HOMEMODULESMODULE_04

Voter Suppression Patterns

Access Control Mechanisms

6 hours4 topicsPrimary sources included
4.1

A Continuous Architecture

Voter suppression is not a relic of history—it's a continuous system that adapts to legal constraints. When one method is outlawed, new methods emerge that achieve similar effects through different mechanisms.

1870

15th Amendment

Prohibits denying vote based on race. Southern states immediately begin finding workarounds.

1877-1965

Jim Crow Era

Poll taxes, literacy tests, grandfather clauses, white primaries systematically disenfranchise Black voters.

1965

Voting Rights Act

Federal preclearance required for election changes in covered jurisdictions. Black registration soars.

2013

Shelby County v. Holder

Supreme Court strikes down preclearance formula. Within hours, states announce new restrictions.

suppression_evolution.pseudo
pseudo
// Voter suppression adapts to legal constraints
HISTORICAL_METHODS (now illegal):
    - poll_tax                 // 24th Amendment (1964)
    - literacy_test           // VRA Section 4 (1965)
    - grandfather_clause      // Guinn v. US (1915)
    - white_primary           // Smith v. Allwright (1944)

MODERN_METHODS (legally contested):
    - strict_voter_id         // Disproportionate ID access
    - registration_deadlines  // Distant from election day
    - poll_closures          // Reduce access in targeted areas
    - voter_roll_purges      // Algorithmic removal of "inactive" voters
    - felon_disenfranchisement // 5.2 million affected
    - limited_voting_hours   // Conflicts with working schedules
    
// Same function, different implementation
FUNCTION suppress_vote(target_demographic):
    method = select_legal_method()
    WHILE method.impact_on(target_demographic) < threshold:
        method = next_legal_method()
    RETURN method
4.2

Shelby County v. Holder: The Rollback

The Voting Rights Act required certain jurisdictions with histories of discrimination to get federal approval before changing election laws. In 2013, the Supreme Court struck down the formula determining which jurisdictions needed preclearance.

shelby_impact.pseudo
pseudo
// Shelby County v. Holder (2013)
BEFORE shelby:
    covered_jurisdictions = 9 states + portions of 7 more
    
    FUNCTION change_election_law(jurisdiction, change):
        IF jurisdiction IN covered:
            REQUIRE federal_preclearance(change)
            // DOJ or DC Court must approve
        APPLY change
        
AFTER shelby:
    covered_jurisdictions = NONE
    preclearance_required = FALSE
    
    // New changes within 24 hours of ruling:
    texas -> implements_strict_voter_id()
    mississippi -> implements_voter_id()
    alabama -> implements_voter_id()
    north_carolina -> omnibus_voting_restrictions()

// Impact (2013-2020)
new_restrictive_laws = 25 states
poll_closures_in_formerly_covered = 1,688
voter_purges_increase = 33%
1,688
Polling places closed
In formerly covered jurisdictions
33%
Increase in purges
After Shelby County
25
States
Passed new restrictions 2013-2020
4.3

Voter ID: Surgical Precision

Strict voter ID laws require specific forms of identification that certain populations are less likely to have. Courts have found some laws were crafted with "almost surgical precision" to target minority voters.

voter_id_analysis.pseudo
pseudo
// North Carolina HB 589 (struck down 2016)
FUNCTION craft_voter_id_law():
    // Step 1: Request racial data on ID types
    data = request_dmv_data(breakdown_by_race=TRUE)
    
    // Step 2: Identify IDs disproportionately held by Black voters
    disproportionate_black = [
        "public_assistance_id",
        "government_employee_id",
        "out_of_precinct_provisional"
    ]
    
    // Step 3: Exclude those IDs from acceptable list
    acceptable_ids = all_ids - disproportionate_black
    
    // Step 4: Accept IDs disproportionately held by white voters
    acceptable_ids.add("hunting_license")  // 90%+ white holders
    acceptable_ids.add("concealed_carry")  // Similar demographics
    
    // Court finding: "target African Americans with almost surgical precision"

// ID possession rates
DEMOGRAPHIC white_voters:
    has_acceptable_id = 0.91
    
DEMOGRAPHIC black_voters:
    has_acceptable_id = 0.73
    
gap = 0.18  // 18 percentage point disparity
THE DATA REQUEST
Before passing HB 589, North Carolina legislators requested data on voting practices broken down by race. They then eliminated or restricted each practice where Black voters participated at higher rates than white voters.
4.4

Voter Roll Purges

States regularly remove voters from registration rolls using various criteria. While maintenance is necessary, aggressive purges often use flawed matching algorithms that disproportionately affect minority voters.

purge_algorithm.pseudo
pseudo
// Crosscheck Program (used by 27 states until 2019)
FUNCTION crosscheck_purge(voter_lists):
    FOR voter_a IN state_a_list:
        FOR voter_b IN state_b_list:
            IF voter_a.first_name == voter_b.first_name AND
               voter_a.last_name == voter_b.last_name AND
               voter_a.dob == voter_b.dob:
                FLAG as potential_duplicate
    
    // Problem: Common names + DOB = false positives
    // "Maria Garcia" born 1/15/1980 matches thousands
    
    // Demographic impact analysis
    name_commonality_by_race:
        white_surnames = more_unique  // Schmidt, O'Brien
        black_surnames = less_unique  // Williams, Johnson
        hispanic_surnames = less_unique  // Garcia, Rodriguez
        asian_surnames = less_unique  // Lee, Kim, Nguyen
    
    // Result: Minority voters 85% more likely to be flagged
    false_positive_rate = 0.993  // 99.3% of matches were errors

// Georgia 2018: 340,000 purged for not voting
// Many discovered on Election Day they couldn't vote
17M
Voters purged
Between 2016-2018 alone
99.3%
False positive rate
Crosscheck program accuracy
4.5

Felon Disenfranchisement

The U.S. is one of few democracies that restricts voting rights for people with felony convictions. Due to racial disparities in the criminal justice system, this disenfranchises Black men at nearly 4x the rate of others.

felon_disenfranchisement.pseudo
pseudo
// Current state of felon disenfranchisement
STATE_POLICIES:
    permanent_ban = [Alabama*, Florida*, Kentucky, Tennessee*, Virginia*]
                   // *with restoration process
    
    post_sentence = [Arizona, Delaware, Nebraska, Nevada, ...]
    
    probation_parole = [California, Colorado, Connecticut, ...]
    
    incarceration_only = [DC, Hawaii, Illinois, Indiana, ...]
    
    no_restriction = [Maine, Vermont]  // Can vote from prison

// Racial impact
FUNCTION calculate_disenfranchisement():
    total_disenfranchised = 5,200,000
    
    black_voting_age_pop = 30,000,000
    black_disenfranchised = 1,700,000
    black_rate = 5.7%
    
    all_other_rate = 1.5%
    
    disparity_ratio = 5.7 / 1.5 = 3.8x

// Florida Amendment 4 (2018)
// Voters restored rights to 1.4 million
// Legislature added fines/fees requirement
// Effect: 80% still cannot vote due to unpaid obligations
THE FLORIDA LOOPHOLE
In 2018, Florida voters passed Amendment 4 restoring voting rights to 1.4 million people. The legislature then required all fines, fees, and restitution be paid first. Since Florida has no central database of what's owed, many cannot verify eligibility.

MODULE_04 // KEY_TAKEAWAYS

  • Voter suppression is a continuous system that adapts methods when old ones become illegal.
  • Shelby County v. Holder removed federal oversight, triggering immediate new restrictions.
  • Voter ID laws and purge algorithms can be designed to appear neutral while targeting minorities.
  • 5.2 million Americans cannot vote due to felony convictions, disproportionately affecting Black men.
shadow_advisor.exe

// SHADOW_ADVISOR INITIALIZED

Try asking:

$
Built with v0