HSBC 2015 Annual Report Download - page 353

HSBC HOLDINGS PLC
351
Strategic Report Financial Review Corporate Governance Financial Statements Shareholder Information
aim is to establish the points where the change in credit risk is considered meaningful in risk management terms and to test
these points against subsequent stage movements and defaults. Where less sophisticated default metrics are used or credit
scores are not available, as tends to apply with the less significant retail portfolios, a consistent but simplified approach is
expected to be used. In particular, for any retail portfolio, days past due will be considered in determining loans transferred to
stage 2 and the more significant portfolios will supplement this information with additional mechanisms linked to PDs. HSBC
expects to finalise the transfer criteria for the more significant portfolios during 2016.
• Stage 1
In accordance with IAS 39 (see page 356), incurred but not yet identified impairment is recognised on individually assessed
loans for which no evidence of impairment has been specifically identified by estimating a collective allowance determined
after taking into account factors including the estimated period between impairment occurring and the loss being identified.
This is assessed empirically on a periodic basis and may vary over time. Similarly, for homogeneous groups of loans and
advances which are assessed under IAS 39 on a collective basis, the inherent loss is determined using risk factors including the
period of time between loss identification and write-off which is regularly benchmarked against actual outcomes. Under IFRS
9, financial assets which are not considered to have significantly increased in credit risk have loss allowances measured at an
amount equal to 12 months ECL. This 12-month time horizon is likely to be equal to or longer than the period estimated under
IAS 39 (typically between 6 and 12 months), which will tend to result in IFRS 9 allowances being larger. In the absence of
models able to calculate IFRS 9 allowances, it is not possible to estimate the difference.
Methodologies applied to measure 12-month and lifetime expected credit losses
ECLs are calculated using three main components, i.e. a probability of default (‘PD’), a loss given default (‘LGD’) and the
exposure at default (‘EAD’). For accounting purposes, the 12-month and lifetime PDs represent the probability of a default
occurring over the next 12 months or the lifetime of the financial instruments, respectively, based on conditions existing at
the balance sheet date and future economic conditions that affect credit risk. The LGD represents losses expected on default,
taking into account the mitigating effect of collateral, its expected value when realised and the time value of money. The EAD
represents the expected balance at default, taking into account the repayment of principal and interest from the balance sheet
date to the default event together with any expected drawdown of a committed facility.
12-month ECL is calculated by multiplying the 12-month PD, LGD and EAD. Lifetime ECL is calculated using the lifetime PD
rather than the 12-month PD.
Credit loss modelling techniques
HSBC plans to base the ECL calculations on the systems used to calculate Basel expected losses (‘EL’s). This is considered to
be most efficient given the similarities in the calculations. However, certain adjustments need to be made to the Basel risk
components (PD, LGD, and EAD) to meet IFRS 9 requirements.
For wholesale portfolios and material residential mortgage and fixed-term loan portfolios, ECL will be calculated at the
individual loan level. The main adjustments necessary to Basel risk components are explained in the table below:
Model Regulatory capital IFRS 9
PD • Through the cycle (represents long-run average
PD throughout a full economic cycle)
• The definition of default includes a backstop
of 90+ days past due, although this has been
modified to 180+ days past due for some
portfolios, particularly UK and US mortgages
• Point in time (based on current conditions, adjusted
to take into account estimates of future conditions that
will impact PD)
• Default backstop of 90+ days past due for all portfolios
EAD • Cannot be lower than current balance • Amortisation captured for term products
LGD • Downturn LGD (consistent losses expected to be
suffered during a severe but plausible economic
downturn)
• Regulatory floors may apply to mitigate risk of
underestimating downturn LGD due to lack of
historical data
• Discounted using cost of capital
• All collection costs included
• Expected LGD (based on estimate of loss given default
including the expected impact of future economic
conditions such as changes in value of collateral)
• No floors
• Discounted using the original effective interest rate of
the loan
• Only costs associated with obtaining/selling collateral
included
Other • Discounted back from point of default to balance sheet
date
IFRS 9 PD and LGD estimates also have to be flexed to capture the effects of forward-looking macroeconomic variables. The
aim is to use existing stress testing models to measure these effects. Transferring between stages will be applied at individual
loan level and will also capture the effects of forward-looking macroeconomic variables.
For material non-term retail loans, transfer between stages will also be applied at individual loan level. However, loans will be
aggregated into segments based on PD or other risk drivers for the purpose of ECL measurement, to make the calculations