There's a fourth possibility:

4. Companies increasingly occupy Extremistan, where a large portion of their financial performance comes from one-off, Black Swan events. Both the GAAP and non-GAAP numbers are true but useless; the future cash flows of the company depend upon unforeseeable events in the future.

The description that the article gives of the non-GAAP results supports this. Take, for example, a lawsuit against the company that results in a $5B judgment, payable out over many years to survivors of an environmental disaster. GAAP results include this lawsuit, non-GAAP results exclude it. Which more closely represents the future financial performance of the company, from the POV of a long-term value investor? Well, that depends whether there will be future lawsuits like this one. It depends whether the verdict will be reversed on appeal. It depends how many survivors survive.

The same goes for things like severance pay and patent lawsuits. Increasingly, it also applies to "normal" business, things like product introductions and the price of oil.

In Zero to One, Peter Thiel mentions (and decries) that an increasingly large portion of America is adopting an "indefinite" worldview, where the future cannot be known and it's futile to try to predict it. I'm not sure folks really understand the potential implications of this shift. Does "investing" have any meaning when the assumption is that returns will be dictated by large, random events? Does "meaning" have any meaning? What does it even mean to be rational, when reason is a tool we use to predict and our assumption is that events are unpredictable?

Treating environmental disasters as 'random' events is too kind. I think in just about every major recent environmental disaster you can without too much difficulty, trace back the disaster to cost cutting measures where decisions were made to put off long term risks in favor of short term profits.

The BP oil spill, shortcuts were made in the safety of the drilling equipment emergency cutoffs, in the Los Angeles methane leak, the company had put off maintenance to try to make rate payers pay for improvements instead of out of profits to the company. The Northern California San Bruno pipe explosion - safety funds had been diverted to executive bonuses.

Smart investors (and insurers of these companies) should be digging into long term risks. It is not that we cannot predict it - we are too cheap and profit oriented to pay to mitigate it.

If losses can be socialized (which they indeed can, if you're big enough), it's smart (albeit sociopathic) not to bother mitigating risks. It's much cheaper to PR your way out of disasters than to actually prevent them from happening.

Though sometimes it seems like it's executives reaping short term gain while the company itself pays the long term cost at a later point after the execs have moved on.

I'm not sure that this is mutually exclusive from random. Random refers to the timing of them, not them being unexplainable.

Maintenance is a form of insurance.

Possibility Five:

5. The companies are incontinent. The rank and file neither know what GAAP or non-GAAP really are within their limited scope. To appease bosses that themselves have bosses, etc, they don't 'cook' the books, they rely on numbers that are fudged or optimistic from 3 divisions down the ladder and across the world and then fudge them some more.

Simple Example: If you have a 90% 'truthfulness' data transmission rate things get bad quickly. Say I tell my 10 underlings to go get staplers. They then go tell their 10 underlings to go buy staplers, etc. If the transmission rate is 90%, then at the first step down I get 9 staplers, not 10. The next layer down I get 81+9 = 90 staplers from 110 folks. The next layer down I get 729+81+9 = 819 from 1,110 people. By layer 8, I am below a 50% data transmission rate to go get staplers. Granted there are now over 111,111,110 people working at the company then, but you get the idea.

Now bubble that upwards instead. Do you think that the folks there are giving a 90% truthfulness rate in what they tell their bosses? I know the situation is a lot more complicated and layers get jumped and the chain of command is not nearly so nice. Still, I think the point stands that the companies maybe are just dumb and it may just be their fate to be so.

As amusing as the idea of large corporations being unable to control their bladders is I think you meant "incompetent", not "incontinent". But that leads to (silly) possibility six: The companies are incontinent. In the rush to go to the bathroom they'll just put whatever numbers they can to get the financial results out the door.

Yeah, the Home Depot thing sounded similar. Sure, this particular data breach has no bearing on their general financial position, and you can make a compelling argument that you should ignore it: it's a one-time weird thing, and they're not going to have data breaches every year.

... or are they? At least, will they have data breaches every few years? What's the annual cost of doing all the work needed to prevent data breaches? If it's less than the averaged-out cost of cleanup, why weren't they doing it all along? If it's more, and they're accepting data breaches as a cost of doing business, shouldn't they be reporting that cost?

And if they don't think it's going to happen again, why not?

Presumably there's an assumption that they are "taking it seriously" and hardening their systems as a result. While it may not make them any less of a target (indeed, people now know they can be hit, and what NOT to try if they plugged this hole), they hopefully have taken steps to limit the legal, financial, brand and customer fallout.

If you think that is a valid assumption, perhaps this risk is not mitigated entirely, but it would certainly change the risk profile of such a black swan event, would it not?

The most common abuse of GAAP isn't Extreme events - it's companies excluding some kind of common event every year. This happened at a company I used to work for: for 4 years out of 5 there was a "one-time-only" exceptional cost taken out of earnings. After a while it got silly.

This isn't extreme good fortune, which would appear in both numbers.

It is futile to predict the future. GAAP is just there to try and make it easier to interpret the past, by applying some rules to compare companies against each other.

Good point.

Although, I never really was able to reconcile Thiels' criticism of an 'indefinite' worldview. It appears to me if the factors that influence things are not only the unknowns but the are unknown unknowns (i.e. you don't even know what you would need to know that will determine the fate of the company/country/whatever). In that case, how could you be anything but uncertain and just try to structure your life to benefit from convexity or as Taleb coined, anti-fragility?

I don't know what an accounting post-modernist is, but I do know that a consistent approach to keeping the books makes it possible to evaluate the positions of companies meaningfuly.

Companies as varied as Bank of America, Groupon, and Kinder Morgan all find creative and fascinating ways to assert that their bullshit financial metrics are more meaningful than GAAP.

Generally speaking, companies that aren't a shitshow and are actually making money don't bother cooking up new accounting schemes. Their CFOs focus on counting the profits, not heating air.

There's a sort of obvious 4th option: The non-GAAP numbers are fake and flattering and most of the investors know this but believe all other investors to be chumps and so everyone acts as if the non-GAAP numbers are true hoping others will buy in and let them cash out handsomely... thus beginning a game of chicken.

Oh that doesn't happen anymore. It's 2016!

Matt Levine is a smart guy but often tends to oversimplify. It is not just one of the three, is all three at the same time.

The number 1 is more or less true for a good number of companies. Number 2 and 3 applies to the rest of the companies depending on the investor you are looking at. The number of gullible investors is non-zero. Throws in that many companies are in option 1, and distinguishing between 1 and 2/3 became a complex problem.

So, we have a subset of companies whose price is determined by the ratio between good and bad investors. Obviously, good investors will fly away from these companies. But these companies could make you a lot of money if you leave soon enough. And then in finance classes, they use to say: "the market could stay irrational longer than you could stay solvent."

The results? Inflated Pro-forma statements stay around for enough time to do damage.

> and all the investors are chumps who believe whatever management puts in the big font at the top of the press release

Chumps, no. Humans, yes.

Better to have something standardised with known quirks and weakness than have a complete free-for-all. Yeah some of it will be an improvement but some won't and you can't easily tell which is which. Might as well not bother reporting anything then.

What's wrong with GAAP that all of those companies feel the need to report non-GAAP numbers?

long story short, different companies have different economic circumstances - different assets, different risks, different cashflow, supplier and customer arrangements, different operating structure, different expenses, etc that will all influence the nature of GAAP statements, being a common set of standards applied to ALL companies. It's like measuring all people with a single score like IQ.

Experienced investors never take GAAP at face value - most of them will work with both the GAAP and pro forma (as well as any other public releases) and try to reverse engineer the individual lines in order to tease out more pertinent (in their opinion) factors that affect the value of the business. different investors value risks differently but the pro forma is created by management for a general set of investors in order to bridge the gap between GAAP and exact economic reality.

Exactly. Gaap doesn't necessarily perfectly reflect economics of the company's business. So, the non gaap measures are there to be used in conjunction with, not instead of, gaap

My understanding is that at least the IFRS standards are developed precisely with the goal to fit all businesses in all industries. Whenever concerns come up that this is not the case, the standard is amended. And so companies who are making claims that IFRS does not adequately describe their business are justly met with suspicion. If GAAP does not work that way then it's one more reason to move on to IFRS.

I think its wrong to think that any accounting system can ever be made to be appropriate for all industries (GAAP, IFRS, or anything else). They each require 100s of decisions and assumptions that may bias or mislead investors one way or another; its not like any one of those decisions is the "right" one either. Its more important that we make sure investors are aware of what those pitfalls are. Not to mention, the cost of asking every US company to switch to IFRS would be enourmous (though i'm sure the big 4 accounting firms wouldn't mind).

As a side note, GAAP definitely isn't industry agnostic. For example, FASB (the financial accounting standards boards) issues guidance on revenue recognition for all sorts of companies and the guidance for a software company is very different than the guidance for your generic widget manufacturer.

? You linked to another Levine article, about Goldman.

Where's the Morgenson article?

what is post-modern accounting?

It involves an Italian bistro.

And dense texts describing how the act of reading that 10K necessarily involves the reader in the narrative of the document.

And some business about talking to a lion, I think.