Bubbles again

I did not mean to suggest in my post on bubbles that the Great Panic of 2008 could have happened without restrictive land-use regulations, exotic securities, lax lending practices or the other likely culprits.   I did mean to suggest that bubbles don't need bad regulation, sleazy lenders, or loose monetary policy.  They develop spontaneously even in simple financial markets where the participants have perfect information.

Although bubbles may be inevitable, they don't normally cause the collapse of the real estate industry and credit markets (and who knows what else).  I'm sure that 100 years from now economists will still be debating the causes of this bubble.

Too much leverage and bad risk models bear at least some of the blame.  But tight land-use regulations in California and the northeast played a role, too.  Economists Ed Glaeser and Joseph Gyourko demonstrated in this May 2006 paper that restrictive land-use controls are associated with more volatile home prices:

In today's still solid housing market, we may have forgotten the historical correlation between housing price growth in one period and decline over subsequent periods.  For example, Joseph Gyourko and I have found that if an area has a $10,000 dollar increase in housing prices during one period, relative to national and regional trends, that area will lose $3,300 in housing value over the next five-year period, again relative to national and regional trends (Glaeser and Gyourko, forthcoming).  Such housing cycles occur almost everywhere, but the dollars involved are far bigger in metropolitan areas with restricted housing supply such as many parts of California, New York, and Boston.  Illustratively, booms and busts in the Atlanta region have been relatively modest while in Boston, the last boom was followed by a 30 percent drop in housing values between 1988 and 1994.  Moreover, this boom-bust cycle was associated with significant dislocation in the regional economy.

It's not hard to figure out why home prices are more volatile in tightly-regulated jurisdictions.  When the supply of new housing is severely limited, even a small increase in demand will cause a spike in home prices.  This spike, when combined with the perception that prices will continue to rise, lures other buyers into the market who want a piece of the windfall, which causes a larger spike, and so on, until the Ponzi scheme comes crashing down.  Tight land-use regulations thus pull double duty:  They cause the initial spike by constraining supply and they reassure buyers that supply will be constrained in the future.  This is why the bubble started in places like California and Boston rather Houston or Dallas.

I admit this can't be the full story, though, because Phoenix and Las Vegas were eventually swallowed by the bubble even though it's easy to build new housing in both cities.   And this does not explain why lenders were so willing to extend credit to bad risks, or why investors were so willing to buy arcane financial instruments that they did not understand, or why sophisticated companies like AIG so grossly miscalculated the risk of default of the securities they insured.  

Hopefully we can figure these things out to avoid economy-crashing bubbles in the future.  Given the experimental evidence, though, we will have bubbles no matter what we do; the best we can hope for is to limit their size and the damage they cause. 

Comments

[M1EK]

Actually, Phoenix and Las Vegas were hit harder by the bubble than anywhere except maybe Southwest and Southeast Florida - both of which also have lots of supply for sprawly crap.

I don't think there's a strong correlation here at all.

[bittergradstudent]

Bubbles can happen anywhere where production capacity has a time delay.

Assume that you have two factories making paperclips. They both find that they are making 200% profit on the paperclips. It takes a year to retool the factories, and neither knows about the other. They both decide to double capacity. Well, then, the result is that the paperclip market collapses and they are both running at a loss.

Couldn't you explain the housing bubbles in rapidly growing neighborhoods this way, and actually expect them there (and if I remember right, Houston and Dallas aren't growing like Las Vegas and Phoenix are)--places where builders are noting a huge profit in home sales are places where the builders are goign to want to aggressively build. But every builder will do so, in an uncoordinated way. Then, they will eventually suddenly find that they vastly overbuilt, and then the bottom will fall out of the market.

Under this model, building restrictions might actually be an insulation against this sort of effect, and you could use it to say why the housing market collapse is less severe in Austin than it is in other similarly growing cities. (and where it has hit Austin, it's been in the exurbs, where they have been aggressively adding supply).

[AC]

A couple of points:

In your paperclip example, the problem is the time lag for new supply. In the housing market, a long lag makes developers slow to react to new prices, which makes the housing supply inelastic. In markets like Houston, developers can be more nimble, which means that the supply is (more) elastic. In a market with inelastic supply, increases in demand will show up in higher prices. In a market with elastic supply, increases in demand will show up in greater quantity.

If the only culprit for price fluctuations was your production timing problem, then banning all new production would indeed solve the problem. But demand for housing fluctuates all the time. In an elastically-supplied jurisdiction, a drop in demand will cause a reduction in new construction. In elastically-supplied jurisdictions, a drop in demand will show up almost exclusively in home prices. This is the model that Glaeser and Gyourko have in mind when talking about the volatility of home prices in tightly-regulated housing markets.

A bubble is a different creature. A bubble means an asset is consistently trading well above its "fundamental" value. Homes, in theory, shouldn't be worth any more than the capitalized value of their rental income. Paying more than that is (or ought to be) a bet that rents will rise. People in the bubble markets were paying prices that could not be justified by any anticipated growth in rents. Buyers were betting that home prices would keep going up simply because home prices had been going up.

Price volatility does encourage bubbles, though. That made tightly-regulated places in California fertile places for housing bubbles.

That still does not explain the bubble in places like Phoenix and Vegas. I don't know what jump-started the bubbles there, unless it was Californians awash in cashed-out equity looking for places to invest. The overproduction model makes a lot more sense in these cities.

[M1EK]

Again, though, the correlation between land-use restrictions and the bubble seems very weak. Phoenix and Las Vegas and Miami and Ft. Myers have nearly zero restrictions and ample supply of buildable land (except Miami). They all got hit worse by the bubble than did anywhere in California, except for the parts of California where regulations were weak! (Imperial Valley, e.g.).

[Mike]

>>That still does not explain the bubble in places like Phoenix and Vegas

You can add places like Miami and Orlando (and probably large swaths of the rest of Florida) to the list of "overproduction cities".

Also seems to me that there was quite a lot of new construction in the suburbs and exurbs of southern California over the past 10 years. But ultimately, giving everyone monopoly money to play with and the psychology of a buyer's market seems to have caused a bubble even in these areas that are supposedly "tightly regulated" - where lack of supply does not seem to have been an issue worthy of mention.

>> In a market with elastic supply, increases in demand will show up in greater quantity.

I think one of the problems is that even in cases where housing has elastic supply, it still takes a significant amount of time to adjust supply. It takes months to build a house. It takes years to build larger condos and apartment complexes. So the idea that we can have a perfect market where suppliers can measure and match the demand seems like wishful thinking.

[bittergradstudent]

But there is a time delay in housing development. It takes years for a new subdivision to fill out with owners, not to mention the fact that housing takes a few months to go from empty lot to actual a built house. it takes time and investment to move in development capacity, buy lots, add infrastructure, etc.

Couldn't a lot of this just be uncoordinated developers just thinking large, overbuilding developments, and then witnessing a crashing market?

[bittergradstudent]

sorry, I didn't properly read your response, which makes sense. Still, wouldn't large tracts of cheap land and a rapidly expanding population make an economy of scale approach to development more attractive, and thus make overinvestment a more significant risk?

[AC]

BGS, sorry for the late reply. Overinvestment happens. We see it all the time. Subdivision developers can slow their build rate, or even shut things down for a while, which mitigates overproduction. This is a bigger problem for multi-family, which can't be constructed a unit at a time.