Wednesday, February 24, 2016

Housing: Part 121 - Michigan's Odd Place in the Housing Bust

Michigan keeps strangely popping up in readings about the housing bust.  It's strange, because I don't think Michigan is the place we usually think of when we think about the housing bubble.  Mian and Sufi seem to use it as an example of how low income households were herded into unsustainable mortgages.  Low income neighborhoods in Detroit certainly paint a certain picture in our heads.

Source
They find correlations at the local level between low incomes and home price volatility.  I would expect gentrifying or improving neighborhoods to experience housing gains at the local level in any market, bubble or no.  And, I would expect neighborhoods with a large number of recent purchases to be more vulnerable to a severe price contraction.  I'm not sure how to separate these issues from signals of predatory credit.  Maybe they address this in ways I haven't seen yet.  I haven't looked at all of their research.

Share of mortgages with principal balance exceeding estimated home value: 2009:Q4
Share of mortgages with principal balance exceeding estimated home value: 2009:Q4
Source: San Francisco Fed
But, at the metropolitan area level, Detroit seems like a strange choice as the lead character in the housing bust drama.  Home prices in Detroit were flat during the boom - even declining in real terms.  This is because depopulation was bringing down rent inflation and expected rent inflation.  Housing was becoming more affordable in Detroit.  It is true that this is not a particularly fitting context for some types of subprime loans, according to Gary Gorton, which tend to depend on stable or rising home prices.  So, there may have been some borrowing on inadvisable terms.  But, it seems likely that the housing boom in Detroit was less about high expectations about capital gains than it was about staking an ownership claim in a neighborhood that wasn't dying.

Negative equity was high in Michigan at the bottom of the bust, which does sort of place it in a category with the places that saw sharp price spikes and contractions.  But, on an aggregate scale, Detroit does not show any signs of the boom.

Mortgage affordability was flat, much as in the Open Access cities of the sunbelt, until it fell to extremely low (affordable) levels after the bust.  It seems especially wrong-headed for us to be afraid of mortgage credit when the median household in Detroit could buy the median home with a conventional mortgage and payments would only require 10% of their income.

Source
I think the problem in Michigan wasn't a problem of too much credit.  It was a problem of too little income.  Mian and Sufi describe the early rise in unemployment in northern Indiana because of the failing RV industry.  In this graph, we can see how unemployment in Detroit was elevated during the boom.


Source
In the next graph, we can see that in late 2005, coincident with the inversion of the yield curve, expenditures on new motor vehicles dropped sharply.  The 2006-2007 period was a sort of minor recession with full-employment for the rest of the country, because deprivation was focused on housing.  Homebuilders are spread throughout the country.  And, households either escaped the cost of rising rents by being owners, or experienced the pre-recession recession through higher costs, not through lower wages or the unemployment that tends to accompany negative wage pressures.  But, this was not the case in Detroit, because the pressures on fixed investment and expenditures on durables were focused on Detroit, where there is a concentration of producers.

By the way, to the extent that employment in Detroit is a signal of our hobbled credit markets, the recent tick up is not such a great sign.

Source
The bust in Detroit, therefore, had a different quality than the bust in the rest of the country.  In Detroit, the collapse in incomes clearly came before the collapse in housing market.  If only the rest of the country looked like Detroit, we might have been willing to accept stimulus in 2006 when we could have avoided a crisis.

Source
Here are two maps of foreclosures in 2007 and 2009.  In 2007, the area around Detroit is the core of the foreclosure problem.  Outside of the Detroit area, foreclosures were not elevated at that time.  We can see here that cities like Dallas and Charlotte - pure Open Access cities where building was strong - had slightly elevated foreclosure rates relative to the rest of the country.  But, nothing outside of the Detroit areas, as of June 2007, would create concern.  Remember, the subprime origination market had already collapsed by this time.

In June 2009, Detroit still looks bad.  Places like Dallas and Charlotte still looked about the same as they had in 2007.  But, now, the core of what really became known as the housing bust is clear in Florida, Arizona, Nevada, and California.

In Detroit and its surroundings, there was never a housing bubble.  Price fluctuations were based on localized factors.  And, in that area, the bust happened in the normal way.  Employment and incomes collapsed first and economically stressed households couldn't make their mortgage payments.

The rest of the country was either Open Access, where home prices and foreclosure rates weren't fluctuating wildly, or was Closed Access (or a contagion area), where foreclosures followed after home prices collapsed.

In this last graph, we can see the relative decline of incomes in Detroit.  In 1979, Detroit still had higher household incomes than San Francisco.  From the peak in 1999 to the peak in 2008, the median income in San Francisco increased by more than 23% while it increased in Detroit by less than 7%.  This, despite the fact that San Francisco was the epicenter of the tech sector, the cause of the drop in incomes after 1999.  Median household income in Detroit - nominal income - has only just now surpassed the 2008 level.

In a few decades, when other cities are able to compete and the source of its economic rents weakens, San Francisco may be the blue line on that last graph - a city with homes in search of jobs instead of a city with jobs in search of homes.

PS: pithom makes a good point in the comments that oil prices were rising at the time that auto sales were falling, so that I shouldn't get too excited about credit as a causal factor.  That's true.  Although, Mian and Sufi using the Detroit housing market as a prominent example of the housing bubble in House of Debt remains an oddity in either case.

5 comments:

  1. Egads, we should give a medal to anyone whi buys a house in Detroit.

    BTW--I mention you and manufactured housing over at Historinhas. Prominently.

    ReplyDelete
    Replies
    1. This is still the case for the Detroit Metro though.

      Mom and Dad bought in in a second-ring burb for $110K in 1991, our neighbors sold out at $190K in '05, our other neighbors sold out at $110K in 2009, and then Mom sold out at $150K in '13.

      Which means that in real inflation-adjusted terms, Mom bought in for $191K, and sold for $152K.

      Yes, those $1 houses in Detroit are sort of famous, but they're missing the much nicer $100-$200K homes scattered across the entire Metro with any commute and amenities of your choosing to suit.

      /And yes, property taxes will kill you.

      Delete
  2. It's interesting you don't mention the effect of competition from East Asia and Mexico on rust belt states' industry.

    Yes, Detroit had begun to have clear economic problems as a result of the 2001 recession, which disproportionately affected manufacturing and high-tech. Detroit was affected by the bust in the former (which was longer), San Francisco by the bust in the latter.

    High oil prices might also have contributed to its decline.

    Divide San Francisco unemployment by national unemployment and do the same with Detroit unemployment. See what you get.

    ReplyDelete
  3. It also appears there was a stagnation in automobile manufacturing output per worker throughout the U.S. in 2004-7 and Detroit auto employment declined more consistently than in the rest of the country, where it grew a little from 2003 to 2006, unlike in Detroit.

    https://research.stlouisfed.org/fred2/graph/?g=3zE2

    ReplyDelete