The cat is out of the bag. CFOs admit that they manipulate accounting rules to “misrepresent earnings”. They also say that Wall Street analysts play along. The gory details are in this recent paper “The Misrepresentation of Earnings”, which was featured on MarketWatch.com.

To our clients, this news is no surprise. But, for far too many others, it is a big surprise, and it speaks to the dangers of investing.

The duplicities of Wall Street research have been a big, hidden danger to investors for decades. As more people realize the dangers of Wall Street research, fewer will heed or pay for it. As Wall Street firms keep cutting research budgets, the quality of research will continue to decline and even fewer people will pay it any attention. We expect this downward spiral to persist until there are very few Wall Street analysts left, and as a result, we’re putting Wall Street analysts in the Danger Zone.

20% of CFOs Admit Intentional Misrepresentations of EPS

If you do not think earnings misrepresentation occurs enough that you need to worry about it, please read this quote:

“CFOs believe that in any given year a remarkable one in five firms intentionally misrepresent their earnings using discretion within generally accepted accounting principles (GAAP). The magnitude of the typical misrepresentation is quite material: about 10 cents on every dollar.”

Analysts Have No Incentive to Detect Earnings Quality

According to some CFOs, it is naïve to expect analysts to rectify the misleading earnings.

“Analysts usually don’t actively detect poor earnings quality.  The good ones do but the sell side has no incentive to detect earnings quality.”

The takeaway: there are too many loopholes in GAAP for investors to rely on GAAP earnings. If you rely on GAAP, you are taking more risk than you might realize.

Ways In Which GAAP Can Misrepresent Earnings

Accounting rules are not designed for equity investors, but for debt investors. Reading the fine print in annual reports is the only way to find the loopholes companies exploit to misrepresent earnings. In the paper, CFOs list a few of the 30+ earnings misrepresentation red flags from which we’ve been protecting clients for years:

1. GAAP earnings do not correlate with cash flows
2. Large/frequent one-time or special items
3. Sudden change in reserves
4. Changes in pension plan assumptions or large differences from peers
5. Changes in stock option expense assumptions or large differences from peers
6. Complex footnotes

The Biggest Problem With Misleading Earnings

One of the biggest and most overlooked problems in today’s markets is that when the basic understanding of the profitability is inaccurate, everything about the valuation of the stock is thrown off too.

Examples of Misleading Earnings

In the past year alone, we’ve made nearly $600 billion of loophole-closing adjustments to the income statements of 3000+ companies to protect our clients from misleading earnings. The following are a couple examples of the misrepresentations of earnings that can be found in 10-K’s. You can find more here.

1. Williams Companies (WMB: $50/share), which gets our Dangerous rating, reported $2.5 billion (33% of revenue) in non-operating income due to a gain on the “remeasurement of investments”found on the income statement. The size of this adjustment speaks for itself and shows why you cannot trust the EPS numbers, especially when it comes to P/E-based valuation metrics.

The big unusual gain actually drives WMB’s P/E down to 18, below the industry average of 44 while WMB is anything but cheap. Rather than using P/E, we prefer to use a DCF model to quantify the cash flow expectations implied by the current stock price. To justify its current price of $51/share, WMB must grow NOPAT by 17% compounded annually for the next 12 years. This expectation seems rather optimistic given that, since 2010, WMB’s after-tax profit (NOPAT) has declined by 6% compounded annually.

Those are the real numbers for WMB’s valuation. See how misleading the P/E of 18 can be?

More misleading, however, are the adjustments you cannot make from just looking at the income statement.

2. T-Mobile U.S. (TMUS: $40/share) also gets our Dangerous rating. Wall Street ratings are far more sanguine with 12 strong buys, three buys, three holds, and one underperform rating on TMUS. We think misleading earnings are a big part of the disconnect. Despite showing EPS growth, T-Mobile’s cash flows, aka economic earnings, are headed in the opposite direction.

Most investors do not realize the EPS and cash flows can diverge. For TMUS, the balance sheet has been growing faster than operating profits, which means cash flows are declining. One of the problems with GAAP is that it almost entirely overlooks the balance sheet. In 2014, the cost for T-Mobile’s adjusted balance sheet, aka invested capital, was nearly $3 billion. Our “adjusted balance” sheet is adjusted to close sizable accounting loopholes: over $11.8 billion for off-balance sheet operating leases and $1.7 billion in cumulative asset write-offs. In total, the cost of the capital that T-Mobile needs to run its business equals $3.66/share. Removing this charge reveals that T-Mobile’s economic earnings per share in 2014 were actually -$2.40, well below the reported $0.30 GAAP EPS.

TMUS is up 49% year-to-date and has become overvalued in large part due to GAAP EPS growth. However, to justify its current price of $40/share, T-Mobile must grow NOPAT by 13% compounded annually for the next 19 years. In this scenario, T-Mobile would be generating over $176 billion in revenue in year 2034, which is greater than AT&T (T) and Sprint (S) combined in 2014 and we feel this expectation is highly optimistic.

3. Deere Company (DE: $81/share) is a great example of how analysts can miss when earnings are depressed and wrongly punish a stock. Deere receives our Attractive rating but Wall Street mostly disagrees with just two strong buy ratings, eight hold ratings, and four sell ratings. The reason for the disconnect could be that Deere’s 2014 GAAP earnings were depressed due to non-operating expenses and changes in reserves.

In addition to nearly $100 million in non-operating expenses due to asset write-downs, Deere increased its trade receivables allowance by $905 million (29% of GAAP net income). Analysts who didn’t adjust earnings for this increase would have materially understated Deere’s core operating profitability. While 2014 GAAP EPS declined year over year in 2014, Deere’s economic earnings per share increased for the fifth consecutive year, highlighting the business’ ability to create shareholder value.

Better yet, Deere’s current valuation does not reflect this long-term value creation. At its current price of $81/share, DE has a price to economic book value (PEBV) ratio of 0.6. This ratio implies that the market expects Deere’s NOPAT to permanently decline by 40%. However, with Deere’s history of growing NOPAT and increasing economic earnings, we feel this expectation is awfully pessimistic.

Even if Deere failed to grow NOPAT going forward, its current economic book value, or no growth value is $143/share – a 76% upside.

4. Pension Assumptions: As we recently discussed, pension assumptions can have a meaningful effect on company earnings. As one CFO states “But when they make changes [to pension assumptions] that lead to them being different than everybody else, it’s kind of a red flag.

In 2014, the average pension expected rate of return across all companies in all sectors under coverage was 6.5%. In the Utilities sector, the average rate of return assumption across all pension plans was slightly higher, at 7.1%. However, since 2000, Dominion Resources (D: $75/share) has, on average, projected a pension expected return rate of 8.8%. This expected rate matters because over this time frame, Dominion’s average annual return has been just 6.8%, well below its lofty expectations. In 2014 alone, Dominion was able to record a non-operating net periodic benefit of $140 million (11% of GAAP net income) due to its assumptions.

A Final Note On Misleading Earnings

One final item that the CFO’s noted was a red flag for misrepresented earnings was complex footnotes. At the end of 2013, the average number of characters in 10-K’s was 409,000 up from 224,000 in 1998. As company reports get longer, the material information gets harder to find. Wall Street firms weren’t finding this information before and it seems hard to believe they will begin doing so as 10-K’s continue to get longer. The added cost and time to do their proper job is just one more reason why Wall Street analysts find themselves in the Danger Zone this week.