I actually own these cufflinks

It’s not difficult to find cases of sell-side analysts recommending a stock at its peak, and then keeping the buy on all the way down to its bottom – at which point it becomes a sell. Josh Brown points out an egregious example and asks:

I have no idea how much brokerage firms pay analysts to cover stocks or whether or not the costs are really offset by revenues from institutional trading. I’m assuming that whatever they pay them, it is money well spent and they make more than enough in reciprocal brokerage profits – because how else to explain it otherwise?

Conveniently the Times this weekend answers both questions. Re: how much they’re paid, the answer is “more than a potted plant, anyway,” although the compensation packages do suggest that if they were really so good at picking stocks they’d do so on the buy side:

Douglas Anmuth was lured to JPMorgan Chase earlier this year with a pay package valued at roughly $2 million. He had been making about $1.3 million at Barclays Capital, an arm of the British bank.

Heather Bellini landed at Goldman Sachs with a remarkable pay package worth almost $3 million. And Mark Mahaney, whom JPMorgan tried to hire with an offer of about $3 million, stayed on at Citigroup — after getting a raise.

The Times also notes that a 2008 study (pdf) found that average bulge bracket analyst comp is around $700,000. The study found that the main determinants of an analyst’s compensation are (1) Institutional Investor rankings (based on a survey of the buy side), (2) transaction volume in the covered stocks (more transactions = more valuable to have an analyst churning them), and (3) coverage of investment banking clients. Accuracy of EPS forecasts and performance of buy/sell recommendations were not statistically significant – suggesting that the conventional wisdom, that buy-siders like reading smart write-ups from analysts but ignore their recommendations, is accurate.

Re: do they make enough in brokerage profits to justify that? Less clear. Here’s a suggestive nugget from the Times story:

But Internet analysts are by far the hottest commodities. That is partly a function of the story, partly a function of supply. At the height of the dot-com boom, no fewer than 616 Wall Street analysts were covering Internet companies. Today, the figure is 362, according to data from Thomson Reuters.

And fewer than a dozen of those specialize in social networking stocks, as Mr. Anmuth does. Mr. Mahaney at Citigroup covers Pandora, and some people expect he will cover Facebook. As long as investors bid up these stocks, Wall Street will keep bidding up the price of its analysts.

Ah, yes, Pandora, with its $1.9 billion market cap. Or Facebook, with its $0 publicly listed float. No wonder he’s worth $3 million – just think of all the trades he can generate to buy-side traders by recommending buying or selling Pandora.*

The real answer is of course:

Hype or not, talk that companies like LinkedIn, Facebook and Groupon will change the way we live and do business — and make their shareholders rich in the process — has Wall Street pining for the fees that come with taking these companies public.

In other words, the most valuable analysts are those covering sectors that are about to see a wave of IPOs. Particularly sectors where any appropriate combination of high-Scrabble-scoring letters can get funding thrown at it after a one-minute presentation.

Of course, the Times dutifully points out that after the Spitzer research settlement analysts can no longer legally be paid for helping their firms obtain banking business. The settlement includes other restrictions that make it difficult for banks to use their sell-side analysts in marketing their capital markets work to prospective issuer clients. At the same time, bankers can and do frequently tout the quality of their – say – social media analysts to potential social media IPO candidates. That requires you to have a social media analyst as soon as there’s one shaky social media IPO – because you’re going to have a tough time getting the Facebook IPO unless your analyst has been saying smart things about LinkedIn (and, fine, maybe Google).

The researchers cited by DealBook didn’t look at this in 2008, but the interesting experiment is how analyst comp relates to the IPO shadow calendar (or, say, the percentage of value of the covered industry that is not publicly traded). Covering investment banking clients is nice (as the 2008 study found), but not that important: once a company is an investment banking client, its relationship is with the bankers, not the analysts – and those two are barely allowed to talk to each other. Bankers take pains to explain to clients that they can’t force, or even try to persuade, their analysts to change a recommendation, and many clients are loyal enough to their bankers that they’ll use them even if their bank has a bearish analyst – or none at all.

But brand-new issuers are where the money is – with IPO fees of 5+% – and have no such loyalty. They’re looking for signals of commitment and insight. And at $2 or $3 million a pop, analysts are a pretty cheap way to signal those things.

Star Analysts Are Back (No Autographs, Please) [DealBook]

* Unfair! Mahaney covers GOOG, AMZN, EBAY, and lots more large-cap tech companies. We leave as an exercise to the reader how much sell-side coverage of GOOG affects buying and selling decisions.