[V,VData,Fitted] = estimate(V,Inp,Range,...)
V
[ VAR ] - Empty VAR object.
Inp
[ struct ] - Input database.
Range
[ numeric ] - Estimation range, including P
pre-sample periods, where P
is the order of the VAR.
V
[ VAR ] - Estimated reduced-form VAR object.
VData
[ struct ] - Output database with the endogenous variables and the estimated residuals.
Fitted
[ numeric ] - Periods in which fitted values have been calculated.
'A='
[ numeric | empty ] - Restrictions on the individual values in the transition matrix, A
.
'BVAR='
[ numeric ] - Prior dummy observations for estimating a BVAR; construct the dummy observations using the one of the BVAR
functions.
'C='
[ numeric | empty ] - Restrictions on the individual values in the constant vector, C
.
'J='
[ numeric | empty ] - Restrictions on the individual values in the coefficient matrix in front of exogenous inputs, J
.
'diff='
[ true
| false
] - Difference the series before estimating the VAR; integrate the series back afterwards.
'G='
[ numeric | empty ] - Restrictions on the individual values in the coefficient matrix in front of the co-integrating vector, G
.
'cointeg='
[ numeric | empty ] - Co-integrating vectors (in rows) that will be imposed on the estimated VAR.
'comment='
[ char | Inf
] - Assign comment to the estimated VAR object; Inf
means the existing comment will be preserved.
'constraints='
[ char | cellstr ] - General linear constraints on the VAR parameters.
'constant='
[ true
| false
] - Include a constant vector in the VAR.
'covParam='
[ true
| false
] - Calculate and store the covariance matrix of estimated parameters.
'eqtnByEqtn='
[ true
| false
] - Estimate the VAR equation by equation.
'maxIter='
[ numeric | 1
] - Maximum number of iterations when generalised least squares algorithm is involved.
'mean='
[ numeric | empty ] - Impose a particular asymptotic mean on the VAR process.
'order='
[ numeric | 1
] - Order of the VAR.
'progress='
[ true
| false
] - Display progress bar in the command window.
'schur='
[ true
| false
] - Calculate triangular (Schur) representation of the estimated VAR straight away.
'stdize='
[ true
| false
] - Adjust the prior dummy observations by the std dev of the observations.
'timeWeights=
' [ tseries | empty ] - Time series of weights applied to individual periods in the estimation range.
'tolerance='
[ numeric | 1e-5
] - Convergence tolerance when generalised least squares algorithm is involved.
'warning='
[ true
| false
] - Display warnings produced by this function.
'fixedEff='
[ true
| false
] - Include constant dummies for fixed effect in panel estimation; applies only if 'constant=' true
.
'groupWeights='
[ numeric | empty ] - A 1-by-NGrp vector of weights applied to groups in panel estimation, where NGrp is the number of groups; the weights will be rescaled so as to sum up to 1
.
Panel VAR objects are created by calling the function VAR
with two input arguments: the list of variables, and the list of group names. To estimate a panel VAR, the input data, Inp
, must be organised a super-database with sub-databases for each group, and time series for each variables within each group:
d.Group1_Name.Var1_Name
d.Group1_Name.Var2_Name
...
d.Group2_Name.Var1_Name
d.Group2_Name.Var2_Name
...