Professional Documents
Culture Documents
Formula Sheet
Formula Sheet
FORMULA SHEET
𝜎𝑖 𝐶𝑜𝑣(𝑖, 𝑚) 𝜎𝑖𝑚
𝛽𝑖 = 𝜌(𝑖𝑚) = 2 = 2
𝜎𝑚 𝑖 𝜎𝑚 𝜎𝑚
𝑅𝑚 = expected market rate of return
CAPM formula
(expected return on 𝐸(𝑅𝑖 ) = 𝑅𝑓 + 𝛽𝑖 (𝑅𝑚 − 𝑅𝑓 ) 𝑅𝑓 = risk-free rate
asset i)
𝐸(𝑅𝑖 ) = expected return on asset i
𝐶𝑜𝑣(𝑖, 𝑚) 𝜎𝑖𝑚
𝜌𝑖𝑚 = =
𝜎𝑖 𝜎𝑚 𝜎𝑖 𝜎𝑚
Page | 1
𝛽𝑃 = portfolio beta
𝐸(𝑅𝑝 ) − 𝑅𝑓
Treynor ratio Treynor ratio = 𝐸(𝑅𝑝 ) = portfolio expected return
𝛽𝑃
𝑅𝑓 = risk-free rate
Page | 2
Variance/Covariance
𝑀2 − 𝑀
for a factor model with 𝑀+ 𝑀 = number of factors in the model
𝑀
M factors
Number of covariances 𝑛2 − 𝑛
𝑛 = number of variances
required 2
The Fama-French 𝐸(𝑅𝑖 ) = 𝑅𝑓 + 𝛽𝑖,𝑀𝐾𝑇 𝐸(𝑅𝑚 − 𝑅𝑓 ) + 𝛽𝑖,𝑆𝑀𝐵 𝐸(𝑆𝑀𝐵) 𝛽𝑖,𝑆𝑀𝐵 = factor-beta for the size factor
Model (FFM) + 𝛽𝑖,𝐻𝑀𝐿 𝐸(𝐻𝑀𝐿)
𝐻𝑀𝐿 = value factor
𝛽𝑖,𝐻𝑀𝐿 = factor-beta for the value factor
𝐸(𝑅𝑚 − 𝑅𝑓 ) = CAPM market factor
𝛽𝑖,𝑀𝐾𝑇 = factor-beta for the market-factor
Page | 3
Conditional 𝑃(𝐴)𝑃(𝐵)
𝑃(𝐴|𝐵) = = 𝑃(𝐴)
Probability 𝑃(𝐵)
𝑃(𝐵|𝐴)𝑃(𝐴)
Bayes’ Theorem 𝑃(𝐴|𝐵) =
𝑃(𝐵)
𝑛!
𝑃(𝑋 = 𝑥) = 𝑝 𝑥 (1 − 𝑝)𝑛−𝑥 , 𝑥 = 0,1,2, … , 𝑛 𝑃(𝑋 = 𝑥) = probability mass function of X
𝑥! (𝑛 − 𝑥)!
0, 𝑦<0
Bernoulli distribution 𝐹𝑋 (𝑥) = {1 − 𝑝, 0 ≤ 𝑦 < 1 𝐹𝑋 (𝑥) = cumulative distribution function
1, 𝑦≥1
Page | 4
1
𝑃(𝑋 = 𝑥) = , 𝑥 = 1,2, … , 𝑛 𝑃(𝑋 = 𝑥) = probability mass function
𝑛
𝑛+1
Uniform distribution 𝐸(𝑋) = 𝐸(𝑋) = expectation of X
2
𝑛2 − 1
𝑉(𝑋) = 𝑉(𝑋) = variance of X
12
𝜆𝑥 𝑒 −𝜆
𝑃(𝑋 = 𝑥) = , 𝑥 = 0,1,2, … ; 𝜆 > 0 𝑃(𝑋 = 𝑥) = probability mass function
𝑥!
|𝑥|
Poisson distribution 𝜆𝑖
𝐹𝑋 (𝑥) = 𝑒 −𝜆
∑ 𝐹𝑋 (𝑥) = cumulative distribution function of X
𝑖!
𝑖=1
1
𝑓𝑋 (𝑥) = ,𝑎 ≤ 𝑋 ≤ 𝑏 𝑓(𝑥) = probability density function of X
𝑏−𝑎
0, 𝑥<𝑎
𝑥−𝑎
𝐹𝑋 (𝑥) = { ,𝑎 ≤ 𝑥 ≤ 𝑏 𝐹𝑋 (𝑥) = cumulative distribution function
𝑏−𝑎
Continuous uniform 1, 𝑥≥𝑏
distribution
𝑎+𝑏
𝐸[𝑋] = 𝐸(𝑋) = Expectation of X
2
(𝑏 − 𝑎)2
𝑉[𝑋] = 𝑉(𝑋) = Variance of X
12
Page | 5
1 1 𝑥−𝜇 2
− ( )
𝑓(𝑥) = 𝑒 2 𝜎 , −∞ <𝑋<∞ 𝑓(𝑥) = probability density function of x
𝜎√2𝜋
1 1 𝑙𝑛𝑥−𝜇 2
− ( )
𝑓(𝑥) = 𝑒 2 𝜎 𝑓(𝑥) = probability density function of x
𝑥𝜎√2𝜋
2 2
𝑉(𝑋) = (𝑒 𝜎 − 1)𝑒 2𝜇+𝜎
Log-normal distribution 1
𝐸[𝑋] = 𝑒 𝜇+2𝜎2
𝑙𝑛𝑋 − 𝜇
𝐹𝑋 (𝑥) = Φ ( )
𝜎
∞
1 𝑘 𝑥
𝑓(𝑥) = 𝑥 2 −1 𝑒 −2 , 𝑥>0 𝛤(𝑛) = ∫ 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥
𝑘
𝑘 0
22 𝛤 ( )
2
Chi-square distribution 𝑘
𝐸(𝑆) = 𝑘
𝑆 = ∑ 𝑍𝑖2
1=1
𝑉(𝑆) = 2𝑘
Page | 6
𝑎 Γ(𝑎)Γ(𝑏)
Beta distribution 𝐸(𝑋) = 𝐵(𝑎, 𝑏) =
𝑎+𝑏 Γ(𝑎 + 𝑏)
𝑎𝑏
𝑉(𝑋) =
(𝑎 + 𝑏)2(𝑎 + 𝑏 + 1)
1 −𝛽𝑥
𝑓𝑋 (𝑥) = 𝑒 ,𝑥 ≥ 0
𝛽
Exponential distribution 𝑥
−
𝐹𝑋 (𝑥) = 1 − 𝑒 𝛽
𝐸(𝑋) = 𝛽
𝑉(𝑋) = 𝛽 2
Page | 7
Marginal distribution
𝑓𝑋2 (𝑥2 ) = ∑ 𝑓𝑋 1 ,𝑋2 (𝑥1 , 𝑥2 ) 𝑓𝑋2 (𝑥2 ) = marginal distribution of 𝑋2
𝑥1 𝜖𝑅(𝑋1 )
Multivariate
Random
Variables
𝑓𝑋 1 (𝑥1 )𝑓𝑋2 (𝑥2 ) 𝑋1 and 𝑋2 are independent
Conditional distribution 𝑓𝑋 1 │𝑋2 (𝑥1 │𝑋2 = 𝑥2 ) =
𝑓𝑋2 (𝑥2 ) 𝑓𝑋 1 ,𝑋2 (𝑥1 , 𝑥2 ) = 𝑓𝑋 1 (𝑥1 )𝑓𝑋2 (𝑥2 )
Page | 8
Conditional 𝐸(𝑋1 |𝑋2 = 𝑥2 )= ∑ 𝑥1𝑖 𝑓(𝑋1 │𝑋2 = 𝑥2 ) 𝐸(𝑋1 |𝑋2 = 𝑥2 ) = conditional expectation of 𝑋1
expectations 𝑖
given 𝑋2 = 𝑥2
𝑉𝑎𝑟(𝑋1 |𝑋2 = 𝑥2 ) = 𝐸(𝑋12 |𝑋2 = 𝑥2 ) − [𝐸(𝑋1 |𝑋2 = 𝑥2 )]2 𝑉𝑎𝑟(𝑋1 |𝑋2 = 𝑥2 ) = conditional variance of 𝑋1
Conditional variance
given 𝑋2 = 𝑥2
Page | 9
𝜇̂ − 𝜇 𝜇 = population mean.
𝜎 → 𝑁(0,1)
The central limit
𝜇̂ = population mean estimator
theorem (CLT)
√𝑛 𝜎 = the population standard deviation
Page | 10
Test statistic
The test statistic (Sample statistic − Hypothesized value)
=
Standard error of the sample statistic
Page | 11
𝑅 2 = coefficient of variation
Measures of goodness 𝑟 = correlation coefficient between X and Y
𝑅2 = 𝑟2
of fit 𝐶𝑜𝑣(𝑋, 𝑌)
𝑟=
𝜎𝑋 𝜎𝑌
Regression with
Multiple Method for regression 𝑛 𝑛 𝑛 ∑𝑛𝑖=1(𝑌𝑖 − 𝑌̂) =2 explained sum of squares (ESS)
Explanatory with one or more than 2 2 2
∑(𝑌𝑖 − 𝑌̅)2 = ∑(𝑌̂𝑖 − 𝑌̅) + ∑(𝑌𝑖 − 𝑌̂) ∑𝑛𝑖=1(𝑌̂𝑖 − 𝑌̅) = residual sum of squares (RSS)
Variables one independent
𝑖=1 𝑖=1 𝑖=1
variables ∑𝑛𝑖=1(𝑌𝑖 − 𝑌̅)2 = total sum of squares (TSS)
Page | 12
𝐸𝑆𝑆
( ) 𝑛 = number of observations
𝐹= 𝑘
F-statistics 𝑆𝑆𝑅
( ) 𝑘 = number of independent variables.
𝑛 − (𝑘 + 1)
Regression
Diagnostics Cariance inflation factor 1 𝑅𝑗2 comes from a regression of Xj on the other
𝑉𝐼𝐹𝑗 =
for the variable 𝑋𝑗 1 − 𝑅𝑗2 variables in the model
−𝑗
𝑌̅𝑖 = fitted value of 𝑌̅𝑖 when the observed value
−𝑗 2
∑𝑛𝑖=1 (𝑌̅𝑖 − 𝑌̂𝑖 ) j is excluded
Cook's distance 𝐷𝑗 = 𝑘 = number of coefficients in the regression
𝑘𝑠 2
model
𝑠 2 = estimated error variance
𝑡 =time
𝛾𝑡,𝜏 = 𝐸[(𝑌𝑡 − 𝐸(𝑌𝑡 ))(𝑌𝑡−𝜏 − 𝐸(𝑌𝑡−𝜏 ))]
Autocovariance function 𝑟 = lag
Stationary 𝛾𝑡,𝜏 = autocovariance
Time Series
Autocorrelation 𝐶𝑜𝑣(𝑌𝑡 , 𝑌𝑡−𝜏 ) 𝛾𝜏 𝛾𝜏
𝜌(𝑡) = = =
Function (ACF) √𝑉(𝑌𝑡 ) √𝑉(𝑌𝑡−𝜏 ) √𝛾0 𝛾0 𝛾0
Page | 13
𝛽 = AR parameter
𝛼 = the intercept
𝑌𝑡 = 𝛼 + 𝛽𝑌𝑡−1 + 𝜖𝑡 𝜖𝑡 = the shock.
𝛼
The 1st order 𝐸(𝑌𝑡 ) = 𝜇 = 𝐸(𝑌𝑡 ) = expectation of 𝑌𝑡
Autoregressive 1−𝛽
(AR 1) Model 𝜎2
𝑉(𝑌𝑡 ) = 𝛾0 = 𝑉(𝑌𝑡 ) = variance of 𝑌𝑡
1 − 𝛽2
Page | 14
𝐽𝐵~𝜒22
2
𝑆̂ 2 (𝑘̂ − 3) 𝑇 = sample size
Jarque-Bera test 𝐽𝐵 = (𝑇 − 1) ( + )
6 24 𝑆 = skewness
𝑘 = kurtosis
Page | 15
𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜 𝑣𝑎𝑙𝑢𝑒
The optimal number of 𝑁 = 𝛽𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜 × [ ]
𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑓𝑢𝑡𝑢𝑟𝑒𝑠 𝑐𝑜𝑛𝑡𝑟𝑎𝑐𝑡
futures contracts needed
to hedge an exposure 𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜 𝑣𝑎𝑙𝑢𝑒
= 𝛽𝑝𝑜𝑟𝑡𝑓𝑜𝑙𝑖𝑜 × [ ]
𝑓𝑢𝑡𝑢𝑟𝑒𝑠 𝑝𝑟𝑖𝑐𝑒 × 𝑐𝑜𝑛𝑡𝑟𝑎𝑐𝑡 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑖𝑒𝑟
𝛽 = portfolio beta
Adjusting a stock 𝑃
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑜𝑛𝑡𝑟𝑎𝑐𝑡𝑠 𝑟𝑒𝑞𝑢𝑖𝑟𝑒𝑑 = (𝛽 ∗ − 𝛽)( ) 𝛽 ∗ = target beta after hedging
portfolio’s beta using 𝐴
stock index futures P = portfolio value
𝐴 = value of the underlying
Page | 16
𝑌 = convenience yield,
𝑇 = time to maturity
1 𝐹 = forward price
Convenience yield, Y 𝑆+𝑈 𝑇
𝑌=( ) (1 + 𝑅) − 1 𝑅 = interest rate
𝐹
1+𝑅 𝑇
𝐹 = (𝑆 + 𝑈) × ( )
1+𝑌
𝑈 = present value of storage cost
𝐹 = forward price
Commodity 1+𝑅
Cost of carry, 𝐶 = 1+𝑄 − 1 ≈ 𝑅 − 𝑄 𝑅 = risk-free interest rate per year compounded
Cost of carry, C
Forwards and annually
Futures
Q = yield
Page | 17
non-dividend paying
stocks 𝑐 ≥ max(S0 − Ke−rT , 0) 𝐾 = strike price
Lower pricing bound for
European calls on: 𝑆0 = value of the stock today.
dividend-paying stocks 𝑆𝑇 = Price of the stock at time T
𝑐 ≥ S0 − Ke−rT − 𝐷𝑒 −𝑟𝑇 𝑐 = value of a European call option
𝐶 = value of an American call option
Lower pricing bound for 𝑝 = value of a European put option
European puts on non- 𝑝 ≥ max(Ke−rT − 𝑆0 , 0) 𝑃 = value of an American put option
dividend paying stocks
Properties of
Options
𝑛 = time in years
𝑚×𝑛
𝑅
Future value 𝐹𝑉 = 𝐴 (1 + ) 𝑚 = compounding frequency per year
𝑚
Properties of 𝑅 = interest rate compounded m times per year
Interest Rates
Future value for 𝐹𝑉 = future value
continuous 𝐹𝑉 = 𝐴𝑒 𝑅×𝑛
compounding, 𝑅 = continuously compounded rate of interest
Page | 18
Macaulay Duration for a 𝑛 𝑐𝑖 𝑒 −𝑦𝑡𝑖 𝑡𝑖 = time in years until cash flow 𝑐𝑖 is received
𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛 = ∑ 𝑡𝑖 [ ]
zero-coupon bond 𝑖=1 𝑃 𝑦 = continuously compounded yield on a bond 𝑃
𝑀𝑎𝑐𝑎𝑢𝑙𝑎𝑦 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛
Modified duration 𝑀𝑜𝑑𝑖𝑓𝑖𝑒𝑑 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛 = 𝐷 ∗ = 𝑦
(1 + 𝑛)
Page | 19
Swaps To the pay fixed, 𝑉𝑠𝑤𝑎𝑝 = value of the swap 𝑃𝑓𝑖𝑥 = PV of fixed-leg
receive floating 𝑉𝑠𝑤𝑎𝑝 = 𝑃𝑓𝑙𝑡 − 𝑃𝑓𝑖𝑥 payment
𝑃𝑓𝑙𝑡 =PV of the floating leg payment.
To the pay floating,
𝑉𝑠𝑤𝑎𝑝 = 𝑃𝑓𝑖𝑥 − 𝑃𝑓𝑙𝑡
receive fixed
Book 4 Measures of
Valuation and Financial Risk Volatility of the return,
𝜎𝑅 = √𝐸(𝑅 2 ) − [𝐸(𝑅)]2 𝜎𝑅 = Volatility of the return, R
Risk Models R
Page | 20
𝑋 = confidence level
Calculating and
Applying VaR
Page | 21
𝑟𝑖 = return on a day i
𝑆𝑖 −𝑆𝑖−1
𝑟𝑖 = 𝑆𝑖−1
𝑚
Simplified formula for 1 𝜎𝑛 = current volatility
2
current volatility 𝜎𝑛 = √ ∑ 𝑟𝑛−𝑖
𝑚 2
𝑟𝑛−1 = squared return on day n-1
𝑖=1
Page | 22
Page | 23
2
𝜎𝑛+𝑡 = expected variance rate on day, t
The expected variance 2
𝜎𝑛+𝑡 = 𝑉𝐿 + (𝛼 + 𝛽)𝑡 (𝜎𝑛2 − 𝑉𝐿 ) 𝜎𝑛2 = variance rate on day n
rate on day t
𝑉𝐿 = long-run variance
Portfolio expected loss 𝐸𝐿𝑃 = ∑𝐸𝐴𝑖 × 𝑃𝐷𝑖 × 𝐿𝐺𝐷𝑖 𝐸𝐿𝑃 = expected portfolio loss
𝑈𝐿 = unexpected loss
Measuring
Credit Risk Unexpected loss 2
𝑈𝐿 = 𝐸𝐴 × √𝑃𝐷 × 𝜎𝐿𝑅 2
+ 𝐿𝑅 2 × 𝜎𝑃𝐷 𝐿𝑅 = loss rate
2
𝜎𝑃𝐷 = 𝑃𝐷 × (1 − 𝑃𝐷)
Portfolio unexpected 2
𝑈𝐿𝑃 = 𝐸𝐴 × √𝑈𝐿2𝑖 × 𝑈𝐿2𝑖 + 2𝜌𝑈𝐿𝑖 𝑈𝐿𝑗 𝑈𝐿𝑖 = 𝐸𝐴 × √𝑃𝐷𝑖 × 𝜎𝐿𝑅 + 𝐿𝑅𝑖2 × 𝜎𝑃𝐷
2
loss 𝑖 𝑖
𝑛 = number of loans
Standard deviation
expressed as the 𝜎𝑃 𝜎√1 + (𝑛 − 1)𝜌 𝐿𝑖 = amount borrowed in the ith loan
percentage of the size of 𝛼= =
𝑛𝐿 𝐿√𝑛 𝜌 = correlation between losses on the ith and jth
the portfolio
loan
Page | 25
𝜌𝑈𝑖𝑈𝑗 = 𝑎𝑖 𝑎𝑗 = 𝑎. 𝑎 = 𝑎2 𝑎𝑖 = 𝑎
𝑎𝑠 Δ𝑋𝑖 → 0 Δ𝐹𝑖
𝑛 𝑄𝑖 = 𝑋𝑖
Δ𝑋𝑖
Euler’s Theorem 𝐹 = ∑ 𝑄𝑖
Δ𝑋𝑖 = small change in 𝑋𝑖
𝑖=1
Δ𝐹𝑖 = small change 𝐹𝑖
Operational
Risk 𝑋 = all losses
Standardized
Measurement Approach = 𝑌 = losses greater than EUR 10 million
(SMA) 𝐿𝑜𝑠𝑠 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡 = 7𝑋 + 7𝑌 + 5𝑍
𝑍 𝑍 = losses greater than EUR 100 million
Page | 26
𝜎 2
𝑉𝑎𝑟𝑖𝑎𝑛𝑐𝑒 = ln(1 + 𝑤) 𝑤=( )
𝜇
𝑣 = random variable
𝑥 = random variable greater than 𝑣
Power law Pr(𝑣 > 𝑥) = 𝑘𝑥 −𝛼
𝑘 = scale factor
𝛼 = a power
Page | 27
1
𝑓𝑎𝑐𝑒 𝑣𝑎𝑙𝑢𝑒 𝑦𝑒𝑎𝑟 𝑡𝑜 𝑚𝑎𝑡𝑢𝑟𝑖𝑡𝑦
YTM of a zero-coupon 𝑌𝑇𝑀 = ( )
bond 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑝𝑟𝑖𝑐𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑏𝑜𝑛𝑑
−1
Page | 28
Page | 29
𝑆 𝜎2
ln ( 0 ) + [𝑟 − 𝑟𝑓 + ( 2 )] 𝑇
𝐾
𝑑1 =
𝜎√𝑇
Options on foreign 𝐶0 = 𝑆0 𝑒 −𝑟𝑓𝑇 × 𝑁(𝑑1 )– 𝐾𝑒 −𝑟𝑇 × 𝑁(𝑑2 )
currency 𝑑2 = 𝑑1 − 𝜎√𝑇
𝑃0 = 𝐾𝑒 −𝑟𝑇 × 𝑁(−𝑑2 )– 𝑆0 𝑒 −𝑟𝑓𝑇 × 𝑁(−𝑑1 )
𝑟𝑓 = foreign risk-free rate
r= domestic risk-free rate
Page | 30
Page | 31