Stewart Information Services: $54 of Book Value per Share — $1.58 of It You Can Touch

There is an investor trap that catches careful people in particular — because it feels like prudence: the book value illusion. It goes like this. You are not chasing the hot growth story; you want something solid. You find a company trading at roughly 1.26 times book value and you think: “For every dollar I pay, there is a dollar of substance sitting on the balance sheet. How wrong can this go?” The instinct is honorable. It just assumes one thing nobody has checked — that the book value consists of things that actually exist. Stewart Information Services Corporation (NYSE: STC) of Houston, Texas, has been insuring ownership rights to real estate since 1893 and is today the fourth-largest U.S. title insurer. Reported book value stood at $54.63 per share on June 30, 2026. Tangible book value — after deducting goodwill and intangibles — works out to roughly $1.58. So let us make a deal: before you write “cheaply valued” in your notebook, we read together what Stewart itself reported to the U.S. securities regulator, the SEC — the annual report on Form 10-K for 2025, the quarterly report on Form 10-Q for the period ended June 30, 2026, and the 2022 annual report that documents the boom years. An SEC filing is honest under penalty of law. And this one tells of revenue that has almost returned, earnings per share that have not, and $332.7 million that landed on the balance sheet almost entirely as goodwill. In the end, the decision is yours.

What Stewart Actually Does — the Insurance You Buy Only Once

Title insurance is an American product with no direct European equivalent, so it is worth an everyday picture. Imagine buying a house. In much of Europe, a public land register carries a state guarantee: what it says, counts. The United States has no such register. Instead, ownership records sit scattered across county archives — old deeds, probate records, tax liens, mechanic\'s liens, divorce decrees. Somebody has to search that paper mountain and guarantee that the seller really owns the property and that no one else has a claim. That is exactly what Stewart does: search, examine, insure. And then the unusual part — the premium is paid once, at closing, and covers losses arising from events that occurred before the purchase, for as long as the owner holds the property. No annual premium, no renewal. A business that earns money only when real estate changes hands or mortgages are written.

Stewart sells this through three channels, and the differences between them are the key to everything that follows. First, direct operations: company-owned offices, company employees, and all of the revenue stays in house — $1,157.5 million in 2025. Second, agency operations: independent title agencies sell policies under Stewart\'s underwriting, Stewart carries the risk and keeps a share of the premium — $1,262.6 million of gross revenue in 2025, of which roughly 83 percent flows straight back to the agencies. Third, real estate solutions: credit reporting, valuations, online notarization, property preservation for mortgage servicers — $438.3 million in 2025, and the only area Stewart actively buys into. As of December 31, 2025 the group employed roughly 7,800 people, about 6,000 of them in the United States and 1,800 at locations outside it, mainly in Canada.

That frames the central tension of this analysis, and it runs through every chapter that follows: Stewart is growing strongly again — but the growth comes from the two channels where the least of it sticks, and a substantial part of it was bought rather than earned. What looks like substance on the balance sheet is 97 percent goodwill and intangibles.

How the Stock Landed on Our Desk

Honesty first: Stewart did not reach our research list through a hit in our in-house stock scanner, but through the ongoing review of the U.S. equity universe — the company was not covered in our database at all before August 21, 2026. What stood out was the combination of metrics: a price-to-earnings ratio near 15, price-to-book around 1.26, and a dividend yield just above 3 percent (data as of August 20, 2026) — at a company whose revenue has grown by almost a quarter over the trailing twelve months. Combinations like that are either an opportunity or a warning, and which of the two never shows up in a screener. It shows up in the notes.

A useful reference point is the market leader, which we have already taken apart: Fidelity National Financial, where the vault is full but the money belongs to others. And because Stewart\'s business hangs on the mortgage cycle, it pays to look at the financing side of the same market, which we described in Rocket Companies and the $19.4 billion nobody ever paid for. Remember the finding now: with title insurers, you do not measure by the price-to-book ratio, but by what the book value is made of.

The Numbers Over the Years — Honestly Credited

First, what genuinely speaks for Stewart — and that is more than this article\'s headline suggests. The company survived a brutal cycle without ever posting a loss. When the Federal Reserve raised rates through 2022 and 2023 and the mortgage market collapsed, Stewart\'s pretax income did not halve — it fell by roughly an order of magnitude. Even so, every single year ended in the black. And the recovery is real: $2,921.6 million of revenue in 2025, after $2,490.4 million in 2024 and $2,257.3 million in 2023, an increase of 18 percent in 2025. The first half of 2026 continued the trend: $1,680.5 million against $1,334.2 million a year earlier, up 26 percent. Net income attributable to Stewart rose from $35.0 million to $54.2 million in the same period. The claims side is behaving too: the title loss ratio fell to 3.2 percent of title revenue in the second quarter of 2026, from 3.6 percent in the prior-year quarter. And the balance sheet carries: $261.6 million of cash, $646.7 million of debt against $1.67 billion of equity, and a debt-to-capitalization ratio of roughly 28 percent (all as of June 30, 2026). The dividend of $2.10 per share cost $58.5 million in 2025, against $132.3 million of free cash flow — coverage of 2.3 times.

Now the chart that puts all of it in perspective — diluted earnings per share over five years:

That gap is the real story. In 2021 Stewart generated $3,305.8 million of revenue and earned $11.90 per share. In 2025 revenue was $2,921.6 million — 88 percent — and earnings per share were $4.05, or 34 percent. Where did the rest go? The second chart shows most of the answer: revenue is nearly back, but it comes out of different pockets than it used to.

Put in everyday terms: picture a baker who used to sell most of his bread in his own shop. Today he books the same revenue, but a large share runs through market stalls he does not own and has to hand 83 cents of every dollar to, plus a café he bought at a steep price. Revenue on paper is unchanged. What sits in his till at month\'s end is not. Remember this: at Stewart, revenue growth and profit growth are not the same thing — and for a structural reason, not a one-time one. Which brings us to the uncomfortable truths.

What the Filings Say — the Uncomfortable Truths

Uncomfortable truth No. 1: of $54.63 of book value per share, roughly $1.58 is tangible

The quarterly report for the period ended June 30, 2026 states the book value itself — and it reads reassuringly:

“Our book value per share was $54.63 and $54.30 as of June 30, 2026 and December 31, 2025, respectively. As of June 30, 2026, our book value per share was based on approximately $1.66 billion of stockholders’ equity attributable to Stewart and 30,449,652 shares of Common Stock outstanding.”

— Stewart Information Services Corporation, Form 10-Q for the quarter ended June 30, 2026, Item 2 MD&A

Now the counter-calculation, every figure from the same balance sheet. Of the $1,663.4 million of equity attributable to Stewart, $1,293.8 million is goodwill and a further $321.6 million is other intangible assets — $1,615.4 million together, or roughly 97 percent. What remains is $48.0 million, or $1.58 per share. What is goodwill? In everyday terms: if you buy a company for $100 million whose machines, buildings and receivables are worth $20 million combined, the remaining $80 million is the price of things you cannot touch — customer relationships, a working team, a name. That amount sits on your balance sheet as goodwill and stays there as long as you can credibly argue it is still worth something. If the acquired business earns less than hoped, it must be written down — and equity disappears without a single dollar leaving the building.

How much cushion sits behind that judgement, Stewart does not say — the company discloses no margin between the calculated fair value and the carrying value of the units it tests. Auditor KPMG explicitly designates the goodwill assessment a critical audit matter, meaning one of the most difficult and judgement-heavy parts of the entire audit:

“Minor changes to those assumptions could have had a significant effect on the Company’s assessment of the carrying value of the goodwill.”

— KPMG LLP, report of independent registered public accounting firm, Form 10-K for 2025

In fairness: Stewart has not had to write down goodwill in either 2024 or 2025, and for a services business without factories a high share of intangibles is normal. A title insurer needs no machines; it needs archives, software and people. The scale remains notable all the same — and it explains why a price-to-book ratio of roughly 1.26 tells a different story than it appears to at first glance.

Uncomfortable truth No. 2: the core business grew 15 percent — and earned less than a year earlier

This number needs reading twice. In the second quarter of 2026, operating revenue in the title segment — the actual insurance business — rose 15 percent to $683.6 million. Pretax income in that same segment fell 1 percent to $48.6 million. The segment margin dropped from 8.1 to 6.9 percent. Stewart lays it out in its own quarterly report:

The reason appears a few lines further down in the same filing, and it is purely a question of mix. Direct operations — the channel where revenue stays in house — grew a modest 5 percent to $306.6 million. Agency operations, by contrast, jumped 25 percent gross, to $377.0 million. That sounds excellent until you read the next line: $314.9 million of it went to the agencies in the same quarter. Net, Stewart kept roughly $62 million rather than $377 million. And because the company carries the same examination, administrative and claims costs either way, expense grows with gross revenue while only a sixth of it arrives as Stewart\'s own income. Title segment employee and other operating costs rose $29.6 million, or 11 percent, in the quarter.

Put in everyday terms: you run a taxi company and celebrate that revenue is up a quarter. Then you notice almost all of the increase comes from rides you dispatch to outside drivers — and they keep 83 cents on the dollar. Your dispatch office, your insurance and your bookkeeper still work for the full revenue figure. At year-end you have worked more and earned less. One bright spot: in the real estate solutions segment the movement went the other way — pretax margin rose from 6.0 to 9.4 percent on 75 percent revenue growth. The acquisition is already contributing profit. Whether that holds, nobody knows after two quarters.

Uncomfortable truth No. 3: net income rose 17 percent, earnings per share only 7

In the fourth quarter of 2025, Stewart raised fresh equity in the market — and at a scale worth knowing before extrapolating per-share growth:

“During the fourth quarter 2025, we issued an aggregate of 2,185,000 new shares of Common Stock, which included shares purchased by the underwriters to the offering transaction. Total proceeds from the offering, net of issuance costs, was $140.8 million.”

— Stewart Information Services Corporation, Form 10-K for 2025, Item 7 MD&A

What that means for you as a shareholder shows up in a two-quarter comparison. In the second quarter of 2026, net income attributable to Stewart rose from $31.9 million to $37.2 million — up 17 percent. Diluted earnings per share rose from $1.13 to $1.21 — up 7 percent. The difference is dilution: the average diluted share count grew from 28.330 million to 30.853 million over the same period, or 8.9 percent. In everyday terms: the pizza is a sixth larger, but it is now split among nine people instead of eight. Your slice did grow — just by less than the headline promises.

Fairness demands the other side here too: the money was not raised to plug holes, it was raised to buy. Stewart spent $370.0 million on acquisitions in 2025, $332.7 million of it on service provider Mortgage Contracting Services in December. Whether that pays off will be settled in segment margins over the coming years — and the first two quarters, as described above, look good there. But: growth paid for with new shares only becomes growth for the existing shareholder once it outpaces the share count.

Uncomfortable truth No. 4: Stewart names artificial intelligence in its own annual report as a threat — not an opportunity

We searched the 2025 annual report systematically for the terms “artificial intelligence,” “AI” and “machine learning.” Result: six occurrences — and all six sit in the risk factors. Not one in the business description, none in management\'s discussion, none in the notes. The clearest reads:

“Further, advances in technologies, including technology such as AI and machine learning, could, over time, significantly disrupt the traditional business model of financial services and real estate-related companies, including title insurance.”

— Stewart Information Services Corporation, Form 10-K for 2025, Item 1A Risk Factors

Why does this matter? Because Stewart\'s core business is, at bottom, document search and risk assessment — precisely the kind of work language models are making cheaper fastest. Where Stewart describes its own efficiency plans, the filing deliberately chooses the word automation rather than AI: margin targets are to be met through “additional automation of manual processes” and the consolidation of systems. That is not a weakness in itself — but it means the technology most likely to reshape the business is, so far, described as somebody else\'s risk rather than the company\'s own lever. One more sentence from the same paragraph belongs here: Fidelity National Financial, First American and Old Republic each have, by Stewart\'s own account, “substantially greater gross revenues.” Stewart is clearly the number four among the big four, and number four spreads technology spending across fewer policies.

Valuation: What the Market Pays for a Cyclical With a Goodwill Balance Sheet

Let us talk orders of magnitude, not intraday prices. At $68.96 on August 20, 2026 and roughly 30.52 million shares outstanding (as of July 28, 2026, per the cover page of the quarterly report), the market capitalization stood at about $2.10 billion. That yields the following magnitudes (data as of August 20/21, 2026): a price-to-earnings ratio of roughly 15 on trailing twelve-month earnings, a price-to-sales ratio of roughly 0.6, and a price-to-book ratio of roughly 1.26. Measured against tangible book value of about $1.58 per share, the market pays roughly 44 times — which shows that this metric is no sensible yardstick for a services business without factories, but is an honest reminder that the low price-to-book ratio does not describe a reserve of substance.

The price-to-earnings ratio of 15 is the more interesting number — and it cuts both ways. For a cyclical in recovery it can be cheap: if Stewart returned to something near its 2021 results in a normalized market, the stock would look very inexpensive today. It can also be expensive: in cyclical industries, low price-to-earnings ratios typically appear at the peak of the cycle and high ones at the trough. Stewart is plainly not at a peak — 2025 delivered a third of the 2021 earnings per share — but it is no longer at the 2023 trough either. The only analyst consensus we can find as of August 20, 2026 is thin: a single rating of hold with a price target of $82.25. That is too little data for a conclusion, and we treat it accordingly — as a footnote, not an argument.

What can be assessed cleanly is the dividend. Stewart has paid $0.525 per share per quarter since the third quarter of 2025, or $2.10 a year — roughly 3.0 percent at the price quoted above. In 2025 that cost $58.5 million against $132.3 million of free cash flow ($205.7 million of operating cash flow less $73.4 million of capital expenditure). That is 2.3 times coverage, and considerably more comfortable than 2023, when $45.2 million of free cash flow faced a $50.5 million dividend — in the trough year, the payout was not earned. One caveat matters for judging this: half-year figures make coverage look worse than it is, because the business runs seasonally stronger in the second half. The full-year comparison is the meaningful one.

Opportunities and Risks at a Glance

What speaks for Stewart:

• A genuine rate lever to the upside. The average 30-year fixed mortgage rate in the United States stood at 6.4 percent in the second quarter of 2026, down from 6.8 percent a year earlier — that alone drove refinancing activity up 45 percent. If rates fall further, transaction volume rises, and Stewart\'s cost base is already in place.
• Commercial is running strongly. Domestic commercial revenue rose 20 percent in the second quarter of 2026 and 28 percent in the first half — driven in part by large data center transactions. The average commercial fee per file was $16,900, against $3,200 on the residential side.
• The claims side is quiet. The title loss ratio fell to 3.2 percent of title revenue in the second quarter of 2026 (3.6 percent a year earlier); the reserve for title losses stood at $519.2 million.
• Cheap, long-dated debt. The $450 million note carries a 3.6 percent coupon and matures on November 15, 2031 — a quiet advantage in a materially higher rate environment. At the end of 2025 its fair value was $397.6 million, below par.
• A covered dividend and a balance sheet that carries. Coverage of 2.3 times in 2025, a debt-to-capitalization ratio near 28 percent, $261.6 million of cash and $97.5 million of undrawn credit line capacity (June 30, 2026).

What speaks against it:

• 97 percent of equity is goodwill and intangibles. $1,293.8 million plus $321.6 million against $1,663.4 million of equity (June 30, 2026). KPMG designates the goodwill assessment a critical audit matter; Stewart discloses no cushion between fair value and carrying value.
• Growth without profit growth in the core. Title segment, second quarter of 2026: revenue up 15 percent, pretax income down 1 percent, margin from 8.1 to 6.9 percent.
• Dilution. 2,185,000 new shares in the fourth quarter of 2025; the average diluted share count rose 8.9 percent within a year, to 30.853 million.
• Liability for outside agencies. Roughly 44 percent of operating revenue comes from agency operations. Stewart itself writes that case law in certain states holds the company liable for the acts or omissions of its agents — “regardless of contractual limitations.”
• The smallest of the big four. Fidelity National Financial, First American and Old Republic each have, by Stewart\'s own account, substantially greater gross revenues, and their holding companies significantly greater capital.
• A cycle that has already halved everything once. Going from $11.90 of earnings per share in 2021 to $1.11 in 2023 took two years. This business has no subscription character to smooth revenue.

A Human Bottom Line

We began with the book value illusion — the reassuring thought that a price-to-book ratio near one means you are getting something tangible for your money. At Stewart Information Services the metric is accurate and the conclusion still is not. Of $54.63 of book value per share, roughly $1.58 is tangible; the rest is the price Stewart has paid for other companies over decades. That is not an accusation — for a services business it is the norm, and Stewart has never had to write the goodwill down. It simply means: this particular metric carries no safety net here. Anyone buying Stewart is not buying an asset play, but a wager on the American real estate cycle — with a solid, but smallest, of four large underwriters.

And the honest ledger of this analysis has two sides. One: a company that came through a collapse from $11.90 to $1.11 in earnings per share without a single loss-making year, earns its dividend 2.3 times over, has financed itself at 3.6 percent until 2031, and whose commercial business is currently doing very well on data center deals, is no turnaround case. The other: revenue is back to 88 percent, earnings per share to 34 percent — and the three reasons behind that (shifted mix, higher interest expense, more shares) do not disappear on their own when the market recovers. Two of them are structural. What you make of that is your call. And that is exactly as it should be.

Sources

• Form 10-Q for the quarter ended June 30, 2026, filed August 4, 2026 — balance sheet, income statement, segment reporting, goodwill note, order table, book value per share
• Form 10-K for 2025, filed February 27, 2026 — business description, risk factors, KPMG audit report, equity offering, MCS acquisition, indebtedness, headcount
• Form 10-K for 2022, filed February 28, 2023 — revenue and earnings figures for fiscal 2021 and 2022
• Form 8-K (Item 2.02) dated July 22, 2026 — second quarter 2026 results
• Form 8-K (Item 7.01) dated June 1, 2026 — declaration of the $0.525 per share dividend for the second quarter of 2026
• SEC EDGAR filing index, CIK 0000094344 — verification that no further filings were submitted after August 4, 2026 (checked August 21, 2026)
• Fundamental data (price, market capitalization, valuation multiples, 52-week range, analyst rating; data as of August 20/21, 2026)

Disclaimer: This article is journalistic research and analysis. It is not investment advice, not a recommendation to buy or sell, and not a solicitation to buy or sell securities. All figures come from the primary sources linked above and carry the as-of date stated with them; operating figures refer to December 31, 2025 or June 30, 2026, market data to August 20/21, 2026. Share prices can fall substantially at any time and a total loss is possible. The author holds no position in Stewart Information Services Corporation at the time of publication. Verify every investment decision yourself and seek independent advice if in doubt.