Академический Документы
Профессиональный Документы
Культура Документы
2
p2 = w12 112 + 2w1w2 12 + w22 22
Example:
1.
2.
3.
4.
Regression/factor approach
In general,
ri = ai + bi1 f1 + K + bik f k + i
ri = excess return of holding i
f k = return of factor j
bik = ri sensitivit y to factor k
i = residual return i
2.
3.
4.
sensitivity +
to steepening
shift
Simulation Approach
Example: Yield Book
1.
2.
3.
4.
5.
6.
Risk duration based calculations for risk factors such as volatilities and
spreads.
Aggregate returns of all securities to obtain portfolio return distribution.
11
y
P
P dy
Factor-based:
dP
dP
dP
dP
+ 2
+ 3
+ 4
excess return = 1
+K
sec
tor
dy parallel
dy steepening
dy curve
dy spread
Simulation-based:
1. Repeated draws from PC
multivariate normal distribution
2. Run MC scenario analysis
using set of PC from above
3. Aggregate to obtain price
distribution
12
14
Distribution of 10,000 paths using Monte Carlo simulation over a one-month horizon
15
Simulation Approach
Example: Yield Book
1.
2.
3.
4.
5.
Identify risk factors for each asset class (security type or grouping)
Perform PCA on covariance matrices of each security type
Build correlation matrix between PCs (i.e., simulate the correlated
occurrences of approximately 800 market variables)
Sample from distribution in 3) to obtain PC (e.g. 2 )
Transform change in PC to bond space (i.e., map PC factors to
market risk factors; two examples below)
Yield Curve:
a. Determine that a 2 PC equates to an 80 bp move in the 10-year tsy rate.
b. Perform horizon scenario analysis (via Monte Carlo simulation) to obtain individual bond
returns (i.e. estimate the returns for each bond for an 80 bp move in the 10-year tsy rate).
Spreads:
a. Determine that a 2 move in the spread PC corresponds to a 100 bp change in the spread for
a short maturity, AA, industrial, US corporate bond.
b. Perform horizon scenario analysis for each such bond to obtain individual bond spread returns
(i.e., estimate the bonds return for a 100 bp change for a short maturity, AA, industrial, US
corporate bond).
16
Simulation Approach
Example: Yield Book (contd.)
6.
7.
Also:
17
YEN 3
0.03
-0.27
0.21
0.14
-0.11
0.17
0.19
-0.11
0.04
0
0
1
18
60
40
20
0
0
10
15
20
25
30
35
-20
-40
-60
-80
Term (YR)
-2 sigma
-1 sigma
0 sigma
+1 sigma
+2 sigma
19
Curve
On The Run
Maturity
3m
6m
1Y
2Y
3Y
4Y
5Y
Japan
UK
Canada
Denmark
Switzerland
7Y
8Y
9Y
10Y
12Y
15Y
30Y
x
Model
Swap
Model
Swap
x
x
x
Model
Swap
Model
Swap
Model
Swap
Model
Model
Swap
Model
Swap
Poland
Swap*
S. Africa
Swap*
Czech
Swap*
Norway
Model
New Zealand
Swap*
Hong Kong
Swap*
South Korea
Swap*
Thailand
Model
Australia
25Y
Swap
Swap
Sweden
20Y
x
Model
Agency
EMU
6Y
Singapore
Model
Hungary
Model
x
x
x
x
20
0.444121
Return
0.5
0
-2
-1
-0.5
-0.677958
-1
-1.5
-2
-2.12611
-2.5
PC1 move (unit: std)
21
Sample output
22
23
S = aS 1 + bS
24
S =
Ss
s
Where:
S: Simulated issue-level spread change
Ss: Simulated spread move of the aggregated high yield industry sub-sector
: Issue-level historical volatility of OAS level
s: (Sub)sector-level historical volatility of OAS level
The bonds return due to spread change remains as for a corporate bond:
(Spread Return) = - P0 * D * (Spread)
25
S =
Ss
s
26
Value-at-Risk
High Grade
Monthly Return Frequency
25.00%
Frequency
20.00%
15.00%
Frequency
10.00%
Normal
95th VaR
-0.90%
99th VaR
-1.54%
5.00%
2.35
1.96
1.57
1.19
0.80
0.42
0.03
-0.36
-0.74
-0.90
-1.13
-1.52
-1.54
-1.90
-2.29
-2.67
-3.06
-3.45
0.00%
Monthly Returns
High Yield
Monthly Return Frequency
25.00%
15.00%
Frequency
10.00%
99th VaR
-2.64%
95th VaR
-1.66%
Normal
5.00%
Monthly Returns
6.17
5.44
4.72
3.99
3.26
2.54
1.81
1.09
0.36
-0.36
-1.09
-1.66
-1.81
-2.54
-2.64
-3.26
-3.99
0.00%
-4.71
Frequency
20.00%
28
29
Incremental TE
Definition:
Incremental TE = TE(all holdings)
TE(holdings in a sector eliminated)
Incremental TE captures the change of total
risk exposure when you sell out an entire
sector in your portfolio.
30
Marginal TE
Definition:
( Tracking Error )
Marginal TE = ( Market " Weight" of a Sector )
Marginal TE captures the change of total risk
exposure when you increase your bet on a
sector by small amount
31
Component TE
Definition:
Component TE =
COV ( wi Ri u i Qi , r )
COV (r , r )
32
33
34
Swaptions or options
Capture delta impact via underlying security or proxy
Use swaps as swaptions proxies
Notional value adjusted by delta
Recalibrate deltas weekly (for weekly runs)
Municipals
Shoe-horn into Yield Book
Issues that needed to be overcome:
Muni security type defines muni curve use at present.
Lack of muni curves (can load ownbut one-at-a-time)
No facility for pre-refunding (high yield bond becomes a
virtual treasury bond)
Need to incorporate inverse floaters (leveraged munis)
Specifications for various sinker retirement provisions (e.g.,
optional double sinker versus mandatory sinker
Tend to be called on first call date (change specs in YB)
Specifications of inverse floaters
Index (BMA not in YB)
Price/performance approximation
Testing in risk model as corporate bonds.
36
Other analyses
Fund complex risks (aggregate all funds)
Counterparty analyses
Estimate VaR of counterparties
Aggregate as one portfolio
Combine fixed income and equities
Credit risk
use right side of the distribution (our gain is our loss!)
Incorporate joint probabilities-of-default across counterparties
37
Future expectations
38