Skip to main content

How Much Money Will Fannie And Freddie Get From Banks For Crappy Mortgages, Use To Make Further Crappy Mortgages?

  • Author:
  • Updated:

The Federal Housing Finance Agency’s lawsuits against every bank paint a pretty dastardly picture of the seventeen big banks - three of which are now BofA - committing all sorts of frauds in securitizing mortgages and selling them to Fannie Mae and Freddie Mac. This in turn caused Fannie and Freddie to have a series of accidents that left them wards of the state under FHFA conservatorship. FHFA's lawyers are thorough, quoting among other people the Financial Crisis Inquiry Commission, SEC investigations and suits, and even Matt Taibbi. One thing they don’t do, however, is give any hint of how much money they’re looking for.

Others have jumped in to do the math for them. A popular approach, taken by FT Alphaville and Nomura, guesstimates that the damages will be around 20% of the original principal amount. The reason for this is that it seems to be the amount claimed by FHFA in its suit filed against UBS in July – there, FHFA sought $900 million on $4.5bn notional of mortgages. As the total claims against all the banks are around $200 billion, that gets $40bn of potential damages.

Keefe Bruyette & Woods has a higher estimate of $60 billion. They note that the potential liability is for "rescission": if the FHFA proves that the banks lied materially in their prospectuses, then the banks are on the hook to buy back the mortgages at par, and thus are at risk even for losses due to the general housing market downturn and unrelated to shoddy underwriting.

Goldman has a note out today that uses two methods – the 20% UBS precedent, and the precedent of the proposed $8.5 billion settlement that Bank of America negotiated with its non-FHFA private label mortgage investors:

The BAC PLS settlement implies aggregate losses of $5 bn, while the FHFA’s suit against UBS (suing for $900 mn on $4.5 bn of securities) implies almost $40 bn. Assuming 20% cumulative losses of which 50% were due to improper underwriting implies $20 bn of losses. Friday’s move down seems to imply 10%-20% losses for the banks (or $20-$40 bn).

Here’s the chart:

Goldman guesses how this will play out / gets in a dig at BofA’s capital position:

Both timing and the manner of resolution are uncertain at this point. However, given the difference in roles played and the complexities of a global settlement (e.g., negotiations for the robo-signing settlement continue without resolution), we would expect the banks to either (1) opt for individual settlements or (2) allow for resolution via the courts. We believe banks will claim that differences in underwriting standards, loss rates and roles will make it difficult to paint everyone in the same light. Additionally, some may be incentivized to “put it behind them,” given their robust reserves (JPM), while others may prefer to grow their capital base for a few more quarters before any resolution (BAC).

Finally, Felix Salmon takes the throw-it-down-the-stairs approach, counting pages in the filings to find JPM and Bank of Amerrillwide to be the big losers.

Two thoughts:

First, you can get a rough upper bound by figuring out how much Fannie and Freddie have lost on private label securitizations, since they're probably not suing for more than they’ve lost on all securities ever. Here’s a quick effort at that: if you take the total of other-than-temporary impairments in Freddie’s available-for-sale securities from 2008 to 2010, you get around $42 billion. Add Fannie’s impairments and fair-value losses and you get another $40bn. That’s a total of $82 billion that FHFA's little sluggers lost on all RMBS securities that they owned since the 2008 crash. That number is too high because (1) they owned non-private-label securities (including their own) that have lost value for non-fraud reasons, and (2) they’re not suing everyone – Wells Fargo, among other big names, has been left out. Assume that they’re seeking to recover half of those losses from the banks sued on Friday, and you get close to Nomura’s $40 billion number for plausible damages.

Second, Goldman calculates what is arguably a lower bound by using the precedent of Bank of America's $8.5 billion settlement with BoNY Mellon and private investors on $424bn principal amount of private label mortgage securities. That's damages of 2.4% of the principal amount. But nobody - including Goldman, which uses 20% damages to estimate effects on capital - seems to believe that the FHFA will cut a deal as cheaply as that.

So why should the BAC private investors? Their settlement is not final yet, and there have been lots of objections arguing that the settlement is much too low. If the FHFA suit ends up working out to damages of 10-20% of principal, that would imply a $40-$80 billion number on the private suit, much higher than the current $8.5bn deal - and a much bigger headache for BofA than the $5-$12bn that they reasonably stand to lose on the FHFA suit directly.