Why is Everyone Still Earnings-Obsessed?
Board: Value Hounds
Why hasn't free cash flow (which is harder to manipulate and isn't as heavily dependent on discretionary accounting practices) replaced these crude yardsticks?
If by "free cash flow" you mean after capex, that has plenty of defects too. For example, if a company buys a large building for headquarters and pays cash, that will crash its free cash flow for that quarter (it wouldn't impact earnings to such an absurd extent because the expense would be properly amortized); vice versa, if the company sells the building to a REIT, with a 20-years contract to lease it back in a triple-net arrangement (plenty of REITs thrive on such deals), that will show a surge in "free cash flow" with no reality or substance.
Etc, etc. Cash-based accounting may work (kinda, sorta) for a small business, but not really for a large company (or government, even though, absurdly, almost all governments use it) -- accrual accounting has its own defects, but, overall, it's a sounder reflection of a company's business. Putting cash flow (free or otherwise:-) upon a pedestal, repudiating 8+ centuries of accounting, isn't the smartest of moves.
Just one example: consider a company that in the past has piled up a very large inventory of raw materials it uses for manufacturing (for example, rumor has it that many Chinese manufacturers did just that for copper, preferring to stockpile huge amounts, even as a "store of value" stabler than currency, and to ward against copper price volatility). Of course at the time that stockpile was being accumulated it depressed the company's free cash flow. But when they start running down the inventory, this suddenly flatters their "FCF" because, by that cash-based-accounting reckoning, quarter after quarter they appear to be getting their copper "for free". Of course this is not sustainable, any more than sell-and-leaseback arrangements -- just a transient boost to that cash flow. Accrual accounting avoid some of these distortions.
You can't really summarize a large company (much less an economy, despite the undue popularity of GDP:-) by any one single number; rather, whether you're trying to run it as a business, or considering it as an investor (to buy part of the business, i.e, equity; or lend it money, e.g, buying its corporate bonds), you really want to look at several "dimensions". And, as you mention Benjamin Graham, let me point out that "Security Analysis" (his immortal textbook for professionals, from the 1930s) makes this point quite transparently.
Looking at many relevant metrics is crucial. Book value may be irrelevant in many cases (as it relies on original purchase price of assets, possibly with write-downs but never write-ups, if the Peter Minuit Corporation had endured and still owned all of the land of Manhattan and no other assets, it might be shown with a book value of $24 -- a pretty useless number to any potential investor, right?-)
This isn't a theoretical issue. Youtube shows up in Google's book value at $1.65B (almost all "goodwill", not "tangible book", since YT's tangible assets at the time Google purchased it weren't much). These days [actually a couple of years ago] I saw an analyst ( http://www.valuewalk.com/2012/03/google-inc-goog-youtube/ ) estimating the value of Youtube at $45.7 billions (smacks of false precision to me:-) -- so while not as crazy as the Minuit case (since after all only a few short years have passed since the purpose, not centuries:-), the "book value" is also pretty useless to a potential investor here.
I could offer similar criticism of just about any financial metric; only looking at a lot of those that are relevant for a given company, and digging deep into modeling the company, is going to help (and definitely not in a way affording easy, useful comparisons across industries and sectors -- at most, if you're lucky, just of companies operating in the same industry).
My job these days is (mostly) to develop, and lead the development of, Business Intelligence software for internal use at my company -- over the years this has reinforced my appreciation for accountants and their challenging, difficult, and fascinating profession. You seem to suspect that, when presenting numbers to investors, accountants (working for a given company, or in charge of standardization across companies) may be distorting things with an agenda. In fact, it's not that hard (short of outright fraud, which ignores the standards as well as reality:-) for "forensic accounting" to spot problems if any are there -- see e.g http://www.amazon.com/Whats-Behind-Numbers-Financial-Chicane... (the book's authors are currently both TMFers).
Rather, even when presenting metrics strictly for internal use, to guide a single company's middle and senior managers, summarizing all the complexities of reality into just a few simple numbers and time series is a really difficult, challenging tasks; doing so across many companies (I can only imagine, as I only work on BI SW for this single one:-) must be far harder -- and standardizing usefully across the spectrum of companies (not just exchange-quoted ones -- also private ones that float public bonds) gets to mind-boggling complexity levels...