Attached follows a data-file with an equation-object.
* The estimation works nicely over the whole sample
* Static fit works nicely over the whole sample
* Dynamic forecast/fit works only for the first few months.
My question is WHY?
The file differs from my file in that in my file all variables are linked to a sheet with higher frequency, which again is linked to several databases.
The equation object is generated with the following code:
Code: Select all
group xi44_BEER_l log(p_oil(-1)) u_gap_nok(-1)-u_gap_hpc(-1) idiff03m_hpcnok(-1) islopediff_hpcnok(-1) _
log(cpi_idx_nok(-1))-log(cpi_idx_hpc(-1))
group xi44_BEER_d dlog(p_oil) d(u_gap_nok-u_gap_hpc) d(idiff03m_hpcnok) d(islopediff_hpcnok) _
dlog(cpi_idx_nok)-dlog(cpi_idx_hpc)
for %ccy i44 'eurnok
for %model beer '[... several other models]
equation eq_{%model}_{%ccy}.ls(cov="HAC") dlog(s_{%ccy}) c log(s_{%ccy}(-1)) x{%ccy}_{%model}_l x{%ccy}_{%model}_d
next
next
Dagfinn