set industry dummies

[Yiyang Weng]

Dear Statalist,

In my panel data, I want to control for time dummies and industry dummies. Industry dummies are set according to 2-digit sic code(1-99). However, my sic codes are 4 digit. How do I set industry dummies according to the first two digits sic code? Do you have any suggestions for the code?
For example, when sic code is 3826, the first two digits code is 38, I set industry dummies based on the first two digit sic code.

Also, when including time dummies, it is ok to use code: reg totalstr femaleratio outsiders1 roe logasset RDintensity risk i.year, robust.
i.year represents year dummies?

Thanks for any helps.

* Example generated by -dataex-. For more info, type help dataex
clear
input double year str10 Ticker float(totalcsr totalstr totalcon)    double NumberDirectors float(femaleratio    outsiders1)    str6    gvkey    str4    sic    float(logasset    roe    RDintensity    risk)
2002 "A"     3  6 -3  8 .125      .875 "126554" "3826"  9.012256    -.2230387   .19450915    .14019261
2007 "A"     9 11 -2  9 .111  .8888889 "126554" "3826"  8.929832    .1972789   .12638377    .27627748
2008 "A"     6  8 -2  9 .111  .8888889 "126554" "3826"  8.914223    .2708089   .12192588     .2857335
2009 "A"     6  8 -2  9 .111  .8888889 "126554" "3826"  8.937481    -.012370312    .1432716     .3815029
2010 "A"     4  6 -2  8 .125      .875 "126554" "3826"  9.179469    .21137206   .11202636    .22586633
2011 "A"     5  5  0  9 .111  .8888889 "126554" "3826"  9.111293    .23447637   .09794748    .21331567
2012 "A"     2  2  0  9 .111  .8888889 "126554" "3826"  9.262553    .22237223    .0975503    .20045558
2013 "A"     7  7  0  9 .111  .8888889 "126554" "3826"   9.27669    .13688788   .10395163    .25257346
2004 "AAON"  0  0  0  8    0      .875 "021542" "3585"   4.65612    .10567506  .006186983   .001587045
2005 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  4.732736    .14418517 .0090769185 .00051933876
2006 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  4.867965    .18705782  .008528471            0
2007 "AAON"  0  0  0  6    0  .8333333 "021542" "3585"  4.921002    .2426745  .009458435            0
2008 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  4.946936    .29619154   .00921262            0
2009 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  5.051208    .2349257  .012638514            0
2010 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  5.076903    .1875466  .014720796            0
2011 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"   5.18728    .1141677    .0180302            0
2012 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  5.265241    .19870996   .01187672            0
2013 "AAON"  0  0  0  7    0  .8571429 "021542" "3585"  5.372701    .22879724  .016192315            0
2002 "AAP"   0  0  0 11 .091  .9090909 "145977" "5531"  7.583362    .13882388           0      .368829
2004 "AAP"   0  0  0  7 .143  .8571429 "145977" "5531"  7.697104    .26025763           0    .19904976
2006 "AAP"   0  0  0 10   .1        .9 "145977" "5531"  7.894572    .22439453           0     .1778717
2007 "AAP"  -2  0 -2  9 .111  .8888889 "145977" "5531"  7.939361    .23277804           0    .18002144
2009 "AAP"  -2  0 -2  9 .111  .8888889 "145977" "5531"  8.030397    .21083935           0    .06603626
2010 "AAP"   0  0  0  9 .333  .8888889 "145977" "5531"  8.117973    .3329437           0    .08969336
2012 "AAP"   0  0  0 10   .2        .9 "145977" "5531" 8.4368105    .3202048           0    .13101113
2013 "AAP"   0  0  0  9 .111  .8888889 "145977" "5531"  8.624211    .25838062           0    .18916635
2002 "AAPL"  1  1  0  5    0        .8 "001690" "3663"  8.747988    .015873017   .07784744    .05017466
2003 "AAPL"  0  0  0  6    0  .8333333 "001690" "3663"  8.826881    .016339095   .07588207            0
2004 "AAPL"  1  1  0  8    0      .875 "001690" "3663"  8.993427    .05437352    .0590651            0
2005 "AAPL"  2  2  0  7    0  .8571429 "001690" "3663"  9.354527    .1788106   .03833178            0
2006 "AAPL"  2  2  0  8    0      .875 "001690" "3663"  9.752955    .19921875   .03686254            0
2007 "AAPL"  1  2 -1  7    0  .8571429 "001690" "3663" 10.140415    .24057253   .03257519            0
2008 "AAPL" -1  1 -2  8 .125      .875 "001690" "3663" 10.585877    .2298621   .03414514            0
2009 "AAPL" -1  1 -2  7 .143  .8571429 "001690" "3663" 10.768506    .26027182   .03106864            0
2010 "AAPL" -1  5 -6  6 .167  .8333333 "001690" "3663"  11.22768    .2932142  .027320813            0
2011 "AAPL"  1  4 -3  8 .125      .875 "001690" "3663" 11.664538    .3383411  .022439007            0
2012 "AAPL" -2  4 -6  8 .125      .875 "001690" "3663" 12.078603    .3530412   .02160273            0
2013 "AAPL"  2  8 -6  8 .125      .875 "001690" "3663" 12.240474    .2997758   .02618337    .08193237
2009 "AATI"  0  0  0  6    0  .8333333 "163679" "3674"  4.991412    -.09597123    .3175051            0
2010 "AATI"  0  0  0  7    0  .8571429 "163679" "3674"  4.933193    -.10421536    .3224078            0
2009 "ABAT" -1  0 -1 11 .273 .54545456 "002831" "3690"  5.061493    .1619016  .005474969            0
2010 "ABAT"  0  0  0 11 .273 .54545456 "002831" "3690"  5.427581    .17897487  .002110595            0
2004 "ABAX" -1  0 -1  6    0  .8333333 "024888" "3845"  4.125488    .4403907   .10148484 .00025848977
2005 "ABAX" -1  0 -1  6    0  .8333333 "024888" "3845" 4.2628064    .07866444   .09761553            0
2006 "ABAX" -1  0 -1  6    0  .8333333 "024888" "3845"   4.41978    .10522537   .08888986            0
2007 "ABAX"  0  0  0  6    0  .8333333 "024888" "3845"  4.631958    .11471097   .07167628            0
2008 "ABAX"  0  0  0  6    0  .8333333 "024888" "3845"  4.794989    .11947558   .06927828            0
2009 "ABAX"  0  0  0  6    0  .8333333 "024888" "3845"  4.946708    .09467106   .07920464            0
2010 "ABAX"  0  0  0  5    0        .8 "024888" "3845"  5.122868    .08852018    .0858081            0
2011 "ABAX"  0  0  0  6    0  .8333333 "024888" "3845"  5.237824    .08619729   .08333334   .003962605
2012 "ABAX"  1  1  0  6    0  .8333333 "024888" "3845"  5.203105    .08192884   .07820123   .004306078
2013 "ABAX"  1  1  0  6    0  .8333333 "024888" "3845"  5.307094    .15584525  .072984815  .0033802036
2013 "ABBV"  1  4 -3  9 .111  .8888889 "016101" "2836" 10.281856    .9189671    .1699308     .4894856
2004 "ABC"  -3  0 -3  9 .111  .7777778 "031673" "5122"  9.363405    .10794771           0    .09928872
2005 "ABC"  -3  0 -3 10   .1        .8 "031673" "5122"  9.339716    .06182779           0    .08360113
2006 "ABC"  -2  0 -2 10   .1        .8 "031673" "5122"  9.455943    .11294283           0    .08557086
2007 "ABC"  -2  0 -2  9 .111  .8888889 "031673" "5122"  9.418173    .15135786           0    .09969875
2008 "ABC"  -1  0 -1  9 .111  .8888889 "031673" "5122"  9.410648    .09245566           0    .09718717
2009 "ABC"  -1  0 -1  8 .125      .875 "031673" "5122"  9.515819    .185313           0      .086713
2010 "ABC"  -1  0 -1  9 .222  .8888889 "031673" "5122"    9.5774    .21553284           0    .09304971
2011 "ABC"  -1  0 -1  9 .222  .8888889 "031673" "5122"   9.61465    .2464803           0    .06493255
2012 "ABC"   0  0  0  9 .222  .8888889 "031673" "5122"  9.644984    .2926619           0    .09367768
2013 "ABC"   1  1  0 10   .2        .9 "031673" "5122"  9.847903    .1869632           0     .0738217
2013 "ABC"   1  1  0 10   .2        .9 "031673" "5122"  9.847903    .1869632           0     .0738217
2013 "ABC"   0  0  0 10   .2        .9 "031673" "5122"  9.847903    .1869632           0     .0738217
2010 "ABCD"  0  0  0  8    0       .75 "176939" "7370"  5.948197    -.12861659   .05824782    .42595455
2011 "ABCD"  0  0  0  9    0  .7777778 "176939" "7370"  5.912638    -.54891753    .0576635    .50437945
2007 "ABG"  -2  0 -2 11 .182  .9090909 "144640" "5500"  7.609019    .0872181           0    .23501016
2008 "ABG"  -2  0 -2 11 .182  .8181818 "144640" "5500"  7.411133    -1.517737           0     .3314393
2009 "ABG"  -2  0 -2 11 .182  .8181818 "144640" "5500"   7.24487    .05500821           0     .3774716
2010 "ABG"  -1  0 -1 11 .182  .8181818 "144640" "5500"  7.304045    .13270637           0      .359887
2011 "ABG"  -1  0 -1 10   .2        .9 "144640" "5500"  7.257989    .20789957           0     .3093561
2012 "ABG"   0  0  0  9 .222  .7777778 "144640" "5500"  7.415416    .2040715           0     .2777176
2013 "ABG"   2  2  0  9 .222  .7777778 "144640" "5500"  7.543591    .22238076           0     .2876734
2013 "ABG"   0  2 -2  9 .222  .7777778 "144640" "5500"  7.543591    .22238076           0     .2876734
2009 "ABII"  0  0  0  8    0      .625 "178855" "2836"  6.973899    -.12184836    .4306225            0
2004 "ABMD"  1  1  0  8 .125      .875 "013619" "3841" 4.0802627    -.17384423    .5391429            0
2005 "ABMD"  1  1  0  9 .111  .7777778 "013619" "3841"  4.111873    -.04168817    .3458761            0
2006 "ABMD"  1  1  0  9 .111  .8888889 "013619" "3841" 4.3635697    -.4237998    .6880009            0
2007 "ABMD"  1  1  0  8 .125      .875 "013619" "3841"     4.914    -.22835498    .4525854            0
2008 "ABMD"  0  0  0  8 .125      .875 "013619" "3841"  4.770947    -.4373144    .4171629            0
2009 "ABMD"  0  0  0  9 .111  .8888889 "013619" "3841"  4.912346    -.27242786   .33942086            0
2010 "ABMD"  0  0  0  9 .111  .8888889 "013619" "3841"  4.864221    -.17621994   .29673445            0
2011 "ABMD"  0  0  0  8 .125      .875 "013619" "3841"  4.879676    -.11222707   .25442162            0
2012 "ABMD"  0  0  0  7 .143  .8571429 "013619" "3841"  5.036375    .011837178    .2034817            0
2013 "ABMD"  1  1  0  7 .143  .8571429 "013619" "3841"  5.135793    .10952728    .1621955            0
2004 "ABR"   0  0  0  7    0  .7142857 "160173" "6798"  6.815963    .09542422           0    .18170877
2005 "ABR"   0  0  0  9 .111  .7777778 "160173" "6798"   7.24142    .1751927           0     .3705381
2006 "ABR"   0  0  0  9 .111  .7777778 "160173" "6798"  7.698186    .17025372           0     .6246259
2007 "ABR"   0  0  0 10   .1        .8 "160173" "6798"  7.972981    .21386772           0     .6591153
2008 "ABR"  -1  0 -1 10   .1        .8 "160173" "6798"  7.855248    -.28906855           0    .55695254
2009 "ABR"  -1  0 -1 10   .1        .8 "160173" "6798"  7.630837    -2.3382268           0     .8148192
2010 "ABR"   0  0  0 10   .1        .8 "160173" "6798"  7.456575    .5470048           0     .7250463
2011 "ABR"   0  0  0 10   .1        .8 "160173" "6798"  7.482521    -.2329352           0     .6160958
2012 "ABR"   0  0  0  8 .125      .875 "160173" "6798"  7.439489    .09297287           0     .5821112
2001 "ABT"  -3  1 -4 14 .143  .7857143 "001078" "3845" 10.056055    .17113546    .1785636    .18610123
2002 "ABT"  -2  1 -3 14 .143  .7857143 "001078" "3845" 10.096547    .26196158   .09440338     .1761802
2003 "ABT"  -3  1 -4 13 .154  .7692308 "001078" "3845" 10.192993    .2106165   .09317376    .12922646
2004 "ABT"  -6  1 -7 13 .154  .7692308 "001078" "3845"    10.267    .22587603   .10039417    .16643555
2005 "ABT"  -5  1 -6 14 .143  .7857143 "001078" "3845" 10.279908    .2339231   .08248034    .15687424
end

[Clyde Schechter]

To get the first two digits of the sic code, all you need is:

gen sic2 = substr(sic, 1, 2)
destring sic2, replace

I have taken the second step of converting the two digits into an actual numeric variable so that it can be easily used as a factor variable (set of "dummy" variables) in your modeling.

As for the regression you are proposing, I see problems. I'm not sure what your panel variable is. None of the obvious candidates (Ticker, gvkey) jointly identify unique observations with year. In fact, you have some observations in your example data that are complete duplicates on all of the variables. Fully duplicate observations like that are usually data errors in any data set. And in a panel data set, you must never have two observations that have the same values for the panel and time variable. So I think the data set you are working with needs some rehab before you can use it for anything. And remember that a data set that contains one error is likely to contain others you haven't found yet. So don't just do some quick fix of this immediate problem: carefully review how the data set was created and scrutinize it for other possible errors. Better to find them now than when you are doing a presentation in front of a group and somebody in the audience points out that some of your results can't possibly be right.

Once you get the data fixed, your regression equation doesn't take into account the nesting of observations within panels (whatever your panel is). That needs to be accounted for. You should be using -xtreg-, not -reg- in order to do that. (And that in turn requires you to -xtset- your data first.) Moreover, you appear to be working with economic or financial data, and in those disciplines one ordinarily uses cluster robust variance estimation unless the data has too few panels to do that.

[Yiyang Weng]

Thanks, Clyde.

I will fix my duplicates issue.