Housing costs and centralization

In my limited experience, commentators who oppose regional land use regulations like urban growth boundaries (or at least worry about the impact of such regulations on housing costs) tend to favor keeping cities constrained within their 1950 boundaries, while people who favor such regulations tend to favor city-county mergers, revenue sharing, and other ways to essentially merge city and suburb. 

So you might think that the cities with the highest housing costs would be "hyper-elastic" cities that have gobbled up their suburbs, while metro areas with hundreds of little suburbs have lower costs.

But this is not necessarily the case- especially at the high end of the cost spectrum.  The most expensive regions* are a mixed bag- New York, San Francisco, and Honolulu (which arguably is an outlier because it is an island).  New York and San Francisco are classic "inelastic" cities; they are trapped in their 1950s boundaries and surrounded by dozens, if not hundreds, of little suburban municipalities with their own zoning regulations. Cities with centralized environmental regulation (such as Seattle and Portland) tend to be somewhere in the upper middle- not among the least expensive, but far less so than New York and San Francisco.

Does this mean that centralized regulation might reduce housing costs by reducing the power of neighborhood NIMBY groups?  Or does it mean that annexation and centralization just don't matter very much?

*For statistics, see http://www.nahb.org/reference_list.aspx?sectionID=135