I got some push-back on my claim that tight zoning restrictions in some jurisdictions helped spark the housing bubble. Below I set out some evidence to support my claim.
Let me first explain in a little more detail how I think the bubble developed:
1. In 2001 or thereabouts, tightly regulated markets like Miami and California experienced a bump in demand. (Yes, the supply of housing in Miami is fairly inelastic, as I show below the jump.)
2. Thanks to tight zoning restrictions, the housing prices in these markets spiked.
3. The bubble did not necessarily start with this spike in prices. A surge in housing prices in a tightly-zoned market might be perfectly rational because rising demand also pushes up rents.
4. Although the initial surge in prices might have been justified by market conditions, the rapid rise in prices triggered the bubble: homes began trading well over their fundamental value, as measured by their rental value, as more and more investors entered the market to exploit the rapid rise in prices.
5. The bubble could not have happened without the securitization of mortgages and lax lending standards; these allowed everyone and his dog to get into the game.
6. Bubbles by their nature are contagious. The California bubble spread to the much more elastic markets of Phoenix and Las Vegas. The bubble in Las Vegas lagged the California bubble by at least a year; the Phoenix bubble lagged California by at least two years.
7. The bubble burst. Naturally, the crash has been worse in places like Phoenix and Las Vegas where developers could quickly add housing.
Did the tight land-use regulations in California "cause" the bubble? If I can mix my metaphors, I think tight land-use regulations were kindling. The initial surge in demand was the spark. And the monopoly money flooding the market acted like lighter fluid which banks kept squirting on the fire.
Now to the evidence. There are three pieces, I think, that support my story: (1) the inelasticity of markets like California and Miami markets where the bubble began; (2) the timing of the bubbles in Las Vegas and Phoenix; and (3) the failure of many markets to follow suit.
Inelastic housing markets
The Census Bureau tracks the annual change in housing stock in each U.S. county. The Case-Shiller index (xls) tracks the year-over-year change in prices of existing homes. Putting these together gives us a rough idea of a housing market's responsiveness to changes in home prices.
It is hard to overstate just how unresponsive some markets are to changes in home prices. Let's take Los Angeles:
The supply of new housing in Los Angeles is remarkably inelastic. Between 2000 and 2007, Los Angeles County's housing stock grew at a glacial rate -- between 3/10ths and 6/10ths of a percent per year. This, during a period when housing prices were rising between 10% and 33% per year.
To calculate the elasticity, we divide the percentage change in quantity by the percentage change in home prices. It takes time for developers to respond to price changes, of course, so we have to include a lag. If we use a one-year lag, we get an average elasticity of .03 between July 2001 and July 2007. Interestingly, we get the same elasticity if we use a two-year lag. (Note that a two-year lag consistently yields elasticities in the 0.03 range).
This mean that in Los Angeles County a 10% increase in home prices spurred only a 0.3% increase in home supply. Home supply in Los Angeles County is almost completely unresponsive to prices. It should be no surprise, then, that any change in demand quickly inflated home prices there.
The Miami market is slightly more responsive than the Los Angeles market, but only slightly so:
If we assume new construction lagged price changes by one year, we get an average elasticity of .121. If assume a two-year lag, we get an average of .157. Again, this means that, in the Miami-Dade market, a 10% increase in home prices yielded only a 1.2-1.6% increase in supply. This is quite inelastic.
The timing of the bubbles
If the housing bubble had been confined to tightly-regulated markets, this would be a pretty straightforward argument. But bubbles also developed in the Phoenix and Las Vegas markets where housing supply is relatively elastic.
The timing of the bubbles in Phoenix and Las Vegas is suggestive. This chart depicts the changes in the Case-Shiller index for Los Angeles, Miami, Phoenix and Las Vegas between July 1998 and July 2008.
(This index is calibrated separately for each market; "100" in Phoenix is not necessarily the same thing as "100" in Los Angeles.)
As is evident from the chart, prices in Los Angeles and Miami began to take off between 2001 and 2002. The increase accelerated between 2002 and 2003. The bubble in the Los Vegas market did not begin to form for another year; the bubble in Phoenix followed a year after that.
Why the lag? Bubbles are fundamentally psychological events, a kind of mass delusion. Which makes them contagious. The lag is at least consistent with the theory that Las Vegas and Phoenix caught the bubble from California. And I doubt it is a coincidence that bubbles developed in two markets bordering California; both markets likely saw spillover from frenzied California investors.
Some markets did not experience a bubble
The final piece of evidence is that there were plenty of markets -- Dallas, Charlotte, the midwest -- that did not experience a bubble. These markets were also flooded with funny money and subprime buyers, and many have experienced a high foreclosure rate. Any story about the bubble needs to explain why bubbles occurred in some markets rather than others. And I don't think there is an explanation if one ignores the reason prices spiked in California in the first place.