Click here for the latest from James J. Cramer.

How corrupt are analysts? Depends on the analyst. How corrupt are firms? Depends on the firm.

During my stint as CNBC "Squawk Box" co-host on Tuesday morning, I participated "peripherally" in a compelling roundtable with Henry Blodget and Steve Milunovich from Merrill Lynch.

Related Stories

They Are Not All Penguins!

Report Card: Henry Blodget

Report Card: Steven Milunovich

Analyst Rankings

With analysts under fire now more than ever for shilling for banking clients, it said a tremendous amount about these two fellows and their firm for allowing them to come on.

I know both gentlemen from my years at the hedge fund, and they are decent, good men who would withstand the pressure of corporate finance were it ill-applied. Merrill Lynch isn't as dependent on underwriting as others -- it has a sizable retail client base that generates lots of revenue -- so corporate finance isn't as powerful as it might be at other firms.

The discussion put the cart before the horse. The pressure to be positive comes from so many different sources that it really may be worth it to give you a primer about why there are conflicts to begin with. Nothing's black and white, and I know plenty of analysts who work their butts off to make you money and are beholden to no one.

Like in politics, with analysts you have to follow the money. If you wanted analysts to be the best stock pickers in the world, then you would incentivize them the way hedge funds incentivize: You give the people who make money a percentage of what they make. If all that mattered were making money for the clients, you would establish benchmarks of time and performance and pay the person with the best batting average the highest price.

That's not how it is done. Research, per se, generates no direct revenue. Brokers tend not to charge by the report (only a few have ever even tried), and there is a bit of a tacit bargain that "if we do well for you, could you please do some commissions with us?"

In fact, Wall Street, which is supposed to be so good at getting paid, has never done a good job monetizing research. Even the best firms haven't been able to link their most valuable analysts with a fee structure. You would think, say, that Chuck Phillips, a money maker from Morgan Stanley Dean Witter, would be able to bill his firm's clients for all of the money he has made them. A good call from Chuck out of research could make someone millions of dollars. But it's not linked at all. I used to peddle research, and good analysts from Goldman wasted a lot of people's time, and never wrote a ticket. (Slang for never did any commission.)

Surely, though, these people don't work for free. They are used internally to generate sales calls that can produce tickets. They are action points. When I was a broker, I would call everybody on any upgrade or downgrade, even if I thought the analyst was a doofus who couldn't make you money if she tried! I didn't want my client to hear or read elsewhere that Goldman had made a big call.

So some of their salary comes from sales. But how much? How do you determine how much? It isn't easy and it is very subjective. I always thought every sale I ever got was "mine" from my clients. My bosses always thought the sales were the firm's, in that it wouldn't happen if I weren't at Goldman. The analysts always thought that the intellectual property behind the idea was theirs! It is awfully hard to figure out a split.

With the reduction in day-to-day commissions to really low levels, however, one thing was certain: There wasn't enough money to go around. However, there is one major area in which the fees haven't declined at all: underwritings.

When we say underwritings on Wall Street, we are talking about initial public offerings and secondaries. These are big tickets. I usually paid a nickel a share for day-to-day trading. But the issuing companies pay giant fees, dollars even, per share to get the deals done.

Here's the rub, though. Underwritings are corporate finance, and the corporate finance department is a breed apart. If sales forces are gigantic, with many people creating a mesh, a giant web, to snare commissions, corporate finance departments are small, and their members powerful.

They are the ones who bring in the big underwriting fees. Of course, they can pitch that the firm has a big sales force to sell the deal, and they can even pitch that they have analysts who will cover the company.

But it is right there, right in that nexus, that the conflict glows and can't be dampened. Sometimes, it is a slow burn, as when the corporate finance titan says: "We sure could use some help getting National Gift's business."

Sometimes it rages out of control, when corporate finance links your bonus directly to how much business you bring in for the firm. As it is extremely difficult to bring business in by advertising to National Gift that you are going to "be objective" and rate it a hold, an analyst gets himself in a difficult spot at one of these incendiary places.

Analysts are often compensated by both corporate finance and sales, and neither of these two paymasters is that interested in how the analyst has done for you, the retail client, because you are simply not as important and don't pay his bills.

Now, none of this process ever received much scrutiny before these folks started going on television. In fact, I think that in many ways, television was the worst thing that could have ever happened to the cozy quiet world of analyst as salesperson for the firm because, somewhat unwittingly, these analysts would come on television and give you the appearance that they were objective.

None of us internally at these places ever regard these peoples as objective. We all recognize the subtle constraints that keep them from "telling the truth." We know how their bread is buttered. But when they went on television, they held themselves out as being the objective stock pickers that they never were.