Monday, March 14, 2016

Housing: Part 127 - Arguing over false premises

The funny thing about trying to read some small portion of the countless books that have been written about the housing bubble is that they are full of facts that are useful for my contrarian point of view.  By about 2001, everyone had agreed that the eventual bust of the so-called housing bubble would be caused by overleveraged speculation.  So, by 2007, when we had finally engineered the bust that would prove all of our intuitions right, and show those lousy speculators how foolish it is to expect 21st century institutions to support stability, we didn't need to convince one another about the premise.

So, frequently, observers will make arguments that are really arguments against the premise, but they are only presented as arguments against one of the stories that are based on the premise, because the premise was canonical before it was falsifiable.

The fact that commercial building mimicked the behavior of residential housing is used to argue that the GSEs couldn't have caused the bubble.  But, advocates of that story want to blame private investors for the bubble, so they don't apply the argument against lenient retail mortgage or subprime securitization.  In fact, they should be pushed even farther, because the commercial market strongly suggests there was no bubble at all.  Investors in shopping malls, multi-unit housing developments, and offices are about the least speculative people you'll ever meet.  Even now, when capitalization rates are strong even while bonds are tickling the zero lower bound, the real estate financial analysts I see generally only talk about the high income potential of real estate investments.  I don't think there were many REITs in the market to flip apartment complexes.

Arnold Kling recently repeated one of these types of arguments when he explained that mortgage relief programs wouldn't have accomplished what their backers were hoping for because many of the late homebuyers who defaulted were speculators, not owner-occupiers.

He adds:
(Y)ou face a trade-off. Give the borrowers too generous a bailout, and you generously reward the profligate.
And, he links to a previous post with the following observations:
Loans to speculators were made by Freddie and Fannie. Loans to speculators were eligible to be laundered into AAA securities that were favored by government capital requirements. The sad fact is that the real estate lobby was so good at playing the violins for "home ownership" that they were able to put a smokescreen over a wave of speculative borrowing....
Anyone who wants to stop mortgage foreclosures needs to have his head examined. How many of the bad loans are investor loans, where the borrower never occupied the house? 20 percent of them? 50 percent? 70 percent? We know that in the last years of the bubble more than 15 percent of mortgages were for non-owner-occupied (the true figure might actually be higher than reported, because it is common to fraudulently claim that you will be using the home as a residence when you will not). Investor loans default at a much higher rate than regular loans, somewhere between 3 and 10 times as much. If it's 4 times as much, then already we can be surmise that a majority of bad loans are investor loans. The best thing to do with those is to foreclose ASAP.
So, the story that the peak of the bubble was facilitated by predatory banks pushing subprime loans on low income owner-occupiers is wrong.  The peak of the bubble was facilitated by investor households with multiple properties.  They were, according to Kling, engaging in massive fraud, also.

But, this is a much different story than the predatory lender/over-leveraged owner-occupier story.  Kling's story points back to the GSE's and fraudulent investors.  Isn't it weird how there are two mutually exclusive versions of the period, but they both include rampant fraud?  Profligacy was the premise.  We know there were villains.  Somebody must have been committing fraud, because prices were so high.

The Mian & Sufi story is that high wealth households lend and low wealth households borrow, and so low wealth households were wiped out when equity values in homes crashed, and the wealthy households that owned their mortgages came out on top.  So, this is just another example of the rich getting richer at the expense of the poor.

But, that kind of sounds the opposite of Kling's story, too.  Doesn't it?  Everybody has a villain.  Everybody has cynicism.  Everybody sees profligacy.  Everybody sees inevitable bust.  And it confirms everyone's sense of injustice or poor policies, even if some of those confirmations contradict others.  So, we argue about those contradictions.

The one thing everyone agrees on is that we absolutely could not have supported a policy of price or monetary stabilization.  Stabilization rewards profligacy.  You'd need to have your head examined to stop foreclosures, says Kling.

Here are some statistics from the Survey of Consumer Finances.  About 75% of mortgages for primary residences were held by the top 40% of households, by income.  About 85% of mortgages for other residential properties were held by the top 40% of households - more than half by the top 10% of households by income.  Aren't these the households that would be least affected by changing credit conditions?  Isn't it strange that marginalized lending due to the CRA or the GSEs or subprime lenders were the cause, and yet it was households least in need of generous credit terms that were the source of demand?

The proportion of mortgage debt held by households that was for non-primary residences did rise from about 8% to 12% from 2001 to 2007.  But, the total mortgages outstanding in the GSE and Ginnie Mae pools grew by less than 1% in 2004 and only by 3% in 2005 before rising by about 8% in 2006.  Part of the reason investor buying took over and private pool securitizations took over was because financing for owner-occupiers through the GSE's and Ginnie Mae was drying up.

One of the worst side-effects of centralized economic policy management is the problem of attribution error and the human bias of attributing outcomes to our innate sinfulness.  The gods have always been angry with us.  There is this terrible, terrible idea that seems to be widely believed among practitioners, academics, and policy makers, that investors become complacent if we have long periods of stability, and that a good dose of pain is, in the end, useful.  So, we all congratulate ourselves for economic malpractice.

There is literature that purports to document many past episodes where this lack of risk aversion leads to a bubble and bust.  But, as far as I can tell, the consensus among the academics that work in that area is that the recent episode was a grand example of that problem.  Is there anyone that believes irrational bubbles are an important recurring phenomenon, but doesn't think the recent episode is a good example?  It seems hard to believe that the entire history of empirical evidence could rest on an error, but the recent crisis certainly suggests the possibility.

In the linked post, Kling points to research that shows a sort of contagion effect of speculation, that when people are surrounded by others who are speculating on an asset, they are more likely to speculate on it too.  I don't have a problem with that research.  I don't even doubt it.  But, there is this big disconnect I see in all of the discussions about the housing market between plausible distortions in human behavior and how those distortions might explain a tripling of home prices in California over a decade.  That's the problem with these behavioral explanations.  Once you plug irrationality into your model, scale goes out the window.  And, when scale ceases to impose discipline on our explanations, then we are free to blame the bad harvest on whatever sinful behavior was bothering us the most beforehand.

Ask yourself a question.  Is there any level of home prices that would have given you pause, where you would have said, "Hm. At this point, irrationality or bubble behavior probably can't explain this anymore."?  If the average home price had risen to $500,000, or $1 million, or $2 million, would you have decided that the scale had outgrown the ability of behavioral explanations to explain it?  It appears to me that what happened was that the higher prices went, the more convinced everyone was in the bubble story.  Now there's a model, huh?

In the version of the story I am working on, I am trying to explain the pricing behavior with supply, and there is a definite point where if prices rose above a certain level, I would conclude that supply could not explain that scale of price change.

There have been hundreds, probably thousands, of books and academic articles written on the so-called bubble.  If any readers know of a single one that attempts to do this for the bubble narrative, please let me know.  I don't know of any.  Would it even be possible?  Can irrationality have confidence bands?


  1. Excellent blogging.

    If you listen to some right wingers, the free enterprise system is an inherently unstable platform on stilts, prone to catastrophic failure after any period of extended stability, prosperity or "low" interest rates.

    And those are the right-wingers....

  2. According to Stephen Williamson, who answers my questions when Scott Sumner bans me, there was a Heloc bubble. When the commercial paper market declined, as Heloc credit lines were pulled, it caused a decline in the economy. Dr Williamson showed me some charts I used in the following article, showing that commercial bank lending never slowed, but the commercial paper market crashed completely:

    1. Sorry, commercial bank lending slowed for a short time but immediately resumed strongly in 2010.

    2. Setting the record straight, Scott Sumner didn't ban me. Maybe I posted late at night or there was some other glitch. Kevin, what do you think about Bernanke's lack of worry about negative rates?

  3. I think that Ed Glaser article ("can cheap credit explain the housing boom?") that you actually linked once was the best effort I've found at quantifying the role of credit in explaining the boom; Glaser concludes at most 20% of the price increase can be so attributed, the rest being attributable to other factors like supply restriction. But he uses a 'non-standard' model for the interest rate's effect on prices, and some of the math escaped me when I read it. In any case, I've yet to see anyone present reasons for rejecting his explanation (or rather refutation of the cheap credit explanation). It's a shame it hasn't gotten more attention. Conventional writing on the housing crisis seems overwhelmingly anecdotal rather than quantitative.

    1. He has to be wrong. Check out this most important housing chart that shows an explosion of mortgages and Helocs that caused prices to doublt and triple in big bubble areas.

    2. Thanks for reminding me of that. That is similar to what I have found. Glaeser's work seems to always make intuitive sense to me.