Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Eviews5 / EViews5 / Docs / EViews 5 Command Ref.pdf
Скачиваний:
91
Добавлен:
23.03.2015
Размер:
5.23 Mб
Скачать

308—Appendix B. Command Reference

garch

Equation View

 

 

Conditional standard deviation graph of (G)ARCH equation.

Displays the conditional standard deviation or conditional variance graph of an equation estimated by ARCH.

Syntax

Equation View: eq_name.garch(options)

Options

vDisplay conditional variance graph instead of the standard deviation graph.

p

Print the graph

Examples

equation eq1.arch sp500 c

eq1.garch

estimates a GARCH(1,1) model and displays the estimated conditional standard deviation graph.

eq1.garch(v, p)

displays and prints the estimated conditional variance graph.

Cross-references

ARCH estimation is described in Chapter 20, “ARCH and GARCH Estimation”, on page 599 of the User’s Guide.

See also arch (p. 206) and makegarch (p. 352).

genr

Command || Object Declaration (Alpha) | Object Declaration (Series) | Pool Proc

Generate series.

Generate series or alphas. This procedure also allows you to generate multiple series using the cross-section identifiers in a pool.

genr—309

Syntax

Command: genr ser_name = expression

Pool Proc:

pool_name.genr ser_name = expression

For pool series generation, you may use the cross section identifier “?” in the series name and/or in the expression on the right-hand side.

Examples

genr y = 3 + x

generates a numeric series that takes the values from the series X and adds 3.

genr full_name = first_name + last_name

creates an alpha series formed by concatenating the alpha series FIRST_NAME and LAST_NAME.

The commands,

pool pool1

pool1.add 1 2 3

pool1.genr y? = x? - @mean(x?)

are equivalent to generating separate series for each cross-section:

genr y1 = x1 - @mean(x1) genr y2 = x2 - @mean(x2) genr y3 = x3 - @mean(x3)

Similarly:

pool pool2 pool2.add us uk can

pool2.genr y_? = log(x_?) - log(x_us)

generates three series Y_US, Y_UK, Y_CAN that are the log differences from X_US. Note that Y_US=0.

It is worth noting that the pool genr command simply loops across the cross-section identifiers, performing the evaluations using the appropriate substitution. Thus, the command,

pool2.genr z = y_?

is equivalent to entering:

genr z = y_us

genr z = y_uk

310—Appendix B. Command Reference

genr z = y_can

so that upon completion, the ordinary series Z will contain Y_CAN.

Cross-references

See Chapter 27, “Pooled Time Series, Cross-Section Data”, on page 823 of the User’s Guide for a discussion of the computation of pools, and a description of individual and balanced samples.

See series (p. 442) and alpha (p. 204) for a discussion of the expressions allowed in genr.

gmm

Command || Equation Method | System Method

 

 

Estimation by generalized method of moments (GMM).

The equation or system object must be specified with a list of instruments.

Syntax

Command: gmm(options) y x1 [x2 x3 ...] @ z1 [z2 z3 ...]

gmm(options) specification @ z1 [z2 z3 ...]

Equation Method: eq_name.gmm(options) y x1 [x2 x3 ...] @ z1 [z2 z3 ...]

eq_name.gmm(options) specification @ z1 [z2 z3 ...]

System Method: system_name.gmm(options)

To use gmm as a command or equation method, follow the name of the dependent variable by a list of regressors, followed by the “@” symbol, and a list of instrumental variables which are orthogonal to the residuals. Alternatively, you can specify an expression using coefficients, an “@” symbol, and a list of instrumental variables. There must be at least as many instrumental variables as there are coefficients to be estimated.

In panel settings, you may specify dynamic instruments corresponding to predetermined variables. To specify a dynamic instrument, you should tag the instrument using “@DYN”, as in “@DYN(X)”. By default, EViews will use a set of period-specific instruments corresponding to lags from -2 to “-infinity”. You may also specify a restricted lag range using arguments in the “@DYN” tag. For example, to use lags from-5 to “-infinity” you may enter “@DYN(X, -5)”; to specify lags from -2 to -6, use “@DYN(X, -2, -6)” or “@DYN(X, -6, -2)”.

Note that dynamic instrument specifications may easily generate excessively large numbers of instruments.

gmm—311

Options

General Options

m=integer

Maximum number of iterations.

 

 

c=number

Set convergence criterion. The criterion is based upon

 

the maximum of the percentage changes in the scaled

 

coefficients.

 

 

l=number

Set maximum number of iterations on the first-stage

 

iteration to get the one-step weighting matrix.

 

 

showopts /

[Do / do not] display the starting coefficient values and

-showopts

estimation options in the estimation output.

 

 

deriv=keyword

Set derivative methods. The argument keyword should

 

be a oneor two-letter string. The first letter should

 

either be “f” or “a” corresponding to fast or accurate

 

numeric derivatives (if used). The second letter should

 

be either “n” (always use numeric) or “a” (use analytic

 

if possible). If omitted, EViews will use the global

 

defaults.

 

 

p

Print results.

Additional Options for Non-Panel Equation and System estimation

w

Use White’s diagonal weighting matrix (for cross sec-

 

tion data).

 

 

b=arg

Specify the bandwidth: “nw” (Newey-West fixed band-

(default=“nw”)

width based on the number of observations), “number

 

(user specified bandwidth), “v” (Newey-West auto-

 

matic variable bandwidth selection), “a” (Andrews

 

automatic selection).

 

 

q

Use the quadratic kernel. Default is to use the Bartlett

 

kernel.

 

 

n

Prewhiten by a first order VAR before estimation.

 

 

i

Iterate simultaneously over the weighting matrix and

 

the coefficient vector.

 

 

s

Iterate sequentially over the weighting matrix and coef-

 

ficient vector.

 

 

o (default)

Iterate only on the coefficient vector with one step of

 

the weighting matrix.

 

 

312—Appendix B. Command Reference

c

One step (iteration) of the coefficient vector following

 

one step of the weighting matrix.

 

 

e

TSLS estimates with GMM standard errors.

 

 

Additional Options for Panel Equation estimation

cx=arg

Cross-section effects method: (default) none, fixed

 

effects estimation (“cx=f”), first-difference estimation

 

(“cx=fd”), orthogonal deviation estimation (“cx=od”)

 

 

per=arg

Period effects method: (default) none, fixed effects esti-

 

mation (“per=f”).

 

 

levelper

Period dummies always specified in levels (even if one

 

of the transformation methods is used, “cx=fd” or

 

“cx=od”).

 

 

wgt=arg

GLS weighting: (default) none, cross-section system

 

weights (“wgt=cxsur”), period system weights

 

(“wgt=persur”), cross-section diagonal weighs

 

(“wgt=cxdiag”), period diagonal weights (“wgt=per-

 

diag”).

 

 

gmm=arg

GMM weighting: 2SLS (“gmm=2sls”), White period

 

system covariances (Arellano-Bond 2-step/n-step)

 

(“gmm=perwhite”), White cross-section system

 

(“gmm=cxwhite”), White diagonal

 

(“gmm=stackedwhite”), Period system (“gmm=per-

 

sur”), Cross-section system (“gmm=cxsur”), Period

 

heteroskedastic (“cov=perdiag”), Cross-section het-

 

eroskedastic (“gmm=cxdiag”).

 

By default, uses the identity matrix unless estimated

 

with first difference transformation (“cx=fd”), in

 

which case, uses (Arellano-Bond 1-step) difference

 

weighting matrix. In this latter case, you should specify

 

2SLS weights (“gmm=2sls”) for Anderson-Hsiao esti-

 

mation.

 

 

gmm—313

cov=arg

Coefficient covariance method: (default) ordinary,

 

White cross-section system robust (“cov=cxwhite”),

 

White period system robust (“cov=perwhite”), White

 

heteroskedasticity robust (“cov=stackedwhite”), Cross-

 

section system robust/PCSE (“cov=cxsur”), Period sys-

 

tem robust/PCSE (“cov=persur”), Cross-section het-

 

eroskedasticity robust/PCSE (“cov=cxdiag”), Period

 

heteroskedasticity robust (“cov=perdiag”).

 

 

keepwgts

Keep full set of GLS/GMM weights used in estimation

 

with object, if applicable (by default, only weights

 

which take up little memory are saved).

 

 

nodf

Do not perform degree of freedom corrections in com-

 

puting coefficient covariance matrix. The default is to

 

use degree of freedom corrections.

 

 

coef=arg

Specify the name of the coefficient vector (if specified

 

by list); the default behavior is to use the “C” coeffi-

 

cient vector.

 

 

iter=arg

Iteration control for GLS and GMM weighting specifica-

(default=“onec”)

tions: perform one weight iteration, then iterate coeffi-

 

cients to convergence (“iter=onec”), iterate weights

 

and coefficients simultaneously to convergence

 

(“iter=sim”), iterate weights and coefficients sequen-

 

tially to convergence (“iter=seq”), perform one weight

 

iteration, then one coefficient step (“iter=oneb”).

Note that some options are only available for a subset of specifications.

Examples

In a non-panel workfile, we may estimate equations using the standard GMM options. The specification:

eq.gmm(w) y c f k @ c z1 z2 z3

estimates the linear specification using a White diagonal weighting matrix (one-step, with no iteration). The command:

eq.gmm(e, b=v) c(1) + c(2)*(m^c(1) + k^(1-c(1))) @ c z1 z2 z3

estimates the nonlinear model using two-stage least squares (instrumental variables) with GMM standard errors computed using Newey-West automatic bandwidth selected weights.

For system estimation, the command:

314—Appendix B. Command Reference

sys1.gmm(b=a, q, i)

estimates the system SYS1 by GMM with a quadratic kernel, Andrews automatic bandwidth selection, and iterates until convergence.

When working with a workfile that has a panel structure, you may use the panel equation estimation options. The command

eq.gmm(cx=fd, per=f) dj dj(-1) @ @dyn(dj)

estimates an Arellano-Bond “1-step” estimator with differencing of the dependent variable DJ, period fixed effects, and dynamic instruments constructed using DJ with observation specific lags from period t − 2 to 1.

To perform the “2-step” version of this estimator, you may use:

eq.gmm(cx=fd, per=f, gmm=perwhite, iter=oneb) dj dj(-1) @ @dyn(dj)

where the combination of the options “gmm=perwhite” and (the default) “iter=oneb” instructs EViews to estimate the model with the difference weights, to use the estimates to form period covariance GMM weights, and then re-estimate the model.

You may iterate the GMM weights to convergence using:

eq.gmm(cx=fd, per=f, gmm=perwhite, iter=seq) dj dj(-1) @ @dyn(dj)

Alternately:

eq.gmm(cx=od, gmm=perwhite, iter=oneb) dj dj(-1) x y @ @dyn(dj,-2,-6) x(-1) y(-1)

estimates an Arellano-Bond “2-step” equation using orthogonal deviations of the dependent variable, dynamic instruments constructed from DJ from period t − 6 to t − 2 , and ordinary instruments X(-1) and Y(-1).

Cross-references

See Chapter 16, “Additional Regression Methods”, on page 459, Chapter 23, “System Estimation”, on page 693, and Chapter 29, “Panel Estimation”, on page 899 of the User’s Guide for discussion of the various GMM estimation techniques.

Соседние файлы в папке Docs