by Tanta on 9/09/2008 10:53:00 AM

People keep sending me this article or bringing up this "fact" in the comments. Because it is such a fine example of a "fact" that isn't actually a fact, but is apparently becoming an article of quasi-religious faith in some quarters, I shall make the attempt to slap it down. I have no particular illusions about how well this will work, but there may be a handful of people who actually care about accuracy and good faith, even (!) when the subject is Freddie Mac. I'm talkin' to you.

*********************

Let me start out with a couple of general observations. This post is about financial accounting matters. If you are one of those people who drove us insane in the comments to yesterday's post on "assets" versus "liabilities" by arguing that "assets" are "really" "liabilities" because you, like Humpty Dumpty, are The Master, then you will find this post unsatisfactory. Tell it to the Marines. The habit of refusing to use standard accounting terms in preference to sloppy "synonyms" is what got these two reporters in trouble in the first place. I'm not going to pander to anyone by doing it myself.

Second, we basically went through a nearly identical version of this brouhaha last November with Fannie Mae. It's deja-vu all over again.

The offending "fact" comes from this article by the world-renowned Gretchen Morgenson and Charles Duhigg, whose willingness to believe anything any unnamed source says about Freddie Mac, whether it makes sense or not, has been documented before on this blog.

Finally, regulators are concerned that the companies may have mischaracterized their financial health by relaxing their accounting policies on losses, according to people familiar with the review. For years, both companies have effectively recognized losses whenever payments on a loan are 90 days past due. But, in recent months, the companies said they would wait until payments were two years late. As a result, tens of thousands of loans have not been marked down in value.

The companies have injected their own capital into pools of securities containing these loans, arguing that their new policies are helping more borrowers. Under conservative accounting methods, changing these policies would not have any impact on the companies' books. However, people briefed on the accounting inquiry said that Freddie Mac may have delayed losses with the change.

What follows is my best effort to discover what the hell these people are talking about. I must disclose to you all that I am really just making an educated guess here. If you possess any expertise at all in financial accounting in general and Freddie Mac's business operations in particular, the foregoing paragraphs do not make any sense whatsoever. (It's like talking to people for whom "asset" "really" means "liability.") So I could be wrong, and they could be talking about some entirely different part of the balance sheet. Anyone with a better guess than mine is hereby invited to share.

My theory is that they are talking about optional repurchases of MBS loans. I cannot think of or find any other part of Freddie's financial statements in which that "two-year rule" or this thing about "injecting capital" would fit. And if they are talking about optional repurchases, they're guilty of terrible reporting. Either their sources are badly informed or they didn't understand what their sources told them or both.

To review the basics of what Freddie does: they buy mortgage loans on the secondary market. These purchases of loans result in two different "portfolios" of loans: the "retained portfolio" and the "guarantee portfolio." The retained portfolio consists of loans and MBS that are owned outright by Freddie. That means Freddie's capital is invested in these loans. Freddie gets the capital to invest in the retained portfolio in large part by issuing notes and bonds--what everyone calls "agency debt." The retained portfolio constitutes an "asset" on Freddie Mac's books (net of the loss reserves), and the debt-funding constitutes a "liability" thereon.

The "guarantee portfolio" consists of various MBS that Freddie guarantees the credit risk of, but does not invest the capital in. The capital to fund these securities is provided by investors who buy MBS. Therefore, the total principal amount of the guarantee portfolio is not an asset on Freddie's books (it is an asset on the MBS investors' books). What shows up on Freddie's balance sheet is the "guarantee asset," which is the fair value of the guarantee fees received, and the "guarantee obligation" (over on the liability side) which reflects the fair value of the projected credit losses.

This distinction between retained and guaranteed portfolios is one reason why Freddie's (and Fannie's) financial statements are complex; each part of the "total portfolio" has different accounting treatment. If you read through these financial statements or any reports having to do with portfolio balances or loan purchase volume, you simply need to pay attention to when a number is given for the "total portfolio" versus one or the other parts thereof. To answer a question that may arise at this point, as of June 30 the principal balance of Freddie's retained portfolio was $792 billion and the guarantee portfolio balance was $1.410 trillion, making a total portfolio of $2.202 trillion (see Table 49 of the 10-K).

So. How do loans get into the retained portfolio? They are either originally purchased as portfolio investments or, in some cases, they were originally purchased in the guarantor program but had to be repurchased out of the MBS. As I said, the current flap seems to be about repurchases of MBS loans. I am going to quote here from Freddie Mac's 10-K. It will help you to know that "PC" means "Participation Certificate," and is just Freddie-speak for "MBS."

We also have the right to purchase mortgages that back our PCs and Structured Securities from the underlying loan pools when they are significantly past due. This right to repurchase collateral is known as our repurchase option. Through November 2007, our general practice was to purchase the mortgage loans out of PCs after the loans became 120 days delinquent. Effective December 2007, we no longer automatically purchase loans from PC pools once they become 120 days delinquent, but rather, we purchase loans from PCs when the loans have been 120 days delinquent and (a) the loans are modified, (b) foreclosure sales occur, (c) the loans have been delinquent for 24 months or (d) the
cost of guarantee payments to PC holders, including advances of interest at the PC coupon, exceeds the expected cost of holding the nonperforming mortgage in our retained portfolio. Consequently, we purchased fewer impaired loans under our repurchase option for the three and six months ended June 30, 2008 as compared to the three and six months ended June 30, 2007. We record at fair value loans that we purchase in connection with our performance under our financial guarantees and record losses on loans purchased on our consolidated statements of income in order to reduce our net investment in acquired loans to their fair value.

Remember that the "guarantee" on the MBS means that Freddie Mac is responsible for passing through interest payments to bondholders as long as those bondholders have principal invested, whether the borrowers make payments or not. The way this usually works is that for the first 90 days of delinquency (120 days since last payment), the servicer is obligated to advance scheduled interest and principal to Freddie Mac, who passes it through to the bondholders. The servicer makes efforts to collect the past-due payments from the borrowers. Generally at around 90-120 days, if the loan is still delinquent, the servicer's obligation to advance payments stops and Freddie Mac is the one obligated to advance payments to the bondholders. The basic contractual terms of the MBS are that Freddie has the right, but not the obligation, to buy a seriously delinquent loan out of the pool at this point. The repurchase price would always be par, since the bondholder must receive 100% of principal invested per the terms of the guarantee. Obviously, a seriously delinquent loan is likely to have a fair value of much less than par, but Freddie has to take that loss, not the MBS investor.

However, that is an option, not an obligation. Alternatively, Freddie can allow a seriously delinquent loan to remain in the MBS, while continuing to advance payments to the bondholders, until foreclosure or modification, for up to two years. To my knowledge the two-year limitation has always been part of the MBS rules--it's just the outside limit on how long Freddie (same for Fannie) can keep advancing on delinquent MBS loans before they have to give up and repurchase them. There have never been many loans that are seriously delinquent for two years without ever getting to foreclosure or workout, but in the strange cases (probate, bizarre title problems) it can happen. In no sense is this "two year rule" about letting loans just stay delinquent with no action by the servicer or Freddie Mac, or no effect on the fair value of the guarantee obligation, for two years. It absolutely does not mean that no credit losses are taken until a loan is "two years late." The two years refers to how long a delinquent loan can stay in the MBS, not how many months past-due it can be before it is impaired.

Why would Freddie elect to repurchase a loan when it doesn't have to? Well, if the cost of capital is cheap, but the interest payments you have to advance to the bondholders are not, it generally makes sense to repurchase the loan. The loan balance then comes out of the "guaranteed portfolio" and into the "retained portfolio." The write-down of the asset occurs immediately, given that the purchase price of the loan was par (100% of unpaid principal balance) but the fair value of a seriously delinquent loan is less than par. So a loss is immediately recognized by the retained portfolio. On the other side of the books, the guarantee asset and obligation are adjusted to reflect the fact that this loan is no longer earning a guarantee fee or reflecting guarantee costs. Any final loss taken on the loan in foreclosure is taken on the retained portfolio side, not the guarantee side.

On the other hand, if the cost of capital--Freddie's borrowing cost, including capital reserve requirements--is expensive, but the interest payments to be advanced to bondholders are relatively cheap, then you leave the loans in the MBS unless and until you are obligated to buy them out, which would be when they are delinquent and they are modified, foreclosed, or hit that two-year limit. If the loans stay in the MBS, they rack up those costs that go into the guarantee obligation, but they do not result in a recognized loss to the retained portfolio because they are not in the retained portfolio.

Now, go back and reread the Morgenson/Duhigg version of this and see if it strikes you as a reasonable paraphrase. As you do this, ask yourself if you've read anything lately in the news about Freddie needing to increase its capital reserves and facing much higher borrowing costs than it had previously. Then ask yourself if this all might be about not "injecting capital" into "pools of securities."

Of course this election not to buy out every seriously delinquent MBS loan means that fewer losses have to be recognized in the retained portfolio. The whole damned idea is to keep these loans in the guaranteed portfolio instead of the retained portfolio. However, it certainly doesn't keep Freddie from having to pay interest to bondholders every month, whether paid by the borrower or not. It still has a major effect on credit losses. It simply keeps the loans' principal balance "financed" by MBS bondholders instead of by Freddie Mac.

Has that been a wise move by Freddie? Well, I don't know we could answer that question in Morgenson/Duhigg terms, since they seem to think that the only "losses" that can be taken are in the retained portfolio. You would have to analyze the effect of the interest advances to the guarantee side of the books to see if this was a smart move or not. But of course Morgenson and Duhigg have no intention of doing that--I suspect they fail to grasp how one might do that--because to do so interferes with the narrative of "cooking the books" that they're peddling.

The interesting question that will arise, of course, is what will happen to this repurchase policy post-conservatorship. Will the government order Freddie to start buying out every delinquent MBS loan at 90 days down--knowing that the government might have to provide the capital for them to do that--in order to book retained portfolio losses "promptly," or will it perhaps decide to let the bondholders continue to finance these loans, just as Freddie has done? I'm really looking forward to finding out, myself.

At any rate, if one more person starts bringing up this canard about "no losses until the loan is two years past due" in the comments, those of us who are actually paying attention are going to jump your case for--wittingly or not--spreading stupid. You gotta stop believing everything you read in the paper.

I am the kind of person who wants to read a long post about financial accounting issues.

by Tanta on 7/01/2008 10:02:00 AM

New-Old meme: FAS 157 is ruining the financial industry. Barry Ritholtz knocks this point of view around, as reported in the New York Times:

Some blame the rapacious lenders. Others point to the deadbeat borrowers. But Stephen A. Schwarzman sees another set of culprits behind all the pain in the financial industry: the accountants.

You see, the magic of securitization during the boom was that it created obscure instruments like CDOs that were "worth" more than the underlying collateral (absurd mortgage loans). Now that the magic of securitization during the bust is that it has left behind obscure instruments--those pesky CDOs--that may well be "worth" less than the underlying collateral, if you can imagine that, foul is cried:

Of course, the purpose of FAS 157 was to make the market more transparent and efficient, which Mr. Schwarzman doesn’t take issue with.

“The concept of fair value accounting is correct and useful, but the application during periods of crisis is problematic,” he said. “It’s another one of those unintended consequences of making a rule that’s supposed to be good that turns out the other way.”

In other words, mark-to-market is great on the way up, but it's not fair to have to mark on the way down.

by Tanta on 5/20/2008 08:53:00 AM

Last week--I think it was last week--CR asked me at one point if I were going to write anything about Freddie's financials and the FAS 157 Uproar and I remember saying that our blog colleague Accrued Interest had just that day remarked that he might well write about the subject. I therefore fervently hoped he would do so, and I could just link to it, which would save me the trouble of having to have my own opinion.

So he finally got around to it. Go read it. It's well worth your time.

by Tanta on 3/26/2008 06:47:00 PM

Apparently, the accounting firms never learn. From Vikas Bajaj in the NYT:

In a sweeping accusation against one of the country’s largest accounting firms, an investigator released a report on Wednesday that said “improper and imprudent practices” by a once high-flying mortgage company were condoned and enabled by its auditors.

KPMG, one of the Big Four accounting firms, endorsed a move by New Century Financial, a failed mortgage company, to change its accounting practices in a way that allowed the lender to report a profit, rather than a loss, at the height of the housing boom, an independent report commissioned by a division of the Justice Department concluded. . . .

The 580-page report documents how New Century lowered its reserves for loans that investors were forcing it to buy back even as such repurchases were surging. Had it not changed its accounting, the company would have reported a loss rather a profit in the second half of 2006. The company first acknowledged that its accounting was wrong in February 2007 and sought bankruptcy protection less than two months later as its lenders stopped doing business with it.

The profit was important because it allowed executives to earn bonuses and convince Wall Street that it was in fine shape financially when in fact its business was coming apart, the report contended. But the report stopped short of saying that the company “engaged in earnings management or manipulation, although its accounting irregularities almost always resulted in increased earnings. . . .

“I saw e-mails from the engaged partner saying we are at the risk of being replaced,” Mr. Missal said in a telephone interview about a KPMG partner assigned to work on the audit of New Century. “They acquiesced overly to the client which in the post-Enron era seems mind-boggling.”

I have to say I have, really, no desire to read a 580-page report on this subject. But doesn't that seem like a lot of report to find one problem--under-reserving for repurchases that doesn't rise to the level of earnings management or manipulation?

(hat tip, sk)

by Tanta on 1/19/2008 10:00:00 AM

We get a lot of questions about accounting issues these days. People are concerned about accrual accounting, particularly as it relates to Option ARMs or negative amortization and the income treatment of accrued but unpaid interest. This issue always butts up against the question of how mortgage holders reserve against losses on loans, or determine the extent to which capitalized interest is or is not ultimately collectable. We've also posted some news stories regarding the uproar over SFAS 114 treatment of modified mortgages, which really get down into the weeds in terms of mortgage accounting and which are, therefore, hard to follow without a basic understanding of general mortgage asset accounting. Finally, a number of people have been asking, as we keep seeing more and more reports of write-downs at banks and investment banks, how long this writing-down is going to go on, and how it is really calculated.

Some of you--bless your lovely hearts--may be entirely innocent of any background in accounting. You may be struggling to follow the conversation in part because you make the common civilian error of forgetting that bank accounting is "backwards." To you, a loan is a liability and a deposit account is an asset. To the bank, the loan is an asset and the deposit account is a liability. It does get more complicated than that, but it never makes any sense at all until you do get past that point.

Some of you, I know, come from the "real economy," or "widget-accounting," and you are stuggling with accounting concepts that make sense to you in terms of widget makers (or retailers), like inventory and receivables and warehouses, but become puzzling when they are applied to financial accounting. Some of you may be small business owners who work on a cash basis, not an accrual basis, so you may find financial accounting even more impenetrable.

I, who was never allowed into the accounting department unescorted am not an accountant, quite often struggle to make these things clear. Fortunately, our regular commenter Lama, who is a real accountant, has offered us a splendid GuestNerd post which walks us through the basics of accrual accounting, reserves, income, and asset valuation. I have added a few comments of my own, strictly from a banking perspective. Another of our regulars, the mighty bacon dreamz, has contributed illustrations to support Lama's text. I think you will find them enlightening.

Lama On Accrual Accounting and Reserves:

To understand how "other amortization" works, I guess you first need to know how accrual accounting works. Most individuals calculate their taxes based on cash basis accounting. You recognize income when you receive the cash (or check); you recognize expenses when you pay cash. Companies with simple cash transactions and not much equipment or inventory do not vary much if they use cash or accrual accounting (think newsstands, maybe a small consulting company).

Accrual accounting means you recognize revenue when earned, expenses when incurred. A gas station would not incur an expense when they purchase gas for resale. That station would incur the expense at the time the gas was sold. That’s because the gas’ cost was a cost to produce the sale. In the time between the purchase and subsequent sale, the company holds the gas as inventory as an asset on its balance sheet.

Another concept to keep in mind is that every asset on a balance sheet has a base and a reserve. The base asset value is the easy part. If someone borrows $100,000, you have a schedule with the $100,000 on it. The loan cost you $100,000 to make. If someone owes you the $100,000 and $5,000 accrued (unpaid) interest, now your schedule will have $105,000 as a current loan value.

Making a loan on the accrual basis means the lender is earning income based on what he is owed, not based on how much cash he receives, but how much the borrower owes. If there’s a $100,000 loan at 5%, the borrower owes $5,000 after one year. If the borrower pays $6,000, the loan balance is $99,000 with $5,000 paying interest and $1,000 paying principal. If the borrower pays $4,000, the loan balance is now $101,000, with $5,000 paying interest and $1,000 increasing the amount of the loan. That’s all there is to calculating the base asset. This is in keeping with every related principle of accrual accounting and has been done the same way since The Mortgage Pig wore short pants.

Now the reserve or fuzzy area. A reserve is the amount by which you will devalue the amount you record as the base asset. There are 3 basic ways to calculate a reserve. It’s possible to use any combination of 1, 2 or 3 within a portfolio. On debt instruments a company intends to hold (until maturity or involuntary termination), the reserve will be based on both Net Present Value of cash flows and collectability. The most accurate method is to specifically identify impaired assets. If you only think you’ll collect $90,000 NPV, the value is $90,000. Repeat the same for each loan. One might have a value of $0. [Tanta: this involves a credit analyst reviewing each loan at each reporting period, generally quarterly, to determine whether or to what extent it is impaired. You will find this method used on commercial portfolios, where there are fewer units of much larger loans which may not be homogenous or easily comparable to each other.]

The second method is a percent reduction of loan values based on historical experience. Say, if instruments of a certain type typically devalue by 5%, you’d apply that to the assets in the classification. Split the classified loans into as much detail as you think appropriate (there’s some detail in this area we don’t need today). [Tanta: this is the method typically used for loans classified as residential 1-4 family mortgages. In any but the tiniest portfolios, there are too many units to examine individually, and the guidelines and underlying collateral for residential 1-4 family are (supposed to be) homogenous enough that classification or grouping of loans for analytical purposes is considered sound practice as well as obviously efficient practice.]

The third type of reserve is called a general reserve. It is simply an educated guess applied to the entire portfolio. A manager might apply this on top of the other two types. This is also known as wetting your thumb and checking which way the wind is blowing. Years ago, the chairman of the SEC decried general reserves as “Cookie Jar” reserves and “Rainy Day Funds” and their use substantially diminished. Oh, in case you do some research on your own, Reserves = Allowances. [Tanta: in the context of banking, you will find loan reserves referred to as Allowances for Loan and Lease Losses or ALLL.]

The loan balance you see on the balance sheet or support schedule is the net of the base less reserves.

So, how does this affect Income? In accounting, everything is a transaction. Hence the Debits and Credits, which are just names. Your credit card company says “we’ll credit your account.” That means they are posting a credit to their Assets account “Loans." Assets are Debit accounts, so increases in Assets are Debits, decreases are Credits. Revenue is a Credit account, so increases in Revenue are Credits, decreases are Debits.

Reserves for Loans Losses is a Contra-Asset account. Contra-assets are credit accounts that piggyback off Asset accounts. That is to say, an increase to the Reserve is a decrease (credit) to the net Asset. The sister account to Loan Reserves on the Income Statement is the expense called "Provision for Credit/Loan Losses." To balance the journal entry, you post a debit to the expense, increasing it. So, in our example loan above, the base Asset is still $100,000 and the Contra-asset is $10,000, net $90,000. Expenses in the Income Statement will take the other side of the journal entry and increase (debit) for $10,000. In theory, there could be income from a reduction of the allowance account. I don’t see that happening these days. [Tanta: that would be the “cookie jar” problem: the temptation to over-reserve in very good times, which reduces current income to just “good,” in order to reduce those unnecessary reserves in future bad quarters, which would increase income in those quarters from bad to “good” (or just “acceptable,”) and therefore make the income over time appear more stable, which makes Wall Street analysts happy. It is not likely that anyone intentionally over-reserves in bad times, since it’s hard to withstand the effect that has on current income, whatever it might do for future quarters. I certainly do not believe that banks and thrifts were over-reserving during the boom, and as each quarter’s reports come out we see they are steadily increasing ALLL. This does not look like “smoothing” income to me.]

Something very likely to happen is that, in addition to principal, jingle mail senders will not be paying interest. If the bank can foresee a future default on our loan’s interest payments of $2,000, then they record a liability for $2,000 (not a reserve to an asset) and record a debit to Interest Income for $2,000 in the current year, reducing revenue.

Now, our total Net Income is reduced by $10,000 + $2,000, $12,000. Unless and until Cash is loaned or received, there is no effect on Cash.

It’s clear that there is some estimating and guessing done within the reserve calculations and bank managers are going to do all in their power to avoid restatements as none of them ever could have known (enter specific affecting crisis here). So if you want some moral outrage, don’t vent it on the accrual method, vent it against the reserves (more specifically, the people who estimate them). [Tanta: this is why I keep saying that the accounting treatment we are seeing for OAs is not "accounting games." The issue isn't the accounting rules for non-cash income; the issue is what assumptions went into estimating how much of that deferred interest is ultimately collectable.]

Lama On Asset Valuation:

First of all, most assets are required to be valued at the lower of cost or market (assuming a market exists). Usually, the more liquid an asset, the closer the market value will be to cost. Cash is the logical extreme as cost always equals market. Then you have Accounts Receivable which is usually proximate. Next, Inventory can closely trace the actual cost. That is, it should, but companies buy things they can’t sell, they redesign products they make, causing component inventory to be worth little to the company. Even items that have value to someone else might be valued at very little because there’s too much cost involved in finding a buyer and transportation. I once audited a company that had hundreds of thousands of titanium pipes valued at cost. Well, they had no use and no customer for them. The best offer they got was from the original vendor, 15% of cost. That was my number. So it goes onto Capital/Fixed Assets. Here, market value is usually not important. Heavy equipment that might make lots of money frequently has a low resale value and huge transportation costs if it was sold.

Our debt instruments are, by their nature, very liquid. If the holder is interested in selling, they should be valued at market. If you don’t like the current market price, then the instruments are not for sale . . . ok, you don’t mark them to market, you mark them to discounted cash flows. This is where the “mark to make believe” has been and is happening.

[Tanta: banks and thrifts in the mortgage business have two categories of mortgage assets: HFS (held for sale) and HTM (held to maturity). The former is “inventory” and the latter is “portfolio investment.” HFS is marked to market. As Lama says, if you aren’t marking to a market price, then apparently you aren’t really trying to sell anything here (home sellers, take notes on this part). Eventually, if you cannot (or will not) sell your HFS pipeline, you will need to transfer it to HTM (if you have the capital necessary to hold loans to maturity), and at that point you record the loan at the lower of cost or market. In this case there is an original “write-down” of the asset if current market value is less than cost. After that, further write-downs will be necessary at each reporting period, as Lama indicates above, if the assets become impaired (or more impaired than they were when you originally put them on the books). So what we are reading in the news about write-downs of mortgages and mortgage-related assets these days involve a combination of mark-to-market adjustments (for anything in “inventory” or being taken out of HFS to HTM) and impairments of assets that have deteriorated since the asset value was originally recorded. No one is allowed to take a “once and for all” write-down of a mortgage asset; ALLL is based on your best projections of realized losses in the next 12 months (adjusted each quarter). Theoretically, every loan you own is subject to further write-down each quarter if in fact your estimate of collectability continues to deteriorate.]

by Tanta on 1/14/2008 07:58:00 AM

Or, "The Revenge of SFAS 114." Or, possibly, "KPMG Can Has Accountants." Choose your own subtitle.

NEWPORT BEACH, Calif., Jan 14, 2008 /PRNewswire-FirstCall via COMTEX/ -- Downey Financial Corp. announced today changes to previously reported levels of non-performing assets. These changes pertain to non-performing asset levels since June 30, 2007.

Rick McGill, President, commented, "As previously reported, we implemented at the beginning of the third quarter of 2007 a borrower retention program to provide qualified borrowers with a cost effective means to change from an option ARM to a less costly financing alternative. We contacted borrowers whose loans were current and we offered them the opportunity to modify their loans into 5-year hybrid ARMs or ARMs with interest rates that adjust annually but do not permit negative amortization. The interest rates associated with these modifications were the same or no less than those rates afforded new borrowers but they were below the interest rates on the original loans. We initially did not consider these modifications of performing loans to be troubled debt restructurings, as the modification was only made to those borrowers who were current with their loan payments and the new interest rate was no less than those offered new borrowers. KPMG LLP, our independent registered public accounting firm, did not object to this assessment during its third quarter review."

Mr. McGill continued, "During December 2007, KPMG advised us that upon further review of the modification program, it was likely the loan modifications should be recorded as troubled debt restructurings. After reassessing our initial analysis, we determined these modified loans should be accounted for as troubled debt restructurings. This conclusion was reached because in the current interpretation of GAAP, especially in the current housing market, there is a rebuttable presumption that if the interest rate is lowered in a loan modification, the modification is deemed to be a troubled debt restructuring unless the modified loan can be proved to be at a market rate of interest based upon new underwriting, including an updated property valuation, credit report and income analysis. We did not perform these additional steps since borrowers who qualified for our retention program were current and we were trying to streamline the process for qualified borrowers to modify their loans at interest rates no less than that being offered to new borrowers. Inasmuch as we chose not to perform these additional measures, we are now required to make this reporting change and, as such, our non- performing assets will increase from what has been previously reported. While periods prior to the third quarter of 2007 are not impacted by this change, it will result in $99 million of loans being classified as non-performing at September 30, 2007."

Brian Cote, Chief Financial Officer, commented, "As required for all loans classified as troubled debt restructurings, loans modified as part of our borrower retention program must now be placed on non-accrual status but interest income will be recognized when paid. If borrowers perform pursuant to the modified loan terms for six months, the loans will be placed back on accrual status and, while still reported as troubled debt restructurings, they will no longer be classified as non-performing assets because the borrower has demonstrated an ability to perform and the interest rate was no less than those afforded new borrowers at the time of the modification."

Mr. Cote further commented, "We believe that when loans modified under our borrower retention program are current, it is relevant to distinguish them from total non-performing assets because, unlike other loans classified as non-performing assets, these loans are effectively performing at interest rates no less than those afforded new borrowers. Accordingly, when performing troubled debt restructurings are excluded from the revised ratio of non- performing assets to total assets, the revised ratio of all other non- performing assets to total assets is not materially different from that previously reported."

The take-away, for those of you unmoved by financial accounting esoterica: KPMG is now conditioned to bark every time it hears "streamlined process." That's progress.

Now we wait to see who else was using Downey's interpretation of "troubled debt restructurings."