The housing market is a mess. Prices might look cheap, but probably aren't finished falling. They've basically fallen back to historic averages. The washing out process that happens after bubbles burst usually takes assets far below average. Euphoria doesn't just end. It turns to hatred. Fool me once ... you know.

Still, housing is going to recover. It might be years from now, but it's going to happen. You can guarantee it. Those are strong words, but the force that will drive recovery is even stronger.

A driver behind the housing bubble was that we built too many homes. For most of the last decade, new home construction grew far faster than population growth or household formation. This wasn't surprising. It didn't even matter that there weren't enough new households to fill all the new homes. Investors would happily purchase a new home -- or 10 -- and let it sit empty. The sole point was to let it appreciate and sell it at a profit.

But today we're in the exact opposite position. New home construction is comatose. It doesn't make any sense for contractors to keep building. "The construction industry is dead right now," Yale economist Robert Shiller said last month. "They don't see any profit in building homes at these prices." That's the strongest force you could ask for to keep new home construction glued to the floor, and it means homebuilding in relation to the size of the economy is nowhere near normal levels.

Here's a good way to visualize this:

Housingstarts

Source: Federal Reserve.

The higher the line on this chart, the higher the odds are that we'll need to build more homes in the future. Obviously, we're in uncharted waters right now.

Some more figures to throw at you: In 2005, about 1 million new households were formed while more than 2 million new homes were built. That created an inventory glut. Today, about 1 million new households are still being formed, yet housing starts (new home construction) are running at about 560,000 per year. That's eating up excess inventory -- quickly.

That trend should stand, too, thanks to the U.S.' strong demographic and immigration trends. Household formation over the next decade should average nearly 1.5 million per year, according to the Joint Center for Housing Studies at Harvard University.

Simply put, the number of homes being built today cannot, and will not, support population growth. Excess from the housing bubble is being removed, and will likely be mostly cleared out in another year or two. After that, one of two things (or a combination of both) has to happen.

The first is home prices will rise. It's simple. At the rate we're going today, demand will not only catch up with, but surpass, supply in the future. And that's a fairly safe forecast to make. Unlike demand for stocks, gold, or bonds, a minimum level of demand for housing can be projected rather safely based off household formation -- itself a relatively safe forecast being based on demographics.

Once prices start rising after new households clear out excess inventory, homebuilders will regain the incentive to build. They'll have the demand to do it. I think that's the most important point to consider when looking at companies like Pulte (NYSE: PHM), NVR (NYSE: NVR), and KB Homes (NYSE: KBH). If these companies can stick it out another two years or so, business is practically guaranteed to improve. Quite substantially, too. A lot of these companies are being priced based solely off of today's derelict housing market without respect for the inevitable construction rebound. Could be an interesting sector if you've got the patience. Most investors don't, and that's why these companies trade where they do.

More importantly, no modern U.S. recession has ever fully recovered without the help of housing. Even in non-bubble times, housing is one of biggest economy drivers for the simple reason that it's the largest investment most people ever make. Its downfall is a major reason our economy is still in a funk. And we'll probably stay there for a while. But if there's light at the end of the tunnel, it's from the chart above. The inevitable rebound in construction is one of the most positive future indicators we have today -- even if that future is a full two years away.

Again, the housing market is a mess and will probably get worse. Count on it. But everything is in place for a recovery. All markets are cyclical, and when you look at the numbers it's hard not to think we're near the bottom of this cycle.