Grocer Function
NAME
becm - performs bayesian error correction model estimation
CALLING SEQUENCE
rbecm=becm(nlag,tight,weight,decay,arg1,...,argn)
PARAMETERS
Input
-
nlag = the lag length of the VAR
-
tight = Litterman's tightness hyperparameter
-
weight = Litterman's weight (matrix or scalar)
- decay = Litterman's lag decay = lag^(-decay)
-
argi = an argument which can be:
- 'jres=xx' where xx is the name of a johansen results tlist (optional: if not given, is estimated by the function)
- 'nbr=xx' where xx is the # of cointegration vectors to keep from the johansen estimation (optional: if not
given, is calculated by the function with a level equal to plevel)
- 'plevel=xx' where xx=0.01, 0.05 or 0.1 is the significance level for the cointegrating vectors (optional: if not given, is set to 0.05; useless if the option 'nbr=xx' is used)
- the string 'noprint' if the user doesn't want to print the results of the regression
- 'endo=var1;var2;...;varn' or 'endo=[var1;var2;...;varn]' with vari the name of an admissible variable for explox
(a real vector, a ts, a real matrix or a list of such variables)
Output
-
rbecm = a results tlist with:
. rbecm('meth') = 'becm'
. rbecm('y') = y data vector
. rbecm('x') = x data matrix
. rbecm('nobs') = # observations
. rbecm('nvar') = # exogenous variables
. rbecm('neqs') = # endogenous variables
. rbecm('tight') = Litterman's tightness hyperparameter
. rbecm('weight') = Litterman's weight (matrix or scalar)
. rbecm('decay') = Litterman's lag decay = lag^(-decay)
. rbecm('resid') = residuals, with rbecm('resid')(:,i): residuals for equation # i
. rbecm('beta') = bhat, with rbecm('beta')(:,i): coefficients for equation # i
. rbecm('rsqr') = rsquared, with rbecm('rsqr')(i): rsquared for equation # i
. rbecm('f') = F-stat for the nullity of coefficients other than the constant with: rbecm('f')(i): F-stat for equation # i
. rbecm('pvaluef') = their significance level with: rbecm('pvaluef')(i): significance level for equation # i
. rbecm('rbar') = rbar-squared with rbecm('rbar')(i): rbarsquared for equation # i
. rbecm('sigu') = sums of squared residuals with rbecm('sigu')(:,i): sum of squared residuals for equation # i
. rbecm('ser') = standard errors of the regression with rbecm('ser')(i): standard error for equation # i
. rbecm('tstat') = t-stats, with rbecm('tstat')(:,i): t-stat for equation # i
. rbecm('pvalue')= pvalue of the betas, with rbecm('pvalue')(:,i): p-value for equation # i
. rbecm('dw') = Durbin-Watson Statistic, with: rbecm('dw')(i): DW for equation # i
. rbecm('boxq') = Box Q-stat, with rbecm('boxq')(i): Box Q-stat for equation # i
. rbecm('sigma') = (neqs x neqs) var-covar matrix of the regression
. rbecm('prests') = boolean indicating the presence or absence of a time series in the regression
. rbecm('namey') = name of the y variable
. rbecm('nx') = # of x variables
. rbecm('nb_coint_relat') = # of cointegration relations
. rbecm('namex') = name of the cointegrating variables (if any)
. rbecm('bounds') = if there is a timeseries in the regression, the bounds of the regression
. rbecm('jres') = the result tlist from the johansen step
DESCRIPTION
Performs error bayesian correction model estimation. The user can specify the names of the tlist resulting from johansen estimation, the number of cointegration relations to take into account or the significance level to use to select the number of cointegration relations, but if these parameters are not provided, the function takes default values and, if necessary, performs johansen estimation. The lags of the var models are estimated with bayesian prior given by the parameters tight, weight and decay.
EXAMPLE
results = becm(2,0.1,1,0.5,'illinos','indiana','kentucky','michigan'...
,'ohio','pennsyvlania','tennesse','westvirginia');
Example taken from function becm_d(); endogenous variables are 'illinos','indiana','kentucky','michigan',
'ohio','pennsyvlania','tennesse','westvirginia'; # of lags set to 2; tight, weight and decay are set to 0.1,1 and 0.5.
AUTHOR
Eric Dubois 2002