Halloween fatal car accidents and deaths
On the ten Halloweens from 2015 to 2024, 1,163 people died in 1,069 fatal car accidents on US roads. Among them were 19 children aged 14 and under, killed walking: more than three times as many as on an average day, and more than on any other date of the year.
Halloween 2026 falls on a Saturday. The last two Saturday Halloweens, 2015 and 2020, had 129 and 161 road deaths, against a ten-year average of 116.
Pedestrians killed on Halloween, hour by hour
From 6 pm to midnight, 130 pedestrians were killed on the ten Halloweens, 27% more than on the same weekday in the three weeks before and after (103). The worst hour was 7 pm to 8 pm, with 37.
Halloween is the calendar day of October 31, midnight to midnight. A crash at 1 am on November 1 is not in it; one at 1 am on October 31 is. 9 Halloween crashes have no recorded hour.
Children aged 14 and under killed walking, hour by hour
Of the 19 children killed walking on the ten Halloweens, 13 died between 6 pm and midnight, the hours of trick-or-treating, against 2 on ten normal days at those hours. The worst hour was 7 pm to 8 pm, with 5. These are small counts: one child more or less changes a bar.
Pedestrians killed on Halloween by age
Children are 8% of the pedestrians killed on Halloween, but they are the group that rises most against a normal day: about three times as many. Adults aged 21 to 34 rise by 41%; people over 50 do not rise at all.
| Age | Ten Halloweens (total) | Normal days (ten-day equivalent) | Change on normal day |
|---|---|---|---|
| 0 to 14 | 19 | 6.2 | about three times |
| 15 to 20 | 12 | 9.2 | +30% |
| 21 to 34 | 59 | 41.7 | +41% |
| 35 to 49 | 54 | 48.8 | +11% |
| 50 to 64 | 41 | 57.2 | -28% |
| 65 and over | 46 | 43.7 | +5% |
| Age not recorded | 5 | 1.7 | Not enough data |
Source: NHTSA Fatality Analysis Reporting System (FARS), 2024 Annual Report File (released April 2026, subject to revision). How we count
States where Halloween is deadlier than a normal day
In Kentucky, Halloween had 47% more road deaths than the same weekday in the three weeks before and after, the largest rise of the 20 states with enough deaths to compare, ahead of Alabama and Georgia. In six states Halloween had fewer deaths than a normal day, most of all Illinois (-40%).
- Kentucky+47%
- Alabama+44%
- Georgia+36%
- Ohio+32%
- Indiana+28%
- South Carolina+26%
- Florida+24%
- Pennsylvania+21%
- Tennessee+17%
- New Jersey+15%
- Louisiana+12%
- Texas+8%
- Arizona+4%
- Virginia+3%
- Oklahoma-1%
- California-1%
- North Carolina-9%
- New York-13%
- Michigan-35%
- Illinois-40%
Ten Halloweens against sixty normal days: the same weekday in each of the three weeks before and after. A state is compared from 20 deaths on each side; the other 31 have too few. With counts this size, a change of 20% can be a handful of deaths.
Source: NHTSA Fatality Analysis Reporting System (FARS), 2024 Annual Report File (released April 2026, subject to revision). How we count
Halloween road deaths in every state, 2015 to 2024
Kentucky has the most Halloween road deaths per resident: 0.71 per million residents per Halloween, against 0.35 across the United States. New York has the fewest of the 32 states ranked. A rate is given from 10 deaths; the 19 states under that say "Not enough data".
| State | Per million residents | Deaths (total) | Fatal crashes (total) | Pedestrians (total) | Change on normal day | |
|---|---|---|---|---|---|---|
| 1 | Kentucky | 0.71 | 32 | 31 | 7 | +46.6% |
| 2 | South Carolina | 0.68 | 35 | 31 | 7 | +25.7% |
| 2 | Alabama | 0.68 | 34 | 32 | 7 | +43.7% |
| 4 | Nebraska | 0.67 | 13 | 10 | 1 | Not enough data |
| 5 | Mississippi | 0.64 | 19 | 16 | 4 | Not enough data |
| 6 | Arkansas | 0.63 | 19 | 15 | 5 | Not enough data |
| 7 | Tennessee | 0.60 | 41 | 38 | 6 | +16.6% |
| 8 | Georgia | 0.54 | 58 | 52 | 10 | +35.9% |
| 9 | Louisiana | 0.50 | 23 | 21 | 3 | +12.2% |
| 9 | Oklahoma | 0.50 | 20 | 19 | 2 | -0.8% |
| 11 | Florida | 0.48 | 103 | 97 | 22 | +24.1% |
| 12 | Indiana | 0.47 | 32 | 31 | 6 | +28.0% |
| 13 | Arizona | 0.44 | 32 | 30 | 11 | +3.8% |
| 14 | Texas | 0.41 | 119 | 102 | 22 | +8.2% |
| 15 | North Carolina | 0.38 | 40 | 38 | 7 | -9.1% |
| 16 | Nevada | 0.36 | 11 | 11 | 6 | Not enough data |
| 17 | Ohio | 0.34 | 40 | 34 | 4 | +31.9% |
| 17 | Utah | 0.34 | 11 | 9 | 1 | Not enough data |
| 19 | Oregon | 0.33 | 14 | 14 | 3 | Not enough data |
| 20 | Colorado | 0.31 | 18 | 18 | 6 | Not enough data |
| 21 | Pennsylvania | 0.30 | 39 | 38 | 5 | +21.2% |
| 22 | California | 0.29 | 113 | 97 | 40 | -1.2% |
| 22 | Virginia | 0.29 | 25 | 25 | 2 | +2.7% |
| 22 | Maryland | 0.29 | 18 | 16 | 4 | Not enough data |
| 22 | Missouri | 0.29 | 18 | 17 | 2 | Not enough data |
| 22 | Wisconsin | 0.29 | 17 | 17 | 2 | Not enough data |
| 27 | Washington | 0.25 | 19 | 18 | 3 | Not enough data |
| 28 | Minnesota | 0.23 | 13 | 12 | 3 | Not enough data |
| 29 | Michigan | 0.22 | 22 | 22 | 6 | -35.3% |
| 29 | New Jersey | 0.22 | 20 | 19 | 7 | +15.4% |
| 31 | Illinois | 0.17 | 22 | 21 | 4 | -39.7% |
| 32 | New York | 0.14 | 27 | 25 | 11 | -12.9% |
| Idaho | Not enough data | 9 | 7 | 0 | Not enough data | |
| Connecticut | Not enough data | 8 | 8 | 0 | Not enough data | |
| Iowa | Not enough data | 8 | 8 | 0 | Not enough data | |
| Massachusetts | Not enough data | 8 | 8 | 3 | Not enough data | |
| New Mexico | Not enough data | 8 | 8 | 1 | Not enough data | |
| Kansas | Not enough data | 7 | 7 | 0 | Not enough data | |
| Maine | Not enough data | 7 | 7 | 0 | Not enough data | |
| Rhode Island | Not enough data | 6 | 5 | 0 | Not enough data | |
| South Dakota | Not enough data | 6 | 6 | 1 | Not enough data | |
| West Virginia | Not enough data | 6 | 6 | 0 | Not enough data | |
| New Hampshire | Not enough data | 5 | 5 | 2 | Not enough data | |
| Montana | Not enough data | 4 | 4 | 0 | Not enough data | |
| Wyoming | Not enough data | 4 | 4 | 0 | Not enough data | |
| Delaware | Not enough data | 3 | 3 | 0 | Not enough data | |
| Hawaii | Not enough data | 3 | 3 | 0 | Not enough data | |
| Alaska | Not enough data | 2 | 2 | 0 | Not enough data | |
| District of Columbia | Not enough data | 1 | 1 | 0 | Not enough data | |
| North Dakota | Not enough data | 1 | 1 | 0 | Not enough data | |
| Vermont | Not enough data | 0 | 0 | 0 | Not enough data |
Source: NHTSA Fatality Analysis Reporting System (FARS), 2024 Annual Report File (released April 2026, subject to revision). How we count
Download Halloween road deaths in every state as CSV. Free to reuse with credit: “Fatal Crash Data, fatalcrashdata.com/halloween-fatal-car-accidents/, from NHTSA FARS (2024 Annual Report File).”
Every Halloween from 2015 to 2024
The deadliest Halloween of the ten was 2020, a Saturday, with 161 road deaths; the least deadly was 2019, a Thursday, with 83. Weekend Halloweens tend to be worse, but so are their normal days: compare the two columns.
| Year | Weekday | Deaths | Fatal crashes | Normal day (average) | Pedestrians | Children walking |
|---|---|---|---|---|---|---|
| 2024 | Thursday | 119 | 113 | 101.3 | 20 | 2 |
| 2023 | Tuesday | 103 | 90 | 96.7 | 22 | 0 |
| 2022 | Monday | 119 | 114 | 110.2 | 26 | 3 |
| 2021 | Sunday | 156 | 145 | 140.5 | 26 | 2 |
| 2020 | Saturday | 161 | 147 | 142.7 | 25 | 0 |
| 2019 | Thursday | 83 | 77 | 98.5 | 26 | 1 |
| 2018 | Wednesday | 95 | 89 | 94 | 20 | 4 |
| 2017 | Tuesday | 91 | 86 | 91.7 | 23 | 2 |
| 2016 | Monday | 107 | 100 | 107.2 | 23 | 2 |
| 2015 | Saturday | 129 | 108 | 117.8 | 25 | 3 |
Source: NHTSA Fatality Analysis Reporting System (FARS), 2024 Annual Report File (released April 2026, subject to revision). How we count
Questions about Halloween fatal car accidents
- How many car accidents happen on Halloween?
- NHTSA's file counts only crashes in which someone died: 1,069 on the ten Halloweens from 2015 to 2024, about 107 a year, which killed 1,163 people. Crashes with injuries or damage only are not in it; those totals come from state police and transportation departments, and no national count of them exists by day. Source: NHTSA's Fatality Analysis Reporting System (FARS).
- How many people die in car crashes on Halloween?
- 1,163 people died on US roads on the ten Halloweens from 2015 to 2024, 116.3 a year. That is 6% more than on the same weekday in the three weeks before and after (1,101 over ten such days). 236 of them were pedestrians.
- Is Halloween the deadliest day for children walking?
- Yes, by date: from 2015 to 2024, 1.9 children aged 14 and under were killed as pedestrians on October 31 in an average year, more than on any other date; the next is December 4 with 1.3. On an average day the figure is 0.54. The counts are small (19 children over ten Halloweens), so one bad year moves them.
- What time of day is most dangerous on Halloween?
- For pedestrians, the evening. From 6 pm to midnight, 130 pedestrians were killed on the ten Halloweens, against 103 on the same weekday in the three weeks before and after. The worst hour was 7 pm to 8 pm, with 37.
- Which state has the most Halloween road deaths per resident?
- Kentucky, with 0.71 road deaths per million residents per Halloween from 2015 to 2024, against 0.35 across the United States. New York has the lowest of the 32 states with enough deaths to rank, at 0.14.
How the Halloween figures are counted
Every fatal crash NHTSA dated October 31, from midnight to midnight, in the 10 years from 2015 to 2024. The normal day it is set against is the same weekday in each of the three weeks before and after (October 10, 17 and 24, November 7, 14 and 21), so the weekday and the season match; their deaths are divided by six to stand for one day per year. Four weeks after would sometimes be Thanksgiving, so the window stops at three. Children are pedestrians aged 14 and under, NHTSA's own age band for children. Rates per million residents use Census Bureau population estimates summed over the ten years. 2024 is the latest year NHTSA has published; the 2025 file is expected in spring 2027. The full method.