How to create big firm's sales share by year & industry

[Chul-Kyoo Jung]

Hi, I am trying the create a new variable called `big_firm_share`, which is computed as:

the sum(sales of all big firms)/sum(sales of all firms) by year and industry together.

Big firm is defined if lbg30 ==1, and industry variable is ksic1.

I want to make sure if this is a correct way of doing so and how to complete the code. Thank you.

sort year ksic1
by year ksic1 : egen big_firm_share= ....

* Example generated by -dataex-. To install: ssc install dataex
clear
input str6 firmid float(year lbg30 sales) str58 ksic1
"000020" 1991 0   8.8382e+10 "manufacturing"                            
"000020" 1992 0  1.02573e+11 "manufacturing"                            
"000020" 1993 0  1.14395e+11 "manufacturing"                            
"000020" 1994 0  1.19775e+11 "manufacturing"                            
"000020" 1995 0  1.28408e+11 "manufacturing"                            
"000020" 1996 0   1.3484e+11 "manufacturing"                            
"000020" 1997 0  1.43815e+11 "manufacturing"                            
"000020" 1998 0  1.41186e+11 "manufacturing"                            
"000020" 1999 0  1.36299e+11 "manufacturing"                            
"000020" 2000 0   1.3823e+11 "manufacturing"                            
"000020" 2001 0  1.26424e+11 "manufacturing"                            
"000020" 2002 0  1.34147e+11 "manufacturing"                            
"000020" 2003 0  1.34383e+11 "manufacturing"                            
"000020" 2004 0  1.23012e+11 "manufacturing"                            
"000020" 2005 0  1.37962e+11 "manufacturing"                            
"000020" 2006 0  1.52822e+11 "manufacturing"                            
"000020" 2007 0  1.48739e+11 "manufacturing"                            
"000020" 2008 0   1.7512e+11 "manufacturing"                            
"000020" 2009 0  1.88608e+11 "manufacturing"                            
"000020" 2010 0  2.15279e+11 "manufacturing"                            
"000040" 1991 0   8.7739e+10 "manufacturing"                            
"000040" 1992 0    9.399e+10 "manufacturing"                            
"000040" 1993 0   8.4703e+10 "manufacturing"                            
"000040" 1994 0   6.8227e+10 "manufacturing"                            
"000040" 1995 0   5.4498e+10 "manufacturing"                            
"000040" 1996 0  1.72138e+11 "manufacturing"                            
"000040" 1998 0   8.4769e+10 "manufacturing"                            
"000040" 2001 0  1.00538e+11 "manufacturing"                            
"000040" 2002 0   9.2186e+10 "manufacturing"                            
"000040" 2003 0   7.9167e+10 "manufacturing"                            
"000040" 2005 0  1.07786e+11 "manufacturing"                            
"000040" 2006 0  1.13905e+11 "manufacturing"                            
"000040" 2007 0  1.33963e+11 "manufacturing"                            
"000040" 2008 0  1.14734e+11 "manufacturing"                            
"000040" 2010 0  1.44521e+11 "manufacturing"                            
"000050" 1991 0  1.16072e+11 "retail"                  
"000050" 1992 0  1.16114e+11 "retail"                  
"000050" 1993 0  1.15466e+11 "retail"                  
"000050" 1994 0  1.50682e+11 "retail"                  
"000050" 1995 0  1.78967e+11 "retail"                  
"000050" 1996 0  1.76528e+11 "retail"                  
"000050" 1997 0  1.93152e+11 "retail"                  
"000050" 1998 0   2.2036e+11 "retail"                  
"000050" 1999 0  2.15739e+11 "retail"                  
"000050" 2000 0  1.72513e+11 "retail"                  
"000050" 2001 0  1.78381e+11 "retail"                  
"000050" 2002 0  1.91323e+11 "retail"                  
"000050" 2003 0  1.85753e+11 "retail"                  
"000050" 2004 0  1.73897e+11 "retail"                  
"000050" 2005 0  1.63968e+11 "retail"                  
"000050" 2007 0  1.52194e+11 "retail"                  
"000050" 2009 0  1.77298e+11 "retail"                  
"000050" 2010 0  2.72271e+11 "retail"                  
"000070" 1991 0  6.52292e+11 "service"
"000070" 1992 0  6.98666e+11 "service"
"000070" 1993 0  7.24981e+11 "service"
"000070" 1994 0  7.69685e+11 "service"
"000070" 1995 0 1.117635e+12 "service"
"000070" 1996 0 1.144018e+12 "service"
"000070" 1997 0 1.128696e+12 "service"
"000070" 1998 0 1.368467e+12 "service"
"000070" 1999 1 1.206235e+12 "service"
"000070" 2000 0  1.20507e+12 "service"
"000070" 2001 0  9.33729e+11 "service"
"000070" 2002 0  7.79117e+11 "service"
"000070" 2003 0    7.877e+11 "service"
"000070" 2004 0  8.90163e+11 "service"
"000070" 2005 0  9.84375e+11 "service"
"000070" 2006 0  1.17002e+12 "service"
"000070" 2007 0 1.145461e+12 "service"
"000070" 2008 0 1.308486e+12 "service"
"000070" 2009 0 1.495948e+12 "service"
"000070" 2010 0 1.665497e+12 "service"
"000080" 1991 1  1.91285e+11 "manufacturing"                            
"000080" 1992 1  2.17408e+11 "manufacturing"                            
"000080" 1993 1  2.55228e+11 "manufacturing"                            
"000080" 1994 1  3.02197e+11 "manufacturing"                            
"000080" 1995 1  4.11659e+11 "manufacturing"                            
"000080" 1996 1  4.36697e+11 "manufacturing"                            
"000080" 1997 1  4.89055e+11 "manufacturing"                            
"000080" 1998 1  4.62326e+11 "manufacturing"                            
"000080" 1999 1   5.1305e+11 "manufacturing"                            
"000080" 2000 1  5.21251e+11 "manufacturing"                            
"000080" 2001 0  5.47127e+11 "manufacturing"                            
"000080" 2002 0  5.90404e+11 "manufacturing"                            
"000080" 2009 0  6.81861e+11 "manufacturing"                            
"000080" 2010 0  7.05579e+11 "manufacturing"                            
"000100" 1991 0   1.1015e+11 "manufacturing"                            
"000100" 1992 0  1.27698e+11 "manufacturing"                            
"000100" 1993 0  1.36927e+11 "manufacturing"                            
"000100" 1994 0  1.48019e+11 "manufacturing"                            
"000100" 1995 0  1.68781e+11 "manufacturing"                            
"000100" 1996 0  1.88294e+11 "manufacturing"                            
"000100" 1997 0    2.029e+11 "manufacturing"                            
"000100" 1998 0   2.0052e+11 "manufacturing"                            
"000100" 1999 0  1.88452e+11 "manufacturing"                            
"000100" 2000 0  2.20495e+11 "manufacturing"                            
"000100" 2001 0   2.5899e+11 "manufacturing"                            
"000100" 2002 0  2.84982e+11 "manufacturing"                            
"000100" 2003 0  3.06596e+11 "manufacturing"                            
end

[Nick Cox]

You've created a set for the movie. You just need some actors. Consider

egen double numer = total(lbg30 * sales), by(ksic1 year)
egen double denom = total(sales), by(ksic1 year)

and then it should be clear what to do next.

[Chul-Kyoo Jung]

Nick Cox thanks for your answers and important points. Using those codes, I want to figure out two things: which industry has high dominance of big firm in terms of sales and then regress on big firm share by controlling other covariates. Then is the next step should be like the following?

 egen big_firm_share = numer/denom

Or alternatively, how about doing without by industry option and then later consider them by industry, like the following?

egen double numer = total(lbg30 * sales), by(year)
egen double denom = total(sales), by(year)
egen big_firm_share = numer/denom
sort big_firm_share, by(ksic1)

[Nick Cox]

Your sort command is illegal. but otherwise it's up to you. Different questions have different answers.