Key Findings

Nationally

Summary:

This section calculates the average number of years individuals spent in prison past their parole eligibility year by race and ethnicity. The analysis excludes states with unreliable or missing data and states that abolished discretionary parole to ensure results are robust and representative.

Finding:

Code

# ---------------------------------------------------------------------------- #
# Years Spent in Prison Past Parole Eligibility
# ---------------------------------------------------------------------------- #
# Combine excluded states
combined_excluded_states <- unique(c(
  states_with_high_missing$state,
  states_with_high_missing_race$state,
  states_to_exclude
))

# Remove states that abolished parole
valid_states <- setdiff(
  ncrp_yearendpop_filtered$state, 
  unique(c(combined_excluded_states, states_abolished_parole$state))
)

# Filter NCRP data to valid states only
ncrp_yearendpop_filtered <- fnc_filter_pe_population_criteria(
  data = ncrp_yearendpop_filtered,
  exclude = states_to_exclude,
  dont_filter = states_nofilter
)

# Filter NCRP data to people currently eligible for parole (also past parole eligibility since it's yearend)
ncrp_past_pe <- ncrp_yearendpop_filtered |>
  filter(state %in% valid_states, parelig_status == "Current")

# Average time served past PE for people still in prison by race and ethnicity and state
avg_past_pe_race_by_state <- ncrp_past_pe |>
  filter(race != "Unknown") |>
  filter(
    state %in% states_use_other_race_eth$state |
      (!state %in% states_use_other_race_eth$state &
         race %in% c("White, non-Hispanic", "Hispanic, any race", "Black, non-Hispanic"))
  ) |>
  mutate(years_to_estimated_pey = abs(years_to_estimated_pey)) |>
  group_by(state, race, rptyear) |>
  summarise(
    avg_years_to_estimated_pey = mean(years_to_estimated_pey, na.rm = TRUE),
    total_years_past_pe = sum(years_to_estimated_pey, na.rm = TRUE),
    people = n(),
    .groups = "drop"
  ) |>
  fnc_filter_by_year(which_overall_year)

# Calculate weighted averages
avg_past_pe_race <- avg_past_pe_race_by_state |>
  group_by(race) |>
  summarise(
    avg_years_to_estimated_pey = sum(avg_years_to_estimated_pey * people, na.rm = TRUE) / sum(people, na.rm = TRUE),
    total_years_past_pe = sum(total_years_past_pe, na.rm = TRUE),
    people = sum(people, na.rm = TRUE),
    .groups = "drop"
  )

# Get states used
included_states <- unique(avg_past_pe_race_by_state$state)
states_using_2018_data <- avg_past_pe_race_by_state |>
  filter(rptyear == 2018) |>
  pull(state) |>
  unique()

# Generate Text Outputs
cat("In 2019, ",
    highlight_number(comma(avg_past_pe_race$people[avg_past_pe_race$race == "White, non-Hispanic"])),
    "White, non-Hispanic individuals collectively spent",
    highlight_number(comma(round(avg_past_pe_race$total_years_past_pe[avg_past_pe_race$race == "White, non-Hispanic"], 1))),
    "years in prison past their parole eligibility year across",
    highlight_number(length(included_states)), "states, serving an average of",
    highlight_number(round(avg_past_pe_race$avg_years_to_estimated_pey[avg_past_pe_race$race == "White, non-Hispanic"], 1)),
    "years beyond parole eligibility.<br><br>")

In 2019, 81,734 White, non-Hispanic individuals collectively spent 244,565 years in prison past their parole eligibility year across 24 states, serving an average of 3 years beyond parole eligibility.

Code

cat("In 2019, ",
    highlight_number(comma(avg_past_pe_race$people[avg_past_pe_race$race == "Black, non-Hispanic"])),
    "Black, non-Hispanic individuals collectively spent",
    highlight_number(comma(round(avg_past_pe_race$total_years_past_pe[avg_past_pe_race$race == "Black, non-Hispanic"], 1))),
    "years in prison past their parole eligibility year across",
    highlight_number(length(included_states)), "states, serving an average of",
    highlight_number(round(avg_past_pe_race$avg_years_to_estimated_pey[avg_past_pe_race$race == "Black, non-Hispanic"], 1)),
    "years beyond parole eligibility.<br><br>")

In 2019, 74,764 Black, non-Hispanic individuals collectively spent 280,147 years in prison past their parole eligibility year across 24 states, serving an average of 3.7 years beyond parole eligibility.

Code

cat("In 2019, ",
    highlight_number(comma(avg_past_pe_race$people[avg_past_pe_race$race == "Hispanic, any race"])),
    "Hispanic individuals collectively spent",
    highlight_number(comma(round(avg_past_pe_race$total_years_past_pe[avg_past_pe_race$race == "Hispanic, any race"], 1))),
    "years in prison past their parole eligibility year across",
    highlight_number(length(included_states)), "states, serving an average of",
    highlight_number(round(avg_past_pe_race$avg_years_to_estimated_pey[avg_past_pe_race$race == "Hispanic, any race"], 1)),
    "years beyond parole eligibility.<br><br>")

In 2019, 34,296 Hispanic individuals collectively spent 111,719 years in prison past their parole eligibility year across 24 states, serving an average of 3.3 years beyond parole eligibility.

Note: The following states used 2018 data due to unreliable 2019 data: Hawaii
Included states ( 24 ): Arkansas, Colorado, Connecticut, Georgia, Hawaii, Idaho, Iowa, Kentucky, Maryland, Massachusetts, Mississippi, Missouri, Montana, Nevada, New Hampshire, New York, North Dakota, Pennsylvania, Rhode Island, South Carolina, Tennessee, Texas, West Virginia, Wyoming

By Offense

Summary:

This section analyzes the offense types of individuals held past their parole eligibility year, distinguishing between violent and nonviolent offenses. Nonviolent offenses are further categorized into Property, Drug, and Public Order offenses.

Finding:

Code

# Categorize offenses into violent vs. nonviolent, keeping "Unknown" offenses for initial analysis
ncrp_past_pe_offense_initial <- ncrp_past_pe |>
  mutate(offense_group = case_when(
    fbi_index %in% c("Murder or Nonnegligent Manslaughter",
                     "Negligent Manslaughter",
                     "Rape or Sexual Assault",
                     "Robbery",
                     "Aggravated or Simple Assault",
                     "Other Violent Offenses") ~ "Violent",
    fbi_index %in% c("Drug", "Public Order", "Property") ~ "Nonviolent",
    TRUE ~ fbi_index))

# Filter out "Unknown" offenses for further analysis
ncrp_past_pe_offense_filtered <- ncrp_past_pe_offense_initial |>
  filter(offense_group != "Unknown") |>
  fnc_filter_by_year(which_overall_year)

# Check if each state has all nonviolent offense types (Property, Drug, Public Order)
states_with_all_offenses <- ncrp_past_pe_offense_filtered |>
  filter(offense_group == "Nonviolent") |>
  group_by(state) |>
  summarise(has_property = any(fbi_index == "Property"),
            has_drug = any(fbi_index == "Drug"),
            has_public_order = any(fbi_index == "Public Order"),
            .groups = "drop") |>
  filter(has_property & has_drug & has_public_order) |>
  pull(state)

# Summarize counts for nonviolent offense subcategories
# Property, Drug, Public Order
nonviolent_summary <- ncrp_past_pe_offense_filtered |>
  filter(offense_group == "Nonviolent") |>
  group_by(fbi_index) |> 
  summarise(total_people = n(), .groups = "drop")

# Extract counts for subcategories
property_offense_count <- nonviolent_summary$total_people[nonviolent_summary$fbi_index == "Property"]
drug_offense_count <- nonviolent_summary$total_people[nonviolent_summary$fbi_index == "Drug"]
public_order_offense_count <- nonviolent_summary$total_people[nonviolent_summary$fbi_index == "Public Order"]

# Total nonviolent count and states with all offenses
nonviolent_offense_count <- sum(nonviolent_summary$total_people)
states_with_nonviolent <- length(states_with_all_offenses)

# Generate output sentence
cat("Over", highlight_number(comma(nonviolent_offense_count)), 
    "people held past their parole eligibility year were convicted of nonviolent offenses in", 
    highlight_number(states_with_nonviolent), "states. Thousands of people who have not committed violent offenses remain incarcerated after their parole eligibility date. This includes approximately",
    highlight_number(comma(property_offense_count)), "people convicted of property offenses,",
    highlight_number(comma(drug_offense_count)), "convicted of drug offenses, and",
    highlight_number(comma(public_order_offense_count)), "convicted of public order offenses.\n\n")

Over 86,711 people held past their parole eligibility year were convicted of nonviolent offenses in 24 states. Thousands of people who have not committed violent offenses remain incarcerated after their parole eligibility date. This includes approximately 33,861 people convicted of property offenses, 28,214 convicted of drug offenses, and 24,636 convicted of public order offenses.