SE & t-stats for descriptive stats

[Laurel]

If I have monthly stock price data and generate monthly returns (percentage changes) descriptive statistics, how do I get the standard errors, t-stats, and p-values for the mean, standard deviation, and Sharpe ratio = mean/standard deviation? Thank you!

[EViews Gareth]

You'll have to be more descriptive. p-value for what test?

[Laurel]

To determine if the mean, standard deviation, and Sharpe ratio = mean/standard deviation are significantly different from zero. Thank you!

[EViews Gareth]

http://www.eviews.com/help/helpintro.ht ... 23ww204681

That will perform the test, you’ll have to then freeze the results and extract the bits you want.

http://forums.eviews.com/viewtopic.php?f=5&t=18

[Laurel]

Is there a way to use robust standard errors for the test, e.g. heteroskedasticity and autocorrelation adjusted similar to the covariance method options in equation estimation (cov=hac)? Thank you!

[EViews Gareth]

Not built in

[startz]

Testing for the standard deviation different from zero doesn't make any sense. If the sample standard deviation is different from zero, then so is the true standard deviation.

For the mean, try

Code: Select all

ls x c