RE Mean Reverts to what?

The monthly Case-Schiller numbers are "smoothed"; they are a three month average.

Revert to the mean means revert to its own mean return....

This is a factor returning to its traditional amount.... not a correlation/casuation thing.

Thanks for the responses. That is sort of my point. If one is looking to determine what the long term price appreciation of residential real-estate has been in NY historically, where would that place a current price target assuming the market has become "overbought" and will revert back to that long-term average, if not overshoot it?

One would assume this long-term average is indeed correlated to something; population growth, inflation, salaries/bonuses, etc. which would allow for further refinement of a price target accounting for current local conditions. Case-Schiller does not really have enough data determine this. Everyone speculates on what prices are going to do but I am attempting to determine a more quantitative price target to guide my own purchase decision.

Shiller has tied it to income.... but he's also noted that the long term real return on RE is traditionally near 0 (under 1%).

I would imagine real estate mean reverts to income, price of credit, and ease of obtaining credit. Of course prices over and under-shoot the mean like any other market.

Shiller's book "Irrational exuberance" (second edition) discusses real estate in great detail. I recall that it said that residential real estate is not well correlated to anything - population, etc. If nyc10022 says he indicated it is correlated to income, then I believe him. I just don't recall that. I also recall other sources that indicated that the long-term return on R.E. is 2-3% over inflation, not 0 over inflation. Of course, it depends on the time frame you're looking at - especially if your tying it to household income, since that has increased/decreased in real terms depending on the time frame you're looking at.

If you want to do another interesting exercise, download the "prestige homes" data from this site and compare it to the broader market for each city. You will find that the high-end housing market tends to lag the broader housing market by 1.5-2 years but follows the exact same path.

http://www.firstrepublic.com/lend/residential/prestigeindex/index.html

I am a couple of weeks ahead of you. Been playing with the data for a while trying to find the holy grail of price.
A couple of thoughts:
My understanding is the NYC data is NY metro area single family homes, and I believe does not include condos or co-ops. So, it may not really show the market we are interested in. I don't think NYC apts are off as much as the index is. NJ or Westchester prices seem more in-line.
As a base line, you might model RE price growth with GDP, as that would be a proxy of what people have to spend. So, I think if you pick some point in the early 1990's and trend it up to today at 3-4%, you might find the current bottom.
As you see from the early 90s data, RE prices tend not to fall much. Different price action than other assets. So, I could see prices coming off a bit and just sitting until the trend line catches up in a few years...
Then again, NYC is a limited supply, so it might have some more upward price pressure than the country in general.
No matter what, I doubt it can grow at 15% like it had been for the past few years... so there has to be some type of correction, be it downward or flat.