SNOW - Mind The GAAP: Does Anyone Report Real Earnings Anymore?

2023-08-08 12:12:53 ET

Summary

• Non-GAAP accounting is on the rise, with roughly 75% of companies reporting non-GAAP earnings and about 20% of companies reporting a non-GAAP profit despite a GAAP loss.
• Aggregate earnings appear to be overstated by at least 10%-15% by companies using earnings adjustments that don't reflect their true costs of doing business.
• This means that stock returns are likely to be roughly 1% per year lower than they would be otherwise in the long run, bolstering the case for bonds and/or cash.
• Some red flags to look for in company earnings releases, and some of the worst offenders with stock-based compensation and earnings adjustments.

Lies, Damned Lies, and Non-GAAP Accounting

During the 1990s, many companies reported dazzling earnings to investors during the dot-com bubble. But the dirty secret of these earnings is that they routinely excluded stock-based compensation and a slew of other costs from their expenses. When the market turned down from 2000-2002, many of these same companies ended up going bankrupt or severely diluting their shareholders because their earnings weren't real. The Sarbanes-Oxley Act was passed in 2002 after high-profile accounting fraud scandals at Enron and WorldCom. Sarbanes-Oxley tightened the definition of earnings and expenses under generally accepted accounting principles (GAAP) to give greater clarity to investors. President George W. Bush declared at the time that "the era of low standards and false profits is over."

But in the 21 years since Sarbanes-Oxley, the percentage of companies reporting non-GAAP EPS has risen from less than 5% to over 75% . To be fair, earnings overstatement lies on a spectrum. Most companies weren't inflating earnings as severely as Enron, or even as much as high-flying tech companies that heavily diluted shareholders. But many were, leading to heavy losses for shareholders. In the 21 years since Sarbanes-Oxley, hard-learned lessons seem to have been forgotten, and many of the old tricks are back.

The research is starting to come in on these non-GAAP earnings, and I doubt you'll be too surprised by the conclusions.

• A Harvard Business Review article , cleverly titled "Mind the GAAP," found that a fifth of firms reporting a GAAP loss reported a non-GAAP profit.
• A Cornell and MIT study found that companies with large spreads between GAAP and non-GAAP earnings tend to have abnormally high CEO pay.
• The WSJ reported that SEC this year has inquired about non-GAAP figures at companies like Lyft ( LYFT ), Sleep Number ( SNBR ), and Graham Holdings ( GHC ), with the view that they're uncomfortable with aspects of the non-GAAP disclosures being made.

Are Non-GAAP Earnings Overstated?

The answer is yes. Widespread use of aggressive accounting by companies means that S&P 500 earnings are likely overstated by 10% or more . And at smaller, younger companies, the effect is greater. Non-GAAP isn't always bad, but the way it is used generally is. Buffett often used to grouse about GAAP and how it understated Berkshire Hathaway's (BRK.A) ( BRK.B ) investment earnings. However, for the typical company, GAAP is a necessary constraint and is often found buried in the bowels of their earnings reports, as required by law.

One (fairly conservative) way of looking at the discrepancy is to compare operating earnings to reported (GAAP) earnings. GAAP forces companies to include stock-based compensation, legal costs, M&A restructuring costs, goodwill write-offs, etc. Operating earnings aren't the same as non-GAAP but I conveniently have longer-term data on operating earnings, so I'll share to illustrate. Here, you can see that GAAP is almost always lower in the aggregate than operating earnings. Here, last quarter's GAAP earnings are about $48.5 per index point, while last quarter's operating earnings were $52.4 per index point. If you annualize this, it implies GAAP earnings of about $195-$200 for 2023, vs. operating earnings of about $210-$215.

Investors should also note the "kitchen sink" tendency, which is for everyone to take write-offs in times like 2008 because of the easy ability to blame the external economy. And a Bloomberg piece from this spring sounded the alarm on a worsening discrepancy between operating cash flows and earnings. However, recent research shows that non-GAAP is even more aggressive than we would think by simply using operating earnings.

The above analysis shows that as of the end of calendar year 2022, non-GAAP earnings were 15.7% higher than GAAP. That's a huge gap (pun intended). It's the difference between the index trading for 20.2x earnings (on the very high side of normal), and 24.1x earnings (dot-com bubble-esque). If we take this difference at face value, the returns on the S&P 500 should be expected to compound at 1% less annually in the long run.

And where does the difference go? To well-paid corporate executives, law firms, investment banks, and the owners of acquisition targets that said bankers, executives, and lawyers manage to overpay for.

Here we see that stock-based compensation is the No. 1 adjustment used in non-GAAP numbers. Second is the amortization of intangibles, which is heavily used by pharma companies and big tech to ignore the effects of patents expiring. Third is goodwill impairment, and fourth is restructuring charges, which are both generally related to M&A done by companies. Goodwill impairment costs companies more than restructuring charges, creating the rebuttable presumption that a lot of M&A ends up destroying shareholder value. Another one I see over and over again is the effects of fines, settlements, and legal expenses. If your company has paid more than one large fine in the past 10 years, it's likely a cost of how you choose to do business, not a one-time occurrence. These are all expenses, and if you're a shareholder in the business, your investment is on the hook.

If Earnings Are Overstated, How Do We Apply This Knowledge?

Macro:

If the true P/E ratio of the S&P 500 is 24x, then you might think twice about your asset allocation. The most basic model for valuing stocks is to take earnings yield + earnings growth to get your long-term expected return. If you rack your brain to go back to your high school calculus class, you should be able to see that a company trading for 20x earnings (a 5% earnings yield) and a 3% growth rate should equal the long-term return of a bond with an 8% fixed dividend.

This is theory of course, and stock returns can be much higher or lower in the short and medium term. But if we assume that the S&P 500 is trading for 20x earnings and can grow those earnings at 4% annually (2% real growth + 2% inflation), then the theory would hold that the long-run expected return should be 9%. And that's about what stocks have returned - you made more if you bought when they were cheap and less if you bought when they were expensive, like during the dot-com bubble. Here, we have to assume that the stock market's earnings yield is closer to 4%, while the overall economic growth prospects don't appear to have meaningfully changed. You could argue that they're higher because of Chat GPT or whatever or lower because 21st Century earnings growth was driven much more by tax cuts and interest rate repression than actual increases in technology. Here, we get a long-term return estimate of 8% annually for the S&P 500 vs. cash which is paying 5.5%. There's not a lot of turkey on the bone by this measure, considering 10-15 years ago cash paid 0% and stocks' implied returns were close to 12% annually.

With bonds, your returns are much easier to forecast, because the companies or governments either pay you or they don't. Bond investors also are generally able to generate capital gains with the passage of time. Stocks are much more uncertain, which is why it's weird right now for stocks to be priced so high relative to other asset classes. If you pump money into a bond fund with a five-year duration or less and hold it for five years, you're very likely to make that return. For example, Vanguard's short-duration investment grade bond fund ( VFSUX ) is offering 5.5% or better for a three-year duration. If investors are forced to reprice stocks due to an inability to maintain the façade of growth and profitability, then stocks could easily lose a third or more of their value.

Micro:

Now, here's the fun part. Many companies report some sort of non-GAAP earnings, but our job as investors is to avoid the worst and find the best. There's a well-known asset pricing anomaly called the accrual anomaly which holds that companies with cash earnings tend to outperform those who rely mainly on accruals.

There are still a fair amount of companies that still report GAAP earnings after all. For the S&P 500, that leaves us with roughly 100 companies that are old-school enough to report GAAP. As a general rule, I think that these companies are likely to be better investments than the rest of the corporate management teams that are keeping up with the Joneses. Their earnings quality is higher, and that's generally a sign of superior returns.

Second, we can find some of the worst non-GAAP offenders and avoid them.

One approach is to look at companies with unusually high levels of stock-based compensation. Seeking Alpha author Confounded Interest did a study late last year on the worst offenders for stock-based compensation. A few of the companies flagged for excessive levels of stock-based compensation reveal some of the flashier, more controversial companies in America:

• C3.AI ( AI ).
• Robinhood ( HOOD ).
• Palantir ( PLTR ).
• Snowflake ( SNOW ).
• The Trade Desk ( TTD ).

Figures for stock-based compensation are not always apples to apples, so SBC alone isn't something you'd base trades on, but it can be a good start.

Another big red flag to watch out for is companies that report GAAP losses but adjusted profits. Since making a profit is a long-term prerequisite to staying in business, losing money but reporting an "adjusted profit" is a red flag. Research shows that companies with large adjustments to income are nearly twice as likely to be forced to restate earnings.

One big red flag to look out for is the phrase "adjusted EBITDA." Adjusted EBITDA is for companies that can't figure out any other way to report an adjusted profit, so it's in a way a code word for losing a lot of money. While it's not impossible for companies in tough positions to turn around, creative earnings announcements are a marker for overvaluation, dilution, and the potential to go out of business. Companies I've noticed reporting adjusted EBITDA lately include many of the heavy SBC users above and also:

• Carvana ( CVNA ).
• Kyndryl ( KD ).
• Coinbase ( COIN ).
• DraftKings ( DKNG ).
• Redfin ( RDFN ).

This is by no means an exhaustive list of companies with non-GAAP issues and abnormally high levels of stock-based compensation. The number of companies with these sorts of characteristics always peak at the top of the bull market. They then get flushed in the ensuing recession and bear market.

Bottom Line

Corporate America has a large and growing difference between actual profitability and non-GAAP reported profitability. As an investor, you should aim to know what the companies you own are actually making and be paying attention to common red flags. Forewarned is forearmed! What companies are you seeing with excessive stock-based compensation and/or earnings adjustments? Share your thoughts below in the comments!

For further details see: