Lazard Inc LAZ

Key data

The business

Lazard runs two businesses under one 177-year-old name. The larger and better one is Financial Advisory: a pure advice shop that tells companies and governments how to buy each other, break themselves up, raise money, and restructure debt they cannot pay. It carries no loan book and takes no market risk, so when a deal closes Lazard keeps almost the entire fee. The smaller engine is Asset Management, which runs stock and bond funds for institutions and charges a slice of the money each year. That business is the steady one, but active managers like it have bled money to index funds for a decade. The moat is the advisory name: a sovereign in a debt crisis or a board facing a hostile approach calls a short list of firms, and Lazard has sat on that list since before most countries on it existed. What the client actually buys is a senior banker's judgment in a room, not a product, which is why the whole model lives and dies on paying those bankers.

The numbers

The revenue line is not the problem here. The pay line is.

Money coming in has held in a tight band near $800 million a quarter, and the spring quarter was up 2.8% on the year. But read the last column. Diluted earnings fell from $0.91 in the first quarter to four cents in the second, and net income dropped to $4.8 million from $55 million a year earlier, even as revenue rose. That is not a demand problem, it is a cost problem: operating income fell about 60% year on year as compensation swallowed the top line. The watch coming into this print was whether the pay ratio would turn back toward its target, and it broke the other way. Both of the last two quarters missed the Street's earnings mark, the June quarter by more than half.

Step back and the decade tells the real story. Revenue has compounded about 4% a year since 2022, but diluted earnings per share have shrunk almost 15% a year over the same stretch, and free cash flow per share with them. The share count went the wrong way too, up 5% in three years, because stock handed to bankers as pay more than offsets the buybacks. So the owner of this stock has watched the business get modestly bigger while their slice of it got smaller. 2023 shows why this name is not a compounder you buy at any price: one bad advisory year and the whole firm posted a loss. What I believe that the tape does not: the earnings you see today sit near a cyclical trough, and the multiple looks high only because it rests on depressed profit. The print that settles it is the awarded compensation ratio in the next filing.

Management

Peter Orszag, the former federal budget director, has run Lazard since 2023, and the record so far is mixed. His 2025 pay of $5.0 million was a modest 2% of net income, reasonable, though his 2023 package topped $30 million on a one-time grant to take the job. More telling is what insiders do with their own cash: nine open-market sales in the last year, no buys, about $14.8 million sold. Orszag himself sold $5.4 million of stock on August 25, weeks after the quarter that halved profits, and the filing does not say whether that sale was pre-scheduled, so I will not assume it was. Buybacks have been poorly timed, $91 million spent in 2025 at an average of $49, above today's $45, and they still did not stop the share count from rising. The dividend, $2.00 a year for a 4.4% yield, now eats 84% of earnings, which is fine at trough profit and a problem if the trough deepens.

How it fails or surprises you

The pay ratio stays broken. Advisory revenue held up this year; earnings did not, because compensation ate the gain. If the awarded comp ratio stays where the June quarter put it, earnings per share sit near this trough no matter how many deals close, and the 84% dividend payout becomes the pressure point. The next filing's comp ratio is the number to read first.

The one that argues against me. Net income fell to $4.8 million on rising revenue in June. If that is structural, a firm that cannot turn record fee revenue into profit, then the cheap-franchise read is wrong and this is a value trap. The two prints ahead have to show margin returning, or the read fails.

The deal cycle turns (right tail). This model has violent operating leverage: fees fall almost straight to the bottom line when volume returns, and restructuring and M&A can run hot at once. A real advisory upcycle could take earnings back toward the $3.50 the firm cleared in 2022, and nobody is paying for that at 21 times trough profit today. Announced-deal conversion over the next two quarters would show it first.

Closing thoughts

The shape of the payoff is a wide fork, and a named print decides which arm you are on. On one side, compensation stays elevated, earnings sit near this trough, the dividend gets squeezed, and you own dead money paying you 4.4% to wait. On the other, the deal cycle turns, the operating leverage that runs in reverse today runs forward, and earnings normalize toward $3 to $4, at which point 21 times trough profit was really 11 or 12 times normal profit. The left tail is a stretched payout, not insolvency: a 177-year-old advice firm with no loan book does not go bust, it goes quiet. The fatter tail from here is up, but only if the comp line breaks the firm's way, which this year it did not.

The bet, in the end, is still that the world keeps hiring Lazard's name for its biggest mergers, breakups, and debt fights, and that the deal cycle turns back up before the payout gives way. What breaks it is the pay line: if the comp ratio and quarterly operating income do not both recover over the next two prints, this trough is the new normal and the stock is fairly priced for it.