1. This seems to have caught people by surprise and resulted in a significant jump in the stock market. Should it be surprising if you understand housing markets?
No! Why - because there's not one housing market so we don't need to wait till all the inventory is run down in say Californian single family homes before we start building condos in Denver. For more details see my blog from January. IT is worth noting the jump in multi family housing starts was mostly in the north east of the country - an area with smaller boom and bust in housing. http://reflexivityfinance.blogspot.com/2009/01/misconception-2-housing-market.html
Probably not. Housing starts have fallen to lowest on record. Even worse than the graph suggests if you factor in population growth. They need to increase 100% from this point to get back to a position where they would be considered terrible in any other postwar recession. So even if we get a turnaround it isn't going have a big dollar impact because it will still mean not a lot of dollars going into building. In addition dollars spent are related more to completions then starts and obviously housing completions lag housing starts considerably. Housing completions are still going to be going down for the next 3 months or more and then increase back to the level where we are now for the next few months. This is because housing starts have been going down rapidly for the last 3 months.
So in summary any rebound isn't going to start to translate into dollar impacts for at least 3-6 months and even then the rebound in dollar terms is going to be small because we are starting from an extremely low base and because many markets have significant inventory, low prices etc that will stop the rebound being a broad based one.
For more detail on the figures graphed above see http://www.calculatedriskblog.com/2009/03/housing-starts-rebound.html