TD Bank 2011 Annual Report Download - page 71

TD BANK GROUP ANNUAL REPORT 2011 MANAGEMENT’S DISCUSSION AND ANALYSIS 69
migrations, defaults and idiosyncratic spread movements. Similar to
GMR, DSR is computed as the threshold level that portfolio losses are
not expected to exceed more than one out of every 100 trading days.
Trading-related revenue is the total of trading income reported in
other income and the net interest income from trading positions
reported in net interest income. Trading related revenue in the graph
below excludes revenue related to changes in the fair value of loan
commitments. The commitments are not included in the Value at Risk
measure as they are not managed as trading positions. In 2011, there
were 63 days of trading losses, with 1 breach of VaR primarily caused
by market events around the U.S. Government debt rating downgrade.
The graph below discloses daily VaR usage and trading-related revenue
within Wholesale Banking.
Calculating VaR
TD estimates total VaR on a daily basis by combining the General
Market Risk (GMR) and Debt Specific Risk (DSR) exposure associated
with TD’s trading positions. GMR is determined by creating a distribu-
tion of potential changes in the market value of the current portfolio.
We value the current portfolio using the market price and rate changes
(for equity, interest rate, foreign exchange, credit, and commodity
products) of the most recent 259 trading days. GMR is computed as
the threshold level that portfolio losses are not expected to exceed
more than one out of every 100 trading days.
DSR measures migration and default risk as well as idiosyncratic
credit spread risk for credit products in the trading portfolio. Monte
Carlo Simulation is used to capture potential changes in value due to
TOTAL VALUE-AT-RISK AND TRADING-RELATED INCOME
(millions of Canadian dollars)
20
(20)
0
$30
10
(10)
(30)
(40)
Trading-related Income
Total Value-at-Risk
Jan 31/11 Jan 24/11
Nov 1/10 Nov 8/10
Nov 15/10 Nov 22/10 Nov 29/10
Dec 6/10
Dec 13/10 Dec 20/10 Dec 27/10
Jan 3/11
Jan 10/11 Jan 17/11
Feb 7/11
Feb 14/11 Feb 21/11 Feb 28/11
Mar 7/11
Mar 14/11 Mar 21/11 Mar 28/11
Apr 4/11
Apr 11/11 Apr 18/11 Apr 25/11
May 2/11 May 9/11
May 16/11 May 23/11 May 30/11
Jun 6/11
Jun 13/11 Jun 20/11 Jun 27/11
Jul 4/11
Jul 11/11 Jul 18/11 Jul 25/11
Aug 1/11 Aug 8/11
Aug 15/11 Aug 22/11 Aug 29/11
Sep 5/11
Sep 12/11 Sep 19/11 Sep 26/11
Oct 3/11
Oct 10/11 Oct 17/11 Oct 24/11 Oct 31/11
3 The aggregate VaR is less than the sum of the VaR of the different risk types due
to risk offsets resulting from portfolio diversification.
4 Not meaningful. It is not meaningful to compute a diversification effect because
the high and low may occur on different days for different risk types.
1998 Russian debt default crisis, the aftermath of September 11,
2001, the 2007 Canadian ABCP crisis, and the collapse of Lehman
Brothers along with the ensuing credit crisis of fall 2008.
Stress tests are produced and reviewed regularly with the Market
Risk and Capital Committee.
MARKET RISK IN OTHER WHOLESALE BANKING ACTIVITIES
We are also exposed to market risk arising from a legacy portfolio of
bonds and preferred shares held in TD Securities and in our remaining
merchant banking investments. Risk management reviews and
approves policies and procedures, which are established to monitor,
measure, and mitigate these risks.
We are exposed to market risk when we enter into non-trading
banking transactions with our customers. These transactions primarily
include deposit taking and lending, which are also referred to as “asset
and liability” positions.
1 On July 22, 2011, the VaR calculation was updated to include additional risk
factors. Upon implementation, inclusion of these additional risk factors increased
VaR by approximately $2 million.
2 Interest rate risk includes credit spread risk results until July 21, 2011. Credit spread
risk is measured separately from interest rate risk as of July 22, 2011. Prior period
comparatives have not been re-classified due to this change.
Validation of VaR Model
For each of our trading portfolios, and for the portfolio as a whole, we
use a back-testing process to compare the actual and theoretical profit
and losses to VaR to ensure that they are consistent with the statistical
assumptions of the VaR model. The theoretical change in profit and
loss is generated using the daily price movements on the assumption
that there is no change in the composition of the portfolio.
Stress Testing
Our trading business is subject to an overall global stress test limit. In
addition, global businesses have stress test limits, and each broad risk
class has an overall stress test limit. Stress scenarios are designed to
model extreme economic events, replicate worst-case historical experi-
ences, or introduce severe but plausible changes in key market risk
factors. The stress testing program includes scenarios developed using
actual historical market data during periods of market disruption. The
events we have modeled include the 1987 equity market crash, the
(millions of Canadian dollars) 2011 2010
As at Average High Low As at Average High Low
Interest rate2 $ 7.5 $ 6.5 $ 10.3 $ 4.0 $ 14.4 $ 12.6 $ 20.0 $ 8.1
Credit spread risk2 9.0 8.8 12.2 4.7
Equity risk 4.1 5.3 9.4 3.8 6.4 7.8 11.3 6.1
Foreign exchange risk 1.3 3.0 5.4 1.3 1.5 2.5 6.1 0.7
Commodity risk 0.8 0.7 1.0 0.4 0.8 1.1 3.4 0.4
Debt specific risk 21.3 20.3 26.1 13.4 22.9 17.2 26.5 10.2
Diversification effect3 (19.4) (20.5) N/A N/A (18.0) (18.9) n/m4 n/m4
Total Value-at-Risk $ 24.6 $ 24.1 $ 29.0 $ 17.1 $ 28.0 $ 22.3 $ 32.0 $ 14.5
VALUE-AT-RISK USAGE1
TABLE 51