Many of you already know that I've long been a fan of Oracle (NASDAQ:ORCL). I'm also a shareholder. I got in because the value investor in me thought the database king's shares were unreasonably depressed by market pessimism over its pursuit of PeopleSoft. I thought there was a compelling case to be made for an "OracleSoft" combination.

A year later, Oracle has accomplished what it set out to do. PeopleSoft is now a part of the company. Oracle's applications business, though still anemic compared with its core database franchise, is now growing. Margins have expanded. Cash flow has improved. CEO Larry Ellison even managed to one-up rival SAP (NYSE:SAP) in stealing retail software vendor Retek (NASDAQ:RETK) with a hostile bid after the German firm had brokered a friendly deal for Retek.

But not all is well if you look at the financials. No, I'm not talking about how the company failed to meet the Street's expectations for sales growth in Tuesday's third-quarter earnings announcement. I'm talking instead about its balance sheet. For the first time that I can remember, Oracle is carrying negative tangible stockholder equity. Yeah, I know, the phrase sounds like it's straight out of NASA. Trust me, it's not.

Instead, it means that Oracle shareholders like me would have no claim on company assets were the database leader to liquidate tomorrow. That's because $9.94 billion, or 42%, of the company's assets are intangible. Subtract that total (because, really, intangibles are about as helpful as a $3 bill to us common investors) from the firm's $9.87 billion in stockholder equity (otherwise known as our claim on the company's fortunes), and you'll find Oracle in the red. More than $74 million in the red, to be exact.

How can that be?
I'll tell you how: When any firm pays a premium to acquire another company, the premium, otherwise known as "goodwill," lands on the assets side of the balance sheet. That's because the assumption, naturally, is that acquisitions are a good thing. But you know what? That's not always true. And when it's not, it hurts, because the memory of acquisitions loved and lost live on in the financial statements. Indeed, intangible assets have to be amortized. That means, not unlike alimony, a company can be stuck with charges that take bite-sized chunks out of earnings for years and years.

Can we expect this to happen to Oracle? Yeah, probably. Intangibles plus goodwill have risen by a total of $9.86 billion since the opening quarter of fiscal 2005. How did that happen? I think it's fair to say that all of it can be attributed to overpaying for PeopleSoft. Consider: As of its last quarterly filing with the Securities and Exchange Commission in November, PeopleSoft had $1.41 billion in intangibles plus goodwill, most of which Oracle likely inherited. Now add to that $8.45 billion in net intangibles, with $6.2 billion of that in goodwill. That's probably close to what PeopleSoft cost shareholders in terms of equity. And that's why the number is negative today.

In the end, this process can look a little like spending your life savings for a wedding ring. If you've found the right person, the lifetime of happiness you achieve will more than compensate for the hole created by emptying your bank account. But if not... well, you know.

There's more to watch
It's not as if Oracle won't recapture positive equity. It absolutely will. And the chief benefit of "OracleSoft" -- increased cash flows from higher-margin maintenance revenue -- may already be starting to take effect. But not to the degree I'd like. Indeed, my estimate of owner earnings for the trailing 12 months comes in at $2.9 billion. That's only 5% better than Oracle's full fiscal 2004 total.

Moreover, the current quarterly earnings statement shows only $132 million in amortization. That's $528 million annualized. At that rate, it would take more than 18 years to pay down the intangibles on Oracle's balance sheet. But don't expect that to happen. It's more likely that the company would write down some or all of its intangibles in a hit to earnings. Unfortunately, it's impossible to know when, if ever, that would occur.

Have I lost faith?
Not at all, though I remain cautiously optimistic. My main source for optimism, surprisingly, also stems from the balance sheet. Oracle's net cash and investments total more than $9.1 billion. Subtract total debt of $7.3 billion, and $1.8 billion remains. That might not seem all that great till you see where Oracle placed the debt -- under short-term borrowings. That means the company expects to repay the debt it incurred for the PeopleSoft acquisition within 12 months, or maybe even sooner. And that, to me, signals enormous confidence from management in the cash flows that "OracleSoft" is capable of generating.

The most interesting question, however, is whether Oracle remains a buy or hold for me based on valuation. Let's take a look. I'm using Oracle's earnings before interest, taxes, depreciation, and amortization -- otherwise known as EBITDA -- because it removes charges related to the acquisition and, therefore, should provide a fairer historical comparison. Now, assuming analysts are correct that Oracle will grow earnings 26% this year and 22% next, then last year's $4.1 billion should translate into $6.3 billion by the end of fiscal 2006. Then factor in the Street's estimated 11% long-term growth rate for the ensuing three years, and you get $8.6 billion in EBITDA by the end of fiscal 2009. That's heady growth, but there's more to valuation than calculating future earnings. Stay with me.

Oracle currently trades for roughly the same enterprise value-to-EBIDTA multiple as does close rival SAP. Indeed, according to Yahoo! Finance, Oracle comes in at 15.4, while SAP checks in at 15.5. That makes sense, of course. First of all, the companies are close in size. They each also generate sizeable cash flows. The main difference between the two is in margins, where Oracle currently has an advantage. But given what we've outlined above, it's probably wise to say that Oracle is the riskier of the two stocks. We can account for that risk by discounting the EBITDA multiple from 15.4 to, say, 13 and then apply our projected 2009 total against it. Do the math, and you end with a projected 2009 enterprise value of $111.8 billion, or a 76% gain from here. That's a 12% annualized return.

Is that a guarantee? Of course not. But it's encouraging to know that the odds still favor my investment in Oracle. That's all you can ever ask for when it comes to stocks.

Fool contributor Tim Beyers owns shares of Oracle. To see what other stocks are in Tim's portfolio, check out his Fool profile . The Motley Fool is investors writing for investors and has a disclosure policy .