# Hamilton Insurance: the underwriting is good, but the return depends on a hedge fund HG, NYSE. Published October 21, 2026. Price as of the October 20, 2026 regular session close. Values in US dollars. ## Verdict: HOLD at $36.51 At $36.51 the shares trade 10.5% below my Base value of $40.81, an upside of 11.8%. Hamilton has built a better underwriting business, with a combined ratio of 92.5% in the first half, but less than a third of its profit in my Base case comes from underwriting. The rest depends on a bond portfolio and a Two Sigma fund, and on reserves for casualty claims that will take years to settle. Confidence is C and the required margin of safety is 40%, which puts the maximum buy price at $24.48. Hold for holders, watchlist for new money. Third quarter results on October 29 are not in this analysis. - Bear value: $22.86 - Base value: $40.81 - Bull value: $58.52 - Upside to base value: +11.8% - Maximum buy price: $24.48 - Required margin of safety: 40% - Price to reported book value: 1.26 times - Combined ratio, first half 2026: 92.5% - Return implied by the price: 13.3% on equity - Modeled annual return, five years: 15.0% - Business quality: 67/100 - Valuation score: 49/100 - Confidence: 15/25, grade C Hamilton earned $86.7 million from underwriting in the first half of 2026 and reported $277.3 million of net income to common shareholders. The gap is mostly investment returns, and the largest single source is a fund run by Two Sigma, which returned $208.5 million in the same six months before the noncontrolling interest took its share. That makes Hamilton a different stock from the one its combined ratio suggests. At $36.51, the October 20, 2026 regular session close, the shares trade 10.5% below my triangulated Base value of **$40.81**, an upside of 11.8%. The Bear value is $22.86, 37.4% below the price, and the Bull value of $58.52 is 60.3% above it. Confidence is grade C (15/25), the required margin of safety (MOS) is 40%, and the maximum buy price is **$24.48**. The price is within reach of my Base and far above the buy ceiling, so the verdict is Hold. Valuation route: a Bermuda based specialty insurer and reinsurer, valued on common book value and the return that book can sustain, not on free cash flow, because claims reserves and regulatory capital decide what can be paid out. Primary method (65% weight): a price to book that follows a five-year fade of return on equity to a normal level. Secondary method (35%): 10 times normalized earnings built from underwriting and investment income. A tangible net asset test is a floor check and carries no weight. Classification: quality growth, 8 of 12 on the insurance scorecard. ## What Hamilton is Hamilton Insurance Group writes specialty insurance and reinsurance and invests the premiums it holds until claims are paid. The listed shares are Class B on the NYSE, but the company has three share classes (A, B and C) that share the economics, so market value built on Class B alone understates it. In 2025 net premiums earned were $2.11 billion, up 21.6%, the combined ratio was 92.9% and book value per share rose from $22.95 to $28.50. The investor presentation puts the premium mix at roughly half insurance and half reinsurance, with casualty close to half of gross premiums written. That matters more than the headline margin. Casualty claims are slow: a policy written today can produce its real losses in five to ten years, so a good combined ratio now is a hypothesis, and reserve development is how it gets tested. The same presentation shows combined ratios above 100% in 2021 and 2022 and profitable underwriting in 2023 to 2025. The second engine is the investment side. A bond and cash portfolio earns a steady yield. Alongside it, Hamilton allocates capital to the Two Sigma Hamilton Fund, which in the first half returned $208.5 million against $26.4 million from fixed income, short-term and cash holdings. Part of the fund is owned by outside investors and the manager earns an incentive allocation, so the common shareholder sees only what is left. ## The latest results | Latest reported period | Result | | --- | --- | | Second quarter 2026 net premiums earned | $586.0 million | | Second quarter combined ratio | 95.0%, with $49.9 million of catastrophe losses | | Second quarter net income to common | $143.8 million, $1.42 a diluted share | | Second quarter operating income | $158.2 million, $1.56 a diluted share | | First half net income to common | $277.3 million, $2.73 a diluted share | | First half operating income | $324.9 million, $3.20 a diluted share | | First half combined ratio | 92.5% | | First half annualized return on equity | 19.6% reported, 22.9% operating | | Book value per share, June 30 | $28.91, up 1.4% from $28.50 at year end | | Book value plus accumulated dividends | $30.91, up 8.5% for the half year | Operating income excludes certain items, so it is not the number I value. Reserve development was mixed. In the first half, attritional prior-year development was unfavorable by $13.1 million. In the second quarter, Bermuda was unfavorable by $13.0 million while International was favorable by $13.8 million. The two roughly cancelled, but a reserve charge in either segment is the line to read each quarter. The company reports third quarter results after the US close on October 29, with a call on October 30 at 9:30 am New York time (2:30 pm in Prague). None of it is in this analysis. ## Which return to value Free cash flow means little for an insurer: operating cash flow rises when premiums are collected before claims are paid, and the cash belongs to policyholders until then. I value common shareholders' book value and the return it can earn, and I take that return after the shares that belong to someone else. | $ million, first half 2026 | Amount | | --- | --- | | Consolidated net income | 471.8 | | Less income attributable to noncontrolling interests | -194.5 | | Net income to common shareholders | 277.3 | The $194.5 million is large, and it includes the fund manager's incentive allocation. A valuation that started from consolidated income, or from the fund's gross return, would overstate what common shareholders earn by more than half. Cash is also counted once: investments, float and the $2.00 special dividend already paid sit inside book value, and I add no cash on top. Unrestricted cash of $717.3 million is not surplus either, because it backs insurance obligations. ## Shares, book value and capital | Share count | Millions | | --- | --- | | Class A | 17.320 | | Class B | 65.891 | | Class C | 15.404 | | Total basic | 98.614 | | Allowance for dilutive awards (my estimate) | 2.038 | | Valuation share count | 100.652 | Common equity of $2,850.8 million divided by 98.614 million shares gives the reported $28.91, so the count reconciles. Dividing by my 100.652 million valuation count gives a diluted book value of **$28.32**, which is the base for every per-share figure below. Market value at $36.51 is $3.60 billion on basic shares and $3.67 billion on the valuation count. A data page that shows about $2.4 billion is counting Class B alone: 65.9 million shares at $36.51 is $2.4 billion. The balance sheet is clean on the surface. The term loan is $149.8 million, 5.25% of common equity, and tangible book value is $28.07 a basic share. Insurance leverage sits elsewhere: gross loss reserves of $4,783.1 million, or $3,319.2 million after $1,463.9 million of reinsurance recoverable. Capital coverage at the end of 2025 was 192% for the group and 206% for Hamilton Re on the transitional enhanced capital requirement. Those are year-end figures, not June or October. In May 2026 AM Best upgraded Hamilton Select, the US excess and surplus lines platform, to A (Excellent) with a stable outlook. The company spent $41.8 million on open-market buybacks in the first half, about $29.97 a share, and $55.3 million on repurchases in all. The $29.97 average was above the $28.91 book value at the time. That can still be a good use of money if the shares were worth more than $29.97, but it does not raise book value per share by itself. ## What the business can earn in a normal year The first half run rate is flattering: annualized operating earnings are about $6.40 a share and GAAP about $5.46. I do not capitalize either. Instead I build a normal year from the parts, with each assumption my estimate. | $ million unless stated | Bear | Base | Bull | | --- | --- | --- | --- | | Net premiums earned | 2,350 | 2,400 | 2,550 | | All-in combined ratio | 99.0% | 94.5% | 92.0% | | Underwriting profit | 23.5 | 132.0 | 204.0 | | Bond and cash income on 3,750 of capital | 150.0 (4.0%) | 168.8 (4.5%) | 187.5 (5.0%) | | Net Two Sigma contribution on 2,300 of capital | 115.0 (5%) | 207.0 (9%) | 299.0 (13%) | | Corporate cost, amortization, interest | -85.0 | -85.0 | -85.0 | | Tax at 10%, common net income | 183.2 | 380.5 | 545.0 | | Normalized earnings per share | $1.82 | $3.78 | $5.41 | A few choices decide this table. The Base combined ratio of 94.5% is worse than the 92.5% and 92.9% Hamilton reported, because it keeps a normal catastrophe load and some reserve drift. The fund return of 9% is after fees and the manager's incentive allocation and below the 13.2% annualized record in the investor presentation, so the fund is not subtracted a second time. The 10% tax rate avoids extending the unusually low interim rate forever. The two capital pools, $3.75 billion in bonds and $2.30 billion in the fund, are rounded allocations based on the presentation's split of about $3.8 billion and $2.4 billion. Each dollar earns one return only. Base earnings of $3.78 a share is a 13.3% return on the $28.32 book value. In my Base, underwriting supplies $132 million of $423 million pre-tax profit after costs, less than a third. The fund supplies $207 million, nearly half (my arithmetic). ## The primary valuation: price to book with a return fade The primary method follows the book value forward five years. Each year earnings equal opening book times the return on equity, a share of that is kept to grow book, and the rest is what could be paid out. After year five the business earns a normal return forever, and the exit price to book is (terminal return - growth) / (cost of equity - growth). In my Base the return fades from 18% to 15%, book grows 5% a year, the terminal return is 14%, terminal growth is 3% and the cost of equity is 11%, my assumption and not a universal hurdle. The exit multiple is 1.375 times book. | Base case year | Opening book | Return on equity | Earnings | Retained | Payout capacity | Closing book | | --- | --- | --- | --- | --- | --- | --- | | 1 | $28.32 | 18% | $5.10 | $1.42 | $3.68 | $29.74 | | 2 | $29.74 | 17% | $5.06 | $1.49 | $3.57 | $31.23 | | 3 | $31.23 | 16% | $5.00 | $1.56 | $3.43 | $32.79 | | 4 | $32.79 | 15% | $4.92 | $1.64 | $3.28 | $34.43 | | 5 | $34.43 | 15% | $5.16 | $1.72 | $3.44 | $36.15 | The exit price at the end of year five is $49.70, a future figure and not today's value. Discounted at 11%, the payouts are worth $12.93 and the exit $29.50, so the primary Base is **$42.43**. The payouts are what the capital could support, not a promised dividend. The exit carries 69.5% of the value, which is the main weakness of this method: most of what you pay for is the return Hamilton earns after the fifth year. A second calculation, discounting the earnings above the 11% cost of equity on each year's book, reproduces $42.43 to the cent. The Bear path (8% rising to 11% return, terminal return 11%, which gives exactly 1.0 times book) is worth $27.32. The Bull path (22% fading to 18%, terminal return 17%) is worth $57.95. ## The secondary valuation: 10 times normal earnings The second method takes the normalized earnings above and applies a multiple: $3.78 times 10 is **$37.80**. A multiple of 10 is a 10% earnings yield and fits a business whose earnings are sustainable but volatile. It is my assumption, not a peer median, and I use 8 times for the Bear ($14.56) and 11 times for the Bull ($59.56). This is not an independent check. It uses the same view of combined ratios, fund returns and tax as the primary method, only built from components instead of a return on book. When both agree it shows my own assumptions are consistent, not that they are right. For a check that needs no forecast, I use tangible common net assets: $2,850.8 million less $83.2 million of intangibles, divided by 100.652 million shares, is **$27.50** a share. Taking a 5% haircut to net claims reserves of $3,319.2 million, with no tax shield, removes $1.65 and leaves $25.85. That is a coverage reference, not a liquidation value and not a second valuation of the business. ## A price check against history Reported book value per share was $16.14 at the end of 2022, $18.58 in 2023, $22.95 in 2024 and $28.50 in 2025, then $28.91 in June 2026. From the end of 2020 ($16.23) to the end of 2025 that is 11.9% a year, but over 2023 to 2025 it was 23.9% a year. The recent pace reflects a very good stretch for both underwriting and the fund, and I do not treat it as the steady state. Hamilton listed in November 2023, so there is no long trading history, and I could not build a clean series of historical price to book to compare with. History gets no weight. The shares trade at 1.26 times reported book and 1.29 times my diluted book. My primary Base value is 1.50 times diluted book. The growth-adjusted check is conditional. A third-party forecast page shows 2025 operating EPS of $4.84 and an estimate of $5.09 for 2026, growth of 5.2%, a forward earnings multiple of 7.2 and a PEG of 1.39. I do not trust that figure: the estimate sample is a single analyst, the net income and EPS rows on the page do not reconcile to one accounting basis, and insurance earnings swing with investment marks. A broker page that shows a PEG of 0.11 and a return on equity of 31.7% has no period bridge, so I disregard both. Free cash flow growth is not meaningful for an insurer. The measure I use instead is book value plus dividends, which grew 8.5% in the first half of 2026 and 24.2% in 2025. ## Triangulating the values | Value per share today | Bear | Base | Bull | Weight | | --- | --- | --- | --- | --- | | Price to book with return fade | $27.32 | $42.43 | $57.95 | 65% | | Normalized earnings times a multiple | $14.56 | $37.80 | $59.56 | 35% | | Triangulated value | $22.86 | $40.81 | $58.52 | 100% | The Base methods differ by 11.5% (difference over midpoint), under the 20% warning level. That earns no confidence bonus, because the two share their drivers. I gave the book method 65% because for a property and casualty insurer common capital and reserves are the binding constraint, and the earnings method 35% because it tests whether the assumed returns can be produced by underwriting and investment income at all. The weights are my judgment and were not tuned to the price. Equal weights give $40.11 and the primary method alone gives $42.43. Neither clears the 40% margin of safety. The Bear methods differ by 61%, which is a warning in itself: the primary Bear assumes capital returns recover to the cost of equity, while the earnings Bear applies 8 times to a poor year. The weighted Bear of $22.86 is 16.9% below tangible net assets of $27.50. It is a bad outcome, not a limit on losses. ## What the price already assumes | Question at $36.51 | Answer | | --- | --- | | Constant return on equity that justifies the price (11% cost of equity, 3% growth) | 13.3% | | Terminal return needed if my five-year path holds (primary method alone) | 11.8% | | Terminal return needed in the weighted model | 11.5% | | Normalized earnings needed at 10 times | $3.65, or $367.5 million | | Fund return needed with Base underwriting and bonds | 8.4% | | Or an all-in combined ratio needed with a 9% fund return | about 95.1% | These are not demanding assumptions. The market is paying for decent underwriting and a net fund return in the mid to high single digits, and it does not need the 22.9% operating return Hamilton reported for the first half to persist. A multiple of 10 is reasonable for a business with volatile profits and slow claims. This is a decomposition of the price, not a prediction that the assumptions will hold. ## Bear, Base and Bull | | Bear | Base | Bull | | --- | --- | --- | --- | | Return on equity, years 1 to 5 | 8, 10, 11, 11, 11% | 18, 17, 16, 15, 15% | 22, 21, 20, 19, 18% | | Book growth, years 1 to 5 | 2% | 5% | 7% | | Terminal return / growth | 11% / 2% | 14% / 3% | 17% / 3.5% | | Cost of equity | 11% | 11% | 11% | | Normalized earnings / multiple | $1.82 / 8 times | $3.78 / 10 times | $5.41 / 11 times | | Value per share today | $22.86 | $40.81 | $58.52 | | Against $36.51 | -37.4% | +11.8% | +60.3% | | Five-year annual return from $36.51 | 3.9% | 15.0% | 23.5% | The Bear has weaker underwriting and fund returns and a recovery only to the cost of equity, with no extra reserve haircut on top. The Base assumes moderate margins and a fund return below the historical record. The Bull assumes strong risk selection and fund returns near the record. Share count, tax, costs and cost of equity are the same in all three. The 8 times multiple in the Bear stands for less durable earnings, not a second cut to the same year's profit. The five-year returns are modeled branch outcomes with the assumed payouts and exit, not probabilities, and no scenario carries one. ## Sensitivity and the case that flips the verdict | Triangulated value, Base | Cost of equity 10% | Cost of equity 11% | Cost of equity 12% | | --- | --- | --- | --- | | Terminal return 12% | $40.60 | $37.32 | $34.76 | | Terminal return 14% | $44.77 | $40.81 | $37.73 | | Terminal return 16% | $48.94 | $44.29 | $40.69 | One point on the all-in combined ratio changes normalized earnings by $0.21, the earnings method by $2.15 and the triangulated value by $0.75. One point on the net fund return moves the triangulated value by $0.72. The inputs that matter most are the terminal return (12% to 16%), the cost of equity (10% to 12%), the fund return (5% to 13%) and the combined ratio (92% to 99%). One input at a time: | Changed estimate | Value | Entry at 40% MOS | Label at $36.51 | | --- | --- | --- | --- | | Cost of equity 10% | $44.77 | $26.86 | Undervalued, margin unmet | | Cost of equity 12% | $37.73 | $22.64 | Hold | | Terminal return 12% | $37.32 | $22.39 | Hold | | Terminal return 16% | $44.29 | $26.58 | Undervalued, margin unmet | | Book value 5% lower | $39.43 | $23.66 | Hold | | Book value 5% higher | $42.19 | $25.31 | Undervalued, margin unmet | The label can flip between fairly valued and undervalued with one reasonable change, and the 5% book test covers the gap since June 30, because the book value after the third quarter is unknown. What does not flip is the action: no single change creates a Buy at this price with a 40% margin of safety, and the entry price ranges from $20.86 to $29.36 across the grid. Even the Bull at the same margin gives $35.11, below the price. ## How the buy price is set Confidence C calls for a margin of safety of 35% to 50%. I use 40%, toward the lower half because the capital is real and reported, and add nothing on top: the pricing cycle, reserves, fund concentration and terminal value are already handled in the scenarios, the confidence score and the sensitivity grid, and charging them again would count one risk twice. The maximum buy price is $40.81 x (1 - 40%) = **$24.48**. It is a decision reference, not an order to place. I would review accumulation between $20 and $25, ask for fresh evidence between $25 and $30, and look for what has broken below $20. Three numbers answer different questions: what the shares are worth ($40.81), the return from buying at $36.51 and following the primary Base path for five years (15.0% a year) and the price that gives a 10% annual return on that path ($44.12). The 10% is an illustration, not a requirement I know you have, and it does not replace value or margin of safety. A 15% modeled return next to a $24.48 buy price looks odd, and it is meant to: the return depends on payouts and a terminal return that could be wrong, while the margin of safety pays for that. ## The dividend Hamilton declared a special dividend of $2.00 a share, about $206 million, in February 2026 and paid it on March 30. It is already out of the business and inside the June book value, so I do not add it to value. Against $36.51 it is a historical 5.48%, not a yield: reliable recurring dividend guidance is not available, and I do not annualize a special payment. For an income investor the question is not the $2.00 but whether repeat earnings can fund a regular payout and still support underwriting growth. ## Reasons to own - **An underwriting record that is improving.** Combined ratios of 92.9% in 2025 and 92.5% in the first half follow losses in 2021 and 2022, and good underwriting builds capital that does not depend on the share price. - **More ways to grow.** A US platform with a stronger rating (Hamilton Select, upgraded to A in May) widens the market without a large acquisition. - **Light financing debt.** Debt is 5.25% of common equity, which removes one source of pressure, although claims leverage stays. - **A real record of book value growth.** Book value per share rose from $16.14 to $28.50 over 2022 to 2025, through a weak earlier period. - **A Bull case with real upside.** It is worth $58.52, and today's price does not need the best recent returns to repeat. ## Main risks - **Casualty reserves.** Claims can surface years after premiums are booked, and repeated adverse development would erase apparent profits and reduce book value. - **Losses and portfolio at the same time.** A catastrophe year and a weak market can arrive together, when liquid assets are also needed to pay claims. - **One fund manager.** Two Sigma's future net returns, fees and operations can differ from the record, and the gross return is not all the common shareholder's. - **Softer pricing.** Premium growth at inadequate rates destroys value later. - **Trapped capital.** Subsidiary solvency, ratings and collateral limit what moves up to the parent, and consolidated cash is not holding company surplus. - **Tax and expenses.** Recent tax credits and a lower expense ratio may not last. - **Dilution and governance.** Awards offset buybacks, and three share classes complicate the picture. - **Model risk.** The terminal return carries most of the primary value, and both methods share the same return drivers. ## Management execution Management's outcomes are visible without guidance to test them against: underwriting improved over three years, book value per share rose from $18.58 at the end of 2023 to $28.50 at the end of 2025, buybacks and the special dividend returned cash, and the Select upgrade supports the expansion. I score management credibility 14/20: execution 4/5, capital allocation 3/5, disclosure 4/5, alignment 3/5. The first half buyback averaged about $29.97, above the $28.91 book value, which is defensible if the shares were worth more, but I have not rebuilt every purchase against the book value on its date. Fund related compensation, voting arrangements and adjusted earnings definitions deserve attention. There is no numeric guidance I could score as delivered or missed, and I do not penalize the absence. ## Stock Analyza scorecard | Business quality | Points | Maximum | | --- | --- | --- | | Underwriting and distribution franchise | 13 | 20 | | Return versus cost of capital | 14 | 20 | | Balance sheet and resilience | 11 | 15 | | Earnings and reserve quality | 8 | 15 | | Growth and reinvestment runway | 11 | 15 | | Management and capital allocation | 7 | 10 | | Dilution and governance | 3 | 5 | | Total | 67 | 100 | On the insurance scorecard Hamilton scores 8/12, quality growth: 1/2 for the return spread (14% to 15% against an 11% cost of equity, not a lasting 5 points), 2/2 for book value growth, 0/2 for earnings quality given the negative years and investment swings, 1/1 for solvency and ratings, 1/1 for underwriting economics, 2/2 for repeat renewals and 1/2 for the runway. It does not qualify as a permanent compounder. The valuation score is 49/100: 20/40 for the discount to value, 4/20 for protection against the Bear, 11/15 for agreement between methods, 4/10 for history, 7/10 for what the price implies and 3/5 for data quality. ## Confidence score | Component | Points of 5 | | --- | --- | | Data quality and provenance | 4 | | Business predictability | 3 | | Valuation robustness | 3 | | Accounting transparency | 3 | | Scenario dispersion | 2 | | Total, grade C (13 to 17) | 15 of 25 | Clean arithmetic does not justify a higher grade. Current regulatory capital, the book value after June and the true size of reserve cushions are all unknown to me. ## What I would watch from here ### Supports the thesis - A four-quarter combined ratio at or below 95% with a sound current-year loss trend, and prior-year reserve development stable or favorable without depending on releases. - Normalized return on equity of 15% or more, and book value plus distributions growing more than 10% a year. - Group capital coverage above 180% on the same definition, an A rating with a stable outlook, and a multi-year net fund return of 8% to 12%. ### Watch closely - A combined ratio of 95% to 100%, or favorable releases hiding weak current-year underwriting, and adverse prior-year development of up to 2% of annual premiums. - Return on equity of 11% to 15%, capital coverage of 150% to 180%, or a negative rating outlook. - Awards that offset buybacks, a larger fund allocation without better disclosure, or fund returns of 0% to 8%. ### Breaks the thesis - A combined ratio above 100% persistently, or adverse development above 2% of annual premiums more than once. - Normalized return on equity below 11% for two annual periods, or capital coverage below 150%, a capital raise or limits on distributions. - A rating downgrade, or material fund losses together with capital stress. ## The most important thesis breakers - **Casualty reserves.** Repeated adverse revisions that push normalized return on equity to about 11% or below. - **The fund.** A net return well under the 9% I assume, at the same time as a weaker combined ratio. - **Capital and ratings.** A fall in capital coverage or a downgrade that limits access to business. ## Stress test and independent check I rebuilt the model from its stated inputs with separate code and reproduced: the five-year Base table and the $42.43 primary value, including a second calculation through excess returns over the cost of equity that matches to the cent; the Bear and Bull values of $27.32 and $57.95; the normalized earnings of $380.5 million and $3.78 a share for the Base; the $37.80 secondary value; the $40.81 blend and the 11.5% divergence; the equal weight value of $40.11; the 15.0% five-year return and the $44.12 entry for a 10% return; the five reverse results; the sensitivity grid and the single-input table; the $24.48 buy price; the $27.50 tangible net assets and the $25.85 stress; and the combined ratio and fund return sensitivities. I found no construction error. From the company's documents I confirmed the second quarter and first half results, the combined ratios, the $28.91 book value and $30.91 with dividends, the balance sheet figures (equity, liabilities, noncontrolling interest, term loan, cash, intangibles, gross and recoverable reserves), the $471.8 million, $194.5 million and $277.3 million bridge, the $55.3 million of repurchases, the $2.00 special dividend and its dates, the 2025 results and book values of $28.50 and $22.95, the 192% and 206% capital coverage (a transitional enhanced capital requirement at the end of 2025), the AM Best upgrade of Hamilton Select, and the October 29 reporting date. The $36.51 price is the October 20 close on a quote page that showed a $0.47 gain; the October 19 close was $36.04. A broker page showed $36.36 at a different moment and the verdict is the same at either price. I could not re-read some inputs, so they come from the supplied analysis: the June 30 counts for Class A and Class C shares, the first half buyback share count behind the $29.97 average, the diluted weighted share counts, the Bermuda segment's first half reserve charge, the investor presentation items (premium mix, combined ratios for 2021 to 2023, bond duration, the 13.2% fund record and the $3.8 billion and $2.4 billion split), the November 2023 listing date, and book values for 2020, 2022 and 2023. The Class A and Class C counts reconcile to the reported $28.91 book value, which supports them. I then ran three tests of my own, the first three rows below. The primary method assumes a terminal return of 14%, above the 13.3% my own bottom-up earnings imply, so I lowered it, and I tested a higher tax rate. The last three rows are robustness tests from the original model. | Test | Triangulated value | Entry at 40% MOS | Effect | | --- | --- | --- | --- | | Terminal return 13.3%, matching the bottom-up earnings | $39.59 | $23.75 | Hold, margin unmet | | Return 13.3% in every year, no fade | $37.36 | $22.42 | Hold, margin unmet | | Tax 15% instead of 10% in the earnings method | $40.07 | $24.04 | Hold, margin unmet | | Consensus 2026 EPS of $5.09 at 10 times in the earnings method | $45.39 | $27.23 | Margin unmet; the estimate basis is unverified | | No new dilution (about 2% fewer shares) | $41.65 | $24.99 | No change to the action | | Bull operations at the same 40% margin | $58.52 | $35.11 | Still below the price | The strongest case against Hold is that recent returns and the fund's record support much more value, and the Bull is 60.3% above the price. My answer is that both need capital to be extracted and returns to persist that the reserve and investment evidence does not make predictable. A smaller margin of safety would lead to a different decision, which is a different risk preference and not a factual error. The strongest case against owning at all is the Bear at $22.86, which combines weaker reserves with weaker fund earnings. Reported capital and low financing debt cushion that, but neither is a floor under the share price. The main data limits: I do not know book value, share count or regulatory capital after June 30, I could not check the fund's return after fees, and I do not know how much of the fund's gain is reversible. These are why Confidence is C. ## Final verdict: hold at $36.51 Hamilton is a better underwriter than it was three years ago, and it carries more capital than a screen based on price to earnings would suggest. But the cheapness is overstated by investment gains and by broker ratios that do not line up with the filings, the price sits close to a normalized Base, and a moderate error in reserves or in the return the fund can keep would absorb the discount. A new purchase is not supported by this model at a price above $24.48. **Verdict: HOLD at $36.51, watchlist for new money. Triangulated Base value $40.81, upside 11.8%, Confidence C (15/25), required margin of safety 40%, maximum buy price $24.48.** For holders this means keeping the position and reading the October 29 report for book value, reserves and the fund's return, because the gap to Base is small. For new money it means waiting for a lower price or for another year of clean reserves. Underwriting is the part of Hamilton you can audit. The fund is the part that moves the answer. ## More analyses on Patreon This analysis is free. Premium members get every in-depth analysis, the full archive and a vote on which stocks get analyzed next: https://www.patreon.com/stockanalyza ## Sources Primary sources include Hamilton's [second quarter 2026 quarterly report on Form 10-Q](https://www.sec.gov/Archives/edgar/data/1593275/000159327526000098/hg-20260630.htm), the [second quarter 2026 results](https://investors.hamiltongroup.com/news/news-details/2026/Hamilton-Reports-2026-Second-Quarter-Results/default.aspx) and its [SEC exhibit](https://www.sec.gov/Archives/edgar/data/1593275/000159327526000094/hg991pressreleaseq226.htm), the [second quarter investor presentation](https://www.sec.gov/Archives/edgar/data/1593275/000159327526000094/hg_993invdeckq226.htm), the [2025 fourth quarter and full year release](https://www.sec.gov/Archives/edgar/data/1593275/000159327526000007/hg991pressreleaseq425.htm) and [2025 annual report](https://s203.q4cdn.com/992304269/files/doc_financials/2025/ar/ARS-HG-March-24-2026-final.pdf), the [2025 Financial Condition Report](https://s203.q4cdn.com/992304269/files/doc_downloads/2026/05/Hamilton-FCR-2025.pdf), the [AM Best upgrade of Hamilton Select](https://www.hamiltongroup.com/press-release/am-best-upgrades-financial-strength-rating-to-a-excellent-for-hamilton-select/), the [third quarter reporting date](https://investors.hamiltongroup.com/news/news-details/2026/Hamilton-to-Report-Third-Quarter-2026-Financial-Results-on-October-29-2026/default.aspx) and the [2023 annual report](https://investors.hamiltongroup.com/files/doc_financials/2023/ar/Hamilton_2023AR_Digital_VF.pdf) for earlier book values. The earnings estimate used in one robustness test comes from a third-party [forecast page](https://stockanalysis.com/stocks/hg/forecast/), was not used as an input to the Base, and its basis could not be verified. The reference market price is $36.51, the October 20, 2026 regular session close, shown on the [StockTitan quote page](https://www.stocktitan.net/overview/HG/), with the October 19 close of $36.04 on the [stockanalysis.com quote page](https://stockanalysis.com/stocks/hg/). Per share values use 100.652 million economic shares. The normalized combined ratio, the investment pools and returns, the tax rate, the cost of equity, the return path, terminal assumptions and multiples are analytical estimates, not company guidance. **Disclaimer.** This analysis is provided solely for research and educational purposes. It is not personalized investment, financial, legal or tax advice. Values are estimates, not forecasts or trading instructions. Hamilton is exposed to reserve development on long-tail casualty business, catastrophe losses, dependence on a single investment manager for a large share of its returns, softening insurance prices, capital held in regulated subsidiaries, and dilution. Investors can lose part or all of their invested capital. Method: Master Investment Analysis Framework v3.4. Approach: price to book with a five-year fade of return on equity, cross-checked with normalized earnings and a tangible net asset test. Confidence: C/15. Data checks: passed with warnings. Conditional research only. Open data gaps: book value and share count after June 30, current regulatory capital and the fund's return after fees.