Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Two Cheers for Ohio Redistricting Reform

November 4th, 2015, 10:16am by Sam Wang

Yesterday, Ohio voters overwhelmingly approved Issue 1, an initiative that sets up a bipartisan commission to oversee redistricting of state legislative districts. This is a great step in combating gerrymandering. However, there are some fine points that have to be resolved for it to work well.

The new commission will draw up districting plans. The commission is bipartisan and has seven members, and is dominated by the party that controls the legislature. Plans are passed by majority vote, and then go to elected state officials for approval. If two members of the minority party dissent, the plan goes into place for only four years instead of the usual ten.

In addition to the fact that the majority party remains in control, which is a concern, let me focus on a part that needs some work. The relevant text is at this link (I got it via this advocacy organization) and is pasted below:

Section 6. The Ohio redistricting commission shall attempt to draw a general assembly district plan that meets all of the following standards:
(A) No general assembly district plan shall be drawn primarily to favor or disfavor a political party.
(B) 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.
(C) General assembly districts shall be compact.

Section 6(B) is a bit tricky to interpret. In a single-member district system, the only way to get proportional representation is to make at least some districts highly uncompetitive. The reason is that the expected number of seats takes the form of an integrated curve as a function of the vote-share V, plotted across all districts. The only way to get proportionality is to have districts distributed like this:

0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100% Democratic.

If <=30% is considered noncompetitive, this means that in this example, only 3 out of 11 districts are competitive. So competition and proportionality are enemies!

On the other hand, if “correspond” means “a seats-to-votes curve that follows historical practice,” that is easier to achieve. Historical practice supports a standard deviation of V of 12-15%, which leads to an S-shaped seats-to-votes curve (details buried  at footnote #59 on page 34 of my paper; sorry, will provide better writeup later). To get there, all one has to do is not pack voters of one party or the other. Surely the Commission could do that.

There is a second problem, rooted in politics. Two minority-party dissensions are needed to trigger the four-year threshold. State-level gerrymandering includes packing the opposing party into a few ultra-safe districts. It is possible for the majority to buy off the minority-party members of the Commission by promising them safe seats – or some other reward.  That has to be watched carefully.

Despite these concerns, this new law should still encourage plans that allow a majority of voters to elect a majority of representatives, which should emerge automatically in a party-blind districting process. For other levels of vote share, the Ohio commission will need to learn a little math to calculate the natural seats-to-votes curve.

Tags: Redistricting

5 Comments so far ↓

  • SF Bays

    That’s really good news. Anytime the public can take this power from legislators is a good thing.

  • Todd S. Horowitz

    I don’t understand. The commission is still controlled by the legislators whose districts it is drawing. Aside from the option for the minority party to call for a mulligan after four years (which, as noted could be avoided by entrenching the minority party in safe seats, which in turn could be interpreted as meeting Section A requirements), the only real improvement over the status quo (ante?) is the requirement that districts be “compact” in Section C. Sections A and B seem open to a great deal of interpretation.

  • 538 Refugee

    As an Ohioan I didn’t see much improvement when I read the thing initially. I’m trying to find the ‘editorial’ that accompanied the actual ballot wording. Makes you think it was a ‘one and done’ ideal solution.

  • Adam R

    Districts (‘electorates’) for the state parliament of South Australia are drawn following something like 6B. It is unusual, even in Australia.

    The practice here is that after each election, electorate boundaries are adjusted to try (where possible) to bring the proportions of left and right voters closer to the statewide average. Obviously, some electorates are heavily weighted one way or another, and cannot be 50-50 without serious gerrymandering (which is contrary to a compactness requirement). However, it generally works OK.

    One downside is that a successful local member of parliament who outperforms their party’s result are penalised by having their electorate made more competitive after each election, by incorporating more and more voters of the other persuasion, until they can no longer hold the seat.

Leave a Comment