Rising home prices are a signal to builders: As housing gets more expensive, it's time to create more of it. Then as new buildings rise, the competition for scarce housing falls. And that keeps prices and rents from going crazy.

This cycle works for the most part around Charlotte, or Atlanta, or Indianapolis. But it's clearly broken in high-cost coastal cities like San Francisco, New York or Boston.

The housing markets there, as economists say, aren't very elastic. New supply doesn't respond well to rising demand. And so, in metropolitan San Francisco, home prices have spiked about 275 percent over the last 20 years. In Boston, they're up nearly 150 percent.

What explains, though, the difference between these two sets of places? Land is obviously part of the problem. San Francisco and Boston, hemmed in by water, have only so much of it left to build on. But while some of the obstacles are natural — the Bay Area's got those hills, too — others are manmade.

Critics have long argued that we make it much harder to build in some places than others. We outlaw denser buildings with single-family zoning. We empower neighbors to block new housing. We leave projects in limbo for months at a time, as officials decide whether or not to approve them.

This last factor — building permit delays — turns out to be a significant one, according to an analysis by Trulia chief economist Ralph McLaughlin. New York, Boston and San Francisco all take much longer on average than Atlanta or Charlotte to grant building permits to developers. That means new housing takes much longer to create. Building it costs developers more. Less of it ultimately gets built. And what does get built doesn't respond very quickly to pressure in the market.

In this chart, based on McLaughlin's analysis, metros with longer permit delays tend to have less elastic housing markets (the higher the number on the y axis, the more readily supply responds to rising prices; an elasticity of 1 here means that every 1 percent increase in housing prices is associated with a 1 percent increase in housing stock):

These delays count just the time from when builders apply for permits — whether they want to rezone or subdivide land, or just receive permission to build on it — to when those permits are approved. So the full design and construction process actually takes much longer.

Zoning laws themselves have gotten a lot more attention among critics who say high-cost cities don't build enough. But McLaughlin's work suggests that these permit delays may be even more important in understanding what's holding back new housing supply. Bureaucracy, he argues, matters even more than zoning, although the two are closely connected (and long delays may stem from the same political opposition as strict zoning).

"It’s this delay that really seems to be explaining the difference between why some places are elastic, and why some places aren’t," McLaughlin says.

The delays add uncertainty for developers, which makes a project riskier. Some builders may be turned off by the prospect and never propose projects in the first place. Year-long delays could also boost the chances that a larger project gets revised in scale, ultimately producing fewer new housing units. The result of the mundane-sounding permit process ultimately affects how much housing gets built and, in turn, what housing costs you.

Nationwide, housing supply is less responsive to rising prices than it used to be, meaning that elasticity is at a historic low, McLaughlin writes. But places like San Francisco and Boston today are even well below that low point.