| |
 January 2005 Analysts or Reporters?
That doesn't mean there haven't been some changes in the investing world, just that they haven't been nearly as dramatic as expected. And now, after four years, we can draw some conclusions about Wall Street's sell-side analysts, too.
Up until late 2000, analysts enjoyed a cozy relationship with the firms they covered. Companies were feeling increased pressure to meet or exceed quarterly earnings estimates so to turn the tables in their favor, CFOs would provide "guidance" in periodic conference calls with select analysts. The next morning, the analysts would issue research reports reflecting this information, often moving the market. Not surprisingly, their estimates would usually prove to be quite accurate when earnings were actually reported. Markets, however, had become much more efficient. With real-time information available through broadcast media and most especially via the Internet, analysts and their customers had an unfair time advantage over the average investor. In an effort to level the playing field, the SEC enacted Regulation FD. It prohibits companies from disseminating any material information to analysts (or any limited group) before making it available to the general investing public. That put an end to the cozy guidance calls. The Actual ImpactSell-side firms warned that without this preferential information, analysts would be unable to fine-tune their estimates and market volatility would rise. We thought volatility would actually decrease if analysts widened their estimate range.
And guess what? It did. In the four years prior to Reg. FD, monthly volatility (as measured by standard deviation) for the Dow Industrials was 5.10%, but in the next four years it fell to 4.76%. Although not as dramatic, the S&P 500 also showed a decline (4.86% vs. 4.76%). So to the sell-side alarmists, volatility didn't increase. Even so, that doesn't mean standard deviations declined because of Reg. FD. Indeed, there's a much more likely explanation. Think about the general market conditions in the two measurement periods. The first, October 1996 - September 2000 includes the volatile bull market as well as the impact from the Asian crisis and the collapse of major hedge fund, Long Term Capital Management. The second period predominantly covers the subsequent bear market when stocks at first fell and then held near their lows. Arguably, volatility would have fallen with or without Reg. FD. Our prediction was correct, but for the wrong reason. But what about analysts' accuracy? How did it fare? To measure this, we turned to the First Call data available through Baseline. Again we compared data from the 15 quarters prior to Reg. FD to the following 15 quarters. This time we weighed analysts' consensus estimates against actual reported earnings. If Reg. FD was truly detrimental, the difference between estimates and reported results would grow and correlation between the two would decline.
Just looking at the quarterly differences, you don't notice much change. Chart 2 graphs the differences between analysts' estimates and the subsequently reported earnings. The black line divides pre-Reg. FD on the left from post-Reg. FD on the right. Just about the time it went into effect, estimate accuracy rolled off the table. In late 2000 and early 2001, the consensus was off more than 25%. It would suggest analysts had a difficult time adjusting to the post-Reg. FD world. But if you throw out that transition period, the quarters leading up to Reg. FD and those after 2001 don't look much different. Again it's tempting to conclude that after taking several quarters to get acclimated, analysts were able to adapt. Charting BehaviorBut is that really what happened? Sometimes the same data tell a different story when graphed differently. This is an example.If you look at the same estimates and actual reported earnings but this time graph them against one another, you get Chart 3. It uses the same data as Chart 2 but shows both series rather than just their difference. There's actually a discernable patter on Chart 3 although it's not a statistical or geometrical one. Instead, its a behavioral one. Ever since Harry Markowitz introduced the Capital Asset Pricing Model (CAPM), critics have (correctly) pointed out that investors don't always act rationally and aren't always risk averse. As human beings, investors are often motivated by factors other than mean-variance optimization. Analysts, being human beings, can also fall prey to these influences.
In the 1990s, academics began to offer "behavioral" market theories. One such observation is that when it comes to estimates, investors tend to think the current trend will continue. It doesn't matter if it's an uptrend or a downtrend, they just assume it will continue. As a result, they often miss turning points when predicting market cycles or share price movements. Isn't that exactly what's happening on Chart 3? In 1996 analysts underestimated earnings growth so increased estimates in 1997, only to see profits level off. They lowered estimates in 1999 only to see actual reported earnings take off again. When the tech bubble burst in 2000, analysts were caught off-guard as earnings quickly fell well below their projections. Again, it's more likely this was a result of what was happening in the market rather than the enactment of Reg. FD. To complete the picture, analysts quickly lowered estimates in the latter part of the bear market just as actual earnings regained their footing. By the time the market cycle troughed and earnings picked up -- especially in 2003 -- analysts were again underestimating their strength. (Although Chart 3 ends in September 2004, it's very likely that marked a new turning point and that estimates will again exceed actual earnings in the next few quarters.) When viewed in this context, the major divergence that occurs around the enactment of Reg. FD was more of a coincidence than a result of the new regulation. Despite the sell-side's warnings, it's still hard to see any definitive impact -- whether positive or negative -- from Reg. FD. Finally, what about the overall accuracy of the estimates? Back in the pre-Reg. FD days, analysts all got the same inside scoop from management, so all their estimates looked alike and not surprisingly, most were quite close to actual reported earnings. Post-Reg. FD, without such precise guidance, one might think the range of estimates would grow and accuracy would decline. To check that out, we broke differences between estimated and reported earnings down into 5% segments. This allowed us to graph their frequency both before and after Reg. FD went into effect. The results are shown on Charts 4 and 5.
The estimate accuracy is quite similar in both instances. Each has a few overly-optimistic outliers, but the pattern (illustrated by the red trendline on Charts 4 and 5) is almost identical. As you would hope, the mean difference for both periods is approximately zero. From the looks of this, Reg. FD didn't really affect estimate accuracy at all. Again as opposed to the sell-side's warnings, there have been few (if any) negative consequences. It's the Same InformationSo why didn't all the dire predictions come to pass? At the very least you would have expected to see at least some sort of impact from such a major change in corporate disclosure.Perhaps the best explanation is the fact that Reg. FD didn't really change the quality corporate disclosure, only its timing. On the face of it, this is obvious: Analysts no longer hear from management before the rank and file investing public -- but there's more to it than that. Back in the pre-Reg. FD days, analysts had the opportunity to ask about arcane accounting and financial details without the company's management having to worry about boring or misleading the public. Presumably, analysts then used this information to fine-tune their estimates and support their buy and sell recommendations. When management finally did speak to the public, they hit the high points without having to go into the same depth as with the analysts. In the post-Reg. FD world, there aren't any cozy conference calls with analysts. Instead, companies make brief announcements or issue precisely worded press releases. Everyone gets the same guidance at at the same time, yet analysts' estimates haven't suffered. Why would they when all analysts have ever done is simply report what companies have told them? That's right, they haven't really been analyzing balance sheets or business plans, they've just been parroting back what CFOs have been feeding them. The only thing Reg. FD changed was when they get the information. Prior to its passage they got it before the investing public, now they get it at the same time. That's why downgrades come after firms lower guidance, not before. That's why earnings surprises are surprises: They weren't foretold by the company so analysts aren't expecting them.
It's hard to give up the image analysts burning the midnight oil toiling over their complex valuation models. Indeed, that's the impression the sell-side wants you to have, but all this suggests it's just not true.
We suspected all along -- even before Reg. FD or the revelation of analysts' ties to their firms' investment banking -- that analysts were more reporters than, well, analysts. Looking back over the past eight years, the data support this conclusion.
Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials: |
||||||||||||||||||||||||||||||
T'S BEEN FOUR YEARS NOW since the SEC's Fair Disclosure regulation went into effect. Contrary to the dire warnings from Wall Street's sell-side, the world is still turning on its axis.
