Tyler Cowen has linked to two items recently that are among the countless articles and studies that are at odds with the findings of my current research.
The first item is a link to an upcoming book by Joel Kotkin, where Tyler suggests that higher housing expenses in the cities I call Closed Access cities are somewhat inevitable and that expanded building will only draw in more high income workers. (1) I reject that claim and (2) even if it is true, it would be an even better reason to build, because it means that the value of density is practically limitless. Maybe this is step one to Robin Hanson's world of ems living in tightly knitted networks. On careful reading, I don't think Tyler is saying this is a problem, per se. He's just saying building won't lower costs. But, even here, I think it would be quite a jump to argue that greatly expanded building in the Closed Access cities would not benefit the current residents who are being stressed by rising rents. Even if expansion only led to more rising incomes and rising rents, the increased local market for non-tradable services would surely raise the incomes of current residents, too. They would likely get some relief from rising incomes, even if rents didn't relent.
The second item is a link to a VoxEU article by Borio, Kharroubi, Upper, and Zampolli that is yet another in a LONG line of articles which take financial expansion as the cause of a boom, and misallocations of capital and labor as a result of that boom, which lead to busts.
I have recently received some great data from Zillow which helps to pull together some more conclusions on these matters. (THANK YOU, ZILLOW! One more reason Zillow rocks. Zillow Data, much of which they publish for free at their website, has been invaluable in my study.) They sent me data with estimates of total residential real estate market values, by MSA, going back to 1998. I had been hoping to get something like this to confirm the total effect on national asset values coming out of the Closed Access cities.
Here are the picture pages:
As I suspected, deprivation was the source of much of the increase in real estate values, not demand-side excess. In this first graph, the US outside the Closed Access and Contagion cities saw a rise of roughly 20% in Property Values / Income from 1998 to 2006. Over this timeframe, real yields on 30 year TIPS fell from about 4% to about 2% - a tremendous drop for rates that usually are very stable, in a short period of time. To give an indication of the power of this change, this caused the price of 30 year bonds to rise by 45%. So, outside the Closed Access and Contagion cities, not only is there no evidence, in the aggregate, of demand-side excess, but there is no evidence of demand-side excess, and the response of home prices to long term real interest rates is tentative. The market price of real estate in 2005 was very conservative, compared to a discounted cash flow valuation (before taking tax effects into account).
In 2003, two shifts began to happen. Pressure on the GSEs from the Bush administration led to a sharp curtailment of conventional mortgage originations.* And, population growth shifts downward in the Closed Access cities.
This caused migration strains in the Contagion cities. But, notice that population growth didn't respond in the Contagion cities, so the Contagion cities began to show some of the symptoms of Closed Access. (Keep in mind that the Closed Access cities have about 15% of the US population, while the Contagion cities only have about 5%.) Instead of building more homes, the Contagion cities were forcing home prices to rise enough to induce a second wave of migration from Contagion cities to the rest of the country. Yet, at the same time, Price/Income levels in the Open Access cities were falling even while real long term interest rates continued to decline.
I know this is a hard sell for a lot of people, but the sudden rise in non-conventional mortgages in 2004 and 2005 was not a product of excess demand. It was a response to the arbitrary curtailment of conventional loans. The price movements in the Open Access cities here should be at least as surprising to us as the movements in the other direction in the Closed Access and Contagion cities. Two trends were pulling in opposite directions in 2003-2005. Supply constraints in the Closed Access cities were creating a high rent refugee crisis and the high home prices from that effect were from high rents. And demand constraints from the GSE's were already creating a negative dislocation in the rest of the country.
By 2006, we had killed demand through credit and monetary policy. This, perversely solved the refugee crisis because an economy with fewer opportunities meant that there was less in-migration into the Closed Access cities to tap into lucrative labor markets. They weren't so lucrative anymore.
Note what happens after 2006. The falling relative population in the Closed Access cities levels off, and suddenly population growth in the Contagion cities (here, Riverside, Phoenix, Miami, and Tampa) notches down from their long term trend. The bust has created a persistent dislocation.
Rents have resumed their rise, but since the bust expanded the supply problem to the whole country, now Closed Access residents have rising rents and nowhere to escape to. As we see in the next graph, Closed Access cities have seen rising incomes over the whole period - much of which is simply routed to landlords. Most of the country has seen incomes rise at a relatively equal pace. Except for the Contagion cities, because we killed the construction industry, so for at least 6 or 7 years, we just sucked the life out of those economies for no reason.
The reason we have imposed this problem on them is because of this giant honking error at the center of everyone's point of view on this. That error is that we have one housing market, and since, in the aggregate, it looked like rising prices and rising housing starts were related, we assumed the exact wrong conclusion - that both of them were caused by too much money. But this didn't happen anywhere. It can't happen anywhere. The reason everyone is so mad at the banks is that we assumed something was happening that would be basically impossible. So, we explained this incongruity with the idea that banks were massively drawing buyers into homeownership to prop up outrageous home prices. But the thing we were trying to explain wasn't happening anywhere.
As the next graph shows, the way to make real estate in your city more valuable is to not build. From 1998 to 2005, real estate in the Closed Access cities increased from 19% of the US total to 24% of the US total. And they did this by NOT BUILDING. And the Open Access cities caused their total real estate value to fall from 4.6% to 3.6% of the US total by building like crazy - their population increased from 4.9% to 5.3% during the same time.
The housing "bubble" had nothing to do with homebuilding. The widespread presumption that misallocation of labor and capital was due to monetary excess is wrong on every count. None of those things happened. But, models that use those assumptions look like they are on to something, because they measure falling productivity along with the supposed misallocation. But, the reason there was misallocation was because of the deprivation. The reason productivity is low and construction activity appears to be high is because Closed Access policies are preventing workers from accessing highly productive labor markets. And, since those markets are closed, the houses we can build in other cities require a lot more material inputs for the same amount of value.
So, these studies get the causation backwards and it creates a whole set of false conclusions to argue about that all have the premise wrong.
Looking back at the first graph, the truth of the matter that will be, at some level, the cause of the next recession, is that if we took the fetters off the mortgage market, it would unleash a massive resurgence of pent up real investment. This would cause interest rates to rise. But, even if real long term rates recovered back to their 2003-2004 levels, and even if all that new building brought rents back down to previous levels, home prices would still rise by 30% across the board just to pull back to reasonable valuations.
Before that happens there will be gnashing of teeth and general kvetching, and there will be nearly universal calls to kill the housing market again. We need years of 2 million annual housing starts, and this can only happen with a market in equilibrium, which means much higher housing prices. And, we aren't about to let that happen.
* This may be hard to believe, given many observers who blame the GSEs for the problem that never happened but that they think happened. Here is one of my older posts with some details.