Filed Under: ,
Tickers in this Article: S, GE, BRCM, AMZN, GOOG,
More companies are reporting pro forma numbers. Companies no longer publish one set of earnings per share (EPS) numbers. Instead, we now get basic and diluted EPS in two or three or four sets. In addition to earnings required by GAAP (generally accepted accounting principles), Sprint's (S) last press release offered two adjusted operating income numbers, a pro forma net income, an adjusted net income, and a free cash flow metric (pro forma excluded, for example, hurricane charges). Thank you Sprint, I think.

It will only get worse because there is an overarching shift in the philosophical priorities that underlie GAAP. The accounting rules are anchored in two primary principles: reliability and relevance. Reliability means that you can measure something with objective accuracy. Relevance means that we care about the cost. But these two principles are in natural tension.

Take stock options, the old rule assigned a zero cost to most "plain vanilla" options. That is perfectly reliable-everybody can calculate zero! The new rule requires a fair value estimate.

But give two experts the same exact circumstances and they will produce different option cost estimates. The price of increased relevance (this could be called accuracy) is decreased reliability (this could be called objectivity). And, as goes the treatment for options, so goes accounting generally.

The Financial Accountings Standards Board (FASB) has embarked on a gradual shift away from historical cost and toward fair value. As I have written before, we are entering a sort of third phase for reported earnings. The first was typified by Jack Welch's famous ability to manage General Electric's (GE) earnings; in this phase, management was actually admired for managing a smooth earnings result. Enron and Worldcom then dethroned, respectively, reported EPS and reported EBITDA.

The second phase saw (most) managers carefully avoid the appearance of smoothing; the mantra understandably became "let earnings fall where it may." The third phase is "fair value" and Wall Street has no history of dealing with this. Reported earnings are going to become simultaneously more accurate, more convoluted, and more volatile. The market wants predictability but it won't get it, even from the best companies.

When we go to value a company, we really care about recurring operating performance. There are two parts to this view. Recurring means that ongoing business activities produced the earnings, and that we can expect them to continue. They are, at least in potential, sustainable. Operating means that core business produced the earnings – not financial engineering, not special transactions, and not asset dispositions. Therefore, many pro forma numbers are entirely useful because they exclude non-recurring and/or non-operating items.

To illustrate the difference between operating income and net income, I totaled each for all traded companies in my database (over 8,000 companies). You can cycle through each sector to see the difference. Also, I totaled non-recurring and unusual items (as classified by the company) to show they can be significant. You will also note that there is often a material difference between net income and the income used for EPS.

Operating income is an important number. But you can't use it "out of the box" to compare, say, valuation multiples (e.g., price-to-operating income). Operating income is before (or above) several items but the most important is interest expense on the debt. Operating income minus interest expense is therefore useful. The problem with net income and "income for EPS" is they contain other less useful items.

This all helps to explain the current renaissance of cash flow measures. Amazon (AMZN), for example, reports free cash flow in their earnings release because "it reflects an additional way of viewing our liquidity that, when viewed with our GAAP results, provides a more complete understanding of factors and trends affecting our cash flows." To see the difference, again by aggregate sectors, I totaled and compared net income to two flavors of cash flow: free cash flow and net cash flow. Free cash flow is cash flow produced for shareholders (i.e., after operating cash flows, investing cash flows and cash paid to service debt). Net cash flow is the "all in" cash flow; the net change in cash flow produced by a company for the year, counting every cash inflow and outflow. Net cash flow is probably the most concrete number on the financial statements (i.e., it is very hard to manipulate) but it is volatile and rarely can you extrapolate it forward.

The problem with cash flows is well-understood. They are lumpy. If your company produces zero cash flows because they invest in a new overseas facility as part of a global expansion, you don't really think their sustainable base is zero. Similarly, if the company received a one-time up-front cash payment for a long-term contract, you don't really think the lift is sustainable. Accrual is sometimes helpful and better than cash flow.

In my opinion, the best you can do for now is to consider both the reported and pro forma numbers. The historical importance of weighing the "quality of reported earnings" will become even more important. Ideally, you would construct your own pro forma based on recurring, operating earnings. But the pro forma provided by the company is a good place to start. You just can't take it at face value.

Broadcom's (BRCM) recent quarter included a pro forma that is illustrative. The difference between their reported and pro forma earnings included over a dozen items. I grouped them in the chart below. If you select the check-boxes in succession, you will see how the net income inflates to the pro forma income. Red notes indicate a dubious exclusion. Orange indicates debatable. Green indicates a favorable exclusion.

Broadcom's decision to exclude (or "back out") stock option expense is becoming common practice. You can make an argument either way, here. It is debatable because option expense is hard to measure. Clearly the option has a cost, but on the other hand, if Broadcom's stock doesn't increase, these options may expire as worthless. In which case, the expense would have been inaccurate (and it would not reverse in future periods, for the most part). But the exclusion of restricted stock is just plain wrong. Broadcom paid people in shares which have an obvious and measurable value. It must be counted as an expense. If these can be excluded, the same logic could exclude salaries!


Google (GOOG) also excludes stock option expense to produce a pro forma earnings number. Additionally, they exclude "in process research and development." In my opinion, Google's pro forma is reasonable. In regard to in process research and development, they bought a company and (as is increasingly the case in technology) some of the purchase was for the target company's ongoing R&D efforts. The rules require Google to expense this cost right-away, but understandably, Google views this purchased R&D as an asset. Therefore, they want to exclude it from a pro forma income.

Good luck in sorting this out. Keep in mind that cash isn't always better than accrual and sometimes it is good to exclude one-time items. We really are trying to discern the earnings that are generated by the core operating business.

comments powered by Disqus

Trading Center