lOMoARcPSD|9327251

CFA level II formula sheet

Finance (Harvard University)

StuDocu is not sponsored or endorsed by any college or university

Downloaded by Smit Gandhi ([email protected])
 lOMoARcPSD|9327251

Formula Sheet Level II 2020

FinQuiz Formula Sheet CFA Program Level II

(for single independent variable R2 = r2) 𝑆𝑆𝑇

d
Reading 4: Introduction to Regression
= `(𝑦c
Total
7. SST = SSE + SSR(or RSS) df = n-1
ce5
∑1 / /
-23(,- .,)(0- .0)
− 𝑦U)B
1. Sample Cov (X, Y) =
4.5 8. Hypothesis Testing:
• Null and Alternative hypotheses
789:;
2. Correlation Coefficient = r,0 = (< or Source of Sum of Mean Sum
: )(<; )
• H0: b1 = 0 (no linear relationship) Variability
DoF
Squares of Squares
789(,,0)
r= • H1: b1 ≠ 0 (linear relationship does
=9>?(,)=9>?(0)
exist) Regression MSR =
! (Explained)
1 RSS
RSS/1
3. t-test (for normally distributed variables) = b1 - b1
• Test statistic = t =
?√4.B
t= t distribution with (n −
s b1 Error
n-2 SSE
MSE =
=5.? C (Unexplained) SSE/n-2
2) deg. of freedom • Confidence Interval = b1 ± t c s b1
SST=
Total n-1 RSS +
4. Linear Regression = Yi = b0 + b1Xi + εi, 9. ANOVA (Analysis of variance) = SSE
ANOVA SS MSS F
• Intercept (b0) = b0 = y - b1 x =
pqq
𝑆𝑆𝑅 lmn ( )
• Slope or regression coefficient = b5 = d 𝑆𝑆𝑅g 10. F-Statistic or F-Test = = r
qqs
𝑆𝑆𝑅 lmo ( )
789(S,T) ∑(S.SU)(T.T
/) 𝑘 turu3
= `(𝑦bc
Regression
or = 𝑘 𝑆𝑆𝐸g
9>?(S) ∑(S.SU)C (𝑛 − 𝑘 − 1)
df = k (df numerator = k = 1)
ce5
− 𝑦U)B (df denominator = n – k – 1 = n – 2)
5. Standard Error of Estimate SEE = SW = 𝑆𝑆𝐸
∑1 [ )C
d
𝑆𝑆𝐸
v ± t 7 sx
11. Prediction Intervals = Y
YYW -23(T- .T
X =X = `(𝑦c
Error
𝑛−𝑘−1 5 / )C
(,.,
4.Z.5 4.Z.5
𝑤ℎ𝑒𝑟𝑒 sxB = sB }1 + 4 + (4.5)<C • and
df = n-k-1
ce5
− 𝑦b)
B :

6. Coefficient of Determination (R2) = 2

YY\.YYW ]YY
sf = s
= =
f
where, 0 ≤ R2≤ 1
YY\ YY\

www.finquiz.com All rights reserved

Downloaded by Smit Gandhi ([email protected])

 lOMoARcPSD|9327251

Formula Sheet Level II 2020

Reading 5: Multiple Regression & Issues in 2. Log-Linear Trend Models = yt = e

b0 +b1t 8. Smoothing Past Values with n-Period
Regression Analysis Moving Average =

3. Autoregressive Time-Series Models: xt + xt -1 + xt -2 + ..... + xt -( n -1)

1. Yi = b0 + b1X1i + b2X2i + … + bkXki + εi,i = • First order autoregressive AR (1) = xt n
1, 2, … n   = b0 + b1 x t-1 + εt 9. Correcting Seasonality in Time Series

Prediction equation = 𝑌•c = 𝑏•ƒ + 𝑏•5 𝑋5c +

• pth-order autoregressive AR (p) = xt Models:
2.
𝑏•B 𝑋Bc +. . . +𝑏•… 𝑋…c + ε‡ , 𝑖
= b0 + b1 x t-1 + b2 x t-2 + …..+ bp x t-p
+εt •For quarterly data = xt = b0 + b1x t-1 +
b0 b2x t-4 + εt
Adjusted R2 = 𝑅UB = 1 − ‰d.….5Š (1 − 𝑅B )
d.5
3. 4. Mean reverting level of xt = • For monthly data = xt = b0 + b1x t-1 +
1 - b1 b2x t-12 + εt
4. Breusch–Pagan test 10. ARCH model =
5. Chain Rule of Forecasting:
• H0 = No conditional eˆ 2 t = a 0 + a1eˆ 2 t -1 + µ t where µt is
Heteroskedasticity exists • One-period ahead forecast =
an error term
• HA = Conditional Heteroskedasticity x̂t+1 = bˆ0 + bˆ1 xt • Predicting variance of errors in period
exists • Two-period ahead forecast= 2 2
• Test statistic = n × R2residuals t+1 = σˆ t+1 = α̂ 0 + α1εˆt
x̂t+2 = bˆ0 + bˆ1 xt+1
5. Durbin-Waston Test = 𝐷𝑊 = 6. Random Walks and Unit Roots: Reading 7: Machine Learning
∑• b Ž .•bŽu3)C
Ž2C(• • Random Walk without drift = xt = x t-
∑Ž23 •bŽ C
•
1 + εt where, b0 = 0 and b1 = 1. LASSO:
1. Penalty term (when l > 0) = 𝜆 ∑’…e5‘𝑏U… ‘
• For Large Sample size DW Statistic • Correcting Random Walk = yt = xt - x
(d) = d ≈ 2 (1 – r) t-1

• Random walk with a drift = xt = b0 + x

2. ∑dce5(𝑌c − 𝑌c )B + 𝜆 ∑’…e5‘𝑏•’ ‘
t-1 +
εt where, b0 ≠ 0 and b1 = 1
• By taking first difference yt = xt - x t-1
3. When l = 0, LASSO penalized regression
= b0 + εt
= OLS regression
Reading 6: Time Series Analysis
7. Using Dickey-Fuller Test = xt - x t-1 = b0 +
Reading 8: Big Data Projects
1. Linear Trend Models = yt = b0 + b1t+ εt (b1 -1) x t-1 + εt
• Predicted/fitted value of yt in period ›œ .›•œt
1. 𝑋c(d“”•–—c˜™š) = ›
•žŸ .›•œt
(T + 1) = yˆ t +1 = bˆ0 + bˆ1 (T + 1)
where Xi = value of observation

www.finquiz.com All rights reserved

Downloaded by Smit Gandhi ([email protected])

 lOMoARcPSD|9327251

Formula Sheet Level II 2020

5
Performance Metrics: • (1 + i° ) = S∫ ¹1 + i∫ º » ½ 7. Forward discount or premium as % of spot
g ®∫ /¼
2. Accuracy = (TP + TN)/(TP + FP + TN + ° rate:
5¾‡∫
FN) • F∫ /° = S∫ ‰
(5¾‡¼ )
Š Ff / d - S f / d
g
F1 score = (2*P*R)/(P + R) ° @ (i f - id )
• Using day count convention: S f /d
3. Receiver Operating Characteristic (ROC): ' ! Actual $* If uncovered interest rate parity holds
)1+ id # ,=
False positive rate (FPR) = FP/(TN + FP)
( " 360 &%+ • Ff /d − S f /d
and = = %ΔS ef /d ≅ (i f − id )
S f /d
True positive rate (TPR) = TP/(TP + FN), ' ! Actual $*' 1 *
S f /d )1+ i f # ,) ,
which is same as recall
( " 360 &%+)( Ff /d ,+
8. Purchasing Power parity (PPP)
4. Root Mean Square Error (RMSE): • Pf = S f/d × Pd
d
(𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑c − 𝐴𝑐𝑡𝑢𝑎𝑙c )B æ é Actual ù ö • S f/d = Pf / Pd
` ç1+ i f ê
360 úû ÷
÷
𝑛 • ç ë
ce5 Ff / d = S f /d
ç é Actual ù ÷ 9. Relative version of PPP = %∆S f/d = πf – πd
ç 1 + id ê ÷
Reading 9: Excerpt from ‘Probabilistic è ë 360 úû ø
10. Ex ante version of PPP = %∆Sef/d = πef –
Approaches, Scenario Analysis, Decision Tree πed
& Simulations’ 6. Uncovered Interest Rate Parity :
i f - %DS e f / d = id 11. Real Exchange Rate
•
%DS e
= i f - id æ S f / d Pd ö æ ö
Reading 10: Currency Exchange Rates •
f /d
ç ÷ = S f / d ç Pd ÷
ç P ÷ çP ÷
qf/d = è ø è f ø
• Forward premium or discount: f

1. Bid-offer Spread = Offer price – Bid price • For one year horizon =
æ CPI d ö
2. Fwd rate = Spot Exchange rate + Ff /d − S f /d = qf /d = S f /d ç ÷
®8?¯>?° ±8‡4²< ç CPI ÷
5ƒ,ƒƒƒ "i −i % or è f ø
3. Forward premium/discount (in %) =
S f /d $ f d ' ≅ S f /d (i f − id ) 12. Fisher effect:
<±8² µS7¶>4·µ ?>²µ.(x8?¯>?° ±8‡4²</5ƒ,ƒƒƒ)
# 1+ id &
−1 • id = rd + πεd
<±8² µS7¶>4·µ ?>²µ • Using day count convention: • if = rf + πεf
( " Actual % +
* $ - • if – id = (rf – rd) + (πεf- πεd)
4. To convert spot rate into forward quote: # 360 '& - • (rf – rd) = (if – id) - (πεf- πεd)
• Spot exchange rate × (1 + % premium) Ff /d − S f /d = S f /d * (i f − id )
* 1+ i " Actual % -
* d$ -
• Spot exchange rate × (1 - % discount) ) # 360 '& ,

5. Covered interest rate parity:

www.finquiz.com All rights reserved

Downloaded by Smit Gandhi ([email protected])

 lOMoARcPSD|9327251

Formula Sheet Level II 2020

Reading 11: Economic Growth & The • Slope of straight line = [δ + n + θ / (1

Investment Decision 9. Contribution of Improvement in – α)]
technology = Labor productivity growth 14. During the transition to the steady state
1. Economic growth = Annual % ∆ in real rate – Capital Deepening growth path:
GDP or in real per capita GDP • Growth rates of output per capita = ∆y
Æ Ç
10. Growth Accounting based on Solow / y = ɉ Š + 𝛼𝑠 ‰’ − 𝛹ŠË =
5. Å
W Á
P = GDP ‰ Š ‰WŠ
Approach = ∆Y /Y = ∆A / A + α ∆K/K +
Æ
2.
¿ÀÁ (1 – α) ∆L/ L ‰5.ÅŠ + 𝛼𝑠 (y/k – Ψ)

3. Expressing in terms of logarithmic rates:

• Capital-to-labor ratio = ∆k / k =
Æ Ç Æ
ɉ Š + 𝑠 ‰’ − 𝛹ŠË = ‰5.ÅŠ + s
11. Labor productivity growth accounting
• (1/T) % ∆P = (1/T) % ∆GDP + (1/T) equation 5. Å
%∆ (E / GDP) + (1/T) % ∆(P / E) • Growth rate in potential GDP = LT g (y/k – Ψ)
• % ∆ in stock MV = % ∆ in GDP + % rate of labor force + LT g rate in labor
∆ in share of earnings (profit) in GDP productivity 15. Proportional impact of the saving rate
+ % ∆ in the P/E multiple change on the capital-to-labor ratio and per
12. Balanced or Steady State Rate of Growth capita income over time:
1
4. A two-factor aggregate production in Neoclassical Growth Theory:
' ! Y $ *α −1
function: Y = AF (K, L) • Growth in physical capital stock = ∆K
)# & ,
= sY – δK knew " K %new ,
=)
5. Cobb-Douglas Production Function = F • kold )!Y $ ,
(K, L) = Kα L1 - α 13. In the steady state: )( #" K &%old ,+
• Growth rate of capital per worker = ∆k
a
6. Under Cobb-Douglas production function: / k = ∆y / y = ∆A / A + α ∆k / k =
ÂÃÄ
y new é k new ù
• Marginal product of capital = MPK =
5. Å
è Steady state growth rate of =ê ú
α AK α-1 L 1-α = α Y/K •
yold ë k old û
labor productivity
• α Y/K = r èα = r (K) / Y = Capital
• Growth rate of Total output = ∆Y / Y
income / Output or GDP 16. Production function in the endogenous
= Growth rate of TFP scaled by labor
force share + Growth rate in the labor growth model = ye = f (ke) = cke
7. Output per worker or Average labor Æ • Growth rate of output per capita =
force = +n
productivity (Y/L or y): 5. Å ∆ye/ye = ∆ke/ke = sc – δ – n
• GDP/Labor input = TFP × capital-to- • Steady state Output-to-capital ratio =
Ç 5 Æ
labor ratio × share of capital in GDP = ‰È Š ɉ5. ÅŠ + 𝛿 + 𝑛Ë = 𝛹
’ Reading 12: Economics of Regulation
• Or y = Y/L = Akα • Gross investment per worker =
8. Contribution of Capital Deepening = Labor Æ
ɉ Š + 𝛿 + 𝑛Ë 𝑘
productivity growth rate – TFP 5. Å

www.finquiz.com All rights reserved

Downloaded by Smit Gandhi ([email protected])