[EL] Toobin and House Results -- Re: ELB News and Commentary 11/18/18

Mon Nov 19 10:10:36 PST 2018

Rick,

I think all those points are good, which is why I made the (obscure) comment at the end about this being particularly relevant to the election of 1974…

That said, the question I was wading into was what to make of the seats/votes relationship as revealed by the 2018 election (if I might translate the question that arises from the Toobin article into political science).  The value of an empirical analysis like this (as you know) is that it abstracts from many of the details you bring into the question.  There are dangers to abstraction; there are dangers in wallowing in the details of the cases.

As you know, there are different ways to depict the empirical seats/votes curve, and different choices to be made in deciding how to implement the methods.  The method I chose --- which is rooted in the beginnings of the modern analysis of this question by Tufte --- picks observed seats/votes pairs for a period of time and looks at the relationship.  Of course, the bigger the time interval, the better the fit of the data to a systematic relationship, but the harder it is to argue relevance.  So, we make trade-offs.

For the record, I re-did my seats-votes curve, for the curious, starting the period at different years:  1946, 1956, 1966, … 2006.  The scatterplots are all in the figure below.  (I’ve added the 95% confidence interval of the prediction to help make the point about variability of small samples.  Some comments:

1.       The regression line describing the relationship seems to shift downward with the 1996-2016 graph.  Of course, from that period on, the “Democratic side” of the graph is dependent on two elections in which Democrats won a majorit of seats.

2.       As we reduce the number of years covered, the standard error of the prediction grows to reflect the reduction in data.  If we confine ourselves to data from the past five elections (graph 2006-2016), virtually any likely seats/votes pair we observe will be within the confidence interval.

3.       For every graph, the 2018 point is below what we would expect, given other elections from that period.  It is closer to expectations if we restrict ourselves to data since 2006.

Obviously, these types of seats/votes graphs won’t convince people who believe that the fundamental result of the election was due to gerrymandering or unrelated to gerrymandering.  Hopefully, it will encourage people to understand where 2018 fits into the context of the seats/votes relationship --- however they choose to bound their historical window.

Cheers,

Charles

[cid:image003.png at 01D48009.3E5FBE30]
-------------------------------------------------------------------------------------------------------------
Charles Stewart III
Kenan Sahin Distinguished Professor of Political Science
The Massachusetts Institute of Technology
Cambridge, Massachusetts   02139
617-253-3127
cstewart at mit.edu<mailto:cstewart at mit.edu>

From: Pildes, Rick <rick.pildes at nyu.edu>
Sent: Monday, November 19, 2018 11:27 AM
To: Charles H Stewart <cstewart at mit.edu>; Election Law Listserv <law-election at uci.edu>
Subject: RE: [EL] Toobin and House Results -- Re: ELB News and Commentary 11/18/18

Charles, the scatterplot is interesting, but I’m not sure what kinds of inferences can legitimately be made from it given such a long time-period:
(1) from the 1940s until the post-Reynolds v. Sims era, many states did not redistrict at all or redistricted in a way that intentionally used non-equal population districts to bias in favor of rural areas;
(2) for most of the “modern” era of redistricting, when the 1 person, 1 vote doctrine applied, the Democrats had unified control of many more legislatures than the Republicans and had vastly greater opportunities to gerrymander, which was true in the 1980 and 1990 round for sure ;
(3) also for much of the time period shown here, the South was mostly a one-party Democratic monopoly, though there were a few districts in the region that elected Republicans, and that will also influence these results.  Some of these elements cut against each other, but without filtering the data to take these kind of variables into account, I’m not sure what legitimate inferences for today can be made from the bare scatterplot of seat/vote proportions over this time period.
Of course, there were significant Republican gerrymanders of Congress in this cycle, such as in NC, and none of these points is meant to deny that.

Best,
Rick

Richard H. Pildes
Sudler Family Professor of Constitutional Law
NYU School of Law, Vanderbilt Hall 507
40 Washington Square South, NY, NY 10012
212 998-6377
Fax: 212-995-4082

From: Law-election [mailto:law-election-bounces at department-lists.uci.edu] On Behalf Of Charles H Stewart
Sent: Sunday, November 18, 2018 9:41 PM
To: Mark Scarberry; Election Law Listserv
Subject: Re: [EL] Toobin and House Results -- Re: ELB News and Commentary 11/18/18

For those who are curious, here is the scatterplot of Democratic seat share in the House against Democratic vote share, from 1946 to the present.  If we just take the seats-votes curve from the post-war era, we would expect the Democrats to have won 59.36% of the seats, or 253, rather than the most likely 233, or 53.56% of seats.  Of course, as the graph shows, for the entire post-war period, there has been a pro-Democratic bias in the House worth about 16 seats, which I think most experts believe no longer exists.  And, of course, there are other ways to calculate what the seat/votes curve is, and should be.  Other methods of calculating the current empirical seats/votes curve (such as the model used by fivethirtyeight) show a clear Republican bias.  Let’s just say, if this were 1974, and the Democrats had gotten this popular votes, the majority would have been must greater.

[cid:image001.jpg at 01D47FFE.E902C2C0]
-------------------------------------------------------------------------------------------------------------
Charles Stewart III
Kenan Sahin Distinguished Professor of Political Science
The Massachusetts Institute of Technology
Cambridge, Massachusetts   02139
617-253-3127
cstewart at mit.edu<mailto:cstewart at mit.edu>

From: Law-election <law-election-bounces at department-lists.uci.edu<mailto:law-election-bounces at department-lists.uci.edu>> On Behalf Of Mark Scarberry
Sent: Sunday, November 18, 2018 8:22 PM
To: Election Law Listserv <law-election at uci.edu<mailto:law-election at uci.edu>>
Subject: [EL] Toobin and House Results -- Re: ELB News and Commentary 11/18/18

Jeffrey Toobin, in the New Yorker article, writes:

"Even the good news from the election comes with a caveat, however. According to an analysis by the Brennan Center for Justice, Democrats won the over-all popular vote in the four hundred and thirty-five races for the House of Representatives by about nine per cent, but they managed to capture only a relatively narrow majority of seats. This is because the district lines are so egregiously gerrymandered, especially in states fully controlled by Republicans."

Assuming my math is correct:

A 9% margin would put the percentages at 54.5 to 45.5 (leaving aside third parties). Out of 435 seats, 54.5% would be 237, and 45.5% would be 198. It appears that, with a few races still to be decided, Democrats will have at least 232 seats and Republicans will have at least 198. If the five other raises split evenly, the division will be 234 or 235 Democrats, and 200 or 201 Republicans. Is this particularly disproportionate?

Mark

Prof. Mark S. Scarberry
Pepperdine Univ. School of Law

On Sun, Nov 18, 2018 at 4:09 PM, Rick Hasen <rhasen at law.uci.edu<mailto:rhasen at law.uci.edu>> wrote:

...

Jeffrey Toobin Expresses Some Optimism About Voting Rights<https://urldefense.proofpoint.com/v2/url?u=https-3A__electionlawblog.org_-3Fp-3D102371&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=E6HCkqFQbLA8N5JmAMOctlvkRMc4qYkPY8Kev4POJjY&e=>
Posted on November 18, 2018 3:17 pm<https://urldefense.proofpoint.com/v2/url?u=https-3A__electionlawblog.org_-3Fp-3D102371&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=E6HCkqFQbLA8N5JmAMOctlvkRMc4qYkPY8Kev4POJjY&e=> by Rick Hasen<https://urldefense.proofpoint.com/v2/url?u=https-3A__electionlawblog.org_-3Fauthor-3D3&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=_5Wb8_qtBxqrXZaX-uGLFLmeCc4bCGRiupLVd7wPoP8&e=>

Not so sure I agree with this one<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.newyorker.com_magazine_2018_11_26_how-2Dvoting-2Drights-2Dfared-2Din-2Dthe-2Dmidterms&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=Qhkr6zu_JWcQl8b5jyYHQ5jpfxEVs6z1qM9KCoBFpSo&e=>.
[Share]<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.addtoany.com_share-23url-3Dhttps-253A-252F-252Felectionlawblog.org-252F-253Fp-253D102371-26title-3DJeffrey-2520Toobin-2520Expresses-2520Some-2520Optimism-2520About-2520Voting-2520Rights&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=p99WVNf642hGPJ6c292lhF05SI-VIBz2kyKtNgyjfCY&e=>
Posted in The Voting Wars<https://urldefense.proofpoint.com/v2/url?u=https-3A__electionlawblog.org_-3Fcat-3D60&d=DwMGaQ&c=slrrB7dE8n7gBJbeO0g-IQ&r=v3oz9bpMizgP1T8KwLv3YT-_iypxaOkdtbkRAclgHRk&m=Ah4p--cpKdkC2UuZ5uZp7ksv06e_e51-y4jJvAzsSRc&s=AXc-b5yAoUqI8rhvk_WofBaqnJZ_dbgQGLWkkmFFoGM&e=>

...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20181119/1935849a/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.jpg
Type: image/jpeg
Size: 23963 bytes
Desc: image001.jpg
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20181119/1935849a/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 2021 bytes
Desc: image002.png
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20181119/1935849a/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.png
Type: image/png
Size: 133376 bytes
Desc: image003.png
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20181119/1935849a/attachment-0001.png>