Blank Stata output for D-in-D analysis

[Sabrina K.]

Hello,
I am attempting to run a difference-in-difference regression, however, once I add more than 3 variables to the command, the Stata output comes out blank with dots.

reg FemaleHousePoverty Treated Time DiD CrimeRate UninsuredRate MinWage FemaleHouseholds

      Source |       SS           df       MS      Number of obs   =         8
-------------+----------------------------------   F(7, 0)         =         .
       Model |   322.86875         7  46.1241071   Prob > F        =         .
    Residual |           0         0           .   R-squared       =    1.0000
-------------+----------------------------------   Adj R-squared   =         .
       Total |   322.86875         7  46.1241071   Root MSE        =         0

-------------------------------------------------------------------------------
FemaleHouse~y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
--------------+----------------------------------------------------------------
      Treated |    19.1934          .        .       .            .           .
         Time |  -.8727765          .        .       .            .           .
          DiD |  -2.757518          .        .       .            .           .
    CrimeRate |   .0018898          .        .       .            .           .
UninsuredRate |   4.377351          .        .       .            .           .
      MinWage |  -.4465446          .        .       .            .           .
FemaleHouse~s |   .4179356          .        .       .            .           .
        _cons |  -15.02389          .        .       .            .           .
-------------------------------------------------------------------------------

However, when I only have three controls, Stata gives me a complete output:

reg FemaleHousePoverty Treated Time DiD CrimeRate UninsuredRate MinWage

      Source |       SS           df       MS      Number of obs   =         8
-------------+----------------------------------   F(6, 1)         =   2610.53
       Model |  322.848138         6   53.808023   Prob > F        =    0.0150
    Residual |  .020611954         1  .020611954   R-squared       =    0.9999
-------------+----------------------------------   Adj R-squared   =    0.9996
       Total |   322.86875         7  46.1241071   Root MSE        =    .14357

-------------------------------------------------------------------------------
FemaleHouse~y | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
--------------+----------------------------------------------------------------
      Treated |   23.20011   .8911143    26.03   0.024     11.87743    34.52279
         Time |  -.6343996   .2285763    -2.78   0.220    -3.538737    2.269938
          DiD |   -3.11115   .2292377   -13.57   0.047    -6.023891   -.1984093
    CrimeRate |   .0045527   .0010012     4.55   0.138    -.0081682    .0172736
UninsuredRate |   4.162944   .3877846    10.74   0.059    -.7643268    9.090215
      MinWage |   -.771734   .1254689    -6.15   0.103    -2.365967    .8224993
        _cons |  -3.240574   3.376672    -0.96   0.513    -46.14526    39.66411
-------------------------------------------------------------------------------

Does anyone know why this may be? It is not only with the control variable FemaleHouseholds. I have about six controls variables, and whatever combination I use of them, Stata will not give me a complete output if I use more than three.
Furthermore, it sometimes omits Treated due to collinearity, as shown below with a different outcome variable.

 reg DeepPovertyRate Time Treated DiD CrimeRate FemaleHouseholds FemaleWelfareF
> am UninsuredRate UnemploymentRate
note: Treated omitted because of collinearity.

      Source |       SS           df       MS      Number of obs   =         8
-------------+----------------------------------   F(7, 0)         =         .
       Model |  28.2740857         7   4.0391551   Prob > F        =         .
    Residual |           0         0           .   R-squared       =    1.0000
-------------+----------------------------------   Adj R-squared   =         .
       Total |  28.2740857         7   4.0391551   Root MSE        =         0

-------------------------------------------------------------------------------
DeepPoverty~e | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
--------------+----------------------------------------------------------------
         Time |  -.3837457          .        .       .            .           .
      Treated |          0  (omitted)
          DiD |   .2797787          .        .       .            .           .
    CrimeRate |  -.0010357          .        .       .            .           .
FemaleHouse~s |   .5253158          .        .       .            .           .
FemaleWelfa~m |   .0646355          .        .       .            .           .
UninsuredRate |   .8741832          .        .       .            .           .
Unemploymen~e |  -.6889338          .        .       .            .           .
        _cons |  -12.19267          .        .       .            .           .
-------------------------------------------------------------------------------

If it is helpful to answer, here is my data summarized: (Treated=1 if State=0, Treated=0 if State=1; Time=0 if Year= 2016&2017, Time=1 if Year= 2018&2019)

sum

    Variable |        Obs        Mean    Std. dev.       Min        Max
-------------+---------------------------------------------------------
       State |          8          .5    .5345225          0          1
        Year |          8      2017.5    1.195229       2016       2019
CashAssis~nt |          8     30.4475     9.58168      21.23      44.79
DeepPovert~e |          8     4.46875    2.009765       2.68       7.62
CashAssis~mt |          8    1.19e+08    1.99e+07   9.96e+07   1.65e+08
-------------+---------------------------------------------------------
FamDeepPov~t |          8    23845.75    17436.09       5908      41659
Unemployme~e |          8      4.9125    1.027393        3.4        6.1
   CrimeRate |          8     769.875    319.7532        454       1204
 DivorceRate |          8        2.55    .1309307        2.4        2.7
     MinWage |          8    11.33125     2.20615       8.75       15.2
-------------+---------------------------------------------------------
FemaleHous~s |          8    26.59125    6.168518       20.5       34.5
        Time |          8          .5    .5345225          0          1
UninsuredR~e |          8      4.9875    1.285565        3.5        6.5
     Treated |          8          .5    .5345225          0          1
         DiD |          8         .25      .46291          0          1
-------------+---------------------------------------------------------
BlackPover~e |          8     15.3875    6.162893        9.3       23.7
MarriedPov~e |          8      2.5125    .5792544        1.3        3.4
FemaleHous~y |          8     22.7875    6.791473       15.4       32.8
WhitePover~e |          8        2.65    1.376331         .8        4.2
 BlackFemale |          8       23.95    7.035421       15.4       33.5
-------------+---------------------------------------------------------
WelfareFam~e |          8     30.9375     12.4492       18.8       47.8
FemaleWelf~m |          8     40.7875    11.32424       30.6       57.7
MeanIncome~t |          8       11198    905.4095      10044      12568
MeanIncome~e |          8    11876.38    1047.027      10282      13734
HispanicPo~y |          8     10.3375    3.566887        5.7       17.6
-------------+---------------------------------------------------------
    fivekids |          8        42.8    23.84521       15.2       80.4
 female5kids |          8     61.3375    23.30732       29.3       97.4
female3or4~s |          8       49.55    6.082763       43.6       61.8
female1or2~s |          8     27.1625    7.751485       18.8       38.1

.

Thank you for your help in advance,
Sabrina

[Carlo Lazzaro]

Sabrina:
welcome to this forum.
It would seem that you have highly correlated predictors that make fitted=observed values (as you can see, some of your output table do not show any clue of residuals).
In addition, sky-rocketing R-sq with many non-statistical significant predictors is suspect, too.
I would go:

estat vce, corr

[Scott Merryman]

You are trying to estimate 8 or 9 coefficients with only 8 observations.

[Carlo Lazzaro]

Scott is clearly correct.
I totally missed this point (my bad, indeed)
Sabrina: you are expected to go nowhere with such a sample size, no matter the statistical procedure that you've in mind.

[Sabrina K.]

Thank you Carlo and Scott. You are correct, I will add more observations. Thank you for your help!