The%20Evolution%20of%20Substantive%20and%20Descriptive%20Representation,%201974-2004

1
The Evolution of Substantive and Descriptive
Representation, 1974-2004

• David Epstein
• Sharyn OHalloran
• Columbia University

2
Georgias Gerrymander
Range Baseline Proposed
0-25 31 26
25-40 11 17
40-50 2 0
50-60 2 8
60 10 5
Plan Reallocate black voters to elect Democrats
3
Is This Retrogression?
4
The Perfect Storm

• DC denied preclearance, saying state didnt prove
  non-retrogression in three districts
• SC overruled in Georgia v. Ashcroft
• Retrogression should be assessed statewide, not
  district-by-district
• States could pursue substantive rather than
  descriptive representation
• Put much weight on testimony of black legislators

5
Consensus View

• A conventional wisdom is forming about the
  meaning and importance of Ashcroft
• It abandoned a previous, relatively mechanical
  retrogression test based on electability
• It did so in favor of an amorphous concept of
  substantive representation that will be difficult
  to administer and
• The crux of the debate revolves around whether
  states should pursue substantive as opposed to
  descriptive representation.

6
This Paper

• We disagree with all three of these statements
• The previous standard for retrogression was
  crumbling anyway, due to political changes
• The Court revised this, too, in the opinion,
  moving to a statewide assessment of retrogression
• Substantive representation is not difficult to
  measure and administer
• Real arguments arent over descriptive vs.
  substantive representation, for the most part
• Rather, the question is on how best to achieve
  secure levels of substantive representation

7
Electability High Polarization
8
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
9
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
10
Electability Low Polarization
11
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
Low Polarization
BVAP
0
50
100
12
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
Low Polarization
BVAP
0
50
100
Coali- tional
13
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
PS
Low Polarization
BVAP
0
50
100
Coali- tional
Unsafe Control
14
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
PS
PP
Low Polarization
BVAP
0
50
100
Safe Control
Coali- tional
Unsafe Control
Packing
15
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
PS
PP
PI
Low Polarization
BVAP
0
50
100
Safe Control
Coali- tional
Unsafe Control
Packing
No Minority Control
Influence
16
Measuring Descriptive Representation
P
High Polarization
BVAP
0
50
100
Minority Control
No Minority Control
P
PS
PP
PI
Low Polarization
BVAP
0
50
100
Safe Control
Coali- tional
Unsafe Control
Packing
No Minority Control
Influence
How to make tradeoffs?
17
Retrogression in Electability

• Forget categories just use the probability of
  electing a minority candidate in each district
• Estimate this using S-curves

18
Low Polarization
19
Retrogression in Electability

• Forget categories just use the probability of
  electing a minority candidate in each district
• Estimate this using S-curves
• Then add up the probabilities to get the expected
  number of minorities elected
• Can consider the variance of this distribution,
  too
• For Georgia, the proposed plan had slightly fewer
  expected minorities elected
• Problem with overpopulated districts

20
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
Substantive
21
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
SQ
Substantive
22
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
SQ
Substantive
23
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
SQ
X
X
Substantive
Pre-Ashcroft
24
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
SQ
X
Substantive
Post-Ashcroft
25
Ashcroft Substantive Representation
Descriptive
Pareto Frontier
SQ
X
P
Substantive
A move to P is now non-retrogressive
26
Measuring Substantive Representation

• Great leaps have been made in the past two
  decades in the analysis of voting behavior
• This is now commonly used as a measure of
  members policy preferences
• Not because voting is the only important act
• But because it correlates with constituency
  service, committee work, etc.
• For substantive representation of black
  interests, define a legislators Black Support
  ScoreBSS of votes cast with the black
  majority

27
White Dem.
Black Dem.
South Carolina State House
Rep.
28
Overall Expected Representation

• Can compare plans by calculating the expected
  substantive representation
• Combines prob. of election and support scores
• For Georgia, this was
• Real argument is over the distribution of these
  scores, not over descriptive vs. substantive
  representation

Mean Median
Baseline 62.3 50.2
Proposed 65.9 69.2
29
Trends, 1974-2004

• Show changes in
• Election probabilities
• Substantive representation
• Maximizing plans
• Results
• Greater crossover in voting means point of equal
  opportunity is under 50 BVAP
• Southern Democrats become more liberal
• A tradeoff emerges between substantive and
  descriptive representation

30
White Dems
White Dems
Black Dems
Probability
Probability
Black Dems
Republicans
Republicans
Black Dems
Black Dems
White Dems
Republicans
Probability
Probability
Republicans
White Dems
31
Substantive Representation, 1974-2000
32
(No Transcript)
33
The Emerging Pareto Frontier
34
(No Transcript)
35
BVAP HVAP Combinations for PEO
36
Georgia State Senate, 1999-2002
37
Descriptive Representation, 1974-2000
38
Black Dem.
White Dem.
Rep.