Block bootstrap codes for time series data

[Maali Alburaies]

I'm looking to do OLS regression for time series data but there is potentially serially correlated errors, So, in term to deal with possible serial correlation errors I should use block bootstrap methods. But really I didn't find any codes for doing block bootstrapping in SATA, and how I can determine the optimal length of the block ?

[Felix Bittmann]

What exact commands are you going to use? Which form of bootstrapping are you looking for? As the Wikipedia article describes, there are several flavours of TS bootstrapping. Depending on what exactly you have in mind you will probably have to adapt the bs-prefix with your own program to resample entire blocks of cases and not single points. Regarding the size of the blocks, I would advise to try several specifications and see what works best for your application.
You might also find more answers here:
https://www.stata.com/statalist/arch.../msg00211.html
https://www.stata.com/statalist/arch.../msg00201.html