Every year in the six-week stretch from mid-February through the end of March, the real earnings season, we parse and analyze 10-Ks for companies with 12/31 and 1/31 fiscal year ends. Featured by Bloomberg and Harvard Business School in “Disrupting Fundamental Analysis with Robo-Analysts”, our research automation technology uses machine learning and NLP to automate and improve financial data collection and modeling.
The Power of the Robo-Analyst
While other firms have been slow to augment human operations with technology, we are collecting data faster than ever - much faster than traditional providers, especially given the disruption to their offshore data collection operations due to COVID-19 .
From February 17 through March 31, our Robo-Analyst parsed and analyzed 2,522 10-Ks and 10-Qs and collected over 254,000 data points. Our analyst team uses this data to make 40,617 core earnings and balance sheet adjustments with a dollar value of $19 trillion.
Figure 1: Filing Season 2020 – Power of the Robo-Analyst
Sources: New Constructs, LLC and company filings.
The adjustments, which HBS & MIT Sloan show are missed by investors and the overall market, are applied as follows:
- 14,422 income statement adjustments with a total value of $1.2 trillion
- 17,629 balance sheet adjustments with a total value of $8.0 trillion
- 8,566 valuation adjustments with a total value of $9.9 trillion
Below, we highlight some of the items investors relying on GAAP or other core earnings metrics may have missed in 2019 10-Ks. Adjusting for these unusual items reveals a much different picture of a firm’s profitability than reported or other firms’ numbers.
Kellogg Company (K) – Understated Earnings
In 2019, Kellogg (K) had -$191 million in net earnings distortion that cause earnings to be understated. Notable unusual expenses hidden in Kellogg’s 2019 10-K include:
- $156 million in pre-tax charges, primarily related to reorganizations – Page 28
- $132 million in pre-tax charges due to withdrawing from two multi-employer pension plans – Page 28
- $83 million in selling, general, administrative costs associated with all restructuring programs – Page 73
- $35 million in cost of goods sold associated with all restructuring programs – Page 73
- $9 million in pre-tax charges to prepare for the potential adverse impacts of Brexit – Page 28
These expenses were partially offset by $188 million in other income reported on the income statement.
In total, we identified $0.56/share (20% of GAAP EPS) in net unusual expenses in K’s 2019 GAAP results. After removing this earnings distortion, K’s 2019 core earnings[1] of $3.35/share are significantly higher than GAAP EPS of $2.80. Furthermore, these adjustments reveal K’s true profits declined much less than reported. In 2019, GAAP net income fell 28% year-over-year (YoY). After removing earnings distortion, we find that K’s core earnings fell just 6% YoY.
With understated earnings, K gets our “Beat” Earnings Distortion Score (as featured on CNBC Squawk Box) and is likely to beat consensus expectations in the coming quarter.
3M Company (MMM) – Understated Earnings
We featured 3M Company (MMM) at the beginning of Filing Season in our report “Hidden Expenses Overstate This Firm’s Profit Decline.”
In 2019, MMM had -$630 million in net earnings distortion that cause earnings to be understated. Notable unusual expenses hidden and reported in MMM’s 2019 10-K include:
- $762 million in litigation-related charges related to PFAS matters and coal mine dust respirator mask lawsuits – Page 19
- $162 million loss on deconsolidation of Venezuelan subsidiary – Page 78
- $137 million in selling, general and administrative expenses related to restructuring actions – Page 76
- $72 million in cost of sales related to restructuring actions – Page 76
- $37 million in research, development and related expenses related to restructuring actions – Page 76
These expenses were partially offset by a $114 million gain on sale of business and an $82 million gain on the sale and lease-back of an office location.
In addition, we made a $276 million adjustment for income tax distortion. This adjustment normalizes reported income taxes and removes the impact of unusual or less persistent items.
In total, we identified $1.08/share (14% of GAAP EPS) in net unusual expenses in MMM’s 2019 GAAP results. After removing this earnings distortion, MMM’s 2019 core earnings of $8.89/share are significantly higher than GAAP EPS of $7.81. Furthermore, these adjustments reveal MMM’s true profits declined much less than reported. In 2019, GAAP net income fell 15% YoY. After removing earnings distortion, we find that MMM’s core earnings fell just 7% YoY.
With understated earnings, MMM gets our “Beat” Earnings Distortion Score and is likely to beat consensus expectations in the coming quarter.
Adjustments Reveal Abnormal Tax Rates
We featured Hertz Global Holdings (HTZ), Boston Scientific Corp (BSX), and Univar Solutions (UNVR) during Filing Season in our report “Abnormal Tax Rates Are Today’s Filing Season Find.” Each of these three firms reported effective tax rates significantly different than their actual cash tax rates.
In 2019, HTZ reported an effective tax rate of 500% due to a significant increase in its tax provision from the prior year. After removing the tax impact of unusual gains and losses in HTZ’s filing, we find the firm’s cash tax rate is just 31% in 2019. When we remove all unusual gains and losses, we see that HTZ’s 2019 net operating profit after-tax (NOPAT) is $583 million compared to its 2019 GAAP net income of -$58 million.
BSX recorded a tax benefit of $4.1 billion related to intra-entity asset transfers of intellectual property rights, which led to a reported tax rate of -584% in 2019. Our calculations show BSX’s cash tax rate is actually 8%. After removing unusual items, including the tax impact of unusual items, we find that BSX’s NOPAT in 2019 is $2.4 billion compared to GAAP net income of $4.7 billion.
Lastly, UNVR reported an effective tax rate of -9,500% in 2019 due to increased international tax impacts and tax gains on the sale of a business. After removing the impact of unusual gains and losses, we find the firm had a cash tax rate of 7%. After all adjustments, we find that UNVR’s 2019 NOPAT was $371 million compared to its 2019 GAAP net income of -$100 million.
No Substitute for Diligence
The adjustments above are just a few examples of the material items found in the footnotes of financial filings. Our technology enables us to deliver this level of diligence on fundamentals at a previously impossible scale. We believe this research is necessary to uncover the true profitability of a firm and make sound investment decisions. Professors at Harvard Business School and MIT Sloan published a paper showing that traditional data providers are not collecting financial data as carefully as needed for accurate earnings calculations. Specifically, these offshore operations miss about $0.45 out of every $1.00 of unusual gains/charges on the typical income statement.
The professors went on to show how our adjustments create a measure of core earnings that is more predictive of future earnings than comparable metrics from Compustat and IBES. Only by reading through the footnotes and making adjustments to reverse accounting distortions can investors and advisors alike get beyond the noise and get the truth about earnings and valuation.
This article originally published on April 9, 2020.
Disclosure: David Trainer, Kyle Guske II, and Matt Shuler receive no compensation to write about any specific stock, sector, style, or theme.
Follow us on Twitter, Facebook, LinkedIn, and StockTwits for real-time alerts on all our research.
[1] Our core earnings are a superior measure of profits, as demonstrated in In Core Earnings: New Data & Evidence a paper by professors at Harvard Business School (HBS) & MIT Sloan. The paper empirically shows that our data is superior to IBES “Street Earnings”, owned by Blackstone (BX) and Thomson Reuters (TRI), and “Income Before Special Items” from Compustat, owned by S&P Global (SPGI).