The Trouble with Quarter Earnings Releases in Business Reporting

Cause and effect is a very misused concept since we desparately want to explain everything that happens to us, both the good and the bad.  Cancer, "accidents", mental health, birth defects, happy marriages, successful people, wealth, intelligence all seek to find the silver bullet, so something can be cured or maximized.  But alas, randomness and chance are with us everyday, clouding our understanding of everyday events.  Blame, rather than causation, are more easily diagnosed with our belief systems rather than with data and understanding.

Although there are many examples of misunderstanding causation, I want to use business performance as represented by financial results and the related press releases which follow in this blog post.  First, I would like to review some key concepts I have used before:

1. Variation.  It exists all around us all the time: Temperature, mood, blood pressure, drive time to work, rate of return.  As Edwards Deming said, "measure it once and you know; measure it twice and you don't know".
2. All the ups and downs in results or performance are caused by only two things.  The first is Common Causes (Systemic) which are present all the time.  They are made up of all kinds of simple things which interact in random ways to influence a result or performance in predictable and measurable ways.  Take your drive time to work each day and think about all the things that influence the total drive time.  A partial list might include what time you leave, how anxious you are, stop light patterns, weather, other drivers, your car's performance, when you last fueled, stop signs, commercial truck traffic, school bus activity, pedestrians, how fast you are diving, and on and on.  The other influence is called a Special Cause (Assignable Cause) which generally is singular and rare.  These events will create a result beyond what is expected from the common or normal causes alone.  These are sometimes called Breakthroughs (when helpful) or Breakdowns (when hurtful).
3. Understanding these two sources of variation are critical to making progress and sustaining any benefits in results.  Having a flat tire on your way to work is likely a special cause in the resulting drive time.  However, a "run flat" designed tire was created with the knowledge that flats do happen and the effect, which would be a special cause, can be minimized.
4. Most people want to either place blame for a negative result or take credit for a positive one, when in fact it was caused by the Common Causes only (complex interaction of many causes). This is how cause and effect relationships are erroneously concluded.
5. There is a simple way to establish the boundaries of variation when only Common Causes are present.  This uses only two calculated numbers: the average and the standard deviation.  Over a period of time, say you record the actual drive time to work each day.  Calculate the average and standard deviation and place two boarders around the average of plus/minus 3*SD.  These are called control limits and represents the spread of the  normal variation in drive times to work.  One day, you decide to leave for work 5 minutes earlier and arrive at work 8 minutes early.  If this is NOT outside the Lower Control Limit of drive times, you should not conclude leaving 5 minutes earlier will  shorten the drive time average.  Another example is a thermostat.  If you are in a large room with many people and everyone has been told to adjust the thermostat to their liking, you will find that the average room temperature will be the same over time, but the variation will increase significantly, making everyone unhappy with the room temperature.  This is considered "tampering" which makes the variation higher without changing the average.  
6. This last point happens everyday when companies (and governments for that matter) report financial results.  There is a tendency to find blame when the results are down or a successful project when the results are up.  In the future, they will try to prevent the actions that were blamed or fund more similar projects that were successful.  HOWEVER, if these originally were NOT special causes, they are now tampering, making variation higher without any change in overall performance.  Business leaders do benefit from this higher variation since it will increase the probability of hitting a higher goal in the future without making any positive changes.  The stock market also thrives on this variation as the buy low, sell high creates higher gains (in the short term) which gives rise to high frequency trading.

Now, take a look at a few examples. I will start with AIG.  The revenue has been in steady decline (only common causes), since June of 2011.  Said another way, they have NOT had a significant change in revenue in the last 21 quarters.

However, from the quarterly earnings release for June of 2016 (the last quarter in this graph) I quote:

“AIG’s second quarter results show strong improvement towards all the goals the Board and I announced in January,"

The word Improvement, to me, indicates there has been a change in the performance for which there is no evidence in this case unless the goal is to reduce revenue!  Also interesting to note that the quarterly earnings release only deals with Net Income in their discussions, not revenue.

Campbell Soup gives us another good example with very stable revenue.

From the July, 2016 Earnings Release comes this quote: 

"Sales of $1.687 billion were comparable to prior year as the benefit from the acquisition ofGarden Fresh Gourmet was offset by the decline in organic sales and the adverse impact of currency translation. Organic sales decreased 1 percent primarily driven by Campbell Fresh,reflecting declines in carrots and carrot ingredients, as well as the impact from the voluntary recall announced on June 22 of Bolthouse Farms Protein PLUS drinks. The estimated negative impact on net sales in the fourth quarter related to the recall and related production outages was approximately one percentage point."  

Looking at the graph, all the items mentioned above are singled out common causes which do not explain the July result.  What explains the July result is the complete set of numerous and complex common causes.

As for the Campbell Soup Earnings per Basic Share graph, you will see that the most recent July, 2016 quarter is clearly a special cause and needs an explanation.

This is their response: 

"As reported EBIT was a loss of $37 million, reflecting the non-cash impairment charge, pension and post-retirement mark-to-market losses and charges associated with cost savings initiatives as previously mentioned. Excluding items impacting comparability, adjusted EBIT decreased 2 percent to $253 million reflecting higher advertising and consumer promotion expenses and a lower adjusted gross margin percentage, partly offset by lower administrative expenses".  

The first part of the first sentence is pretty logical as a special cause, but all the remaining words sound very much like list just a few of the common causes.

Using the SOME common causes to explain a special cause will contaminate the institutional memory as these couple of common causes might be used again to create a needed breakthrough in the future, which will NOT materialize.  This would then begin a hunt for the "guilty" for why this project did not work.

Coca Cola is another example of trying to explain something that does not exist.  Both the Revenue and EPS graphs are shown below.

Notice that both graph are stable with only common causes present for the last 5.5 years.  Here is what they are saying in their earnings release for September 2016: 

"I am pleased to report that we delivered results in line with our expectations,"; "While our year-to-date reported net revenues declined 5%, our core business organic revenues* have grown 4% despite continued global economic and political volatility."; and "Full Year Organic Revenue and Comparable EPS Outlook (Both Non-GAAP) Remain Unchanged".

So declining sales and EPS are your expectations!!  Also, look over here at our business unit that is doing well because of these two common causes!?  Finally, saying your outlook remains unchanged means that you do not have any breakthroughs in your business strategy that would turn around your declining results.  Speaking of "outlook", remember if your results have been stable for many years, forecasting future results is a statistical piece of cake.  And these forecasts are what the stock analysts use to establish stock price recommendations.  If you miss your forecast, your stock price takes a hit.  Having stable business results, and therefore, good forecasts, is the best way to promote higher stock prices.  Knowing what causes real breakthrough in results is what can drive reliably raising future forecasts.

Delta Airlines revenue is an interesting case.  Notice that the 7 years, ending September 2016, were show a  stable trend of 3.3% growth.  

But look at what they talked about in their release on September 2016:

"Delta’s operating revenue for the September quarter decreased 5.6 percent, or $624 million, of which $100 million was due to the [computer] outage and $70 million was from prior year Yen hedge gains." 

Having a computer outage that reduced $100 million in revenue IS NOT A SPECIAL CAUSE like they think it is; nor is the Yen hedge.  Both are just among the normal things that affect their quarterly results, within the bounds of the control limits (blue and yellow lines above).  The more the business leaders use isolated common causes to explain their results, the more they will be "tampering" which makes the quarterly variation higher without changing the average result!  

You can see that I am only on the "D" companies in my list that I regularly track.  I am sure you can imagine there are many more examples of of explaining results of a stable business with one or two common causes or totally missing a special cause that needs explaining.  THE ROOT CAUSE OF ALL THESE ERRORS IS USING "INDEX VERSUS YEAR AGO" WHEN REPORTING AND ANALYZING FINANCIAL RESULTS.  In most cases, a few points percentage change, up or down, is due only to the complex set of common causes and needs understanding but not explanation.