I need to do a block-of-blocks bootstrap: pairs bootstrap for dynamic
Basically, if one considers the pair Z_t ≡ [y_t, X_t], I need a random
resampling of n-b+1 blocks
(Z_1...Z_b), (Z_2...Z_b+1), ......, (Z_n−b+1...Zn)
with b >= 2.
In Gretl, as far as I know there's an option "resample" in the dataset
command for pairs bootstrap and the option "block" in the command
resample useful for block RESIDUAL bootstrap.
Is there a simple way to do block PAIRS bootstrap?