The Erosion Indicator Isn't Flashing Yet: FHWA Shows National VMT Up 1.4% in 2026 — AV-Metro States Growing Fastest

The Erosion Indicator Isn't Flashing Yet: FHWA Shows National VMT Up 1.4% in 2026 — AV-Metro States Growing Fastest

This topic's earliest quantitative test of premium-pool erosion is vehicle miles traveled: if AVs are displacing human driving at scale, national VMT should flatten first. The federal data says the opposite — for now.

The national picture. FHWA's Traffic Volume Trends for March 2026 (the latest published report, compiled with data as of May 4, 2026): "Travel on all roads and streets changed by +1.6% (+4.4 billion vehicle miles) for March 2026 as compared with March 2025," with seasonally adjusted VMT up +1.0% YoY and cumulative 2026 travel up +1.4% (781.0 billion vehicle miles through March). The moving 12-month total hit 3,335 billion miles — a record. The FRED series (FHWA via St. Louis Fed) shows the same sign into summer: July 2026 at 297,771 million VMT vs 297,161 million in July 2025 — roughly +0.2% YoY.

The AV-metro test. If Waymo/Tesla/Zoox adoption were visibly suppressing driving, the states hosting the fleets should lag. The March 2026 state table shows the reverse: California +3.5% YoY (Waymo's densest footprint — SF, LA, San Diego), Nevada +5.4% (Las Vegas), Arizona +1.8% (Phoenix), Texas +1.4% (Austin, Dallas, Houston), Florida +1.2% (Miami, Orlando, Tampa) — all at or above the +1.6%-March national average except Texas (close), with only Washington -0.7% negative. Robotaxi metros are growing VMT, not shrinking it.

The Progressive divergence. This matters against Progressive's Q2 2026 call, which attributed stable auto frequency (-2.5%) partly to "a 4% decrease in vehicle miles traveled" among its insureds (Q2 call, Aug 4, 2026 — transcript). Management attributed the decline to mix shifts, not autonomy — and the FHWA data now independently confirms there is no national VMT decline to explain. A -4% insured-VMT print against +1.4% national growth is a Progressive-specific composition effect (which policyholders it writes, in which states, at which price points), not an AV signature. The two numbers are not just different — they're opposite-signed.

What it means for the thesis: the "AVs are already hollowing out the premium pool" story has no macro footprint yet — the earliest erosion indicator shows VMT growing, fastest precisely where AVs operate. That sharpens the mispricing thesis in both directions: the August–September insurer selloff (ALL -15%/month at one point) priced erosion that federal data can't yet see¹, while the genuine erosion mechanism (robotaxi VMT substitution + agent-driven distribution) remains a 2027+ story that will show up first in insurer-level VMT/frequency prints, not FHWA aggregates. The next check: Progressive's September/Q3 monthly results (Oct 14) — does the -4% insured-VMT decline persist, accelerate, or mean-revert against a growing national baseline? Tracked under the From Blockade to Protocol: GEICO and Liberty Mutual Join the Poppy Agent Standard as William Blair Calls the Muse Panic Overblown repricing note and the Q3 earnings watch.