Monday, January 25, 2016

Housing, A Series: Part 107 - A Brief Review of a Simple Point

There is one basic, introductory point that I think stands as a simple response to the volumes of pages that have been written about the housing boom of the 2000s - the mis-named subprime crisis, or housing bubble.  With regard to all of the descriptions of fraud, over-confidence, greed, predatory lending and irresponsible borrowing, there is this simple graph.
Source

Source: Financial Crisis Inquiry Commission
I have drawn a dividing line, roughly at the end of 2004.  This happens to be where homeownership rates had peaked.  Nationally, home prices had roughly doubled over the previous decade.  They were within 16% of their peak.  Six years later, home prices would be 14% below this level.  And today, they are back to 10% above this level.

By this time, subprime originations were near their peak, as a proportion of the total mortgage market.

Source
In other words, a large proportion of recent mortgages were subprime, and for any transaction that happened by then, almost all of the price increases were behind it, a nominal crash in home values unheard of in modern US history was in front of it, and by 2015, even after experiencing that unprecedented crash in market value, it had provided a moderate level of capital appreciation along with healthy and growing cash flows or imputed rental value.

And, yet, with all of that turmoil, the mortgages made up to that point had healthy, low default rates.  Here are graphs of Alt-A default rates and subprime default rates, by vintage.




The mortgages made before the end of 2004, as a group, performed very well, despite a lot of subsequent volatility that might have caused them to perform poorly.

Source

Even the 2005 vintage of mortgages performed fairly normally until late 2007, when home prices really began to drop.  So, even the year with the highest level of subprime originations and housing starts, with homes bought at the top price of the market, were preforming within the range of the previous five years.  Think about how much this contrasts with rhetoric about this period of time.  These were households who, as of 2008 or 2009, were experiencing the worst home price performance of the post-WW II era, by a wide margin.  They are the unluckiest set of homeowners in modern US experience, and they had used non-conventional financing at a scale far outside any previous ranges.  And loan performance was normal.  Even at the depth of the price collapse, the 2004 cohort was similar to earlier recession era cohorts that had experienced no price declines.

Clearly, the mortgages made before 2005 were not a part of the problem.  Maybe there was rampant fraud and misrepresentation by both lenders and borrowers.  But, practically speaking, just about everyone who signed on the dotted line was willing and able to make good on their financial promises.  We really can say this, also, about mortgages made in 2005.  But, let's say, for the sake of argument, that the eventual rise in defaults among 2005 borrowers was a time-bomb waiting to go off and that the eventual kink up in defaults on the 2005 vintage was inevitable.  Here is the only part of the housing market graph above that could possibly contain any systemically important excesses.  Any levels of home ownership or home values above a sustainable level have to have happened somewhere here.

Oh, and by the way, the Fed Funds Rate was above 4% by the end of 2005 and above 5% by the summer of 2006, with an inverted yield curve.  None of the mortgage cohorts with high default rates were made when the Fed Funds Rate was low, and none of the mortgage cohorts originated when the Fed Funds Rate was below 2% had high default rates.  So, if there was some sort of Austrian Business Cycle capital misallocation going on in 2003 and 2004, someone needs to tell the borrowers and lenders who purchased homes at the time.  They may not have noticed.

If your response to this is that I am na├»ve - that you knew an unemployed guy who was flipping houses in 2003, that you knew a guy who worked for a predatory subprime lender in 2001 or 1996, and that they were doing terrible, irresponsible things - then you need to think about evidence, what it is, what makes it relevant, and what could possibly falsify your explanation of events at the macro level.

Here is a copy of the S&P downgrade announcement of Residential Mortgage Backed Securities on July 11, 2007.  From the report:
Although property values have decreased slightly, additional declines are expected. David Wyss, Standard & Poor's chief economist, projects that property values will decline 8% on average between 2006 and 2008, and will bottom out in the first quarter of 2008.
While our LEVELS model assumes property value declines of 22% for the 'BBB' and lower rating category stress environments (with higher property value declines for higher rating category stress environments), the continued decline in prices will apply additional stress to these transactions by increasing losses on the sale of foreclosed properties, as well as removing or reducing the borrowers' ability to refinance or sell their homes to meet debt obligations.
As lenders have tightened underwriting guidelines, fewer refinance options may be available to these borrowers, especially if their loan-to-value (LTV) and combined LTV (CLTV) ratios have risen in the wake of declining home prices.
And:
The loan performance associated with the data to date has been anomalous in a way that calls into question the accuracy of some of the initial data provided to us regarding the loan and borrower characteristics. A discriminate analysis was performed to identify the characteristics associated with the group of transactions performing within initial expectations and those performing below initial expectations. The following characteristics associated with each group were analyzed: LTV, CLTV, FICO, debt-to-income (DTI), weighted-average coupon (WAC), margin, payment cap, rate adjustment frequency, periodic rate cap on first adjustment, periodic rate cap subsequent to first adjustment, lifetime max rate, term, and issuer. Our results show no statistically significant differentiation between the two groups of transactions on any of the above characteristics. Reports of alleged underwriting fraud tend to grow over time, as suspected fraud incidents are detected upon investigation following a loan default.
They seem to take the lack of explanatory power from all the typical sources of default as a sign of fraud.  But, wouldn't this also be a sign that the source of the defaults is not buyer quality?  If the source of stress in these mortgage pools was from some outside influence, wouldn't we still expect fraud to increase as the market distress increased and wouldn't we expect those instances where fraud was involved to be more frequently noticed (much like with S&Ls in the 1980s)?  "Our results show no statistically significant differentiation between the two groups of transactions on any of the above characteristics." That is a very strong effect from underwriting fraud from a market that just 2 years prior appears to have created a cohort of mortgages that performed very well.  In two years, with similar rates of originations, subprime loans went from being benign in the face of extreme volatility to being so devoid of honest underwriting that FICO, LTV, DTI, and many other reported variables had no explanatory value at all?  And the fraud was so universal that even the issuer wasn't explanatory?  Fraud somehow meant even the terms of the mortgages weren't explanatory?

Also, note that at the time this report was published, home prices had basically been flat for about 18 months.  There were some local markets that were dropping by then, but nationally, prices were still at the plateau they had been on since early 2006.  But, S&P projects a decline in home prices of 8% nationally, and up to 22% in some areas.  At a point where home price trends are still stable, they predict future price trends far outside any previous experience.  This must have had a large effect on the implied value of the securities in question.

And, what did the Federal Reserve have to say about the effect of these extreme expectations of nominal collapse, when they met just a couple of weeks later on August 7?:
The Federal Open Market Committee decided today to keep its target for the federal funds rate at 5-1/4 percent.
Economic growth was moderate during the first half of the year. Financial markets have been volatile in recent weeks, credit conditions have become tighter for some households and businesses, and the housing correction is ongoing. Nevertheless, the economy seems likely to continue to expand at a moderate pace over coming quarters, supported by solid growth in employment and incomes and a robust global economy.
Readings on core inflation have improved modestly in recent months. However, a sustained moderation in inflation pressures has yet to be convincingly demonstrated. Moreover, the high level of resource utilization has the potential to sustain those pressures.
Although the downside risks to growth have increased somewhat, the Committee's predominant policy concern remains the risk that inflation will fail to moderate as expected. Future policy adjustments will depend on the outlook for both inflation and economic growth, as implied by incoming information.
In their defense, although the minutes of the meeting mention a softening even in the jumbo loan market, at that time, mortgage levels outside the subprime market were still growing modestly.  But, I wonder how much of a difference it would have made if the Fed had simply made a rhetorical statement that they didn't expect home prices to fall, or that they would expect to add liquidity if home prices appeared to begin to slide.  They didn't make that statement because they had no intention of supporting nominal stability as home prices fell by nearly 1% per month for the next year.  This nominal collapse, after all, was not a problem, it was the "correction" of a problem.

8 comments:

  1. Kevin, your story about the 2003-2004 cohorts makes sense and as you say undercuts the traditional narrative. Clearly, the rise in home ownership in 2003 and 2004 was based on qualified buyers at affordable prices.

    But it begs the question, why did the 2006 and 2007 cohorts perform so poorly? I get that the horrible macro conditions of 2008-2009 would cause a lot of defaults. But I don't see why the impact should be concentrated in the 2006-2007 cohorts. Why did the 2003 cohort see a tiny rise in defaults in 2008 (60 to 72 months out) while the 2006 and 2007 cohorts saw much larger rises in the 24-to-36 month and 12-to-24 month windows respectively? And these bad mortgages were being originated in a climate where home ownership rates were actually falling, meaning the buyers were mostly real estate investors? Color me confused.

    -KEn

    ReplyDelete
    Replies
    1. It is confusing. I think I touched on it a little in part 104, and I hope to have another post up soon, although it only gets more complicated. I think the short version is that we have been saying that the Fed was too tight in late 2007 and 2008, but that, really, nominal contraction began in early 2006. NGDP growth began falling, housing starts collapsed, etc. Home prices, relative to incomes and affordability started falling in 2006-2007, even in Open Access cities where prices weren't high. These should be clear signs of a nominal and credit shock, but we were too fixated on the demand-side explanations to notice. All of these cohorts saw defaults rise at the same time - mid-to-late 2007, because that is when everyone concluded that prices were going to collapse, and the Fed said, "Yep. You should probably figure on that." They were already constricting currency growth, and now credit growth stopped too.
      Isn't it funny how so many Austrians claim that we should account for home prices in inflation measures in 2003-2005, but yet I don't think I've ever heard anyone use that logic to conclude that the last half of 2007 was highly deflationary.

      The first wave of defaults were a product of the price collapse, not the cause of them.

      I haven't been able to completely work out the details in the numbers, but non-shelter core inflation dropped to near 1% in the last half of 2007, and NGDP and RGDP had declined. I think, to the extent that NGDP and RGDP seemed as healthy as they were, it was because we were already seeing the national supply problem due to the credit collapse, so rent inflation and rental income were high. We treat this as inflationary, when really it is a transfer of rents from wage-earners to real estate owners (frequently the same people, of course). I don't think I have gone into it on the blog, but I wonder if this delayed the employment response, which is why the recession happened so much later than the yield curve inversion. Normally, wage stickiness would lead to unemployment. But, here, we still had 2% wage growth, but it was just being sent to the landlord. In a way, the housing supply problem reduced the sticky wage problem, because it moved the recessionary level of nominal wage growth higher, above the problematic zero level.
      My dilemma here is that these seem like important insights to consider, but I am afraid that this just keeps getting more complicated.

      Delete
  2. "The first wave of defaults were a product of the price collapse, not the cause of them."

    -Pretty much. But why didn't something similar happen during the 2001-3 downturn?

    And would real home prices have fallen after 2006 had the Fed committed to an NGDP growth rate of 5% when Bernanke took office? Why didn't Canadian housing prices fall during its NGDP shock? These are all very interesting questions to ponder.

    I also think you should focus on the international phenomenon of housing price increases between 2001 and 2006-8. Canada, Australia, UAE, Spain, Ireland, etc. Something which would make a naive observer turn to the Chinese Savings Glut theory.

    And while the homeownership rate may have fallen after 2004, a lot of the home price appreciation after that was driven by investors purchasing multiple homes.

    ReplyDelete
    Replies
    1. I tend to look at it as some range of reasonable stability rather than a strict target. Any number of targets could have worked. The problem is that we had an explicit policy of demanding instability, and to this day even in hindsight that is the only policy the public would have stood for. Any country that avoided that low bar - don't target instability - avoided some of the worst aspects of the downturn.

      I would say that the savings glut - whether Chinese or otherwise - explains about 1/3 of the nominal price rise from 1995 to 2005.

      That's the great thing about the bubble explanation - everything is evidence for it. Subprime mortgages to new homeowners with low rates? Too much money! Investors and wealthy non-primary residential buyers with rising rates and declining mortgage growth? Too much money! Rising rents with a decade of negligible housing starts? Too much money!

      Delete
  3. Very easy explanation to the top FRED graph. Multiple house owners appeared. They used to go around in buses buying up everything they could see in Nevada. I remember them. You may want to check on that Kevin. There was a bubble. But then, in mid 2005 there was going on a huge increase in inventory. Now this is for the bubble areas. Once these folks who invested in the bubble areas, (they came from SoCal to Nevada), and they realized they were stuck, they may have been forced to sell their houses in SoCal. That was, after all, a margin call, Kevin.

    My daughter and son in law were offered an 800,000 dollar house with an income under thirty dollars an hour in Las Vegas and they would have qualified. If you had a pulse you could buy anything. It was a bubble, clearly.

    Even charts and you see NGDP falling, in 2005, as these loans were increasing. http://www.talkmarkets.com/content/us-markets/nominal-gdp-targeting-and-market-monetarism-made-easy-with-graphs?post=82536

    ReplyDelete
  4. And look how many were from California buying up investment houses in Las Vegas in 2005: http://www.doctorhousingbubble.com/las-vegas-looks-to-california-for-home-sales-demand-1-out-of-3-vegas-home-buys-from-california/ "Las Vegas was unique in the sense that run-off money from the California bubble entered into their market in large numbers. At one point roughly 40 percent of investment homes were being purchased by those with California ties."

    ReplyDelete
    Replies
    1. So, Kevin this is the chart of price action from that article, for Las Vegas. The bubble was rapid and it was popped rapidly as well: http://www.doctorhousingbubble.com/wp-content/uploads/2011/12/house-prices-for-las-vegas-united-states1.png

      Delete
    2. I want to clarify something. I have come to agree with Kevin's point of view. But from a different angle. I believe that there was a crisis, that models didn't work, that risk was mispriced. So, I believe that the S&P actually downgraded before things got out of hand, due to their realization that their risk models would not hold up. Securitization was going to end and commercial paper would start to crash but only after those downgrades. Add to that the government had a bill introduced into the House, bill HR 1424, that would give the Fed power to buy bad paper. Well, there was no or little bad paper when the bill was introduced. It is like they planned on the crash and waited and did nothing so it would be even bigger. So, bubbles were serious, but not everywhere. Some markets could have been saved. But the Fed and S&P chose not to save them. They could have warned congress they needed the power to buy that soon to be failing bad paper but they did not warn congress. The bill sat for a long time after the summer of 2007 initial subprime crash.

      Delete