Friday, March 20, 2015

Housing Tax Policy, A Series: Part 21 - More about the Crisis Timeline

I laid out a timeline of the financial crisis in this post.  I followed up in this post.  But, I tend to throw a bunch of stuff at these graphs, so for clarity, I thought I'd revisit the timeline, simply showing home prices (inverted) and mortgage delinquencies, with max & mins roughly lined up, to get a sense of scale and time.

Here it is.  From early 2007 to the present, delinquencies have followed price with about a 6 to 12 month lag, even falling with the same lag and scale as prices have rebounded.


Source: Calculated Risk
We can see in the Calculated Risk graph of real home prices, that in 2 of the previous 3 dowturns, real home prices fell in the range of 10%.  There is this mythology about the 2000s housing boom that everyone suddenly got stupid and thought houses could only go up.  But, by the end of 2007, nominal house prices had dipped by about 10%, and delinquencies were at 3.1% - less than they had been in 1991.  When real home prices dipped, in 1982 and 1992, YOY currency growth bottomed out at about 6%, and the Fed started pushing the money supply back up.  In the 2000s, currency growth hadn't been as high as 6% since 2003, and by April 2008, was down to 0.5%.

So, during previous housing downturns, inflation was around 5%, so nominal home prices remained level.  Yet, in this downturn, the Fed wouldn't give us any inflation.  Core minus shelter inflation (estimated here by treating shelter inflation as 40% of core CPI and subtracting it) had not been persistently above 2% since 1997.  By the fall of 2007, it was down to 1.0%.

In the next graph, there are more general real estate delinquency measures that go back far enough to compare to the 1991-1992 episode.  They didn't reach the peaks of that period until the end of 2008 or 2009.

Yet, we describe this period as one where reckless lenders were bailed out by the Fed?  This narrative was already in place.  It would have been the Fed's job to create stability.  But, the country basically has the position (and here I can really say "we" - practically everyone seems to feel this way) that, "If the Fed does it's job, how will we ever learn?".  And, again, I truly mean "we".  Several people I know who firmly believe this bought homes in the 2004-2007 time frame.  These are smart people.  And, as I have tried to show in other posts, the prices they paid were reasonable, given their investment alternatives.

So, first, delinquencies didn't come any faster than they had come before.  And, second, home prices were orders of magnitude outside any previous experience before banks began to have widespread problems.

Bear Stearns had to extend credit to two subprime mortgage hedge funds in June 2007.  By that time, nominal home prices were down about 3% - about the worst of any housing corrections in previous decades.  They would fall more than 10% more in the following 6 months.  They would, by some measures, eventually fall by 40%.

In theory, there are systemic risks from lending with low downpayments.  If we had seen 10% delinquencies in 2007, after nominal home prices had bottomed out at 3% or even 10%, this episode might have lent evidence to that theory.

In theory, there are systemic risks from securitizing loans with low down payments and less documentation.  (Although, (1) typical FICO scores did not decline during this period, and (2) as the rough measure in this graph shows, banks were still holding plenty of real estate loans on their books.  The notion that banks were suddenly giving out mortgages without any standards and unloading them on the MBS market, in some sort of frenzy, is greatly overstated.)  If nominal home prices had bottomed out at 3% or even 10%, and some other funds had followed those Bear Stearns funds into trouble, this episode might have lent evidence to that theory.

But, this episode saw real home prices fall by 40%!  What this episode proved - the only thing this episode proved was that when home prices fall by 40%, bad things are going to happen.  This was beyond unprecedented.  Standard deviations can't even really describe the Fed's failure here.

Oh, so you think that the formulas for the CDO securities "devastated the global economy"?  Please, Mr. Hindsight 20/20, please show me where you, or anyone, was saying that these models needed to be able to take a 40% hit to real home prices.  Don't give me any nonsense about how these greedy speculators thought home prices could never go down.  That is not remotely a description of what happened.  Show me any skeptic who, before 2007, said models should be able to handle a 40% downturn.  How about 20%?

Source
Yet, even by the end of 2009, when that 40% drop in home prices had already come to pass, look at the impairment levels of these securities.  AAA tranches on Subprime MBS did remarkably well.  Even a decent amount of CDO's, which had the infamous Gaussian copula function and whose main weakness was a nation-wide correlation in defaults, were still hanging on without impairment.  After a nationwide collapse that was 10 times the magnitude of anything we had seen since the Great Depression?

The blame game here is absurd.  This Wikipedia article on the "Causes of the Great Recession" would be funny if it wasn't so sad.  Among dozens of causes discussed, monetary policy is barely mentioned, and then, only to blame it for being too accommodative (!) before the collapse. (Who knew that 6% NGDP growth was the path to hell?  Is the 20th century completely expunged from the economic history books?)    And one sentence about how increased interest rates might have contributed to lower home prices.  The 40% fall in home prices is simply treated as inevitable.  This is especially outrageous, considering that home prices are climbing back to the pre-recession levels, without any credit growth whatsoever.  People just refuse to believe that home prices could be efficient.  Our intuition regarding prices is useless, yet we insist on believing it.  Here we have a black swan the size of an elephant, and it's in the room.

Next is a graph from the Richmond Fed that breaks out delinquency rates for subprime and adjustable rate loans.  Even adjustable rate subprime loans didn't have a delinquency rate higher than delinquency rates in 2001 until 2Q 2007.  But, keep in mind, look at the home price graph again.  There was no housing slump in that cycle.  Delinquency rates booked at domestic banks, as shown in the graph above, topped out at 2.4% in that recession.  The Richmond Fed graph doesn't go back far enough to compare to the 1982 or 1992 housing dips, but clearly, none of these delinquency rates reached the levels we would see from those earlier corrections in home prices until the end of 2007.  The behavior of delinquencies doesn't appear to have been different than it had been in those cycles.



PS.  I was trying to find a paper I had seen that discussed the finding that FICO scores for mortgage borrowers didn't really fall below typical levels during the boom.  I didn't find the one I was looking for.  But, google has page after page of this kind of nonsense:
July 8, 2014—Mortgage bankers fear another real estate bubble, according to the latest quarterly survey of North American bank risk managers conducted for FICO (NYSE:FICO), a leading predictive analytics and decision management software company. In the survey, 56 percent of respondents directly involved in mortgage lending expressed concern that “an unsustainable real estate bubble is inflating.”
That is a survey of mortgage bankers, by FICO, in July 2014!  Here's a graph of the YOY growth rate of mortgages held by US households.  And bankers are reporting to FICO that we are in a mortgage lending bubble!  Can we, for the love of all that is holy, at this late date, please, please, I'm begging you, at least let something stop freaking declining before industry insiders start speculating about a bubble?

30 comments:

  1. This is really fascinating.

    I'm trying to figure out how to explain it to smart laypeople (or heck, even economists). The hard part for me is explaining the run-up in real house prices between 2002 and 2006. It just "looks" like excessive price growth followed by an inevitable collapse (a la dot-com boom). You argue that is efficient pricing. I'm trying to grasp that argument. Maybe it would help to convert your second graph to log scale and show a long-term trend line? Maybe it would help to show the return-on-equity of housing (imputed rent, with the tax deduction factored in etc) compared with other fixed-income investments over that time, and show that housing was well in line? Probably you've already done those things and I am growing too intellectually feeble to keep it in my head.

    The other thing that would be really interesting from my perspective is to show (nominal) home values, delinquencies, and NGDP on the same graph, and argue that had the Fed simply maintained the level-path of NGDP, home values would have avoided falling so far below efficient levels, and hence delinquencies would have been much lower.

    Regardless, thanks for the great blogging.

    -Ken

    ReplyDelete
    Replies
    1. I'll take a look at some of these ideas.

      Delete
  2. Let me start off by saying the word 'bubble' might be the most overused & misunderstood word in the popular financial press.

    I'm not sure if your point here is bigger than "it wasn't the subprime lending issues that created this problem." I think there was definitely malinvestment in real estate generally during this period. But if prices dropped and people defaulted, it could have been much more easily sorted out without government intervention.

    Jeff

    ReplyDelete
    Replies
    1. Well, there is a lot going on here. At the base of it, I think it's about causation and scale. There might have been an unusual amount of malinvestment in real estate (there is always some, in hindsight). But, I think part of it is that in 2007, everyone was thinking, "Man, delinquencies could get to 3% because of this subprime thing." And subprime was fingered as the culprit. Then, when delinquencies hit 10%+, we kept the culprit in our narrative, but didn't account for the fact that when it was the culprit, what we thought was bad was nothing like what was to come. And what came - a 40% drop in home prices - came after the money supply dried up but before delinquencies rose.
      Neither the timing nor the scale fit the story that fingers housing as the cause. As the Survey of Consumer Finance shows (in earlier posts in this series), neither debt payments / income nor housing consumption (measured by BEA imputed rent) increased from 1998 to 2007. And, when we look at income quintiles, it was the 60%-90% range that saw an increase in debt payments from 2004 to 2007.
      http://idiosyncraticwhisk.blogspot.com/2015/02/housing-tax-policy-series-part-11-low.html
      I think if someone, who didn't know the narrative already, looked at the SCF data, they would not guess that families who would have trouble making payments overbought real estate between 1998 and 2004.

      My version of events is being vindicated by the continued high housing prices in many developed countries and by the return of home prices in the US, with no credit growth behind it. But, that's just our own lying eyes. It will just be called a bubble. A bubble that everybody saw coming and that was funded by foreign buyers, all-cash buyers, and institutions. There is a distrust of markets and a wholesale denial of EMH. (Is there a financial paper somewhere that explains how forcing households to buy large assets in whole, without divisibility (unlike things like equities) could possibly lead to prices above the prices of other, more divisible assets? Everybody has agreed, implicitly, to accept things that simply turn all we knew about finance on its head.)

      Delete
    2. Right, I agree with the subprime lenders story, in that there's not really any "there" there. I don't think that generally lets housing off the hook.

      U.S. home prices, I would say, seem pretty reasonable still: http://research.stlouisfed.org/fred2/series/SPCS20RSA. As always, there are pockets of very high prices (Williston, ND).

      Have you taken a look at either "The House that Uncle Same Built" by Horwitz and Boettke: http://fee.org/files/doclib/HouseUncleSamBuiltBooklet.pdf or "Gambling with Other Peoples' Money" by Roberts: http://mercatus.org/publication/gambling-other-peoples-money? If not, I think you would like them. They came out early on to debunk the subprime myths being concocted at the time.

      Jeff

      Delete
    3. Thanks. I will read these. I think the SCF data also undermines the complaints about the CRA, etc., too, though. Both sides assume there was a problem, and are arguing about the cause, and I am looking at home prices in context and data on household finances, and I'm saying, "Hey, everybody. Back up. There's nothing to fight about. There isn't a problem."

      But, I'll read these and get back to you.

      Delete
    4. Yes, if subprime wasn't a problem, then CRA wasn't a problem.

      Delete
    5. Maybe save a couple of paragraphs, I think I would literally like "The House that Uncle Sam Built" better if you ran a macro to replace every word with its antonym. If they are going to write things like, "When one builds a 70-story skyscraper on a foundation made for a small cottage, the building should come down." they better be damn sure they are right about everything. And, I mean 100% right about everything. Because if they aren't, then they are basically pushing economic asceticism. I'm about 95% sure they are wrong.

      Delete
    6. I have a lot of respect for Russ Roberts, but I don't find his paper compelling either. It's not as bad as the first one, but it's basically question begging with a lot of very compelling just-so stories, in hindsight.
      A 40% drop in home prices was inevitable. We had it coming. How does he know? Because there are all these satisfying stories that were just sitting around waiting to be used to explain something. So they explain a 150% rise and subsequent 40% fall in home prices. They could also have explained a 300% rise, or a rise 5 years earlier or 5 years later, or, with a little shuffling, a boom and bust in wheat futures or oil. But, the housing boom and bust is what we got, so that's what they explain.
      They really are compelling stories. I find myself rooting for all of them. I agree with Russ in spirit. I'd love to believe them. It would support my worldview. I just think we need more than stories.
      In this episode we don't need more than stories, because there are enough compelling stories for every conceivable worldview - there are dozens that are mixed, matched, and shuffled, depending on the story teller. It's mad lib economic history.

      Delete
    7. Fair enough. I'll have to read them again and get back to you.

      Another side of the coin is the work by Gary Gorton. His thing is that there was a run on the repo market. This is a working paper version: http://www.nber.org/papers/w15223. You might find the his story more compelling.

      Delete
    8. I have a feeling I will like this one much better. I think repos are nuts. It's an accounting fiction just made for systemic risk - especially the overnight market. I can't believe long term creditors don't contractually disallow them. Repos are the shadow banking version of FDIC protection. They totally screw up the natural balance of liability costs. I'll read it.

      Delete
    9. I haven't really studied the repo market, other than to know (more or less) how it works.

      Delete
    10. Don't mistake my bluntness for expertise. :-)

      Delete
    11. That Gorton paper was interesting. Great information about the repo market and timing and influence of the subprime collapse. Thanks for the reference.

      One quibble I have with it is in the conclusion when he describes the Lehman Brothers failure as the trigger for the final part of the crisis. Here are a couple of posts I have done on that timeline:

      http://idiosyncraticwhisk.blogspot.com/2014/03/the-fed-in-2008.html

      http://idiosyncraticwhisk.blogspot.com/2014/05/interest-on-reserves-in-2008-and-now.html

      I have complaints about the Fed from about 2006 on, but the meeting in September 2008 the day after Lehman Brothers collapsed is astounding. If you haven't read the transcript, it is worth a read. Consider how long all those spreads had been elevated in the Gorton paper, and the day before their scheduled meeting Lehman failed. And, they spent the meeting congratulating themselves on managing the crisis well, trading notes about how qualified borrowers in their districts were unable to secure credit, and discussing whether they should be really hawkish about inflation or just kind of hawkish.

      Delete
    12. I know Gorton doesn't say why he thinks Lehman was the shoe to drop, but I remember that, at the time, a lot of us in the bullpen (I was a ph.d. student at the time) were arguing that the differential treatment of Bear and Lehman was the real problem, as it increased uncertainty regarding what the Fed and other agencies would do the next time a bank(-ish firm) was in trouble.

      I'll read those Fed minutes. Btw, have you seen the Lastrapes, Selgin, and White paper on the Fed, where they show that the Fed has been worse for stability than the national banking system that preceded it? If not, you might rather enjoy it.

      Delete
    13. I really like Selgin and White's stuff. I don't know if I've read the whole paper, but I've seen them talk about it.

      The Lehman/Bear Stearns thing seems to me to be another example of a just-so story. There are a million little things with our economy or our governance that I agree are problems. But, we are doing ok in spite of those things. One problem could be the treatment of Lehman/Bear Stearns. But, there needs to be a specific connection between that problem and a 50% collapse in the stock market. Has anybody even bothered to try to make that connection? I see a lot of people who just routinely refer to the Lehman failure as if it is the obvious reason. I don't think anyone would come close to saying, prospectively, that if Lehman went bankrupt it would cause the worst stock market crash in 40 years. In fact, I think the scale of the crisis is so large, that if Lehman was the cause, it would take a very strong and specific review of the matter to make that connection.

      Interestingly, in the transcript of the Fed meeting the next day, there isn't very much discussion of Lehman. Because the Lehman failure was a post-hoc reason for the fall.

      Now, at that meeting, after 2 years of disastrous monetary tightening (which is why Lehman failed), the Fed resolved to continue to starve the economy of liquidity, even though either the Lehman failure or the state of the economy that led to their failure, by themselves, should have led the Fed to err on the side of accommodation. In the face of both of those issues, they continued to refuse to simply perform basic OMO purchases. Then, over the next 2 months, they implemented interest on reserves - at a rate lower than the Fed Funds target, but higher than the effective Fed Funds rate, taking in half a trillion dollars in excess reserves, before any QE had happened.

      Now, THAT is something that we could intuit might cause a 50% crash in the stock market. And, in fact, the timing of the fall in equity prices matches those actions. There was some lag between Lehman and the eventual final decline.

      And, the lesson everyone seems to take from this, even while they complain that this whole episode was an example of the federal government bailing out powerful financial interests, is that the one thing the Fed did wrong was to let Lehman fail.

      Delete
    14. I simply don't buy that money was tight. I think the main culprit behind the recession is regime uncertainty. Regime uncertainty explains why it took so long for real non-residential investment to recover. It makes good sense from a real options view of investment.

      Delete
    15. Well, surely you think it was tight at some point. Isn't it a question of when as opposed to if?

      Delete
    16. Tight and loose are really relative terms, but I don't know what they are relative to. Under different rules for monetary policy, one will get different answers. The Bank of Canada or ECB is a little easier to evaluate, because they are mandated to do things that central banks, at least theoretically, can do. But the Fed's dual mandate is based on a false premise (the Phillips curve), so I really don't know how to evaluate it in terms of tight or loose.

      Delete
    17. Well, if that means you discount any analysis that says loose money caused a housing bubble, then I would see this as a great attitude to have. :-)

      On the other hand, if 10 months into a recession, with unemployment 1.5% above it's lowpoint, 5 year inflation expectations falling at 0.5% per month, extended unemployment insurance already in place, house prices persistently falling by more than 1.5% PER MONTH, the day after the latest in major investment bank failures, and with FOMC voting members saying things like:
      "I am hearing that credit is harder to come by for many borrowers who in the recent past would not have thought twice about their creditworthiness.....One of my directors, who heads a very large regional banking organization, reported at our board meeting last week that many banks are shedding assets and that in some cases they are walking away from longstanding customer relationships in order to do so."
      If the FOMC decides not to lower the FFR target and to note that the next move will a rate increase, after which inflation expectations fall sharply negative along with NGDP.
      If that is not tight monetary policy, then truly the position of monetary policy could never be established.

      Delete
    18. Well, I suppose I put a fair bit of weight on real factors, including uncertainty.

      Delete
  3. When I was playing poker for a living years ago there was a group of "investors" that posted on forums about the free money they got from credit card companies. It went something like this (from memory)- you get a credit card (with a small stash of cash already) and take out a cash advance, then pay back that advance and sacrifice a bit of your own cash for the interest. Your credit score goes up, and someone offers you a larger credit line (variations included getting a 0% rate for X months on transfers so you max out your first card then transfer to #2)- after several rounds of doing this you are getting bigger offers frequently enough that you can roll them over well before paying back all the money you have borrowed. Now these kids (mostly college aged) had access to a large (relative to their income) unsecured, 0% interest rate loan, that they (think) can be paid off by a larger loan (essentially a bridge loan) at any time. So they gambled heavily with the money (poker, the market, sports, whatever- some guys were even trading on margin if they could get it). Eventually they maxed out their total credit and if their gambling was down money (as it was for most) they would be staring at a hole that was 3-4X (at least) more than they had ever made in a year (or worse they had started this off with college loan money and were about to be kicked out of school). These guys would generally go for broke- put whatever they had left on something super risky and (obviously) almost always lose.

    If you looked at this circumstance from one perspective you would see a rise in credit, an (apparent) rise in credit worthiness, then a restriction of credit, followed by a wave of defaults and you might conclude that the problem was a restriction of credit. I think what you are doing at this point is similar.

    ReplyDelete
    Replies
    1. The SCF data I have been reviewing shows that 80% of mortgages before, during, and after the housing boom were held by households in the top 40% of incomes. Debt payments per household were higher in 1998, before home prices began to move up, were higher than they have been at any time since - during or after the boom.
      Your comment is exactly the kind of thing I'm pushing back against. There is no shortage of analogies, anecdotes, and plausible sounding stories. But, you're basically assuming that home prices were bid up, across the board, by 40% and that it was inevitable that they must fall. You live in a country full of people making those assumptions, so most places that are not this blog, nobody is going to notice the lack of actual data. I used to believe that story. I looked at the data.

      There is no shortage of plausible sounding stories. But, stories trick us into ignoring data. The housing equivalent of your poker story simply can't cause home prices to triple. There is no way you would prospectively tell that story. It only seems plausible because you're applying it to the known outcome.

      It's crazy to me how driven everyone is to find stories to defend economic dislocation.

      Delete
    2. Kevin,

      What do you mean by "economic dislocation?" Because I think there have been structural changes in the economy, if that's what you mean.

      On another note, he's an anecdote to support the SCF data. When I was at Syracuse University, I knew a former bank VP from a large southern city. She told me that banks were only approving 1 in 3 low-income mortgage applicants. The reason was because they lived in neighborhoods with essentially flat or dropping house prices.

      Delete
    3. By dislocation, I mean the nominal crash that leads to firms being overleveraged and to their being a surplus of labor (unemployment). I'm referring to cyclical problems. I would agree that there are structural problems, but I think they tend to be harder to pin down, in terms of the effect on living standards.

      Delete
  4. TravisV here.

    Kevin, I'm studying corporate profit margins (and whether they're unsustainable). Please look at the graph at the end of this post you wrote:

    http://idiosyncraticwhisk.blogspot.com/2014/10/returns-to-capital-arent-high.html

    In particular, please look at the green line. It really doesn't look like increasing leverage was driving down profit margins in the late 1960's, late 1980's and late 1990's. If it were, then the green line should have been flat or increasing. But no, during those periods, the green line was decreasing, parallel to the blue line.

    I'm curious what the primary factors were that drove profit margins lower in the 1960's, late 1980's and late 1990's. That graph leads me to conclude that it wasn't "increasing leverage."

    P.S.: I'm curious why Krugman didn't simply graph "Corporate Profits / GDI." Instead, he graphed "Corporate profits with inventory valuation and capital consumption adjustments, domestic industries / GDI."

    ReplyDelete
    Replies
    1. TravisV here.

      Another (potential) technical issue: not sure whether or not profits / GDI is a good proxy for profit margins (which is really corporate net income / corporate sales)......

      Delete
    2. I'll look into these things next week. I think one thing you have to be careful of is the global footprint of US corporations. That's why it was appropriate for Krugman to use the measure he did. The "D" in GDI is important. The capital consumption adjustments don't seem to change things that much. Once you've made that adjustment, domestic corporate sales and GDI tend to be decent proxies. And, I find that returns to capital tend to be fairly stable over time (philosophicaleconomics has been looking at EPS, and I think his work is basically saying something similar), so corporate sales, GDI, and the market value of corporate domestic assets tend to move together.
      One of the implications I think this has is that the corporations do have a growing base of foreign profits, and on an aggregate domestic level, these basically fund the trade deficit. There is no debt that needs to be repaid for that consumption.
      I will try to look at this with some more depth, though.

      Delete
  5. TravisV here.

    This Evan Soltas post is getting attention (Brad DeLong's blog, at least):

    http://esoltas.blogspot.in/2015/03/the-rent-hypothesis.html

    But I thought you did a better job than Soltas on the same topic here:

    http://idiosyncraticwhisk.blogspot.com/2014/10/returns-to-capital-arent-high.html

    ReplyDelete
    Replies
    1. Thanks Travis. I saw your comment at TMI. The ironic thing is that the income is actually going to rents - housing rents.

      Delete