This study did not appear to control for other variables very well
For instance, it has been known for decades that people are moving from the rust belt to the sun belt for a combination of reasons, but mainly due to the decline in our industrial employment base. But thats also likely a high tax to low tax move, for other reasons. How dbaseid the study factor that in?
Also, as cosmo pointed out, the mass of hispanic immigration has swelled the ranks of lower tax southern/border states who have been gaining migrant populations.
If there was significant movement from higher tax border states or higher tax sun belt states, then youd have a possible correlation with an actual causal connection.