http://cityobservatory.org/what-filtering-can-and-cant-do/
The economics are clear and intuitive: the "filtering" effect of new construction happens on a decades-long basis. Building today won't have a downward impact on prices on any timeframe relevant to the people involved in this discussion. So fine, build more...but don't expect it to solve your problems.
Considering that the current housing crunch began in ~2012, it's far more likely that we can impact prices on a timeframe that matters by discouraging the (acute) crisis in demand. We can start by reversing the ill-conceived tax incentives we give tech companies to move to San Francisco.
http://www.bloomberg.com/news/articles/2016-06-01/apartment-...
http://sf.curbed.com/2016/7/12/12163544/san-francisco-rent-p...
That's what's so irrational about this debate: the construction market may be inefficient on the margins, but it's basically functioning as you would expect. The "build up" crowd has a poor understanding of the practical impacts of their proposals, and the "affordable housing only" crowd has a poor understanding of local real estate economics.
Meanwhile, everyone ignores the elephant in the room.
Housing assistance would obviously help, but longer term, you need to start the filtering cycle that that article describes.
What evidence?