lOMoARcPSD|9600418

Formuleblad

Investment Analysis (Tilburg University)

StuDocu is not sponsored or endorsed by any college or university

Downloaded by Denis Dumea ([email protected])
 lOMoARcPSD|9600418

FORMULAS IA
Part 1
5% probability that portfolio will fall in value by more than VaR over one year period.
𝑉𝑎𝑅$%%&$' = 𝑊+ 1.64𝜎$%%&$' − 𝜇$%%&$'
𝑉𝑎𝑅$%%&$' = 𝑊+ 12 ×1.64𝜎56%+7'8 − 12 ×𝜇56%+7'8

GUISE: 𝑥:.;: = 0.3125𝑥:.:; + 0.4375𝑥:.:A + 0.2500𝑥:.;: 𝑥B = 𝑊: ×𝑒𝑥𝑝 𝐻𝜇 + 𝑧B 𝜎 𝐻

𝑟H = 𝑟I + 𝛽H 𝑟5 − 𝑟I
𝜎K = 𝑦𝜎M

𝐶𝑜𝑣 𝑟H , 𝑟5
𝛽H =
𝑉𝑎𝑟 𝑟5
U
𝜌ST = VW 𝜎ST = 𝜌ST × 𝜎S × 𝜎T
UV UW

𝑟K = 𝑤S 𝑟S + 𝑤T 𝑟T
Y
𝜎ST = 𝑤SY 𝜎SY + 𝑤TY 𝜎TY + 2𝑤S 𝑤T 𝜎ST

1
𝑈 𝐶 = 𝐸 𝑟K − 𝐴 ×𝑉𝑎𝑟 𝑟K
2

Matrix notation
𝑤; 1
b
𝑤 = ⋮ 𝜄 = ⋮ weight of risk-free asset 𝑤`a = 1 − Hc; 𝑤H = 1 − 𝑤 d 𝜄
𝑤^ 1

𝐸 𝑟; − 𝑟I
Return of combination portfolio: 𝑟K = 𝑟I + 𝑤 d 𝑟Me where 𝑤 d = 𝑤; … 𝑤^ and 𝑟Me = ⋮
𝐸 𝑟^ − 𝑟I
𝐸 𝑟; − 𝑟I
d e ⋮ e
Expected return: 𝐸 𝑟K = 𝑟I + 𝑤 𝜇 where 𝜇 =
𝐸 𝑟^ − 𝑟I
𝜎;Y ⋯ 𝜎;b 𝑤;
d
Variance: 𝑉𝑎𝑟 𝑟K = 𝑤 𝑤 = 𝑤; … 𝑤^ ⋮ ⋱ ⋮ ⋮
𝜎b; ⋯ 𝜎bY 𝑤^

;
Find optimal complete portfolio weights (maximize U): 𝑈 𝐶 = 𝐸 𝑟K − 𝐴 ×𝑉𝑎𝑟 𝑟K
Y
;
= 𝑟I + 𝑤 d 𝜇 e − 𝐴 ×𝑤 d 𝑤
Y
e
𝜇 − 𝐴Σ𝑤 = 0
𝐴Σ𝑤 = 𝜇 e
𝑤 ∗ = 𝐴k; Σ k; 𝜇 e

Downloaded by Denis Dumea ([email protected])

 lOMoARcPSD|9600418

Optimal portfolio with exogenous risk

𝑟K = 𝑟I + 𝑤 d 𝑟Me + 𝑞 × 𝑟m
𝐸 𝑟K = 𝑟I + 𝑤 d 𝜇 e + 𝑞𝜇m
𝑉𝑎𝑟 𝑟K = 𝑤 d Σ𝑤 + 𝑞 Y 𝜎mY + 2𝑤 d 𝜎`n 𝑞
𝑤 ∗ = 𝐴k; Σ k; 𝜇 e − 𝑞Σ k; 𝜎`n

Part 2
Equity
𝑚𝑎𝑟𝑘𝑒𝑡 𝑝𝑟𝑖𝑐𝑒
𝑇𝑜𝑏𝑖𝑛𝑠 𝑞 =
𝑟𝑒𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑎𝑠𝑠𝑒𝑡𝑠
T S• € T •• k •‚
Expected stock return: 𝐸 𝑟 = 𝐷𝑖𝑣𝑖𝑑𝑒𝑛𝑑 𝑌𝑖𝑒𝑙𝑑 + 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐺𝑎𝑖𝑛 𝑌𝑖𝑒𝑙𝑑 =
•ƒ
T S• € T ••
Intrinsic value: 𝑉: =
;€^
1. DDM:
S• S„ S… € •…
a. Finite horizon: 𝑉: = + + ⋯ +
;€^ ;€^ „ ;€^ …
‡ S
b. Infinite horizon: 𝑉6 = +c; † †
;€^
c. Dividend growth:
i. Constant dividend growth model: 𝑔 = 𝑅𝑂𝐸 × 𝑏
1. 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠+€; = 𝐵𝑜𝑜𝑘𝑣𝑎𝑙𝑢𝑒 +€; ×𝑅𝑂𝐸
2. 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠+€; = 𝐵𝑜𝑜𝑘𝑣𝑎𝑙𝑢𝑒+ + 𝐵𝑜𝑜𝑘𝑣𝑎𝑙𝑢𝑒+ ×𝑏 ×𝑅𝑂𝐸 ×𝑅𝑂𝐸
a. 𝑅𝑒𝑡𝑎𝑖𝑛𝑒𝑑 = 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠 ×𝑏
b. 𝐷𝑖𝑣𝑖𝑑𝑒𝑛𝑑 𝑝𝑎𝑦𝑒𝑑 𝑜𝑢𝑡 = 𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠 × 1 − 𝑏
Ž ;€ Œ• † S• ;€ Œ„
ii. Multistage growth models: 𝑃: = 𝐷: +c; ;€^ † + ^k Œ ;€^ •
„
S• ;€Œ
1. Determine price at end with 𝑃: = , add and discount
^kŒ
2. P/E ratios and growth opportunities:
𝑃𝑟𝑖𝑐𝑒 = 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐴𝑠𝑠𝑒𝑡𝑠 𝑖𝑛 𝑃𝑙𝑎𝑐𝑒 + 𝑉𝑎𝑙𝑢𝑒 𝑜𝑓 𝐺𝑟𝑜𝑤𝑡ℎ 𝑂𝑝𝑝𝑜𝑟𝑡𝑢𝑛𝑖𝑡𝑖𝑒𝑠
•ƒ ; ••‘’
a. = 1 + T
T• ^ ^
T
i. If PVGO=0, 𝑃: = •
^
S• T
ii. 𝑃𝑉𝐺𝑂 = 𝑃𝑟𝑖𝑐𝑒 𝑝𝑒𝑟 𝑠ℎ𝑎𝑟𝑒 − 𝑁𝑜𝑔𝑟𝑜𝑤𝑡ℎ 𝑣𝑎𝑙𝑢𝑒 𝑝𝑒𝑟 𝑠ℎ𝑎𝑟𝑒 = − •
^kΠ^
•ƒ ;k”
b. =
T• ^k•’T×”
• ;k”
c. =
T ^kŒ

Empirical Evidence CAPM
1. Estimating betas
a. Theory: 𝐸 𝑟H = 𝑟I + 𝛽– 𝐸 𝑟— − 𝑟I
b. Data: 𝑹𝒊𝒕 − 𝑹𝒇𝒕 = 𝜶𝒊 + 𝜷𝒊 𝑹𝒎𝒕 − 𝑹𝒇𝒕 + 𝜺𝒊𝒕 estimate using time series regressions
2. Regressing returns on betas (average returns on constant β, for every month on time-varying β)
a. Theory: 𝐸 𝑅H+ − 𝑅I+ = 𝜆+ 𝛽+ cross-sectional return-beta relationship
b. Data: 𝑹𝒊𝒕 − 𝑹𝒇𝒕 = 𝝀𝟎𝒕 + 𝝀𝟏𝒕 𝜷𝒕 + 𝜺𝒊𝒕 estimating SML

Downloaded by Denis Dumea ([email protected])

 lOMoARcPSD|9600418

If CAPM true, no other factors should explain time-series and cross-sectional difference in returns à all
other variables should be 0. If CAPM holds, returns only depend on β of market.

Equity premium puzzle: 𝐸 𝑅— − 𝑅I = 𝐴𝐶𝑜𝑣 𝑅— , 𝑅¤ investors care more about consumption

Bonds
¤ ;€Œ %
Annuity: 𝑃 = 1 −
`kŒ ;€`
Ž ¤ •$`•$'&e
𝑃¥ = +c; ;€` † + .
;€` •

¦§†
Ž •¨© †
Duration: 𝐷 = +c; 𝑡 × 𝑤+ , 𝑤+ =
•`HKe
∗ S
Modified duration: 𝐷 =
;€8
ƥ
Duration approximation: = −𝐷 ∗ × ∆𝑦
•
Good for small changes in yield, overestimation of losses and underestimation of gains
•
ƥ
• = −𝐷 ∗ × ∆𝑦 + 1 2 𝐶𝑜𝑛𝑣𝑒𝑥𝑖𝑡𝑦 × ∆𝑦 Y
•
; % ¤«† Y
Convexity: „ × +c; † 𝑡 + 𝑡
• ;€8 ;€8

Macroeconomic risk
Demand shock: affects goods and services, Y↑ r↑ π↑ overestim
Supply shock: affects production capacity/cost, Y↓ r↑ π↑

Certainty
Relationship spot rates and short rates: 1 + 𝑦Y Y = 1 + 𝑟; 1 + 𝑟Y geometric average short rates
𝑟Y > 𝑟; : yield curve slopes up (rates expected to rise, more competition for capital à good economy)
𝑟Y < 𝑟; : yield curve slopes down (rates expected to fall, less competition for capital à bad economy)

Uncertainty
Relationship: 1 + 𝑦Y Y = 1 + 𝑟; 1 + 𝐸 𝑟Y à 1 + 𝑦% % = %Hc; 1 + 𝑟H

;€ 8® ®
Forward rate: 1 + 𝑓% = expectations today locked in
;€ 8®¯• ®¯•
%
1 + 𝑦% = 1 + 𝑦%k; %k; 1 + 𝑓%
;::: ;:::
• 𝑃% = =
;€ 8® ® ;€ 8®¯• ®¯• ;€ I®
«•
• 𝑃; = = 𝑃Y 1 + 𝑓%
;€8 ®

Expectations hypothesis: fn = E(rn)
• Upward sloping yield curve implies expected increase in short rate.
Liquidity preference: fn ≠ E(rn) LT bonds are riskier
• Increase in yield might be caused by premium (decreasing liquidity premium unlikely)

Downloaded by Denis Dumea ([email protected])

 lOMoARcPSD|9600418

Portfolios
Active part:
²³
• w±: = divide by total amount to find weight of each share within active part
´„ µ³
·¸
¹„ º¸
• w¶: = » ¼½

¹„½
o 𝛼¿ = %Hc; 𝑤H 𝛼H
o 𝜎 Y 𝑒¿ = %Hc; 𝑤HY 𝜎 Y 𝑒H
o 𝛽¿ = %Hc; 𝑤H 𝛽H
ƒ
ÀÁ
• 𝑤¿∗ = ƒ beta correction
;€ ;k ÂÁ ÀÁ
• Final weights:
o 𝑤H∗ = 𝑤¿∗ 𝑤H
o 𝑤— ∗
= 1 − 𝑤¿∗ portion in passive portfolio
• Data of optimal risky portfolio:
o 𝛽• = 𝑤— ∗
+ 𝑤¿∗ 𝛽¿ = 1 − 𝑤¿∗ 1 − 𝛽¿
o 𝐸 𝑅• = 𝛽• 𝐸 𝑅— + 𝑤¿∗ 𝛼¿
o 𝜎•Y = 𝛽• 𝜎—Y
+ 𝑤¿∗ 𝜎 𝑒¿ Y add idiosyncratic risk of active part
BÄ Y
%
o 𝑆•Y = 𝑆—
Y
+ Hc; U e ultimate goal to improve SR
Ä

Treynor-Black
Basic idea: allow for more idiosyncratic risk in order to get α à market/passive + actively managed part
(exposed to idiosyncratic risk).

Black-Litterman
Core idea: 𝑤; ×𝑃𝑟𝑖𝑜𝑟 + 𝑤Y ×𝐷𝑎𝑡𝑎 = 𝑃𝑜𝑠𝑡𝑒𝑟𝑖𝑜𝑟

Performance measures: capture risk-return tradeoff
• Portfolio is added to existing portfolios, systematic risk:
•Å k •§
o 𝑇𝑟𝑒𝑦𝑛𝑜𝑟 = slope of SML
ÂÅ
o 𝛼• = 𝑅• − 𝑅« − 𝛽• 𝑅— − 𝑅«
Performance attribution:
• Portfolio in isolation, total risk:
1. Allocation across
•Å k •§
o 𝑆𝑅 = asset classes
UÅ
o 𝑀 Y = 𝑟•∗ − 𝑟— portfolio with same volatility as market to compare 2. Security choice
within sector
• Adding small number of stocks to passively managed portfolio, idiosyncratic risk:
𝜶
o 𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑟𝑎𝑡𝑖𝑜 = 𝑷
𝝈 𝒆𝑷

Test for market timing ability: regress return on market
1. 𝑟• − 𝑟I = 𝑎 + 𝑏 𝑟— − 𝑟I + 𝑐 𝑟— − 𝑟I 𝐷 + 𝑒• two betas (Henriksson and Merton)
Y
2. 𝑟• − 𝑟I = 𝑎 + 𝑏 𝑟— − 𝑟I + 𝑐 𝑟— − 𝑟I + 𝑒• quadratic relation (Treynor Mazuy)

Carhart: 𝑅M+ − 𝑅I+ = 𝛼M + 𝛽 5 𝑅5+ − 𝑅I+ + 𝛽 Ë5” 𝑅Ë5$'',+ − 𝑅”HŒ,+ + 𝛽 75' 𝑅7HŒ7,+ − 𝑅'6À,+

Downloaded by Denis Dumea ([email protected])