05/31/2026
“Creating the Gerrymandering Partisan Index”
“The GPI tracks how far seats deviate from the votes that produced them. Applied to every House election since 1976, the GPI tells a clear story: three decades of rough proportionality, a sustained upward turn after 2008, and in 2026 a single-cycle jump driven not by changes in how voters voted but by changes in how districts were drawn. The projection for the post-2030 redistricting cycle points to a country whose democracy no longer resembles the one Americans have known. The shift is no longer just qualitative. It is measurable, and the measurement is sobering.
“The GPI counts seats won by the majority party beyond what closest-to-proportional rounding would award, summed across all states and expressed as a share of the maximum possible for each cycle on a 0-to-100 scale.
“Imagine a state with 10 districts and a 61 to 39 vote split. In this scenario, proportional representation would deliver six or even seven seats to the majority party. Any other outcome would depart from the principle of proportional representation. The further the gap between how seats are distributed and how people voted, the larger the distortion. Importantly, distortion in either direction counts: a Republican-engineered Tennessee and a Democratic-engineered Maryland both add to the total amount of gerrymandering.”
“How the GPI has changed over time”
“Since 1976, the GPI has ranged from 5 in 1986 to 34 in 2024. In order to get a general baseline for how much distortion is ‘normal,’ thresholds were set at one and four standard deviations above the average levels from 1976 to 2008, the period preceding the post-2010 rise in partisan gerrymandering. Then, I divided the scale into three tiers of escalating partisan distortion: proportional representation, partisan skew, and single-party dominance.
“From 1976 through 2008, the GPI was low and stable, averaging 13, but the composition shifted markedly over time. In 1976, 97% of displaced seats over-represented Democrats, and in 1978 and 1988, 100% did. This is the historical asymmetry that today's Republican rebalancing argument cites. The 1994 Republican Revolution offset that asymmetry through votes rather than maps: Republicans won the House for the first time in 40 years, taking 53% of the seats on 54% of the vote, a proportional outcome. That equilibrium was sustained not by the courts or the Voting Rights Act but by behavior. Both parties chose to operate within a tighter latitude than the formal rules required. The arrangement reflected a Madisonian premise: that legislative majorities should be checked and minority political interests deserve representation.
“Then, things shifted in 2010. After Republicans won state legislative majorities in roughly a dozen states and used new precinct-level mapping technology, the 2012 GPI jumped to 28 and never returned to pre-2010 levels. Democrats followed with aggressive maps in Illinois, Maryland, and (until courts struck them down) New York. The partisan shares returned to near-parity, but the total stayed elevated. The system escalated symmetrically rather than rebalancing.
“Now, the country is experiencing a significant increase in gerrymandering. The 2026 election cycle is projected to produce a GPI of 45, a 30% jump, driven entirely by new maps rather than any change in voter behavior. It is the largest one-cycle move in the modern record, the product of several developments converging in rapid succession.”
America needs a Gerrymandering Partisan Index. Here’s what it reveals.