVA') reflects the potential future cost or benefit of funding the uncollateralised derivative portfolio at rates other than overnight ('OIS') rates. The impact of FFVA adoption in 2014 was a US$263m reduction in net trading income, reflecting the incorporation of a funding spread over Libor. FFVA is measured from an OIS base, and the total FFVA balance of US$460m also reflects the difference between OIS and Libor which had been previously reflected in the fair value of the uncollateralised derivative portfolio.

Risk-related adjustments

Bid-offer

IFRS 13 requires use of the price within the bid-offer spread that is most representative of fair value. Valuation models will typically generate mid-market values. The bid-offer adjustment reflects the extent to which bid-offer costs would be incurred if substantially all residual net portfolio market risks were closed using available hedging instruments or by disposing of or unwinding the position.

Certain model inputs may be less readily determinable from market data, and/or the choice of model itself may be more subjective. In these circumstances, there exists a range of possible values that the financial instrument or market parameter may assume and an adjustment may be necessary to reflect the likelihood that in estimating the fair value of the financial instrument, market participants would adopt more conservative values for uncertain parameters and/or model assumptions than those used in the valuation model.

Credit valuation adjustment

The CVA is an adjustment to the valuation of OTC derivative contracts to reflect within fair value the possibility that the counterparty may default and that HSBC may not receive the full market value of the transactions (see below).

Debit valuation adjustment

The DVA is an adjustment to the valuation of OTC derivative contracts to reflect within fair value the possibility that HSBC may default, and that HSBC may not pay full market value of the transactions (see below).

Funding fair value adjustment

The funding fair value adjustment is calculated by applying future market funding spreads to the expected future funding exposure of any uncollateralised component of the OTC derivative portfolio. This includes the uncollateralised component of collateralised derivatives in addition to derivatives that are fully uncollateralised. The expected future funding exposure is calculated by a simulation methodology, where available. The expected future funding exposure is adjusted for events that may terminate the exposure such as the default of HSBC or the counterparty. The funding fair value adjustment and debit valuation adjustment are calculated independently.

Model-related adjustments

Model limitation

Models used for portfolio valuation purposes may be based upon a simplifying set of assumptions that do not capture all material market characteristics. Additionally, markets evolve, and models that were adequate in the past may require development to capture all material market characteristics in current market conditions. In these circumstances, model limitation adjustments are adopted. As model development progresses, model limitations are addressed within the valuation models and a model limitation adjustment is no longer needed.

Inception profit adjustments are adopted when the fair value estimated by a valuation model is based on one or more significant unobservable inputs. The accounting for inception profit adjustments is discussed on page 378. An analysis of the movement in the deferred Day 1 P&L reserve is provided on page 395.

HSBC calculates a separate CVA and DVA for each HSBC legal entity, and within each entity for each counterparty to which the entity has exposure. HSBC calculates the CVA by applying the probability of default ('PD') of the counterparty, conditional on the non-default of HSBC, to HSBC's expected positive exposure to the counterparty and multiplying the result by the loss expected in the event of default. Conversely, HSBC calculates the DVA by applying the PD of HSBC, conditional on the non-default of the counterparty, to the expected positive exposure of the counterparty to HSBC and multiplying the result by the loss expected in the event of default. Both calculations are performed over the life of the potential exposure.

For most products HSBC uses a simulation methodology to calculate the expected positive exposure to a counterparty. This incorporates a range of potential exposures across the portfolio of transactions with the counterparty over the life of the portfolio. The simulation methodology includes credit mitigants such as counterparty netting agreements and collateral agreements with the counterparty. A standard loss given default ('LGD') assumption of 60% is generally adopted for developed market exposures, and 75% for emerging market exposures. Alternative LGD assumptions may be adopted when both the nature of the exposure and the available data support this.

For certain types of exotic derivatives where the products are not currently supported by the simulation, or for derivative exposures in smaller trading locations where the simulation tool is not yet available, HSBC adopts alternative methodologies. These may involve mapping to the results for similar products from the simulation tool or, where the mapping approach is not appropriate, using a simplified methodology which generally follows the same principles as the simulation methodology. The calculation is applied at a trade level, with more limited recognition of credit mitigants such as netting or collateral agreements than is used in the simulation methodology.

The methodologies do not, in general, account for 'wrong-way risk'. Wrong-way risk arises when the underlying value of the derivative prior to any CVA is positively correlated to the probability of default by the counterparty. When there is significant wrong-way risk, a trade-specific approach is applied to reflect the wrong-way risk within the valuation.

With the exception of certain central clearing parties, we include all third-party counterparties in the CVA and DVA calculations and do not net these adjustments across Group entities. We review and refine the CVA and DVA methodologies on an ongoing basis.

Historically, HSBC has valued uncollateralised derivatives by discounting expected future cash flows at a benchmark interest rate, typically Libor or its equivalent. In line with evolving industry practice, HSBC changed this approach in the second half of 2014. HSBC now views the OIS curve as the base discounting curve for all derivatives, both collateralised and uncollateralised, and has adopted an FFVA to reflect the funding of uncollateralised derivative exposure at rates other than OIS. The impact of adopting the funding fair value adjustment was a reduction in trading revenues of US$263m. This is an area in which a full industry consensus has not yet emerged. HSBC will continue to monitor industry evolution and refine the calculation methodology as necessary.

Financial instruments measured at fair value using a valuation technique with significant unobservable inputs - Level 3

1 Designated at fair value through profit or loss.

Level 3 instruments are present in both ongoing and legacy businesses. Loans held for securitisation, derivatives with monolines, certain 'other derivatives' and predominantly all Level 3 asset-backed securities are legacy. HSBC has the capability to hold these positions.

Private equity including strategic investments

HSBC's private equity and strategic investments are generally classified as available for sale and are not traded in active markets. In the absence of an active market, an investment's fair value is estimated on the basis of an analysis of the investee's financial position and results, risk profile, prospects and other factors, as well as by reference to market valuations for similar entities quoted in an active market, or the price at which similar companies have changed ownership.

Asset-backed securities

While quoted market prices are generally used to determine the fair value of these securities, valuation models are used to substantiate the reliability of the limited market data available and to identify whether any adjustments to quoted market prices are required. For ABSs including residential MBSs, the valuation uses an industry standard model and the assumptions relating to prepayment speeds, default rates and loss severity based on collateral type, and performance, as appropriate. The valuations output is benchmarked for consistency against observable data for securities of a similar nature.

Loans, including leveraged finance and loans held for securitisation

Loans held at fair value are valued from broker quotes and/or market data consensus providers when available. In the absence of an observable market, the fair value is determined using alternative valuation techniques. These techniques include discounted cash flow models, which incorporate assumptions regarding an appropriate credit spread for the loan, derived from other market instruments issued by the same or comparable entities.

Structured notes

The fair value of structured notes valued using a valuation technique with significant unobservable inputs is derived from the fair value of the underlying debt security, and the fair value of the embedded derivative is determined as described in the paragraph below on derivatives.

Level 3 structured notes principally comprise equity-linked notes which are issued by HSBC and provide the counterparty with a return that is linked to the performance of certain equity securities, and other portfolios. The notes are classified as Level 3 due to the unobservability of parameters such as long-dated equity volatilities and correlations between equity prices, between equity prices and interest rates and between interest rates and foreign exchange rates.

Derivatives

OTC (i.e. non-exchange traded) derivatives are valued using valuation models. Valuation models calculate the present value of expected future cash flows, based upon 'no-arbitrage' principles. For many vanilla derivative products, such as interest rate swaps and European options, the modelling approaches used are standard across the industry. For more complex derivative products, there may be some differences in market practice. Inputs to valuation models are determined from observable market data wherever possible, including prices available from exchanges, dealers, brokers or providers of consensus pricing. Certain inputs may not be observable in the market directly, but can be determined from observable prices via model calibration procedures or estimated from historical data or other sources. Examples of inputs that may be unobservable include volatility surfaces, in whole or in part, for less commonly traded option products, and correlations between market factors such as foreign exchange rates, interest rates and equity prices.

Derivative products valued using valuation techniques with significant unobservable inputs included certain types of correlation products, such as foreign exchange basket options, equity basket options, foreign exchange interest rate hybrid transactions and long-dated option transactions. Examples of the latter are equity options, interest rate and foreign exchange options and certain credit derivatives. Credit derivatives include certain tranched CDS transactions.

The following table provides a reconciliation of the movement between opening and closing balances of Level 3 financial instruments, measured at fair value using a valuation technique with significant unobservable inputs:

Movement in Level 3 financial instruments

1 Included in 'Available-for-sale investments: fair value gains/(losses)' and 'Exchange differences' in the consolidated statement of comprehensive income.

Purchases and sales of Level 3 available-for-sale assets predominantly reflect ABS activity, particularly in the securities investment conduits. Transfers out of Level 3 available-for-sale securities reflect increased confidence in the pricing of certain emerging markets corporate debt, in addition to improved price discovery of some ABSs. Transfers into Level 3 largely relate to other ABSs where price discovery has deteriorated. New issuances of trading liabilities reflect structured note issuances, mainly equity-linked notes. Transfers out of Level 3 trading liabilities principally relate to equity linked notes as certain model inputs became observable. Transfers into Level 3 trading assets primarily relate to loans in the process of syndication.

The following table shows the sensitivity of Level 3 fair values to reasonably possible alternative assumptions:

Sensitivity of fair values to reasonably possible alternative assumptions

1 Derivatives, trading assets and trading liabilities are presented as one category to reflect the manner in which these financial instruments are risk-managed.

The reduction in the effect of both favourable and unfavourable changes in significant unobservable inputs in relation to derivatives, trading assets and trading liabilities predominantly reflects greater certainty in some emerging market foreign exchange volatility, as markets have developed. The reduction in the effect of both favourable and unfavourable changes in significant unobservable inputs in relation to available-for-sale assets during the period primarily reflects a decrease in the Level 3 balances.

Sensitivity of fair values to reasonably possible alternative assumptions by Level 3 instrument type

Favourable and unfavourable changes are determined on the basis of sensitivity analysis. The sensitivity analysis aims to measure a range of fair values consistent with the application of a 95% confidence interval. Methodologies take account of the nature of the valuation technique employed, as well as the availability and reliability of observable proxy and historical data. When the available data is not amenable to statistical analysis, the quantification of uncertainty is judgemental, but remains guided by the 95% confidence interval.

When the fair value of a financial instrument is affected by more than one unobservable assumption, the above table reflects the most favourable or the most unfavourable change from varying the assumptions individually.

The table below lists key unobservable inputs to Level 3 financial instruments, and provides the range of those inputs as at 31 December 2014. The core range of inputs is the estimated range within which 90% of the inputs fall. A further description of the categories of key unobservable inputs is given below.

Quantitative information about significant unobservable inputs in Level 3 valuations