cforecast: Conditional forecasts

In joergrieger/bvar: Estimation and forecasting of bayesian VAR models

Description

Conditional forecasts

Usage

cforecast(obj, forecastHorizon, id_obj, cfconds, interval = c(0.05,
  0.95), ...)

## S3 method for class 'bvar'
cforecast(obj, forecastHorizon, id_obj, cfconds,
  interval = c(0.05, 0.95), ...)

Arguments

obj	Estimated model
forecastHorizon	forecast horizon
id_obj	Identification
cfconds	An h\times n matrix containing the forecast conditions with h being the forecast horizon and n the number of variables in the VAR. Unconstrained variables should be NA.
interval	forecast bands
...	not used

Details

Conditional forecasts are forecasts conditional on given values for a subset of variables and are obtained by pre-determining the path of certain variables. Waggoner and Zha (1999) show that the distribution of future shocks is normal with

η\sim N(\bar{η},\barΓ)

where

\bar{η}=R'(RR')^{-1}r

and

Γ=I-R'(RR')^{-1}R

with η the s\times1-vector of structural shocks and r is the vector of differences between predicted and conditional values. Instead of drawing directly from the above distribution we use a singular value decomposition of R as proposed by

Value

returns an S3 object of the class fcbvar

References

Waggoner, Daniel F. and Tao Zha, Conditional Forecasts in Dynamic Multivariate Models, The Review of Economics and Statistics, Vol. 81, No. 4 (Nov 1999), pp. 639-651

Jarocinski, M., Conditional forecasts and uncertainty about forecast revisions in vector autoregressions, Economics Letters, Vol. 25, No. 3 (2010), pp. 257-259

joergrieger/bvar documentation built on July 3, 2020, 5:34 p.m.