Wednesday, October 17, 2018

Housing: Part 325 - Lending and the Housing Market

I don't know if I have shared this graph before.  I think there are some interesting things to see here regarding lending and housing.

Surprisingly, the number of mortgage accounts outstanding continues to decline.  In early 2008, there were 98 million mortgages outstanding.  That dropped to about 81 million in 2013.  Today there are just under 80 million.  Since 2015, the number of owner-occupied homes has increased by about 3 million and the homeownership rate has finally leveled off.  This has happened in spite of lending markets.  That net gain in homeowners consists roughly of 4 million additional households with no mortgage and a decline of 1 million households with mortgages.


Sources: New York Fed Quarterly Report on Household Debt and Credit,
Census (HVS), Fed Financial Accounts of the United States
The average mortgage size has been growing, but home prices were rising more quickly, which has helped home equity levels recover to pre-crisis levels.  Now, home prices and mortgage sizes are rising at about the same rate.  But, since there continues to be a shift to owners with no mortgage at all, average home equity levels continue to rise.

It certainly could be the case that there is a baby boomer effect here, and that there is some growth in new mortgaged ownership, but that it is matched by baby boomers who are making the last payment on old mortgages and moving from mortgaged ownership to unmortgaged ownership.

But, I'm going to step out on a limb here and suggest that this doesn't look like a lending market in a healthy recovery, let alone a lending market that is in need of a macroprudential clamp down.


Thursday, October 11, 2018

September 2018 CPI

Still limping along.  Non-shelter core inflation continues to be pacing along at around 1%.  Possibly shelter inflation might be starting to wane.  This continues to suggest that further rate hikes are unnecessary and potentially disruptive, but rising market rates since the Trump election continue to buffer rising policy rates.

Friday, October 5, 2018

Housing: Part 324 - Commercial Real Estate, Dean Baker, etc.

Arnold Kling points to this post by Dean Baker.  The last paragraph of the Baker post gives Baker's basic conclusion:
The basic story is that demand plummeted first and foremost because of the collapse of the housing bubble, along with the collapse of the bubble in non-residential construction that arose as the housing bubble began to deflate. The financial crisis undoubtedly hastened these collapses, but a steep drop in demand was made inevitable by these unsustainable bubbles that had been driving the recovery from the 2001 recession.
He is arguing against Bernanke's recent posts where Bernanke claims the recession was deepened more by the financial panics than by the housing bust.  (I basically agree with Bernanke, and I would say that the panics were largely caused by Fed policy choices in 2006-2008, and the losses were made permanent/justified by the extremely tight lending standards imposed by the post-conservatorship GSEs and CFPB.)

Bernanke points to the post-crisis drop in non-residential investment as evidence of the importance of the financial crisis in creating the deep recession.  Baker counters that the drop in non-residential investment was mostly a drop in non-residential construction, and was simply a part of the same bubble that had infected residential building.

I'm not sure if I have that much new to add here.  The entire thing pivots basically on this comment by Baker: "Again, the collapse of Lehman hastened this decline, but the end of this bubble was inevitable."  Whether the bust was truly inevitable or not is beside the point.  The bust was inevitable because the zeitgeist had deemed it inevitable.  The conclusions are a product of the presumptions.

And, looking at the CEPR paper that forms the basis of Baker's post, we can see the source of the false presumptions.  In the bullet points that summarize the paper, he notes, among other things:

The decline in residential construction during the downturn was mostly just a return to trend levels of construction, along with a predictable reduction due to the overbuilding of the bubble years. Any impact of the financial crisis was very much secondary.
 ….
The bubble and the risks it posed should have been evident to any careful observer. We saw an unprecedented run-up in house prices with no plausible explanation in the fundamentals of the housing market. Rents largely rose in step with inflation, which was inconsistent with house prices being driven by a shortage of housing.
Unfortunately, these assertions are broadly accepted as canon.  Obviously, taking opposition to the overbuilding issue is central to my work.  In the paper, Baker includes figures for residential construction as a % of GDP, which begins at 1980, and for non-residential construction as a % of GDP, which begins at 2002.  Here is a graph of those two measures, dating to 1960.

I agree with Kling that Baker seems to be an independent thinker. But his choice of start dates seem especially useful for magnifying the level of these measures during the "bubble" years.  I don't think he is trying to be misleading.  The bubble is canonized and setting the timeframe to maximize the apparent excess is part of the public hypnosis in support of the false canon.

In addition to the long-term view, there are a couple of points that might be made about these measures, which it is possible that Baker missed.  Within the non-residential category, "mining exploration, shafts, and wells" increased from 0.3% to 0.9% of GDP from 2003 to the end 2008.

Also the residential category includes brokers commissions on real estate transactions.  From 2000 to 2005, that increased from 0.9% to 1.4% of GDP.  If you subtract that from the residential investment measure, the peak level is at about the same % of GDP as the peaks in the 1970s.  Brokers commissions have nothing to do with building.  In fact, they were bloated specifically because of under-building.  They were bloated because of existing homes in coastal California selling for a million dollars.

But, nonetheless, building was strong at the same time prices were rising, which brings us to the second canonized false presumption that Baker references above: the idea that rising prices were unrelated to rising rent.  This is, again, a product of the public hypnosis on this issue.  Even looking nationally, rent inflation had been above non-rent core CPI inflation for the entire period from 1995 to 2008 - far above non-rent inflation for much of that time.  From the end of 1994 to Sept. 2008, non-shelter core CPI averaged 1.8% and shelter CPI averaged 3%.  But, the problem is even worse when you look at the MSA level instead of the national level.  Generally where prices increased, rents had increased.  So, it's more like there were many places with moderate prices and normal rent inflation and places with high prices and rent inflation persistently well above general inflation.  And, those were places that definitely were not over-investing in construction.  At the MSA level rent explains almost everything.  And, on this point, the public hypnosis is striking.  Open coastal urban newspapers or twitter and the topic is high rents in the coastal metropolises.  It isn't as if this is a secret.  But, hypnosis is strong enough to create mental silos on this issue.

This is part of the story on rising prices.  Before the mid-1990s, if rent affordability got worse in a city, it tended to revert to the mean.  But, beginning in the 1990s, the economy became characterized by this new regime, where urbanization has new value, the urban centers that would create that value do not grow, and workers must segregate by skill and income into and out of those cities.  So, now migration patterns exacerbate the rent inflation problem rather than causing rents and incomes to revert to the national mean.  Prices in 2005 reflected this regime shift.

But, the bubble was canonized before this realization was made.  One might argue that rents should still revert to the mean and that bubble prices still reflected over-optimism, even if rents had been rising for a decade.  (That would be wrong, as rent inflation has resumed after the crisis, but at least it would be an argument that addressed the facts.)  Instead, a false reality is invoked, rent inflation is ignored, and discussions of housing market sentiment revolve largely around price expectations, which, by presumption, leads to behavioral explanations.  Again, I don't say this to be harsh regarding Baker.  To treat rent correctly would be a radical, contrarian position.  Until this correction gets made in the zeitgeist, you might as well complain that he references gravity without engaging in an experimental proof.  It's canon, and the canon is wrong.

Regarding prices, here is a graph of various property types.

The thick orange line is non-residential commercial real estate.  The thick blue line is residential commercial real estate (multi-family buildings).  These are from CoStar.

Figure 5 in Baker's CEPR paper, published in September 2018, shows commercial real estate prices from 2002 to 2010 in order to show how strong the bubble was in commercial building.  He writes:
The plunge following the collapse of Lehman is not a surprise. Non-residential construction is largely dependent on bank credit, and when this dried up with the financial crisis, it was inevitable that it would take a serious hit. But the financial crisis was only the proximate cause of the drop in non-residential construction. The bursting of the bubble was inevitable in any case, the only question was the timing and specific events that set it in motion.
I do not disagree about the importance of credit.  This is all about presumptions.  This entire discussion hinges on one word: "inevitable".

A diversified basket of multi-family real estate bought at the peak of the "bubble" at the end of 2006 would have returned 44% of capital gains in addition to rental income over the following 12 years.  In the 9 years since the nadir at the end of 2009, it would have returned 124%.

In the graph above, I have also included the national Case-Shiller home price index (black), price levels from Zillow for the top and bottom of the Atlanta market (gray) and the LA market (red/orange).  Maybe I am confusing matters by including LA.  Many will see the clear signs of a credit-fueled bubble in the low tier prices in LA, but the truth of the matter is too complicated to go into here.  But, these measures tell a more complete story about what happened.

Once we recognize that rising rents are the main difference between LA and Atlanta and that credit at the extensive margin was not an important factor in the boom (which is clear in Atlanta and most other cities where price appreciation was not very different between top and bottom tier homes during the boom), we can see a different story.

Remove "inevitable" from your presumptions.  The consensus around "inevitable" led to acceptance of, even demands for, a negative credit shock in owner-occupier housing markets that continues to this day.  Nothing was inevitable.  Residential housing markets look like they track along with commercial markets for the entire period.  But, they are a chimera.  They contain open markets and closed markets.  Before 2007, prices in most markets, like Atlanta, were benign, and prices were very high in housing constrained markets.  Sentiment and credit access began to turn in a series of trend shifts and events from the end of 2005 to 2008.  Housing starts started to collapse in 2006 and prices eventually fell sharply after mid-2007.  This happened in Atlanta and LA and everywhere in between.  Since intrinsic value remained strong, commercial building, even in residential, remained strong until 2008, and rent inflation across the country spiked in 2006 and 2007 as new single family supply dried up.

Then, we imposed the "inevitable" bust on the owner-occupier housing market.  Instead of looking for ways to stabilize mortgage markets, lending was largely cut off to the bottom half of the market from 2008 on, and we can see the devastating effect if we look within cities, most of which look like Atlanta, where low tier prices took a post-crisis hit to valuations, frequently of 30% or more.  This has caused the market price of low tier homes to drop below the cost of construction, causing new building to dry up in low tier housing markets.  The lack of supply in those markets has been a boon to commercial residential builders, who have access to equity and borrowed capital.  Ample building is happening there, but it can't make up for the tremendous hit that owner-occupied single family homes have taken, and it can't create ample coastal urban supply.  So, the boon to multi-family builders continues for the same reason prices were high in 2005: there aren't enough units, especially where demand is greatest.

The national multi-family market reflects a price level that is not credit constrained, but is supply constrained.  The national home price reflects a price level that is credit constrained, which is a mixture of cities like LA, which is supply constrained, and Atlanta, which is especially credit constrained, and is only supply constrained now because it is credit constrained.

Tuesday, September 25, 2018

Housing: Part 323: Construction Employment during the crisis

I have dug into employment numbers a bit, and I think there are some interesting things here.

What happened?  A reasonable person might say this: There was overbuilding by the end of 2005.  This required a shift out of construction employment as the housing bubble wound down.  Federal officials underestimated how much the housing correction would bleed into the rest of the economy.  Eventually the collapse of the construction market in the bubble cities metastasized and caused employment and consumption to contract more broadly.

As I have so often found, the truth may be closer to the opposite of that.

Here, I am using state data, and I am using two independent variables to describe each state.  The first variable is the level of construction employment in December 2003 as a percentage of total state employment.


Construction as % of Total Employment: higher vs lower construction states
Here is a graph of construction employment as a percentage of total, over time, for states that had construction employment one standard deviation above or below normal in 2003.  There are several interesting points here.

(1)  States with high construction employment in 2003 are states that typically had high construction employment in the past(in other words, building lots of homes and growing).  And, notice that from the end of 2003 to early 2006, states with less construction employment continued to have flat construction employment but construction employment in the high construction states went higher. This supports the idea that the housing boom was just an acceleration of longstanding patterns of migration.

(2) When the CDO panic hit in the summer of 2007 and the recession officially started in December 2007, construction employment was still near the peak of the boom.  There was no re-sorting out of construction into other forms of employment before the recession.  Then, during the first year of the recession, construction employment did start to drop somewhat in the growing states.  But, it was only after the financial crisis that construction employment saw its steepest decline.  The recession caused the contraction in construction employment.  Construction continues to run below the pre-boom levels in the states that had previously been high construction/high growth states.

1 Year Change in Total Employment, by State
But, interestingly, there was a contraction that preceded the recession in total employment growth in high growth states.  It's just that that contraction was not focused on construction employment.  It was a general contraction.  And it was sharp.  Employment growth in states that had low construction employment continued along at its (low) flat rate of just under 1% annually.

So, there was an employment slowdown in the states that had been building a lot of houses, but it wasn't a slowdown in construction employment - even though housing starts were dropping sharply.

Also, this graph shows that there was an especially wide gap in employment growth in 2004-2006 between high construction and low construction states, but, compared to the 1990s, the gap wasn't wider because the high growth states were growing more.  It was wider because the low growth states were growing less.


Construction employment as % of total: bubble and non-bubble states
The second independent variable I used was the change in construction employment from December 2003 to March 2006 after accounting for the correlation between the pre-existing level of construction employment and rising construction employment that is visible in the first graph.  In other words, the first independent variable is a measure of persistent migration patterns and the second independent variable is a measure of "bubble" activity from 2003-2006.

The pattern is similar during the boom and early crisis.  Construction employment peaked in 2006 and remained relatively high until the recession began, then started to decline, and especially declined after the financial crisis.

But, notice the difference after the crisis.  The "bubble" states - states that had unusual growth of construction employment during the boom - never saw construction employment contract below the level of construction employment in non-bubble states, and then recovered more strongly after the recession, so that now, they are back near the construction levels of 1996-2003.

So, the states that had accommodated persistent in-migration for decades have been permanently hobbled by the housing bust (They continue to have higher construction employment than other states, but not as high as they previously had.), while the states that actually had unusual construction employment growth during the boom continue to have unusual construction employment growth compared to other states.


As with other data, this suggests that we didn't bust a housing bubble.  Instead, there was an acceleration of long-standing migration patterns, and those migration patterns have been hampered by a crippled housing market.

Unemployment rate: high vs. low construction states
There is a similar pattern in the unemployment rate among states.  The unemployment rate was lower in both the long-term growth states and the bubble states until mid-2008, then the unemployment rate in all states rose together through 2009, regardless of their previous construction employment levels.  As shown above, it was then that construction employment really collapsed.  In the bubble states (the states with unusual construction employment growth from 2003-2006, not shown in this chart) the unemployment rate continued to move in line with other states.  But, in the states where there had been long-term high levels of construction employment, the time period where their unemployment rates were especially high was 2010 through 2012.

This last graph is of the unemployment rate for the states that had construction employment one standard deviation above and below average in 2003.  And, here I have added a hypothetical state with zero construction employment, which I think gives an interesting baseline for thinking about construction and non-construction employment before, during, and after the crisis.

The crackdown on lending in 2008 and after, and the consensus view that new construction was problematic were the primary causes of dislocation.  Imagine if construction employment in high growth states had managed to bottom out at even 5.5% of total employment in 2009 and recovered from there.  Or, if it had recovered more quickly, as it had in the 1990s (which wasn't exactly a building boom decade, itself).

I am hoping to get a chance to look more thoroughly at this, but in the meantime, this seemed worth sharing.

Monday, September 24, 2018

Housing: Part 322 - The strange American housing morality play

One of the overwhelming tendencies one finds in the popular literature about the housing market is the nearly universal cynicism about housing consumption and housing finance:
  • Everyone buys too much house.
  • The real estate lobby has Washington on a leash.
  • The GSEs have spent decades lobbying for special treatment and excess lending.
  • The Fed is pumping up bubbles.
  • We lionize homeownership.
  • In the aggregate, homebuying decisions are characterized by speculative thinking.
I could go on and on.  Every book that purports to explain the housing bubble becomes a litany of decades of activities, all meant to get too many homeowners to buy overpriced houses with too much debt.  The rabidity and ubiquity of this treatment of the real estate asset class defines the topic.

The United States does not, on net, subsidize housing.

As with so many issues on this topic, the distance between the consensus and reality is extreme.

Let's look at the subsidies to housing.  For a sense of scale, the BEA estimates total annual rental value of about $2.1 trillion, and net operating surplus (rent after depreciation, expenses, and taxes) of about $1.1 trillion:

There are two biggies.  (Well, one biggie in reality and one in rhetoric):
1) income tax benefits.  This includes untaxed rent, mortgage interest deduction, and untaxed capital gains.  The Treasury estimates that in 2018, these are worth about $230 billion.

2) the GSE subsidy (this is a little cloudier with conservatorship).  When the GSEs were semi-private, it seems that reasonable estimates of their effect on mortgage rates was about 0.25%.  In other words, the implied federal guarantee on their debt led to mortgage rates about 0.25% lower.  Before the bust, they guaranteed about $5 trillion in mortgages.  $5 trillion x 0.25% = $12.5 billion.

Number 1 is much more of a biggie than number 2.  This is why the focus that so many people have on the GSEs baffles me.  As a subsidy to housing, it's a pittance.  I would prefer to get rid of the income tax benefits.  They are regressive and destabilizing.  The GSEs, on the other hand, comported themselves quite well during the bubble, and were a stabilizing factor in the crisis, where they were allowed to be.  As a result of my research, I have become more supportive of the idea that the federal government should provide a credit guarantee on conventional mortgages.  That function is a public good which can only be provided by the government.  The federal agencies should be retained in some form, and the inflationary effect they have on home prices is small.

What about taxes on housing:

There is one biggie:
1) Property tax on residential housing, which is about $250 billion per year, according to the BEA.

On net, these primary factors put negative pressure on housing demand.  Income taxes represent a subsidy of more than 20% of the aggregate net income, property taxes represent a tax of more than 20% of the aggregate net income, and the GSEs amount to a percent or two.

What about other factors?

The realtor lobby wants a lot of housing demand, and they push for maintaining things like the mortgage interest deduction.  But, probably their primary input here is protecting the realtor cartel that charges 6% for realtor services.  High transaction costs clearly pull down the market prices of homes and the demand for housing.

And, what about the 30 year fixed rate mortgage that is supported by the GSE framework?  The 30 year mortgage with a prepayment option puts peculiar risks on lenders, and they require a premium for taking prepayment risk.  That makes mortgages more expensive, which pulls down the market prices of homes and the demand for housing.

There have been other programs related to encouraging home ownership in various ways, but the effect on aggregate demand for shelter or on home prices is marginal.  Certainly not close to the scale of the effects of property taxes and income tax benefits.  Programs meant to increase homeownership can't amount to much.  Consider a very aggressive program that would increase ownership by 5%.  Those households wouldn't necessarily increase their housing consumption by that much, and any pressure they might create in home prices would also be marginal.  So, that program would affect aggregate consumption or prices by 5% x some small percentage representing the marginal new capacity of those buyers to consume more housing.

Nothing else can really come close to the effect of income tax benefits on housing consumption and home prices, and income tax benefits are cancelled out by property taxes.  The only way to rectify this is to argue that property taxes should be ignored.  So, the entire case for claiming that there is some sort of American public mania for housing consumption comes from observer bias.  You have to ignore a very large tax that is imposed specifically on this asset class.  Don't get me wrong.  I think there are a lot of good reasons for having healthy property taxes.  I just don't think you can have them and also claim that real estate is being heavily subsidized relative to other forms of spending.

How do you feel about additional marginal consumption of, say, health care, or education?  Or, for that matter, bananas, or books, or boots?  Compare public expressions about these forms of marginal new consumption to public expressions about marginal new consumption of housing.  At the risk of being a bore, I must say that when I read any history of the housing bubble, it is this universal attitude that strikes me as the fundamental source of irrational public sentiment that caused the crisis.  Once you see it from a different perspective, so that you notice it oozing and dripping rhetorically over every description of the history of American housing, it becomes fairly oppressive.

It's sort of an interesting problem.  For writers, it is a posture that one must take to establish credibility with the audience.  But, starting from that prior, every marginal increase in housing consumption is automatically suspect.  That can only lead to one conclusion.  It should destroy the credibility of the writer, because the conclusion has been predetermined.  For some reason, though, that predetermined conclusion, fundamentally, is the product that the American public wants to consume.  And, the extreme bias this creates is clear.  The national conversation for 20 years has been about what to do about the overconsumption of housing, and real housing consumption has been declining relative to incomes for more than 30 years.

Thursday, September 20, 2018

Housing: Part 321 - What about those naive bubble investors?

There is a story from "The Big Short" where a stripper in Las Vegas explains that she has several highly leveraged investment properties.  This is also a response I have heard and that I see frequently when people take umbrage with my assertions about the housing boom.  "Look, this is all very interesting, but I remember what it was like back then.  The janitor at my office had 7 properties.  It was nuts."  This is especially true of Phoenix and Las Vegas.

This a great example of how some basic factual truths can completely turn your conclusions upside down with just some subtle changes in interpretation.

These people existed - in some significant number.  But, let's think about this.  The housing stock is a big, slow-moving beast.  It doesn't change by more than a few percentage points a year, at most, in a fast growing city.  If you know, say, 4 people that have purchased 5 homes as speculators in the past couple of years, then you should also know about 20 people who have sold homes.  Every home has one owner and every transaction has a buyer and a seller.  So, if you say that you suddenly knew 4 people that each owned 5 speculative properties, then that is basically the same statement as saying you knew 20 people that had sold out of the real estate market.  Two sides of the same coin.

Outside of extreme circumstances, if you know four people that are deep into property speculation and you don't know 20 people who have sold out of properties, then you have stumbled into a deep case of observer's bias.

It happens that in 2006, there were extreme circumstances.  In 2005, annual population growth in Phoenix was over 3% and it was over 4% in Las Vegas.  Between 2005 and 2009, it fell to less than 1% in both cities - a rate of growth slower than either city had seen in decades.  This was a combination of more people moving away and fewer people moving in.  Builders were actually pretty sensitive to this shift, and permits for new construction fell sharply along with population growth.  But, in addition to those migration shifts, tens of thousands of potential new home buyers had entered into contracts to build new homes, and upon seeing the turn in the market, they reneged.  They let the builders keep their small escrow deposits, and they left those homes with the builders.  There was a massive shadow inventory of homes left to builders long before owners were defaulting and leaving homes with the banks.

So, if you knew 4 people who owned 5 homes, you came upon your observer's bias honestly.  Those 20 housing shorts weren't in your frame of vision.  They had either left town or had never moved to town. When you were sitting at the barber shop listening to the guy talking about the seven condos he was flipping, the seven housing shorts that were an integral part of that story were getting their hair cut in LA and Chicago.

This is one reason why migration is such an important corrective to our conception of what happened.

Note that the truth is even there in the conventional telling of the story.  In the scene in The Big Short, when the stripper tells Mark Baum that she has six leveraged properties, he warns her, "Well, prices have leveled off, though."  The trigger for expanding investor share was the negative change in sentiment among homeowners, and the leveling off of prices in the Closed Access cities which reduced the rate of tactical Closed Access sellers.  That scene immediately cuts from the strip club to him making a phone call and saying, "Hey, there's a bubble."  What he had actually just seen  was evidence of the bust, not a bubble.

Were many of those new speculators na├»ve?  Were they late to the party?  Could we bemoan their lack of judgment?  Sure.  Can we blame them for high prices?  No.  Investor buying was somewhat elevated in 2005.  Maybe it could have added a few percentage points to the average home price at the peak.  Investor buying share was highest in 2006-2007 when prices were stable - and investor buying was, by then, a stabilizing influence on prices.  Investor share declined in 2008 and 2009, and during that period, investor defaults were probably also destabilizing, because investors are quicker to default in declining markets than homeowners are.  Then, investor activity settled in at levels in 2010 that were still above 2004 levels, again providing support in markets where homeowners were now credit constrained.

It's possible to dissect the different types of investors and speculators and to point out where there were more reckless or even fraudulent speculators, which appears mostly to involve investors who claimed to be homeowners, which would cause lenders to underestimate their tendency to default on high LTV loans during a crash.  And it is possible to point to some brief points in the timeline where those investors may have been a bullish force in markets that were rising already.  But, their activity just can't be pushed back far enough in the timeline of events to attribute much of the aggregate national value of real estate to them.

Through the main characters in The Big Short, we can see how easily this can lead public sentiment astray.  There were many people who had been calling the market a bubble for years by 2006.  They identified themselves as people who new the value of things and who could be more wise than the average investor about avoiding poor investments.  That's a great identity to have, and for the main characters in The Big Short, it appears to be plausibly accurate.  But, it is just a short step from that to a posture of attribution error - I do things because of the constraints I face, but other people do things because they are greedy or reckless.  Multiply that by a few million people who sit down and watch The Big Short, and think, subconsciously, "I know value.  I identify with these characters.  We all recognize the greed and recklessness of all those background characters, which created the bubble."

That sentiment was part of a positive feedback loop that led to widespread blame on speculating and lending, and that blame only strengthened with each year of rising prices.  So, when migration stopped and these investors and speculators became a noticeable part of the market, it didn't look like a sentiment shift.  It looked like more of the same.  More of those other people acting on greed and recklessness.  And, the tricky part is, many of them were acting on greed and recklessness.  But, that doesn't change the fact that they didn't cause the bubble and that, in reality, they were a red flag signaling a coming crisis.

Instead of clamping down on credit and money growth, we should have been aiming for stability.  We should have been adding nominal support for these markets that were about to be hit with a migration whiplash.  What about moral hazard, you ask?  I suppose that if we had done that, some of this "dumb money" would have been somewhat better off (although, even a moderately accommodative credit market and monetary policy at the time would not have been likely to reignite the migration event.  Las Vegas and Phoenix would have likely still seen some price retraction.)  But, there is no benefit to punishing the "dumb money".  "Dumb money" didn't cause the bubble.  The bubble drew in "dumb money".  Hurting those late-cycle speculators did nothing to prevent a future bubble.

It seems to most people like it would, because those late speculators just seem like one more fish in a school that includes Alt-A homeowners in San Francisco in 2004, Fannie and Freddie borrowers in 2002, and new young first-time buyers in 1999.  Moral hazard is not why the median home in Los Angeles was selling for over $600,000 in 2006, though.  In fact, it is the opposite.  Prices in LA were that high because anyone who wants to build some housing units must first spend the better part of a decade addressing every single possible objection to building housing units.

Tuesday, September 18, 2018

The Wall Street Journal gives my work a shout out.

Holman Jenkins at the Wall Street Journal has noticed my work.    He does a good job in the first couple of paragraphs of laying out the basics of the work and tying it into Scott Sumner and the market monetarists' point of view on the Fed and the crisis.

Looks like word is starting to get out.

Friday, September 14, 2018

Housing: Part 320- Debt Growth and Home Price Appreciation

When I was thinking about the previous housing post, I decided to revisit some basic data on debt outstanding to get a sense of the relationship between debt and home prices.  It appears that, as with many measures pertaining to this subject, the story the numbers tell flips upside down, depending on if you look at it from a national level or from a more local level.

Here I am using the New York Fed Quarterly Report on Household Debt and Credit (Total Debt Balance Per Capita By State, Chart 20) for the debt measure, and the All Transactions House Price Index from the FHFA for the home price measure for various states.


www.idiosyncraticwhisk.blogspot,com   2018
In this graph of national measures, I have also added the national S&P/Case-Shiller home price index.  Since the housing boom was largely facilitating the movement of Americans away from expensive cities, the S&P/Case-Shiller index of all existing homes rose more than indexes based on sales of homes, especially during the boom.  The S&P/Case-Shiller measure moves more in line with rising per-capita debt levels until 2006.  It is probably the case that the all-transactions measure understates changing home values, because Federal Reserve Flow of Funds data during the boom don't point to rising leverage, suggesting that prices and debt were rising in parallel.


www.idiosyncraticwhisk.blogspot,com   2018
In any case, using the FHFA average price measure, it appears that, at the national level, before 2004, debt was rising faster than the prices of homes for sale.  Then, after 2005, debt continued to rise, even though prices levelled off.  This appears to support the idea that rising debt fueled rising prices and also that rising price, then, led to more rising debt, in the classic positive feedback of a bubble.

The New York Fed provides debt data on several states, and comparing per capita debt among states seems to confirm this story.  States where per capita debt was the highest in 2008 were "bubble" states - California, Nevada, Arizona, New Jersey, Florida.

www.idiosyncraticwhisk.blogspot,com   2018
But, what happens if we compare debt levels to home prices?

For the following graphs, I have indexed both home prices and debt levels to 1 in January 1999 to compare relative changes over time.  In this next graph, in each state, I index the ratio of per capita debt / average home price to 1 in January 1999.  On this measure, the relative order of the states is flipped upside down from the basic measure of debt-per-capita.  Here, which is a broad estimate of changing leverage, it is the non-bubble states where leverage increased during the boom - Illinois, Michigan, Ohio.  It's only after prices fall in the bubble states that debt/price levels rise.  In fact, in the bubble states, debt/price levels were slightly declining during the boom.

Research has shown that, in the aggregate, homeowners harvest about a quarter of new home equity gains.  Comparing the change in home prices over time to the change in debt, it appears that this data reflects that tendency.  The next graph is a scatterplot comparing the change in debt to the change in home prices in each of these states over various periods of time.  In each case, for each percentage point increase in home prices in a given state, debt rises by between 0.2% and 0.3%.


www.idiosyncraticwhisk.blogspot,com   2018
Over time, we should expect the relationship between debt and price to be close to 1:1.  This is not because of equity extraction, but simply that if leverage levels remain fairly stable over time, then if, over a long period of time, property values double, we should generally expect debt outstanding to double too.

What's interesting is that leverage over the 1999-2005 time period was relatively stable.  But, what we can see here is that, apparently, the stable level of national leverage was really a mixture of places where prices were relatively stable but leverage was rising; and places where prices were rising and leverage was declining.

A hypothetical state with no change in home prices would have expected debt per capita to rise by about 50% from 1999 to 2005 and about 20% from 2005 to 2011, for a total of about 70% over the total period.

(An aside: Remember back to the first graph.  It could be that the all-transactions measure of home prices was understated by about 20% in 2005, and reconverged with other price measures by 2011.  In that case, if we could use other price measures, the 1999-2005 plot (blue) would move right by about 20% and the 2005-2011 plot (red) would move left about 20%.  This would pull their y-axis intercepts closer together.)

So, according to this measure, if there was a debt bubble, it was concentrated in the places with the least price appreciation.  The subtle issue here is that there never would have been a moral panic in favor of watching home prices drop by 20% or 30% based on higher leverage in states with stable home prices.  The early rise in foreclosures in 2007 did emanate from Michigan and Ohio.  But, this was related to local economic problems, and, in fact, debt growth in those states from 2005-2007 was quite a bit lower than the US average.  It was speculation in places like Phoenix that led to complacency about "disciplining" the housing market.  In fact, if the focus had been more on working class households losing their homes in the rust belt in 2007, maybe public sentiment would have been more counter-cyclical.  Maybe that would have fed a more typical populist response in favor of inflation.

This lines up with a separate analysis I have done regarding price and rent.  Across metro areas, changing prices from the 1990s to 2005 correlate pretty strongly with changing rents.  And, if you assume that home prices have a moderate sensitivity to real long term interest rates, then interest rates basically explain the change in home prices across the country and rent inflation explains price changes in local hot spots.  Using such a model, prices in 2005 are not out of line.  They reflect interest rates and rent inflation.  In order to make prices in 2005 look high, in the aggregate, you must assume that home prices are not sensitive to real long term interest rates.  The relationship between rent inflation and price remains, regardless of the sensitivity to interest rates.  And, since rent inflation has little effect on home prices in Texas and Ohio, then, interest rate sensitivity is more important to prices in low priced places than it is in high priced places.

In other words, given the undeniable relationship between rent inflation and price inflation, in order to believe prices were too high in 2005, you must believe that it was places like Texas and Ohio where prices were especially too high, not California and Massachusetts.  And, similarly, as shown above, leverage rose more where prices were more moderate.  It appears that if one is to believe that debt was the driving force in the housing market, that would need to be squared with the fact that prices didn't seem to be highly sensitive to changing debt levels during the boom, and, as researchers like Mian and Sufi have pointed out, some of the effect is from causation in the other direction, where rising prices lead to cash out refinancing.
www.idiosyncraticwhisk.blogspot,com   2018

Here, it may be worthwhile to look at the 2005-2011 period more closely.  There was still an expansion of debt from 2005-2007.  Here, we can see that there was a large difference between states during that period that was unrelated to the concurrent change in home prices (the blue dots in this graph).  But, in a plot of the 2005-2007 change in debt against the 1999-2005 change in price (red dots), the correlation is quite strong, and similar to other periods.

During that time, there was a combination of two factors - the harvesting of equity, as noted by Mian and Sufi; and the slow tendency toward an equilibrium leverage level that I mentioned above.  Surely, over long periods of time, we should expect debt levels to rise at a similar rate as price levels as new mortgaged buyers replace older owners.  In other words, even after prices stopped rising, we should expect debt to continue to rise in the places where prices are the highest as new leveraged owners enter the market and as existing owners harvest home equity.  In any natural long term scenario, this should continue over time until the regression line approaches a slope of 1.

Then, from 2007-2011, prices and debt both sharply moved into negative territory and the relationship steepened.  This was the period dominated by foreclosures, short sales, etc.

Looking back at the previous graph, for the entire period from 1999-2017, the relationship remains fairly stable.  If prices have gone up an additional percentage point over that long period of time in a given state, per capita debt has only risen by about 0.2%.  This shouldn't be the case.  That should tend toward 1:1.

The lack of a strong relationship is clear if we look at each individual state, over time.  In each of the following graphs, the black line is the US aggregate number, and the colored lines are individual states.  I have categorized them, roughly, by the type of market, although the "Closed Access" states are really a mixture of Closed Access metros, Contagion, and Rust Belt areas.  In each graph, I have clearly marked the 4th Quarter of 2003 and 2005, to get a sense of where each state was at those points in time.


www.idiosyncraticwhisk.blogspot,com   2018
Remember back to the first graph, also.  These graphs are based on the all-transactions home price measure, which may be somewhat understated from 2003-2007, so the aggregate US measure is probably the useful comparison to use as an estimate of how leveraged households were becoming in each state, and with a more comprehensive measure, it may have moved more in line with the 45 degree line until 2006 (stable leverage).

By 2003, leverage (relative to 1999) was high in Michigan and Ohio (where home prices were especially low) and was low in Texas (where prices were near the national average).  During that period, it appears that price and debt were inversely related, if anything, and where debt was high, it was likely the result of households in economically challenged locations using home equity as a financial safety net.  Note also that Nevada and Arizona had roughly moved with the national average in terms of both debt and price over that time.


www.idiosyncraticwhisk.blogspot,com   2018
From 2003 to 2005, prices accelerated in all states.  In Ohio and Michigan, debt growth since 1999 retracted back toward the national average and home prices didn't rise as sharply as in other places.  In Texas, both prices and debt levels remained relatively low.

During this period, debt levels did rise more quickly than average in Nevada and California.  In other states, debt rose at a rate similar to the national average.  There is no systematic difference here, regarding debt, between the Closed Access and Contagion states and the other states.  In general, the "bubble" states didn't move up during this period, or even move diagonally along the 45 degree line.  They moved horizontally to the right.  Valuations changed in those states, but there is little sign of systematic differences in debt levels.

After that, leveling off and declining prices dominate the behavior, so where prices had previously risen more, debt continued to rise more as prices stabilized, then where prices declined more, debt declined more along with them.  This creates a pattern of concentric circles around the national average, especially in the Contagion states.  It is the lagging nature of debt growth that creates that counterclockwise shape.  And, after all of that, generally, states that had higher appreciation from 1999 to 2005 have had prices rebound so that they have more price appreciation today than the US average (with the exception of Nevada) and they have less debt than the US average (with the exception of New York).

The only other states with per capita debt growth higher than the national average from 1999-2017, besides New York, are Texas and Pennsylvania.  Those are also the two states who had the smallest price shocks after 2005.

There is little evidence here that debt was an important causal factor in systematic differences between states.  There is some evidence of price as a causal factor in the differences in debt growth between states.  I would suggest one other factor that seems important here, and that is property taxes.  There seems to be a relationship between higher property taxes and less volatile housing markets, mostly because higher property taxes moderate prices when they are rising.

Thursday, September 13, 2018

August 2018 CPI

CPI of all items less food, energy, and shelter for the previous 6 months (annualized) is about 0.6%.  I suspect that the year-over-year measure will fall back to below 1% over the next 6 months, but rent inflation will keep core CPI near the 2% target.

This increases my confidence that the remaining rate hikes in 2018 will be contractionary.  Forward rate markets seem little affected, though.

Wednesday, September 12, 2018

Housing: Part 319 - "Talk of a Coming Decline"

Here is a Wall Street Journal article from February 2002, headlined, "Housing Market's Sustained Growth Prompts Talk of a Coming Decline" (HT: Nick Timiraos)

It's a good example of how strong the idea of a predetermined bubble/bust was.  Even by 2002, this was becoming the canon, presumptively.  By the time the bust came, there was a broad exasperation with the supposed bubble that had refused to bust, in spite of years of insistence that it must.

Here is a graph from OECD data of home price/income measures for several countries.  The difference between the US and the other countries is that when we became enamored with a communal case of attribution error, we were singularly capable of imposing a crisis on ourselves in order to play out the national narrative we had constructed.  And, as we imposed that narrative on our housing market, the attribution error that led us to create the crisis also led us to blame the scapegoats that attribution error had created.  So, to this day, practically any article from any point of view about what happened during the crisis will begin with the presumption that lenders and speculators did this to us by creating a bubble that inevitably would crash.

There are a handful of countries that have had stable home prices - Germany, Switzerland, Japan, etc.  But the countries most like the US, in a number of ways (higher income growth, trade deficits, lower manufacturing employment, etc.), are the countries shown above.

Thursday, August 30, 2018

Housing: Part 318 - Affordability

Almost anyone you hear talk about the housing market today, and the recent levelling off we have seen in some measures, will mention "affordability" as a headwind on the market.  In some ways, this is true, in the same way that there is an "affordability" issue in bonds.  To get $5 of cash flow next year, you have to invest a lot more in bonds than you used to.  The same is true for housing.  If you are buying a house with cash, in many markets, you have to invest more cash in order to earn the same level of rent (or imputed rent) than you used to.

But, since this is true of both homes and bonds, then for the borrowing homebuyer, the returns are better than ever.  This is especially true in low tier neighborhoods.  Zillow.com tracks affordability by tier (bottom, middle, and upper).  Here are the measures for the US and for Atlanta. (Most cities look, more or less, like Atlanta, in this respect.)

For borrowers, homes are more affordable than they were at any time before 2009, especially in bottom tier markets.  In the Atlanta bottom tier, before 2009, the median mortgage required about 30% of the median income.  Now, it requires about 19%.  It isn't even close.  Affordability is absolutely not a problem.

Here is a random home in Atlanta, that, as of today, is for sale for $99,700. That equates to a monthly mortgage payment on a conventional mortgage of about $415.  The Zillow rent estimate is $1,150.  The posting states: "Expand your rental portfolio. Long term tenant wants to stay. Don''t miss out!"

(This isn't the case for middle or upper tier homes, where mortgage and rent payments tend to be more similar.  Some of this might be maintenance on older properties.  But, much of it is regressive tax benefits for owners that inflate upper tier prices, higher management and vacancy costs in low tier rental markets, and higher tenant risk.  Many of those costs and risks can be eliminated by a low-tier owner-occupier who can remain in a property for a long period of time and treat it well.)

I ask you, for the long term tenant who wants to stay, is there an "affordability" problem with the $99,700 price tag?  I would argue that the primary goal of a functional housing financing system should be that it provides a way for a long term tenant who wants to stay in a home like this one to do just that, as an owner.  And while our financial system fails aggressively at that function, the conventional beliefs about the housing crisis are so wrong, they prevent these patterns from being acknowledged or noticed.

The relative affordability of bottom tier housing stock is a natural social leveler - a way for households with low incomes to get higher returns on their investment than savers with more capital can.  In order to do this they need to borrow from those wealthy savers who must accept lower returns on their mortgage securities than the homeowner receives on their bottom tier home equity.  But, the "predator" narrative and "the business model of Wall Street is fraud" narrative, along with the "bubble" narrative and the "keeping up with the Jonses" narrative all lead us to impose public obstructions that prevent millions of households from engaging in this prudent, lucrative form of savings.

When you see stories about the housing market, notice how often you hear "affordability" mentioned as a problem today.

Friday, August 24, 2018

Housing: Part 317 - Unsold inventory during the financial crisis

To follow up on the previous post, here is a Fred graph comparing various measures of homebuilding.

Source

 Sorry, it's a bit messy.  The orange line at the bottom is homes built speculatively which have not been started yet.  The dark blue line at the bottom is completed homes that have not been sold.  This is one source of the "overbuilding" story.  Speculative building increased, basically in line with general growing sales.  The trend in speculative buying reversed soon after general sales started to decline, but it wasn't enough, and so homes that were completed without a buyer started to rise.  The green line is speculative homes under construction, and it begins to decline in 2006, but not quickly enough.  That is the extra inventory that might be blamed on overbuilding.

But (You knew there would be a "but"), notice the scale of the problem.  Even at 200,000 units, the level of unsold inventory should have amounted to just a few months' sales.  The only reason they amounted to much more than that was because sales had collapsed.  It was the collapse in demand that was the shock, not an oversupply.

In the next graph, the blue line (left scale) is months of inventory of new homes.  The red and green lines (right scale) are new homes sold (red) and inventory of vacant homes among existing homes.  Notice that the inventory of vacant homes also shot up at exactly the same time that sales began to collapse.  There was an event at the end of 2005 that led to a sharp collapse in demand for homeownership.  Remember, though, there was no shift in demand for the service of housing (relative to supply) because tenant vacancies remained low and rent inflation was high and increasing.
Source
Source
Or, maybe, this was the result of all those buyers who had been suckered into buying homes they couldn't afford.  Maybe they were getting foreclosed on and that was the source of vacancies.  But, again, notice the timing.  Vacancies started to rise in 2005.  They were already at the top of the range, over 2 million units, by the 4th quarter of 2006.

Here is data from the New York Fed Report on Household Debt and Credit.  When the vacancy level of units for sale had already reached more than 2 million units, in the 4th quarter of 2006, there were about 220,000 foreclosures.  That compares to an average of about 176,000 foreclosures from 2003 to 2006.  The foreclosure crisis mainly happened from 2008 to 2010, when every quarter more than 400,000 units were foreclosed.  When the vacancies developed in 2006, foreclosures had just barely begun to rise.

By the time the foreclosure crisis happened, builders had already contracted to practically nothing.  New building had contracted so much by the time foreclosures were happening in large numbers, that the number of unsold new homes was actually declining back toward normal levels by then.  Builders were actually quite responsive to the shock in demand.  The flooding of foreclosures onto the market didn't eat into builder sales.  Their sales were already collapsed by the time that happened.

But, here's the kicker.  The Census Bureau doesn't track cancelled orders.  So, many of those homes in the first graph that are measured as new houses sold and completed were actually homes that were sold, completed, and then cancelled.  Cancellation rates are normally about 15-20%, and they roughly doubled over the course of 2006.  So, a little more than a million homes were sold in 2006, but about 400,000 of them turned into unsold inventory instead of new, occupied stock.

Where the Census Bureau showed 200,000 finished but unsold homes in 2007, it was actually probably more than double that amount, if cancelled contracts are included.  Now, that inventory really doesn't have anything to do with subprime lending terms, etc., because new home buyers generally just have to put down a small escrow, and the mortgage isn't closed until the house is finished and the buyer takes ownership.

The homebuilders were quite disciplined, really.  From peak to trough, annual sales declined by about 1 million units, twice the change of the housing contractions of the 1970s.  Yet, months of inventory didn't rise much above the months of inventory that had developed back then, and it quickly declined from its peak, even though there were also hundreds of thousands of units of shadow inventory from cancellations were also on the market.

It would be interesting to see detailed research on this.  Was the rise in cancellations because the buyers were tactically cancelling or because their lenders were pulling out before the homes were finished?  In either case, the point is that the rise in cancellations, just like the other measures, predated the rise in foreclosures by at least a year, and was responsible for most of the rise in new home vacancies.  There has been research showing that falling prices generally led to defaults, and not the other way around.  But, here we can see that cancellations were creating hundreds of thousands of new vacancies before prices declined significantly.  A demand shock led to a change in sentiment which led to vacancies which eventually led to price declines, which led to defaults.

I have commented before on the perverse self-destructiveness of public sentiment in 2007, where stabilizing policies were routinely avoided specifically because stabilization would have prevented some people from taking devastating capital losses.  But the perversity is even deeper than that.  Consider the homebuilders.

I looked at annual SEC filings from Meritage Homes to collect sales and cancellation data from the Arizona market.  Actually, in hindsight, I'm not sure if Arizona looks much different than the rest of the country on this matter.  Oddly, it appears as if sentiment shifted downward to the extreme, in some cases well in advance of actual price shifts.  For instance, for Meritage, in Arizona, cancellations jumped to 37% in 2006.  From 2006 to 2008, the average sales price in Arizona dropped by more than 40%.  By 2008, Meritage cancellations were declining back to the norm in Arizona because nobody was bothering to sign a contract in Arizona in the first place, by 2008.

In Texas, the cancellation rate was 35% in 2006 and it increased to 40% by 2008, but the average price of Meritage homes over that period was basically flat.  Finally, from 2008 to 2010, Meritage home prices in Texas declined by a little bit less than 10%, but Texans had been getting cold feet for several years before that happened.

Maybe this is because buyer quality had become so bad that suddenly nearly half of potential buyers realized by the time the house was completed that they couldn't possibly afford their mortgages? But, after 2005, when this was happening, sales were declining - first time buyers were declining pretty sharply.  It doesn't make sense that the period of time where borrower quality was deteriorating was the period where the number of buyers was dropping like a stone - unless the decline in borrower quality was from a liquidity shock.

Here are the Meritage numbers for Arizona.  In 2006, they had new contracts for 2,910 units.  Of those, 1,833 were built and sold and 1,077 were cancelled so that Meritage had new, unplanned inventory.  Then, in 2007, net orders after cancellations were 1,203.  So, after accumulating 1,077 unplanned vacant units in 2006, total net sales in 2007 were only 1,203.  In the meantime, they accumulated another 642 cancelled units, and 2008 net sales were only 884.  For years, in Phoenix, there were developments full of empty homes.

I was planning on ending this post with a discussion of systemically risky terms on subprime loans and how, even if there wasn't an oversupply of housing, those type of loans by investors were likely an important part of the panic.  And, they probably were.  But, this is a part of the story few talk about.  These cancellations have nothing to do with lending terms, or owners vs. investors.  In all cases, they would have created a small escrow fund to initiate building.  In all cases, they would have been likely to tactically put the home back on the builder.  Neither owner nor investor would have had an emotional history with the property at that time, and the size of their down payment or their interest rate terms were irrelevant to the decision.  However, lower interest rates in 2006 would have probably helped to reduce this problem, because buyers would have been more optimistic about housing markets and they would have been able to close the sale with better terms.  Surely, that would have improved the cancellation rate.

But, here is where I ask you to think about the perversity of public sentiment in 2006 and 2007.  The public discussion was all about subprime borrowers and homeowners who were in over their heads.  The public discussion was all about overbuilding and oversupply.  It was about corporations using predatory tactics against regular people.

And, the primary initial shock in 2006 and early 2007 was among homebuilders.  What were homebuilders doing?  First, in Phoenix, they couldn't get permits fast enough to meet demand.  So, rather than overbuilding, in 2005 they were holding lotteries at the new developments to determine which buyers could order a new house.  They were clearly not able to build at a fast enough pace to meet demand.  Then, they got signed contracts from qualified buyers who ordered new homes.  Only after they had a contract and an escrow account for a new home did they begin to build.  And, at that time, it was probably generally several months before they even broke ground, since they were running at capacity.  Far from being predatory, builders take a generous approach.  They take the construction risk.  They finance the construction, and they generally give buyers an easy out to walk away if there is some reason why they don't want to complete the transaction in several months when the house is ready.  Even in good times, cancellations frequently run over 10%.

So, the homebuilders were eating all of this inventory and taking it on the chin.  The Fed had steered the economy in this direction and had seen the decline in residential investment as appropriate, then in 2007, they held off on lowering rates because doing so would let speculators off the hook and would fail to discipline the marketplace.  At that point, even though prices in much of the country remained stable, new home sales had collapsed.  And, homebuilders were sitting on inventory of hundreds of thousands of homes.  Their sin was that they had built homes for paying customers who had signed contracts to order them.

What were they supposed to do?  Should we have a policy that homebuilders should refuse to build homes for paying, qualified customers who fund escrow accounts?  Should we have a policy that homebuilders must require buyers to put down escrow amounting to 20% of the home prices and wait for 9 months before they can move in, so that builders have more power over buyers during the construction process?

The only policy that fixes this problem is to provide accommodation sometime over the 18 months between early 2006 and the 2007 panic.  Lower rates to, say, 4%, instead of 5.25%.  Or, lacking that, in 2007, let people off the hookActually help stabilize markets that are panicked.  God forbid, maybe even allow some speculators and lenders to avoid financial ruin in the process.  By the time the Richard Fisher of the Dallas Fed told the FOMC in September 2007, "I’m very concerned that we’re leaning the tiller too far to the side to compensate risk-takers when we should be disciplining them.", the homebuilders had been disciplined for a good 18 months in the face of sharply declining nominal investment and sentiment.  To the extent that speculators were involved in that shift in activity, they had already taken their small losses by surrendering their escrow funds.  And, by September 2007, the homebuilders were far along passed the time where they needed to be reminded to be careful about speculative building.

PS.  Here is a chart from the Calculated Risk link that shows cancellations across several builders, which has the same pattern I found at Meritage.

Wednesday, August 22, 2018

Housing: Part 316 - The phases of the bust.

There was one answer from my podcast interview with David Beckworth at Macro Musings that I feel like I sort of flubbed, so I am clarifying the issue in this post.  David asked me about the early shifts in the migration event and what caused it to stop.

I have covered the timeline before, so some of this might be a repeat, but I think my thinking has evolved a little bit on this as new information comes in, so there might be subtle differences and additions to previous posts.  For this topic, my moving chart of home prices over time is helpful.


1990s to 2003:
For this period of time, supply constraints dominated.  Cities that were more expensive to start with were generally becoming even more expensive over time, because the reason they were expensive was that they lack sufficient housing and so rents are high and rising.


November 2003 to June 2004:
For this brief period of time, home prices in Los Angeles accelerated relative to other places.  It is plausible that the newly expansive private securitization market, which was especially active in California, was the source of this unusual price appreciation.  Added demand from new homebuyers in Closed Access cities created a surge in the out-migration that tends to come out of these cities.  Since the housing stock is relatively stagnant in Closed Access cities, more housing demand from one household means less housing stock is available for other households, and someone has to move away.  The price appreciation appears to have triggered a boost in the outmigration of homeowners, who were cashing out as prices increased.  So, there was an increase in migration among both owners and renters.

July 2004 to November 2005:
The Federal Reserve started to raise interest rates in July 2004.  This was probably appropriate purely as a monetary policy shift, and the economy, in general, was still clicking.  So, Phoenix had two forces pushing on its housing market.  Aspirational migration was strong.  In a typical year, about 50,000 households move into the Phoenix area from places that aren't "Closed Access".  During this time, the rate of migration from other places moved up slightly to about 60,000 by 2005.  Those were probably mostly households moving up from less expensive places or moving to Phoenix for retirement, as is typical.  That increase was due to a strong economy, flexible lending, etc.

From 2002 to 2005, migration from Closed Access and other Contagion cities increased from about 12,000 to 21,000.  These were households moving down-market.  They were moving to Phoenix to lower their housing expenses.  There would have been several forces at play with this group of movers.  Loose lending in California was a force that was inducing the outflow.  The high prices that had developed in California were inducing more selling and migration from homeowners.  And, rising interest rates were likely adding to that selling pressure as homeowners expected rising rates to eventually be a drag on rising prices.  So, there was an extended period of time where continued economic growth and changing sentiment in the California housing market were both putting upward pressure on the Phoenix population and on its housing market.

This is clearly visible in the moving chart, as Los Angeles price appreciation moved back down to the national norm, even though private securitization mortgages were very active there.  At this point, the added demand those mortgages created in California was mostly creating new local housing supply by inducing out-migration and selling by existing owners.  Rising rates probably had something to do with that.  So, rising interest rates and selling pressure were pulling down prices in Los Angeles, where credit-fueled demand might have otherwise pushed prices a little higher and they were pushing prices up in Phoenix because of the migration that ensued.  (Note that any effect here was purely from expectations or sentiment, because 30 year mortgage rates and long term interest rates in general did not actually rise until the Fed had flattened the yield curve in late 2005.)

November 2005 to August 2007
By November 2005, the Fed had increased the Fed Funds Rate high enough to flatten the yield curve.  I would argue that at these levels (about 5% at the time) that is basically an inversion.  As is clear in the moving chart, there was a uniform national reaction in the housing market.  All measures began to collapse everywhere.  Sales, housing starts, etc., and inventory of homes for sale rose.

Now, from this new perspective, which says that there were never too many homes anywhere and that home prices in 90% of the country generally reflected fundamental factors like rising rents and low long term real interest rates, this is a huge red flag.  Clearly, the economy was already in a state of dislocation.

This was clearly a monetary event.  I don't even think that is a point that is open to debate, for several reasons.  First, as I said, the premise is doing all the work here.  If the premise changes so that we don't presume that there were too many houses selling for too many dollars, then the shifts in the housing market were extreme and moving in unison across the country in an extreme contraction that wasn't called for.  This was not a subtle shift.  Second, the Federal Reserve, itself, considered the downturn in the housing market to be a product of their policy choices, and they were pleased by the downshift at the time.  They were very focused on housing markets at the time and mentioned them in nearly every FOMC press release.  Third, many economists who have commented on Fed policy at the time attribute high home prices to loose monetary policy, they attribute some power to the Fed to be able to pull prices down, and many of them argue that the Fed should have targeted home prices earlier and tightened policy earlier in order to do it.

There is a question as to what mechanism was at work.  Sentiment? Money supply? The effect of yields on valuations or lending?  (One interesting facet here is that mortgage growth was a fairly late factor, still growing slightly even when the 2007 panic hit.  But, home equity had been declining for some time, and much of that decline appears to have come from unmortgaged owners selling and exiting the market.  According to the Survey of Consumer Finances, in 2001, 23.1% of American households owned homes with no mortgage.  That declined to 19.9% by 2007.  So, even though there is an abrupt shift in price trends when the yield curve flattened, it appears that sentiment and money supply were more important than lending in the early phases of contraction.  I consider an inverted yield curve to be an important forward contractionary signal, but what exactly causes the signal and what the inversion itself causes to happen remain a mystery.  The 2005-2007 period, to my mind, both confirms the importance of yield curve inversion and adds to the mystery of what is actually happening.  Maybe these flows out of home equity as an asset class and into other forms of long term fixed income are an important part of that puzzle.)  But, as to whether those mechanisms were related to monetary policy, there is no argument left here.  Once the premise changes, the conclusion changes with it.  The Fed put a stop to the migration event as a side effect of causing a housing contraction.  That is stipulated.  The argument is whether they should have.  Before blaming the Fed too much, keep in mind that in 2007, nearly everyone thought they should have.

Source
Source
So, the collapse in migration that led to an especially tough housing collapse in Phoenix was actually part of a national negative monetary shock.  Both nominal and real GDP growth were marginally at levels that have normally been recessionary.  And, as shown in the chart here, employment growth, which had not been particularly strong, started to decline in 2006.  Nationwide, it didn't decline sharply like it usually does in a recession.  There is a typical recessionary downshift in the Contagion cities in 2006. (Here I show the Phoenix metro area and the state of Florida.)  Because the economy had become so characterized by these inter-metro migration patterns, the initial result of this wasn't a rise in unemployment.  It really is quite striking.  With such a negative shock to employment growth in Phoenix in 2006, the unemployment rate didn't bottom out until June 2007!  All of the shocks to employment markets were buffered by a reversal in migration flows!  In Contagion cities, the shock didn't lead to higher unemployment.  It led to an end of the migration event.  In other cities, employment didn't collapse in 2006 the way it normally does at the beginning of a recession because at the same time that employment growth was slowing, population growth got a boost because households stopped moving to the Contagion cities.

In the Contagion cities, the inflow of households slowed down.  I would expect that the retrenchment among renters with lower incomes was related to declining employment opportunities.  The retrenchment among home sellers was likely related to the fact that home prices had levelled off.  As prices had been rising, on the margin, more homeowners had been selling, collecting their capital gains, and moving away from Closed Access cities.  But, when prices levelled out or started to fall slightly, that would have induced many fewer tactical sales.  The inflow of both households and capital to the Contagion cities slowed down as a result.

At the same time, the outflow of households out of the Contagion cities to other places that were now more affordable continued to rise, so net inmigration fell dramatically.  The reason for the outmigration is because the liquidity shock had thrown the construction market into disarray, and the Contagion cities, who already were struggling to build enough homes to meet the demand from in-migration now had a terrible shortage of homes themselves.  Inventories of homes for sale had shot up, because the liquidity shock had created barriers to home buying, and vacancies of owned homes increased in the Contagion cities, related to that.  But, rental vacancies remained low for some time after the shock in 2006.  In Phoenix, rental vacancies were low until 2008.

Prices, nationally, should never have declined at all, and when they finally did in the summer of 2007, it was only after this had been going on for a year and a half, housing starts had retracted as far as they could to buffer the decline in demand triggered by the disrupted buyers' market, and when declining sentiment led to the panic in privately securitized mortgages, the general mood of the country was that it wasn't anyone's job to right the ship.  Political sentiment made the unnecessary price collapse inevitable.

So, what specifically caused the migration event to end?  Monetary policy has a lot to do with it.  Essentially, a proto-recession had already been induced, and it hit the Contagion cities the hardest because their home prices had been driven up by that migration event.  For a year and a half, during that proto-recession, relative to previous trends, recessionary conditions were mitigated in the rest of the country because they were matched by a temporary population boost.  Whereas Austrian school economists might argue that loose money induces unsustainable demand that leads to a boom and bust, what might have been happening in 2006 and early 2007 was that money was relatively tight, but that non-Contagion markets were induced into unsustainable demand that kept them at the margin of recessionary conditions because they had a temporary boost in demand through population flows.

Home prices, even in the Contagion cities, remained relatively strong as housing starts collapsed along with those migration flows.  The reason is that the Contagion cities didn't have enough homes.  It is possible that if things had turned out differently, and the migration event had wound down without being induced by a monetary shock, that home prices in the Contagion cities would have dropped a bit.  It is clearly plausible that home prices there had been boosted by cyclical pressures related to the migration event.  And, it is plausible that if the Fed had been accommodative in 2006, home prices in California would have still peaked, cutting off the flow of wealthy home sellers into Phoenix, and plausibly, the economy might have grown while the Phoenix market corrected.  But, that isn't what happened.  The Contagion cities never had too many houses, so in 2007, rent inflation was high and gross out-migration was rising, and prices remained relatively stable until the country accepted the development of a series of financial panics.

Accommodative monetary policy in 2006 would have been very useful, and one important result of it would have been stabilizing the housing market in the Contagion cities.  Don't get me wrong.  I'm not saying that the Fed should be in the business of stabilizing prices in various asset classes.  The primary reason that it would have been important to see improvements in the Phoenix housing market wasn't to ensure that investors or lenders didn't have losses.  It wasn't even to create employment growth in construction.  The reason it would have been important is because the Contagion cities needed houses and the fact that they couldn't manage to build them and that buyers couldn't manage to fund them was already a signal of disequilibrium.  Given public sentiment at the time, it would have been impossible for the Fed to have seen this in real time, or to have acted on it.  But, in hindsight, we must understand what happened.

Given that in 2007, we had the premise wrong, and we thought that a credit bubble that was destined to bust was our problem, one can imagine how Federal Reserve officials could have read these signals as positive.  They had managed to slow down the housing boom.  They thought that was the right thing to do.  And they thought that their biggest challenge was going to be limiting the damage to the rest of the economy.  In May 2007, when Ben Bernanke was assuring the public that the subprime crash would be contained, he was looking at the economy in Phoenix that had just had a massive shock.  It's housing market was DOA.  And unemployment there was at 3.2%.

Follow up.