Question

ForecastX Regressions

Exhibit #1

Audit Trail — Coefficient Table (Multiple Regression Selected)

Series Description

Included in Model

Coefficient

Standard Error

T-test

P-value

F-test

Elasticity

Overall F-test

SALES

Dependent

−51.24

54.32

−0.94

0.36

0.89

8.98

PRICE

Yes

30.92

10.32

3.00

0.01

8.98

1.46

Audit Trail — Correlation Coefficient Table
Series Description	Included in Model			SALES			PRICE
SALES		Dependent			1.00			0.63
PRICE		Yes			0.63			1.00

Audit Trail - Statistics
Accuracy Measures	Value		Forecast Statistics	Value
AIC	130.02		Durbin Watson(1)	0.34
BIC	130.80		Mean	111.19
Mean Absolute Percentage Error (MAPE)	10.67	%	Standard Deviation	17.49
R-Square	39.07	%	Ljung-Box	39.71
Adjusted R-Square	34.72	%
Root Mean Square Error	13.22

Exhibit #2

Audit Trail — Coefficient Table (Multiple Regression Selected)
Series Description	Included in Model			Coefficient			Standard Error			T-test			P-value			F-test			Elasticity			Overall F-test
SALES		Dependent			123.47			19.40			6.36			0.00			40.51						154.86
PRICE		Yes			−24.84			4.95			−5.02			0.00			25.17			−1.17
INCOME		Yes			0.03			0.00			13.55			0.00			183.62			1.06

Audit Trail — Correlation Coefficient Table
Series Description	Included in Model			SALES			PRICE			INCOME
SALES		Dependent			1.00			0.63			0.94
PRICE		Yes			0.63			1.00			0.83
INCOME		Yes			0.94			0.83			1.00

Audit Trail - Statistics
Accuracy Measures	Value		Forecast Statistics	Value
AIC	86.56		Durbin Watson(1)	1.67
BIC	87.34		Mean	111.19
Mean Absolute Percentage Error (MAPE)	2.22	%	Standard Deviation	17.49
R-Square	95.97	%	Ljung-Box	15.22
Adjusted R-Square	95.35	%
Root Mean Square Error	3.40

Consider the two regressions shown above.

Multiple Choice

• The simple regression probably suffers from multicollinearity.

• The multiple regression probably suffers from specification error.

• The multiple regression probably suffers from autocorrelation.

• The simple regression probably suffers from specification error.

Answer 1

Option 1: The simple regression probably suffers from multicollinearity. Since there is only one independent variable, there can not be multicollinearity since multicollinearity arise if there is high correlation among the independent variables. It means, there must be at least 2 independent variables to check for multicollinearity. Hence, this option is not correct.

Option 2: The multiple regression probably suffers from specification error. The results do not show the RESET test results which is used to asess the validity of specification. Hence, it should not be correct.

Option 3: The multiple regression probably suffers from autocorrelation. Looking at the value of DW = 1.67 in the multiple regression, I think it lies in the region where it may be inconclusive and hence the multiple regression probably suffers from autocorrelation. Hence, this option should be correct.

Option 4: The multiple regression probably suffers from specification error. The results do not show the RESET test results which is used to asess the validity of specification. Hence, it should not be correct.

The tables are not organised properly and hence not easy to read them. I can not find the number of observations for the regression. Hence, I may not be able to direct to the exact cell of the DW table.