Explaining Southern California's economy

Freddie Mac scandal: What the heck is an 'inverse floater?'

While Sales Of Existing Homes Rise In July, Prices Continue To Fall

David McNew/Getty Images

A recent ProPublica/NPR report on Freddie May refusing to refinance mortgages for struggling homeowners shows that the market is still coming to terms with new ways of measuring risk.

This morning, NPR and ProPublica fired a broadside at the government side of the embattled mortgage market, with a report on how Freddie Mac is refusing to refinance loans, reportedly because Freddie has developed a financial derivative that would suffer if interest rates on loans were lowered. Here's a sampling, and please note that there's something called an "inverse floater" at the core of the alleged problem. The Silversteins are a couple who have been unable to refinance their mortgage after doing a short sale on another property:

Here's how Freddie Mac’s trades profit from the Silversteins staying in “financial jail.” The couple’s mortgage is sitting in a big pile of other mortgages, most of which are also guaranteed by Freddie and have high interest rates. Those mortgages underpin securities that get divided into two basic categories.

One portion is backed mainly by principal, pays a low return, and was sold to investors who wanted a safe place to park their money. The other part, the inverse floater, is backed mainly by the interest payments on the mortgages, such as the high rate that the Silversteins pay. So this portion of the security can pay a much higher return, and this is what Freddie retained.

In 2010 and '11, Freddie purchased $3.4 billion worth of inverse floater portions — their value based mostly on interest payments on $19.5 billion in mortgage-backed securities, according to prospectuses for the deals. They covered tens of thousands of homeowners. Most of the mortgages backing these transactions have high rates of about 6.5 percent to 7 percent, according to the deal documents.

Between late 2010 and early 2011, Freddie Mac’s purchases of inverse floater securities rose dramatically. Freddie purchased inverse floater portions of 29 deals in 2010 and 2011, with 26 bought between October 2010 and April 2011. That compares with seven for all of 2009 and five in 2008.

In these transactions, Freddie has sold off most of the principal, but it hasn’t reduced its risk.

First, if borrowers default, Freddie pays the entire value of the mortgages underpinning the securities, because it insures the loans.

It’s also a big problem if people like the Silversteins refinance their mortgages. That’s because a refi is a new loan; the borrower pays off the first loan early, stopping the interest payments. Since the security Freddie owns is backed mainly by those interest payments, Freddie loses.

That last paragraph is important, because what Freddie has done is keep the "refinancing risk" to itself, and the implication is that because Freddie controls the levers and can obstruct refis, it can package securities that are attractively low-risk for investors while retaining the high-risk stuff and "game" the risk-management for its own benefit.

Matt Levine at DealBreaker struggles mightily to crack this nut open and does as close to a masterful job as I think is possible, given the ornate game of hot-potato risk-shifting that's going on here:

All you need to know is that Freddie is taking a disproportionate amount of refinancing risk on the mortgages it guarantees and packages, which means that the buyers of those packaged mortgages are getting disproportionately less refinancing risk.

And why would Freddie do this?

One possible reason might be “Freddie has the ability to change its rules to make it harder for the Silversteins to refinance, so it thinks it has the best ability to outperform market expectations on refinancing risk, so it hangs on to that risk itself because it believes it’s being overpaid for taking the risk.” This is the view that ProPublica takes and it’s pretty plausible because Freddie apparently did change its rules to make it harder to refinance, tightening its credit checks to exclude borrowers who have a recent short sale. That’s at least awkward given Freddie’s whole role as, whatever, government guarantor of homeownership. 

The tricky thing here, which Levine zeroes in on, is that somebody has to absorb the refinancing risk, and if Freddie (and also Fannie Mae, the other big government guarantor) takes it, then the private sector doesn't have to worry about it — and Presto! this means that refis by the private sector will be much easier to come by.

I should point out that Levine is clearly racking his brain over what exactly is going on here. I've also been trying to determine why inverse floaters are necessarily a bad thing, given that you don't want to be in the business of selling refi risk if you don't have to, and the whole point of Freddie (and Fannie) is to create a market for mortgages that allows the government to soak up risk and then distribute it. This implies keeping the worst of that risk itself.

The title of Levine's post kind of says it all: "Freddie Mac Profits From The Misery Of American Homeowners So That Banks Don’t Have To." The troubling aspect to this is that Freddie wants to play both sides. On the one hand, it performs its duty by taking on the least-desirable risk. On the other, it then manipulates (maybe) that risk because it figures it should make more from it. Good and bad in the same bucket, if you will.

Mind you, ProPublica and NPR could have done a better job of explaining this. As it stands, we get this:

These inverse floaters burden Freddie with entirely new risks. With these deals, Freddie has taken mortgage-backed securities that are easy to sell and traded them for ones that are harder and possibly more expensive to offload, according to mortgage market experts.

So something funny is going on here. But as I and others have ascertained, something funny might be going on for a good reason. 

Follow Matthew DeBord and the DeBord Report on Twitter.

blog comments powered by Disqus