Business

Know the Business — The Progressive Corporation

Progressive is a scaled, data-rich auto insurer that compounds book value through two stacked profit pools — disciplined underwriting and a $97.4B float invested at AA−. The right way to underwrite this stock is not as a "cheap insurer at 12x earnings"; it is as a high-ROE auto specialist whose moat is per-driver pricing accuracy and whose central risk is regulatory ceilings on segmentation and rate. The market is most likely overestimating how durable a 40% ROE is across the cycle and underestimating how much of FY25's headline result came from a benign cat year, a $1.4B reserve release, and a 1.7-point boost from policy life trends that have not stabilized.

1. How This Business Actually Works

Progressive is two profit machines stacked on top of each other. Underwriting earns a margin equal to 100 minus the combined ratio; investments earn a yield on the float — the policyholder money sitting between premium collection and claim payment. Together, the two pools produced $11.3B of net income in FY25 on $87.7B of revenue. Neither pool works without the other: underwriting alone would be a low-margin business, and the investment pool can't exist without disciplined writing of premium to fund it.

What actually drives incremental profit. Two levers, in this order. First, rate accuracy by cell: when PGR can quote a 38-year-old preferred driver more accurately than State Farm or Allstate, it wins that policy at a price that still leaves a margin and lets a competitor write it at a loss. The customer is sticky, and the float compounds. Second, expense ratio leverage at scale: every dollar of additional NPW spreads $5.1B of advertising and a roughly fixed corporate base across a larger book, so each preferred-tier policy added is worth more than its average. Bottlenecks are regulatory (state rate filings cap how fast price can move; Florida caps total profit) and capital (premiums-to-surplus 2.9:1, so growing $1 of NPW requires roughly $0.34 of surplus).

Bargaining power. PGR has very little against a single customer — auto coverage is mandatory and shopping is frictionless. It has substantial power against an agent (40,000 of them, none indispensable) and meaningful pricing leverage against catastrophe reinsurers because of its size, but those reinsurers re-set every June 1. The TNC (Uber) contract is the one place a single counterparty has real leverage: it is 14% of Commercial Lines premium.

2. The Playing Field

PGR earns the highest ROE in the U.S. listed P&C peer set while trading at the lowest combined ratio variability. Allstate posts a comparable headline ROE but only after a brutal FY23 (combined ratio 104.5); Travelers and Hartford are commercial-skewed multi-lines; Chubb runs a more diversified, lower-ROE global P&C; Erie is a management-fee structure that prints a different kind of multiple. The largest competitors — State Farm, GEICO (inside Berkshire), USAA, Liberty Mutual, Farmers — are mutuals, members, or conglomerate subsidiaries and do not appear in this table at all, which is why national share matters less than relative position inside each state.

What this peer set actually reveals. Chubb is the cleanest comparison for underwriting discipline: it ran sub-90 combined ratios through the same 2022–2023 industry blowout that pushed Allstate to 106.6 and PGR to 95.8, but it earns a 14% ROE because its book is diversified and capital-heavy. Allstate is the cleanest comparison for personal auto cyclicality: same business, less segmentation, and the FY22–23 cycle nearly broke its earnings. PGR's accomplishment is sitting between those two — Chubb-like discipline with Allstate-like ROE — and it does that by writing where it segments best (auto) and staying small where it does not (homeowners, #12 nationally with the book actively pruned in volatile states). The cleanest read of "what good looks like" in this industry is: combined ratio below 96 across a five-year window with PIF growth at or above industry, and PGR is the only listed peer that hits both consistently.

3. Is This Business Cyclical?

Yes, materially — and the cycle hit where it always does: underwriting margin. PGR's revenue rarely declines (auto insurance is mandatory; rate hikes inflate the top line even as customers shop), but operating income collapsed 78% from $7.4B in FY20 to $1.2B in FY22 because loss-cost inflation outran rate filings. Net income fell from $5.7B to $0.7B in two years, and book value per share dropped from $31.0 to $27.1 — a -13% real drawdown in the equity an insurer is supposed to compound. Anyone underwriting this business at the FY25 ROE of 40% needs to know FY22's ROE was 4%.

Where the cycle physically hits. Not demand (mandatory, low-elasticity), not channel mix (slow-moving), and not investment income (which has actually been a tailwind as bonds roll into higher coupons — book yield rose from 3.5% in FY23 to 4.1% in FY25). The pressure point is the loss ratio: severity inflation in body shops, medical, and attorney-represented bodily injury claims runs faster than the state-by-state rate filings approval process allows price to move. When that happens, combined ratios spike, reserves develop adversely (FY23 took a $1.1B unfavorable reserve hit — a 1.9-point CR drag), and surplus shrinks; carriers then take aggressive rate, shoppers flood the market, frequency falls, and combined ratios snap back hard (FY24–FY25). This is the playbook for U.S. personal auto. The next test is whether 2025 tariffs on imported auto parts re-ignite severity inflation in 2026 — PGR explicitly flagged this risk.

What is not cyclical at PGR. PIF growth (FY24: +18%, FY25: +11%), float compounding (portfolio grew from $80.3B to $97.4B in one year), and the segmentation moat. These are the through-cycle compounders. Underwriting margin is the cyclical layer on top.

4. The Metrics That Actually Matter

Five numbers tell you more about Progressive than the P&L does. Skip ROE, P/E, revenue growth — they mix three things together and obscure the actual operating story.

Why these five matter more than the obvious ratios.

• Combined ratio is the only number that tells you whether the core business is profitable. It strips out investment income, share buybacks, and tax. PGR runs to a self-imposed 96 ceiling — when CR drifts above 96, growth slows by management design.
• Policies in force growth strips rate hikes out of revenue growth. NPW can rise 19% off pure rate increases (as in FY23); only PIF growth confirms PGR is actually winning customers — and PIF growth turned negative in early FY23 before recovering hard in FY24.
• NPW growth is the leading indicator: premium written this year becomes earned premium next year, and it grows the float that funds the second profit pool.
• Investment book yield is the durable tailwind in the current rate regime. Unlike the underwriting margin, the book yield moves slowly and is locked in for years on existing bonds.
• Cat points on combined ratio is the honest number behind a flattering headline. FY25's 87.4 combined ratio was helped by 1.8 cat points; a normal cat year of 3–4 points would put underlying CR closer to 89–90. Always normalize.

What to de-emphasize: P/E ratio (distorted by reserve releases and one-time items — see FY22's 109x P/E), ROA (irrelevant for a leveraged spread business), and dividend yield (PGR pays a variable annual dividend tied to Gainshare).

5. What Is This Business Worth?

The right lens is normalized underwriting earnings plus investment income on float, capitalized at a multiple that reflects regulatory and cycle risk. Not P/E (FY22 vs FY25 are not comparable), not P/B (book value sits below intrinsic because reserves are conservative), not EV/EBITDA (insurance EBITDA is meaningless). Two value drivers do the work. Driver one: how cheaply PGR can produce a 96-or-better combined ratio on a growing premium base. Driver two: the spread between what its float earns and what it costs to acquire (effectively zero — float is "free leverage" if underwriting is profitable). Everything else — segments, geographies, products — is texture.

Sum-of-the-parts is not the right lens here. Personal Lines is 87% of NPW, Commercial Lines 13%, and the property book is tiny and shrinking — these segments share a customer-acquisition machine, a reserves base, a regulatory framework, and the same float. Valuing them separately would falsely imply they could be cleanly separated; in practice the commercial-auto book benefits from the same telematics expertise, the same actuarial talent, and the same brand spend as personal auto. The one quasi-SOTP question worth thinking about is the $97.4B investment portfolio: at a 4.1% pretax yield the recurring investment income alone is ~$4B per year, which against a $30B book of equity is a 13% ROE before any underwriting profit at all. That is the floor under the business and the reason PGR is unlikely to become a value trap unless the underwriting machine truly breaks.

What would make the stock cheap. A return to combined ratios in the mid-90s coupled with PIF growth falling below 5% — in that scenario, the FY25 ROE compresses fast and the multiple has historically migrated toward Allstate's 5.5x P/E and 1.8x P/B band. What would make it expensive. Sustained sub-90 combined ratios, PIF growth holding above 10%, and the segmentation moat surviving California-style rating-variable scrutiny. The market is currently pricing roughly in the middle (4.4x P/B, 11.8x P/E) — which suggests it neither believes the FY25 result is the new normal nor that the cycle is about to break.

6. What I'd Tell a Young Analyst

Watch combined ratio first; ignore the P/E. P&C insurers are spread businesses; the headline P/E is distorted by reserve releases (FY25 favorable $1.4B = 1.7 pts), catastrophe variability, and capital actions. Build your model around a normalized combined ratio of 92–93 (4-pt UW margin) and a book yield assumption that decays as the curve flattens — not 12.6 pts of UW margin and 4.1% book yield extrapolated forever.

The real question is whether 40% ROE is repeatable. The honest answer: probably not. FY25 had three tailwinds simultaneously — benign cats (-1.8 pts vs ~3 pts normal), $1.4B reserve release (a 1.7-pt tailwind that is, by definition, not repeatable in the same direction every year), and 4.1% book yield that lags but reflects a high-rate environment. Strip those and the underlying ROE is closer to 25–28% — still excellent, but a different stock. Underwriting at 40% is how you lose money.

Where the market is most likely wrong. Not on the segmentation moat — that is well-understood. More likely on the regulatory risk: Florida's excess-profits cap already cost PGR $1.2B in FY25, and California's posture toward rating variables is the long-cycle threat to the entire pricing-accuracy thesis. Track NAIC AI/algorithm bulletins quarterly. Also track Policy Life Expectancy — it fell 7% YoY in personal auto and 12% in personal property; if shopping intensity stays elevated, the renewal margin (which is structurally higher than new-business margin) compresses.

What would change the thesis. A two-year stretch of combined ratios above 95 with PIF growth below industry. That combination would mean the segmentation advantage has narrowed, the cycle has turned, or both. Any one of those alone is noise; together they reset the framing.

The single number to put on the cover. Combined ratio less catastrophe load. FY25 = 85.6 pts. If that holds through a normal cat year (3 pts → 88.6), this is a remarkable business. If it drifts toward 92, this is an ordinary one.