Financial stocks crashed in the first half of July and then rallied in the second half. Shorting these stocks had become widespread (often paired with going long oil and commodities), resulting in the shorts became very crowded. Thus, in mid-July when the government stepped in to eliminate naked shorting, prevent Fannie Mae and Freddie Mac from failing (at least temporarily; we are short both) and pass (yet another) housing bailout bill, financial stocks experienced a quick, sharp reversal that fed into itself as momentum-driven short sellers rushed for the exits and a clumsily unwound their trades.
We believe this is yet another bear market rally in this sector, of which there have been many over the past year, as this chart of the Financial Select Sector SPDR Fund makes clear - click to enlarge:
We remain short many financial stocks, primarily the bond and mortgage insurers, the GSEs and a few banks, because we believe that, in general, their fundamentals remain awful and they have not yet fully fessed up to big losses, which will require them to raise substantial additional capital on highly dilutive terms (if they’re lucky). At their lows a month ago, perhaps this scenario was built into their stock prices, but at much higher levels today we believe there’s limited upside and substantial downside.
The State of the Housing Market
In general, we are bottoms up stock pickers, but on rare occasions we develop a high degree of conviction on a particular macro view and apply this to our stock picking, both long and short. One such view is that a worldwide debt bubble of unprecedented proportions – the largest component of which was the U.S. mortgage market and its various derivatives (RMBSs, CDOs, etc.) – built up over a number of years, peaking in mid-2007, that is now in the process of unwinding. We believe that the unwinding of this bubble – and the resulting credit crunch – will last many years, the dollar amounts involved will eventually be measured in the trillions and the total losses will likely exceed $1 trillion, of which only about $400 billion has been recognized to date by the world’s financial institutions.
There are many factors impacting this ghastly mess, but surely one of the biggest is the unprecedented decline in U.S. home prices. Simply put, we do not believe that the financial sector will stabilize until U.S. home prices do. So, the key questions are: how much further will home prices fall and when will they reach a bottom? Our best guess is that home prices, which through May were down 18.4% from their peak in July 2006 (based on S&P Case Shiller data), are only about halfway finished declining – meaning that home prices will eventually fall 30-40% from their peak – and that a bottom will not be reached until 2010 at the earliest. Needless to say, our view is significantly more pessimistic (we prefer to think realistic) than that of most analysts, CEOs of financial companies and other “experts”.
Presentation on the Housing Market
We have put together an 18-page slide presentation with our latest analysis of the U.S. housing market, which is posted at www.valueinvestingcongress.com (click on the link on the right side of the page that reads: “READ the presentation or VIEW THE VIDEO from T2 Partners on the Bursting of the Housing and Credit Bubbles.”).
Here are our comments on each of the 18 slides:
Slide 1) This chart shows that prior to this decade, the average American household was able to borrow approximately 3x its pre-tax income to buy a house. As the decade progressed, lenders became increasingly irresponsible and, at the peak from early 2006 to early 2007, they were willing to lend 9x a household’s income! This madness was exacerbated when lenders frequently didn’t bother to verify a borrower’s income or assets – these low/no-doc loans are only somewhat jokingly referred to as NINJA loans: no income, no job, no assets.
The far right side of the graph shows that, as of the beginning of this year, due mainly to lenders tightening their debt-to-income ratio limit back to historical levels around 35%, the amount that can be borrowed to buy a home had fallen by 39.4%, to 5.2x income (as of today, the decline has been even greater, as Amherst Securities estimates that the leverage has fallen to 4.6x). Keep in mind that these figures are optimistic, as they assume an interest-only loan, which few lenders are willing to make anymore and, in addition, lenders are once again requiring meaningful down payments.
The massive decline in leverage available to home buyers will likely be permanent and put home prices under pressure for many years to come.
2) As the bubble has burst, existing home sales have tumbled and the inventory of homes to be sold has reached an all-time high of nearly one year. Econ 101 will tell you that prices cannot stabilize until excess inventory is sold off. It took three years for the excess inventory to accumulate and we wouldn’t be surprised if it took three years to work it off.
3) Not only are many homes and condos for sale, but even worse is the fact that many are vacant, which can result in distressed sale prices, homes falling into disrepair and depressing the prices of nearby homes, etc. This chart shows that 2.9% of all American homes and condos were vacant as of the end of Q1, more than 50% above historical levels (and the number is surely higher today). The most shocking statistic on the page is the fact that of all homes and condos built this decade, 10.2% are vacant.
4) Foreclosures are major drivers of home price declines because they typically result in an auction of the home, often resulting in a highly distressed sale price, whereas someone selling their own home typically won’t accept such a price. This chart shows that foreclosure filings have skyrocketed over the past two years and there’s no sign of a let-up. Last month, foreclosure-related sales accounted for 42% of home sold in California and here’s what an article in the New York Times article said today about what the nation’s two largest home mortgage players are doing:
Daniel H. Mudd, Fannie’s chief executive, told investors that Fannie was “using a lot of innovative ways to get the property out the door.” This is not, he said, a good time to be holding on to foreclosed properties “and hoping for a better day.”…
Foreclosures are soaring. In 2006, Fannie and Freddie between them acquired 52,967 properties, a figure that leaped 36 percent, to 71,961, in 2007. During the first half of this year, there were 66,420 foreclosures by Fannie and Freddie.
Not only are there more foreclosures, the losses are much larger. In 2005, Fannie lost an average of 7 percent of the unpaid principal when it sold a foreclosed property. So far this year, the figure is 26 percent, and in California it is up to 40 percent. In all of 2006, Fannie foreclosed on 93 California homes. The current rate is almost 1,000 a month.
Here are some additional statistics:
• At the end of Q1, 2.47% of mortgage loans were in foreclosure, equal to 1.3 million homes or 2% of U.S. households
• 1/3 are in California and Florida
• Nearly three million homeowners were behind on their mortgages at the end of 2007 and 1-2 million are at risk of foreclosure in 2008
• 8.8 million homeowners were underwater on their mortgages (balances equal to or greater than the value of their homes) as of the end of March according to Moody’s Economy.com
• 30% of subprime loans written in 2005 and 2006 are already underwater
5) Fueled by loose lending standards and a bubble mentality by all market participants, home prices nationwide from 2000-2006 rose more than 50% – and now need to fall approximately 34% from their peak to return to trend line.
6) In the top 20 metropolitan areas tracked by the S&P Case Shiller Index, home prices more than doubled – and, through May, had only fallen 18.4% from their peak.
7) This chart that appeared in the Wall Street Journal last month shows that home prices in the 20 metropolitan areas tracked by the S&P Case Shiller Index, having declined 17.8% from their peak through April, need to fall an additional 29.0% (41.6% total) to reach the level they were at in January 2002.
8) These two charts measure home prices relative to rent and income levels. The former shows that, assuming steady rent levels, home prices need to fall about twice what they have fallen so far to reach the historical average. The latter shows that home prices have already fallen back to the long-term price-to-income ratio. This is the only data point we’ve been able to find that shows home prices have already fallen back to normal levels, and it doesn’t make any sense to us, as income levels for average Americans have been stagnant.
9) This page shows the month-to-month (sequential) changes in U.S. home prices from March 2005 to May of this year. In the first 17 months of this chart, home prices continued to rise, albeit at a slower and slower rate, until July 2006, when home prices peaked. At that point, the line on the chart goes negative, where it has remained, meaning the home prices have fallen in every month since their peak.
The most interesting and relevant part of the chart is the line in the last three months, in which the rate of decline slows sharply: from January to February 2008, home prices fell an unprecedented 2.6%, but then "only" fell 2.2% in March, 1.3% in April and a mere 0.8% in May. Looking at this data, one might be tempted to extrapolate the recent trend and conclude that home prices might stabilize in the new few months – and if that's true, then many financial stocks will likely be great investments from here. We don’t think this trend will continue, however, as the next two slides make clear.
10) This graph shows month-to-month home price data since the inception of the S&P Case Shiller 20-city index at the beginning of 2000. It makes clear the enormously seasonality in the month-to-month numbers – and that in the eight previous years going back to 2000, home price changes in April, May and June of each year (highlighted by the red circles) are much higher than the rest of the year.
11) This is the same graph, using the S&P Case Shiller 10-city index, which goes back to January 1987. Again, in virtually every year, one can see the better performance in April-June. Mathematically, the simple average monthly increase in April-June of each year since 1987 is 0.76% (9.6% annualized) vs. only 0.29% in the other 9 months (3.5% annualized).
So does this mean that home prices are likely to return back to falling 2%+ per month later this year once this seasonal effect wears off? Probably not – 2% per month is a very high number. But we'd guess – and it’s only a guess – that the average monthly decline will average at least 1% for at least another year and then another year in the zero to -1% range on average for before home prices finally bottom in mid-2010.
12) Many mortgages written at the peak of the bubble (from early 2005 through the middle of 2007) had low “teaser” interest rates that reset after two years, which triggers payment shock and, in many cases, default. It’s not a coincidence that the mortgage crisis began in early 2007 which, as you can see in the chart, is when the first wave of resets hit, from (mostly subprime) mortgages written in early 2005.
This chart shows that the resets taper off by the end of the first quarter of 2009, which is leading many to predict that home prices will bottom shortly thereafter. There are two reasons why we believe this prediction is incorrect. First, it takes an average of 15 months from the date of the first missed payment for a home to go through the process of going into default, being foreclosed upon, the lender taking possession and, finally, liquidating the asset. It is primarily in the last phase, typically when the home is auctioned off, that home prices are pressured. So, to determine the timing of when home prices might bottom, one shouldn’t look at when the resets taper off, but rather when the wave of foreclosures and auctions do. Thus, add 15 months to the end of Q1 2009 and that’s mid-2010. But the story is even more grim…
13) This chart picks up what the other chart misses: that a huge wave of Option ARMs and, to a lesser extent, Alt-A mortgages will reset (or “recast”) in 2010 and 2011 – a wave as big as the one in 2007 and 2008 that has caused so much pain to date.
14) This slides describes what an Option ARM is. In short, an Option ARM is an adjustable-rate mortgage that was given to a prime borrower, usually based on the borrower’s FICO score and appraised value of the home, with little concern for the borrower’s income or assets – hence, more than 70% of Option ARMs were low/no-doc. The key benefit to the borrower was that, until the loan resets (the technical terms is “recasts”), the borrower only had to pay a small fraction of the interest owned on the loan, with the difference being added to the principal balance of the loan.
So, for example, the Option ARM might have an interest rate of 7%, fixed at this level for the first five years, but the borrower would only have to pay interest only at, say, a 2% rate (with very small increases each year). Each month that the borrower chooses to pay the lower rate, the unpaid interest is added to the balance of the loan – this is called negative amortization, and approximately 80% of all Option ARMs have incurred some degree of negative amortization since inception.
A typical Option ARM resets either after five years elapses or – this is key – the loan, due to negative amortization, hits a trigger point, generally at 110%-125% of the original loan balance. This means that a borrower paying 2% rather than 7% for three years will see the loan balance increase by roughly 15%, which often triggers a reset.
Slide 13 shows the scheduled five-year Option ARM resets, but because nearly all Option ARMs are negatively amortizing, many are resetting right now, which is triggering a sharp rise in defaults.
15) This slide has quotes from a Wall Street Journal article about rising Option ARM defaults. Here is the key quote:
’My sense is that many option ARM borrowers are in a worse position than subprime borrowers,’ says Kevin Stein, associate director of the California Reinvestment Coaliton, which combats predatory lending. ‘They wind up owing more and the resets are more significant.’
16) This slide shows that until March 2005, the interest rate on an Option ARM was lower than that of a fixed-rate 30-year mortgage, which attracted rational borrowers to Option ARMs and deterred those already holding Option ARMs from refinancing.
But after March 2005, rising short-term interest rates pushed the Option ARM interest rate above the rate of a 30-year fixed-rate mortgage – at one point, nearly 200 basis points higher. This should have dampened demand for Option ARMs and led to a surge of refinancings into the fixed-rate mortgages, but this didn’t happen – in fact, demand for Option ARMs surged.
How could this be? What could account for such unprecedented, irrational behavior? There are two likely explanations:
A) The borrowers had high FICO scores, but weren’t really prime borrowers (one of the major lessons from the mortgage bubble is that FICO scores have proven to be very poor predictors of whether someone would default on a loan). Recall that most Option ARMs didn’t require any documentation of income or assets (unlike most traditional fixed-rate loans), so if someone had, for example, recently lost his job, an Option ARM would be the product of choice. Of course, such borrowers are likely to have subprime-level default rates…
B) The borrowers were prime, but were buying their dream home at a peak-of-the-bubble price, were stretched to the limit, and simply couldn’t afford the full-interest-plus-amortization payments that traditional fixed-rate loans require. What do you think will happen to these borrowers when the Option ARM resets, the interest rate jumps (to a level far higher than what it would have been under a fixed-rate loan), and not only must all of the interest be paid each month, but the loan also becomes fully amortizing?
Even if the homeowners can afford the new, dramatically higher payments, will they be willing to make them? As the next slide shows, the majority Option ARMs were written in bubble states in which home prices have already fallen by 25% or more, meaning a huge number of Option ARM loans are far, far underwater.
Consider this example: Imagine a $600,000 mortgage that negatively amortizes to $690,000, which triggers a reset (assuming a 15% trigger). At the same time, the value of the home has fallen 25% to $450,000. How many homeowners will simply walk away from their homes (or demand that the lender write down the loan to market value) rather than make full interest and amortization payments on a mortgage that’s 53% above the value of the home?
17) This slide shows that Option ARMs were concentrated in the states that had the biggest housing bubbles – and are suffering the worst consequences.
18) This is an email from a Federal Senior Bank Examiner about Option ARMs, which he predicts will be “the next tsunami to hit the housing market”.
Disclosure: The author manages funds that are short Fannie Mae, Freddie Mac, Ambac, MBIA, General Growth Properties, PMI Group, and Radian.