posted 04-19-2000 05:42 PM         

A couple of thoughts:
How many analysts out there are both "right" and able to please companies? It seems to me that people that honestly try to be "right" will lose business for their banks because, at some point, they will have to be negative about a stock. At that point their ability to move markets is actually dangerous to a company--they will not want to give their business to an "honest" analyst because they have the power to hurt the stock. But, once the analyst tries to please the company, he will no longer be "right." I guess Tom's argument is at its strongest right here: the analysts who inflate their ratings artificially will lose their value to companies because they will no longer move markets after people know they are wrong most of the time. The dirty little game should end.

My other concern is about self-fulfilling prophecies. At some point, when an analyst's reputation has been built up, isn't it possible that they will end up being "right" almost all of the time? People will take for granted that what they say is pretty on the money, and the price will move in accordance. Pretty soon, everyone will make the same predictions as those analysts, and everyone will be "right" most of the time. Then, the analysts with the "good" reputations will again be susceptible to the pressures of business. Hmmm.

Now, if analysts were focused solely on predicting earnings and growth, that would be one thing, because their statements can't effect the data. But this is like a lousy science experiment where the variables aren't controlled properly-->inconclusive results.

posted 04-19-2000 02:14 PM         

quote:

---

Originally posted by hecubus:
I get the feeling that the Analyst game is almost too far gone to be saved by accountability, though. At least from your presentation of the matter, they have already sold out and are not doing much besides drumming up business for their I-banks. If the people they are appealing to are the companies themselves, and not to investors, will analysts care so much about bad "ratings?" It seems to me that as long as they keep pleasing the companies, it won't matter to them all that much.

---

posted 04-19-2000 11:43 AM         

The internet has increased the need for stock picking accountability, for two important reasons. First, the internet has created opportunities for commentators of all shapes and sizes to voice their opinions. With the increase in the number of stock pickers comes an increase in the importance of being able to differentiate between the good and the bad. Second, the internet is contributing to the popularization of stock picking. Conversations at your coffee shop are just as likely to be about favorite stocks as favorite sports teams, a very recent development that the internet has accelerated by enabling individuals to do their own stock research. With an increase in the number of people looking for stock recommendations comes an increase in the importance of being able to differentiate between the good and the bad.

Everyone's rating stocks, but who's rating the raters? Until recently, the answer has been: virtually no one. Fortunately, while the internet is making stock picking accountability more important, it's also making it more feasible. A new breed of upstarts are using the internet to bring order to the chaotic world of stock recommendations. In this three-part essay, I will review each major category of stock picker, summarize that category's current level of accountability, its performance, and why that performance is suboptimal. I will also describe the web sites that are working to bring accountability to the area and how much success I expect them to have.

- - - Analysts - - -

Current Level of Accountability:

Low. Who are the best analysts? You probably don't know, and neither do I. Magazines and newspapers occasionally publish rankings of the best analysts, but the lists are usually based more on reputation (deserved or undeserved) than on results. There is some accountability, in that you can record what a given analyst said and check back later to see whether they were right. Unfortunately, analysts don't make predictions that are specific enough to examine later for accuracy. Every brokerage has its own buy/sell/hold terminology, comprising in total more than 40 different terms (if you don't believe me, check out http://biz.yahoo.com/f/bc.html#ratings.) This is done primarily in order to make comparisons between brokerages very difficult. What's worse, the terms don't mean what they appear to mean: "hold" sounds like a positive term, but is actually quite negative. To their credit, many brokerages include price targets in their recommendations, but they rarely specify the date on which they expect the target to be met. In an extended bull market such as they one we've been experiencing (at least until last week), it's very easy to set a price target 15-20% above the current price and just wait for the stock to creep up to meet the target. Although a few brokers make explicit performance predictions for each of their ratings (for example, Wit Capital's are at http://www.witcapital.com/research/undrstnd_ratings.jsp), these are in the minority.

Performance:

Due to the ambiguity of analyst predictions described above, I can't quantify aggregate analyst performance, but I can cite a few recent examples:
- None of the 38 analysts covering Lucent notified their clients that the company's inventories and receivables were skyrocketing prior to the stock's 35% drop in January.
- B2B portal VerticalNet hit an all-time high of 148 on March 7. That day, JC Bradford initiated coverage with a buy rating. The stock has lost three quarters of its value since then and now trades in the high 30s.
- Three analysts maintained "buy" ratings on Iridium even as it lost 95% of its share value and went into bankruptcy.
- Rite Aid, amid yet another reorganization, a cash crunch, and accusations of racketeering from the State of Florida, was rated by 11 out of 12 covering analysts as "hold" or better; the remaining analyst rated it a "moderate sell".
- DrKoop.com has fallen more than 90% and is rapidly running out of money, but three of the five analysts who cover the stock have recommended it throughout its descent.
Please note that this list is by no means exhaustive. I could have made it much longer, but I think these five representative examples will suffice.

Analyst recommendations usually arrive too late to be useful. When bad news is revealed, analysts downgrade the stock. Good news, analysts upgrade. But the stock has already moved. Nevertheless, since accountability is low, this technique of recommending works to the analyst's advantage, since anyone who later checks what the analyst says about various stocks will see that the analyst has said good things about stocks which are doing well and bad things about stocks which aren't. Without exact tracking, it's hard to separate cause and effect.

To those of you who use a full-service broker: have they ever given you at the end of the year a graphical report of when their analysts made recommendations and revisions and what happened to those stocks afterwards? If you're using a full-service broker instead of an online broker, it's probably because you feel the information and recommendations they give you justify the extra cost, but you deserve to know exactly how much those recommendations are worth to you. It would be extremely easy for them to provide this information, but doing so would probably show you that they generally downgrade only after the bad news has already been factored into the stock price, and they hop on board the hot stocks only after they've already experienced significant run-ups. In fact, some smart investors actually use analyst revisions as a contrary or a lagging indicator. If you do ask your full-service broker for such a report, request that they calculate the performance starting the day after (and not the day of) the recommendation, since the recommendation itself often moves the stock price.

In addition to making stock recommendations, analysts also enjoy predicting upcoming earnings numbers. But they have proven to be so consistently inaccurate that investors and commentators have started paying less attention to them, and more attention to 'whisper numbers', available at sites such as EarningsWhispers.com and WhisperNumber.com. In fact, if you watch closely, you'll find that when a company announces earnings, the stock price tends to move based on how the actual numbers differ from the whisper numbers, rather than how they differ from the consensus analyst numbers.

Explanation for Performance:

Why is their performance worse than you might expect? Some of it could be blamed on a lack of skill, but the more reasonable explanation is conflicts of interest. Analysts are the primary way that a public company draws attention to its stock. Positive attention helps the stock go up, negative attention drives it down. So analysts have something of value that companies want: power. Similarly, companies have something that analysts want: underwriting business. Secondary stock offerings and bond offerings generate big bucks for underwriters, so there is a compelling opportunity for mutual back-scratching, at the expense of the investor who naively believes that the analyst is an objective source of information. Additionally, analysts who don't give a positive rating to a company will often find it difficult for them to get information from that company and to get in on its conference calls. Jeffrey Hooke, author of Security Analysis on Wall Street, says: "If an analyst gives a negative report on a company, he might be more candid than others, but he'll get cut off from information. If he did that to multiple companies, he'd be unemployed." The result is that "buy" ratings are common and "sell" ratings are rare. In fact, buy ratings became so common that investment banks started using "strong buy" to try to differentiate the real winners, and now "buy" actually sometimes has a negative connotation. And among the top Wall Street firms, less than 1% of all analyst recommendations are "sell" or its equivalent. The exceptions to this conflict of interest problem are the analysts who work for independent research companies, like Soundview Technology Group... oh wait, they were just acquired by Wit Capital. From Jeffrey Hooke: "Any analyst whose firm does major investment banking work--and nearly all of them do--is suspect. I don't know why the SEC doesn't ask these firms to spin off their research operations."

As described in the section above, analysts' earnings estimates tend to be predictably inaccurate. Why is this so? Most analysts have a herd mentality. No analyst wants to be the only one who was wrong, but if all analysts are wrong in the same way it won't seem like their fault; rather, the essential information causing the discrepancy simply must not have been available. If all analysts are wrong in the same way, they won't all lose their jobs. If one analyst goes out on a limb with a bold prediction that turns out to be wrong, his/her career could be jeopardized. So they all watch what numbers each other publishes and they try to stay in the middle of the pack. Because of this, it takes time for the herd to move in response to new information, and so the average expectations lag behind the latest news. Additionally, analysts often publish lowball earnings estimates that they are confident the company will have no trouble exceeding (so that when the company does report better than 'expected' numbers, the stock can get a boost). The rationale is the same as for giving a strong buy recommendation for a less-than-stellar company: the analyst is providing the company with something of value, in the hope of future reciprocation.

I should point out that as with most rules, there are exceptions. Some analysts do provide valuable recommendations and good earnings estimates, and some analysts do not cave in to conflicts of interest. If you rely on an analyst for advice and/or stock picks and are confident that you're getting your money's worth, then by all means I encourage you to stick with them. The point I'm trying to make is that without a sufficient level of accountability it's very difficult and time-consuming to determine which analysts are earning their keep.

Sites Working to Improve Things:

There are a few sites working to bring accountability to analysts' recommendations. The most notable of them is BigTipper.com. This site aggregates analysts' price targets and tip sheets, and then reveals their track records. Choose from among the hundreds of analysts and see a list of stocks they recommended, when the recommendation was made, and what the stock has done since then. This isn't completely scientific: it weights all picks equally rather than giving extra weight to the analyst's top picks, and it doesn't enable analysts to later sell positions when the outlook changes. However, these enhancements would require additional input from the analysts, so the current system is probably the best that can be done under the circumstances.

Additionally, several sites and magazines occasionally invite a group of analysts to participate in a stock picking competition. The best example is MSN MoneyCentral's Strategy Lab (http://moneycentral.msn.com/articles/invest/derby/derbsumm.asp?primer). Each analyst explains his/her strategy and manages a mock portfolio in accordance with that strategy. Transactions are recorded along with explanations for the trades, and performance is tracked to see how well each analyst is doing. I applaud these analysts for having the confidence to put their reputations on the line, and hope others will follow their lead.

Also, some magazines and newspapers are adding to analyst accountability, by mentioning whether the analyst's firm does underwriting business for any of the companies mentioned in the article. This implicit acknowledgement of a potential conflict of interest alerts readers to take what the analyst says with a grain of salt. The practice isn't as common as it should be, but hopefully it will be soon.

I expect that BigTipper.com will continue to grow in popularity, and that other sites will soon offer similar information, enabling more individual investors to gain access to analyst performance data. The benefits will be twofold: analysts will be more careful and less likely to be influenced by conflicts of interest, knowing that their recommendations are going on their 'permanent record' for all to see; and individual investors will be better able to differentiate between high-quality and low-quality analysts, and will be able to immediately know which ones are worth listening to.

Coming Soon:

Part two of this essay will be posted next Wednesday, and will look at two more groups of stock pickers: mutual fund managers and columnists. We'll review their performance, provide reasons why it's worse than you may have thought, and describe some sites that are trying to bring increased accountability to these groups. The third and final part of this essay will be posted the following Wednesday, and will look at individual stock pickers, and a new breed of sites that are hoping to prove that some individual investors can outperform the so-called experts.
_____________________________________________________________________