NCSL Virtual Meetings, 16 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Racial Bloc Voting Analysis Determine If A Redistricting Plan Complies With Voting Rights Act

Description

To stay connected with legislators and legislative staff, the NCSL Elections Team will begin holding Office Hours—a virtual meeting to speak with our team on any elections or redistricting-related topics. Please feel free to hop on with your question, and stay for as little or as long as you like. You can also send us your questions in advance, and we'll be sure to find the answer or find an expert who knows the answer. All persons interested are welcome to attend.

A

B

Okay, what I want to do is tell you why you need to think about doing a racial block voting analysis. At least you want to think about doing this. If you um have a sizable minority community in your state or jurisdiction,.

C

So the voting rights act section two of the voting rights act- prohibits.

D

Any voting standard practice or procedure, including a redistricting plan that results.

E

In the denial or dilution of minority voting strength.

B

Unlike section 5, which you probably know, is no longer operational because of the supreme court decision in 2013, I think it is.

C

um Section 5 only applied to some states.

D

And local jurisdictions, section 2, has always applied to the entire country.

F

And it still applies to the entire country, so everybody should be aware of what section.

G

2 means for them in the context of redistricting.

C

It was amended in 1982 to make it clear that plaintiffs, bringing a challenge did not have to prove that the jurisdiction had um intentionally discriminated against.

E

Minority plaintiffs, instead, the idea.

D

Was that they simply had to show that, for example, the redistricting plan had the effect.

E

Of denying or diluting minority voting strength.

E

The first case to consider the amended voting rights act was thornburg.

D

B jingles: this was decided in 1986. It was a challenge to the north carolina state legislative plans and.

H

In that, the supreme court held that plaintiffs had to satisfy three preconditions before it.

D

Could qualify for relief under.

C

Section two and the first was that the minority community had to be sufficiently large and geographically.

E

Compact to form a majority in a single member district. This goes to the remedy. Even if votes are diluted, if you don't have a remedy, you can't draw a district. Then there is.

I

No possibility of bringing a successful suit that would in fact produce a successful remedy.

G

The second is that the minority group must be politically.

D

Cohesive, if the minority.

G

Group is not politically cohesive. Obviously drawing a single member district isn't going to.

J

Allow the minority group to elect a candidate of choice, because the minority.

D

Group is not supporting the same candidates. Finally, whites must vote as a block to usually defeat.

E

Minority preferred candidates so.

J

You may have an instance where voting is polarized, but the minorities might be in a majority and.

I

Actually, consistently.

E

Win, in which case you wouldn't need to create a remedy.

E

Now, how do we know how the minority group votes.

B

How do we know if they're politically cohesive? How do we know if whites are block voting to usually defeat the minority preferred candidate? The way that we know this is by doing a racial block voting.

G

Analysis so the racial black voting analysis is going to analyze voting patterns by race and.

C

It's going to tell us if the plaintiffs.

G

Can prove the second and the third prong of genes.

B

What do you need to do this? Well, we have a secret ballot in this country, so we're not going to know from the ballot how.

G

People what the races of the person who cast the ballot so we're going to have to do some estimation.

D

And we're going to do this by creating a database that combines election returns.

J

This is how we're going to know what the voting behavior is with the demographic composition of the precinct by election returns. We're going to focus.

G

On the most relevant elections being those for the opposite issues, so.

D

If you're talking about state legislative districts or congressional districts, you're going to focus on state legislative.

G

Or congressional elections and the most probative according to the supreme court.

H

And other courts is those contests that include minority candidates.

K

So that's going to be the focus in terms of the election returns.

H

In terms of the demographic composition of the precincts, we want to get as close as we can to the electorate.

C

So there are actually a number of states that.

D

Collect registration by race and you can even take the registration voter lists.

E

And indeed, and determine in many of these states turn out.

F

By race, so this is what you want. This will bring you closest to the actual electorate.

H

But this is only possible in a handful of southern states.

G

So most of us will be using voting age population by race or ethnicity, and this.

F

Will be provided by the pl data which we're going to see very very soon. This has vtd data in it. So this this vtd.

J

Data gives us the racial composition.

D

Of the precincts in which the elections occurred, if you are in a jurisdiction with a high non-citizen.

I

Voting citizen population- you might want to look at citizen voting age population by race, since non-citizens can't vote and.

F

You would be closer to the electorate, but of course we won't know in the pl.

J

D

Citizen voting age population so we're going to use an algorithm from the most recent.

F

Acs american community survey data to produce citizen voting age population data by race for our precincts.

F

So that's the data that we're going to use.

J

Now, if we wanted a.

C

If we had precincts.

B

In which um one race was predominant, if we had precincts.

H

F

Of the voters were say, black or all of the voters were hispanic. We could use those precincts to.

G

Produce estimates of how.

F

These groups, these racial groups voted, but there are two drawbacks to that approach and the first is very few jurisdictions: have a lot of homogeneous precincts, or at least not a lot of homogeneous, say black or hispanic.

K

Precincts, they might have a lot of homogeneous white precincts and the second is it's quite possible that voting in homogeneous precincts um is different than voting in precincts that are more integrated.

F

So we have developed over the last 40 years a couple of different techniques for estimating voting behavior by race, and these are called ecological.

E

Regression, analysis and ecological inference analysis.

I

Ecological regression analysis has been around a long time in.

E

Fact, homogeneous precinct analysis and ecological regression analysis were the two techniques used by the.

G

Plaintiff's experts in thornburg b, jingles and the.

D

Court clearly approved of these two methods and drew.

K

Their conclusions about voting behavior on the basis of these two methods.

H

Now this is how ecological regression works. Each of these points on this scatter plot is a precedent and the precinct has been put on the scatter plot on the basis of two pieces of information, the percentage black turnout in that precinct and the percentage of votes for warnock. This is the 2021 runoff.

G

In georgia, in a single county.

I

Will remain nameless, but this is an actual election result for that county.

K

And you can see a very clear pattern. Looking at the scatter plot, you can see that the higher.

J

The proportion black turnout, the higher the votes for warnock.

J

And we use this to produce.

I

What's called a regression line, the computer will find for us the best fitting line for this pattern and where.

D

The line hits the zero point is going to be our estimate of the white votes.

E

Because, of course, that's zero black journal.

K

And where it hits, the one point is going to be our estimate.

H

Of the percentage of black votes for warnock.

K

I'm simplifying this a little bit and I'm using a.

D

Jurisdiction in which we actually have black turnout.

C

I'm also using a jurisdiction in which, because.

J

Voting is polarized. If you take the line over, I think you're going to find you're a little bit above one.

E

One of the drawbacks.

I

To ecological regression is, it produces estimates that are outside the logical bounds. You can get something like 101 or.

F

105 of blacks, turning out voted for warnock, which is.

C

Not possible so in the 1990s.

K

A professor at harvard named gary king, developed another method for estimating voting patterns by race, and he called this ecological inference.

C

This is just a little more complicated. What we're going to do here is each of these lines represents a precinct, so rather.

D

Than a single point.

I

Each line represents a precinct and it represents all the possible values.

D

That uh the percentage black could be given that what we know the percentage black of the precinct is and what we know.

F

About the votes for the particular candidate, so.

C

Each of these lines represents everything, that's possible.

H

Given the the values that we know- and you can see.

J

That, um where the lines gather the.

E

Most is over on the right hand, side of what's called.

F

This is called the tomographic plot and.

K

All of our points fall between 0.8 and point and 1.0, so we.

J

Know that the estimate is going to be very high of the percentage of.

G

Black voters, who supported warner, we're going to do the same thing for white voters.

I

This is what you get when you conduct this analysis.

J

These are er and ei analyses that came from the tomographic plot and the regression scatter plot. So we estimate that warnock got either.

K

94.9 or 94.4 of the black vote and um only 23.5.

D

To 25.5 percent of the white votes estimates are.

J

Always going to be slightly different since they are produced.

G

Using slightly.

K

Different methods um but they're all it doesn't matter if it's off by a percentage point or.

F

J

Three, because, of course, what we see here is this contest is polarized that the vast majority of black voters.

E

Supported warnock and the strong majority of white voters supported his opponent, lawful, so.

C

When you do a racial black voting analysis, this is the kind of thing that you're going to find, and these are the pieces.

D

Of information that you want to get you.

G

Want to know are black voters. Cohesive vary in this particular instance. Anyway. Are.

J

Whites voting against the black preferred candidate. Well, the black preferred.

G

Candidate is we're not the vast majority of whites are voting against warnock, so this uh indicates that.

I

Voting is polarized.

G

Black voters are very cohesive and white voters are black voting to defeat the minority prefer.

J

Candidate, I mean you're not going to do this for one election you're going to do this for many elections, because.

I

You want to look for a pattern across elections now, if you find a pattern of polarized voting across a series.

J

Of elections that you look at then you're going to know that you need to create a minori.

E

A district in which minority voters have an opportunity to elect their candidate.

F

Of choice, if such.

J

Districts already exist and the minority proper candidates are winning, because these.

E

Districts exist, then you have to make sure to maintain such districts. So.

G

We need to do this to determine. If voting is we need to do a rational analysis to.

K

Determine if voting is polarized.

I

K

E

Need to draw districts or maintain districts that allow minority voters to elect their candidates of choice.

G

C

G

J

Else with this information, now that we've, let's say, we've determined.

G

That voting is polarized.

E

And that we have to create minority districts, we.

F

J

To figure out, if the district will in fact perform for minorities, we have to figure out if we've created an effective minority district, and we can use the analysis that I just described to determine.

E

This the first approach we can use before any draft.

K

Districts are drawn and what we're going to do is we're going to use basic.

J

Algebra to determine using the degree of minority cohesion, the degree of.

D

White crossover vote and the participation rates of minorities and whites to determine about the percentage.

J

E

You would need in a district for a minority preferred candidate to be elected after you've drawn districts. You will have already done.

H

Your racial block voting analysis. You will have selected what I'm going to call some bellwether elections, that is, elections that.

E

Are polarized and that.

K

You have about a consistent amount of white crossover, vote and minorities.

H

Are cohesive and you can recompile election results.

J

To determine if the minority preferred candidate will actually carry that proposed district. So here's an example of calculating the minority.

E

Percentage needed to win.

H

I should have said this: only this is calculated on the basis of getting 50.

D

Plus one of the vote, so, if you're in a jurisdiction that has an outlaw collection system and you can win.

K

uh With 30 of the vote, this particular method isn't going to be very helpful. This is for general elections or primaries that are limited, and this will help you determine the percentage that you would need.

F

This is how you can do it before you start drawing, and this is how you can do it after you start drawing you're going to disaggregate.

D

The election results of previous elections.

J

Down to the block level and re-aggregate them up to conform with proposed boundaries to determine how the minority preferred candidates do in the.

F

Proposed districts.

C

So that's that's why you're going.

E

To put that's why you want to do a racial black voting.

I

Analysis and who needs to do.

K

It well anyone with a significant minority population. So if it's possible to draw a minority district in your jurisdiction, you better figure out. If you need to draw a minority district, if it isn't possible, then this.

J

Analysis probably can't be done anyway, because you have to minority.

G

Voters in order to actually conduct the analysis.

B

What it is you're going to take election precinct data, that's your voting behavior information and you're going.

K

To tie it to the precinct demographics and do the statistical techniques I just described, you should do this analysis before you begin drawing.

E

um Or at least before you adopt a redistricting plan.

J

Unfortunately, all too often this.

K

Is done after you're, sued and you're already in trouble at that point, because it's been quite clear that you didn't determine if the voting rights act applied to you or not, and.

J

Why you're going to do this to determine what you need to do to comply with the voting rights act.

C

And that's all I have uh in 20 minutes.

K

And I am happy to answer any questions.

A

Well, thank you so much dr hamlin. This was terrific and I I know that um even if the audience doesn't have a question right off of the bat, the more they think about this, it just seems like one of those sub.