Friday, October 26, 2012

Corporate Quarterly Earnings Report's Negative Effect on Wall Street

In the last week, the Dow has suffered a 202 and a 240 point single day drop on 10/19 and 10/23 respectively.  The media coverage of these events headlined the weak Corporate Earnings as a cause of the poor performance in the Stock Market.  This rekindled my long held belief that any of these explanations do NOT have any statistical relevance to the real business results.

To check this out, I read several business sites like CNBC and Morning Star, and recorded the companies that were mentioned as explanations of these two drops in the Dow.  About 80% of the ones mentioned are also in my database of companies that I have been tracking since 1993.  So I updated results with the quarterly results announced in October for these companies to understand if any the JAS 2012 quarter were statistically relevant in comparison to previously reported quarters.

There were 15 companies in my database that were also reported in the media as contributing to drop in the Dow.  I track both Revenue and Basic Earnings per Share (EPS) for these companies which gives me 30 possible areas of concern in Corporate performance.  After analyzing all 30 areas of performance with Control Charts, only 6 of these areas showed a statistical change in the last quarter which might have negative impacts on the Dow.  However, there were also 6 which showed a positive statistical change which should have "helped" the Dow.  But key for me is that 18 of the 30 results (60%) showed NO STATISTICAL DIFFERENCE from past performance and 5 companies showed no change in either Revenue or EPS!

For example, here are a several graphs from these studies.  First, lets look at IBM which has not had a statistically important change in either Revenue or EPS over the last 11 years, including the JAS 2012 quarter!  First, the graph of Revenue:

As you can see the last red result is slightly above the average performance of 2.2% annual growth (green line), but not anything statistically different or outside the normal limits (blue and yellow).

Again, the last red result for JAS 2012 is following the previous 11 year pattern.  This result is approaching the Upper Control Limit, UCL (blue) but as the absolute numbers continue the increase, so should the RSD or Relative Standard Deviation which will increase the width of the control limits.

So what shakes up Wall Street with these results??  You will read comments that express disappointment with these JAS results that "miss" the previous forecast estimate of performance either from "analysts" or the company's previous quarter report.  IF YOUR PAST 11 YEAR PERFORMANCE HAS BEEN THIS STEADY, HOW COULD YOU MISS A FORECAST?  The answer is simple: for the past 11 years, these analysts and executives have made their success by explaining every up and down in these results with "precise" singular causes.  These may range from new marketing programs, new products, acquisitions, economic conditions, material sourcing, organization changes and the like.  The problem is that when there are only normal (common) causes effecting results, the above chart is statistically stable, and has been for the last 11 years.  The "common causes" are a complex set of activities that influence results, and randomly interact in a way to create growth and variation that stay within these statistically calculated limits.  For example, below are some quotes out of the IBM most recent quarterly report that reflect this erroneous explanation:

Third-quarter net income was $3.8 billion, flat year-to-year; or $3.9 billion, up 3 percent excluding the impact of UK pension-related charges. Operating (non-GAAP) net income was $4.2 billion compared with $4.0 billion in the third quarter of 2011, an increase of 5 percent.

Total revenues for the third quarter of 2012 of $24.7 billion were down 5 percent (down 2 percent, adjusting for currency) from the third quarter of 2011. Currency negatively impacted revenue growth by nearly $1 billion.

“In the third quarter, we continued to drive margin, profit and earnings growth through our focus on higher-value businesses, strategic growth initiatives and productivity,” said Ginni Rometty, IBM chairman, president and chief executive officer.

You will first notice the use of Index versus Year Ago which shows up as a percent increase or decrease versus the same quarter in 2011.  Interesting but useless.  These changes are just chance (common cause related) and therefore cannot be explained by a single event or project.  Said another way the "currency adjustment", "initiatives" and "productivity explanations" could rightfully be used in any quarter!  However, since they have been explaining every up or down for years, their institutional memory would suggest to them that if they are working on a similar style project in the future, that they should be expecting another 5% increase in the future.  Problem is, when the future comes, the common cause system is just as likely to cause a 3% decline which in turn produces the forecast "miss" that creates the dip in the Dow.  WOW, what a huge, non-productive routine that does nothing more create more buying/selling, increasing the variation in the stock market which in turn creates winners and losers even though nothing has really changed!!  The quarterly report should have read: "Nothing has changed and IBM continues to reliably deliver a 2.2% compound growth in Revenue which in turn is generating an 18.1% growth in EPS."  These reports should go into detailed explanations only when there is a statistically relevant change and in my database of 85 companies, these changes only occur once every 7 years on average!

Here is another example of the "blamed" companies, Amazon.  First Revenue.

As you can see, consistent 30% growth for around 9 years with only variation increasing as the Relative Standard Deviation of the actuals increases over time.  No statistically relevant changes.  Now look at EPS.

Here you an see that EPS has shown a statistical change from the historical 29.1% growth.  However, it is clear that this change did not just occur in JAS 2012, but rather OND of 2011.  Why didn't the sky start falling a year ago??  Technically, nothing in the last 3 quarters has changed since a Special Cause in OND 2011.  Rest assured that Amazon has declared what terrible things have happened to them effecting each one of the last 4 quarters rather than just the one thing that happened in OND 2011.

Finally, an example of a company whose results really should have effected the markets with statistically relevant changes in the JAS 2012 quarter, UPS.

As you can see, the JAS 2012 quarter has two results in a row closer to the LCL (yellow) than to the Average (green) which is an indication that something has statistically changed outside the common causes.  This needs an explanation.

The JAS 2012 quarter is uniquely low compared to the last 2 years performance and is of concern.  This does have a unique cause which is explained in the quarterly report as seen below:

On a reported basis, third quarter 2012 earnings per share were $0.48. In August, the company announced a decision to restructure pension liabilities for certain employees. As a result, UPS recorded an after-tax, non-cash charge of $559 million during the quarter.

How could the entire quarterly reporting process be improved for all companies in such a way as to reduce unnecessary gyrations in stock market reporting and forecasting?  Simply report only when there is a special cause in the results, which will be about once every 7 years!  Using the statistically relevant explanation, a company (or analyst) would then update their forecast and continue to use this same forecast every quarter until the next, rare special cause comes along.  But alas, this would put a significant number of analysts and executives out of work and also reduce the number of winners/losers in the stock market game, which the "winners" will surely resist!