topic Re: How to implement a bootstrapped skewness-adjusted t-statistic? in SAS Programming

How to implement a bootstrapped skewness-adjusted t-statistic?

kwil — Tue, 06 Oct 2020 00:16:10 GMT

Hi,

I am trying to implement the bootstrapped skewness-adjusted t-statistic originally developed by Johnson (1978) and outlined in Lyon et al. (1999) to test the significance of my average buy-and-hold abnormal returns. I am just looking to use this for a one sample test, not as a two sample test.

Is this as simple as using the bootstrap method in proc ttest? Any help would be really appreciated.

Cheers.

Re: How to implement a bootstrapped skewness-adjusted t-statistic?

Ksharp — Tue, 06 Oct 2020 11:10:35 GMT

Calling @Rick_SAS

Re: How to implement a bootstrapped skewness-adjusted t-statistic?

Rick_SAS — Tue, 06 Oct 2020 11:26:35 GMT

Yes, you can use the BOOTSTRAP statement in PROC TTEST. Ask for the BC (bias-corrected) interval if you are concerned about skewness. The TTEST doc shows formulas for all bootstrap intervals. For an example of how to use the BOOTSTRAP statement, see "The BOOTSTRAP statement for t tests in SAS" (although the article uses two-sample tests).