TD Bank 2014 Annual Report Download - page 79

TD BANK GROUP ANNUAL REPORT 2014 MANAGEMENT’S DISCUSSION AND ANALYSIS 77
The Bank’s Commercial Banking and Wholesale Banking businesses
use credit risk models and policies to establish borrower and facility
risk ratings, quantify and monitor the level of risk, and facilitate its
management. The businesses also use risk ratings to determine the
amount of credit exposure it is willing to extend to a particular
borrower. Management processes are used to monitor country,
industry, and borrower or counterparty risk ratings, which include
daily, monthly, quarterly, and annual review requirements for credit
exposures. The key parameters used in the Bank’s credit risk models
are monitored on an ongoing basis.
Unanticipated economic or political changes in a foreign country
could affect cross-border payments for goods and services, loans,
dividends, and trade-related finance, as well as repatriation of the
Bank’s capital in that country. The Bank currently has credit exposure
in a number of countries, with the majority of the exposure in North
America. The Bank measures country risk using approved risk rating
models and qualitative factors that are also used to establish country
exposure limits covering all aspects of credit exposure across all busi-
nesses. Country risk ratings are managed on an ongoing basis and
are subject to a detailed review at least annually.
As part of the Bank’s credit risk strategy, the Bank sets limits on the
amount of credit it is prepared to extend to specific industry sectors.
The Bank monitors its concentration to any given industry to ensure
that the loan portfolio is diversified. The Bank manages its risk using
limits based on an internal risk rating score that combines TD’s indus-
try risk rating model and detailed industry analysis, and regularly
reviews industry risk ratings to ensure that those ratings properly
reflect the risk of the industry. The Bank assigns a maximum exposure
limit or a concentration limit to each major industry segment which
is a percentage of its total wholesale and commercial exposure.
The Bank may also set limits on the amount of credit it is prepared
to extend to a particular entity or group of entities, also referred to
as “entity risk”. All entity risk is approved by the appropriate decision-
making authority using limits based on the entity’s borrower risk rating
and for certain portfolios, the risk rating of the industry in which the
entity operates. This exposure is monitored on a regular basis.
The Bank may also use credit derivatives to mitigate industry
concentration and borrower-specific exposure as part of its portfolio
risk management techniques.
The Basel Framework
The objective of the Basel Framework is to improve the consistency
of capital requirements internationally and make required regulatory
capital more risk-sensitive. The Basel Framework sets out several
options which represent increasingly more risk-sensitive approaches
to calculating credit, market, and operational RWA.
Credit Risk and the Basel Framework
The Bank received approval from OSFI to use the Basel Advanced
Internal Ratings Based (AIRB) Approach for credit risk, effective
November 1, 2007. The Bank uses the AIRB Approach for all material
portfolios, except in the following areas:
• TD has approved exemptions to use the Standardized Approach
for some small credit exposures in North America. Risk Management
reconfirms annually that this approach remains appropriate.
• TD has received temporary waivers to use the Standardized
Approach for the majority of its U.S. credit portfolios and for
some small credit portfolios. The Bank is currently in the process
of transitioning these portfolios to the AIRB Approach.
To continue to qualify using the AIRB Approach for credit risk, the
Bank must meet the ongoing conditions and requirements established
by OSFI and the Basel Framework. The Bank regularly assesses its
compliance with these requirements.
Credit Risk Exposures Subject to the AIRB Approach
The AIRB Approach to credit risk is used for all material portfolios
except in the areas noted in the “Credit Risk and the Basel Framework”
section. Banks that adopt the AIRB Approach to credit risk must report
credit risk exposures by counterparty type, each having different under-
lying risk characteristics. These counterparty types may differ from the
presentation in the Bank’s Consolidated Financial Statements. The
Bank’s credit risk exposures are divided into two main portfolios, retail
and non-retail.
Risk Parameters
Under the AIRB Approach, credit risk is measured using the following
risk parameters: PD – the likelihood that the borrower will not be able
to meet its scheduled repayments within a one year time horizon;
LGD – the amount of loss the Bank would likely incur when a borrower
defaults on a loan, which is expressed as a percentage of EAD – the
total amount the Bank is exposed to at the time of default. By applying
these risk parameters, TD can measure and monitor its credit risk to
ensure it remains within pre-determined thresholds.
Retail Exposures
In the retail portfolio, including individuals and small businesses, the
Bank manages exposures on a pooled basis, using predictive credit scor-
ing techniques. There are three sub-types of retail exposures: residential
secured (for example, individual mortgages and home equity lines of
credit), qualifying revolving retail (for example, individual credit cards,
unsecured lines of credit and overdraft protection products), and other
retail (for example, personal loans, including secured automobile loans,
student lines of credit and small business banking credit products).
The Bank calculates RWA for its Canadian retail exposures using the
AIRB approach. RWA for U.S. retail exposures are currently reported
under the Standardized Approach. All Canadian retail parameter
models (PD, EAD, and LGD) are based exclusively on the internal
default and loss performance history for each of the three retail
exposure sub-types. For each Canadian retail portfolio, the Bank
has retained performance history on a monthly basis at an individual
account level beginning in 2000; all available history, which includes
the 2001 and 2008-2009 recessions in Canada, is used to ensure that
the models’ output reflects an entire economic cycle.
Account-level PD, EAD, and LGD parameter models are built for
each product portfolio, and calibrated based on the observed account-
level default and loss performance for the portfolio.
Consistent with the AIRB Basel Framework, the Bank defines default
for Canadian exposures as 90+ day delinquency/charge-off for all retail
credit portfolios. LGD estimates used in the RWA calculations reflect
economic losses, and as such, include direct and indirect costs as well
as any appropriate discount to account for time between default and
ultimate recovery. EAD estimates reflect the historically observed utili-
zation of undrawn credit limit prior to default. PD, EAD and LGD
models are calibrated using logistic and linear regression techniques.
Predictive attributes in the models may include account attributes, such
as loan size, interest rate, and collateral, where applicable; an account’s
previous history and current status; an account’s age on books; a
customer’s credit bureau attributes; and a customer’s other holdings
with the Bank. For secured products such as residential mortgages,
property characteristics, loan-to-value ratios, and a customer’s equity
in the property, play a significant role in PD as well as in LGD models.
All risk parameter estimates are updated on a quarterly basis based
on the refreshed model inputs. Parameter estimation is fully automated
based on approved formulas and is not subject to manual overrides.
Exposures are then assigned to one of nine pre-defined PD segments
based on their estimated long-run average one-year PD.
The risk discriminative and predictive power of the Bank’s retail credit
models is assessed against the most recently available one-year default
and loss performance on a quarterly basis. All models are also subject
to a comprehensive independent validation prior to implementation
and on an annual basis as outlined in the Model Risk Management
section of this disclosure.