Re: st: re: Automating coefficient plots.

[andreas nordset <andreas.nordset@gmail.com>]

On Sun, Nov 21, 2010 at 10:45 AM, Christopher Baum <kit.baum@bc.edu> wrote:
> <>
> Andreas said
>
> I am studying the pattern of income for households around unemployment
> spell starts.
>
> The dataset is a balanced panel of households from 1995-2007, and I
> use households that are first time hit by unemployment in years
> 1999-2003. That allows me to track them 4 years prior to, and 4 years
> after unemployment incident. In the analysis I generate a variable
> called "relative_year", that tells me how far away from the
> unemployment spell start the observation is.
> E.g. relative_year == 3, means we observe the household in year 3
> after unemployment spell start.
>
> I want to plot the average income development from relative_year -4 to
> relative_year +4, net of calendar year fixed effects
>
> So far I have done this by running the following:
> reg income relative_year-dummies calendar-year-dummies
> , then exporting the results to excel where i plot the
> relative_year-dummies plus the constant to get the plot.
>
> Is there a way to automate this procedure?
>
>
> -----------------------
> webuse grunfeld, clear
> qui reg invest i.year i.company
> margins i.year, post
> // Roger Newson's parmest from SSC
> parmest, saving(yeareff, replace)
> use yeareff, clear
> // Kit Baum's tsmktim from SSC
> tsmktim year, start(1935)
> tsline estimate || tsrline min95 max95, legend(col(1)) scheme(s2mono)
> -------------------------
>


Just a short followup:

the data I analyse are on a secure server, where the only version
available is Stata10. Hence, I cannot just use -ssc install ..... - to
install the userwritten packages above.

I can install these packages manually if i have the "ado" and the
"help" file, but for e.g. -parmest- I am only able to get the version
written for Stata11. Hence while trying to execute these commands,
Stata10 refuses saying they are exclusive to Stata11.
I have searched the web, but am not able to find the packages for
previous versions of Stata to install. Is there any way around this?


Best

Andreas
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/