In its 13-F SEC filing for the quarter ended December 31, Berkshire Hathaway revealed that it had sold its entire $2.1B holding in Oracle. The position was first established only one quarter ago, and while we have no specific insight into the reason for Berkshire to buy or sell Oracle stock, we looked through the lens of the Amenity Analytics text analytics platform for insights from each of the last two quarters. 

We first saw warning signs on the September call, while the December earnings call was more benign. However, one emerging hot button issue may have been the culprit.

NLP Analysis of Q&A on September Earnings Call First Raised Warning Signs

The Amenity Analytics NLP model has the ability to focus analysis on the most critical element of earnings calls: the Q&A section. One of just a handful of times per year when CEOs and CFOs of public companies have to go off-script to take questions from analysts. We applied our standard Fundamental Analysis model to analyze Tone in the Q&A section as well as our unique Deceptive Language analysis, which identifies potential problem areas that management is trying to dodge.

Berkshire may have seen something in the 9/17 earnings call to trigger the $2 billion purchase, but the NLP model flagged two warning signs from the Q&A:

1. Deteriorating sentiment: the Q&A Section scored a value of 2 compared to 13 in the prior quarter (on a -100 to +100 scale)
2. Significant increase in Deception Score: 52 vs 22 in the prior quarter; the higher score indicated higher prevalence of evasive and vague commentary

A Deeper Dive Into What Drove the Caution in September

Insights in the Amenity Analytics platform always trace back to the Why behind the numbers, and ultimately to the specific text highlighted by the NLP model. We found the spike in Deception Score on the 9/17 earnings call was driven by an increase in comments characterized as Evasive by our model, where questions seeking additional supporting data on the Cloud business were met with little to no detail.

Oracle (9/17/18):

• "I just don’t know what words I can use to get – to show more confidence in the apps business than what I’ve done."
• "In the Fusion business, I don’t know what more I can do to read you this list of customers."
• "I don’t know what’s in everybody’s model, but relative to our view, we did better in that line and we did based on our forecast at the beginning of the quarter."

Did Something Spook Berkshire in December?

The above chart showed little change in the Q&A tone and a significant decrease in the Deception Score from the September to December earnings calls.  Although the headline analysis was benign, when we dove into the hot button issues of the Q&A section, the database business and competitive threats in the Cloud (Amazon) did garner additional attention compared to prior quarters. After its painful IBM experience re: Amazon, was that enough to spook Berkshire?

Oracle (12/17/18):

• "Now the beauty of what Amazon did is they put them in Amazon Cloud and they made them (databases) available on the cloud. They did that long before we made the Oracle database available on the cloud. But in terms of technology, there is no way that someone can move, a normal person would move from an Oracle database to an Amazon database."
• "So we think we’re – as I said in my opening remarks, we think we’re not only going to hold onto our 50% share, we’re going to expand it.  Nobody, save maybe – Jeff Bezos gave the command, “I want to get off the Oracle database,” and they’ve been working on this for a few years to try to get off the Oracle database and get onto the Amazon databases. It’s taken Amazon, who’s dedicated to doing this several years and they’re not there yet."

Request access to Amenity Key Drivers, the NLP language modeling and scoring system. Analyze earnings call transcripts based on key fundamental factors that drive stock performance and uncover areas of risk, exposure, and opportunity hidden in financial documents.

‍

This communication does not represent investment advice. Transcript text provided by S&P Global Market Intelligence.

Copyright ©2019 Amenity Analytics.