For the 10 years prior to Jan 2000, both the Case-Shiller price index and the OFHEO price index indicate that housing prices increased about at the rate of inflation. If real housing prices had followed inflation since then, they would have been 47% below their actual peak in summer 2006, according to Case-Shiller. If real housing prices had followed inflation through July 2008, they would have been 27% below their actual values in July 2008. In other words, as of July 2008 housing prices had 27% more to fall in order to reach the real value they had for several years prior to the “bubble.” To put it yet another way, as of July 2008, the house price decline was not yet even half way complete.
Capital theory suggests that housing prices have a long run relationship to construction costs, not necessarily relative to the overall CPI. The PPI for residential construction increased about 10% (depending on which components are considered) relative to the CPI over the years 2000-2008, after decreasing somewhat relative the CPI over the prior 10 years. If today’s construction costs are expected to remain, then July 2008 housing prices had only 20% more to fall. On the other hand, it can be argued that construction costs themselves will fall once the energy price spike recedes. To a rough approximation, this means that, as of July 2008, housing prices had fallen about half way from their peak to their ultimate real value. This process is closer to completion (more precisely, closer to being evident in actual housing market transactions) in some regions than in others, but perhaps by summer 2009 it would be complete in most of the nation.
The basic logic of this approach implies that these basic trends – namely that housing prices had increased much more than construction costs – were readily discernible in 2006, and before. So where was I with the warnings? First of all, it is perfectly consistent with capital theory that housing prices would temporarily exceed construction costs because, in the short run, there is a limited supply of houses until the construction industry can catch up with demand. Persons who insist on purchasing a house when prices exceeded construction costs were, in effect, (probabilistically) paying for the privilege of owning a house earlier rather than later (when the construction industry would have produced all of the houses that were demanded). Second, no one knew for sure when supply would catch up with demand. In, say, 2005, market participants might have rationally forecast that prices would remain above construction costs through 2010 (by which time the persons purchasing houses then would either have grown into their mortgages or have sold their house without a significant capital loss) even while they understood that eventually house prices would fall back to construction costs. That forecast was ultimately wrong, but hindsight is 20-20.
[Added Oct 10: Two days after I wrote this, the Wall Street Journal had an article showing a graph of real housing prices. You can see there how real housing prices are still above historical levels.]
[Added Oct 11: Whenever I have a tough question about urban/housing economics, I ask Professor Glaeser. You can learn from him directly on his blog. For example, he says that housing prices will keep falling. I don't know whether he agrees with my forecast as to the amount of the reduction.]
 The OFHEO index shows the peak in 2007, which tells me that it is not suited for high-frequency analysis. Relative to January 2000 (and in nominal terms, the Case-Shiller peak is 30% higher than the OFHEO peak).