Display the significance of T-stat in a report

[Aktar]

Hi,

I would like to display in a report the signifiance of my betas coefficient, indeed the signifiance associated to the T-stat.

I tried %signif and with other %**signif functions but it doesn't work.

Thank you.

[TomDoan]

You can do that with %ttest(%tstats(i),%ndf) for coefficient i.

[asmith05]

Hi Tom,

Quick question. This works great for homoscedastic standard errors.

See below:

Code: Select all

Linear Regression - Estimation by Least Squares
Dependent Variable SLOPE
Daily(5) Data From 1999:01:04 To 2012:02:29
Usable Observations                       107
Degrees of Freedom                        106
Skipped/Missing (from 3433)              3326
Centered R^2                       -0.0000000
R-Bar^2                            -0.0000000
Uncentered R^2                      0.0075226
Mean of Dependent Variable       -0.004531168
Std Error of Dependent Variable   0.052290697
Standard Error of Estimate        0.052290697
Sum of Squared Residuals         0.2898376003
Log Likelihood                       164.4262
Durbin-Watson Statistic                1.7741

    Variable                        Coeff      Std Error      T-Stat      Signif
************************************************************************************
1.  Constant                     -0.004531168  0.005055132     -0.89635  0.37209696

Use '%ttest(%tstats(1),%ndf)' to calculate p-value       0.37210

However, RATS seems to be doing something different when I use robust standard errors. In particular, the "Signif" value in the regression output differs from my calculation. See below:

Code: Select all

Linear Regression - Estimation by Least Squares
With Heteroscedasticity-Consistent (Eicker-White) Standard Errors
Dependent Variable SLOPE
Daily(5) Data From 1999:01:04 To 2012:02:29
Usable Observations                       107
Degrees of Freedom                        106
Skipped/Missing (from 3433)              3326
Centered R^2                       -0.0000000
R-Bar^2                            -0.0000000
Uncentered R^2                      0.0075226
Mean of Dependent Variable       -0.004531168
Std Error of Dependent Variable   0.052290697
Standard Error of Estimate        0.052290697
Sum of Squared Residuals         0.2898376003
Log Likelihood                       164.4262
Durbin-Watson Statistic                1.7741

    Variable                        Coeff      Std Error      T-Stat      Signif
************************************************************************************
1.  Constant                     -0.004531168  0.005031455     -0.90057  0.36781793

Use '%ttest(%tstats(1),%ndf)' to calculate p-value       0.36986

Could you shed some light on what RATS is doing to calculate the "Signif" values when using the "robust" option?

Many thanks!
Lee Smith

[TomDoan]

It's a N(0,1). Use %ztest(%tstats(i)) to reproduce exactly the information from the output.