Tobit model with independent variable in lof form

[Angelo Rias]

Hi everyone, hope u are all safe. This is a small part of my database.

I have a quick question.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(hours wage lwage)
 192   12.5  2.525729
1344  48.36  3.878673
1300   2.88 1.0577903
1971  52.77  3.965943
 860  26.16 3.2642314
1748  17.45   2.85934
 936  12.82  2.551006
2080  19.23 2.9564714
2000  128.5  4.855929
2174  18.11  2.896464
   0      0         .
1920   27.6  3.317816
2080  24.04  3.179719
2250  17.78  2.878074
   0      0         .
2210  41.86  3.734331
   0      0         .
2600     10 2.3025851
2600  19.23 2.9564714
 450  88.89 4.4873996
2080   12.5  2.525729
2600  21.15   3.05164
   0      0         .
1810   45.3  3.813307
3920  20.41  3.016025
2340  25.64 3.2441535
2570  13.27  2.585506
2450 132.65  4.887714
2032  14.44  2.670002
1944  18.24  2.903617
3120  38.46 3.6496186
1750  50.17  3.915417
2560  22.29  3.104138
 400  16.25  2.788093
1520  19.74  2.982647
 606    9.9 2.2925348
1612  29.16  3.372798
1275  13.33  2.590017
2262  10.61 2.3617969
2880  69.44 4.2404633
 630   7.94 2.0719132
2100  24.76 3.2092295
1175   5.53  1.710188
1900  21.05 3.0469005
1946  14.68  2.686486
1936  13.95 2.6354795
2794   8.79 2.1736147
2018   22.3 3.1045866
1350  48.15  3.874321
2080   9.13 2.2115657
1748   9.73  2.275214
2223  29.01 3.3676405
1888  20.54  3.022374
1952  14.34  2.663053
2640  38.26  3.644405
2040  20.73  3.031582
3120   5.45 1.6956155
1756  12.53  2.528126
2304  49.91  3.910221
 880  10.23 2.3253245
   0      0         .
1888  68.86 4.2320757
1256  22.29  3.104138
 561   7.13 1.9643112
1845  16.91  2.827905
3000      7   1.94591
1868  14.99 2.7073834
2530  27.15  3.301377
1960  20.41  3.016025
4000      3 1.0986123
1575  68.57  4.227855
1645  12.16  2.498152
 868   9.22  2.221375
1170     18  2.890372
2856  13.11  2.573375
   0      0         .
1470  36.05  3.584907
2205  16.79 2.8207836
   0      0         .
 273   6.96 1.9401795
 376  18.62  2.924236
 828   6.04  1.798404
1631   21.2  3.054001
2600  14.62 2.6823905
1952  10.32 2.3340838
 355  12.65  2.537657
3036  11.86 2.4731715
 426   4.23  1.442202
1500   4.67  1.541159
   0      0         .
3060   5.76 1.7509375
   0      0         .
   0      0         .
2000   13.5   2.60269
1862  28.46 3.3484995
2080   20.9  3.039749
2100   7.86 2.0617867
   0      0         .
1587  21.42  3.064325
1784  46.52 3.8398824
end

I want to estimate a Tobit model since hours of work often are zero for a part of my sample. And for every observation where hours=0 of course wage=0. But since my independent variable- wage( in most of the studies and of course makes sense) is in logarithm, when i run tobit all the zero values of hours are being dropped from the regression since the logarithm of 0 wages is missing value. and iam left with no censored observations.
I want to underline that the tobit model is required from my teacher, and i cannot choose another model.

tobit hours lwage, ll(0)

What can i do in this case. is there any way i can solve this?
Thanks for any help

[FernandoRios]

you need to impute wages before running the tobit model
That is the only way i know of to estimate this type of model.
Otherwise, what you are looking at is a Heckman type of model.

[Angelo Rias]

Thanks! If it is not to much to ask, how can i impute wages?

[FernandoRios]

there is a large literature on that.
But for a class, i suppose a simple linear regression would do.

[Phil Bromiley]

Please read the FAQ on asking questions-we don't generally help with homework assignments.

If zero is a legitimate value in your data, it is seldom a good idea to use the log transform. One can easily use square roots or cube roots and get very similar results (making sure negatives are handled appropriately).