Tuesday, October 23, 2007

I want them alive... No modifications!

I've been thinking a lot about loan modifications, which is basically when a bank voluntarily agrees to alter the terms of a loan, and here I'm talking about a mortgage loan, in an attempt to avoid foreclosure.

No one has covered this issue better than Calculated Risk (for example: here but CR and Tanta have done many posts on the subject.) CR and others have spilled a fair amount of virtual ink on the problem that most mortgage loans which have been securitized either cannot be modified, there are severe limitations on modifications, or there is no single party motivated to choose modification.

This leads us to an unfortunate reality. In 2007, given that such a large percentage of loans are held in securitized form, loan modifications are harder than ever to achieve. And yet, given that rate resets are a big part of the subprime problem, its likely that loan mods would be more successful in limiting lender losses than in times past.

Other sites have done a fine job in covering this issue from the consumer's perspective. But AI is all about the cold hearted capitalists, and after all, bond holders are de facto lenders here. So I'm going to give some thoughts about loan modifications from the perspective of a CDO/ABS holder.

In the olden days, back when men were real men, women were real women, and banks were real banks, the negotiation of a loan modification could theoretically be held between all interested parties: the borrower and the lender could actually meet together and hammer something out. If the borrower fell hopelessly behind, the bank could decide whether foreclosure or the modification would result in the minimal loss to the bank.

The securitization business thrives on homogenization. So when it came to the issue of loan modifications, investors wanted strict rules about what could and could not be done. Remember that ABS and RMBS (and later CDOs) were built on complex mathematical models about default and recovery. If a servicer was too aggressive about making mods, then that would mess up the nice neat models. You know, quants get very cranky when you mess with their models...

Loan modifications were often not successful in avoiding eventual foreclosure. But in an investor pool with both senior and junior note holders, the timing of payments becomes a huge issue. Say a loan was modified such that the borrower stayed current for another 8 months but then fell back into arrears and was eventually foreclosed upon. If that loan was in an investor pool, then some of the payments made during the 8 months of modified payments likely went to junior note holders. Senior note holders had a legitimate gripe: had the loan simply been foreclosed upon, payments would have flowed to the senior note holders first, likely leaving nothing for junior note holders. In effect, the loan mod benefited junior note holders at the expense of senior note holders. This is particularly true in pools where there is any kind of trigger. Increased foreclosures may trip the trigger, which usually causes more cash to flow to senior note holders, where as loan mods may prevent a trigger, keeping the junior classes around for longer, soaking up cash.

This view was reiterated on a recent dealer conference call. The traders were advocating senior tranches of subprime pools from the worst originators. The argument went that pools with larger early payment defaults would cause the payment triggers to be tipped. This results in all cash flow being paid sequentially, with the most senior tranches getting all payments until completely paid off. Given that these kinds of bonds are trading at deep discounts and that the structure had substantial subordination to begin with, there is an opportunity for strong returns.

On the other hand, deals with fewer initial defaults and did not breech the trigger levels would continue to pay pro rata (proportionately to senior and subordinate tranches alike). However, despite relatively low initial defaults, odds are good that defaults will keep rising. Any cash paid out to junior tranches just leaves less for senior tranches down the road. If you are going to get to a high level of defaults anyway, the trader reasoned, why not do so where your tranche is getting all the cash flow?

OK so back to loan mods. Investors objecting to loan mods really need to think this through, regardless of where you are on the capital structure. Say I own the senior most piece (originally rated Aaa) of a 2006 vintage subprime RMBS currently trading at $95. We know its trading at $95 because of default fears. Note that a $5 loss indicates about 125bps in spread widening. So for reference sake, let's say that the pool originally had a coupon of 5.5%, but at a dollar price of $95 would have a yield to average life of 6.75%.

Another way to think of the 125bps of spread widening is that if the same security was created today and priced at par, the coupon would have to be 125bps higher, or 6.75%. Or put another way, investors would pay $95 for a bond with a coupon 125bps lower, yet no default risk. Make sense? So if the senior bond holders were given the option of eliminating default risk in exchange for reducing their coupon by something less than 125bps, they'd likely accept.

Well subprime pool holders, this is your lucky day, because through the magic of loan mods, just such an exchange is possible. Say that half of a pool of MBS is going to reset in 2008 at levels 400bps above the teaser rate, and that these borrower will struggle to make. If these loans were all modified such that the reset was merely 200bps higher, then the net coupon on the deal will only be impaired by 100bps. And since the senior tranche is, say, 90% of the deal total, its coupon is only 90bps impaired!

I know the comments will be filled with two primary objections. First is moral hazard, but frankly we have two bad actors in this case. The borrower who probably should have known better, and the investor who really should have known better. Well, the originator too, but he's probably out of business anyway. The borrower is being given a little of a free pass, but the investor isn't seeing his loss taken away, merely reduced. I don't think investors in Aaa securities are going to have their losses reduced from 5% to 3.5% and consider themselves bailed out. A 3.5% credit loss in a Aaa security is still awful. I also don't think the borrower will walk away from this experience saying "What a great choice that ARM was..." If you see your interest rate go from 5% to 7%, I think that's plenty painful for borrowers to reconsider the ARM idea, the fact that it might have gone to 9% is besides the point. So I'm not buying the moral hazard issue.

Second is the fact that many modified loans wind up defaulting anyway, as stated earlier. This time is different, though. Because the borrower who got in trouble entirely because of a rate reset isn't the same as the classic delinquent borrower who merely can't keep a job or handle credit cards. Remember that all these subprime deals were priced assuming a certain level of defaults: a level consistent with the "classic" reasons for delinquency. If we could do enough loan mods to make the "classic" reason for default the most common reasons, then we'd have a better chance of containing the subprime contagion.


Anonymous said...

So what are the statistics on defaults? Are the subprime issues mainly defaulting because of rate resets or because of poor underwriting? Why are we seeing tons of defaults on recent issuance(06,07)? Hard to think that stuff has reset yet? So is the real problem that homeowners really can't afford their home even at the teaser rates? Look at debt to income ratios....Can't see how loan mods solve the problem...they only delay the inevitable...

Robert said...

"Second is the fact that many modified loans wind up defaulting anyway, as stated earlier. This time is different, though. Because the borrower who got in trouble entirely because of a rate reset isn't the same as the classic delinquent borrower who merely can't keep a job or handle credit cards."


Loan mods are not the evil that many people think that they are. Everyone benefits from preventing defaults in this case. Your post was an excellent analysis of this issue.


I agree with the ANON that this may only delay the inevitable. This is not due to the "classic" reasons that you mention, but more to do with HPA.

Modifications will allow the payments to come in line with what most owners can reasonably pay. However, at some point certain borrowers (especially in the case of subprime) will not want to continue paying on a home/condo that is worth 20-25% less than the existing mortgage.

While most borrowers would attempt to keep up with the payments to preserve their credit, a subprime borrower has little skin in the game.

Why keep paying $2500 for a house that you can rent for $1200? That may seem extreme, but I have seen many such examples of exactly that situation where I am from (admittedly, ground zero).

Now, combine the above with some of the "classic" reasons for default and we have the ingredients for a perfect storm of delinquencies.

Fortunately, through continued job growth, rate cuts, loan mods, etc., etc., that storm is likely some time away.

Accrued Interest said...

To both Anon and Robert:

I think higher defaults are inevitable, due mostly to poor underwriting. Nothing will solve the problem. But I think that structured product holders would benefit greatly if the default rate on subprime was held to, say, 15% and not 20%.

The June subprime delinquency rate is 14%. Not all delinquencies will turn into foreclosures, historically its usually been less than half. But I think we'd all agree that default rates are more likely to climb than subside within the subprime market.

Note that the 15% figure doesn't tell the whole story for subprime ABS/CDO holders. Those securities tend to own a certain vintage of subprime loans. The 2006 vintage of subprime is already over 10% "seriously delinquent."

Robert said...


I completely agree. My only point was that we could be talking about 7%-10% of these loans becoming "seriously delinquent" in a relatively strong economy (housing notwithstanding).

Modifications prevent borrowers from attempting to refi into a more traditional product (to the extent that it would possible), essentially keeping prime and near prime borrowers in a product that will be very difficult to service in more difficult times.

I, for one, do not think that we will see a recession in the near term (although the media loves to throw around the R-word these days). I think that we can all agree that recessions DO occur and that an economic slowdown is likely at some point in the future. Who knows when, I certainly don't.

To the extent that borrowers reject refinancing into a more conventional product (while it is still possible) and instead choose a modification, I think that that is a recipe for more housing pain.

We need a good ole' bear market rally in housing to set up the next stage of the slope of hope. Modifications can provide the breathing room to allow that to happen.

Robert said...

My point was this only helps current holders of structured products in the near term. These assets will also experience a potential bear market rally (with modifications) but are at great risk should a recession occur.

Also, most of these models allowed for a certain level of defaults, and then a worst case scenario. The "worst case" wasn't 10%-15% in many cases from what I understand. If too many defaults occur within too short a period of time, it "blows up". Modifications will prevent many of these securities from "blowing up" in the near term, but the risk is still there (Massive hit to Aaa super senior type of paper). Am I wrong on this AI?

It certainly gives current holders some breathing room to stop the bleeding and find some buyers for their paper. Which might be a good idea...

PNL4LYFE said...

First of all, great blog. The posts are great and the titles are even better.

Second, could AI or someone else please explain the math in the below section? How does the 200bp 'decrease' in rate only change the coupon by 100bp. And how does the size of the senior tranche affect this (i.e. 100bp to 90bp)? Thanks in advance!

"Say that half of a pool of MBS is going to reset in 2008 at levels 400bps above the teaser rate, and that these borrower will struggle to make. If these loans were all modified such that the reset was merely 200bps higher, then the net coupon on the deal will only be impaired by 100bps. And since the senior tranche is, say, 90% of the deal total, its coupon is only 90bps impaired"

Anonymous said...


Half of the pool will reset 400bps higher. Thus, the whole pool will average a coupon increase of 200 bps.

If these (resetting) loans were all modified such that the reset was merely 200bps higher, then the overall pool will average a coupon increase of 100bps instead of 200 bps.. the "only" 100bps in your first question.

90% of the now-100 bps increase on the entire security coupon = 90 bps coupon on the senior tranche.

Personally, I'm firmly in the foxhole with Calculated Risk and Tanta that the US is losing 4 trillion dollars of home equity. Of course, that's only 2 iraq wars, so maybe that's not all bad.

Anonymous said...

Maybe this is not related with the current discussion but Nouriel Roubini has been talking about the difference between liquidity and insolvency crisis for a long time. Shortly he says that cutting rates won't be an effective tool as when we consider the growing housing inventory. I was wondering how do we know what kind of crisis we are experiencing currently. As far as I know half of the subprime loans are cash-out refis. So the borrowers are already distressed and the coming resets will pull the trigger.

And is there a way to run these default (loss) scenarios on CDOs or ABS on Bloomberg? Thanks

bond rookie

Accrued Interest said...

I believe the only thing you can do on Bloomberg is alter the constant default rate. And I don't think that works on CDO deals, only ABS deals.

As for Roubini, we have different things going on here. We have a bursting housing bubble, we have defaults on subprime MBS, we have a liquidity crisis, we have stress on banks, we have weakening consumer balance sheets.

Obviously these things are all related, but its entirely possible for some of those problems to be solved (or perhaps ameliorated) while other problems continue.

For example, it may be that the awful earnings we saw the last couple weeks from brokers and banks represented a "fessing up" about where their CDO/ABS/MBS positions should really be valued. I'm just saying its possible. If that's the case, then the worst is behind us for those banks and brokers. Maybe they'll need to find something to replace the profits they had been making from MBS, but at least there won't be continuous calls on their capital.

But it could be that subprime defaults continue. I mean, let's say that Merrill Lynch wrote down all their subprime assets to reflect an expected default rate of 50% on 2005-2007 vintage assets. But foreclosure rates are currently in the 10-15% area, so those foreclosure rates could move much higher without Merrill taking any more writedowns.

Again, I'm not saying that's what MER actually did, I'm just trying to make an example of where some of our problems could improve while others get worse.

Accrued Interest said...


Certainly the expected default rate on 2006 vintage MBS is going to be far higher than the expected levels. Whether they get large enough to cause real losses in Aa or Aaa tranches remains to be seen. I think its going to depend on the deal. Some deals may see 15% foreclosures, some might see 50%.

I think we should all remember that the senior tranche on any securitized deal is going to get whatever cash flow is available. So even if losses are quite high, the senior-most tranches will keep getting paid something. I think this fact is ignored in the media.

Anonymous said...


What rights do bond insurers have in accepting or rejecting modifications?

I assume that they probably don't care if they are insuring the entire CDO, but in the case where they are wrapping a top tranche they might claim that adverse modifications would waive the policy.


Accrued Interest said...

I don't think bond insurers have any rights at all. I'm not totally sure on this, but I think the policy on a CDO looks a lot like a CDS contract,

If any readers know this better than I, please speak up.