The process of redrawing political boundaries, or redistricting, has major implications for our democracy. Redistricting in such a way that favors a party or group is known as gerrymandering. The precision of new social science models has led to increasingly ambitious gerrymandering that undermines the will of the voting majority. On the flip side of the equation, new simulations can also determine the extent to which a district is subjected to abusive redistricting. One such set of models compared legislatively determined districts and judicially determined districts in North Carolina. The results are striking.
Notorious gerrymandering cases in Wisconsin and Maryland -- the former perpetuated by Republicans, the latter by Democrats -- are textbook examples of how to steal an election from the majority, or diminish the voice of the minority. The Wisconsin redistricting was particularly notorious, so much so that the Supreme Court recently took up the case in Gill v. Whitford. Upon capturing the political majority in 2010, Wisconsin's Republican legislators proposed a redistrict that ensured they would lock-in the electoral majority in nearly any electoral scenario. In 2012, Republicans won 60 out of 99 seats in the Wisconsin Assembly despite garnering only 48.6 percent of the two-party state-wide vote.
Gerrymandering is an optimization problem. The two most commonly used gerrymandering techniques are packing and cracking. Packing is an attempt to pack as many opposition voters as possible into a single district. Cracking occurs when opposition voters are spread across multiple districts, and are therefore unable to secure the majority anywhere. Both packing and cracking operate by taking advantage of the majority principle -- regardless the voting makeup, the majority takes the vote. This may be likened to a sort of geographic inequality in which every vote is not equally impactful. In many ways, additionally complex algorithms made it possible for politicians to better crack and pack to their liking.
Utilizing quantitative tools for bias detection should therefore not be dismissed out of principal, as these same tools are already being utilized to exacerbate the problem of gerrymandering. One academic study compared two partisan redistricting plans in North Carolina to a redistricting proposal developed by a panel of judges, then compared all three to a set of 24,000 randomly generated district maps. Researchers found that the redistricting proposal developed by the panel of judges was 75 percent less gerrymandered than the randomly generated maps, which in turn was less gerrymandered than the two redistricting plans developed by the partisan North Carolina legislature.
As demonstrated in 1986 Davis V. Bandemer, the courts have already found themselves able to rule on cases of political gerrymandering. At that time, the courts abstained due to a lack of conclusive evidence that such tampering was occurring. The irony is worth noting -- while the increased complexity of models has increased redistricting abuses, many new models may also provide the courts the evidence they need to strike against gerrymandering.
To think applied mathematics is the problem, or the solution, is folly. Ultimately, a model is a tool, and a tool may be used for good or bad depending on who wields it. Nonetheless, academic findings make clear that bipartisan or independent redistricting commissions, and not partisan redistricting commissions, avoid the heavily engineered political tampering that strikes at the core of voter fairness.
Mahdi Al-Husseini is the volunteer organizer of TEDxDouglasville, a senior at Georgia Tech studying biomedical engineering and public policy, and a U.S. Army cadet.