The latest issue of Legislative Studies Quarterly contains a neat piece (ungated, here) by Jason Kelly on the extent to which state legislators strategically use prison populations for partisan gain in redistricting at the state level. The ingenuity of state mapmakers knows no bounds.
The census data used to redraw legislative districts counts the country's nearly two million prisoners in the location of their incarceration, rather than their previous place of residence. By drawing these phantom populations into districts that lean heavily toward the majority party, legislators can free up eligible voters from those districts to be distributed among neighboring marginal ones, thereby increasing that party's likelihood of winning additional seats in the state legislature. An analysis of state senate districts finds that prison populations shift systematically from districts controlled by one party to districts controlled by the other following a switch in partisan control.