Daily Speculations The Web Site of Victor Niederhoffer and Laurel Kenner


The Chairman
Victor Niederhoffer



2005 All content on site protected by copyright

Write to us at: (address is not clickable)

An Earnings Analysis

The following table is from a daily summary of earnings reports of S&P 500 companies that Bloomberg publishes each day.

POSITIVE SURPRISES:    218/331 = 65.9%
% of Surprises:                        EPS Differences (Actual-Estimate):
     (0% to 10%):     154             (1 cent):                     58
    (10% to 20%):      33             (2 cents):                    31
    (20% to 30%):      11             (3 to 4 cents):               52
    (30% to 40%):      11             (5 to 9 cents):               36
    (40% to 50%):       3             (10 to 19 cents):             25
0% SURPRISES:           44/331 = 13.3%

NEGATIVE SURPRISES:     69/331 = 20.8%
% of Surprises:                        EPS Differences (Actual-Estimate):
      (-10% to 0%):    47             (-1 cent):                    15
    (-20% to -10%):    12             (-2 cents):                   14
    (-30% to -20%):     4             (-3 to -4 cents):             10
    (-40% to -30%):     1             (-5 to -9 cents):             13
    (-50% to -40%):     0             (-10 to -19 cents):            8
  (less than -50%):     5             (-20 to -49 cents):            5
        % chg n/a       0             (-50 cents or less):           4

Source: Bloomberg LP

One of the amazing things each quarter is to see the market go down overnight when a company misses its earnings estimates or gives a warning, and every commentator says, "Stocks down on earnings shortfall in this one or that one." It's as if the public really expects a 1,000% batting average for companies. Yet earnings of the companies in the S&P 500 are up about 18% this quarter, 80% showed increases, and two-thirds beat estimates. Eight-nine beat estimates by 1 or 2 cents a share, and 29 were below estimates by 1 or 2 cents.

I do admire the ability of companies to eke out those accruals on the positive side versus their inability to do so on the negative side to an inordinate degree. Along these lines note that for companies reporting more than 10 cents above estimates, about 1.5 times as many beat the estimate as fell below (25 who beat versus 17 that fell). But among the companies that just beat or just missed earnings by 1 to 4 cents, the ratio of companies beating to losing was 3.6 to 1 (141 beating versus 39 that missed).

I can only express amazement at such a nonrandom and asymmetric distribution. Along these lines, various articles have come down the pike that show earnings statements this year are very much better than ever before. Actual GAAP earnings are apparently only 15% worse than pro forma earnings this year, whereas in previous years the GAAP earnings have been about 40 % worse. And yet, the animal spirits, the ability to guide propitiously and to squeeze would not seem to be dead. What kind of world would it be if it were dead.

Perhaps, in view of everything, our readers will excuse me for doing something unmentionable and truly uncharacteristic. What I'd like to do is to pat myself on the back. You see, about 35 years ago, in an article with Pat Regan, I analyzed the influence of earnings surprises and earnings changes on stock prices. The results are on p. 272 of my 1997 book, "The Education of a Speculator". The key step that cleared the way for everything else was to measure the movements in earnings as earnings change per dollar of price rather than percentage earnings change. For example, if a company is at $10 and it increases its earnings to $1.10 from $1.00 a share, that's a 1-cent increase per dollar of price. This paves the way for actually using earnings declines and earnings changes from a base close to plus or minus zero, instead of throwing out these observations as do 95% of the retrospective studies of the value practitioners and the academics of this persuasion. It also opens up some very meaningful regressions that separate out the influence of actual magnitude of change and surprise on future returns.

Estimated EPS for the S&P 500 for 2005 is $77 a share, which corresponds to a 6.5% yield. With a 6% increase predicted for 2006, that would make a 6.85% estimated earnings yield for next year, compared with a 4.54% yield on the 10-year Treasury note. That 2.31% differential will and has figured more prominently in the decisions of investors than the individual woes or triumphs of this or that shooting star. A detailed analysis of the predictive properties of earnings yields versus 10-year yields is contained in the joint continued work of Vic, Collab and the artful simulator Tom Downing.

I wish to praise Bloomberg for providing this sophisticated, analytical and practical summary. The founder, now a politician, once complimented me for continuing to be a customer despite my grievances against certain parties there, and the company's ability to come up with a detailed Baedeker on things like earnings, corporate actions and returns of the individual aspects of indexes (with retrospection) between any two dates prevents me from cutting off my nose to spite my face. One sometimes must live to fight another day.

More by Victor Niederhoffer