Creating summary statistic tables with multiple variables with multiple super-columns.

[Joshua Evans]

Hi all,

I am trying to replicate the results of a paper, and I am having trouble creating nicely formatted tables.

Specifically, Table "3" is in the picture below.

I am unsure if this table came from Stata or not, but I'd like to replicate it in Stata.

As is stands, my code is currently this:

by terc: tabstat fd_s ndi_s nei_s future_fd at_growth                 ///
            future_at_growth firm_age at PPE MB ROA returns_vol Leverage     ///
            Lev_delta Leverage_2 if terc != 2 & terc!= . ,                    ///
            statistics(me sd p25 p75) columns(statistics) format(%10.3g)

For reference:

• terc

  is the super-column variable, splitting the data into terciles based on the probability of a firm having a bond rating. For Table 3, I am only interested in terciles 1 and 3 (excluding tercile 2 and missing values)

•  fd_s ndi_s nei_s future_fd at_growth future_at_growth firm_age at PPE MB ROA returns_vol Leverage Lev_delta Leverage_2

  are the variables of interest (first column in the picture above)

The code above produces the following output:

I essentially want my tables to be side-by-side with the super-column variable "terc" splitting the two tables above for terc == 1 and terc == 3.

Please see .do file attached. For reference, this table refers to lines 708-739.

I appreciate any help and look forward to responses.


Regards,
Joshua Evans