This article is part of our Rising Star Portfolios Series.

If I had to guess, I'd say that 99% of discounted cash flow models are wrong. Here are four reasons why I think that's a reasonable estimate, in order of importance. By being aware of these flaws, you'll not only be more skeptical about trusting fancy-sounding DCF valuations, but you'll also be able to fix your models yourselves. As an added bonus, I'll also provide an alternative approach to valuation. Let's get started.  

1) All cash flows are remitted to owners? Not likely
This is the glaring flaw. Calculating free cash flow is the right thing to do, but after hunching over a keyboard for hours, when analysts finally arrive at the FCF estimates (and discount them back), they simply stop. Hurrah! We have the magical free cash flow number, and therefore the value of the firm. Period, full stop, we're done.

But analysts overlook what happens to that cash flow. It doesn't get mailed to you in a check. It stays inside the company until management does something with it. And management can spend that cash on acquisitions, stock buybacks, unproductive new projects, or even an increase to their salaries.

What is management's track record in this area? Most would say it's pretty poor. Most acquisitions (and certainly the largest ones) add little value, and stock buybacks often take place at the wrong time. In 2006 and 2007, right before the financial crisis, buybacks soared.

So how can you be confident in a valuation when the key driver is a slippery target? It's like holding a wet fish! This is one reason why stocks like Dell (Nasdaq: DELL) have remained dormant for years. On paper, the cash flow looks good, but it's been poured into debilitating stock repurchases that have done absolutely nothing for shareholders. Dell is a prime example of how most DCFs fail to deliver.

If you have to construct a discounted cash flow model, make sure to factor in what management will likely do with all that cash.

2) There is never a down year
I've never seen a free cash flow model with a down year. In fact, in almost any forecast across industries, you rarely see down years. But just because you can't predict when the next downturn will come doesn't mean one won't come.  If your model doesn't include at least one down year, you're meaningfully overstating the resulting valuation.

3) Stock-based compensation expense is usually excluded
Stock-based compensation works out well from a reporting basis; although it's included in earnings, it gets added back into operating cash flow, since it's not technically cash. But that overstates the true cash earnings potential of a firm; without stock-based compensation, companies would have to pay an equivalent amount of cash compensation to retain employees (roughly speaking).

4) The terminal growth rate is usually too high
Typically, folks model out 10 to 20 years, then apply some sort of growth rate for eternity after that. Often, it's 3% -- the growth rate of the economy. But it can't be 3% for all companies, simply because growth in an economy comes from new firms, not existing firms. Old firms slow down and eventually collapse, while new firms excel and pick up the slack. A more proper terminal growth rate would be a negative one.

Another downside of a DCF model
A model not only takes a long time to build, but also even more time to maintain. Imagine cleaning up a broken jar of molasses -- a cumbersome morass that takes forever to manipulate. Now imagine that the jar keeps breaking every three months.

In the process of updating, analysts spend way too much time fretting over the little details that will make their model "perfect," while ignoring the bigger issues -- management, return on capital, competition, and other intangible factors that don't appear in a model. 

My approach
To avoid all these pitfalls and hassles, I prefer to stress-test an annual free cash flow estimate. I want to know what I can expect year in, year out, on average for a company. It takes much less time than building a time-consuming, inaccurate model. And in my opinion, it's much more precise and conservative, because you're factoring in points one through three while avoiding point No. 4. Once you have the number, you can simply compare it with the current price to determine the yield.

In fairness, this approach doesn't work as well for fast-growing companies. But for many, it could reduce the chronic overstating of cash flow that plagues discounted cash flow models, giving you a more realistic valuation and preventing you from overpaying for a stock.

For example, with a moderate grower like L-3 Communications (NYSE: LLL), I forecast $1 billion of free cash flow -- an attractive 11% yield on the market capitalization of $9 billion. I wouldn't be nearly as confident with an estimate for NVIDIA (Nasdaq: NVDA), for example, because it's growing quickly and its earnings are less predictable.  

Find a salt block
To conclude, if you ever see an analysis based on any sort of "discounted free cash flow," find a salt block and lick it, because a grain simply won't do. Even if the analysts have overcome the standard DCF pitfalls, they probably haven't even thought about mine. As Dell shareholders know, the right approach to sizing up free cash flow can mean the difference between rags and riches.

This article is part of our Rising Star Portfolios series, where we give some of our most promising stock analysts cold, hard cash to manage on the Fool's behalf. We'd like you to track our performance and benefit from these real-money, real-time free stock picks. See all of our Rising Star analysts (and their portfolios).

Andrew Sullivan, CFA, owns options on L-3 Communications. The Motley Fool owns shares of L-3 Communications. Try any of our Foolish newsletter services free for 30 days. We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy.