12,000/year is obviously too many homes to add every year, and the 4,000-5,000 years seem to happen when we need to absorb the excess. I would guess that if the cycles would level out some we’d probably split the difference in an average year and come up with a smaller variance between the fat and lean years. Say, 6,000 – 9,000 per year.
Of course, if these cycles were to ever flatten out that would basically remove the speculators from the market. There wouldn’t be any big short term gains to be made.
I think the effects of the internet age could result in flatter swings in the market, especially after this party/hangover cycle is completed. Enough blogs have popped up, including some data-centric ones like this one, that more people are going to realize that they could have seen this coming had they just taken the time and effort to look.
Unlike the stock market, which in some ways really is rigged in favor of the insiders, the RE markets are basically transparent. Especially so when considering the rise of public access to readily available data from multiple sources. The average buyer can now see about 95% of what’s happening from the outside, independent of listening to what truths or untruths an insider might tell them.
I think more and more buyers are going to tune into the fact that the they can tell which way it’s going to go (although not necessarily the timing) if they just watch the numbers. Couple that with the relative illiquidity and the slow pace in which changes can occur, and I think it’s possible the RE markets could eventually stabilize with movement only occurring when its economically warranted.