This sounds like it'll be a real joy to keep track of. Did Fidelity (Alltel to Tanta) release a whole new workstation for this?

Did any of the compliance vendors write code to do the TIL? I mean, a step loan with a normal balloon plus a zero-interest balloon? What's the effing APR on that? I hope the FDIC ran that past the Fed . . .

So, what is the time value between a first year 3% and a fifth year 6.5% over 40 years?

Oh, and they get a five year option on a 6.5% fixed? What if rates in year 5 are 8%?

I'm not sure I understand your first question. You were comparing what Indy was paying on 5-year deposits to these loans. The question would therefore be what the effective yield is on the loans over the first five years, to match assets/liabilities. The average over five years is 4.9%.

They don't get a "five year option," they get a fixed rate loan (that has an initial rate break in it). If rates in year 5 are 8%, these loans will do just what regular old 30-year fixed rates originated at 6.50% will do: sit there yielding 6.50%. Somebody is going to have to hedge that risk. Somebody will undoubtedly start that process by lowering those 5-year CD yields--no hedge works well if you've already got a cost of funds gap.

Then again, this is why very very few depositories put fixed-rate loans in their portfolios. They put ARMs in their portfolios, and they securitize the fixed-rate stuff. Insofar as the FDIC is putting fixed rates onto Indy's books, they're adding to the interest rate risk of that portfolio. My own guess is that the unannounced step two of this plan is to securitize all those modified loans to get the bank out from under that rate risk.

Umm, Tanta, just for today, can we have a politics thread?

No. I'm sorry, but I'm not going to start that.

There are only about a million political blogs out there for this to be discussed amongst people who self-select for "giving a shit about political opinions."

People come here, in part, to get away from political blogs. We self-select for "giving a shit about finance and economics."

So the answer is "Go bug Atrios."

i thought the interest rate step down could go as low as 1.5% (6.5% - 5%) based on the fdic press release???

I don't know what press release you're reading.

So the bonds drop their payout:
1. Holders of the bonds see less per Q.
2. Bond values drop.
3. Writedowns due to lower bond values.
4. Some CDS triggered due to payout reduction.

Yes, but they're still doing the "least loss" test vis-a-vis foreclosure. The idea is that in the absence of this workout effort, the bondholders would be worse off because all these loans would go right to FC and those losses would be more than the mod-related write-downs.

I guessing that the spread between what somebody would pay for those bonds and what that bank is taking in would mean that they'll just try to keep 'em and mark 'em to their imagination.

Let's not get into the whole "mark" problem here. Dammit.

If these loans are in Indy's portfolio, then they are "held to maturity." That means they don't get marked to market. They take the impairment write-down at the time the mod is done, then they only get further write-downs if and when they actually start going delinquent or further underwater. Just like any old portfolio loan. Those write-downs taken now hit current income, now.

What I was talking about is interest rate risk, which is a question of income, not balance sheet valuations. The problem with long-term fixed rate loans is, obviously, that they can end up yielding less than your current cost of funds. The only way for a bank to hedge that risk is to offset it with shorter-term assets (like T bills) and/or paying lower rates to depositors for longer-term deposits and/or finding some other source of funds besides deposits. As I have observed before, one of the most reliable ways to find other ways of funding loans is to securitize the things.

Yes, they don't negatively amortize, but you have the same situation of the payment doubling over five years and the "owner" ending up deep underwater. It's a virtually guaranteed future default. Because the ramp-up is gradual the default may come sooner but seriously, how many families will have their income double in the next 5 years?

But the payments don't "double" with a 1.0% per year rate increase for five years.

Kevin did the math for you on an example loan:

4% $2083.79 (44%)
5% $2392.93 (51%)
6% $2714.75 (58%)
6.5% $2879.07 (61%)

The HTI there assumes flat income.

Returning to my previous question, how would FDIC/Indy securitize this paper? The private-label market is basically moribund, no?

Yes, it is. So who do ya think is going to securitize it? Starts with an "F" . . .

But would the interest rate rise if the home price appreciated?

You mean contractually? Could IndyMac yank up a fixed rate just because you now have equity? Uh, no. A FIXED rate is FIXED. FROZEN. STUCK. UNCHANGING. If the contract says the rate caps out at 6.50% and never goes more or less than that, then that is what it means.

If you are asking whether long-term interest rates will rise in an economic environment in which home prices are appreciating again, then I don't know what to tell you. WHY WOULD THEY?

You're just getting confused listening to people like MS. 6.5% is the current 30-year fixed rate because that's what the bond market's inflation expectations are telling them to demand.

Some of you people need to spend less time cluttering up my threads with political bloviating and more time trying to understand finance and economics. I swear. You're worried about Fannie and Freddie buying these loans, but you're not worried about depositories loading up on fixed rates without an effective hedge? You're worried about market interest rates hitting 8-12%, but not worried about the fact that everyone is making new loans today at 6.50%? I don't get it.

Wouldn't this complicate the little issue of deleveraging/recapitalizing those beasts?

I guess it depends on what you mean by "complicating" it.

After everybody including the FDIC finds out what happens when depositories load up on fixed rate loans and run into their own asset/liability gaps, I strongly suspect that this fixation on Fannie and Freddie will fade a little bit. Another generation will learn why we have a secondary market for fixed rate paper.

I am missing something or did you answer your own question. CR posts last night were about FRE/FNM paper which will affect mortgage rates. With all due respect aren't all these things related?

Yes, you are missing a whole lot.

The "paper" being referred to is debt. The article says the GSEs are paying 1.23 over TSY for three-year paper. They use the proceeds from issuing that debt to buy mortgage loans for portfolio, if they want to do that. They could buy a boatload of 6.50% 30-year loans at the kind of discount we're talking here (for previously modified loans) and still have a positive net interest margin.

Or, of course, they could issue MBS backed by these loans. Which means some end investor earns 5-5.5% on the loans, the rest (after the servicing fee) goes to the GSE to guarantee the credit risk, and so they are not debt-funded by the GSEs.

If the GSEs cost of funds get too high, of course, then that market rate will go up to 7.00% or something. Or those loans modified at a 6.50% cap will be sold at a steep discount. That means the Fs would buy the loans at a price less than 100% of the outstanding principal (the interest-earning principal, in this case).

99% of the snarky comments I read about the GSEs are incredibly ill-informed, which annoys me.

Can they even buy something with a zero-interest lump attached on the tail end, such as you described? Who gets the joy of having this in their pool?

No, they won't "buy" that part. That part's a freebie.

As I said in the post, the first thing anyone who owns one of these loans (security or IndyMac) has to do is write down the loan by the amount of that forborne principal. It is as close to a classic case of "uncollectible" as the accountants ever get. So that's an immediate loss.

If they then securitized these loans, the UPB would be the interest-earning part.

Now, the cost of the credit guarantee on these loans, if in fact a GSE put them into the MBS, would be offset by potential recovery of the forborne balance. I don't know exactly how they'd do that math, but the point is that the recovery of such principal, if it occurred, would be the GSE's, not the end investor's. So it would reduce the cost of guaranteeing the credit on MBS.

If the GSEs bought the loans for portfolio, they would simply adjust the price they paid based on whatever they estimate for future losses, net of any recovery of "balloon" principal.

Again, though, I was trying to talk about rate risk, not credit risk. Any bank can keep loans like this in portfolio and make their own adjustments for collectibility of principal. But there's a separate issue about long-term fixed rates on the loans that perform. Conceptually, you don't have rate risk on loans that either fail or prepay. The rate risk is what you get if these modifications work.

So Tanta, you're saying that the loans are left in the bond pools if they were securitized?

ABSOLUTELY.

That is this whole folderol about following investor rules (less loss to the security) for modified loans!

Note that we are talking about non-GSE private-issue securities.

If any of the IndyMac loans in question were in GSE MBS, they would have to be either bought out of the pool by the GSE or by Indy, depending on how their specific contract reads. But IndyMac had the big private-issue Alt-A servicing portfolio, not the big GSE MBS portfolio.

Tanta has me convinced that I am A dumb ass. So I will leave. Bye y'all and good luck.

This blog is actually a very hospitable place for people who want to learn more about how this stuff actually works, you know. Which was kind of the point of the original post.

But if you are going to throw off a bunch of uninformed one-liners about "Hank" and the GSEs and the bond market without having a single clue what the issue actually is, then yes, I'm capable of making you feel like a dumbass. And it's fine with me if you go back to Mish or Ken or Yahoo or wherever else you've received your half-baked ideas lately. Nobody forced you to come here and throw out slogans like "let the market decide" when you seem to have a problem understanding how the market functions.

We all need to do something about the signal-to-noise ratio here, my friends. And that includes giving it a rest with the VP comments. Some of you regulars know better than to participate in letting the threads here degenerate into posturing.

I'm curious - what percentage of loans do lenders anticipate will go the full term? Is it simply that these are 'temporary' patches to keep things afloat until the market stabilizes, expecting sales and re-fis to be the final 'cleanup' later?

You may have to break down and read Tanta's famous post on convexity.

Short version: loan life shortens when prevailing rates fall and lengthens when prevailing rates rise. Also, loan life shortens when home values are rising and lengthens when home values are falling.

Investors in a 6.50% Freddie Mac MBS, which is what this rate is based on, are getting a coupon of about 5.5-6.0%, depending on the kind of loans in the deal. Remember that these are credit-risk free to the investor, because Freddie takes the risk. So you can back into an estimate of how long investors expect these loans to be around (average loan life) by comparing 5.5% or so to the duration of alternative investments that have no credit risk.

Calculated Risk: Options Theory and Mortgage Pricing 

Why can't the lenders just add the three or four missing payments to the end of the loan?

"Adding principal to the end of the loan" is EXACTLY WHAT THEY ARE DOING.

It's called a "balloon payment." We talked about that in the post. If the unmodified balance (including any past-due payments) is too high to result in affordable monthly payments by fully amortizing it, then you can balloon part of the balance.

That's what the post is about.

So if I understand this correctly, they're not writing down the principle, but instead they are giving an under-market rate, teaser-rate, balloon payment loan which is effectively a negative amortization loan.

You do not understand this correctly.

I do not know how to get people to stop using the term "negative amortization" incorrectly. I pointed out several times in the post that these loans fully or partially amortize. I am not lying to you about that. It's really true.

I also pointed out that very few borrowers are going to get principal forbearance, given the effect of the rate reductions. Some of you blew through that part too.

Of course these borrowers are getting a rate reduction; that's what workouts do. But they are also being forced to pay down principal every month of the loan.

but mathematically speaking how is offering 6.5% with an amortizing on some of the principal and 0% on the rest of the principal with a balloon any different from offering a lower rate on all of the principal with non-amortizing payment and a balloon?

In the first case, the loan partially amortizes. In the second case, the loan does not amortize. This, mathematically, is the difference that matters.

Because they WANT people to start repaying principal.

Furthermore, you need to review the post again. They only forbear principal after they've tried the stepped rate and 40-year amortization thing. They are not therefore changing the interest rate to 6.50% and then forbearing principal. They're forbearing principal only after they've had to go down to a first-year step of 3.00% and still couldn't hit a 38% payment ratio.

Please note that the lower the interest rate on a loan, the faster principal pays down. These "stepped" loans don't just amortize, they amortize FAST, because more principal is being paid in the earliest years of the loan. That speeded-up amortization on the interest-bearing balance offsets the lack of amortization on the forborne balance. So the deal still pays down principal from inception of the mod.

is there a magic % of the principal that becomes at some point "the line in the sand" where the modification cannot be made? IOW, is there a pre-determined max. percentage of forborne principal that they will allow?

I guess I just assumed that you all had read the earlier posts on the program.

So you would remember that the "least loss to the investor" test is in place for these loans. The loss in a mod has to be less than the loss in foreclosure.

This implies that there is indeed a limit to how much principal you can forbear: you have to stop when the loss gets worse than foreclosure would be.

Another way to look at it, of course, is to go back to JP Morgan's point that a loan with a HTI of 60% would only need principal forgiveness of 17-18%. It is extremely hard to imagine a loan with an HTI greater than 60% that ever made a single payment. So since you are trying to hit a 38% HTI with these steps, and you do them all in order, to imagine that they could ever have to forbear more than 17-18% of the principal on a given loan is to imagine that there are lots of loans that started out with a house payment greater than 60% of pre-tax monthly income and such people occupy these homes and they want to keep them.