This column originally posted on RealMoney.com at 9:03 a.m. EDT. For more information about subscribing to RealMoney, please click here.

With homebuilder stocks down about 30% from their recent highs in April, it's tempting to bottom-fish in a search for quality buys in the sector.

The problem with bottom-fishing, however, is that you run the risk of catching the

bottom feeders

-- you know, the ugly, flat fish with the sideways mouths that suck up all the garbage that accumulates in fish tanks and on the ocean floor.

Ryland

(RYL)

and

Pulte Homes

(PHM) - Get Report

-- two homebuilders I've been flagging as overvalued in the Bricks and Mortar mock portfolio -- are good examples of bottom feeders. Both companies are working overtime to clean up the waste created by the excessive feeding frenzy of the U.S. housing boom, and their stocks should be avoided for now.

Buying either of these stocks today amounts to a bet that housing prices will improve in 2009, leading to the homebuilding industry's return to profitability. There are, however, numerous reasons to be skeptical of such thinking.

The tone from the Bank of America homebuilding conference last week was gloomy. Pulte executives said "pricing power in the industry is non-existent." Several builders talked about how reining in costs was the strategy for 2008. Meanwhile, the price of oil has generally increased costs in the near term, a Ryland executive said.

In its typical volatile fashion, the

SPDR S&P Homebuilder ETF

(XHB) - Get Report

is now down 5% this year, a far cry from the 24% gain it sported in early April.

To show why the sector remains a value trap, today I'm going to focus my analysis on Ryland (although Pulte and other builders likely to survive the downturn sit in a similar boat).

The two main questions for investors to ask about homebuilders are: When will earnings return and how do we value these companies?

Trying to project future profits is a difficult task that relies on three main variables: housing prices, sales volumes and the overall future trend in profit margins for the industry.

Wall Street analysts on average expect Ryland to post a loss of $2.04 per share in 2008 but return to a profit of 19 cents per share in 2009, according to Thomson Reuters estimates. If you look at the range of 2009 EPS estimates -- from a loss of $1.55 to profit of $2.49 -- you realize analysts really don't have much of a clue as to where results are actually headed.

I think Ryland's 2008 earnings loss estimate seems doable, since it is essentially based on current industry conditions, which stink. However, the 2009 consensus number is way too optimistic, since it forecasts an improving housing market and Ryland's return to profit.

Analysts expect Ryland's revenue to remain roughly flat from 2008 to 2009. Implicit in the bullish 2009 EPS consensus estimate is a sharp increase in margins, which makes little sense. Ryland's gross margins in the first quarter were 5%, or 12% excluding impairments. The gross margin in 2007 was -2.5%, or 17.2% excluding impairments. This shows how margins are deteriorating, even if you don't count the one-time land charges.

So long as housing prices remain in a downward trend, then impairment charges will likely continue for the industry, so I expect Ryland's gross margins should stay closer to 5% than 12% in the 2008-09 period.

Selling costs remain the other giant expense that builders are having a tough time managing. Ryland's SG&A as a percent of revenue spiked to 16% in the first quarter. The company is planning to get this ratio down to 10% by the end of the year (good luck with that!).

Assuming housing prices remain under pressure, builders are unable to get cost cuts from suppliers, and selling expenses remain somewhat high (since it is much harder to sell a home these days), then I don't expect revenue or margins to improve at Ryland until 2010.

Here are my estimates for Ryland:

Two inputs go into revenues: the number of home closings and the average selling price of a home. In Ryland's first quarter, closings were down 33% and prices were down 14%. There's no great reason to assume closings or prices will increase until after the glut of inventory on the market works itself out, which probably won't be until late 2009.

How do my generous estimates differ from Wall Street analysts? Sell-side estimates call for $2.91 in EPS in 2010 because of even higher margin assumptions.

While the earnings side of the equation requires significant assumptions, so do valuation multiples. Builders trade under two valuation methods: price-to-earnings and price-to-book value, and it's worth eyeing history to see where they should trade today.

Data from SNL Research shows that Ryland traded at an average P/E multiple of 7.4 during the period of 1999 to 2006 (earnings became negative in 2007, so I've excluded that multiple). During the same period, the stock traded at an average price-to-book multiple of 1.7.

To derive an estimate for Ryland's stock price, I apply a P/E multiple of 9.5 (which matches the peak multiple from the 1999 to 2006 period) to the company's 2010 EPS. I also apply a 1.7 book value multiple to my estimate for that year's book value. Discounting the average of these two stock prices back to today using Ryland's 16.6% cost of equity estimate from Bloomberg, then the stock is worth $21.60 per share.

On Friday, Ryland shares traded around $23.70. The only reason to buy shares today is if you think Ryland will return to profitability in 2009, which I think is a foolish proposition.