Great contributions from everyone….all of us trying to put in 2D text what should be in a multi-factor model of great complexity (and in fact impossible to build).
One analytical method that must now be deemed impossible is any kind of model for mortgage supply and demand (which feeds into the house price model). The impact of structural change in the supply of mortgages and their types has been so extreme in the last five to ten years (of which the arrival of ARMs in their various guises is only one of many factors) that no sensible time series could be constructed now in my view. The chaotic effect of the current credit crisis (and legislative reactions as well) will just muddy the regression waters some more.
That is partly why I instinctively favor looking at the rental/buy decision and landlord investment buy/sell decision in simple terms to try to feel for inflexion points in the residential real estate market.
As various contributors have pointed out , we cannot be sure what rent levels will be. Rents should rise with repos (as happened in the mid 80s) varying by area as well, but a recession would tend to restrain growth of income and hence rents as has been pointed out . Hope people feed back on the rental market over time .
