AFL — Q2 2026 Financial Report Analysis

What happened

Aflac's reported profit for the three months to June 30, 2026 rose sharply — net earnings of $825 million versus $599 million a year earlier — while the insurance business itself earned less. Adjusted earnings, the measure that strips out investment gains and losses management doesn't control, fell 7.7% to $883 million. The gap between those two numbers is the whole story of the quarter.

Underneath, the two operating segments moved in opposite directions on margin. In Aflac Japan, measured in yen, profitability improved: pretax margin of 34.3%, up 2.3 percentage points, because claims came in lighter than a year ago. In Aflac U.S., premiums grew 2.3% but the share of premiums paid out as claims rose, and segment pretax profit fell 4.6%.

Two terms that carry most of the weight below. The benefit ratio is claims and benefits paid out divided by premiums collected — for a supplemental health insurer it is the single most direct read on whether policies are priced correctly, and a lower number means more of each premium dollar stays with the company. Net earned premiums are the premiums Aflac keeps after amounts handed to reinsurers; it is the revenue line that reflects the size of the book of business actually being carried.

Why reported profit jumped 38% while the business earned less

Three things pushed GAAP earnings up, none of them insurance results.

First, investment losses shrank. Aflac booked net investment losses of $153 million this quarter ($0.30 per diluted share) against $421 million ($0.78 per share) a year ago. That single line accounts for roughly $0.48 per share of the $0.52 GAAP EPS increase.

Second, the tax rate dropped. Pretax earnings rose 21.0% to $995 million, but net earnings rose 37.7%, because the combined U.S. and Japanese effective tax rate fell to 17.0% from 27.0%. The 10-Q attributes this to "the exclusion of foreign currency translation gains and losses on Aflac Japan U.S. dollar-denominated investments held in the Delaware Statutory Trust" — currency swings on those assets move pretax earnings but are not taxed, so a quarter with large currency gains mechanically lowers the reported rate. It is a structural feature, not a rate cut.

Third, and most visibly on a year-to-date basis: first-half total revenues rose 12.0% to $8,463 million from $7,558 million, which reads like strong growth and is not. Net investment losses for the half were $104 million against $1,384 million a year ago — a $1,280 million swing that sits inside the revenue line. Net earned premiums over the same six months fell 4.2%, to $6,562 million from $6,851 million. The headline revenue growth and the underlying premium trend point in opposite directions.

Takeaway: Almost none of the 38% jump in reported profit came from insurance. What actually held adjusted earnings per share roughly flat (-1.7%) while adjusted earnings fell 7.7% was the buyback: the diluted share count dropped 6.1% year over year, to 505.6 million from 538.4 million. Aflac is currently converting capital return, not premium growth, into per-share results — and with U.S. premium growth guidance just cut and Japan's premium base deliberately shrinking, that dependence is unlikely to ease in the near term.

Aflac Japan: better margin, smaller book — mostly on purpose

Japan is roughly two-thirds of group revenue, so how it is measured matters. In yen, net earned premiums fell 3.7% to ¥245.1 billion. The 10-Q is specific about the cause: approximately ¥4 billion from "an external reinsurance transaction established in the first quarter of 2026" and approximately ¥3 billion from "limited-pay products reaching premium paid-up status." Reinsurance means Aflac handed part of its risk — and the matching premium — to another insurer; paid-up status means policyholders finished paying on policies that remain in force. Both shrink the premium line without indicating lost customers. CFO Max Brodén gave the cleaner figure: Japan's underlying earned premiums, excluding reinsurance, paid-up policies and deferred profit liability, fell 1.4%, not 3.7%.

Margin improved for a reason with a shorter shelf life. Japan's benefit ratio fell 250 basis points to 64.0%, which the filing credits to "the impact of annual cash flow assumption updates performed in the third quarter of 2025 and higher reserve remeasurement gains" — in other words, reserves set last year are proving conservative against actual claims, partly helped by what Brodén described as continued favorable trends in cancer and hospitalization claims. Reserve releases are real earnings, but they are a comparison against last year's assumptions, not an improvement in current pricing, and the next annual assumption review lands in Q3 2026.

The FX distinction here is worth being exact about. Japan's yen pretax adjusted earnings rose 3.4% as reported, but the 10-Q's currency-neutral table shows a 2.1% decline. The difference is that Aflac Japan holds a large block of U.S. dollar-denominated bonds; when the yen weakens, the dollar interest income from those bonds translates into more yen and flatters the yen-based growth rate. Translated the other way, into dollars, Japan pretax adjusted earnings fell 6.2% to $741 million. Three readings of the same quarter — +3.4%, -2.1%, -6.2% — and only the middle one describes operating performance.

Also of note: Japan's annualized premiums in force fell 2.7% to ¥1.16 trillion, again driven by policies reaching paid-up status, and premium persistency slipped 100 basis points to 92.7%. Brodén attributed the persistency decline to "somewhat elevated lapse and reissue activity on recently launched products" — existing customers cancelling and rebuying into newer, broader coverage. That is a milder problem than customers leaving, but it does depress the reported figure.

Japan's sales mix is rotating out of cancer insurance

New annualized premium sales in Japan fell 5.6% in the quarter to ¥19.6 billion, against a high base: Miraito cancer insurance launched in March 2025 and inflated the prior-year comparison. Across the first six months, sales were up 7.0% to ¥37.3 billion.

The mix shift is larger than the headline decline suggests:

Cancer insurance dropped from nearly three-quarters of new sales to just over half in a single year, with the slack taken up by Anshin Palette (a medical product launched December 2025) and the refreshed Tsumitasu savings-type life policy. Cancer and medical products — what Aflac calls "third sector" — carry the margin structure the company is built on; Tsumitasu is a first-sector savings product with a different, generally thinner economic profile, and it is the product being ceded under the new external reinsurance arrangement. Management's stated logic is that Tsumitasu reaches a younger demographic it can later cross-sell third-sector cover to. That is a reasonable strategy and an unproven one; for now the mix rotation is a reason to watch Japan's margin rather than extrapolate this quarter's 34.3%.

Aflac U.S.: growth is fine, claims are the problem

U.S. net earned premiums rose 2.3% to $1,539 million on "improved sales and continued strong persistency," and annualized premiums in force rose 4.5% to $6.8 billion. Sales grew 2.6% to $349 million, driven by group voluntary benefits and network dental and vision products.

Profit went the other way. Pretax adjusted earnings fell 4.6% to $370 million and the segment's pretax margin fell 160 basis points to 20.9%, because total benefits and claims rose 7.0% — more than three times the premium growth rate. The 10-Q attributes this to "higher incurred claims for certain group products largely associated with the growth in net earned premiums," and Brodén was more specific: "an increase in incurred group disability claims in the quarter relative to favorable results in the previous quarter." Part of the 220 basis point benefit-ratio increase is therefore a comparison effect against an unusually good prior quarter rather than a new claims trend — but group disability is the line to watch, since it behaves more like an economically sensitive employment-linked exposure than Aflac's traditional accident and critical-illness products.

The quarter also contained a one-off help: a $26 million expense contingency release lowered U.S. segment expenses and added roughly $0.04 to adjusted EPS. Strip that out and the U.S. margin decline is slightly worse than reported.

The $4.8 billion portfolio switch, and what the investment line actually shows

The investment-loss detail explains what looks alarming at first glance. Gross losses on sales of available-for-sale bonds were $837 million in the quarter, against $98 million a year ago — but offset by $599 million of foreign currency gains on the same activity. This was deliberate: Brodén said the investment teams repositioned $4.8 billion of the portfolio through "switch trades" designed to "capture foreign currency gains to minimize market losses on lower yielding assets, reduce the risk of future FSA impairments, improve our ALM, and boost net investment income," with an expected run-rate benefit of over $50 million of additional annual net investment income and limited capital impact. In substance, accumulated currency gains were used to absorb the accounting loss on selling low-yielding bonds and rotate into higher-yielding ones.

Two items in the same disclosure are less comfortable and are not currency-related:

• Commercial mortgage loan credit losses of $76 million in the quarter, $126 million for the half — real provisioning against the loan book, up from $61 million and $114 million respectively.
• $24 million of impairments in the first half on real estate owned, taken because of "a material adverse change in occupancy," writing those properties down to a $179 million fair value.

Separately, variable investment income — the return on holdings whose income fluctuates rather than paying fixed coupons — came in $72 million, or $0.11 per diluted share, below Aflac's long-term return expectations. That is a meaningful drag inside the $1.75 adjusted EPS figure and, netted against the $0.04 expense-contingency benefit, means underlying adjusted EPS was running roughly $0.07 better than reported. Portfolio yields tell a mixed story: Japan's new money yield fell to 4.19% from 5.26%, while the overall portfolio book yield rose to 3.39% from 3.26% — reinvestment is still accretive, just less so than a year ago. Total investments and cash stood at $103.0 billion.

Capital: the buyback is doing the per-share work

Aflac returned $1.3 billion in the quarter — $983 million of share repurchases and $309 million of dividends — and $2.6 billion in the first half, against $2.3 billion in the first half of 2025. First-half repurchases totalled 17.5 million shares for $2.0 billion, with 96.8 million shares still authorized. The board declared a third-quarter dividend of $0.61, up 5.2% year over year, after reaching 43 consecutive years of dividend increases in 2025.

Capital metrics support the pace. Aflac Japan's estimated regulatory Economic Solvency Ratio — Japan's new economic-value-based capital test, first disclosed as of March 31, 2026 — was 226%, or 240% including the undertaking-specific parameter, down quarter over quarter "primarily driven by significant subsidiary dividends," i.e. cash moved up to the parent to fund buybacks. Combined U.S. risk-based capital was estimated slightly above 600%, adjusted leverage 21.8% against a 20–25% target, and holding-company unencumbered liquidity $3.3 billion versus a $1 billion minimum.

Reported book value needs the same FX care as earnings. Shareholders' equity was $30.3 billion, or $60.35 per share, up from $50.86 — but that includes a cumulative $10.4 billion increase from changes in the discount rate used to value insurance reserves (versus $5.6 billion a year ago), a $2.8 billion net unrealized investment loss, and a $5.0 billion unrealized currency translation loss. On the measure management itself uses, adjusted book value per share excluding currency remeasurement, the figure fell 4.1% to $41.22. Annualized return on equity was 10.9% reported; adjusted ROE excluding currency remeasurement was 16.6%.

Guidance and trajectory

Management's own numbers from the quarter:

• Aflac Japan benefit ratio: now expected at the high end of the 60–63% guidance range for full-year 2026, excluding the Q3 annual actuarial assumption review. Year-to-date is 63.4% — already above that range — so the guidance implies claims improve, or reserve releases continue, over the back half.
• Aflac U.S. net earned premium growth: cut to just below the 3–6% guidance range for 2026, from prior guidance of the low end of that range. The 2025–2027 three-year compound growth target of 3–6% was reaffirmed, which pushes the required catch-up into 2027.
• Internal reinsurance: the Japan target was reset from up to 10% of U.S. GAAP assets to up to 30% of FSA reserves — a materially larger allowance that will keep compressing Japan's reported net earned premiums while supporting return on equity.
• Portfolio repositioning: expected to add over $50 million of annual net investment income on a run-rate basis.

Our read: the earnings base is holding up better than the revenue line, but for reasons that are mostly non-repeating. Japan's margin gain came from reserve releases measured against 2025 assumptions, and those assumptions get refreshed in Q3 — the single most consequential near-term event for this company's reported profit. The U.S. segment has the cleaner growth story and the deteriorating margin, and its premium guidance was just cut while a three-year target was left intact, which is where the pressure now sits. Meanwhile the reinsurance programme is being expanded specifically to shrink the reported premium base further. That combination leaves per-share earnings growth resting on the buyback, which is being funded by moving capital out of Japan — visible in the ESR falling quarter over quarter. It is sustainable at current capital levels and it is not a substitute for premium growth. The two things worth watching in Q3 are the assumption review's effect on Japan's benefit ratio, and whether U.S. group disability claims normalise or hold at the elevated level.

Source: Aflac Incorporated Form 10-Q for the quarterly period ended June 30, 2026 (filed August 7, 2026), and the Form 8-K dated August 6, 2026, Exhibits 99.1 and 99.3.