The market is pricing the next combined ratio. The multiple is paying for something else.

• The premium isn't cyclical. Progressive trades at 4.4× book versus Allstate's 1.8× — at near-identical FY25 ROEs of 40.4% and 39.5%. That spread is not paying for margin; it is paying for one variable: the regulatory durability of per-driver pricing accuracy.
• The floor is structural. A $97.4B float at a 4.1% book yield delivers ~$4B of recurring investment income against $30.3B of equity — a ~13% ROE floor before any underwriting profit, locked in by a 3.4-year portfolio duration through 2028.
• The break is regulatory. California already excludes Snapshot from rating. Florida's excess-profits formula cost $1.22B in FY25. A propagation event in Texas, New York, New Jersey, or Colorado would put the durability premium at risk regardless of the next combined-ratio print.

March's monthly snapshot widened margin and accelerated growth — but the Q1 YTD print was flatter.

• The freshest data point cuts the other way — for one month. The April 15, 2026 release showed a March-month combined ratio of 88.8 (improved 2.1 points YoY), March NPW +10%, March net income +36%, with PIF +9% and Direct auto PIF +12% at March 31. Q1 YTD was less clean: CR 86.4 (0.4 pts worse than Q1 2025), NPW +6%, net income +10%. The monthly print breaks the cycle-roll signature; the quarter does not yet confirm it.
• The bear's mechanism is real. PIF growth has halved over five quarters (18% → 9%), trailing 12-month Policy Life Expectancy ran −7% in personal auto, and management cut personal-auto rates ~1% in FY25 into elevated tariff-driven severity risk. That sequence preceded FY22's collapse from $7.4B to $1.2B of operating income.
• The print refutes the mechanism, not the destination. The lean: long, wait for confirmation. Conversion to high conviction requires two more clean quarters — combined ratio sub-92 with PIF growth above 7% — across the Atlantic hurricane season ending November 30, 2026.

FY25 was the cleanest year in company history — and three of the inputs don't repeat at scale.

The 40% ROE pulled three temporary tailwinds: $1.39B of favorable prior-year reserve releases worth 1.7 points of CR, a 1.8-point catastrophe load versus a 3-to-4-point normal worth another 1.5 points, and a 4.1% book yield from bonds bought during the 2023-24 rate spike. Strip those out and underlying ROE runs 25-28%. The float is the part that does repeat — $97.4B at 3.4-year duration locks roughly $4B of recurring investment income through 2028 regardless of how underwriting prints.

The 32% drawdown isn't a market move — it's the market starting to price Allstate-style mean reversion.

• Stock-specific repricing. Peaked $291 in March 2025, troughed $190 in May 2026 — a ~35% drawdown over ~14 months against a flat tape. Death cross active since August 1, 2025. Price sits 9% below the 200-day at $217, with the 50-day at $199 still rolling.
• Distribution signature. The new 52-week low printed on 5.4M shares — 1.8× average daily volume. RSI 48 is neutral after touching 23 oversold in late October 2025. MACD oscillates without thrust. None of it looks like a constructive bottom.
• The catalyst the chart is digesting. Allstate's FY25 Auto combined ratio of 85.0 (total Property-Liability 85.2) ran below Progressive's companywide 87.4 — the first time Allstate's auto book has come in tighter than Progressive's companywide ratio in five years. Whether that's a mechanical snap-back from Allstate Auto's 103.4 FY23 / 95.0 FY24 prints or a structural segmentation catch-up is the bull-versus-bear hinge.