(Fortune) -- For the most accurate, up-to-date view of where America's housing market stands today, the best source is Mike Castleman. This tall, lean Texan rancher who fits the Hollywood image of a retired bronco buster is the founder of Metrostudy, a firm that dispatches researchers to inspect 4.5 million building lots every ninety days in the busiest cities for homebuilding, from Washington, D.C., to Chicago to Phoenix to Sacramento.
Metrostudy keeps a constant watch over two-thirds of the housing construction in America for its clients, chiefly America's top homebuilders. So to learn if his latest numbers point to the robust recovery we've long awaited, and that strong recent numbers in March -- new home sales rose 18.7% over March of 2009 -- might foretell, I called Castleman at his 460 acre cattle ranch in Dripping Springs, Texas, where the deer and wild turkey roam.
Castleman had just finished compiling the numbers from his latest inspections the evening before. His on-the-ground research shows two major forces shaping the housing market: one favorable, the other sobering. The combination points to a market bumping along at a steady but plodding pace.
On the positive side, Castleman says that the burden that wrecked the market, the excess inventory of both finished homes and houses under construction, has finally evaporated in places from Indianapolis to the California's Inland Empire. The negative is that the gains didn't come from what the market needs most: a resurgence in sales showing that Americans are getting optimistic and hence more willing to buy homes again.
"This is no comeback"
"Sales simply stopped dropping after they hit an extremely low level," says Castleman. "This is no comeback. There is no reason for celebration. No matter what you hear about 'job creation', the economy remains stagnant." The improvement, he adds, comes from a drop in housing production that's even deeper than the staggering drop in sales.
Still, Castleman's analysis shows that housing is finally turning one crucial corner: For the first time in four dismal years, demand for new homes is outstripping supply. So builders are increasing their production, with strong confidence the houses will sell. Amazingly, homebuilding has been so weak, for so long, that a few markets are even suffering from shortages. As a result, says Castleman, prices will actually rise in some places, including Washington, D.C., and Indianapolis. And prices should be stable in most others, such as Houston, Naples, Florida, Charlotte and Denver. That's no sterling comeback, but it is progress.
Let's follow Castleman's explanation of how housing sank so fast, and the forces that are restoring balance: The biggest threat to the market is excess inventory, measured as total finished houses for sale, plus those under construction. In a decent market, Castleman says it should take builders around six to seven months to sell their stock of completed homes and the ones they just started building. Why is any time period longer than that one a danger?
"It usually takes homebuilders four months to complete a house," says Castleman. "They can usually carry it for two months without a problem. But four months or more of paying big interest and insurance costs puts them under tremendous strain, and forces them to slash prices." That's what happened in the crisis.
How the excess housing inventory piled up
Starting in 2004, builders began erecting far more houses each quarter than they were eventually able to sell. In the areas Metrostudy covers, housing starts rose to around 800,000 a year -- 300,000-plus more than the market could absorb. "They over-built because of the investors who flooded into California, Phoenix and Las Vegas, signing contracts on ten houses at a time," says Castleman. "We'd preach to clients not to sell to investors, but they couldn't help themselves."
By early 2007, the builders had an incredible 10 months of inventory either for sale or in the pipeline, when tens of thousands of investors and other buyers simply walked from their contracts as rates rose and jobs disappeared. "The price war we predicted happened," says Castleman. "It cost the homebuilders $40 billion in market cap, and cost smaller builders their businesses."
To stop the bleeding, builders sharply curtailed their production. "In two quarters, they cut their output by 40%, from 200,000 houses per quarter to 120,000," marvels Castleman. "We've never seen anything like it." The problem was that the builders had poured so many foundations during the boom that home construction already in the pipeline kept flooding the already oversaturated market. The stock of unsold homes remained huge until the start of 2008. But the market was working: The number of sales, though extremely weak, remained far higher than the even feebler number of housing starts. So that excess inventory steadily eroded during 2009.
One turnaround leads to another
That brought a dramatic change in the first quarter of 2010. "All of a sudden," says Castleman, "demand stabilized at around 60,000 units a quarter, and stayed there." To be sure, that's an extremely low number. Castleman estimates that in a normal economy, around twice that many units should be selling. But housing starts, by the end of 2009, had dropped to an astounding 30,000 a quarter, an extraordinary one-fourth of what the market normally requires. "Now builders are seeing, for the first time in years, that they don't have enough houses either finished or under construction to meet demand," says Castleman. Result: In Metrostudy's markets, housing starts spiked by 44% in the first quarter of 2010, versus the same period last year.
So it's the steep fall in homes for sale -- now, and for the future -- rather than the return of the buyer, that's restoring balance, and finally supporting prices. The picture is extremely different from market to market. In Sacramento, Salt Lake City, Minneapolis and Washington, D.C., the inventory now stands at a healthy 6 to 7 months of supply. So in those cities, prices should remain firm or even rise from their current levels. By contrast, Los Angeles, San Diego, Las Vegas and Phoenix still suffer from more than ten months of inventory, meaning prices are likely to keep dropping.
How Texas and D.C. housing are alike -- by being completely different
The two healthiest markets, says Castleman, are Texas and Washington, D.C., for totally opposite reasons. In the Lone Star state, prices never spiked, so builders weren't lured into turning out a flood of new homes. Texas mostly avoided the crash. The reason is simple: Builders can put up new houses in just a few months in Texas because the zoning laws are far less restrictive than in say, California or Arizona. Texas has also managed to preserve far more jobs than most states.
In the Maryland and Virginia suburbs around Washington, the restrictions on building are among the most onerous in the entire nation. As a result, only a trickle of new housing is coming on the market, despite the good economy and strong job growth in the Washington area. Castleman actually thinks prices will rise in the Washington suburbs: "The area will see substantial price increases by the end of the year," he says. And he even sees slow, modest increases in his home state of Texas.
Castleman does think the worst is over. "The good news is that builders will need to build a lot more houses than last year to keep up with demand. That will help the economy by creating jobs." The big leap, he says, will only come when robust job growth and rising consumer confidence bring back the buyers. To get the first view of when that happens, we'll keep checking in with this Texas rancher who is charting the new trends in housing.