Importing dummy variables from Excel

[Fabian Rieger]

Dear Statalists,

I am importing a data set on the Duration of sovereign debt restructurings from Excel. In that data set included is a dummy variable that is =1 if the debt restructuring was a so called Brady deal and =0 if not. If I now regress e.g. Duration of debt restructuring = GDP growth in default year + BradyDummy + E, Stata omits all non Brady restructurings (Dummy=0) from the regression, which gives me the effect of GDP growth in default year on Duration of debt restructuring of all Brady restructurings. However, I would like to have the general effect of GDP growth on Duration of debt restructuring in that regression and then see the prolonging or shortening effect of Brady deals through the Brady Dummy in that same regression. I can't see how this creates multicollinarity issues.

Anyone got a hint?

Cheers
Fabian

[Clyde Schechter]

I don't think anybody can help you with this without seeing an example of the data. Do not post your Excel data set. Use -dataex- to post an example of your data from the Stata data set you imported it into. You can install the -dataex- command by running -ssc install dataex-. Then read the simple instructions for using it by running -help dataex-. Use -dataex- every time you post example data.

Without seeing the data, my best guess is that your indicator ("dummy") variable is coded as 1/missing as opposed to 1/0. 1/missing is not a useful coding of dichotomies in Stata as observations with missing values for any model variable are excluded from regression commands. If this is what is happening, you will have to recode the missing values to zeroes.

[Fabian Rieger]

Can't belive I overlooked that. The Dummy was actually coded as 1/missing. Thanks a bunch for your help! Sometimes I do not see the wood for the trees.