Back when I entered the stock market, and when the CFA Institute was run by professional investors, that organization published an industry-by-industry guide to securities analysis written by veteran industry specialists.
The section on technology was a real eye-opener. It was a litany of virtually every accounting fraud known to man. The abuses ran in two related areas:
–capitalizing (and therefore delaying recording as expense) almost everything, especially research and development expense, and
–setting unrealistically lenient schedules for depreciating or amortizing these balance sheet accounts.
The result (and intent) of this chicanery was income statements that showed profits far in excess of the “real” amounts–even paper “profits” when a company was actually bleeding red ink.
The abuses were so blatant that the accounting rules were changed in the US to force virtually all companies to expense R&D costs as incurred, rather than store them up on the balance sheet. Virtually all companies still write off against income the cost of their plant and equipment much more slowly in their financial reporting to shareholders than they do in their tax reporting to the IRS.
Fast forward to the present.
What effect does the long-ago change in accounting rules for R&D have on today’s publicly traded companies? For hardware-dependent firms, not that much. Yes, computers may wear out or become obsolete much faster than, say, a cement plant. But the presentation of profits to actual and potential shareholders of a semiconductor foundry, a server farm or a steel mill allow for a more or less apples to apples comparison.
Not so for software-oriented firms.
Let’s take AMZN as an example and do a back-of-the-envelope calculation.
During the first half of 2013, AMZN spent $2.97 billion on technology and content. If we assume that all of that were capitalized instead of expenses (an aggressive assumption) but that any resulting financial reporting profits were subject to tax at the very-high US rates, then I figure AMZN would have shown earnings of $4.50 per share so far in 2013 instead of the $.46 actually reported. This would suggest, in ballpark figures, $10 a share in EPS for the full year.
Maybe the multiple isn’t so crazy.
Just as important, the same adjustment should apply to any other R&D-centric firm. R&D may be creating important long-term intellectual property assets for a company, but the accounting rules (for sound historical reasons) portray this activity as a minus.