Friday, July 31, 2009
A short note on GAAP fetishism
Glenn Reynolds links to a bit of commentary by David Pauly who says that "Wall Street analysts ... continue to promote earnings lies." The supposed evidence is that analysts are adjusting their own models to reflect earnings that are not precisely according to generally accepted accounting principles (GAAP):
Stock analysts continue to promote corporate earnings lies, insisting that net income isn’t really what investors need to know.
Instead, their earnings estimates ignore often huge expenditures that can’t help but affect a company’s health.
In analystspeak, Intel Corp. wasn’t hit with a $1.45 billion fine from the European Union in the second quarter for anticompetitive practices.
After setting aside funds to cover the fine, which Intel is appealing, the semiconductor-maker had a quarterly loss of $398 million, or 7 cents a share. Disregarding the fine altogether, analysts maintain the company earned 18 cents a share, beating their average estimate of 8 cents....
General Electric Co., which makes jet engines and electric power equipment and has a financial services arm, had a second- quarter profit of 24 cents a share. GE and the analysts emphasized earnings from continuing operations, which at 26 cents a share, exceeded their estimate by 2 cents. A $194 million loss from discarded businesses was discarded.
Wall Street’s big earnings lies must exasperate investors. They already have lost faith in the reported earnings of banks that are the center of the financial system.
Both the substantive complaint -- that analysts value companies on a basis other than GAAP -- and the conclusion -- that "Wall Street's big earnings lies must exasperate investors" -- are uninformed at best, and typical business journalism at worst.
Accounting is, by its nature, retrospective. It tells you what a company did. Investing, the business of analysts and the portfolio managers who read what they write, is prospective. Investors in liquid public stocks do not really care what happened, they want to know what will happen. Accounting does not tell them that; it can only give clues about future performance by describing past performance in a particular way. In order to turn historical results -- inherently useless information to investors -- into a useful prediction of future results, investors require analysis, not the superficial regurgitation of GAAP results. Securities analysis is an art, but at its center is the idea that some information about past performance is less predictive of future performance than other information. So, for instance, if a company earned $500 million last year and lost $200 million this year, it is important to know why. If the reason is that it had to pay a $800 million fine for an offense that it is unlikely to repeat, then it is reasonable to say that for the purposes of predicting future performance we should regard the company's core earnings as having increased 20%, notwithstanding the requirements of GAAP.
Furthermore, there is powerful evidence outside Wall Street "sell side" analysis that untweaked GAAP is not useful for making predictions: Lenders, who do not care a wit about the stock price, do not use strict GAAP measures when assessing the creditworthiness of a prospective borrower or in the calculation of financial covenants under loan documents. GAAP, it turns out, is no more perfect a device for credit analysis than it is for stock analysis.
Then there is Pauly's idea that "Wall Street's big earnings lies must exasperate investors". This is silly for at least two reasons. First, the only investors who really matter (in the sense of dollars involved and ability to affect a stock's price) are institutions (mutual funds, pension funds, and hedge funds), and they perform their own analysis. Sure, institutional investors are interested in what Wall Street analysts have to say, but they take it with a huge grain of salt. They use Wall Street analysis as an input into their own work, and often write internal reports that far exceed the quality of the "sell side" stuff that Pauly is complaining about. Pauly would no doubt be appalled to learn that "buy side" institutional analysts often further adjust GAAP results, both by rejecting some of the adjustments suggested by companies and sell-side analysts and, no less frequently, making adjustments of their own. Why? Because institutional investors have very different philosophical approaches and time horizons. You are going to make a different prediction about a stock and care about different things if you are willing to hold it for five years rather than five months. The former investor might be happy about heavy spending on a clinical trial for a product that will not launch for three years, and the latter might hate it. That's what makes markets.
Second, this idea that one analyst's shoddy argument amounts to a "lie" that might actually deceive somebody is a fantasy of prosecutors and journalists, but silly in the real world. Even the now diminished community of sell-side analysts is a cacaphony of different voices and competing opinions (opinions being the operative word, by the way); fourteen different analysts write coverage on my own company, and there are at least ten different opinions among them about the relative importance of different financial measures. Is it revenue growth that matters, gross margin, net margin, operating cash flow, EBITDA, or EBITDA adjusted to exclude equity-based compensation? Any investor who pays attention to just one voice is an idiot, and deserves to lose his money.
The recent claim, popular among journalists and prosecutors, that financial information that is not GAAP is somehow a "lie" is itself, well, a lie. GAAP sets a baseline, a minimal standard for financial information that every public company must produce. That attribute of GAAP results does not magically render them into a perfect device for predicting future results. Other information is required to do that, and some of that information is financial. An opinion about a company's prospects based on non-GAAP information, financial or otherwise, is not a "lie," it is an opinion. There is still a difference between the two.
8 Comments:
By 
TOF, at
Fri Jul 31, 10:25:00 AM:
And that is why I'll take an economist's analysis over an accountant's analysis hands down.
By 
, at
Fri Jul 31, 10:28:00 AM:
Some people think there is objective "truth" in accounting, and that therefore any report that doesn't conform even slightly from GAAP must therefore be deviously false.
These are not people who live in the real world of business.
By , at
Fri Jul 31, 10:38:00 AM:
Good post. Like a lot of good posts, it was all stuff that I kinda-sorta knew, but had never really put together like that.
By Billy Bob Corncob, at
Fri Jul 31, 01:30:00 PM:
GAAP and non-GAAP measures of performance
I would put it a bit differently.
Starting from the simple. Most analysts tend to focus on multiple metrics. Reported net income is one. It provides useful information in part because the FASB tries to have the GAAP categories reflect current results. It also provides useful information because the standards are uniform. (This is, for example, a problem were you to allow banks to start fiddling too much - they will always fiddle - with marking to market. They found during the Savings and Loan debacle that they needed to tighten the rules to avoid leaving investors with hard to compare results. And in economics, all decisions are made based upon what the alternative is, not in a vacuum.)
So GAAP tries to reflect the current condition of a company, earnings, balance sheet, and cash flows.
Analysts tend to focus on free cash flow, not earnings. Over time, the two should converge, but net income tends to be subject to some issues that free cash flow gives a better picture of. Free cash flow is not, however, a GAAP measure.
In the Intel example, the proper valuation method is probably something like what they did. You take the company's free cash flow (and exclude the one time charge); you project it into the future (in some sensible, sound way); and you discount using what you believe to be the cost of equity. The number is then adjusted by adding cash on hand. So in our example, the EU fine probably should not be viewed as affecting the pattern of earnings Intel can expect. It should be, and would be, viewed as a reduction of cash on hand. Thus, reducing one's valuation. It's not forgotten, or fiddled with. It just, with good analysts, enters into the estimate of the value of company stock through a different door, so to speak.
Analysts, by the way, are not bound to use any GAAP measures. Except they must provide honest advice.
Companies can use non-GAAP performance metrics. In some industries, for example, certain types of operating income, it is thought, give a better picture of company performance than net income itself, with all its one time charges and the like. Companies are, however, limited by the SEC in the ways they can use and disseminate non-GAAP measures.
This is called "Regulation G." It can be found at 17 C.F.R. Part 244.
Have a good investing day.
By PD Quig, at
Fri Jul 31, 01:40:00 PM:
Glenn Reynolds linked to an article by T.J. Rodgers that is worth reading.
http://pajamasmedia.com/edgelings/2008/12/03/tj-rodgers-my-financial-statements-are-a-mystery-even-to-me/
I agree with Rodgers. In the past ten years FASB and other regulators have made company financial reporting more opaque and less understandable to the normal human being.
By Billy Bob Corncob, at
Fri Jul 31, 01:50:00 PM:
Not so fast my friend
I have been reading financial statements for about 10 years, with a more attuned eye in the last 6 years.
I think TH would tell you that I know a fair amount about them. Not as much as some of the better accountants I work with. But a fair amount.
My strong belief is that financial statements are as easy to read as the company and their auditors want to make them. Period. They are obtuse when the CFO wants it that way. Or doesn't push to make them understandable. Nothing has changed. Except that in the 1990s there was essentially no enforcement by the SEC of the rules for GAAP accounting (count the number of financial fraud cases), and now there is.
Why? Because businessmen - not all but enough to cause harm - did not follow the rules. And the result of misbehavior that causes significant harm is always more rules.
Having said that, the rules allow companies to make things very transparent - if they want to. Very transparent.
I won't characterize Rodgers' statement, except to say that it is really, if one knows enough and thinks about it enough, reckless.
GAAP is a cost of having a pretty efficient capital raising and captial allocation system in this country. It works much better with rules than without, because the centerpiece of any impersonal market - any impersonal market - is some alternate source of trust.
Even a guy from Iowa knows that.
By PD Quig, at
Fri Jul 31, 03:24:00 PM:
Fair enough, but I've worked for three Silicon Valley software startups that never even considered going public due to the overhead required by regulatory regimes. I didn't read Rodgers as saying that regulation was bad per se. Just that the likes of FASB and SOX have overcorrected so as to begin hindering greatly the capital formation process.
By , at
Fri Jul 31, 05:32:00 PM:
From Link,
I don't know where to begin on this. I could go on for days pissing on GAAP.
To start from first principles, GAAP necessarily requires accruals which get away from the "cash in - cash out" way I ran my paper route. This is basic and necessary, but many sophisticated investors actually want to reverse engineer GAAP so they can create their own cash-flow discount models. They're often frustrated.
Businesses are very different from one another -- GAAP wants to pretend they're the same. This is a flawed intellectual conceit. As businesses have gotten both bigger and more diverse, this conceit has led to no end of problems. One size doesn't fit all. If I'm wrong on this, then why do p/e ratios vary so wildly ... but often have a consistency within industries, etc, etc.
A very basic difference -- even before getting industry specific -- is that some businesses are short-cycle and some are long. Many think that Jack Welch walked on water. But for years he created a black box at GE where he fucked with the numbers among the smorgasbord of GE short and long cycle businesses, along with timed M&A deals. There's a reason GE changes their segment reporting every two or three years. When this maxed out, Jack left current CEO Immelt a bag of goods. I used to feel Immelt's pain -- until he put on his Gucci knee pads for The One. Sandy Weill had a similar game going with M&A-related cookie car reseverves -- Citigroup is his legacy.
What I'd instead ask of GAAP is a process to create hard auditable numbers -- as much as possible -- so investors can better make their own judgments on future prospects. Instead, the shitheads* at FASB have taken us in the other direction. Integration with the principle-based approach of the IASB will only make this worse.
Here's one proven example of FASB stupidity -- FASB 125, which I once had to know, but since rescinded, should have been a warning signal over subprime securitization. It's a lesson in shithead* accounting theory trumping common sense ... with bad results. Don't get me going on stock option accounting.
Link, over
* shithead (Bronx definition): someone who's thinks they're smart, but they're actually stupid. I was called this a lot growing up ... I still hear it from many.
