Ohio Bipartisan Redistricting Commission Amendment

Ohio Bipartisan Redistricting Commission Amendment

This actually passed by popular referendum in November 2015. It's an interesting bill with a few wrinkles I haven't seen in other "independent commission" model plans. It has a lot more rules that bind that commission to produce a specific map. I guess we'll really know how it goes by September 2021, when the first result of this process is due.

Paragraph 3.B.1 allows for 95% to 105% of the ideal population of a district. This is an important legal precedent (unless it is maybe someday overturned by SCOTUS deciding that a stricter rule would be needed to satisfy "one person, one vote"). My maps hold to plus or minus 0.5% in almost all cases.

Sections 3.C-E define an algorithm! It might be unsatisfiable, but it has an escape clause. If a county happens to have 95% to 105% of the ideal population, it must be kept whole as one district. More populous counties are supposed to contain whole districts and one leftover region that merges with a neighboring district. It ends with, “Where feasible, no county shall be split more than once,” which I expect to be unenforceable.

3.D.2 “Representative districts shall be drawn so as to split the smallest possible number of municipal corporations and townships…” This is an interesting measure. I could automatically count this in my programs.

3.D.3 “… not more than one municipal corporation or township may be split per representative district.” This is another fascinating requirement, and I also expect it will be necessary to break it. But, it is easy to count.

And then there are a bunch of tedious tiebreaker-type rules for when the rules have to be broken.

Section 6.B calls for Fake Proportional Representation! “The statewide proportion of districts whose voters, based on statewide state and federal partisan general election results during the last ten years, favor each political party shall correspond closely to the statewide preferences of the voters of Ohio.”

I could probably write a program to do this whole process automatically. Lower down on Ohio's list of requirements is to "make compact districts." My code can do that quite well. Taking all of these requirements together, there's approximately zero wiggle room for the commission to actually do anything. We may as well save ourselves some money and automate them out of a job. Hire me for six months or a year, and I'll write the program to do it. It'd be a bargain.


  1. Ohio has had a redistricting commission for over 150 years. For about 100 years, they had one of the neatest reapportionment schemes ever.

    Representatives and senators were apportioned to counties, and rounded to the nearest 1/5th, so a county might have 1.4 representatives. Over a 10-year cycle of 5 2-year legislatures, a county apportioned 1.4 representatives, would elect one representative for three legislatures, and two representatives for two legislatures, for an average of 1.4 representatives over the decade.

    Reapportionment was almost automatic. The commission would have to split multi-county districts if one of the counties was entitled to one whole representative or senators. They were so strict on keeping districts, but adjusting the apportionment, that senate districts split the newest county in the state, which had been added after the constitution went into effect (this newest county had been formed from parts of counties that were in different districts).

    The redistricting commission would simply need to crunch the numbers, and make an occasional adjustment to county combinations.

    There were some oddities, such as a county apportioned n.8 representatives would get n+1 representatives. And then around the turn of the century, Mark Hanna got an amendment passed that provided that every county would have at least one representative. This would counter the increasing population in the cities due to industrialization (and would also keep the legislature Republican). Amazingly, this amendment passed with 97% support.

    At the time of the OMOV decisions, they recognized these flaws. There was also concern that at-large elections in large counties might violate the right to vote of minorities (eg Hamilton or Cuyahoga counties might have 8.6 representatives, and elect 8 or 9 representatives from the entire county.

    They panicked over the fractional representatives - figuring that an ambitious judge would overturn it, and switched to the current system.

  2. Most of the rules in the 2015 amendment have been in place since 1967 (see Article XI of the Ohio Constitution - the version on the legislature's web site doesn't have the 2015 amendment yet).

    The old commission had three statewide elected members (governor, secretary of state, and auditor), and two members appointed by the leaders of the two parties in the legislature. This would often lead to 4-1 or 3-2 partisan split.

    The 2015 version adds two more members chosen by legislative leaders, ensuring at least two members from each major party, and required concurrent partisan approval, with the interesting deadlock provisions.

    The provision for keeping the maximum number of whole districts in larger counties, reflects the spirit of the 1851 constitution. Instead of a county having 2.6 representatives, with 2 elected every election, and a 3rd for 3 of 5 elections; 2 whole districts will be formed in the county, and the fraction will be placed in a single district shared with other counties.

    They switched from handling the fraction temporally, to handling it spatially, while switching to single-member districts.

    The part about splitting large counties is mandatory, and is not that hard to practice, and has generally been followed.

    A range of -5% to 5% deviation is per SCOTUS decisions, for ***legislative districts***. The actual limit is 10% total - the decision on which it is based from Texas had deviation ranging from 5.1% to -4.9%.

    The 1967 version, but not the 2015 amendments even had a special rule that a single county with a population from -0.90% to 1.10% of the quota, having its own representative district. This has been upheld in a case involving Ohio.

    The 10% range defined by the SCOTUS is a burden-shifting threshold. If the range is below 10%, the burden is on a challenger to prove it violates equal protection. The Arizona case that the SCOTUS heard in December involves such a challenge, since the "independent" redistricting commission overpopulated Republican-leaning districts, and underpopulated Democratic-leaning districts.

    If the range is greater than 10%, the state has to justify the greater range. In that previous case, Ohio justified the deviation based on respecting county borders.

    The requirements are a lot weaker than they appear to be, and could be. At first blush, it would appear that the constitution requires minimal feasible county splitting. You would of course have county splitting of larger counties, but can generally avoid splitting of smaller counties (perhaps 2 or 3 in the state), while maximizing the number of districts wholly within a county.

  3. A better method would be to divide the state into regions that are comprised of whole counties and have the population equivalent to a whole number of districts. Then each region could be districted independently. You could even have a local redistricting commission.

    We can divide the population of each county, city, block, etc. by the ideal population of a district. Then the target population is one. This can be expressed as a mixed decimal fraction, calculated to three decimal points.

    For a county, the whole number represents the number of districts which should be wholly within a county. The fraction is the surplus for a county.

    For smaller counties, the surplus is the entire population of the county.

    Ideally, we can form districts (1) wholly within a county; or (2) comprised of whole smaller counties, and/or the surplus portion of larger counties.

    Individuals can propose apportionment regions:

    An apportionment region is made up of whole contiguous counties, with a population equivalent to a whole number of districts. The population per district should be in the 95% to 105% range, and the total number of districts for all apportionment regions.

    For each apportionment region, calculate the number of districts that can be wholly within a county, and the total of the county surpluses.

    If the surplus is approximately zero or one, a single district can be comprised from whole smaller counties and the surplus of the larger counties.

    If the surplus is close to two, then one additional split is needed, if the surplus is close to three, then two additional splits, are needed.

    Proposed plans for apportionment regions can be compared based on the total number of regions (more is better), excess splits (fewer is better), and population equality (if we assume perfect equality within each region, we can estimate the standard deviation for the state).

    We can select the best plans, and let a representative sample of the population choose the one they like best.

    The individual regions can then be independently districted.

  4. Pennsylvania has rather specific rules on how to apportion state legislature districts to counties and cities:

    And apparently a 2012 court case found that the rules had not been sufficiently followed: