T. Rowe Price

T. Rowe Price manages money for other people and takes a percentage of it every year. That percentage is currently 38.1 basis points — thirty-eight hundredths of one per cent. On $1.89 trillion, that comes to roughly $7.2 billion of annual revenue, and about thirty cents in every dollar of it drops through to operating profit.

It is a wonderful business model in the abstract. There is no inventory, no factory, no receivable that might not be paid. The fee is deducted automatically from the assets before the client ever sees them. Costs are mostly people, and people can be reduced. When markets rise, revenue rises with them for no additional effort at all.

It has exactly two problems, and they are the same problem. The percentage can fall, and the assets can leave.

This is the whole analysis, and it took no interpretation to produce — it is the company's own table. A record was announced. The record was delivered by the stock market. The firm's own contribution was negative six and a half billion dollars.

We want to be fair about this, because there is a version of the criticism that is unfair. Every asset manager's assets go up when markets go up, and there is nothing dishonest about reporting the total. Rob Sharps, the chairman and chief executive, also said in the same statement that "fundamental active equity remains under pressure" — which is true, direct, and not what a man hiding something says.

But an investor needs to separate the two engines, because only one of them belongs to the company. Markets will do what markets do. Flows are the scoreboard for whether clients want what you sell.

Three of the four asset classes are growing. The fourth is half the company.

This is not a T. Rowe Price problem. It is the defining structural fact of the last two decades in fund management: money has moved from actively managed equity funds, which try to beat an index, into index funds, which do not try. The index fund charges three or four basis points. T. Rowe charges considerably more, and must therefore beat the index by more than the difference, after tax, reliably, for decades.

Rob Sharps put it plainly on the call: "active equities lost meaningful share to passive", and the migration of clients from mutual funds into ETFs and separately managed accounts has "put pressure on fees".

★ That last point is the one that makes this business genuinely difficult, and it deserves stating twice. T. Rowe's ETF business took in $4.4 billion in the quarter. That is a real success. But money arriving in an ETF at a low fee, while money leaves a mutual fund at a high one, shrinks revenue even when total assets hold. Winning the flow and losing the revenue are entirely compatible, and this company is doing both at once.

Against a headline announcing a record, this is what the chief executive told analysts a week later.

We give management credit for saying it. A company that wanted to ride the record headline would not have volunteered "meaningfully more challenging" or "air pocket" eight days later. ★ But an investor reading only the press release would have the wrong picture entirely, and that gap between the headline and the guidance is worth more than any ratio in this analysis.

Having made the structural case as hard as we can, here is the other half, and it is stronger than the bears allow.

The largest thing T. Rowe Price owns is not a fund. It is a position inside the American retirement system. Multi-asset assets are $690 billion — 36.4% of the firm — and a substantial part of that is target date funds: the default option in thousands of corporate retirement plans, into which employees' contributions flow automatically, every payday, without anybody making a decision.

The honest measure of all this: $30bn of ETFs and $61.7bn of alternatives together are about 4.9% of the firm. They are growing quickly and from a base too small to offset $13.5 billion of quarterly equity outflow. That may change over a decade. It will not change over three years.

We have spent five analyses this week explaining why standard metrics are category errors — for a manufacturer with a captive bank, for two net-lease REITs, for an industrial landlord with a $625m impairment. It is a pleasure to report that for T. Rowe Price the screen is broadly correct.

★ One consequence worth doing the arithmetic on. Strip the net cash out and you are paying roughly $21.3 billion for a business consensus expects to earn about $2.2 billion next year — around 9.7 times, against a headline multiple of 11.0.

Even the discounted cash flow figure from our data feed — $125.51, an implied +11.8% — is, unusually, a plausible number rather than a category error. An asset manager's free cash flow really is free: there is no rate base to fund, no warehouse to build, no lending book distorting the capital expenditure line. We report it here without the health warning we attached to every other company this week.

Note two things. First, the buyback was cut hard in the bad years and restored as cash flow recovered — the correct behaviour, and the same pattern we praised at Deere five days ago. Second, 2025's total return of $1,764m exceeded free cash flow of $1,479m. The difference came out of the cash pile, which the company can afford and which is a deliberate choice rather than a strain — but it is not repeatable indefinitely.

★ The pace this year is what matters for the investment case. Buybacks were $157m in the second quarter and $497m year to date — roughly 2.5% of the shares outstanding in six months. Share count is down to 213.3 million. At that rate the company retires about five per cent of itself a year.

★★ And that is the entire bull case in one sentence, so here it is explicitly: a 4.63% dividend plus a 5% annual reduction in share count is roughly 9.6% a year of shareholder return, before any growth at all and before any change in the multiple. The question is whether the business melts more slowly than that.

Revenue in 2025 was above the 2021 peak. Operating income was 41% below it, and the margin had fallen from 48.4% to 29.9%. That is what fee compression and rising costs do to an operationally levered business, and it is why the same revenue is worth so much less than it used to be.

To be fair to management, the adjusted numbers are better than the statutory ones. On the company's own adjusted basis, second-quarter operating income was $709.1m on net revenues of $1,907.4m — an operating margin of 37.2%, against 28.3% on a statutory basis, with the gap driven largely by carried interest, acquisition-related amortisation and consolidated fund accounting. Adjusted earnings per share were $2.57, up 14.7% on the year.

But look at what the analysts expect next, because it is the least ambiguous statement in this file.

Flat in 2027 and lower in 2028. Nobody covering this company is forecasting growth. ⚠️ The 2028 figure rests on a single estimate and should be treated as illustrative rather than as consensus — but the direction of the 2026-to-2027 line, which eight analysts do cover, is unambiguous: nothing.

★ Which means the investment case cannot rest on earnings growth, and we are not going to pretend otherwise. It rests on the dividend, the buyback, and the melt being slow.

T. Rowe Price yields 4.63% and has raised its dividend every year for four decades. It is one of a small group of American companies with a record that long. The payment is not in any danger. The increases are another matter.

Eleven per cent, then four consecutive years of roughly two. A two per cent increase from a company earning twice its dividend is not a capital-allocation decision. It is a streak being maintained. There is nothing dishonourable in that — but an investor should price a 4.63% yield growing at 2%, not one growing at the 11% the record implies.

★ We have now found this pattern four times in a single day — at Zoetis (15% increases for four years, then 6%, five months before a guidance cut), at Rexford (31% decaying to 1.2%), at NNN (steady at 3%, which is honest), and here. The character of a dividend increase is one of the most reliable pieces of forward information a company publishes, and almost nobody reads it.

What would force a cut. Almost nothing on any horizon we can see. Free cash flow would have to fall by roughly half and stay there, with the cash pile exhausted. Given a 51.6% payout, net cash and a business that remains solidly profitable even after a nineteen-point margin decline, we regard the dividend as one of the safest in this month's analyses.

The realistic outcome is what the table above shows: a 4.63% yield growing at two per cent, indefinitely. That is a perfectly respectable income holding. It is simply not what forty years of increases makes people imagine they are buying.

So what does 11.0× assume? It assumes that earnings are roughly flat for several years, that the equity outflow continues at something like the present rate, that the fee rate keeps grinding down a basis point or so a year, and that none of it accelerates. That is not a pessimistic assumption. It is close to a description of what is already happening.

There is no hidden value here to uncover, and we are not going to invent one. What there is, is a straightforward proposition that can be checked with a calculator: a 4.63% dividend, a share count shrinking about five per cent a year, and a business that must decline more slowly than the sum of those two for the arithmetic to work.