The Second Number

I · One Hundred and Eleven Per Cent

On the eleventh of August, 2026, a lens manufacturer in California filed its quarterly report with the Securities and Exchange Commission and reported that net sales for the three months ended 3 July had risen 111.0 per cent. The company is STAAR Surgical, and it makes implantable collamer lenses—small foldable discs that a surgeon inserts into the eye through a very small incision, correcting vision without removing corneal tissue. Everything else in the summary was of a piece with the headline. Gross margin, 74.5 per cent. Net income of $8.1 million, against a loss of $16.8 million in the same quarter a year before. Cash and investments of $181.5 million, up from $163.9 million three months earlier. No debt at all. It is difficult to construct a more encouraging paragraph out of true sentences, and every sentence in it is true.

Some pages later, in the same document, is the statement of cash flows. For the six months ended 3 July 2026, the ordinary business of the company consumed two million dollars in cash. Net income of $13.3 million for the half, plus $30.1 million of non-cash items added back, less $45.4 million absorbed by working capital, comes to minus two. The single largest line in that working-capital drain is accounts receivable, and the balance sheet shows the size of it: trade receivables of $98.5 million against $50.1 million at the start of the year. The company sold a great deal more and collected rather less of it.

Neither passage is a lie. They are both in the filing, they are both audited to the same standard, and a reader who stops at the first has not been deceived so much as unaccompanied. This is the ordinary condition of financial reading: the number that leads a report is almost never the number that decides it, and the number that decides it is generally printed some pages later in a smaller typeface, in a statement nobody quotes.

For a New York reader of financial news, the companies can be anywhere while the reading problem remains close at hand. A market screen brings their headlines into one column; the filings keep the denominators, obligations, and reporting clocks in separate documents. The discipline is to resist treating the screen as if it had already done the comparison.

What follows is an attempt to say what the difference consists of, using seven securities that happened to be examined in the same week of August 2026 by a private investor who wanted to know what they were looking at. They are a deliberately awkward set: an Israeli shipping line in the middle of a takeover, an American biotechnology company having the best product launch of its year, a genomics laboratory, the lens maker above, a Texan middle-market lender, a London-listed office landlord, and an exchange-traded fund advertising a distribution rate of 48.55 per cent. Between them they cover most of the ways a security can be misunderstood. Every headline figure attached to them is arithmetically correct. Not one of the seven is enough to act on, and the reason is the same reason in each case, wearing seven costumes.

II · The Borrowed Denominator

Return to the 111. It is a growth rate, and a growth rate is a fraction, which means it has two ends and the headline shows you one. STAAR’s June-quarter sales were $93.5 million against $44.3 million in the corresponding quarter of 2025. The difference is $49.2 million. In the same filing, two sentences below the 111, the company reports that its two distributors in China took $52.3 million of product in the quarter just ended, and $5.3 million in the quarter it is being compared with. Which means that $47.0 million of the $49.2 million of additional revenue—the overwhelming majority of the most impressive number in the document—is the difference between two Chinese quarters, one of which was very nearly empty.

Take those two distributors out of both ends and the same filing yields a different sentence. Sales to the rest of the world went from about $39.0 million to about $41.2 million, which is growth of roughly six per cent. That subtraction is this essay’s arithmetic, not the company’s, but every figure in it is the company’s own. Six per cent and 111 per cent are both accurate descriptions of the same three months. They differ only in what they are being compared with, and the choice of comparison was not made by an analyst. It was made by a distributor’s inventory decision in the spring of 2025.

This is not a modern discovery, and the classical remedy is older than the filing requirement. When the security analysts Benjamin Graham and David Dodd published Security Analysis in 1934, in the wreckage of a market that had believed a great many single-year comparisons, their prescription for earnings was to average them over a span of years long enough that no single year could carry the conclusion. The point was not that a long average is more accurate. A long average is often less accurate about the present. The point was that a base year is a choice, and a method that lets somebody else make the choice for you is not a method.

The modern version of the discipline is smaller and more portable. When a percentage is startling, find its denominator before you find its story. Sometimes the denominator is a weak quarter, as here. Sometimes it is a small number of units, so that a doubling means eight customers became sixteen. Sometimes, as in Part V, it is one week’s payment multiplied by fifty-two. The startling figure is generally not the thing being disputed; it is the thing being borrowed against.

None of this settles whether STAAR is doing well. It might be. The company’s gross margin genuinely improved, its factory is genuinely scaling, its balance sheet genuinely carries no debt, and 56 per cent of a quarter is a real business even if it is a concentrated one. What the decomposition settles is narrower and more useful: the 111 per cent is not evidence about demand. It is evidence about a comparison. To learn anything about demand one has to go somewhere else in the document, which is what the rest of this essay is about.

III · What the Cash Says

The reason the cash-flow statement keeps rescuing the reader is that profit is an opinion about timing and cash is a fact about a bank account. Revenue is recognised when it is earned, which is not when it is received; costs are matched to the period they belong to, which is not when they are paid. The gap between the two—accruals, in the trade—is not fraud and is not optional. It is what makes accounting more informative than a cash register. But it is also the part of the accounts most exposed to judgement, and judgement has a direction.

The professional literature settled this some time ago and in an unusually clean way. In 1996 the accounting scholar Richard Sloan split reported earnings into their cash and accrual components and asked which of the two better predicted the following year’s earnings. The cash component won, comfortably: earnings driven by accruals were less persistent, reverting faster. He then showed that share prices behaved as though investors did not distinguish between the two, so that a portfolio buying low-accrual firms and selling high-accrual ones earned an abnormal return. That paper, in The Accounting Review, is the origin of a discipline that a great many people practise without knowing whom to credit.

Honesty requires the sequel. In 2011 the accounting researchers Jeremiah Green, John Hand, and Mark Soliman published a study in Management Science whose title asked whether the accruals anomaly had gone: they found that the hedge returns to Sloan’s strategy had decayed in United States markets to the point of being, on average, no longer reliably positive, and attributed that in part to an increase in the capital hedge funds had put into exploiting it. What died, though, was the free money, not the fact. Accruals remained less persistent than cash; it simply stopped being possible to be paid for noticing. For a reader trying to understand a company rather than to beat a market, that distinction is the whole of the matter, and it leaves the tool intact.

Applied to the seven, the tool is immediately loud. STAAR has already been done: $13.3 million of half-year profit and minus $2.0 million of half-year cash, the difference sitting almost entirely in a receivable. But the most instructive case is Insmed, which by any commercial measure is having the best year in the set. Its June-quarter product revenues were $425.5 million against $107.4 million a year earlier, a rise of 296 per cent, of which $308.6 million was United States sales of BRINSUPRI, an oral treatment for non-cystic fibrosis bronchiectasis approved there the previous August. Cost of product revenues was $67.2 million, which is a gross margin around 84 per cent. Launches of that shape are rare enough to be worth watching for their own sake.

The quarter’s operating loss was $1.5 million—within a rounding error of break-even, on a company that spent $210.0 million on research and $247.5 million on selling and administration in the same three months. The reconciliation is one line above: a credit of $99.8 million from the change in fair value of contingent consideration, which is the estimated future money owed to the former owners of a business Insmed once bought. The 10-Q explains what moved it. The change was “primarily due to the decrease in our share price.” The shares fell; the estimated liability fell with them; the fall was booked as a credit against operating expenses; and the operating loss came in ninety-nine million dollars smaller than the operations produced. Strip the credit out and the quarter loses about $101 million at the operating line. The half-year statement of cash flows, which cannot be moved by a share price, reports $311.6 million consumed.

It is worth being precise about what that is and is not. It is not an irregularity; contingent consideration is required to be remeasured, the movement is disclosed on the face of the income statement, and the company states its cause in plain English. It is an illustration of something more interesting than misconduct, which is that an income statement can be moved by an event that made shareholders poorer, in the direction of looking better. No cash-flow statement can do that.

The check runs both ways, which is why it is a check and not a prejudice. GeneDx, the rare-disease sequencing laboratory, reported a June-quarter net loss of $17.7 million under generally accepted accounting principles alongside an adjusted profit measure of its own definition—the sort of pairing that invites suspicion, and which the Securities and Exchange Commission has regulated since 2003, when it adopted Regulation G and Item 10(e) of Regulation S-K under the Sarbanes-Oxley Act, requiring any non-standard measure to be presented alongside the standard one and reconciled to it. The rule is the reason the honest version is always in the same document as the flattering one. In GeneDx’s case, however, the cash statement is kinder than the profit statement: a half-year net loss of $81.1 million against $61.4 million of operating cash consumed. The company is losing less cash than it is losing money.

What is not kind is the direction of travel. In the first half of 2025 the same business generated $20.6 million of operating cash. A swing of eighty-two million dollars in a year is the sort of thing a reader wants explained, and the explanation is in the operating expenses: selling, general and administrative costs rose 62 per cent, to $76.0 million from $46.9 million, in a quarter when revenue rose 11 per cent. Underneath, the unit economics are moving in the same direction. Exome and genome test volumes grew 32 per cent; the revenue those tests produced grew 17 per cent; average reimbursement per test fell 12 per cent. The laboratory is running nearly twice as fast as the money it is being paid for running.

IV · Seven Instruments, Not Seven Stocks

The commonest error among beginners is not credulity. It is applying one ratio to everything. A price-to-earnings multiple is a serviceable question to ask of a manufacturer and a meaningless one to ask of a fund that has no earnings, a lender whose earnings are an accounting estimate of credit, or a company that is about to be bought for cash at a price already printed in a contract. The seven securities in this file are not seven versions of one thing. They are five different kinds of legal and economic claim, and each kind fails in its own way.

Three—the lens maker, the biotechnology company, the sequencing laboratory—are operating companies, which fail by selling things for less than they cost, or by not being paid, or by spending the future faster than the present arrives. One is a lender. One is a takeover in progress, which fails by not completing. One is an office landlord, which fails by owing more than the buildings turn out to be worth. And one is a derivative fund, which fails by being mistaken for something else.

The lender is the instructive case, because it is the one most often read with the wrong instrument. Capital Southwest is a business development company—a creature of the Investment Company Act of 1940, a listed vehicle that lends to businesses too small for the bond market and, being taxed as a regulated investment company, distributes nearly all of what it earns. At the end of June its portfolio was $2.2 billion at fair value, of which a $2.0 billion credit book was 99 per cent first-lien senior secured debt, yielding a weighted average of 10.9 per cent. Net asset value was $16.61 per share. Its regulatory debt-to-equity ratio was 0.91 to one.

Every one of those figures answers a question that a price-to-earnings ratio cannot ask. Take the last. Since the Small Business Credit Availability Act of 2018—which arrived, in the American manner, inside a spending bill—a business development company that obtains the right approvals may operate at 150 per cent asset coverage rather than 200, which is to say at two turns of debt to equity rather than one. Capital Southwest is running at 0.91. It is using less than half the leverage the law now permits, and a reader who did not know the limit had moved would not know that the restraint was a choice.

Or take the non-accruals, the loans on which the lender has stopped booking interest because it no longer expects to receive it. They were 1.1 per cent of the portfolio at fair value and 2.9 per cent at cost. The pair is more informative than either figure. A cost basis of $65.7 million carried at $23.4 million means the impaired positions have already been written down by roughly two-thirds; the bad news, whatever else is true, has largely been taken. That is a sentence about credit conservatism, and it is invisible to every multiple in ordinary use.

V · Two Yields

Capital Southwest paid its shareholders $0.64 for the June quarter and earned, before tax, $0.57 of net investment income per share. On the face of it the dividend is not covered, and the shortfall has a name: the $0.64 is two payments wearing one number, a regular quarterly dividend of $0.58 and a supplemental of $0.06. Against the regular dividend alone, $0.57 is coverage of 98 per cent. Against the total it is 89 per cent. Those are two different promises and they deserve two different figures, which is precisely why the company declares them separately.

There is a further turn of the screw in the same press release, and it is the sort of thing that decides arguments. Two lines below the pre-tax figure, the company reports net investment income after tax of $0.58 per share—a tax benefit in the quarter having lifted it. On that measure the regular dividend is covered exactly, and the total at 91 per cent. Nothing has been concealed; both numbers are printed, both are labelled, and the difference between them is a deferred tax entry. But whether the sentence “the dividend is covered” is true depends entirely on which of two adjacent lines the reader divides by, and no headline yield will tell them which one was used.

Now set beside it a yield of a different species altogether. The YieldMax NVDA Option Income Strategy ETF, ticker NVDY, showed on its own fund page in the third week of August 2026 a distribution rate of 48.55 per cent and a thirty-day SEC yield of 2.61 per cent, printed side by side and both as of the same date. The gap between them is a factor of about nineteen, and it is not a discrepancy. It is a definition, and the sponsor prints it: the distribution rate “is calculated by annualizing the Fund’s most recent distribution and dividing the resulting amount by the Fund’s most recent NAV,” and “represents a single distribution from the ETF and does not represent its total return.” The thirty-day SEC yield, by contrast, is a standardised calculation prescribed by the Commission in Form N-1A, which is why every fund’s version of it is comparable with every other fund’s and no fund advertises it.

The composition question is answered on the same page, and the answer is a statute. Section 19(a) of the Investment Company Act makes it unlawful for a registered fund to pay a distribution from any source other than net income without a written statement disclosing the sources; Rule 19a-1 prescribes what the statement must contain, and the purpose, in the Commission’s own words, is to prevent shareholders from confusing income dividends with distributions made out of capital. YieldMax accordingly publishes a table with an estimated return-of-capital column and a footnote pointing to its 19a-1 notices. For the distribution paid on 19 August 2026, the fund’s estimate was 10.87 per cent return of capital and 89.13 per cent income—a good month, by the standards of the same table’s earlier rows, one of which estimates 94.58 per cent.

None of that makes the fund a fraud, and the criticism it usually attracts—that a return of capital is your own money handed back—is true but not the most important thing. The most important thing is structural: the fund writes call options against synthetic exposure to a single share, which means it exchanges part of any rally for a premium while keeping the whole of any fall. What it converts is not risk into income. It is upside into income. Whether that is a good exchange is a question about volatility, strike prices, the 1.09 per cent expense ratio, and the reader’s own tax position; it is not a question that 48.55 per cent can answer, and the sponsor, to its credit, says so in the footnote under the figure.

VI · The Missing Probability

On 16 February 2026 Hapag-Lloyd agreed to buy ZIM Integrated Shipping Services for $35.00 a share in cash. ZIM’s shareholders approved the transaction on 30 April. On the evening of 21 August the shares closed at $28.25. The gap to the contractual price is 23.9 per cent of what a buyer would pay for them, and that figure is generally described as the spread—and, in conversation, as upside. It is not upside. It is upside multiplied by a probability that nobody has written down.

The arithmetic is short enough to do in a paragraph. If the transaction completes, the holder receives $35. If it fails, the holder owns a shipping line at whatever price a shipping line without a bid is worth—call it B. A price of $28.25 is then the market’s weighted average of those two outcomes, and the probability of completion that makes the two sides balance is the difference between the market price and the break price divided by the difference between the offer and the break price. At a break price of $20, the answer is 55 per cent. At $15 it is about 66 per cent; at $25, about 33. Those are not forecasts. They are the arithmetic already contained in the quotation, and their spread across a plausible range of B is a fair description of how little the quoted number determines.

What determines it is in the documents, which in this case run to a couple of hundred pages of merger agreement. Three clauses matter more than the rest. There is an Outside Date of 17 February 2027, after which either side may walk away, extendable in specified regulatory circumstances. There is a Parent Termination Fee of $160,000,000, payable by the buyer—but only on the particular failures the agreement enumerates, which is not the same as every way the transaction can die. And there is the Special State Share, an instrument held by the State of Israel that is explicitly carved out of the shares being converted into cash and given a section of the agreement to itself. A reader who has understood those three has a view about B and about the calendar. A reader who has only understood 23.9 per cent has a number.

The underlying business, for what it is worth, had a good quarter: revenue of $1.78 billion, up nine per cent; net income of $64 million; free cash flow of $386 million; carried volume of 922,000 twenty-foot equivalent units. But shipping is the most cyclical industry in the world and the company carries $2.77 billion of net debt, most of it leases on vessels it must operate whether or not there is cargo. The standalone value is exactly the quantity nobody knows, and it is the quantity on which everything depends.

The finance literature has a mordant description of the resulting return profile. In 2001 the economists Mark Mitchell and Todd Pulvino examined 4,750 mergers between 1963 and 1998 and found that risk-arbitrage returns were essentially uncorrelated with the market in flat and rising conditions and sharply correlated with it in severely falling ones—a payoff, as they put it, resembling a written uncovered index put. Small premiums, collected steadily, and then a loss that arrives on the same day as everybody else’s. Their sample ends in 1998 and nothing here says it describes 2026. It does describe the shape of the bet, which is that the reward is bounded by a contract and the punishment is not.

VII · The Leveraged Remainder

CLS Holdings owns offices in Britain, Germany, and France, and by the arithmetic of its own accounts it is the cheapest security in this file by a distance. At 46.5 pence, the shares were quoted at about a quarter of the 177.7 pence of European Public Real Estate Association net tangible assets the company reported at 30 June 2026—a discount of some 74 per cent. There are readers for whom that sentence is the end of the analysis, on the theory that a thing at a quarter of book cannot fall much further. It can, and the reason is arithmetic rather than sentiment.

Equity in a leveraged asset is a remainder. CLS reported a balance-sheet loan-to-value ratio of 51.6 per cent, which means that of every hundred pounds of property, the lenders have first claim on 51.6 and the shareholders own what is left. Suppose the portfolio falls ten per cent and the debt does not move. The loan-to-value ratio goes to about 57 per cent, which sounds survivable. The shareholders’ remainder goes from 48.4 pounds to 38.4, which is a fall of about 21 per cent. That calculation is this essay’s, but it uses only the reported ratio, and it generalises: at half loan-to-value, every movement in the asset arrives at the equity roughly doubled. The discount is not a cushion against that. The discount is a description of it.

The rest of the half-year statement is in the same key. Interest cover—earnings before interest, tax, depreciation and amortisation, divided by net finance costs—was 1.6 times. Net debt was 12.2 times the same earnings measure. Vacancy was 14.5 per cent. The portfolio was written down 4.6 per cent in constant currency in six months; the statutory loss after tax was £69.6 million against £24.4 million a year earlier; European Public Real Estate Association earnings per share fell 32.5 per cent to 2.7 pence; no interim dividend was declared. And £182.6 million of loans fall due by September 2027, which is why the directors set their going-concern assessment to run exactly that far, and why the announcement contains the following sentence: the timing and value of the planned refinancings and disposals being outside management’s control, “a material uncertainty exists that may cast significant doubt on the Group’s ability to continue as a going concern.”

That sentence is the single most consequential line in all seven documents, and it is worth noticing how it reaches a reader. Four of these securities are domestic registrants and file the quarterly report on Form 10-Q that Rule 13a-13 requires of them. ZIM, as a foreign private issuer, is not obliged to file 10-Qs at all and furnishes its results on Form 6-K; the exchange-traded fund is a registered investment company and reports to its shareholders on a cycle of its own. CLS is a London company and reports twice a year, under the half-yearly obligation in chapter 4.2 of the Financial Conduct Authority’s Disclosure Guidance and Transparency Rules. The quarterly interim management statement that used to sit alongside it was deleted from those rules in November 2014, when the amended European Transparency Directive removed the requirement. So the going-concern uncertainty was published on 12 August and will not be revisited by the company until the annual results. In the same window a reader of any of the four American filers will have received two further quarters of numbers.

The measure carrying the discount is itself worth a sentence, because it is not a statutory figure at all. European Public Real Estate Association net tangible assets is defined by a Brussels trade body, in Best Practices Recommendations whose October 2019 edition replaced a single EPRA net asset value with three—reinstatement, tangible assets, disposal—on the reasoning that one number could not answer three different questions about what a property company is worth. CLS reports 177.7 pence of net tangible assets and 164.2 pence of statutory net asset value on the same page. Neither is wrong. They are answers to different questions, and the 74 per cent discount is quoted against whichever of them the quoter preferred.

A New York reader of these securities receives disclosures on different clocks. The rhythm at which they may ask whether the story has changed is set partly by where each security is listed and by the reporting rules attached to it. That observation is argued rather than measured—no one has counted the cost of the mismatch to a reader holding both securities—but the published cadences differ by a factor of two.

VIII · The Wall Street Gate

A New York market screen can put all seven securities on one watchlist, but it cannot make their businesses or reporting duties alike. The common task on Wall Street is to write down the second number that would change one’s mind before the first number changes the price. This is a reading discipline, not a local law of returns.

The seven documents have, between them, one structural feature in common, and it is the feature that makes the whole exercise possible rather than merely gloomy. In every single case the qualifying number is in the same file as the flattering one. STAAR prints the China concentration two sentences after the 111 per cent, and the cash statement a few pages later. Insmed prints the contingent-consideration credit on the face of the income statement and names the share price as its cause. GeneDx reconciles its adjusted profit to its statutory loss because the Commission has required that since 2003. Capital Southwest separates the regular dividend from the supplemental and reports its net investment income both before and after tax. YieldMax defines its own distribution rate in a footnote directly beneath it. Hapag-Lloyd’s lawyers wrote the Outside Date and the Special State Share into a public exhibit. CLS states its going-concern uncertainty in the language the accounting standards require. Nobody hid anything. The concealment, where there is any, is done by the reader.

What the seven do not contain is a verdict, and this is where the reading has to end in something other than a feeling. The useful terminal object is not an opinion about the security but a condition attached to a document: a statement of what would have to appear, and where, for the view to change. Receivable days falling and cash arriving from China would say something about STAAR that the 111 per cent cannot. A reimbursement rate that stops falling would say something about GeneDx that a volume figure cannot. A completed refinancing at CLS, an operating cash burn at Insmed that narrows without help from its own share price, a regulatory approval list at ZIM with items crossed off, a coverage ratio at Capital Southwest computed on a stated basis: each of those is checkable by somebody who does not care what the answer is. That is the property that matters. A condition that only its author can evaluate is not a condition.

It is worth saying plainly what this method does not deliver, because the promise usually attached to it is false. It does not produce certainty, or a price target, or an edge. Sloan’s accrual signal made money for about a decade after it was published and then stopped, which is roughly what one should expect of any regularity that becomes widely known. The reader who follows every step of the discipline set out here will still be wrong a great deal of the time, because the future of a Chinese distributor’s inventory, a regulator’s calendar, and a London office market are all genuinely unknown, and no amount of reading converts an unknown into a known.

What it delivers is smaller and more durable: knowing which sentence you are betting on. The lens maker’s quarter may yet turn out to have been the beginning of a real recovery, in which case the receivable will convert to cash and the 111 per cent will look, in retrospect, like an understatement. Or the distributors will be found to be holding lenses that Chinese surgeons are not implanting, and the receivable will come back as a credit note. The filing does not say which, and neither does this essay. What the filing says—the whole of what it says, if one reads to the end—is that the company sold ninety-three and a half million dollars of lenses in three months and finished the half-year two million dollars poorer in cash. Both numbers were printed on the same day, by the same company, under the same signature. The first one travelled. The second one is the one that will decide.

Notes on Sources

STAAR Surgical Company. Quarterly Report on Form 10-Q for the quarterly period ended 3 July 2026, filed 11 August 2026. Source of net sales of $93.5 million and the 111.0 per cent increase from $44.3 million; sales to the two Chinese distributors of $52.3 million and $5.3 million; gross margin of 74.5 per cent against 74.0; net income of $8.1 million against a loss of $16.8 million; cash and investments of $181.5 million; trade receivables of $98,475 thousand against $50,064 thousand; and the six-month cash reconciliation of $13.3 million of net income plus $30.1 million of non-cash items less $45.4 million of working capital, of which the receivables movement was $49,230 thousand.

Insmed Incorporated. Quarterly Report on Form 10-Q for the quarterly period ended 30 June 2026. Source of net product revenues of $425,486 thousand against $107,415 thousand (an increase of 296 per cent), of which $308.6 million was United States commercial sales of BRINSUPRI; cost of product revenues of $67,212 thousand; research and development of $210,034 thousand and selling, general and administrative expense of $247,467 thousand; the operating loss of $1,548 thousand; the $99,760 thousand change in fair value of contingent consideration and the quoted attribution of it to the decrease in the share price; the half-year net loss of $176.8 million; and net cash used in operating activities of $311,566 thousand against $467,657 thousand.

GeneDx Holdings Corp. Quarterly Report on Form 10-Q for the quarterly period ended 30 June 2026. Source of revenue of $114.4 million, up 11 per cent from $102.7 million; the 32 per cent rise in exome and genome test volumes against the 17 per cent rise in the revenue they produced and the 12 per cent fall in average reimbursement rates; selling, general and administrative expense of $76.0 million against $46.9 million, up 62 per cent; the quarterly net loss of $17,740 thousand and the half-year net loss of $81,056 thousand; and net cash used in operating activities of $61,389 thousand against $20,602 thousand provided a year earlier.

Capital Southwest Corporation. Earnings release for the quarter ended 30 June 2026, dated 3 August 2026. Source of the $2.2 billion portfolio and $2.0 billion credit portfolio at 99 per cent first-lien senior secured; the 10.9 per cent weighted average yield; non-accruals of $23.4 million at fair value (1.1 per cent) and $65.7 million at cost (2.9 per cent); pre-tax net investment income of $35.0 million, or $0.57 per share, and net investment income after tax of $0.58 per share; the $0.58 regular and $0.06 supplemental dividends totalling $0.64; net asset value of $16.61 per share; and the regulatory debt-to-equity ratio of 0.91 to one.

ZIM Integrated Shipping Services Ltd. Agreement and Plan of Merger among Hapag-Lloyd AG, Norazia (Israel) Ltd, and ZIM, dated 16 February 2026, filed as an exhibit to a report on Form 6-K. Source of the $35.00 per share merger consideration, the Outside Date of 17 February 2027 in section 8.1(d) and its conditional extension, the $160,000,000 Parent Termination Fee and the circumstances conditioning it, and the treatment of the Special State Share, which is excluded from the shares converted and addressed in section 6.2. ZIM’s second-quarter 2026 results, furnished on Form 6-K on 19 August 2026, are the source of revenue of $1.78 billion (up nine per cent), net income of $64 million, adjusted EBITDA of $491 million, free cash flow of $386 million, carried volume of 922 thousand containers, and net debt of $2,773 million; and of the shareholder approval of 30 April 2026.

CLS Holdings plc. Half-yearly results announcement for the six months ended 30 June 2026, dated 12 August 2026. Source of EPRA net tangible assets of 177.7 pence (from 200.7) and statutory net asset value of 164.2 pence; the loan-to-value ratio of 51.6 per cent (from 50.0); interest cover of 1.6 times and net debt to EBITDA of 12.2 times; net debt of £808.3 million; vacancy of 14.5 per cent; the 4.6 per cent constant-currency portfolio decline; the statutory loss after tax of £69.6 million against £24.4 million; EPRA earnings per share of 2.7 pence, down 32.5 per cent; the absence of an interim dividend; the £182.6 million of loans expiring by September 2027; and the quoted material-uncertainty statement on going concern.

YieldMax NVDA Option Income Strategy ETF (NVDY), sponsor’s fund page, consulted 22 August 2026. Source of the 48.55 per cent distribution rate and 2.61 per cent thirty-day SEC yield, both as of 19 August 2026; the quoted definition of the distribution rate and the statement that it does not represent total return; the 1.09 per cent gross expense ratio; net assets of $1.40 billion and a net asset value of $12.61 as of 20 August 2026; the estimated composition of the 19 August distribution as 10.87 per cent return of capital and 89.13 per cent income; and the estimated 94.58 per cent return of capital on an earlier distribution.

Sloan, R. G. (1996). Do stock prices fully reflect information in accruals and cash flows about future earnings? The Accounting Review, 71(3), 289–315.

Green, J., Hand, J. R. M., & Soliman, M. T. (2011). Going, going, gone? The apparent demise of the accruals anomaly. Management Science, 57(5), 797–816.

Mitchell, M., & Pulvino, T. (2001). Characteristics of risk and return in risk arbitrage. The Journal of Finance, 56(6), 2135–2175. The sample is 4,750 mergers between 1963 and 1998.

Graham, B., & Dodd, D. L. (1934). Security Analysis. New York: McGraw-Hill, for the multi-year averaging of earnings.

United States Securities and Exchange Commission. Conditions for Use of Non-GAAP Financial Measures (Regulation G and Item 10(e) of Regulation S-K), adopted January 2003 under section 401(b) of the Sarbanes-Oxley Act of 2002. Exchange Act Rule 13a-13, for the quarterly report on Form 10-Q. Form N-1A, Item 26, for the standardised thirty-day yield. Investment Company Act of 1940, section 19(a), and Rule 19a-1 thereunder, for the written statement disclosing the sources of a distribution; the characterisation of the rule’s purpose follows the Commission’s own description of it. Small Business Credit Availability Act, enacted 23 March 2018 within the Consolidated Appropriations Act, amending section 61 of the Investment Company Act to permit 150 per cent asset coverage.

Financial Conduct Authority. Disclosure Guidance and Transparency Rules, chapter 4.2 (half-yearly financial reports) and the deleted chapter 4.3 (interim management statements), following the amending Transparency Directive; the deletion is recorded in policy statement PS14/15 and the Handbook page carries a last-updated date of 6 November 2014.†

European Public Real Estate Association. Best Practices Recommendations Guidelines, October 2019, introducing net reinstatement value, net tangible assets, and net disposal value in place of EPRA net asset value and triple net asset value, for accounting periods beginning on or after 1 January 2020.

† marks the one regulatory chronology taken from a handbook page and a policy statement rather than from the amending instrument itself. Market prices—$28.25 for ZIM and 46.5 pence for CLS—are closing quotations of 21 August 2026 and are not executable. Figures identified in the text as this essay’s arithmetic are the ex-China growth rate of about six per cent, the 56 per cent China share of STAAR’s quarter, the $47.0 million of the increase attributable to the two distributors, the roughly $101 million operating loss at Insmed before the contingent-consideration credit, the coverage percentages of 98, 89, and 91, every point on the curve in Fig. 4, and the equity sensitivity in Fig. 5; all of them use only reported inputs. The research note that prompted this essay reported BRINSUPRI revenue of $309.2 million; the 10-Q states $308.6 million of United States commercial sales, and the filing has been preferred. Nothing here is a recommendation, and no conclusion about any of the seven securities is offered or implied.

Drafted with AI (Claude, Anthropic) and fact-checked against the primary filings listed above, August 22, 2026; every company figure was re-read at its source rather than taken from the research note that prompted the piece, and where the two disagreed the filing won. Nothing here is investment, tax, or legal advice. Set, with affection and without affiliation, in the manner of a certain Manhattan weekly.New York revision drafted with AI on October 12, 2026; editorial review pending. ◆ ◆ ◆