Scribd
Upload
8 views3 pages

Autocorrelation and Homoskedasticity Analysis

The document discusses performing several statistical tests on a time series regression model. It reports that: 1. An autocorrelation test of the residuals rejects the null hypothesis of no autocorrelation, confirming autocorrelation exists. 2. A test of homoskedasticity fails to reject the null hypothesis of constant variance, confirming the residuals have constant variance. 3. There is no serious problem of multicollinearity according to variance inflation factors and correlation matrices.

Uploaded by

mathibettu
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
8 views3 pages

Autocorrelation and Homoskedasticity Analysis

The document discusses performing several statistical tests on a time series regression model. It reports that: 1. An autocorrelation test of the residuals rejects the null hypothesis of no autocorrelation, confirming autocorrelation exists. 2. A test of homoskedasticity fails to reject the null hypothesis of constant variance, confirming the residuals have constant variance. 3. There is no serious problem of multicollinearity according to variance inflation factors and correlation matrices.

Uploaded by

mathibettu
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

1.

Null hypothesis: No autocorrelation in residuals


Since p value<0.05, we reject the null hypothesis and
confirm the existence of autocorrelation in residuals
log(s_p_price_index) c dividend d(log(cpi))

log(s_p_price_index) c log(dividend) d(log(cpi))

log(s_p_price_index) c dividend d(log(cpi)) AR(1)


log(s_p_price_index) c dividend d(log(cpi))
dividend(-1) dlog(cpi(-1)) log(s_p_price_index(-1))

Model Model Model Model


Model
1 II III V
IV
AIC -3.86 -4.01 -3.75 1.2 2.3
SIC
HC
RMSEA
Mean
Absolut
e

Yt = a +bxt +bxt-1+ZYt-1
2. Null hypothesis: Variance is constant (presence of
Homoskedasticity)
BP test
Since p value>0.05, we fail to reject the null
hypothesis and confirm that residuals have constant
variance
3. No serious problem of Multicollinearity
a. VIF almost equal to 10
b. Correlation matrix (no correlation was significant
or higher than 0.5)

D(log(x)) = log x2- logx1


Log(D(X)) =log(X2-x1) Not possible

Y = A + BX
Log(Y) = A + BX
Relative change in Y/absolute change in x
Log(Y) =A + Blog(x)
Change in Y/Y
New=d(cpi)
New1=log(cpi)
New2=cpi+dividened
CPI DIVIDEND
11 .48
.44
10
.40
9 .36
.32
8 .28
.24
7
.20
6 .16
2000 2005 2010 2015 2000 2005 2010 2015

EARNINGS S_P_PRICE_INDEX
.8 12
.7
10
.6
.5 8

.4 6
.3
4
.2
.1 2
2000 2005 2010 2015 2000 2005 2010 2015

You might also like