Unum's Group Disability Benefit Ratio Breaks Guidance on PFML Incidence

Unum's group disability benefit ratio landed at 65.8% in the second quarter of 2026, above the 62% to 64% range management had guided to, and 360 basis points worse than the 62.2% posted a year earlier (Unum, July 28, 2026). The overage traces almost entirely to short-term disability incidence in newer paid-family-and-medical-leave states, not to the long-term disability book, which the company's own numbers show holding steady.

Chief Financial Officer Steven Zabel put a number on the split on the earnings call: short-term disability experience added roughly two points to the loss ratio relative to guidance, and "I would size that up as about 60%-70% of that was driven by PFML" (Steven Zabel, Unum CFO, Q2 2026 earnings call, July 29, 2026). Group disability segment adjusted operating income fell 17.4% to $103.1 million from $124.8 million, even as premium income grew to $827.5 million from $797.1 million (Unum, July 28, 2026). Growing premium alongside falling segment income in the same release is the signature of a loss-ratio miss riding on top of a book that is still selling well, a position a pricing actuary has to reprice out of rather than grow out of.

Where the Miss Actually Sits

The release itself is specific about the mechanism: "higher incidence in the short-term disability product line, primarily related to our paid family and medical leave products," with prior period pricing actions cited as a second, smaller contributor (Unum, Q2 2026 earnings release, July 28, 2026). Long-term disability, the line where a claim-severity or duration deterioration would show up as a genuine reserve problem, was not implicated. Zabel was direct about it on the call: LTD recoveries are running consistent with expectations and underlying LTD trends remain stable (Investing.com transcript, Unum Q2 2026 earnings call, July 29, 2026).

Persistency backs that read up rather than contradicting it. Group LTD persistency ran 91.1% through the first half of 2026 versus 90.6% a year earlier, a 50 basis point improvement, while group STD persistency jumped to 91.1% from 88.2%, up 290 basis points (StockTitan, citing Unum's Q2 2026 results, July 2026). A book retaining more of its existing clients while incidence experience worsens is not shedding bad risk through lapse. The deterioration is sitting inside blocks the carrier is keeping, which is precisely why it shows up as a benefit-ratio miss rather than a volume story.

A Second Benefit-Ratio Break, Filed the Same Day

The same earnings release carries a second, unrelated disability story easy to miss when reading only the US numbers. Unum's UK segment reported a benefit ratio, excluding reserve assumption updates, of 82.2% in the second quarter, up 720 basis points from 75% a year earlier, attributed to higher average claim size and increased claim incidence in the UK's group long-term-disability equivalent, Group Income Protection (Unum, Q2 2026 earnings release, July 28, 2026). Segment adjusted operating income in the UK fell to £15.3 million from £29.4 million. Management pinned the severity component on claims concentrated among higher-income employees, a population where income-replacement products with percentage-of-salary benefit structures generate disproportionately large individual claim payments once a claim occurs (Unum Q2 2026 earnings call, July 29, 2026).

Where the US group disability book broke on frequency in a young statutory product, the UK book broke on severity and incidence together in a mature one, at the same company, in the same quarter, disclosed on the same day. That is a useful natural experiment for a pricing desk that writes both product types, because the two breaks require entirely different fixes.

Statutory Immaturity Cuts Both Ways on the PFML Book

PFML is a young product line by design, not by accident. Thirteen states now run mandatory paid-family-and-medical-leave programs, with Delaware, Maine, and Minnesota reaching full operational status in 2026 alone and Virginia enacting its own mandate in April 2026 ahead of a later launch (The Standard, 2026). Each new state effectively resets the credibility of the incidence data underneath it, and Group Benefits president Chris Pyne flagged pent-up demand in newly launched states as claims activity management expects to normalize as awareness of the benefit matures, rather than persist at initial levels (Unum Q2 2026 earnings call, July 29, 2026). That is a testable claim, not a settled one, and a carrier underwriting business in a state that launched in the last 18 to 24 months has a thinner evidentiary basis for the pent-up-demand explanation than one underwriting in California or New Jersey, where the programs have run for the better part of a decade.

The mitigant is contractual rather than actuarial. Most PFML business carries short, largely one-year rate guarantees, which is precisely why Unum can move on the miss as fast as it says it will. Pyne pointed to January 1, 2027 as a key renewal date for resetting rates across a meaningful share of the book (Unum Q2 2026 earnings call, July 29, 2026), and management has already begun implementing double-digit rate increases at new business and renewal, with the fuller earnings benefit expected in 2027 once the bulk of the book has cycled through that reset (Investing.com transcript, Unum Q2 2026 earnings call, July 29, 2026). A frequency miss in a product carrying annual rate guarantees is a two- or three-quarter cash-flow problem. A frequency miss in a product carrying multi-year guarantees compounds toward something closer to a solvency-adjacent one, and Unum's own disclosure that only a small share of the PFML portfolio carries multi-year locks is the detail that keeps this quarter's overage from running into 2027 (BigGo Finance, Unum Q2 2026 earnings call summary, July 29, 2026).

The Assumption That Actually Moved

The two breaks call for different responses because they are different assumptions failing in different ways. The PFML miss is a frequency and credibility problem: state-level incidence tables built on one to three years of claims history in markets where awareness, take-up, and even the practical definition of a qualifying leave event are still shifting. It resolves the way credibility problems resolve, through more experience, tighter geographic and cohort segmentation, and rate guarantees short enough to let pricing catch up before the miss compounds across renewal cycles.

The UK Group Income Protection break is a severity and incidence problem inside a product that is otherwise mature, concentrated in a specific claimant segment where the benefit design, a percentage of pre-disability salary up to a monetary cap, means a shift in claimant income mix moves average claim size directly. That is closer to the kind of assumption drift group LTD pricing actuaries in the US watch for in their own book, one this site examined from a different angle in its review of the mental-health limitation and discount-rate pressure on group LTD rates: a change in who is filing, not just how many are filing, that a termination-rate or benefit-size assumption calibrated on last decade's claimant mix will not catch until the loss ratio already shows it. This site's earlier look at how disability incidence tends to run countercyclical to employment growth is the macro backdrop against which both of these company-specific breaks are playing out. And Unum's own group disability numbers arriving with a clean, product-level attribution of the miss is a reminder of how much a benefit-ratio decomposition can conceal or reveal, depending on whether a carrier discloses down to the driver or stops at the headline number.

What the Next Quarter Should Show

Unum's next disclosure point is the third-quarter call, expected in late October 2026, where the PFML rate actions should start showing partial benefit and where the UK's targeted pricing and underwriting response gets its first quarterly test against the 82.2% baseline. Pricing actuaries at competing group disability and group income-protection writers have two items to raise with reinsurers and brokers before then. First, whether their own PFML books carry the same short rate guarantees that let Unum reprice into 2027 rather than being locked into an inadequate rate for another full cycle; a book with a heavier multi-year-guarantee mix does not get the same escape hatch. Second, whether the claimant income mix in their own income-replacement products has drifted enough since the last experience study to justify pulling a severity review forward, rather than waiting for the next scheduled one, the way Unum's UK numbers suggest it should have.

Lincoln National, Guardian, and Standard Insurance had not yet reported second-quarter group disability results with comparable segment detail as of this writing. When they do, the read-through question is whether Unum's PFML pressure is a company-specific state-mix issue, tied to which of the thirteen active PFML states each carrier is overweighted in, or an early signal for every carrier writing into the same statutory footprint. Given how thin the claims history still is in the states that launched in 2026, that is a question the market will need at least two more quarters of disclosures to answer with any confidence, and it is one worth watching specifically in each carrier's newest-state loss ratios rather than in the blended segment number a headline release tends to report first.