The Bubble and the Balance Sheet

Format for Printing

Format for printing

Request Reprints


By howardroark
August 3, 2001

Posts selected for this feature rarely stand alone. They are usually a part of an ongoing thread, and are out of context when presented here. The material should be read in that light. How are these posts selected? Click here to find out and nominate a post yourself!


Pardon the intrusion. Landed here via Cheeze's link from the Berkshire Board. Before I comment on some of the more theoretical points you made, let me try to answer a more concrete question.

When we look at the list of assets AOL has that you listed in your post, you raise the very pertinent question whether the assets that are listed there could ever command that kind of price anywhere else but where they happen to be: on the left side of AOL's balance sheet. So either AOL just carries those assets on the balance sheet and hopes everything goes okay, or that asset has to be written down on the occasion when it is determined, as you put it, to be "impaired." But when does that occasion occur? How do you know when your asset is "impaired"?

I guess the accountants at JDSU looked at the balance sheet and figured, yep, this is the time for a write down, uncomfortable as that kind of realization might be. Should AOL's accountants do the same?

Again, I don't know. I'd sure like to see someone attempt an explanation, though.

I can tell you the FASB position on goodwill impairment under the new Business Combinations Statement, which eliminates both pooling accounting and goodwill amortization. In general, companies are directed in certain circumstances to test the value of the goodwill carried for each identifiable reporting unit. For those reporting units, you are supposed to measure the value of the reporting unit (e.g., what future cash flows does your model predict, what interest rates, etc.) and compare that assessment of the unit's worth with the carrying value. If the original acquisition valuation (i.e., existent goodwill) is no longer accurate, you write down the difference as an impairment of goodwill.

The sort of things that indicate a need to test your goodwill include a current period operating cash flow loss in the unit, product changes, increased competition, missed expectations, employee attrition, an "other than temporary" decline in the market value of your company, etc. (or, if nothing like that happens, then goodwill should still be tested annually).

In the end, the FASB is telling the company to value its units much like the way a typical investor would, through traditional valuation techniques like DCF, option analysis or comp deals, to determine what a willing seller and willing buyer would likely agree upon in a free market. This standard gives companies substantial leeway, since we know that a given asset can have a fairly enormous "reasonable" valuation range.

I'm still not quite clear how this all works. AOL can just trade a chunk of itself -- X number of shares -- for another company, and the price of that aggregate chunk of shares gets added to the balance sheet, marked down as goodwill, as a real asset on the balance sheet, whether the market price for the shares it traded is warranted or not. So in the magic of the marketplace, perceived value becomes actual value. AOL can exchange the perception of the value of its stock for an actual asset, in this case, the Time Warner company. Truly, Thoughts are Things.

I think the problem you run into here is the inherent imperfection of any particular method of accounting for stock-based acquisitions, or many other events for that matter. If your goal is to create standards that are useful to users of financial statements, you run across particularly acute problems when you are trying to account for the use of potentially inflated capital.

What are your options? You can ignore the acquisition price (pooling accounting) and simply combine the asset bases of the two companies. This has the effect of providing very low quality information to users about capital allocation decisions of the acquirer, and fails to distinguish well between bargain purchases and gross overpayments. Alternatively, you can force management to subjectively estimate the actual value of their purchase, as is the case now for purposes of the impairment test. But that approach has serious problems. First, what management is going to make an acquisition with stock and immediately announce that, because their stock was overvalued, they paid $1.20 on the dollar for their target? Second, you face serious comparability problems, since acquisition accounting is now in the subjective hands of every individual CEO, who is suddenly facing off against the market consensus as security analyst.

What's important, in the end, is to provide enough information so we can tell what's going on from at least two different perspectives. One, you want to be able to evaluate the ongoing business. To do that, you've got to have access to the return on capital the businesses are generating post-acquisition, regardless of the acquisition price. For example, if AOL is generating 20% ROIC and Time Warner is generating 20% ROIC, but AOL pays 10X to acquire Time Warner, you want enough information to be able to understand that the underlying operating businesses are still earning more than their respective costs of capital, even though the overall business is earning much less once goodwill is included. But you also need a way to measure the company's acquisition decisions. This is important not only to judge management's ability, but also to judge the businesses' underlying, sustainable growth rate in relation to its capital requirements.

That last issue is relevant to the JDSU discussion you initiated to begin this thread. A potential additional consequence of these huge write-downs that can get lost in the shuffle is their tendency to make revenue and earnings growth rates misleading. That is, once one-time charges wipe away the balance sheet evidence of prior acquisitions, it can be easy to forget that some of the current growth is created by previous acquisitions.

Say Steve Case painted a portrait of Ted Turner and sold it to Bill Gates for $60,000. Then AOL bought it back from Ted for $70,000, and they went back and forth until AOL had purchased the painting for $1,000,000. Then it sells the painting to Mary Meeker for a swamp in Florida, with a book value of $20K but great future swamp potential. The swamp market suddenly gets hit with overcapacity, and the Swampdex gets creamed, so AOL writes down the entire $980K in goodwill attributable to the swamp. Alas, two years later, the swamp market straightens out a bit, and although it will never justify having paid an apparent 50X book value, the swamp turned out to be worth around 3X book ($60K), and is generating around $6K per year. The new AOL investor, seeing only the $20K on the books after the long past write down (which is also mentally excluded from ongoing earnings as a one-time non-cash charge), thinks...hey, AOL only needed $20K of additional capital to generate $6K in annual NOPAT, that's a 30% incremental ROIC -- and look at those nice revenue growth rates. Or, to be more realistic, the new investor simply overestimates AOL's continued ability to earn returns on its overall incremental capital and its ability to grow earnings and revenues.

But measuring the future implications of the past costs is thorny and subjective. You may not think AOL will need to pay $1 million for $6K a year (because it didn't really pay that much in the first place), but you also don't necessarily think it is going to earn 30% on its incremental capital. I don't think you can put a method in place that objectively solves this problem, but instead must aim to provide enough information for investors to easily make their own estimates.

Personally, I could do without the impairment requirement, and would prefer companies simply carry goodwill as an indefinite asset, with information on when the expenditures for acquired goodwill were made. (Complications arise when a division is completely sold, but I'll show restraint).