As Countrywide Financial Goes, So Goes the Mortgage Sector

Bear Stearns Cos., already forced to shut two hedge funds that bet heavily on the risky subprime-mortgage market, is now facing big losses in a third fund that has roughly $900 million in mortgage investments, according to people familiar with the matter.

Kate Kelly, The Wall Street Journal

In Los Angeles where IRA's HQ is located, we hear increasing anxiety from locals concerned about the real estate market. Residents of CA have seen this movie before, you understand, and their fears are not assuaged by the smiling countenance and confident demeanor of real estate industry deities like Countrywide Financial (NYSE:CFC) CEO Angelo Mozillo. But CFC's performance in future quarters will provide an accurate barometer of industry trends.

When CFC announced Q2 numbers and revealed that prime home equity collateral was deteriorating along with the subprime production, we heard a number of people use the term "surprise" to describe the event. But had you been watching the default experience of the lead unit of CFC, Countrywide Bank FSB, a $100 billion asset mortgage originator which had just $1 billion in assets at the start if 2002, you'd be not surprised at all. In fact, you might know a couple of things most people don't:

First, you might know that CFC's gross loan defaults actually troughed in 2006 at a five-year low of just 4.7bp vs 16bp for the peer group. Even with the Q2 bad news, CFC's defaults are still very low and, along with the rest of the mortgage specialization peer group, have a long, long way to go before we hit 20-year averages much less declare a significant "peak" in terms of loan defaults.

In December 2001, for example CFC's far smaller lead unit, which at the time had a national bank charter instead of the current thrift vehicle, reported 91bp of loan defaults to the FDIC, 3x the 30bp for the mortgage specialization peer group. [Click here to see a chart from The IRA Bank Monitor showing the default experience of Countrywide Bank FSB since 2001.]

Notice how low the default experience of CFC's lead unit fell over the last five years, reporting zero defaults in some quarters, and this during the high tide of the US mortgage credit boom. As the chart suggests, in the past five years CFC generally has outperformed the peer group in terms of default experience even as it has grown to be the leading originator of US mortgage loans.

Second, you would know (or at least suspect) that loan defaults were likely to bounce sharply in 2007-2009 as CFC and the rest of the mortgage peer group revert to the LT mean -- figure at least two standard deviations from year-end 2006 levels.

Loan defaults for the larger units of the mortgage specialization peer group peaked at 26bp in 2003 following the 1999-2001 US economic hiccup, an illustration of how long it takes micro economic events like retail loan defaults to work their way through the banking system. On that basis, in 2008 working in the recovery function of most banks looks to be no fun at all.

Of note, the Loss Given Default of CFC's lead unit at the end of Q1 2007 (latest data available) was 92%. The LGD has been well above the peer average of around 75% for several years, a troubling metric for a supposedly secure mortgage portfolio. By comparison, look at a giant like Wachovia Bank (NYSE:WB), which consistently turns in an LGD around 50-60%.

Now given the low absolute default rate, the high LGD is unremarkable; you'd expect the few defaults seen in good times (or, better, completely irrational boom times) to be particularly nasty events. But CFC's management team better hope that LGD falls as the bank's default rate climbs in coming months. An above 90% LGD is 50% above the peer group and says to us, at least, sloppy.

Bottom line: the unwinding process in the US real estate market has quite a long way to go, years in fact. Given how far below historical averages CFC and its peers remain from peak or even average loan default levels, this roller coaster's got a ways to go yet. Keep your eye on the LGD's for all the mortgage specialization peers in coming quarters for an indication of where aggregate US credit quality is really heading -- and how fast.