This is a copy of something I wrote on Calculated Risks page. I thought it at least covered an area that was too important to lose in a comment post, so I have put it here for save keeping. Those who don't go to Calculated Risks page, you are missing some of the most important information about the economy and the housing bubble. I am going to add a little more to the bottom of the page because I didn't finish my thought.
Risk, you have a great page and I have used it for at least 2 years for reference of data you have accumulated. I have been around housing in some capacity most of my working life, having been both a Realtor and a mortgage broker. I have also been though a housing bust, the 1980's Texas bust being one of the worst seen in the USA for a long time prior to the current mess. In any case, your data doesn't support what the recent data is showing.
The big deal is that home sales really never exceeded 4 million prior to this bubble on an existing basis. I do know from your data and other data I have mined from the government that new home sales were at a record of 819K until somewhere around 1997 from which we didn't see sales that low again until 2008. I believe your chart on the story above says more than you are mining out of it. I will elaborate after I mention the obvious.
You are using a ratio of existing sales to new sales to I believe net out distressed sales. The last time housing really went bust was the early 1980's interest rate shock. It is also the only time new home sales were at a bottom compared to this recent action, dropping to 401K. Existing home sales in that recession/ bust were well below 3 million. You might note that existing home sales have remained above the prior record, despite the worst economy in a long time. This doesn't support the idea that home prices or building are falling because of a bust, but because of a collapse in market fundementals. There has been speculation all the way down. The population argument doesn't hold water because the real game is how many new households are needing houses and the boom generation was providing as many as ever. Thus the 4 million/800K of 1978 was a peak. There was also a speculative bubble going on in 78 to support such a sales figure. The bust that followed dropped sales to 50% on both accounts. Thus our totals here should be more in the 3.5 million range, not the over 5 million range we have seen as a bottom.
The other matter is the new home sales figures and how they stayed so high for so long during the 1997 to 2007 period. Your chart "comparing peaks and troughs for starts, new home sales and residential investment shows the truth. Though you can say a bottom might be in because new home sales are bouncing, remember that not all markets were glutted. They were building roughy 50K in units here in the DFW area before 2007 and scaled back early. I think DFW could probably consume 25,000 units this year, which would maybe by itself make up the change in construction last month. Other areas aren't coming back.
If you look at that chart, don't only look at the peak, but look at the duration of the climb and the size of the top. A big top means a big bottom and we are going to see something that is totally unexpected. Remember, prior overconstruction was always corrected every few years and only the 1972 era and 1978 era peaks lasted over a very short period of time, 2 and 3 years respectively and followed by plunges. it was the plunge that created the following peak and the 2 peaks played themselves out in the 1980's where we only saw a more moderate peak.
This is evidence that the supply of housing is so large that it will take years to work off the excess. Also, the high existing home sales figures tell me the speculators haven't quit speculating. I don't care to listen to NAR statistics as they have been lying for years and using bubble statistics to hide the fact that the bubble hasn't left. This is the true nature of the problem as I see it.
Thus we are going to see one of two things happen. Either we are going to see a rebound and another speculative bubble (the government has an interest in encouraging and recreating a bubble and would do nothing to stop it, including making FHA a subprime financing outfit) or we are going to see this rebound fizzle to lower prices and lower sales. We spent 5 years above the 5% GDP point in housing, which equated to all the time in the 30 plus years prior. We also spent 11 years in record annual sales territory, peaks that had only lasted 2 years prior. Thus, we are looking at a lot of years of surplus housing construction to wipe out.
If this does become a new bubble, I expect it to either bankrupt the US government or force them to drop FNM and FRE and let the market take the losses. If it doesn't, it will force the knife catchers to liquidate or join the groups of foreclosed. In any case, the economy won't be the way it was for a long time.