Scribd
Upload
208 views

Eviews Commands

EViews commands covered include: 1. Importing data from Excel, choosing sample periods, and generating/transforming series such as taking logs, differences, and dummy variables. 2. Calculating descriptive statistics, correlations, and estimating equations like OLS, ARMA models, and multivariate systems. 3. Forecasting with estimated models and plotting forecasts with confidence intervals. 4. Diagnosing residuals and stability with tests for structural breaks, unit roots, and cointegration.

Uploaded by

tjsami
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
208 views

Eviews Commands

EViews commands covered include: 1. Importing data from Excel, choosing sample periods, and generating/transforming series such as taking logs, differences, and dummy variables. 2. Calculating descriptive statistics, correlations, and estimating equations like OLS, ARMA models, and multivariate systems. 3. Forecasting with estimated models and plotting forecasts with confidence intervals. 4. Diagnosing residuals and stability with tests for structural breaks, unit roots, and cointegration.

Uploaded by

tjsami
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Summary of important EViews-Commands

Import of data from EXCEL:


if the xlsx-format does not work, use File.xls
Choice of sample period:
Sample /
@all
@first @last
1990 2010
1981Q3 2005Q1
1960M1 2000M11
in command line e.g.: smpl @first 1990
Univariate statistics:
Click series / View /
Spreadsheet
Graph
Descriptive Statistics&Tests
Correlogram

data as numbers
Graphics
z.B. histogram, mean, etc.
autocorrelationen

Generation/Transformation of series:
Generate / x = 0
generates a series with zeros
Generate / pi = (pc pc(-1))/pc(-1)*100
Generates the inflation rate in % based on prices pc
Generate / x = log(y)
taking logs
Generate / dlx = dlog(x)
dlx = log(x) log(x(-1))
Growth rate in continuous time
Generate / y = exp(x)
exp(x)
as command: series x=0
Trend variable (linear):
Generate / t = @trend
Standard normal distributed realizations:
Generate / x = nrnd
Lags, lagged variables, taking differences:
Generate / x1 = x(-1)
x1(t) = x(t-1), Lag 1 of x
Generate / dx = d(x)
dx(t) = x(t) x(t-1) = (1-B)x(t)
first difference
Generate / d2x = d(x,2)
d2x(t) = dx(t) dx(t-1) = (1-B)^(2)x(t)
taking first differences twice
Generate / d12x = d(x,0,12)
d12x(t) = x(t) - x(t-12) = [1-B^(12)]x(t)
seasonal difference for monthly data
Generate d12_1x = d(x,1,12)
d12_1x(t) = (1-B)[1-B^(12)]x(t)
Geneartion of dummy variables:
seasonal dummies: s=1,2,3,...
Generate / ds = @seas(s)
as command: series ds = @seas(s)
Generate / d1 = 0 and manually in View/Spreadsheet
use Edit+/p-value for x of a test statistic as command:

(N-, t-,
scalar p
scalar p
scalar p
scalar p
scalar p

Chi2-, F-distribution)
= 1 - @cnorm(x)
= 1 - @cnorm(abs(x))*2
= 1 - @ctdist(x,df)
= 1 - @cchisq(x,df)
= 1 - @cfdist(x,df1,df2)

1-sided, right
2-sided
1-sided, right

df ... degrees of freedom


Determinant of correlation matrix: (as command)
group grpx x1 x2 x3 x4
matrix x = @convert(grpx)
group assigned to a data matrix
scalar det = @det( @cor(x) )
@cor(X) makes correlation matrix
Statistics for 2 or more series:
Correlation matrix:
Mark the series by klicking / right mouse / Open as Group / View /
/ Covariance Analysis (z.B. Correlation)
as command e.g.: scalar corrxy = @cor(x,y)
Estimation:
Quick / Estimate Equation / y c x1 x2
Option: LS ... OLS
Enter the estimation period

bivariate correlation coeff


y = a1 + a2*x1 + a3*x2 + u

as command: ls y c x1 x2
with storage in equation eq3: eq3.ls y c x1 x2
ARMA Model:
Option: LS
Model specification:
y c AR(1) AR(2) MA(1) MA(2) MA(3)

ARMA(2,3) with non zero mean

with heteroscedasticity robust standard errors:


Options: Covariance Coefficient Matrix / White (z.B.)
System estimation:
Open a multivariate equation system:
Object / New Object / System
Equations e.g.:
y1 = c(1) + c(2)*x1 + c(3)*y2
y2 = c(4) + c(5)*x2 + c(6)*x3
Instruments if needed e.g.:
inst x2 z7
Choose estimation method
Forecasting with LS, ...:
in LS Output (Equation) / Forecast
Forecast period, enter series name for forecasted values (e.g. yf) and
forecast error stdev (e.g. syf).
Plot of the forecast:
Object in Workfile menu / New Object / Group /
(or mark series with mouse / right click Open / as Group)
Enter the elements:
yf
forecasted values
y
observed values
(yf - 1.96*syf)
forecast interval, lower border
(yf + 1.96*syf)
forecast interval, upper border

Testing for structural breaks:


in Equation / View / Stability Diagnostics /
/ Chow Breakpoint Test / Enter: Start of 2., 3., ... period
/ Chow Forecast Test
/ Enter: Start of 2.period, ...
/ Recursive Estimates / CUSUM
Statistics for residuals: autocorrelation, heteroscedasticity
in Equation / View / Residual Diagnostics / ...
Autocorrelation, Heteroscedasticity
Storing residuals:
in Equation / Proc / Make Residual Series
Testing for Unit Roots:
View / Unit Root Test (Augmented Dickey Fuller)
Testing for cointegration:
Collect the variables in a group /
View / Cointegration Test (Johansen)
Estimation of a VAR or CIVAR:
Mark dependent variables / Open as VAR
(with right mouse-button)

You might also like