Sie sind auf Seite 1von 9

Test to perform

JB Test: Test for normality


H0: Data ≈ N ; H1: Data ≠ N

JB
N =Anzahl(Data)
Skewness =SCHIEFE(Data)
Kurtosis =KURT(Data)
JB Test = n * (Skew^2/6 + Kurt^2/24)
Critical value =CHI.SQU(Confidence Intervall; 2)

If JB > critical value reject H0

F Test: Test for multiple Beta analysis


H0: beta1 = beta2 (= 0); H1: beta1 ≠ beta2 (≠0)

ANOVA
Freiheitsgrade Quadratsummen Mittlere Quadratsumme Prüfgröße
(df) (SS) (MS) (F) F krit
Regression 5,00 64,39 12,88 79,77 0,00
Residue 520,00 83,94 0,16
Gesamt 525,00 148,33

Formula: F = ((RSSre – RSSunre)/(DFunre – Dfre)) / MSunre

Critical Value: CHI.SQU(Confidence; # of restrictions [# of = signs])

If F > critical value reject H0

T-Test, P-Test & Intervall: Test for statistical relevance of coefficients (beta)

Koeffizienten Standardfehler t-Statistik P-Wert Untere 95% Obere 95% Untere 95,0% Obere 95,0%
Schnittpunkt 0,41 0,10 4,17 0,00 0,22 0,61 0,22 0,61
educ 0,08 0,01 11,93 0,00 0,07 0,09 0,07 0,09

T-Test: H0: beta = 0 ; H1: beta ≠0

Critical Value: z(a/2); z(1-a/2) . =NORM.INV(Confidence level;0;1)

If T-value > / < “-“ reject H0


P-Test: H0: beta = 0 ; H1: beta ≠0

Critical Value: alpha

If p-value > alpha reject H0

Interval: If zero is NOT within the upper and lower boundary, reject H0

How to analyze regression results:

Regressions-Statistik
Multipler Korrelationskoeffizient 0,41
Bestimmtheitsmaß (R^2) 0,16
Adjustiertes Bestimmtheitsmaß (adj. R^2) 0,16
Standardfehler (std. er) 3,38
Beobachtungen (observations) 526,00

Multiple R.: Tells you how strong the linear relationship is. For example, a value of 1 means a perfect
positive relationship and a value of zero means no relationship at all. It is the square root of r
squared

R^2: „How much of my value can be described by my regression“

Adj. R^2: Same as R^2 just adjusted for the impact of adding new variables “Cut out the bullshit”

Std. Er.: the precision that the regression coefficient is measured; if the coefficient is large compared
to the standard error, then the coefficient is probably different from 0.

Frequently asked questions:

Koeffizienten Standardfehler t-Statistik P-Wert Untere 95% Obere 95% Untere 95,0% Obere 95,0%
Schnittpunkt 0,41 0,10 4,17 0,00 0,22 0,61 0,22 0,61
educ 0,08 0,01 11,93 0,00 0,07 0,09 0,07 0,09

“How much does y change for 1 unit of x?”  Delta is represented by beta
“What value for y when x=4?”  Coefficient Intercept + Coefficient Variable (BETA) * 4

“Is Beta significant at the 90%, 95%, 99% confidence level?”  T or p test look above

“Does X (represented by dummy) has an influence”  Check for significance of variable, then state
the coefficient ALTERNATIVE do the F-Test
How to calculate Portfolios and how to plot them

1) Create weights with ZUFALLSZAHL() or ZUFALLSBEREICH() für w1 und w2 und 1 – w1 – w2 für w3

2) constraint weights = 1

3) Expected return: SUMMENPRODUKT(weigths, returns)  if both are represented in lines

=MMULT(weights [as line]; returns [as column])

4) Volatility: =WURZEL(MMULT(weights [as line]; MMULT(Cov Matrix; MTRANS(weights


[as line]))))

Graf:

X-Axis  volatility;

Y-Axis  Return

Maybe it the Pseudo Sharpe Ratio is asked  Volatility / Return

Solve for optimal Portfolios:


1) Copy paste portfolio with formulae

a) Max return (i) changeable items weights (ii) Control sum [SUMME(weights)] is equal to 1 (iii)
Volatility is equal to target (iiii)

b) Min volatility (i) changeable items weights (ii) Control sum [SUMME(weights)] is equal to 1 (iii)
Return is equal to target (iiii)
MVP: Min volatility (i) changeable items weights (ii) Control sum [SUMME(weights)] is equal to 1 (iii)
 DOES NOT WORK FOR MAX RETURN

Efficient Frontier:

Efficient Frontier weights (ALL THE SAME FORMULA): MMULT(Lambda; weights high return [as line])
+ MMULT(1-Lambda; weights low return [as line])

Efficient Frontier with Risk Free Asset and Tangent Portfolio:

1) Enter random variables for weights

2) Copy formulae for Volatility and Return


3) Sharpe’s Ratio: (Expected Return – Return Risk free) / Volatility

3) Use solver max for Sharpe’s Ratio (i) change weights (ii) Control Sum weights = 1 (iii)

- Risk calculated: Invest in Tangent Portfolio*Risk Tangent Portfolio + (1-Invest in Tangent Portfolio) *
Risk Risk Free Asset

- Return calculated: Invest in Tangent Portfolio*Return Tangent Portfolio + (1-Invest in Tangent


Portfolio) * Return Risk Free Asset

Descriptive Analysis:

Mean  MITTELWERT; Median  MEDIAN; Mode  MODUS (Zeile oder Spalte); Variance 
VARIANZA; Skewness  SCHIEFE; Kurtosis  KURT; Percentiles  QUANT
IQR Analysis:

95% confidence interval


Q1 1.930,01
Q3 5.811,76
IQR 3.881,75

Q1: = QUANTIL.EXKL(alpha/2;DATA)

Q2: = QUANTIL.EXKL(1-alpha/2;DATA)
How to plot a joint density function:

How to derive to Cost of Equity:


1) Compute Excess Return of the market (market return – risk free) and the stock (stock – risk free)

2) Regress Excess Return stock = beta1*excess market + beta2*SMB + beta3*HML

3) Generate monthly means

4) Sum up [Annual Returns of Factors given by model (mean monthly simple returns * 12) *
respective beta] + Risk free rate

Das könnte Ihnen auch gefallen