The first-time homebuyers credit juiced the housing market earlier this year. Its April expiration caused every housing statistic to collapse like Rome shortly thereafter. With those hangover effects now wearing off, these next few months are giving us a view of what the unfettered market looks like.

Here's round one: Housing starts -- homes on which new construction began -- increased to a 610,000 annual rate in September, up from 608,000 in August. That's encouraging. From the looks of it, unfettered housing starts appear to be stabilizing at about the 500,000-600,000 level -- the same level they've been at for most of the past year. The cliff diving, in other words, looks likes it's over. U.S. homebuilder confidence is now at a four-month high. Stability seems to be clawing its way back.

But here's what's important: We've stabilized at a ridiculously low level. Look at the long-term view of housing starts, and you can see how far we've fallen:


Keep in mind how much population has grown during the period this chart covers. In 1960, the U.S. population was 180 million; today, it's 307 million, and yet we're building roughly one-third as many new homes.

Frankly, that's good news. One of the chief causes of the housing crisis was that too many homes were built. Now that new home construction has collapsed to unprecedented lows, that extra inventory is getting soaked up. Put your hands together and cheer for that. It's the single most effective stimulus housing can ask for right now.

Think about this. During the housing bubble, housing starts were over 2 million units per year. Yet during that time, household formation was about 1.2 million per year. The net result was a glut of pointless construction, which wreaked havoc on home prices in 2008-2009, and caused misery aplenty for everyone from Lennar (NYSE: LEN) to Beazer Homes (NYSE: BZH) to Citigroup (NYSE: C).

Today, the tables have turned. Housing starts have leveled off at around 600,000 annually, yet a Harvard study recently estimated that "household growth will average about 1.48 million annually in 2010-20. Even if immigration falls to half the Census Bureau's currently projected rate, household growth will still average about 1.25 million annually."

We're building far fewer homes today than households are forming. Put simply, there's no realistic way to support our current population with the current level of housing starts. Something has to give. Unless we undergo a repeat of the Black Death, housing starts will eventually be forced higher than they are today. Much higher.

Warren Buffett, whose Berkshire Hathaway (NYSE: BRK-A)(NYSE: BRK-B) owns several housing-related subsidiaries, commented on this in his 2009 letter to shareholders. There are only three ways to cure our housing agony, he said: 1) Blow up a lot of houses. 2) Encourage teenagers to cohabitate. 3) Reduce housing starts to a number below household formations.

"Our country has wisely selected the third option, which means that within a year or so residential housing problems should largely be behind us, the exceptions being only high-value houses and those in certain localities where overbuilding was particularly egregious," he said.

Here we are a year or so later. Are housing's problems largely behind us? With housing starts appearing to stabilize, I'm starting to think the answer is yes.

And what makes housing in a sort of Goldilocks state is this: Housing starts have leveled out at a rate homebuilders appear to be getting used to, or at least have come to terms with. You see this in homebuilder confidence. Yet the level of starts is so low that there's really only one place for them to go long term: up. That's reason for optimism if I've seen it.

Disagree? Hit back in the comment section below.