In Depth Analysis: CalculatedRisk Newsletter on Real Estate (Ad Free) Read it here.

Wednesday, October 31, 2007

Accounting For Negative Amortization

by Tanta on 10/31/2007 10:00:00 AM

Accounting for negative amortization is a perennial favorite amongst those who follow banks and thrifts with large Option ARM portfolios. It outrages a lot of folks that the neg am balances, which represent interest that has been earned but not paid, is considered noncash interest income. It also seems to outrage a lot of folks that OA portfolio holders do not simply declare these capitalized balances as "uncollectable." The idea seems to be that 1) the fact of negative amortization itself should mean that the loan is substandard or doubtful, regardless of timely payment of the contractually allowed minimum payment, and therefore 2) the accounting treatment for reserve purposes should be the much more onerous one, "all estimated credit losses over the remaining effective lives of these loans," rather than the standard for non-classified loans, "all estimated credit losses over the upcoming 12 months."

I think this is the argument Jonathan Weil is trying to make in this piece on WaMu:

As for the fourth quarter, Washington Mutual predicted that provisions would be $1.1 billion to $1.3 billion and that charge-offs would increase 20 percent to 40 percent.

To see why even $1.3 billion in provisions looks light, consider Washington Mutual's $57.86 billion of so-called option- ARM loans, which make up 24 percent of Washington Mutual's loan portfolio. These adjustable-rate mortgages were popular during the housing bubble, because they give customers the option of postponing interest payments, which the lender then adds to their principal balances.

As of Sept. 30, the unpaid principal balance on Washington Mutual's option ARMs exceeded the loans' original principal amount by $1.5 billion, meaning the customers owed $1.5 billion more in principal than what they originally borrowed. By comparison, that figure was $681 million a year earlier, when Washington Mutual had $67.14 billion, or 16 percent more, option ARMs on its books.

Look to the end of 2005, and the trend becomes even starker. Back then, Washington Mutual had even more option ARMs on its balance sheet, at $71.2 billion. Yet the unpaid principal balance exceeded the original principal amount by only $160 million -- and that was up from a mere $11 million at the end of 2004.

Deferring Pain

The deferred interest from option ARMs also boosts Washington Mutual's earnings, part of a process known as negative amortization, or ``neg-am.'' That's because option-ARM lenders recognize interest income when customers postpone their interest payments, even though the lenders got no cash.

For the nine months ended Sept. 30, Washington Mutual recognized $1.05 billion in earnings as a result of neg-am within its option-ARM portfolio. That represented 7.2 percent of Washington Mutual's $14.61 billion of total interest income year-to-date. By comparison, neg-am contributed 1.8 percent of Washington Mutual's interest income for all of 2005 and just 0.2 percent for 2004.

What's going on here? Either the borrowers postponing their interest payments are doing so as a matter of choice, by and large, or they can't afford to pay them. Common sense suggests it's the latter -- and that there's serious doubt Washington Mutual ever will collect the $1.5 billion of postponed interest that its option-ARM customers have added to their original principal balances.
If you have any sense of how the OA market worked in the last three or four years, you have to find Weil's approach here a bit odd. Back in 2004-2005, when WaMu had more OAs on its books, it had younger OAs on its books. It takes some time for OAs to build up neg am balances, and as origination of this product really didn't take off until 2003-2004, you wouldn't expect any portfolio of OAs to have large capitalized balances for at least a few years.

What Weil is doing is trying to find a negative trend in the performance of these loans: his "common sense" says that the very fact of negative amortization means the borrowers are in trouble, and the very fact that neg am balances are growing in WaMu's portfolio means that a reasonable person would assume that this pattern should be projected into the foreseeable future. What that implies, of course, is that WaMu should "classify" all of these loans, regardless of LTV, timely payment, etc., on the recognized accounting basis of the "negative trend." If they did that, they would have to reserve a lot more against these loans, since allowances for classified loans are required, by the OTS, to be for the life of the loan. Reserves for nonclassified loans are for the next twelve months.
Yet the $1.1 billion to $1.3 billion of fourth-quarter provisions that Washington Mutual predicted -- for the company as a whole -- wouldn't even cover the $1.5 billion of tacked-on principal. The trend among Washington Mutual's option ARMs shows no sign of slowing, either.

Through a spokeswoman, Libby Hutchinson, Washington Mutual officials declined to comment. She said the company's executives aren't fielding questions until their next meeting with investors on Nov. 7.

Then there's the bigger picture. While Washington Mutual's loan-loss allowance rose 22 percent to $1.89 billion during the 12 months ended Sept. 30, nonperforming assets rose 128 percent to $5.45 billion. So even if Washington Mutual adds $1.3 billion in provisions next quarter, its loan-loss allowance still won't be anywhere close to catching up.

To be sure, Washington Mutual executives have some latitude over the timing of the company's loan-loss provisions. Yet they also may have a monetary incentive to push losses into 2008.
Surely not even the greatest OA skeptic believes that WaMu could conceivably face default of every last one of its OAs with a neg am balance in the next 12 months. Without saying so explicitly, Weil is suggesting that WaMu pack lifetime estimated losses on this portfolio into current reserves, for no other reason than that the loans are negatively amortizing.

To me, that is the crux of all of this upset over OA accounting. I personally would not make OAs nor would I hold them in any portfolio over which I had control, so don't think I'm defending the product. However, we just went through a major regulatory effort on "nontraditional mortgage products," and the upshot of that was that the regulators did not and would not deem the negative amortization ARM an unacceptable product for depositories, in and of itself. There is plenty in the Nontraditional Mortgage Guidance about what underwriting practices and so on should be followed with these loans, and certainly the guidance isn't a carte blanche for writing any old dumb OA an institution can think of. But they are not, explicitly, "classified" just because they're OAs:
When establishing an appropriate ALLL and considering the adequacy of capital, institutions should segment their nontraditional mortgage loan portfolios into pools with similar credit risk characteristics. The basic segments typically include collateral and loan characteristics, geographic concentrations, and borrower qualifying attributes. Segments could also differentiate loans by payment and portfolio characteristics, such as loans on which borrowers usually make only minimum payments, mortgages with existing balances above original balances, and mortgages subject to sizable payment shock. The objective is to identify credit quality indicators that affect collectibility for ALLL measurement purposes.
I do not know how to read this except that there must be specific indicators of credit quality in the analysis besides the fact that the loans are "nontraditional" and that the issue is capital adequacy, not just reserves (capital being expected to cover long-horizon potential losses, and reserves being expected to, well, cover the short term). Weil's "common sense" may tell him that neg am = loan distress by definition, but the regulators' common sense didn't tell them that, and whose common sense do you think matters to WaMu?

What this means is that the federal regulators have said that the fact that a loan accrues neg am balances does not, in and of itself, make the loan unacceptable, substandard, or uncollectable. How, precisely, banks are supposed to get away with reserving for them as if they were beats me: the regulators can get on your case just as much for over-reserving as for under-reserving, as this can smack of "cookie jar" accounting. The only way a bank could defend itself against the charge of over-reserving would be for it to define OAs as unacceptable as a product, without regard to any other facts or characteristics. Why does Weil or anyone else expect an originator of OAs to do this?

Similarly with the issue of treating neg am as income: what else would you treat it as? To argue that deferred interest is never in fact collectable is to argue that OAs always default and the recovery is never enough to cover the balance due. If you believed that to be true, you would never make such loans.

And the issue of a "trend" suffers from the same problems. OAs allow for negative amortization up to some limit. That is established in the legal documents when the loan is closed. When you make those loans, you must assume that any and all borrowers may elect to make the minimum payment. That means that a "trend" over time will occur in a highly predictable way, known as an "amortization schedule." Certainly some industry participants have expressed some pearl-clutching surprise over the fact that making the minimum payment seems to be near-universal among outstanding OA loans, but we can file that under the "stunned but not surprised" heading. You do not make neg am loans unless you're prepared for neg am.

The point: the accounting treatment for these loans--reserves, asset classification, income--isn't going to change as long as they're "legal." And nobody is going to reserve for an OA portfolio today assuming that it will be a total loss in the next twelve months. And nobody is going to stop treating accrued but unpaid interest as income. If you have problems with that--and you surely might--then what you have problems with is allowing banks and thrifts to originate and hold these loans tout court, because you have implicitly defined them as substandard to the extent that they do what they are designed to do.

I realize that it is, in some quarters, more entertaining to speculate about accounting shenannigans than it is to face up to the implications of your rhetoric, but there it is. If you want OAs to be illegal, say so, and let me know what happens when the free-marketers jump all over your case. Otherwise, this business of implying that WaMu is reserving against its performing OA portfolio only for losses expected in the next twelve months because it's playing bonus games, not because that's what the reserve rules are, is really disingenuous.

One can make the case that it is simply impossible to accurately treat a portfolio of OAs: they're either always under-reserved or always over-reserved. Fine. I have some sympathy with that argument. But it's an argument for the abolition of OAs, not a criticism of any one bank's application of accounting rules. In any case, I challenge anyone who has a problem with WaMu's accounting for its OA portfolio, but who does not think the product should be outlawed, to explain to me what it is you do want.