Hartford Calls It a Step Change. The Tape Calls It a Question.
A 17.9% core ROE and a 12% dividend hike are doing the talking — but workers' comp frequency and the first small-business new-business decline in ~20 quarters are doing the telling.
HIG · Earnings Call · 2026-10-27
The quarter that looked like all the others
On the surface, The Hartford's third quarter was another iteration of a machine this management team has trained the market to expect. Core earnings of $956 million, or $3.50 per diluted share, a trailing-twelve-month core ROE of 17.9%, and book value per share ex-AOCI of $82.47, up 16% from a year ago. Capital return was mechanical: $475 million of buybacks for 3.4 million shares, $4.37 billion left on the authorization, and a common quarterly dividend raised 12% to $0.67 — continuing a decade in which the annual increase has averaged roughly 11%.
That is the flywheel. It is also why the stock is only 8.4% below its July 29 peak of $145.68 and up 3.1% over the last 90 days, rather than gapping on bad news. But two things on this call genuinely changed, and both lean against the growth narrative the company has spent six quarters polishing.
The word management chose: 'step change'
This quarter's curated keyword set is dominated by loss-frequency vocabulary — increase in frequency and claim frequency both crack the top 30 for the first time, alongside the step change framing Chris Swift used to describe it. This is a new theme, not boilerplate. Six months ago, on the April call, Swift was asked directly about workers' comp pricing and said flatly: “I don't see a price increase coming anytime soon because, quite honestly, the book is behaving pretty well.” — Christopher Swift, Chairman and Chief Executive Officer · 2026-04-24 Go back further, to October 2025, and it was “steady as she goes” — Christopher Swift, Chairman and Chief Executive Officer · 2025-10-28 on a highly profitable line.
Now the line is not behaving. Adin Tooker, the incoming CEO, was emphatic that this is not a Hartford problem:
We see it across the book in all segments. It is not specific to any industry. It is not specific to any geography. It is not specific to size or risk. No specific injury type. Really what we are seeing we think is a broad based move in frequency that is not unique to our book.
His causal story — rising US healthcare costs, the expiration of ACA subsidies, worker affordability — is plausible and, notably, the same mechanism the market is debating in health insurance. The accounting: workers' comp and non-CAT property each contributed 'a little less than 4 points' to the year-over-year deterioration in Middle & Large's underlying combined ratio of 99.5. Pulled through to total Business Insurance, comp is a little over a point this quarter — but CFO Beth Costello was careful to flag that the quarter “included a bit of a true up for first and second quarter” — Beth A. Costello, Chief Financial Officer · 2026-10-27, and the nine-month run rate is about half a point.
Swift's forward view is the crux of the argument with the market:
Our belief is it is a 1-time step change here in 2026, and we will have to see how things play out in 2027. But we are not expecting sort of another year of elevated frequency in 2027 as we sit here today.
Long-tail reserving is precisely the discipline where 'one-time' is a word you use before you know. The company is putting price into 2027 filings, and severity is still assumed at a 5% starting point — untouched. That's the tension: management says the frequency move is exogenous and temporary; the response is a pricing and underwriting action, which is what you do when you are not entirely sure.
Small business: the engine that stopped adding
The call opened on the franchise's greatest hits — 78% of small business submissions now bindable straight through, an eighth straight year ranked first for small business digital experience with a 20-point lead, growth at twice industry levels, and a small-business underlying combined ratio of 85.2. Then Tooker quietly took the shine off:
“You will see our new business went negative for the first time in probably in 20 something quarters. It is the first time we did not grow our new business.” — Adin Morris Tooker, President and Chief Operating Officer (incoming CEO) · 2026-10-27
That is a real break from July, when the same executive said “I really feel confident about our ability to maintain margins and grow at a similar pace.” — Adin Morris Tooker, Head of Middle & Large Commercial and Global Specialty · 2026-07-24 The 85.2 print also benefited from favorable non-CAT property inside the segment, and Costello declined to extend it into the fourth quarter. Tellingly, Small Business was a top-15 keyword for this company last quarter and top-3 through most of 2025; it does not appear in this quarter's top 30 at all. The celebratory vocabulary has receded while the substance moved to loss picks.
What management stopped talking about
The prior quarter's vocabulary was casualty-reserve vocabulary: Non CAT and CAT property ranked first and second, with general liability reserves, commercial auto and umbrella lines all in the top ten after a $46 million GL strengthening. On this call that theme simply vanishes. Asked directly, Costello answered twice that the quarterly reserve review “did not adjust our booked reserves as we looked at the activity” — Beth A. Costello, Chief Financial Officer · 2026-10-27. Core favorable prior-year development was $97 million — modest — while the headline $594 million included a $497 million benefit from commuting the A&E adverse development cover, excluded from core. A quiet reserve quarter is good news; it also means there is no cushion left to absorb a frequency surprise.
The quieter levers
Personal Insurance is the improving story: an underlying combined ratio of 87.1, better by 2.9 points as earned pricing outran loss trend, with homeowners new business up 20% as the modern product rolls into the agency channel. Top line is the problem — written premium fell 4%, auto down 9% — and Melinda Thompson was candid that the plan is retention plus selective new business, not a manufactured growth push. On the Meta 'Muse' agentic-AI question, Tooker noted only 3% of small business is direct and has been static for years: very early days.
Employee Benefits delivered a 6.4% core earnings margin with a 76.3 group disability loss ratio, up 5.7 points on higher short- and long-term claim incidence, elevated behavioral health, and still-high paid family leave utilization — a theme that has now run for several quarters, and disability loss ratios remain a top-five keyword. What is new is the closed acquisition of Equitable's benefits business: roughly $500 million of annual premium, dental and vision added, about $30 million of annual intangible amortization in core, and modest dilution in 2027 before scaling. Management reiterated the 6-7% long-term margin range and 15-16% return on tangible equity.
The tape and the multiple
The stock's full history is a long compounding line now consolidating: the 90-day tape shows a modest grind higher and an 8.4% drawdown from the summer peak. Meanwhile the fundamentals — still second-quarter figures, filed July 23 — show revenue of $7.26 billion, up 4% year over year and a price-to-earnings multiple of 8.3x, down 25% year over year. Interest coverage of 24.6x is the quiet enabler behind the 'we will not chase growth' refrain. Record earnings plus a compressing multiple is the market's way of saying it believes the ROE is cyclically high — and that the workers' comp frequency question is not fully answered by the word 'one-time.'
None of this is a global theme. The market's tape this quarter is crowded with AI data center and tariff-refund noise; Hartford is not riding any of it. This is a company-specific, idiosyncratic quarter: an excellent operator telling you, credibly but unverifiably, that the only thing that went wrong is temporary. Watch two numbers from here — fourth-quarter non-CAT property, and whether the 2027 comp filings actually get filed.