More and more companies are fudging their numbers, and it's becoming a big problem

"Adjusted" earnings are becoming an epidemic.

According to analysts at Bank of America Merrill Lynch, the percentage of companies reporting adjusted earnings has increased sharply over the past 18 months or so.

About 90% of companies now report earnings on an adjusted basis.

These earnings often exclude items a company deems "special," or "one-time," or "extraordinary." Investors, then, are left with a muddier picture of what companies are really making.

A company will argue these are "clean" numbers that exclude things investors need not concern themselves with. In this sense, then, companies prefer that you to look at less of themselves to make a decision about their real value.

Why?

"We are increasingly concerned with the number of companies (non-commodity) reporting earnings on an adjusted basis versus those that are stressing GAAP accounting, and find the divergence a consequence of less earnings power," BAML wrote in a note to clients on Tuesday.

BAML added (emphasis ours):

Consider that when US GDP growth was averaging 3% (the 5 quarters September 2013 through September 2014) on average 80% of US HY companies reported earnings on an adjusted basis. Since September 2014, however, with US GDP averaging just 1.9%, over 87% of companies have reported on an adjusted basis. Perhaps even more telling, between the end of 2010 and 2013, the percentage of companies reporting adjusted EBITDA was relatively constant and since 2013, the number has been on a steady rise.

We are increasingly concerned with this trend, as on an unadjusted basis non-commodity earnings growth has been negative 2 of the last 4 quarters, representing the worst 4 quarter average earnings growth in a non-recessionary period since late 2000.

GAAP earnings, or earnings that follow generally accepted accounting principles, are what many executives and analysts would have you believe is the rough cut of a company's financial condition (and hence, less representative of the company's underlying condition).

GAAP earnings, for example, include any charges a company may have incurred during a quarter - as, say, part of selling a unit. Companies adjust, or clean up, earnings to outline what management argues is the true picture of the company's actual operations.

(A popular "clean" number to cite is a company's Ebitda, or earnings before interest, tax, depreciation, and amortization. Of that, Charlie Munger has said, "I think that, every time you see the word Ebitda, you should substitute the word 'bullshit' earnings." So there's that. There's also non-GAAP, which is not exactly the same thing but is also a proxy for adjusted earnings.)

And as we've written before, this idea that there is an actual operation to measure is itself just a mirage.

Ben Graham, the godfather of value investing, wrote at length about his concerns regarding corporate accounting in his book "The Intelligent Investor." Using Alcoa's four different earnings figures from a quarter in 1970 as an example, Graham argues that the "true" value of a company doesn't exist the way some on Wall Street would have you believe.

The overarching point of Graham's analysis is that a company could claim anything that didn't happen exactly the same way in the prior quarter was a "one-time item" for a company. The problem is that this sort of ignores how businesses actually work. Which is to say: things happen, sometimes once, sometimes more than that. Sorting out which of these things matters is the whole challenge of being an investor (perhaps, even, an intelligent one).

And in today's environment, a related phenomenon is cropping up.

In a note out late last year, Bank of America highlighted what it called "worrisome gaps" (lots of gap imagery here) in earnings reports. Basically, the number of companies beating on earnings has been increasing at the same time the number of companies missing on sales is ticking higher.

The firm also noted that these adjusted earnings were beating GAAP earnings by about 30% on average, the most since the financial crisis.

This is a conundrum.

Common sense says growing profits while losing sales takes some kind of accounting wizardry to accomplish - outside of, say, a one-time (!) transformation that makes this possible - and the fact that roughly 60% of companies were beating on profit while 60% were missing on revenue is a concern. (It's unclear what the overlap is here on a company-specific level, but the bigger trend is the issue here.)

And as with Graham's discussion of the four alternatives for Alcoa's quarterly earnings, we could argue about whether certain dynamics in corporate America have changed enough to make a credible case that it is possible, on a non-shenanigans basis, for industries to grow earnings while losing sales on a consistent basis.

But this is all sort of beside the point.

Earnings, like records, are made to broken. Or at least, played with. This is the fun part.

The challenge for the investor community, then, is to make decisions about which earnings are the "real" ones, which earnings have been altered too much or too little, and what is the actual value of the company's business.

And the more companies that appear to be changing things, the more you might be inclined to think there's something to hide.