Thursday, November 19, 2015

Housing, A Series: Part 87 - Erosion of Conventional Mortgages Comes From Supply Constraints

I want to think through this issue of rent vs. home prices some more.  Because the US housing market was divided between a few cities with sharp supply constraints and the rest of the country, we have this deceptive statistical issue where, nationally, average rents from 1995 to 2005 were relatively flat while home prices rose sharply, but if we disaggregate by city, we find that in the places where home prices increased the most, rents also were high and increasing.

Many researchers have found a connection between subprime mortgage expansion and the scale of the housing boom and bust.  They have interpreted this to mean that excess or predatory lending was the causal factor.  I have been looking at evidence that housing supply constraints may have been the causal factor in these markets, and I think this issue of how subprime mortgages were related to rising home prices and subsequent defaults is yet another case where subtle and complex factors have caused us to misinterpret the relationship.

During the 1995-2005 period, national rent was relatively stable, but rent in the closed access cities (cities where public policy prevents housing expansion) was rising.  Even in the largest metropolitan areas (MSAs) that have open access policies, rents were rising somewhat during that period.  During the last years of the housing boom, rent was claiming 5% more of the median income in LA and San Francisco than the typical historical levels.  Since then it has risen to claim another 10%.  Rents aren't just rising in these cities; they are rising faster than incomes.

In addition, in a city with persistent rent inflation, home prices will be pushed even higher, much like a growth stock will have a higher Price/Earnings ratio than a mature stock, because owning the home is a hedge against future rent increases.  So, in cities with high rent inflation, and especially cities where long-running supply constraints would cause us to expect rent inflation to remain high, home prices will be very high.

And, we see from 1995 that rent affordability (rent/income) was rising in the closed access cities, and mortgage affordability was rising even more.

(By the way, notice that rent continued to climb after 2007 while mortgage affordability dropped to the lowest levels since at least the 1970s.  That is when a crisis of disequilibrium struck our housing markets, not when both rents and mortgage costs were rising.)

Thinking about the revealed preferences long term household behavior implies, it looks like in the absence of income shocks and housing constrictions, the median household will tend to find a comfort zone where they calibrate their housing consumption to around 20% of income.  If they live in an area with strong housing supply constraints, they will eventually reach a point where they will be reluctant to reduce their real housing expenditures any more, and they will allow their housing expenditures to rise as a portion of their incomes.

Thinking of the higher cost of homes in the constricted areas as a problem of excess demand misses this detail about rents, I think.  I submit that it is more accurate to think of the rising cost of housing in these cities as a measure of the amount of distress households are willing to accept before they decide they have to migrate.  In the worst cities, gross incomes have risen sharply compared to the rest of the country, but income after rent expenses has actually declined, relative to the rest of the country.  Given localized housing constraints, if households held no preferences for location, then prices would not have risen at all, and households would have simply moved to the places where building was allowed without the intervening behavior of bidding up local rents.  The rise of home prices in the closed access cities comes entirely from that preference.  The prices of homes in the open access parts of the country is a reflection of demand, lending, etc.  The prices of homes in the closed access cities is a measure of "location stickiness", if you will.

Our conventions about home buying reflect the expectations of a household with a comfortable cost of housing in an open access area.  So, conventional mortgages usually call for mortgage expenses of less than 30% of income.  Conventional mortgage guidelines are not calibrated for the typical housing costs of a closed access city.  The rise of subprime loans coincided with the rise of rent affordability in the high cost cities because of the supply problem.  Marginal households were buying homes on terms that were sometimes creative, but the trigger for this wasn't that they were enticed by mortgage originators into increasing their housing expenditures.  They had already increased their housing expenditures, passively, by living in cities with rent inflation.  In a city where the median house rents for 35% or 40% of median income, how does a potential homebuyer purchase a home with a conventional mortgage?

These households were faced with the decision to either (1) use creative financing to remain in their home city, (2) move down market into a neighborhood they weren't willing to rent in, so they could buy a home in their city, or (3) move away.  Of course, many chose number 1.  So, we call the lenders predatory and we call the homebuyers speculators.  They didn't ask to be speculators.  They were forced to be speculators by the housing policies of their cities.  Unless they chose option 3 and moved away, they were forced to be speculators.  If they continued to rent, they had to spend (at the median) more than 30% of their income on housing.  And this would keep rising each year, eating up all of their future income gains.  If they bought, they had to pre-commit to paying that rent, based on today's expectations about those future rent increases, which was baked into the price of homes.  Buying was the more expensive option in those cities, but households who were buying in those markets were taking the less speculative position.  They were in a position where their housing budgets were at the frontier of what was personally sustainable.  Thousands of families were being hit with rent increases and forced to move away.  Buying a home would create a stable path of expenses.

Thinking about the conditions brought about by the supply constraints, the creative financing products don't seem so outrageous.  If a family has been living in the same home for a decade, and they have seen their income rise by 3% per year while their rent has risen by 5% per year, with uncertainty each year about the year to come, then is it really that speculative for them to save up a little money for a down payment and take out a mortgage with 95% LTV and a negative amortization schedule so that their monthly payments increase by a set 2% per year?  That looks like the hedge position to me, not the speculative position.  And if rent already claims 40% of their income, is it really unreasonable for them to apply for a subprime loan that doesn't verify income?  What would you have them do?

Furthermore, in the closed access cities, we can infer that housing demand was inelastic because rents were high as a percentage of income.  Households in those cities didn't have higher rents because households in San Francisco and LA were living in mansions compared to households in Houston.  They had high rent budgets because home prices were so high that their demand had become inelastic.  They had made significant cuts in their real housing consumption, and were unwilling to cut more - their housing demand was inelastic.  So, it seems unlikely that households who decided to be buyers would use generous mortgage terms to increase their housing expenses (in terms of rent).

There were some marginally closed access and open access cities that saw significant subprime activity and did see a price spike later in the boom, coincident with the largest increases in private subprime lending - inland California, Florida, Nevada, and Arizona.  It is possible that some of that late price spike was related to demand side excess.  But, on the national level, that probably amounted to 10% or less of the total increase in real estate prices over the 1995-2005 time frame.

There were some open access areas, like Houston, that saw significant subprime activity.  But there was not a price spike there.  In Houston, housing expenses were low.  In Houston, housing demand for the median household would have been elastic.  Houston is the place that we would expect to see overconsumption in housing because of lenient mortgage terms.  In Houston, a reckless family offered a questionable mortgage might say, "Hey, now we can move into that mansion down the street, and still only spend 25% of our income on it."  In San Francisco, they would say, "Hey, we can use these questionable terms to get 5% of our budget back."  In other words, if we are just thinking about the effect of generous credit on individual homebuyers, in Houston easy credit might lead to a price bubble.  It would not in San Francisco.

I'm surprised that this hadn't occurred to me until now.  And, I'm surprised it apparently hasn't occurred to anyone else that I know of.  When we recognize the clear separation between these high rent/high price cities and the rest of the country, this is almost definitional.  Households in the cities where home prices shot up had inelastic housing demand.  The price boom had to be dominated by supply factors.

Based on 30 year fixed rate, 20% down, Median House & Income
Here is a chart of mortgage affordability in Houston, Dallas, and Atlanta (all the data I am using here is from Zillow).  A decent number of subprime loans were originated in these cities in the 2000s.  With a cumulative population of 14 million, they built over 1.6 million homes over the decade.  New York, San Francisco, LA, San Diego, and Boston, with a total population of 42 million, managed to build just under 1.2 million homes.  Of the top 20 MSAs, these 3 cities accounted for 13% of the population and 27% of the housing permits over that decade. So, where was there a housing boom?  Where were marginal new households becoming homeowners?  I think we can consider these cities the baseline cities for how demand and excess lending affected the housing market.

I conclude that there were relatively stable demand forces in these cities until mid-2006, when an unprecedented negative demand shock started an unprecedented decline in real estate prices, where, by the end of the decline,  the median homebuyer in low cost cities could buy the median home for 10% of their annual income - half the cost of renting.  That negative demand shock remains operative today.

But, I think, in the end what we have here is a peculiar interplay between supply and demand, which is due to the migration and income patterns that come from the extreme level of supply deprivation in the closed access cities.  I will save that for the next post.


  1. So I'm not sure about Florida, but Inland California, Nevada, and Arizona sure sound like places where overflow from SF/LA/SD picking Option #1: GTFO edition would be headed.

    So between the anchoring effect of "I know that a house doesn't cost a million bucks, but I haven't seen anything less in the last 5 years", and the overflow from Metro Areas of 30 or 40 Million overwhelming metro areas of a couple million (if that), that would explain a lot about why those 3 regions in particular would have seen excess subprime lending.

    1. Yes. I agree. To be honest, the evidence against the lending bubble narrative seems so overwhelming, I'm kind of throwing it a bone there. There are a few places where there was a brief correlation between the growth of subprime loans, very high price increases, and no long term supply explanation. Your explanation is probably the basis for a very good argument. And, there were short term supply issues. In Phoenix, builders were literally holding lotteries for new homes because they couldn't get new permits fast enough to meet demand. But, the price movements were so strong, and the supply constraints shouldn't have been considered permanent, that I think reasonable readers could balk at an attempt to write off those peculiarities totally as part of the supply problem.


    It occurred to Tom Sowell too.

    1. Thanks, Patrick. That book is now on my to-do list.