Wall Street Analysts Give Investors What They Want
Jan 20, 2017 7:00 AM EST
By Matt Levine
A while back I wrote this:
Here are two models of sell-side equity research.
Model 1
- Sell-side analysts are in the business of finding out what stocks will go up and then telling you.
- They tell you to Buy stocks that will go up, Hold stocks that will stay flat (why?), and Sell stocks that go down.
- You believe them, and do that.
- Sometimes they lie to you, but it is always a shock when they do.
Model 2
- Sell-side analysts are in the business of helping institutional investors get access to corporate management teams.
- They flatter management teams by giving most companies good ratings, to maintain access.
- They help their clients, because investors who meet with management tend to outperform investors who don't.
- The clients aren't too worried about the Buy/Sell/Hold stuff.
I more or less believe Model 2, though I recognize that it is imperfect and incomplete. But there is plenty of empirical evidence for it.
Today's Wall Street Journal has a terrific article demonstrating, beyond any real doubt, that Model 2 is right and Model 1 is wrong. If you believe that the job of a sell-side analyst is to tell people which stocks to buy and which ones to sell, you need to stop believing that right now, because it is not true. You shouldn't be embarrassed about getting this wrong; lots of people did. For instance, this guy thought that, and he actually was a sell-side analyst:
David Strasser, a former retail analyst at Janney Montgomery Scott LLC, says some investors told him they had little interest in his research and were only paying for meetings he could set up with companies.
“I wanted to be valued for my analytical abilities, but arranging meetings became such a critical part of the job,” says Mr. Strasser, adding that he was sometimes asked to sit outside the room so investors could ask questions without him.
But now he knows better ("he left the research industry to join a venture-capital firm"), and so do you. The evidence is overwhelming:
Many securities firms tally the number of times their analysts take company executives on the road to meet clients and use the number to help decide analysts’ annual bonuses.
At some firms, as much as one-third of analysts’ yearly pay can be tied to corporate access, says James Valentine, the founder of training and consulting firm AnalystSolutions LLC.
I mean, look. To be fair: That leaves another two-thirds. Analysts aren't just schedulers. They do analyze. They go to the meetings themselves -- usually! -- and listen to what the managers say and then try to find useful insights for their readers. They put together financial models, and project companies' earnings and stock prices. And yes, yes, yes, they put out Buy and Hold and -- rarely -- Sell ratings. And that stuff can be useful to investor clients. A clever insight from an analyst can give a client a good trade idea. An industry background piece from an analyst can help a client get up to speed on a new sector. An analyst's model can help the client think about how the company makes its money. There is a lot of potential value there, beyond just the scheduling.
But there is also just the dumb simple narrow question of: Which is more important, the corporate access, or the Buy/Sell/Hold recommendation? And on that question, there is just no doubt. Clients want the access, and are willing to sacrifice Buy/Sell/Hold accuracy to get it:
Analysts’ relationships with company executives, including the ability to line up private meetings for investor clients, have become an increasingly vital revenue source. And that is increasing the pressure for analysts to be bullish on the publicly traded companies they follow.
This makes sense: The investors are in the business of making investment decisions. They want information -- management meetings and analysts' research -- that will help them make those decisions. They don't particularly want someone else to make the investment decisions for them. If you run a big mutual fund, your job is to decide which stocks to buy. You might use a research analyst for background or insight or management meetings, but you're not going to buy because she says Buy.
And investors do have to sacrifice some Buy/Sell/Hold accuracy for the access, because some companies will explicitly refuse to do client meetings with research analysts who have Sell ratings on their stocks.
“It’s a decision I have to make on my sell-rated stocks: whether I will forgo the opportunity for corporate access, which clients will explicitly pay for,” says Laura Champine, a retail analyst at Roe Equity Research. Some previous bosses at other firms told her to “just drop coverage” instead of putting out sell ratings, she says, while declining to comment on where that happened.
Notice that the analysts don't really have to sacrifice anything else. They can still write insightful industry background pieces, and have thoughtful conversations with clients about a company's future, and summarize their takeaways from management conversations. They can even write negative reports, really. They just have to slap the word "Buy" on top. And so "just 6% of the roughly 11,000 recommendations on stocks in the S&P 500 index are sell or equivalent ratings."
Now, one potential reaction here is: This is terrible. The analysts are lying to people. Instead of telling investors their true feelings -- that they think investors should sell a company's stock -- they tell investors to hold, or even buy, that stock, just so they can get the investors a meeting with the company's management. The investors are deceived. They are losing money because of this conflict of interest.
But it isn't exactly a conflict of interest, is it? This isn't a story of analysts lying to investors in order to win investment banking business from big companies. This is a story of analysts (maybe) lying to investors because that's what the investors want. The investors want the corporate access. You know that because they tell the analysts that, and because they explicitly pay for it. The investors don't care so much about the Buy/Sell/Hold stuff. You know that because they tell the analysts that (sometimes), and because their demand for corporate access has led to more bullish research. If they cared more about ratings accuracy, you'd see less bullish ratings and less corporate access. But the reverse is true.
If the investors are asking to be lied to, then they probably aren't deceived, and we shouldn't feel worry about them. They are getting what they want, and pay for: access to corporate managers. The fact that they are also getting pieces of paper with "Buy" written on top of them is irrelevant.
But of course those aren't the investors that anyone is worried about. The worry is that small-time investors don't understand this system, and don't benefit from it. Big institutional investors get to meet with corporate management, which they like, because it is useful; they discount the Buy recommendations. Small-time retail investors -- and even small institutions -- don't get the benefit of corporate access, so they just assume that the research is about the Buy/Sell/Hold recommendations, and that they should buy all the stocks labeled Buy.
Analyst recommendations often carry weight with small investors, says John Bajkowski, president of the American Association of Individual Investors, a nonprofit group with 180,000 members. Most retail investors tend to lack sophisticated financial data and seldom dig through corporate filings, he says.
My own view is that buying individual stocks in your spare time based on research analysts' recommendations is an incredibly weird niche leisure activity and that there is no reason to organize any part of our capital markets around it, though I am aware that that is a minority view. But, yes, those retail investors may be deceived, if they think that a "Buy" recommendation represents the deepest convictions of a research analysts with profound insights into which stocks will go up, rather than a compromise to give that analyst's big clients what they really want, a meeting with corporate management. And it isn't just retail investors who think that. It's also the Securities and Exchange Commission, which ... doesn't so much think that as mandate it. Regulation AC requires analysts to certify "that the views expressed in the report accurately reflect his or her personal views," which is awkward if you are censoring your personal views to improve management access.
One simple solution here would be to just tell retail investors about Model 2, which solves the problem of them being deceived, though it doesn't really address the SEC problem. Really, while you're at it, you might want to tell retail investors that they are very unlikely to beat the market by trading individual stocks based on analysts' published recommendations, and that if that was their plan they should probably just index. It seems unlikely that the current research-analyst system works well for retail investors, but: What system would work well for them?
There is a lot of Wall Street stuff that looks sort of dumb and shady, and it is tempting to point at it and say "this is dumb and shady" and think that will make it go away. But I prefer to assume that most of it comes from some rational economic place, and to try to understand the rational economic explanation rather than write it all off as shady nonsense. People have been crowing for years about the fact that research analysts rarely give Sell ratings. It is a very well-known fact. If it bothered investors, they could easily do something about it. They could sit down with the analysts' bosses and say: Look, we need better ratings; we need at least a third of your ratings to be Sells, and if you can't do that, we are going to stop giving you commissions. But they didn't do that. Instead, they sat down with the analysts' bosses and said: Look, we need more meetings with corporate managers; your analysts need to get us those meetings, and if they can't do that, we are going to stop giving you commissions. The market demanded a product, and the banks responded by supplying it, and the market is happy with it. It's just that the product -- the actual value provided by research analysts -- is an access/analysis melange, not just a bunch of Buy and Sell recommendations. It's a little different from, and a little more complicated than, what regulators, and the press, and retail investors think it is.
I am fond of the analyst who wrote a report titled "Not All Is Good In Buckeye Land," slapped a Buy recommendation on it, and was fined by the Securities and Exchange Commission for not saying what he meant. He meant "Not All Is Good In Buckeye Land." He didn't mean "Buy."
That story happened! It led to the big analyst research settlement of 2003. People remain suspicious, but I am fairly confident that that conflict is mostly gone.
By way of allocating commissions to banks that provide them more access. "U.S. investors paid $2 billion in brokerage commissions for corporate access in 2016," reports the Journal, "or more than a third of all the money spent on stock research and related services, according to consulting firm Greenwich Associates."
Or, I don't know, maybe all companies are good now and so you should buy them all? Something like 84 percent of the S&P 500 stocks were up over the last 12 months, according to Bloomberg data, which is not that far off the 94 percent of ratings that are Buy or Hold.
This exaggerates. Analyst recommendations do matter. "Upgrades and downgrades by analysts often move stock prices," notes the Journal. But that is in part just because most ratings are so bullish, and Sells are so rare. If everyone has a Buy rating, and you move to a Sell, it's news.
More disturbing is that some small professional money managers also may rely too uncritically on sell-side research recommendations: "As much as we can screen the fundamentals of a company, those analysts are going to know far more than me and my colleagues," says a guy who runs "$800 million in investment trusts held mostly by retail investors."
When bond-rating firm Standard & Poor's paid $1.5 billion to the Justice Department to settle charges that it "engaged in a scheme to defraud investors by knowingly issuing inflated credit ratings for CDOs that misrepresented their creditworthiness and understated their risks," I pointed out that S&P had never really lost any market share despite the exposure of that scheme. Why is that? Shouldn't investors have stopped trusting it? My own guess is that investors -- many investors, anyway -- wanted inflated ratings on collateralized debt obligations. They wanted to own AAA-rated stuff, but they wanted it to have high yields. They weren't idiots, and they knew that higher yields came with higher risks. They didn't really think that AAA-rated CDO-squareds were as safe as Treasuries. But they wanted the AAA rating to show their regulators, or their board of directors, or whomever, and so they were happy that S&P helped supply it.
bloomberg.com |
