Monday, June 27, 2016

Housing: Part 163 - Some notes on Fannie Mae

I have been digging into Fannie & Freddie's SEC filings recently.

Remember, private securitizations (subprime, Alt-A, etc.) really only were dominant from 2004 to early 2007, and that market had died by late 2007.  Homeownership peaked in early 2004.  The market outside the GSEs is surprisingly unrelated to growing homeownership rates.  I have come to see the rise in private securitizations as a response to the decline in the GSE's that came after the Bush administration pressured the GSE's with regulatory scrutiny, leading to accounting scandals, impositions of new capital controls, etc.  So, it is likely that without the rise in private mortgages, defaults would have risen in 2004 and homeownership would have fallen more sharply then.  The lack of correlation doesn't mean that those markets didn't provide some credit support for homeowners, but it does seem like the GSE's are the primary conduit, on net, for sustained ownership via the entry of first time homebuyers.

While private securitizations had provided support for mortgage markets when the GSE's pulled back in 2003 and 2004, the GSE's didn't provide support for mortgage markets when private pools collapsed in 2007.  In fact, they joined in the collapse.

I have shown in other posts that there was no change in the overall credit-worthiness of homebuyers in the 2000s.  Homeowner incomes were flat or rising until the collapse, then rose after the collapse because the mortgage market for low income buyers has dried up.  We see a similar pattern at Fannie with FICO scores.  FICO scores were very stable until 2008.  Since then, Fannie has basically closed down originations for the bottom third of what was previously a longstanding portion of the origination market.

In the second graph, here, we can see that all of the growth in mortgages outstanding at Fannie Mae have been at FICO scores of 740 or more.

This highlights something that I think may be underappreciated.  I have pointed out many times that the mortgage market has been closed to the bottom half of the population, by income or credit-worthiness.  But, notice here that all FICO categories under 740 had been declining or treading water since 2007.  Originations have been negligible for FICO scores under 660 and low for scores under 700.

Many of those households are basically grandfathered in.  Many probably cannot qualify to refinance existing loans.  So, even existing owners at the low end of the credit spectrum have been disadvantaged by the closed off mortgage market, and find it harder to refinance at lower rates compared to households at the higher end of the credit spectrum.

In the next graph, we can see the sharp drop in mortgages for purchase that began in 2006  2008. (edit: the original x-axis was off on this graph.)  I will probably follow up on this in a future post, but one interesting thing going on here is that first-time buyers at the GSE were strong in 2006 and 2007.  The pain during that period was felt by existing owners who were being squeezed by falling home prices in some cities.  The new buyers in 2006 and 2007 would eventually be the most vulnerable to that problem when prices really collapsed.

In the next graph, we can see that LTV's were relatively normal throughout the period and that when home prices started to collapse, LTV's first ticked up slightly, but by 2008 and 2009, average LTVs (loan-to-value) were falling.  By 2008 the main problem was that market LTVs had risen above the original LTVs for Fannie loans.  In the midst of that problem, Fannie was decreasing the average LTV.  This makes sense from a lender point of view.  Part of what was happening was that there was broad expectation of continuing price collapses.  That expectation was being encouraged by the Fed.  Lenders don't lend on collateral they expect to lose value.  This becomes a self-fulfilling prophecy.

Look at what we see in the last graph.  Even in 2007, Fannie was pulling back on the low end of the market.  Market prices of the homes in their portfolio were falling, and as that happened, they were cutting off the bottom end of the market.

The bottom of the market was collapsing after late 2007 because we stopped providing credit to it.  Low priced neighborhoods didn't lose value because of defaults on risky loans to households that couldn't afford them.  They lost value because from 2006 into late 2007, the market for mortgages to those neighborhoods, whether from private markets or the GSEs, disappeared.  The market disappeared before the defaults piled up.

Since that time, we can see the sharp divergence between the existing pool of mortgages and the marginal new pool of mortgages here.  The typical down payment, which had been steady or rising throughout the boom, kicked up in 2008 and 2009.  The average loan size for the small number of new buyers kept marching up even as the typical home value in the portfolio was collapsing.  Larger loans with larger down payments means that in 2008 and 2009, Fannie was only supporting the very high end of the market.

Before that, the average home value of the existing portfolio and new originations followed along fairly closely with one another.  The moral error of public perception and policy is stark.  The CEO's of Fannie and Freddie at the time are currently being sued by the SEC for having put too much money into the low end of the mortgage market.

I find that the inability of the GSE's to provide support for the mortgage market when it was needed largely comes down to expectations.  We really did will this collapse into existence.  I'll probably have more on that later.


  1. This is one of those strange circumstances in which the conspiracy-monetary nuts have a leg to stand on. Who gets credit---that is, the new money?
    We see new money creation extended to the well-connected, or the already affluent.

    No it is not as simple as that. After all some people have low FICO scores deservedly.

    But to a guy who can't get a loan....

    1. Careful, though. There is a difference between a loan and a transfer. But, yeah, you are right. Between credit access and the related drop in property values, the Federal government has imposed a huge cost on middle class and lower middle class neighborhoods. The irony is that the residents in those neighborhoods seem to have been convinced that the banks were at fault for loaning them money when the shock clearly came from the various developments that removed that credit access.

  2. Kevin---
    Access to credit (new money in the US system) is highly lucrative in times of rising asset values. Leverage up and huge returns on equity are possible.

    I am not a Jekyll Island conspiracy nut.

    On the other hand, federal agencies are often captured by interest groups and class perspective frames the argument.

    Sheesh the "libertarian" blogs rant about the minimum wage on a daily basis...but go mute on property zoning. Not a topic.

    The story line about predatory lending is false. The real story, that you are telling, about federal policies that direct new money away from lower income homebuyers, is not.

    In fact, if the Fed would do its job, lending on property, including low cost housing, would be relatively safe.

    Well, at least I am not blaming UFOs....