Progressive published its September results and its third-quarter summary on Wednesday, October 14, before the market opened. Net premiums written: $22.3 billion, up 4%. Combined ratio: 88.7, against 89.5 a year ago. Net income fell 3%, to $2.54 billion, or $4.37 per share.
There may be errors or inaccuracies. Found something that looks wrong? Write to me and I will fix it.
What Was Reported
| Q3 2026 | Q3 2025 | Change | |
|---|---|---|---|
| Net premiums written | 22,301 | 21,384 | +4% |
| Net premiums earned | 21,869 | 20,849 | +5% |
| Total revenues* | about 23,192 | about 22,512 | +3% |
| Underwriting income* | about 2,473 | about 2,190 | +13% |
| Investment income* | 1,035 | 924 | +12% |
| Realized gains (losses) before tax | Loss of 170 | Gain of 295 | |
| Net income | 2,542 | 2,615 | -3% |
| Diluted EPS (dollars) | 4.37 | 4.45 | -2% |
| Combined ratio | 88.7 | 89.5 | -0.8 pts |
| Policies in force at period end (thousands) | 40,489 | 38,078 | +6% |
In millions of dollars, except per-share figures, combined ratio and policies. Source: Progressive's results release for September and the third quarter, October 14, 2026. Rows marked with an asterisk are not a quarterly column in the release: they were calculated from the nine-month figures less the first half, taken from the June report.
Progressive reports monthly, so the release covers September and the quarter together. September on its own: premiums of $7,255 million, up 2%, net income of $630 million against $305 million a year ago, and a combined ratio of 90.0 against 100.4. The release contains no guidance, no management quote and no cash-flow statement.
The Refund That Left the Base
The combined ratio improved by 0.8 points, but it only reads correctly against its base. In September 2025, Progressive booked a $950 million expense for a refund to private-passenger auto policyholders in Florida, on profit above the statutory cap. There is no such item in 2026, and the release flags no other one-offs.
the refund accounts for 4.6 points of the combined ratio in the third quarter of 2025. Without it, the ratio a year ago would have been about 84.9, against 88.7 now, and underwriting income a year ago would have been about $3.14 billion, against about $2.47 billion this year - a decline of about 21%. These are not company metrics but calculated from the release's figures.
The breakdown shows where it comes from: the loss and loss adjustment expense ratio (loss/LAE) rose to about 68.2 from 64.5, and the expense ratio stayed at about 20.5 in both quarters. Together they come to about 85.0 a year ago, and the gap up to 89.5 is the refund. Against the second quarter of 2026, when the ratio was 87.3, this is an increase of 1.4 points. Year to date the ratio stands at 87.5.
Catastrophe losses added about 1.8 points to the ratio in the quarter, and 2.2 points year to date. Reserve development was favorable: about $534 million in the quarter, of which $143 million in September, and $1,536 million year to date.
Growth: More Policies, Slower Premium Growth
Policies in force rose 6% to 40.5 million at the end of September. Direct auto rose 8%, to 16.8 million, and agency auto 7%, to 11.3 million. Special lines (special lines) rose 6%, to 7.4 million, and commercial 5%, to 1.25 million. Property was flat at 3.65 million. Net premiums written grew more slowly than the number of policies: 4% against 6%, and 2% in September alone.
Profitability by line, year to date: agency auto 85.1, direct auto 89.6, property 76.3, and 87.2 across the whole personal segment. The faster-growing line, direct auto, is also the one with the higher combined ratio. In September alone the gap stood at 92.7 against 88.6.
In commercial lines the picture is different: a combined ratio of 89.0 year to date and about 92.6 in the quarter alone, against about 88.0 in the personal segment. Net premiums earned in commercial fell 2% year to date, even though its policy count rose 5%.
Investments, Capital and Buybacks
Investment income rose 12%, to $1,035 million. On the line beneath it the gap is large: a pretax realized loss of $170 million, against a gain of $295 million a year ago - a gap of $465 million between the two quarters. In September alone the realized loss was $231 million, against a gain of $138 million.
The total return on the investment portfolio was negative in September, 1.5%, and also year to date, 0.3%, against a positive return of 6.1% a year earlier. The net unrealized pretax loss on fixed maturities stood at $3,455 million at the end of September, and worsened by $1,656 million compared with August.
Diluted shares fell 1%, to 582.2 million. In September the company repurchased 969,262 shares at an average price of $207.92. Shareholders' equity stands at $34,485 million, book value per share at $59.51, debt-to-capital at 19.6%, and trailing-12-month return on equity at 34.4%.
What the Release Leaves Out
There is no guidance for the fourth quarter or for the full year, and the monthly commentary adds no remarks on September's results. The full quarterly report, the 10-Q, is due on November 2, and that is where the cash-flow statement will appear. The management Q&A call is scheduled for November 3 at 9:30 a.m. New York time, and October results will come out on November 18, before the market opens.
Versus Expectations, and the Stock
According to Zacks and AP data published on Yahoo Finance, nine analysts expected earnings of $4.46 per share. Progressive does not publish adjusted earnings, so the conclusion depends on the basis of comparison: AP presents adjusted earnings of $4.60 and describes them as above expectations, but reported GAAP earnings, $4.37, are below them. $4.60 comes from adding the realized loss, after 21% tax, to reported earnings.
On revenue the picture is reversed: six analysts expected $23.41 billion, and both reported revenue, about $23.19 billion, and adjusted revenue, $23.36 billion, are below that. Zacks' preview from October 12 estimated net premiums earned at $21.9 billion, and that is indeed what was reported, rounded, and the combined ratio at 89.4. The actual ratio, 88.7, is better than that forecast.
The stock closed the report day at $213.59, down 4.17% from $222.88 the day before. As of the October 13 close, the eve of the report, the stock was down about 2.1% year to date.
הזווית שלי
דעה אישית של אילן אברמוב - לא ייעוץ ולא המלצה
What stopped me in this report is not the 88.7 combined ratio, but what it is compared against. The ratio of a year ago, 89.5, contains 4.6 points of a one-off refund to policyholders, and without it the picture flips: from 84.9 to 88.7, with the loss ratio rising and the expense ratio staying put. The headline says improvement, and the breakdown says the improvement comes from the base and not from operations.
You can read this in two ways. One: a combined ratio of 88.7 is still very profitable, return on equity is 34.4%, and the policy count grew 6% - underwriting margin is narrowing from an exceptionally high level while the company grows, and that is a normal cycle. Two: premiums are growing more slowly than the policy count, only 2% in September, losses per premium dollar are rising, and underwriting income without the refund fell. Both readings fit the same numbers.
What will decide between them is not this quarter but the next: the 10-Q and the call in early November, and October results on the 18th. There I will look first at the loss ratio and at premium per policy.
(To be clear: this is my personal opinion only. Nothing here is a recommendation to take any action, and there is no forecast here of the price of any security.)
Ilan Abramov, the author, or anyone on his behalf may hold some of the securities or funds mentioned in this article.
