Yes,
I’m referring to time dummy variables to control for a (previously) known (or unknown) structural break at a specific point um time, whatever the frequency of our dataset – daily, weekly, monthly,….
In particular, consider the fooling equation to be estimated
y(t)=a0+b0t+a1DM+b1DT+x(t)
where t is time, ai and bi (with i=0,1) are parameters, y and x are observed variables – for example, inflation and unemployment, – and DM and DT are the needed dummies. The dummies take the following values: DM =1 if t>TB and DM=0, otherwise, and DT = (t-TB) if t>TB and DM=0 otherwise. Note that TB is the time of a possible structural change.
Clearly, it is quite simple to generate and include these time dummies in our dataset using excel format files.
My point is how to generate these dummies using gretl.
Regards
José
On Sat, 5 Dec 2015, Riccardo (Jack) Lucchetti wrote:
On Fri, 4 Dec 2015, Allin Cottrell wrote:
If you want a dummy variable for each year (unusual, but hey) you can do
list DT = dummify(time)
Not so unusual with panel datasets, in fact,
True, and for that reason the "Time dummies" option would be active under the Add menu in the panel data case.
Apparently José is talking about annual time series data, and wishes to construct "custom" dummies for particular periods. This is not an automatic action but is easy enough. Example
series WW2 = year > 1939 && year < 1946
If you don't have a "year" variable you can generate one as, e.g.,
series year = time + 1979
for data that start in 1980.
Allin Cottrell