Q1.

Import the data given in an excel form in to EViews data sets (1pts)

Before importing the data, the first step in any project is creating a new work file by
following File > New > Work file. Since we have annual data from 1971to 2011 in regular
frequency, we adjust new work file like frequency, “Annual”, start date, “1971” and end date
2011

Once we created a new work file, we converted given PDF data into excel format in order to
import data into Eview software. Then it is imported and saved as it is indicated in the
following
 2. Draw the cross plot of inflation against its determinants individually (3pts)
A. The Cross plot of CPI against Broad money supply
As indicated on figure below, the line graphs of CPI and Broad money supply in their log
forms indicate that these graphs are upward trending as time increases. This implies that
these series are non-stationary over 1971-2011 periods.

A. Cross plot of inflation with money supply

12

10

2
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNMS

B. The Cross plot of CPI against Expected inflation

As indicated on figure 7below, the line graphs of CPI and Expected inflation in their log
forms indicate that these graphs are upward trending as time increases. This implies that these
series are non-stationary over 1971-2011 periods.
7

2
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNEP
 C. The Cross plot of CPI against Nominal effective exchange rate

As indicated on figure below, the line graphs of CPI and Nominal effective exchange rate
in their log forms indicate that these graphs are upward trending as time increases. This
implies that these series are non-stationary over 1971-2011 periods.
6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.0
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNNEER

D. The Cross plot of CPI against with Interest rate

As indicated on figure below, the line graphs of CPI and Interest e rate in their log forms
indicate that these graphs are upward trending as time increases. This implies that these series
are non-stationary over 1971-2011 periods.
6

-2

-4
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNR
E. The Cross plot of CPI against Real gross domestic product

As indicated on figure below, the line graphs of CPI and Real Gross Domestic Product in
their log forms indicate that these graphs are upward trending as time increases. This
implies that these series are non-stationary over 1971-2011 periods.
14

12

10

2
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNRGDP

F. Cross plot of CPI with Budget deficit

As indicated on figure 8 below, the line graphs of CPI and Budget deficit

in their log forms indicate that these graphs are upward trending as time increases. This
implies that these series are non-stationary over 1971-2011 periods

0
1975 1980 1985 1990 1995 2000 2005 2010

LNCPI LNBDY
 Q3. Conduct an OLS regression for price /inflation against its determinates and
interpret the result (3pts)

Dependent Variable: LNCPI

Method: Least Squares
Date: 08/02/21 Time: 18:11
Sample: 1971 2011
Included observations: 41

Variable Coefficient Std. Error t-Statistic Prob.

C -3.772654 1.447285 -2.606711 0.0135

LNBDY -0.009636 0.105509 -0.091327 0.9278
LNEP 0.091479 0.028900 3.165390 0.0033
LNMS 0.541633 0.062609 8.651101 0.0000
LNNEER 0.042924 0.082682 0.519148 0.6070
LNR 0.013036 0.026811 0.486206 0.6299
LNRGDP 0.178379 0.158800 1.123291 0.2692

R-squared 0.983974 Mean dependent var 3.766301

Adjusted R-squared 0.981146 S.D. dependent var 0.853793
S.E. of regression 0.117234 Akaike info criterion -1.295045
Sum squared resid 0.467286 Schwarz criterion -1.002484
Log likelihood 33.54843 Hannan-Quinn criter. -1.188511
F-statistic 347.9316 Durbin-Watson stat 0.914391
Prob(F-statistic) 0.000000

In this model, all covariates except LnMS and lnEP have negative and insignificant effect on
log of CPI
 Broad money supply (LnMS2) is statistically significant at 1% significance level. This
means that broad money supply (LnMS2) has effect on inflation. It has positive effect on
inflation of Ethiopia. This indicates that holding other things constant, a unit changes in
broad money supply (LnMS2) leads to 0.541633 changes in LnCPI during the study
period.

 Expected inflation or inflation inertia (LnEP) is statistically significant at 5% significance

level. This means that Expected inflation or inflation inertia (LnEP) has effect on
inflation. It has positive effect on inflation of Ethiopia. This indicates that holding other
things constant, a unit changes in Expected inflation or inflation inertia (LnEP) leads to
0.091479 changes in LnCPI during the study period.
 Durbin-Watson stat value of 0.914 is less than 2. This shows evidence for the existence of
strong positive serial correlation in the residuals.
 R2 which measures goodness of fit is 0.983974. This means that 98.39% variation of
LnCPI is explained by included explanatory variables such as broad money supply,
government expenditure and real GDP. Only 1.61% is unexplained variation. But R2 is
greater than Durbin-Watson stat which is symptom of spurious model. This suspicion can
be proved or disproved by testing unit roots of Residual for stationarity.
 F-statistic: Probability value of F-statistic is less than 5%. This implies that explanatory
variables are jointly significant in influencing inflation.
 The mean dependent variable is 3.77 which indicate the average value of LnCPI in the
data. The standard deviation of dependent variable is 0.85 which indicates the deviation
from the average value of Ln CPI the data.
 In our model, of the given information criteria, we select the one with the lowest value of
the three. Thereby, we select Akaike information criterion which is 1.295045 to adopt the
best model.

4.Does the OLS estimate satisfy the Normality, homoscedasticity, no autocorrelation

tests? (1pts)

6
Series: Residuals
Sample 1971 2011
5
Observations 41

4 Mean -3.28e-16
Median -0.011017
3 Maximum 0.184259
Minimum -0.251773
Std. Dev. 0.108084
2
Skewness -0.151773
Kurtosis 2.169933
1
Jarque-Bera 1.334468
0 Probability 0.513126
-0.2 -0.1 0.0 0.1 0.2

Fig 4.1 Normality test

If the null hypothesis is normality error distribution but when we see the result the jarque bera
probability value is insignificant that means not accept the null hypothesis sothat there is
normality distributed
Breusch-Godfrey Serial Correlation LM Test:

F-statistic 8.961106 Prob. F(2,32) 0.0008

Obs*R-squared 14.71911 Prob. Chi-Square(2) 0.0006

Test Equation:

In the figures above The null hypothesis: There is no serial correlation. Since both F and chi-
Square values are insignificant, the model is free from autocorrelation or serial correlation
problem.

Heteroskedasticity Test: Breusch-Pagan-Godfrey

F-statistic 1.750666 Prob. F(6,34) 0.1393

Obs*R-squared 9.676971 Prob. Chi-Square(6) 0.1389
Scaled explained SS 3.892785 Prob. Chi-Square(6) 0.6912

Heteroskedasticity problem is detected by Breusch-Pagan-Godfrey Test. Here, null

hypothesis states that there is constant error variance. Since both P-value of F(6, 34) and Chi-
squared are greater than the standard significance level (5%) which means p-values are
insignificant. Accordingly, the null hypothesis cannot be rejected which confirms there is no
heteroskedasticity problem in this model.
5. Recommend on the stability and specification of the model. (1pts)

Ramsey RESET Test

Equation: UNTITLED
Specification: LNCPI LNBDY LNEP LNMS LNNEER LNR
LNRGDP C
Omitted Variables: Squares of fitted values

Value df Probability
t-statistic 0.038459 33 0.9696
F-statistic 0.001479 (1, 33) 0.9696
Likelihood ratio 0.001838 1 0.9658

F-test summary:
Sum of df Mean
 Sq. Squares
Test SSR 2.09E-05 1 2.09E-05
Restricted SSR 0.467286 34 0.013744
Unrestricted SSR 0.467265 33 0.014160

Part II: Vector Autoregressive model

1. Using the same datasets determine the lag length (1pts)

Based on the the figure below

VAR Lag Order Selection Criteria

Endogenous variables: LNCPI LNBDY LNEP LNMS LNNEER
LNR LNRGDP
Exogenous variables: C
Date: 08/02/21 Time: 18:36
Sample: 1971 2011
Included observations: 39

Lag LogL LR FPE AIC SC HQ

0 -92.51064 NA 3.88e-07 5.103110 5.401698 5.210241

-
1 150.6593 386.5779 1.92e-11 -4.854325 2.465621* -3.997278
94.98703 - -
2 227.8363 * 5.80e-12* 6.299297* -1.820477 4.692335*

* indicates lag order selected by the criterion

LR: sequential modified LR test statistic (each test at 5% level)
FPE: Final prediction error
AIC: Akaike information criterion
SC: Schwarz information criterion
HQ: Hannan-Quinn information criterion

Q2. Test for unit root against all variables included in the model (comment on
the nature of variables) Unit Root Price Equation (3pts) fill the answer on the
provided table.
The absence or presence of unit roots for all-time series variables in the model is detected
mostly by using Augmented Dickey Fuller (ADF) test and Philips Perron test (PPT). Here,
Augmented Dickey Fuller (ADF) test is used since it Analyzes auto correlated data by
including lag values as per the frequency of data. It is the most prominent in economics and
used by many scholars.
Null hypothesis against alternative hypothesis is done to test existence of unit roots.
Null Hypothesis: variable has a unit root
Alternative Hypothesis: Variable has not unit root

The Null Hypothesis can be rejected by comparing ADF Test Statistic with the critical
values. If test Statistic is greater than Critical value at 1%, 5%, or 10% with significant P-
value, then null hypothesis is rejected and accept otherwise. Rejection of the Null
Hypothesis implies that a given series has no unit root which means the series is
stationary.

The results in table 3 indicate that all series in their log form have unit roots (non-stationary)
at level but have not unit roots (stationary) after first difference i.e. they are all integrated of
order one processes, I[ ] This is shown by the following figures using lnCPI as example’
Table ADF Unit Root Test for CPI at level with Intercept

Null Hypothesis: LNCPI has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.435027 0.9821

Test critical values: 1% level -3.605593
5% level -2.936942
10% level -2.606857

*MacKinnon (1996) one-sided p-values.

Table : ADF Unit Root Test for CPI at level with Intercept and Trend

Null Hypothesis: LNCPI has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*
Augmented Dickey-Fuller test statistic -1.419418 0.8398
Test critical values: 1% level -4.205004
5% level -3.526609
10% level -3.194611

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for CPI at level with none

Null Hypothesis: LNCPI has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 5.133280 1.0000

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

Afterr First Difference

Table: ADF Unit Root Test for CPI after first difference
with Intercept

Null Hypothesis: D(LNCPI) has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.897960 0.0003

Test critical values: 1% level -3.610453
5% level -2.938987
10% level -2.607932

*MacKinnon (1996) one-sided p-values.

Table: ADF Unit Root Test for CPI after first difference
with Intercept and Trend

Null Hypothesis: D(LNCPI) has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.855346 0.0018

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for CPI after first

difference with None

Null Hypothesis: D(LNCPI) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -3.310929 0.0015

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for BDY at level with

Intercept

Null Hypothesis: LNBDY has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -3.253491 0.0241

Test critical values: 1% level -3.605593
5% level -2.936942
10% level -2.606857

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for BDY at level with Intercept and Trend

Null Hypothesis: LNBDY has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -3.993378 0.0169

Test critical values: 1% level -4.205004
5% level -3.526609
10% level -3.194611

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for BDY at level with None

Null Hypothesis: LNBDY has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -0.937366 0.3047

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

Table ADF Unit Root Test for ms at level with intercept

Null Hypothesis: LNMS has a unit root

Exogenous: Constant
Lag Length: 1 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 1.160781 0.9973

Test critical values: 1% level -3.610453
5% level -2.938987
10% level -2.607932
*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for ms at level with intercept and trend

Null Hypothesis: LNMS has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.645004 0.9994

Test critical values: 1% level -4.205004
5% level -3.526609
10% level -3.194611

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for ms at level with None

Null Hypothesis: LNMS has a unit root

Exogenous: None
Lag Length: 1 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 4.617782 1.0000

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

EP ADF Unit Root Test for EP at level with intercept

Null Hypothesis: LNEP has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.586581 0.4800

Test critical values: 1% level -3.605593
5% level -2.936942
 10% level -2.606857

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for EP at level with intercept and trend

Null Hypothesis: LNEP has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.751066 0.7094

Test critical values: 1% level -4.205004
5% level -3.526609
10% level -3.194611

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for EP at level with none

Null Hypothesis: LNEP has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -0.177084 0.6161

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for BDY after first difference with Intercept

Null Hypothesis: D(LNBDY) has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -8.391080 0.0000

Test critical values: 1% level -3.610453
 5% level -2.938987
10% level -2.607932

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for BDY after first difference with Intercept and Trend

Null Hypothesis: D(LNBDY) has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -8.348513 0.0000

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for BDY after first difference with None

Null Hypothesis: D(LNBDY) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -8.471342 0.0000

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for MS after first difference with Intercept

Null Hypothesis: D(LNMS) has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.530519 0.0008

Test critical values: 1% level -3.610453
5% level -2.938987
10% level -2.607932

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for MS after first difference with Intercept and Trend

Null Hypothesis: D(LNMS) has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.672460 0.0030

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for MS after first difference with None

Null Hypothesis: D(LNMS) has a unit root

Exogenous: None
Lag Length: 1 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.078770 0.7018

Test critical values: 1% level -2.627238
5% level -1.949856
10% level -1.611469

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for EP after first difference with Intercept

Null Hypothesis: D(LNEP) has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.875602 0.0000

Test critical values: 1% level -3.610453
5% level -2.938987
 10% level -2.607932

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for EP after first difference with Intercept and Trend

Null Hypothesis: D(LNEP) has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.840437 0.0001

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for EP after first difference with None

Null Hypothesis: D(LNEP) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.954544 0.0000

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for BDY at level with Constant

LNNEER is not stationary at level
Null Hypothesis: LNNEER has a unit root
Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.141771 0.9651

Test critical values: 1% level -3.605593
5% level -2.936942
 10% level -2.606857

*MacKinnon (1996) one-sided p-values.

ADF Unit Root

Test for LNNEER
at level with
Constant

Null Hypothesis: LNNEER has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 1 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -2.551146 0.3035

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: LNNEER has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.932101 0.0519

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNNEER after first difference with
Intercept
LNNEER is stationary at level
Null Hypothesis: D(LNNEER) has a unit root
Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.505866 0.0009

Test critical values: 1% level -3.610453
 5% level -2.938987
10% level -2.607932

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: D(LNNEER) has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.515937 0.0046

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

Null Hypothesis: D(LNNEER) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.255352 0.0001

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNRGDP at level with none

LNRGDP is non stationary at level
Null Hypothesis: LNRGDP has a unit root
Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 4.136943 1.0000

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNRGDP at level with Constant and linear
Null Hypothesis: LNRGDP has a unit root
Exogenous: Constant, Linear Trend
Lag Length: 2 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 0.942366 0.9998

Test critical values: 1% level -4.219126
5% level -3.533083
10% level -3.198312

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNRGDP at level with Constant

Null Hypothesis: LNRGDP has a unit root

Exogenous: Constant
Lag Length: 2 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic 1.884479 0.9997

Test critical values: 1% level -3.615588
5% level -2.941145
10% level -2.609066

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNRGDP after first difference with none
LNRGDP is stationary at first difference
Null Hypothesis: D(LNRGDP) has a unit root
Exogenous: None
Lag Length: 2 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.002188 0.2781

Test critical values: 1% level -2.628961
5% level -1.950117
10% level -1.611339

*MacKinnon (1996) one-sided p-values.

ADF Unit Root

Test for LNRGDP
after first
difference with
constant and linear
Null Hypothesis: D(LNRGDP) has a unit root
Exogenous: Constant, Linear Trend
Lag Length: 1 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -4.247695 0.0093

Test critical values: 1% level -4.219126
5% level -3.533083
10% level -3.198312

*MacKinnon (1996) one-sided p-values.

it is stationary

ADF Unit Root Test for LNRGDP after first difference with constant

Null Hypothesis: D(LNRGDP) has a unit root

Exogenous: Constant
Lag Length: 2 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -2.108753 0.2425

Test critical values: 1% level -3.621023
5% level -2.943427
10% level -2.610263

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNR at level with none

LNR is not stationary at level
Null Hypothesis: LNR has a unit root
Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.601115 0.1021

Test critical values: 1% level -2.624057
5% level -1.949319
10% level -1.611711

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNR at level with Constant and trend

Null Hypothesis: LNR has a unit root

Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -2.109040 0.5252

Test critical values: 1% level -4.205004
5% level -3.526609
10% level -3.194611

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNR at level with Constant

Null Hypothesis: LNR has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -1.674430 0.4362

Test critical values: 1% level -3.605593
5% level -2.936942
10% level -2.606857

*MacKinnon (1996) one-sided p-values.

ADF Unit Root Test for LNR after first difference with constant and trend
LNR is stationary at first difference
Null Hypothesis: D(LNR) has a unit root
Exogenous: Constant, Linear Trend
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.208104 0.0007

Test critical values: 1% level -4.211868
5% level -3.529758
10% level -3.196411

*MacKinnon (1996) one-sided p-values.

it is stationary
 ADF Unit Root Test for LNR after first difference with constant

Null Hypothesis: D(LNR) has a unit root

Exogenous: Constant
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.282273 0.0001

Test critical values: 1% level -3.610453
5% level -2.938987
10% level -2.607932

Null Hypothesis: D(LNR) has a unit root

Exogenous: None
Lag Length: 0 (Automatic - based on SIC, maxlag=9)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -5.320381 0.0000

Test critical values: 1% level -2.625606
5% level -1.949609
10% level -1.611593

*MacKinnon (1996) one-sided p-values.

it is stationary

Q3. Is there a long run relationship between inflation and its determinants or
test for co integration (write also the long run equation with its prob) (2pts)

Date: 08/02/21 Time: 20:37

Sample (adjusted): 1973 2011
Included observations: 39 after adjustments
Trend assumption: Linear deterministic trend
Series: LNCPI LNBDY LNEP LNMS LNNEER LNR
LNRGDP
Lags interval (in first differences): 1 to 1

Unrestricted Cointegration Rank Test (Trace)

Hypothesized Trace 0.05

No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None * 0.800090 161.4011 125.6154 0.0001

At most 1 * 0.657792 98.61536 95.75366 0.0313
At most 2 0.426206 56.79420 69.81889 0.3468
At most 3 0.322779 35.13027 47.85613 0.4412
At most 4 0.269829 19.92972 29.79707 0.4276
At most 5 0.151570 7.665161 15.49471 0.5019
At most 6 0.031663 1.254836 3.841466 0.2626

Trace test indicates 2 cointegrating eqn(s) at the 0.05 level

* denotes rejection of the hypothesis at the 0.05 level
**MacKinnon-Haug-Michelis (1999) p-values

Unrestricted Cointegration Rank Test (Maximum Eigenvalue)

Hypothesized Max-Eigen 0.05

No. of CE(s) Eigenvalue Statistic Critical Value Prob.**

None * 0.800090 62.78570 46.23142 0.0004

At most 1 * 0.657792 41.82116 40.07757 0.0315
At most 2 0.426206 21.66392 33.87687 0.6336
At most 3 0.322779 15.20055 27.58434 0.7322
At most 4 0.269829 12.26456 21.13162 0.5218
At most 5 0.151570 6.410325 14.26460 0.5611
At most 6 0.031663 1.254836 3.841466 0.2626

Max-eigenvalue test indicates 2 cointegrating eqn(s) at the 0.05 level

* denotes rejection of the hypothesis at the 0.05 level
**MacKinnon-Haug-Michelis (1999) p-values

As indicated in the above table, null hypothesis of no cointegrating equation in the model is
rejected because Trace Test statistic value of 161.4011 is higher than critical value of
125.6154 or, P-value of 0.0001 is lower than the standard significance level of 0.05 which is
significant. This indicates that there is the existence of cointegrating equation at the
5%significance level. This means that there is linear combination exists between the series
that force these indices to have a relationship over the entire year time period, despite
potential deviation from equilibrium levels in the short-term. This means that Trace test
indicates 1 Cointegrating equation in the model.
Trace Test statistic value in Table 1 above indicated, the hypothesized number of
cointegrating equation of at most one in the model is rejected because Trace Test statistic
value of 98.61536 is higher than critical value of 95.75366 or, P-value of 0.0313 is lower
than the standard significance level of 0.05 which is significant. This indicates that there is
the existence of cointegrating equation at the 5%significance level.

Trace Test statistic value in Table above indicated, the hypothesized number of, at most two,
at most three and at most four, at most five at most six in the model is not rejected because
the Trace Test statistic are lower than critical values or P-values of the hypothesized number
of cointegrating equation of at most one ,at most two, at most three ,at most four at mos t five
and at most six are are higher than the standard P-value of 0.05 which is insignificant. That
means, there is no the existence of cointegrating equation at most one, at most two, at most
three, and at most four at the 5%significance level. That means that there is no linear
combination exists between the variables that force these indices to have a relationship over
the entire 41 year time period, despite potential deviation from equilibrium levels in the
short-term.

Decisions: Given the results generated from Trace statistic values and hypothesized
sentences, above, we observed that there is at least one co-integrating equation in the model
which confirms the long run relationship among series. That means the series are cointegrated
and they exhibit long run relationship.

The second method of Johansen Cointegration test is by observing Max-Eigen statistic value.
As indicated in the above table, null hypothesis of no cointegrating equation in the model is
rejected because Max-Eigen statistic value of 62.78570 is higher than critical value of
46.23142 or P-value of 0.0004 is lower than the standard significance level of 0.05 which is
significant. That means, there is the existence of cointegrating equation at the 5%significance
level. This cointegrating equation means that there is linear combination exists between the
series and the Max-Eigen statistic test shows 1 Cointegrating equation in the model

The Max-Eigen statistic test in Table above indicated, the hypothesized number of
cointegrating equation of at most one in the model is rejected because Max-Eigen statistic
value of 41.82116 is higher than critical value of 40.07757or, P-value of 0.0315 is lower than
the standard significance level of 0.05 which is significant. This indicates that there is the
existence of cointegrating equation at the 5%significance level.

The Max-Eigen statistic test in Table above indicated, the hypothesized number of
cointegrating equation of ,at most two, at most three ,at most four,at most five and at mos six
in the model is not rejected because the Max-Eigen statistic values are lower than critical
values or P-values are higher than the standard P-value of 0.05 which is insignificant.. That
means, there is no the existence of cointegrating equation at most one, at most two, at most
three, and at most four at the 5%significance level. This no cointegrating equation means that
there is no linear combination exists between the series. Thus, Max-Eigen statistic test
indicates 1 Cointegrating equation in the model

Decisions: Given the results generated from Max-Eigen statistic values and hypothesized
sentences, above, we observed that there is co-integrating equation in the model which
confirms the long run relationship among series. That means the series are cointegrated and
they exhibit long run relationship

General Decision: Generally, given the results generated above, it is observed that both
Trace statistic and Max-Eigen statistic confirms rejection of the null hypothesis (Ho) of no
cointegrating equation but fail to reject for hypothesized number of cointegrating equation.
Hence, it is concluded that there is at least one co-integrating equation in the model which
confirms the long run relationship among variables. That means the series are cointegrated
and they exhibit long run relationship meaning that ,if there are shocks to the system in the
short run, then the model is likely to converge with time in the long run. Now, Cointegration
test is completed and VECM Estimation procedure has to follow it.

Table: Estimation of Long run model of VECM

Dependent Variable: D(LNCPI)

Method: Least Squares (Gauss-Newton / Marquardt steps)
Date: 08/02/21 Time: 21:02
Sample (adjusted): 1973 2011
Included observations: 39 after adjustments
D(LNCPI) = C(1)*( LNCPI(-1) - 1.09606161615*LNBDY(-1) -
0.263426740017*LNEP(-1) - 0.107210224329*LNMS(-1) +
0.363971608984*LNNEER(-1) + 0.0560906029411*LNR(-1)
-
 1.25631281412*LNRGDP(-1) + 11.4272388234 ) +
C(2)*D(LNCPI(-1))
+ C(3)*D(LNBDY(-1)) + C(4)*D(LNEP(-1)) +
C(5)*D(LNMS(-1)) + C(6)
*D(LNNEER(-1)) + C(7)*D(LNR(-1)) + C(8)*D(LNRGDP(-
1)) + C(9)

Coefficient Std. Error t-Statistic Prob.

C(1) -0.262978 0.063728 -4.126542 0.0003

C(2) 0.237884 0.124728 1.907222 0.0661
C(3) -0.132116 0.067762 -1.949720 0.0606
C(4) 0.011537 0.026131 0.441524 0.6620
C(5) 0.667610 0.210230 3.175617 0.0034
C(6) 0.024445 0.086959 0.281113 0.7806
C(7) 0.030096 0.018389 1.636653 0.1122
C(8) -0.490917 0.157741 -3.112165 0.0041
C(9) -0.002266 0.030296 -0.074791 0.9409

R-squared 0.567456 Mean dependent var 0.079091

Adjusted R-squared 0.452111 S.D. dependent var 0.094905
S.E. of regression 0.070249 Akaike info criterion -2.274380
Sum squared resid 0.148046 Schwarz criterion -1.890481
Log likelihood 53.35040 Hannan-Quinn criter. -2.136640
F-statistic 4.919635 Durbin-Watson stat 2.092994
Prob(F-statistic) 0.000598

Estimation of Short run model or ECM

Dependent Variable: D(LNCPI)

Method: Least Squares
Date: 08/02/21 Time: 21:12
Sample (adjusted): 1974 2011
Included observations: 38 after adjustments

Variable Coefficient Std. Error t-Statistic Prob.

C -0.030601 0.046343 -0.660311 0.5159

D(LNCPI(-1)) 0.973640 0.422590 2.303980 0.0310
D(LNCPI(-2)) -0.315293 0.228924 -1.377281 0.1823
D(LNBDY(-1)) 0.085278 0.066130 1.289554 0.2106
D(LNBDY(-2)) 0.043189 0.063746 0.677516 0.5051
D(LNEP(-1)) 0.057097 0.031559 1.809229 0.0841
D(LNEP(-2)) 0.003104 0.030719 0.101048 0.9204
D(LNMS(-1)) 0.584514 0.226825 2.576942 0.0172
D(LNMS(-2)) -0.317435 0.485383 -0.653989 0.5199
D(LNNEER(-1)) -0.198031 0.131129 -1.510203 0.1452
 D(LNNEER(-2)) 0.156936 0.110453 1.420833 0.1694
D(LNR(-1)) 0.012366 0.020105 0.615048 0.5448
D(LNR(-2)) -0.052018 0.019780 -2.629820 0.0153
D(LNRGDP(-1)) -0.319928 0.257668 -1.241630 0.2274
D(LNRGDP(-2)) 0.762805 0.236931 3.219524 0.0039
ECT(-1) -0.980118 0.527581 -1.857759 0.0766

R-squared 0.687960 Mean dependent var 0.082823

Adjusted R-squared 0.475205 S.D. dependent var 0.093233
S.E. of regression 0.067541 Akaike info criterion -2.256607
Sum squared resid 0.100359 Schwarz criterion -1.567097
Log likelihood 58.87553 Hannan-Quinn criter. -2.011284
F-statistic 3.233580 Durbin-Watson stat 2.334330
Prob(F-statistic) 0.006266

Q4 which determinants of inflation in Ethiopia are statistically significant in the

long run? Or test for significance of the coefficient of the predictors. fill the table
below : (2pts)

As indicated by long run Model of VECM estimation are explained as follows. First
difference LnCPI is statistically significant at 5% significance levels. It has negative effect on
Inflation of Ethiopia in the long run. This indicates that holding other things constant, a unit
changes in one periods lagged LnCPI changes inflation by 0.262978 amounts indifferent
direction in the long run.

As indicated by long run Model of VECM estimation are explained as follows. one periods
lagged LnMS is statistically significant at 5% significance levels. It has positive effect on
Inflation of Ethiopia in the long run. This indicates that holding other things constant, a unit
changes in one periods lagged Lnms changes inflation by 0.667610 amounts in the same
direction in the long run.

One period lagged LnRGDP is statistically significant at 5% significance level in the short
run. This variable has negative impact on inflation of Ethiopia in the short run. The negative
sign of coefficient of this variable implies that a unit changes in one period lagged LnRGDP
reduces inflation by 0.490917 amounts in the short run.

Q5.Test for weak exogeniety of the variables used in the model? What did you recommend if
the explanatory variables are endogenous? (2pts)
 Variables LOGCPI LOGBDY LOGRGDP LOGMS LOGNEER LOG ∏e LOG R

- Coefficient

LR-Testχ2 (1)

p-value

Q6.Which determinants are statistically significant or responsible for the short run
dynamics of inflation in Ethiopia? Furthermore conduct all necessary tests (normality
,autocorrelation ,hetrocedasticity, stability and model specification ) and interpret the
result . or fill the table (4pts)
Dependent Variable: D(LNCPI)
Method: Least Squares
Date: 08/02/21 Time: 21:12
Sample (adjusted): 1974 2011
Included observations: 38 after adjustments

Variable Coefficient Std. Error t-Statistic Prob.

C -0.030601 0.046343 -0.660311 0.5159

D(LNCPI(-1)) 0.973640 0.422590 2.303980 0.0310
D(LNCPI(-2)) -0.315293 0.228924 -1.377281 0.1823
D(LNBDY(-1)) 0.085278 0.066130 1.289554 0.2106
D(LNBDY(-2)) 0.043189 0.063746 0.677516 0.5051
D(LNEP(-1)) 0.057097 0.031559 1.809229 0.0841
D(LNEP(-2)) 0.003104 0.030719 0.101048 0.9204
D(LNMS(-1)) 0.584514 0.226825 2.576942 0.0172
D(LNMS(-2)) -0.317435 0.485383 -0.653989 0.5199
D(LNNEER(-1)) -0.198031 0.131129 -1.510203 0.1452
D(LNNEER(-2)) 0.156936 0.110453 1.420833 0.1694
D(LNR(-1)) 0.012366 0.020105 0.615048 0.5448
D(LNR(-2)) -0.052018 0.019780 -2.629820 0.0153
D(LNRGDP(-1)) -0.319928 0.257668 -1.241630 0.2274
D(LNRGDP(-2)) 0.762805 0.236931 3.219524 0.0039
ECT(-1) -0.980118 0.527581 -1.857759 0.0766

R-squared 0.687960 Mean dependent var 0.082823

Adjusted R-squared 0.475205 S.D. dependent var 0.093233
S.E. of regression 0.067541 Akaike info criterion -2.256607
Sum squared resid 0.100359 Schwarz criterion -1.567097
Log likelihood 58.87553 Hannan-Quinn criter. -2.011284
F-statistic 3.233580 Durbin-Watson stat 2.334330
Prob(F-statistic) 0.006266
One period lagged ECT is derived from estimated dynamic long run model. It indicates the
speed of adjustment to restore equilibrium in the dynamic model. It has been shown in the
table 9 below, the coefficient of ECT (-1) is -0.980118. This coefficient is negative and
significant at 1% significance level. This coefficient implies that 98% deviation from long
run equilibrium level of inflation will be adjusted in a year. The whole system can come back
to (converging to) long run equilibrium at speed of 98.6% per year when there is a shock to
steady state relationship

Q7. Generally give good explanation regarding determinants and status of inflation dynamics
in Ethiopia (3pts)

Dr the time up that is why I am not tried Q7.