©2006 All content on site protected by copyright
Earnings: A Penny Here, A Penny There, by Victor Niederhoffer
You can always hit the earnings target, just sell more of the business. It's not something you can say, 'I can't do that'. That's a management decision of whether you want to do that. -- Jeffrey Skilling in his Enron testimony reported by Fortune .
The earnings numbers in the first quarter are always a full-of-information contest since it's the first report for the year, spending and pricing is sticky, and many aspects of financial life are on a budget. So far of the169 S&P 500 companies reported: 119 were positive surprises, 34 were on the mark or no surprises, and just 16 companies reported negative surprises. In terms of percentage surprise, the median positive surprise for the 119 that were positive was 7%, and the median negative surprise was -5% for the 16 that had a negative surprise.
Turning to the reported earnings versus last year's corresponding quarter, the average earning's increase was up 14% to 22% depending on the weighting. And of the 169 reported earnings: 139 were increases, 3 were unchanged , and 27 were below last year's figures. There are many aspects of the earnings reports that interest me, but I would like to focus on the companies that just missed or just beat the estimate. Given there are numerous ways that a company can adjust its earnings up or down, it takes a very honest company to report a one cent decrease relative to estimates. And the vast preponderance of companies that report an earnings increase of a penny or two, must be stretching on average. I hypothesize that the companies that beat by a penny or two are less trustworthy than those that missed by a penny or two. It is good to do this contemporaneously because there are too many adjustments and retrospections involved in looking back at what the estimates were at a given time and how they actually came out.
Here is a list of the positive surprises of two cents (18) and one cent ( 22): Two cent positive surprises: IMS Health, Broadcom, Leggett and Platt, Robert Half, Equifax, Gilead, Baxter, Costco, Regions Financial, Bed Bath, Danaher, Darden, National City, Lilly, Amgen, PNC Financial, Sensient and UnionBanCal. One cent positive surprises: NY Times, Oracle, Schwab, Texas Instruments, Qualcomm, Huntington, North Fork, Coca Cola, Stryker, Am South Banc, Bellsouth, Bank of New York, US Bancorp, Fifth Third, Marshall & Ilsley, Compass Banc, United Parcel, Johnson & Johnson, Torchmark, Auto Zone, North Fork Bank, and Parkvale Fin.
Here's a list of companies that missed their estimate by one cent: Dow Jones, Ford Motor, Motorola, Paychex, Hershey, Walgreen, Supervalu, United Health, Zions Bancorp, Concurrent Computer, Pacific State Bank.
Here's an approximate distribution of surprises:
Positive Surprises thru 4/21/06 of S&P 500 companies:
Surprise Number of Surprises 50 and up 05 25-49 04 10-24 12 05-9 31 04 10 03 16 02 18 01 22 0 05 negative 1 09 negative 2 0 neg 3 01 neg 4 02 neg 5 to 9 03 neg 10 or more 01
What can we say about a distribution like that. There seems to be many too many one cent and two cent positive surprises (39) relative to one and two cent negative surprises (9). Particularly striking is that there were not one two cent negative surprises, but 18 positive two cent surprises.
Most striking of all, is the accuracy of the Thomson earning's estimates that were used: 79 of 169 were within 4 cents of actual and 69 were within 3 cents of actual. (Differences in totals and enumerations occur because of the unchanged being fuzzy as to positive or negative surprises in various compilations) (These figures are based on pencil and envelope calculations based on Bloomberg reports) (Comparable figures for the % of companies reporting improved sales relative to the corresponding quarter last year and relative to estimate also show this 10 to 1 favorability ratio) ( A much better metric to measure all these numbers is the earnings change or the surprise in cents per share per dollar of price, or equivalently the total earnings change or total earnings surprise per dollar of price. However, as I must play as well as count, that will be left as an exercise for those who want to go beyond the norm.)
There are many interesting queries that strike one as one performs these tabulations. The most important question is what are the predictive qualities of these first quarter results relative to last year's actuals and this year's estimate, subsumed in these numbers. How has that relation changed over time, especially in the post Sarbanes Oxley era. What are the quantum numbers vis-a-vis surprise and actual versus last year that are good and bad for future performance? Are the reversals in actual or surprises a special class in themselves? What other special classes exist?