In May, America’s Safety Streak Cracked. Twenty-Five States Are Going the Wrong Way.

I ran the state-level numbers three times because the first two felt wrong. They weren’t. NSC’s preliminary May 2026 crash data shows 25 states with more traffic deaths than May 2025, against 20 states and D.C. that improved.^[1] Four states held flat. That is the first month this year where a majority of states moved backward.

Rewind four months. NHTSA released the Q1 2026 estimate in July: a fatality rate of 0.99 per 100 million VMT, the second-lowest first quarter in American history.^[2] Thirty states improved. The press releases practically glowed. Then April’s state-by-state data landed. Twenty-eight states still improving, but 20 had flipped to increases. By May, the majority crossed over. More states are now adding deaths than subtracting them.

The arithmetic of national averages hides this because the states still declining are big ones. Illinois shed 129 deaths through May, a 28% drop. Tennessee cut 60. Oklahoma cut 51.^[1] Those reductions carry enough absolute weight to keep the national line pointed downward even as the map fills with red. It is the statistical equivalent of a stock index rising while most stocks fall: a handful of large-cap names dragging the headline number in a direction that misrepresents the broader market.

California is the sharpest contradiction. Through May, it posted 242 more deaths than the same period last year, a 25% increase.^[1] This is not a rounding error. California has more registered vehicles than 22 other states combined, and its absolute increase alone is larger than the total annual traffic death toll in seven states. If you believed the national number meant your state was getting safer, and you live in California, you believed a fiction.

Smaller states exaggerate the percentage picture in ways that are real but misleading. D.C. shows +129% on nine additional deaths. Rhode Island is up 56% on five. North Dakota added 11 and landed at +42%.^[1] A single multi-fatality crash can swing those percentages by double digits. But California at +25%, Mississippi at +13%, Kentucky at +11%? Those are large-population states with enough crashes to smooth out statistical noise, and the trend is real.

Nobody can explain the mechanism. Speed enforcement intensity varies by state. ISA (intelligent speed assistance) remains unregulated in the U.S., and IIHS just confirmed this year that no minivans and no small pickups earned safety awards under 2026 criteria.^[3] Fleet age continues to rise nationally, but there is no evidence that aging fleets are concentrated in the states going backward. The correlation between state-level improvement and any single policy variable is weak enough that nobody willing to be honest can claim credit for the declines or assign blame for the increases.

The Strongest Counterargument

NSC estimates are preliminary. May 2026 data has the lowest maturity in the dataset. Historical patterns show that state-level estimates often revise upward over 12 months as fatalities are confirmed, but the comparison base (May 2025) has had a full year to mature. Comparing a raw preliminary number against a mature one can make the present look worse than it is. This caveat is real. It does not, however, explain a 25-to-20 split, and the maturity bias runs in the same direction for every state, which means the relative ranking is more stable than any individual state’s percentage.

Limitations

NSC preliminary estimates are not final FARS data. Substantial revision is expected over the coming year. Small states can swing wildly on a few crashes. The 25-vs-20 split is unweighted: it counts Wyoming and California as equals. A population- or VMT-weighted analysis would tell a different story. No causal mechanism is identified here because none is defensible with monthly preliminary data. This article reports the state-level trend, not its explanation.

What This Means for You

If you live in California, Mississippi, Kentucky, Iowa, or Delaware, your state’s roads are measurably more dangerous than they were a year ago. The national headline does not apply to you. Check your state’s DOT or SHSO for local enforcement campaigns and adjust accordingly. If you drive a vehicle without standard AEB, lane-keeping assist, or blind-spot monitoring, those features are the single largest risk-reduction purchase available to you, regardless of which direction your state’s trend line is moving.