vector of control variables for panel data

[kazi naher]

i want to know the command to get vector of control variable. I want to get vector of country level control variables - democratic accountability, log gdp, investment freedom and uncertainty index. These are single values for each country. I want to regress stock returns in this and a fixed effect dummy variable.

[kazi naher]

i am attaching the article i am referring to.

[Emna Trabelsi]

* For pooled OLS regression with robust standard errors:

reg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex, robust

* For a panel model with random effects and robust standard errors:

xtreg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex, re robust

To select the appropriate model among the above, you type:

xttest0

If the test is significant, then choose the random effect model and go to the next step.

* Fixed effect versus random effect panel model:

xtoverid

If the test is significant, then the final model should the one with fixed effects and robust standard errors and you type:

xtreg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex, fe robust

Please note that the names of your variables should be the same as they appear in your .dta file.
If you suspect cross sectional dependence, then you can use -xtcd2- to test for it. When there is cross sectional dependence, then you should use -xtscc- command instead of-xtreg-. See Stata Help for xtscc after installing the command if not installed.

ssc install xtscc
help xtscc

Hope this helps and answers your query.

[kazi naher]

i am doing a fixed effect with panel pooled data regression

[kazi naher]

I have only on value for these control variables whereas my stock returns are daily and so are the deaths and confirmed cases. How to i fix that?

[Emna Trabelsi]

So your control variables are either time-invariant or you are working on just one country. To have panel data, you need time-varying observations for each variable (or daily as in the case of the stock returns) by country.

[Emna Trabelsi]

i am doing a fixed effect with panel pooled data regression

If you want to follow the model estimated in the attached paper, you type:

reg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex i.time, robust

[kazi naher]

this is showing r(2001) error as i have only one value for democratic accountability ,gdp, investment freedom and uncertainty avoidance index for each country

[kazi naher]

i am working with 35 countries and the model in the paper is exactly what i am trying to run. But each country has only one value for each data for the control variable. How to i balance them to match with my daily stock returns,confirmed cases and deaths for each country as in the paper table 2?

[Emna Trabelsi]

i am working with 35 countries and the model in the paper is exactly what i am trying to run. But each country has only one value for each data for the control variable. How to i balance them to match with my daily stock returns,confirmed cases and deaths for each country as in the paper table 2?

When you run

reg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex i.time, robust

, what do you obtain? Please you need to correct the names of all variables and of the time variable (I named here it time).

You also need to give an excerpt of your data. Maybe, someone else can help.

[Emna Trabelsi]

Here an example of a (imaginary) dataset of two countries and with 7 daily obervations for each country. Your data should look something like this:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(stockreturns democraticaccountability uncertaintyindex deathcases id time)
  30 5 2  10 1 1
2514 5 2  12 1 2
  12 5 2  36 1 3
 250 5 2 150 1 4
  14 5 2 100 1 5
 140 5 2  10 1 6
  32 5 2 115 1 7
  30 6 3  20 2 1
  25 6 3  40 2 2
  36 6 3 536 2 3
 900 6 3  23 2 4
  40 6 3  50 2 5
 365 6 3  12 2 6
  50 6 3 123 2 7
end

Here, democraticaccountability and uncertaintyindex are time-invariant (one value for each country).

[kazi naher]

here are the excel files - first sheet shows my control variables for each country - the years depending on the latest data available. The second sheet shows the stock returns of major index of each country and the daily confirmed cases and deaths in each country. I would like to know how to I balance them so that the number of observations are equal for all since that is the main error for getting the vector

[kazi naher]

I want to get the command to balance my unbalanced data, that is single observation for control variables to daily observation.

[kazi naher]

When you run

reg stockreturns democraticaccountability loggdp investment freedom uncertaintyindex i.time, robust

, what do you obtain? Please you need to correct the names of all variables and of the time variable (I named here it time).

You also need to give an excerpt of your data. Maybe, someone else can help.

it says omitted for the control variables. I have used DATE as time

[Emna Trabelsi]

it says omitted for the control variables. I have used DATE as time

You need to rearrange your data in your Excel file and obtain a final dataset before running any regression. The problem is in the look of your data. I have no knowledge of a stata command that can balance your data in just one step for instance, but I will see what we can do to achieve the final result without a promise.