← Back to adamtroy.net

Two causes of death moved in opposite directions.

Watch annual, county-level mortality estimates shift across the United States from 1980 through 2014. Both maps use age-standardized rates per 100,000 people.

Between 1980 and 2014, two epidemics moved in opposite directions across American counties. Homicide rates fell by roughly a third. Drug death rates rose fourteen-fold. Over the same period, state and local governments consistently spent far more on public safety (police, fire, and corrections) than on public health. I built this visualization to make that history visible, county by county and year by year, because the spending choices communities are making today with opioid settlement funds will echo for decades, just as the choices of the last forty years did.

  1. Drug deaths rose about 14-fold on average across counties while homicide fell about 35 percent, 1980 to 2014.
  2. Public safety spending exceeded public health spending every year. The dollar gap widened from about $50 billion to about $129 billion in 2014 dollars, even as the ratio narrowed from 3.06x to 2.41x.
  3. Public safety spending (police, fire, corrections) ran far ahead of public health spending throughout the period, while the overdose crisis grew. These parallel trends do not prove causation, but the imbalance is the question communities now face as they allocate settlement funds.
1980 2014
1980

Homicide

IHME cause: interpersonal violence

7.7county mean
Loading county shapes…

Point to or select a county to inspect its rate.

Drug deaths

IHME cause: drug use disorders

0.7county mean
Loading county shapes…

Point to or select a county to inspect its rate.

Government spending over the same 35 years

National state and local direct expenditure. The year marker follows the map timeline; chart lines show inflation-adjusted 2014 dollars.

Adjust spending for population growth

Public health spending

Census Health function, excluding hospital care

$24.1B $8.4B nominal · 1980
$24.1B in 1980 $91.4B in 2014 3.79× real growth

Public safety spending

Police, fire, and corrections combined

$73.7B $25.7B nominal · 1980
$73.7B in 1980 $220.3B in 2014 2.99× real growth

The gap between public safety and public health

Safety spending as a multiple of health spending, and the difference between the two

3.06× $49.6B gap · 1980

Public health spending grew faster in percentage terms, 3.79× versus 2.99× in real dollars, but from a much smaller base. The ratio narrowed while the dollar gap widened substantially.

Scope: State and local government direct expenditure on a fiscal-year basis. Federal spending is excluded. Public health is the Census Bureau’s broad Health function, including administration, laboratories, categorical programs, inspections, environmental regulation, animal control, and separately run ambulance services; it excludes hospitals, Medicaid vendor payments, and nursing homes. Public safety combines police, fire, and corrections and excludes courts, prosecution, and public defense.

Per-capita view: Per-person figures adjust for U.S. population growth from 227.2 million in 1980 to 319.3 million in 2014.

Interpretation: These national spending totals are shown beside county-level age-standardized mortality estimates for context; the parallel trends do not establish that spending caused changes in mortality. Coverage ends in 2014 to match the maps.

Source: U.S. Census Bureau, Annual Survey of State and Local Government Finances, via the Urban Institute’s State and Local Finance Data tool, accessed October 13, 2026. Inflation adjustment uses annual CPI-U averages from the U.S. Bureau of Labor Statistics.

The national pattern, at a glance

7.7 → 5.0

Unweighted mean across counties, 1980 to 2014

0.7 → 10.0

Unweighted mean across counties, 1980 to 2014

Why this matters now

The maps end in 2014, before the fentanyl era and before the opioid settlements. The question they raise is live right now. Billions of dollars in opioid settlement funds are flowing to states and counties, and communities are deciding how much goes to treatment, prevention, and recovery versus law enforcement and other uses. The record above is the baseline: four decades in which public safety spending ran far ahead of public health spending, and the overdose crisis grew anyway. The Census spending categories on this page are broader than settlement spending categories, so this history is context for the choice, not a direct measure of it. Whether settlement spending repeats that pattern or changes it is a choice being made in public meetings today. My work on participatory allocation of settlement funds (MAAPPS) is about making sure the people most affected have a say in that choice.

Methods and data

County mortality estimates come from IHME’s 1980–2014 small-area model, using age-standardized rates per 100,000 for interpersonal violence and drug use disorders. Point estimates are shown; uncertainty intervals are not included. The mortality CSV contains one county-year per row.

Spending comes from the U.S. Census Bureau Annual Survey of State and Local Government Finances via the Urban Institute’s State and Local Finance Data tool. Public health is the Census Health function, excluding hospitals, Medicaid vendor payments, and nursing homes. Public safety is police, fire, and corrections combined. Federal spending is excluded. Totals are fiscal-year direct expenditures; 2014-dollar values use BLS CPI-U annual averages. The spending CSV includes total, nominal, real, and per-capita series.

Important limitations include IHME model smoothing in small counties, the GBD drug-use-disorders definition, the 1999 ICD revision, 32 county-equivalents that share grouped estimation series, and an endpoint before the fentanyl-era surge. These national spending trends and county mortality estimates are descriptive and do not establish causation.