[ACGL] Arch Capital: Q3 2026 Earnings Test for Reinsurance Margins

Arch Capital (ACGL) is a Bermuda-based specialty insurance group that writes commercial property and casualty insurance, reinsurance and mortgage insurance, and the Drillr earnings calendar shows it reporting Arch Capital Q3 2026 earnings on 2026-10-28, covering the Third quarter 2026 (three months ended September 30, 2026)[1]. The latest disclosed quarter is the second quarter of 2026: net premiums written were $4.049 billion, down about 6.9% from $4.348 billion a year earlier, the consolidated combined ratio was 83.5%, and the ratio excluding catastrophes and prior-year reserve development was 82.5%, up from 80.9% a year earlier[2]. After-tax operating income available to common shareholders was $893 million, or $2.56 per share, for an annualized operating ROE of 15.3%; the company bought back $1.2 billion of stock in the quarter and ended June with book value per share of $68.04[3]. Arch Capital does not issue quarterly or full-year earnings or premium guidance; on the July 29 call management gave only quarterly interest expense of $60 million to $63 million for the next two quarters and said the operating expense ratio would return to historical levels in the second half[4]. Drillr's analyst estimates put third-quarter operating EPS at $2.31 and revenue at $4.06 billion[5]. For comparison, operating EPS was $2.77 in the third quarter of 2025[6], and because the calendar's revenue line matches the company's net premiums written, the relevant comparison is $3.964 billion of net premiums written in the third quarter of 2025[2].

Three things matter most in this report, and each maps to one source of profit. The first is the reinsurance segment's underlying margin after the mid-year renewal price cuts: reinsurance is the largest profit engine, its combined ratio excluding catastrophes and reserve development rose to 79.9% in the second quarter, 270 basis points higher than a year earlier, and net premiums written fell 10.4%[7], so the third quarter will show whether rate cuts have moved from a mix effect to broad margin erosion. The second is whether the insurance segment's expense ratio comes down: it was 35.5% in the second quarter versus 33.6% a year earlier, which the company attributed to transition costs from the Allianz middle-market and entertainment business (MCE) and a smaller premium base[8], and the first quarter after the systems migration will test whether that is a one-time cost or a lasting scale problem. The third is whether mortgage profit holds: mortgage underwriting income fell from $260 million in the third quarter of 2025 to $220 million in the second quarter of 2026[9], while prior-year reserve releases fell from $57 million to $49 million over the same span[10], so the third quarter can separate a normalization of releases from early signs of weaker credit quality.

Company Background and Business Structure

Arch Capital is a Nasdaq-listed Bermuda exempted company in the S&P 500 with about $26.9 billion of capital at the end of 2025[11]. Through platforms in Bermuda, the United States, the United Kingdom, Europe, Canada and Australia, it writes specialty property and casualty insurance and reinsurance, as well as mortgage insurance and reinsurance, worldwide[11]. On August 1, 2024, it paid $450 million in cash for Allianz's U.S. MidCorp and Entertainment insurance businesses (MCE), the main source of the insurance segment's 13.4% net premium growth in 2025[12].

The company runs three underwriting segments, measures them on underwriting income and does not allocate investment income to them. In 2025, net premiums earned were $7.771 billion in insurance, $8.122 billion in reinsurance and $1.172 billion in mortgage, for a total of $17.065 billion; underwriting income was $375 million, $1.558 billion and $1.000 billion, respectively, for a total of $2.933 billion[13]. That means insurance supplied about 46% of earned premium but only about 13% of underwriting income, reinsurance supplied about 53% of underwriting income, and mortgage supplied about 34% of underwriting income from only about 7% of earned premium.

The insurance segment sells specialty commercial lines to mid-sized and large companies through independent retail and wholesale brokers; North American lines include other liability, property and short-tail specialty, commercial auto, commercial multi-peril and workers' compensation, and North America made up 74.2% of segment net premiums written in the second quarter, with international business at 25.8%[14]. The reinsurance segment provides specialty, property excluding catastrophe, property catastrophe, casualty, and marine and aviation cover to other insurers, which made up 30.5%, 24.5%, 21.3%, 16.9% and 3.1% of second-quarter net premiums written[15], and it cedes heavily to third-party capital. The mortgage segment insures U.S. lenders against borrower default, participates in Fannie Mae and Freddie Mac credit risk transfer transactions, and writes mortgage insurance in Australia and Europe[16]; at the end of June, U.S. primary insurance in force was $286.399 billion across roughly 1.04 million policies[17].

Financial History and Current Position

For full-year 2025, Arch Capital wrote $22.878 billion of gross premiums, $16.476 billion of net premiums and earned $17.065 billion of net premiums; on top of $2.933 billion of underwriting income, it reported $1.625 billion of pre-tax net investment income, $504 million of equity-method income and $464 million of realized gains[13]. Net income available to common shareholders was $4.359 billion, slightly above $4.272 billion in 2024; after-tax operating income was $3.700 billion versus $3.542 billion, and operating ROE fell from 18.9% to 17.1%[18].

On cash and capital returns, operating cash flow was about $6.2 billion in 2025, the pre-tax investment income yield at amortized cost was 4.11% versus 4.25% in 2024, and the fixed income portfolio had a duration of 3.34 years[19]. Book value per share reached $65.11 at the end of 2025, up 22.6% from $53.11 a year earlier; the company repurchased $1.9 billion of stock during the year, while its main capital return in 2024 had been a $1.9 billion special dividend[20].

In the first half of 2026, the business shrank in size but earned more because catastrophe losses fell. Net premiums written were $8.397 billion versus $8.863 billion a year earlier, and underwriting income rose to $1.385 billion from $1.235 billion, mainly because catastrophe losses dropped from $701 million to $375 million[2][10]. First-quarter net income was $1.037 billion, or $2.88 per share, with a combined ratio of 81.7%[21]; second-quarter net income was $1.047 billion, and first-half operating EPS was $5.06 versus $4.11 a year earlier, but second-quarter operating ROE was only 15.3%, below 18.5% and 18.9% in the third and fourth quarters of 2025[6].

On the balance sheet, the company used most of its first-half profit for buybacks. First-half operating cash flow was $2.510 billion and repurchases were $1.949 billion; second-quarter borrowing proceeds of $1.977 billion lifted debt and preferred stock to 18.1% of total capital from 13.2% at the end of March, and shares outstanding were 341.0 million at the end of June versus 375.4 million a year earlier, with $2.158 billion of repurchase authorization remaining[22]. The company said it repurchased about 94% of its first-half net income[23], and on April 19 the board raised the repurchase authorization by $3.0 billion[24].

Operating Model

Arch Capital's revenue is driven mainly by net premiums earned, which lag changes in pricing and volume by roughly four to five quarters. Consolidated revenue is the sum of net premiums earned, net investment income, equity-method income and realized gains, which in 2025 were $17.065 billion, $1.625 billion, $504 million and $464 million[13]. Net premiums written equal gross premiums less ceded premiums, and gross premiums depend on renewal shares, new business and rates; reinsurance renews mainly at 1/1, 4/1, 6/1 and 7/1, and property catastrophe premium is concentrated in the second quarter, with $392 million of reinsurance property catastrophe net premiums written in the second quarter of 2026 against only $64 million in the third quarter of 2025[15]. In a soft market the company deliberately writes less through non-renewals, smaller shares and more retrocession, which first lowers net premiums written and then lowers earned premium over the following quarters.

Operating income rests on the three segments' underwriting income plus investment income. Segment underwriting income is roughly net premiums earned times one minus the combined ratio, plus other underwriting income, and the combined ratio is the sum of the current-year non-catastrophe loss ratio, the catastrophe loss ratio, prior-year development, the acquisition ratio and the operating expense ratio. Calculated from the annual segment table, the 2025 combined ratio was about 96% in insurance, about 83% in reinsurance and about 17% in mortgage[13]. Quarterly swings come mostly from catastrophes and reserve releases, so underlying profitability is best read through the combined ratio excluding both, which was 82.5% on a consolidated basis in the second quarter of 2026[2].

Cash flow comes from collecting premiums before paying claims, and that float funds investment income. Operating cash flow roughly equals net income plus increases in loss reserves and unearned premiums, less increases in premiums receivable; it was about $6.2 billion in 2025[19] and $2.510 billion in the first half of 2026[22]. Pre-tax net investment income was $417 million in the second quarter at a 3.91% yield[25], so shrinking premiums and falling rates act on both underwriting income and investment income, with the investment side reacting more slowly.

Capital beyond what underwriting and regulators require is returned through buybacks, and regulatory capital is the constraint on that loop. The mortgage business is subject to the Private Mortgage Insurer Eligibility Requirements (PMIERs), and Arch MI U.S. had a sufficiency ratio of 165% at the end of June versus 179% at the end of 2025[17]. Management said on the call that buybacks remain the main capital allocation tool as long as organic growth opportunities are limited[4], which means capital released by shrinking underwriting returns to shareholders mainly through book value and earnings per share rather than premium growth.

Industry and Competitive Position

Arch Capital operates in a highly cyclical market with many competitors. In property and casualty insurance and reinsurance it competes with Chubb, AIG, W.R. Berkley, Markel, Everest, RenaissanceRe, Swiss Re, Munich Re and Hannover Re, among others; in U.S. mortgage insurance its main rivals are Essent, Enact, MGIC, NMI and Radian, and it also competes with government programs from the Federal Housing Administration (FHA) and the Department of Veterans Affairs[26]. The annual report describes the industry as a classic underwriting cycle in which a hard market attracts new capital and is followed by a soft market of falling rates and looser terms[11].

The market is in the early stage of a soft cycle, and property and short-tail lines face the most pressure. At the January 1, 2026 renewals, property catastrophe and short-tail excess of loss rates fell 10% to 20%[27]; mid-year renewals brought mid-teens rate cuts, although the rate index remains above pre-2020 hurricane levels[7]. On the second-quarter call management described an early soft market driven by new capacity, with competition fiercest in property and short-tail lines while casualty rates continue to rise[4]. The first-quarter call made the same point: the U.S. book broadly earned rate at trend, casualty above trend and short-tail property below trend[28].

The company's main distinction is its three-segment mix and its habit of adjusting volume through the cycle. Steady mortgage profit partly offsets property cycle swings, and the mortgage segment exceeded $1 billion of underwriting income for the fourth consecutive year in 2025[16], a cushion that a pure reinsurer lacks. In 2026 the company has cut property and short-tail business, bought more retrocession and shifted capital toward casualty and buybacks; the reinsurance expense ratio fell from 24.4% to 22.9% in the second quarter because of higher profit commissions on retrocessions[8]. The limit of the available material is that it lists competitors but gives no same-quarter combined ratios or growth rates for peers, so Arch Capital's relative performance cannot be quantified.

Core Debates

With property catastrophe rates down roughly 15% again at mid-year, can Arch Capital keep its reinsurance underlying combined ratio within 80%?

This question decides the direction of the company's largest profit pool. Reinsurance contributed $1.558 billion of underwriting income in 2025, about 53% of the three-segment total[13], and $851 million in the first half of 2026, about 61% of the $1.385 billion total[29][2]. On roughly $1.82 billion of quarterly reinsurance net premiums earned in the second quarter, each 1-point rise in the underlying combined ratio cuts quarterly underwriting income by about $18 million, and falling net premiums written keep pressing on earned premium in the following quarters.

Second-quarter data show the underlying margin deteriorating even as reported profit stayed strong. Reinsurance gross premiums written rose 0.2% year over year while net premiums written fell 10.4%, which the company attributed to non-renewals, share reductions and targeted increases in retrocession[8]; net premiums written fell from 64.4% to 57.6% of gross, and net premiums earned dropped from $2.087 billion to $1.820 billion, a decline of about 12.8%[29]. The combined ratio excluding catastrophes and reserve development was 79.9%, 270 basis points above a year earlier, which management attributed to line mix shifts and a more competitive pricing environment[7]. The reported combined ratio actually improved from 78.5% to 77.5%, because catastrophes added only 2.8 points and prior-year reserve releases totaled $95 million[29][10]. The baselines are 76.8% in the third quarter of 2025, 78.1% in the first quarter of 2026 and 79.9% in the second quarter for the underlying ratio; $1.737 billion of net premiums written in the third quarter of 2025 against $2.176 billion and $1.844 billion in the first two quarters of 2026; and $482 million of underwriting income in the third quarter of 2025, when catastrophe losses were only $29 million, against $410 million in the second quarter of 2026[29][10].

The financial transmission runs from new capacity to underwriting income. New capital pushes down property catastrophe rates, which fell 10% to 20% at 1/1[27] and roughly another 15% at mid-year; the company responds with non-renewals, smaller shares and more retrocession, so net premiums written and then net premiums earned decline. As low-loss-ratio property catastrophe business shrinks and rates fall across lines, the underlying combined ratio rises and underwriting income falls. What remains unresolved is how much each force contributes: one reading holds that the higher ratio mainly reflects a mix effect from less property catastrophe business, with casualty and specialty pricing still keeping pace with loss trends, while the other holds that margins are narrowing across lines at the same time. The company does not disclose rate changes by line, and quarterly data cannot fully separate mix from rate.

Three numbers matter in the third quarter. First, whether the reinsurance combined ratio excluding catastrophes and reserve development rises above 80.5%, and how management splits mix effects from rate effects; second, whether the net premium decline exceeds the second quarter's 10.4% and whether net premiums stay below 63% of gross; third, whether management describes 10/1 and January 1, 2027 renewal pricing as falling faster than at mid-year. If the ratio exceeds 80.5% and the premium decline widens, rate cuts are eroding earned premium faster than mix changes can offset; if the ratio falls back below 78% and non-catastrophe lines grow net premiums, the margin damage from rate cuts has been overstated.

With the Allianz middle-market business now migrated onto its systems, can Arch Capital's insurance segment bring its expense ratio down in the third quarter?

The insurance segment has the lowest margin of the three, and its expense ratio is the most direct repair lever. In 2025 it supplied about 46% of the company's net premiums earned but only about 13% of underwriting income[13]. On $1.880 billion of second-quarter net premiums earned, each 1-point drop in the expense ratio adds about $19 million of quarterly underwriting income, while second-quarter underwriting income was only $27 million because both expenses and catastrophes weighed on results[30].

The second-quarter rise in the expense ratio had both a one-time component and a shrinking-premium component. The segment's underwriting expense ratio was 35.5% versus 33.6% a year earlier, which the company attributed to MCE transition expenses and lower net premiums earned[8]; the first-quarter 10-Q had already said the data and systems migration was substantially complete, with some remaining transition expenses expected to extend into mid-year[31], and on the second-quarter call management said the operating expense ratio would return to historical levels in the second half[4]. Net premiums written fell 5.1% year over year, or 1.8% after adjusting for the non-renewal of certain MCE programs[8], and commercial multi-peril net premiums written fell from $205 million to $130 million[14]. The combined ratio excluding catastrophes was 91.6%, 100 basis points above 90.6% a year earlier[30]; management said there were no systematic changes to loss picks this quarter and that exiting lower-quality program business modestly reduced the segment loss ratio[7]. The baselines are an underwriting expense ratio of 34.4% in the third quarter of 2025 (a 19.6% acquisition ratio plus a 14.8% operating expense ratio), 36.3% in the first quarter of 2026 (20.0% plus 16.3%) and 35.5% in the second quarter (19.9% plus 15.6%); a combined ratio excluding catastrophes and reserve development of 91.3%, 92.7% and 91.6% over the same three quarters; and quarterly net premiums written of $1.953 billion in the third quarter of 2025 and $1.933 billion in the second quarter of 2026[30].

Whether the expense ratio falls depends on transition costs disappearing faster than the premium base shrinks. MCE joined in August 2024[12], and the systems migration was completed in the first half of 2026 at the cost of transition expenses; meanwhile some high-loss MCE program business was not renewed and short-tail property shrank on rate cuts, so net premiums earned fell about 4.5% year over year and fixed costs were spread over a smaller base[30]. If transition expenses disappear and exiting weaker business lowers the loss ratio, the expense ratio could return to about 34% to 34.5%, and in a quarter with normal catastrophes the segment could again earn more than $100 million of underwriting income, as it did with $129 million in the third quarter of 2025[30]. The other reading is that the higher expense ratio mainly reflects a smaller premium base, and with short-tail property rates still falling and MCE programs still exiting, the ratio will struggle to return to about 34% even without transition costs; on the first-quarter call management also said the benefits of the acquisition would show more in 2027[28].

In the third quarter, the test is whether the expense ratio falls below 35.5% and moves toward the 34.4% of the third quarter of 2025; whether the combined ratio excluding catastrophes stays at or below 92%, and how management describes casualty loss trends; and whether the net premium decline stays within 7% with casualty lines growing. If the expense ratio stays at 35.5% or higher while net premiums keep falling, the deleveraging of expenses is persistent rather than a one-time transition cost; if the expense ratio returns below 34.5% and the combined ratio excluding catastrophes is no higher than 91.5%, the drag from the MCE integration is largely over.

As reserve releases shrink quarter by quarter, can Arch Capital's mortgage insurance keep delivering about $220 million of underwriting income a quarter?

Mortgage insurance is the company's steadiest profit source in a property downcycle, but a meaningful share of it comes from prior-year reserve releases. The segment supplies only about 7% of net premiums earned, yet it delivered $1.0 billion, about 34%, of underwriting income in 2025[13]. Prior-year releases were $57 million in the third quarter of 2025 and $49 million in the second quarter of 2026[10], more than a fifth of that quarter's $220 million of mortgage underwriting income[9].

Second-quarter data suggest the profit decline came mainly from smaller releases, with no clear deterioration yet in current-year underwriting. Mortgage underwriting income fell from $260 million in the third quarter of 2025 to $221 million and $220 million in the first two quarters of 2026, and the combined ratio excluding prior-year development was 39.8% in the second quarter, 42.2% in the first quarter and 32.7% in the third quarter of 2025[9]. The company said the releases came mainly from better-than-expected cure rates, the second-quarter loss ratio excluding releases was relatively flat year over year, and the releases' impact on the loss ratio fell from 22.8 points to 15.7 points[32]. The U.S. primary delinquency rate was 2.07% at the end of June, persistency fell from 81.9% a year earlier to 79.9%, and refinancing reached 20.6% of new insurance written in the first quarter[17]; management said mortgage pricing competition is very muted and expects profitability to remain steady, and attributed second-quarter premium growth to a large new client in Australia and less reinsurance purchased[7]. The baselines are quarterly mortgage underwriting income of $260 million in the third quarter of 2025 and $221 million and $220 million in the first two quarters of 2026[9]; prior-year releases of $57 million in the third quarter of 2025 and $56 million and $49 million in the first two quarters of 2026[10]; and a U.S. primary delinquency rate of 2.04% on September 30, 2025, 2.17% at the end of 2025, 2.06% at the end of March 2026 and 2.07% at the end of June[17].

Whether profit holds depends on whether the current-year loss ratio stays flat as releases decline. High mortgage rates and home prices limit new originations, but borrowers hold ample home equity, so delinquent loans cure better than expected, releases continue, the mortgage combined ratio stays around 13% to 23%, and underwriting income runs at about $220 million to $260 million a quarter[9]. If releases fall while the current-year loss ratio stays the same, underwriting income falls by the same amount; if lower rates bring more refinancing, lower persistency also reduces premiums in force. The annual report noted that lower mortgage rates were beginning to support originations, though affordability still constrains the market[16]. The other reading is that the decline is mainly a normalization of releases, that delinquencies and the ratio excluding releases have not worsened, and that underwriting income will settle around $200 million rather than keep falling; reserve releases are changes in estimates and cannot be predicted from public data.

In the third quarter, the test is whether mortgage underwriting income falls below $200 million and whether prior-year releases fall below $35 million; whether the U.S. primary delinquency rate at the end of September exceeds 2.25%, and how new default notices compare with cures, which were 11,122 and 10,824 in the second quarter[17]; and whether persistency drops below 79% and insurance in force begins to decline. If releases fall below $35 million while delinquencies rise, the mortgage profit decline turns from release normalization into weaker underwriting quality; if releases stay above $40 million and the combined ratio excluding releases is no higher than 40%, mortgage profit can hold at about $220 million.

Risks and Falsifiers

The first risk is hurricane-season catastrophe losses against an unusually low prior-year base. Catastrophe losses in the third quarter of 2025 were only $72 million, or 1.7 points of the consolidated combined ratio[10][2]; based on in-force exposure as of July 1, 2026, a windstorm affecting the Florida Tri-County region carries a net probable maximum pre-tax loss of $1.8 billion, about 8.0% of tangible shareholders' equity, followed by Northeastern U.S. and Gulf of Mexico windstorms at $1.5 billion and $1.4 billion[33]. Catastrophe losses flow straight into quarterly insurance and reinsurance underwriting income, and $201 million of catastrophe losses in the second quarter already put the consolidated combined ratio about 1 point above the ratio excluding catastrophes and reserve development[3]. If third-quarter consolidated catastrophe losses stay within about twice the prior-year level, or under $150 million, and management does not revise its July 1 exposure, the risk has not materialized.

The second risk is a falling investment yield as interest rates decline. The pre-tax investment income yield at amortized cost fell from 4.25% in the second quarter of 2025 to 3.91% in the second quarter of 2026[25], while the company used about 94% of first-half net income for buybacks[23], and shrinking premiums also reduce investable cash inflows. Second-quarter pre-tax net investment income was $417 million, about 63% of that quarter's $657 million of underwriting income[25][2], so a lower yield directly reduces after-tax operating income. If the third-quarter pre-tax yield is no lower than 3.85% and net investment income is no lower than the $408 million of the third quarter of 2025, the current view of resilient investment income still holds.

The third risk is further rate cuts in property catastrophe and short-tail reinsurance. Management said rates would decline further if catastrophe activity stays low through the 2026 hurricane season, and pricing and mix had already raised the reinsurance underlying combined ratio by 270 basis points year over year in the second quarter[7]. On about $1.82 billion of quarterly net premiums earned, each 1-point rise in the underlying ratio cuts quarterly reinsurance underwriting income by about $18 million, and lower net premiums written depress earned premium for the following four to five quarters[29]. If the reinsurance ex-catastrophe accident-year combined ratio stays at or below 80% in both the third and fourth quarters and the net premium decline narrows to within 5%, this risk is falsified.

The fourth risk is that insurance premiums shrink faster than expenses adjust while casualty social inflation has not eased. Continued non-renewal of MCE programs and shrinking short-tail property could make net premiums earned fall faster than fixed costs, leaving the expense ratio above 2025 levels even after transition costs end[8]; management also acknowledged that the benefits of tort reform and more defense pushback have not yet appeared in reported loss data[7]. On about $1.88 billion of quarterly net premiums earned, each extra point of expense ratio or ex-catastrophe combined ratio cuts quarterly insurance underwriting income by about $19 million[30]. If the insurance expense ratio stays at or below 34.5% and the ex-catastrophe accident-year combined ratio at or below 91.5% in both the third and fourth quarters, this risk is falsified.

The fifth risk is shrinking mortgage reserve releases combined with refinancing that weakens persistency. Refinancing driven by lower rates reduces persistency and premiums in force, and if employment weakens, rising delinquencies would also lift current-year losses[17]. In the second quarter, the $49 million release was about 22% of the $220 million of mortgage underwriting income, and each $10 million drop in releases reduces underwriting income by the same amount[10][9]. If mortgage underwriting income stays at or above $210 million in both the third and fourth quarters and the delinquency rate at the end of September is no higher than 2.2%, this risk is falsified.

What to Watch Next

• Reinsurance margin: the combined ratio excluding catastrophes and reserve development was 76.8% in the third quarter of 2025 and 79.9% in the second quarter of 2026. A reading above 80.5% with a wider premium decline would deepen the erosion case; a reading below 78% would mean the erosion was overstated.
• Reinsurance volume: second-quarter net premiums written were $1.844 billion, down 10.4%, at 57.6% of gross. A decline larger than 10.4% or a net-to-gross ratio below 63% keeps pressure on; a decline within 5% would ease it.
• Reinsurance pricing: mid-year property rates fell about 15%. Commentary on 10/1 and January 1, 2027 renewals showing steeper cuts would extend pressure on future earned premium.
• Insurance expense ratio: 34.4% in the third quarter of 2025 and 35.5% in the second quarter of 2026. At or below 34.5% the integration drag is over; at 35.5% or higher it is a scale problem.
• Insurance underlying results: the second-quarter ex-catastrophe combined ratio was 91.6% on $1.933 billion of net premiums written, down 5.1%. A ratio at or below 92% with a premium decline within 7% keeps the base stable, and at or below 91.5% with casualty growth strengthens it.
• Mortgage profit: second-quarter underwriting income was $220 million with a $49 million release. Income below $200 million or releases below $35 million weaken the case; releases above $40 million support profit near $220 million.
• Mortgage credit: the U.S. primary delinquency rate was 2.07% and persistency 79.9% at the end of June. Delinquencies above 2.25% or persistency below 79%, combined with smaller releases, would signal weaker underwriting quality.
• Cross-debate risks: catastrophe losses were $72 million in the third quarter of 2025 and the second-quarter investment yield was 3.91%. Catastrophe losses within $150 million and a yield of at least 3.85% would keep the non-underwriting part of operating income stable.

Conclusion

Arch Capital's profit has three parts: reinsurance provides the most underwriting income, mortgage insurance provides steady high margins, and the insurance segment is the largest by premium but the least profitable, with roughly $400 million a quarter of investment income on top[25]. In the first half of 2026, helped by light catastrophe losses, the company earned $1.385 billion of underwriting income and $5.06 of after-tax operating EPS[2][6], while deliberately shrinking net premiums and returning about 94% of net income through buybacks[23]. The key unresolved relationship is how much of the rate decline and deliberate shrinkage can be offset by mix changes, retrocession and lower expenses: the reinsurance underlying ratio is rising, the insurance expense ratio is waiting to come down, and mortgage profit must hold above $200 million as releases shrink, and together these three will set operating income in the third quarter and beyond.

From the July 28, 2026 second-quarter release to the time of writing, the coverage found consisted only of results recaps, share-price reports, executive appointments and bond issuance, with no outside analysis containing independent argument, so this article cites no outside views and does not treat those reports as a shared market judgment. The two readings presented under each core debate come from the company's own disclosures and management's statements, not from a comparison of third-party opinions.

From here, if the reinsurance ex-catastrophe combined ratio holds within 80%, the insurance expense ratio returns below 34.5%, mortgage underwriting income stays at or above $210 million despite smaller releases, third-quarter catastrophe losses stay within $150 million and the investment yield holds at or above 3.85%, the view that the company can protect margins through the soft market would strengthen materially. Conversely, if the reinsurance ratio breaks above 80.5% with a wider premium decline, the insurance expense ratio stays above 35.5%, and mortgage releases fall below $35 million while delinquencies rise, the soft market would be building pressure on all three businesses at once rather than producing a one-time mix adjustment.

Sources

[1] ACGL earnings calendar updated 2026-10-24 · 2026-10-24 · Drillr earnings calendar

[2] ACGL 8-K filed 2026-07-28 · consolidated underwriting five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[3] ACGL 8-K filed 2026-07-28 · 2Q26 results · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[4] ACGL Q2 2026 earnings call 2026-07-29 · 2026-07-29 · earnings-call · https://gateway.drillr.ai/mcp/private

[5] Drillr analyst estimates ACGL Q3 2026 retrieved 2026-10-24 · 2026-10-24 · Drillr analyst estimates · https://gateway.drillr.ai/mcp/private

[6] ACGL 8-K filed 2026-07-28 · operating income five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[7] ACGL Q2 2026 earnings call 2026-07-29 · segment Q&A · 2026-07-29 · earnings-call · https://gateway.drillr.ai/mcp/private

[8] ACGL 8-K filed 2026-07-28 · insurance and reinsurance segment commentary · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[9] ACGL 8-K filed 2026-07-28 · mortgage segment five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[10] ACGL 8-K filed 2026-07-28 · prior year development and catastrophe losses by segment · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[11] ACGL 10-K filed 2026-02-26 · overview · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[12] ACGL 10-K filed 2026-02-26 · insurance segment outlook and MCE acquisition · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[13] ACGL 10-K filed 2026-02-26 · segment results 2025 · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[14] ACGL 8-K filed 2026-07-28 · insurance net premiums written by line five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[15] ACGL 8-K filed 2026-07-28 · reinsurance net premiums written by line five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[16] ACGL 10-K filed 2026-02-26 · mortgage operations and outlook · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[17] ACGL 8-K filed 2026-07-28 · U.S. mortgage insurance operating statistics five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[18] ACGL 10-K filed 2026-02-26 · operating income and returns 2025 · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[19] ACGL 10-K filed 2026-02-26 · investment income 2025 · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[20] ACGL 10-K filed 2026-02-26 · book value and capital return 2025 · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[21] ACGL 8-K filed 2026-04-28 · 1Q26 results · 2026-04-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[22] ACGL 8-K filed 2026-07-28 · capital, cash flow and share repurchases five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[23] ACGL 10-Q filed 2026-08-04 · current outlook · 2026-08-04 · 10-Q · https://www.sec.gov/Archives/edgar/data/947484/000094748426000124/

[24] ACGL 8-K filed 2026-04-20 · repurchase authorization · 2026-04-20 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[25] ACGL 8-K filed 2026-07-28 · investment income five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[26] ACGL 10-K filed 2026-02-26 · competition · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[27] ACGL 10-K filed 2026-02-26 · reinsurance segment outlook and 2025 loss ratio · 2026-02-26 · 10-K · https://www.sec.gov/Archives/edgar/data/947484/000094748426000017/

[28] ACGL Q1 2026 earnings call 2026-04-29 · 2026-04-29 · earnings-call · https://gateway.drillr.ai/mcp/private

[29] ACGL 8-K filed 2026-07-28 · reinsurance segment five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[30] ACGL 8-K filed 2026-07-28 · insurance segment five quarters · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[31] ACGL 10-Q filed 2026-05-05 · segment outlook · 2026-05-05 · 10-Q · https://www.sec.gov/Archives/edgar/data/947484/000094748426000058/

[32] ACGL 8-K filed 2026-07-28 · mortgage segment commentary · 2026-07-28 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000947484&type=8-K&dateb=&owner=include&count=40

[33] ACGL 10-Q filed 2026-08-04 · catastrophe and severe economic event exposure · 2026-08-04 · 10-Q · https://www.sec.gov/Archives/edgar/data/947484/000094748426000124/