Monday, July 3, 2017

Housing: Part 239 - Homes in Contagion Cities during the housing bubble were Inferior Goods

The motion chart from the previous post really helps to visualize the difference between the Closed Access cities and the Contagion cities.  The bubble in Phoenix happened entirely after the Fed began to hike the Fed Funds rate.

The out-migration from the Closed Access cities had been growing since the late 1990s, and peaked in 2004 and 2005.  This surely was facilitated by nonconventional mortgages which helped households living in the Closed Access cities with high incomes to purchase homes and spread their elbows a bit.  But, you would think this would show up in gross migration flows.  You would think that during this time, more potential in-migrants would be able to buy Closed Access homes, so that there would be an increase in both in- and out-migration among the Closed Access cities.  But, according to IRS data, this wasn't the case.  Closed Access in-migration was low and flat throughout the housing boom.  This is one of several oddities that I think creates some doubt about the centrality of the private securitization boom as a cause of bubble prices and migration patterns.  ACS data, which only goes back to 2005, suggests maybe Closed Access in-migration in the top income quintile increased by about 10,000 households during the peak boom years, with little change among other income quintiles.  This compares to net out-migration at the peak of more than 200,000 households, annually, from the Closed Access cities.

One of the interesting things that I think the motion graph helps to show is that as soon as the Fed began to hike interest rates, price appreciation in LA - especially in the top tier markets - moderated.  And, it was after this moderation that Phoenix prices shot up.  But, we can also see that prices in Phoenix are much lower than in LA.  Here is a line graph of home prices over time, by price quintile.  In 1999, the top quintile of prices in Phoenix were similar to the 4th quintile of prices in LA.  By 2004, top quintile prices in Phoenix were lower than 2nd quintile prices in LA.

We can see the downshift in price appreciation in LA here in 2004, when rates began to rise.  And, at the same time, prices in Phoenix shot up.  But top quintile prices in Phoenix peaked at the end of 2005, still below the median quintile in LA.

Of the more than 200,000 households, net, that migrated out of Closed Access cities in 2005, about 85,000 were homeowners from the top two income quintiles.

Technically, we can call what happened in Phoenix a bubble.  It had the classic ingredients of a bubble - temporarily inelastic supply and demand.  But, this had nothing to do with "easy money" and I'm not sure that it really had much to do with easy credit.  The bubble in Phoenix, ironically, was the very early first signal of the bust.  Homes in Phoenix were inferior goods.  As counterintuitive as this is, this should be uncontroversial when one thinks about it for a moment.  That massive inflow of homebuyers in Phoenix in 2005 were buying downmarket.  They were buying way downmarket.  For the migrant households as a whole, it would have been mathematically impossible to do anything else.

Homeownership and rate of first time buyers were declining at the time.  But, how does this square with prices that continued to rise and mortgages outstanding that continued to rise?

Mortgages continued to rise because those Closed Access sellers were very lightly encumbered.  If your home increases in value from $450,000 to $1,000,000 in just a few years, it would be very difficult to be leveraged even if you tried really hard to be.  So the sellers were mostly claiming equity in those sales.  But, the new buyers would have naturally been more leveraged.  When those 85,000 homeowners left town, they had to be replaced by about 85,000 new homeowners.  Homeownership rates were starting to drop, but they weren't dropping by that  much.  About 2% of homeowners were leaving the Closed Access cities annually.  Homeownership rates shift by fractions of a percent.  Those new homeowners are naturally more leveraged.  That is why mortgage levels continued to grow.

Prices continued to rise in Phoenix because there was a migration surge of buyers massively reducing their housing expenditures.  Prices continued to rise in LA because prices in LA were a rational reflection of future rent values.  Today, it is easier to say that, because as we see in the graph, the most expensive homes in one of the most expensive cities in the country, have risen to new highs, even as mortgage markets have remained suppressed.  Those homes in top tier LA markets in 2006 turned out to be decent investments over the following decade.

This is because it is only in extreme and temporary circumstances that a shift in the number of buyers and sellers will move market prices.  Intrinsic value rules.  In Phoenix, briefly, the number of buyers might have pushed prices out of sustainable levels.  That was not the case in Los Angeles.  Consider that, if mortgage expansion could create a sustained increase in prices at the scale we have seen in the Closed Access cities, that the expansion had to have been so far outside normal ranges that it still pushed prices out of rational valuations, even though before that could happen, it first had to make up for those 85,000 fleeing homeowners.  That seems highly unlikely to me.  That's the jab to the credit fueled explanation for Closed Access home prices.  And the uppercut is the strong price trends in those cities in the decade since the mortgage market collapsed.


  1. Does IRS data capture immigration to closed access cities from outside the nation?


    1. That's a great question. The IRS does track foreign migration, and the number of immigrants fell sharply after the bust, and their average reported incomes shot up.

      According to the Census Bureau, most major cities take in some amount of net foreign migration, but the Closed Access cities don't appear to be unusual in this regard in net terms.