Yes, a big problem with predicting what effects large building projects will have on a specific area within a metropolitan area is that they impact both supply and demand, not only supply. At the level of a whole metro area, demand is impacted more by the general economy, but whether people choose to live in Manhattan vs. Brooklyn vs. New Jersey (or SF vs. Oakland vs. San Jose) is intertwined with type of housing available, perceived neighborhood features, etc. You can model the net effects of an intervention, but the models are somewhat sensitive to assumptions.
I don't think that SF has a problem attracting demand. Neither do Manhattan or Brooklyn for that matter -- both are prohibitively expensive at this point unless you're a millionaire. Manhattan demand is a bit different in that it's driven globally: a Manhattan apartment (and one in London, and one in Hong Kong) is seen as a must-have among the newly-wealthy Chinese upper class. You know how there are now more millionaires in China than in the US? They all have a Manhattan apartment that they rarely use.
SF doesn't quite have that issue yet -- but IMO housing prices are a big reason why startups are looking outside the bay area more seriously these days.
You're right, but given the prices in the SF market, there are only so many people able to buy a $1 million+ home. It wouldn't be hard to increase the supply to the point where prices start to decrease significantly, especially if you start to build a decent inventory of non-luxury units (not everyone needs or can afford a doorman).
