by Anonymous on 8/31/2007 08:26:00 AM

This issue of the GSE portfolio caps has been tossed around a lot lately. In the spirit of improving the quality of the discussions, I offer a simple version of what the deal is here. If you're new to UberNerdity, a primer on GSE MBS is here.

The GSEs can "provide liquidity" to the secondary mortgage market in two ways: they can buy loans outright for their own investment portfolios, which means they keep those loans, funded with their own money (raised through debt issues, generally), and earn the interest income from them, taking the credit risk as any investor does. Or, they can buy loans to pool for MBS issues. When they do that, outside investors buy the MBS, fund the loans, and earn the interest income. Because the GSEs guarantee their securities, they charge a guarantee fee to the lenders who sell them the loans, and they have the guarantee obligation left on their books while the MBS is outstanding. Occasionally, the loans that end up in the portfolio are the ones the GSEs bought out of those MBS pools in honor of their guarantee obligation.

Also, the GSEs buy two general types of loans: what we might call "inventory" and "flow." You will usually see the inventory stuff described as "bulk" purchases. That means the GSEs are buying loans, usually in a big chunk (a "bulk" deal), that the lenders already originated, using some guidelines that may not be exactly the same as what the GSEs require in their standard MBS programs. So these deals involve negotiation of pricing. But since the loans have already been originated, you negotiate over the exact pile of loans you have, not over some guidelines that might generate some unknown pile of loans in the future.

The flow business is the forward business. This is a matter of the GSEs publishing guidelines and putting out prices that allow lenders to originate new loans into a forward commitment. Clearly, for the GSEs and the lenders, this stuff is harder to price. You might see reference to "TBA" deals. That means the actual composition of a pool of loans is "to be announced." This is true "rep and warranty" business: the price is established based on the representation that the pool of loans that we end up with will follow all the published guidelines. (The "warranty" means you pay back some or all of that price if your representations were not true.)

As a general rule, the GSEs buy bulk for their portfolios and flow for their MBS programs. There isn't really a law about this, it's just the way they do business most efficiently. If you want to think about it in terms of their "liquidity" functions, they use their portfolios to liquify lender inventory, and their MBS business to liquify lender current production. Small banks and most non-bank mortgage companies don't have the capital to build up "inventory," so all of their business is flow. It's usually the big banks who accumulate "inventory," and either sell it in bulk to the GSEs or securitize it privately or sell it to some insurance company or hold it in their own portfolio, as the market and the bank's investment needs may warrant. Of course, for years now those "inventory" trades generally didn't go to the GSEs, they went into this private security market. A great deal of that ended badly.

It does not have to be that way, and in fact we saw Freddie announcing a few weeks ago that they were shifting some portfolio dollars from bulk to forward on Alt-A purchases. They were more concerned about the liquidity crunch at the level of current production than at the level of inventory. This required them to limit the forward Alt-A purchases to only those lenders from whom they have purchased bulk Alt-A deals in the past, because that offers some kind of reference for a forward commitment. It was a matter of saying, if you originate stuff that is sufficiently like the stuff you have sold us in the past, we will buy it on a forward basis at a predetermined price level. Without some kind of reference point like that, you've got a pig in a poke, and nobody rational can price such a thing in advance.

The portfolio caps for the GSEs are in place in an attempt to control their risk-taking by limiting the dollar amount of loans they can own outright. The caps do not limit what they can buy for their MBS programs: they can buy as much of that as they can 1) find investors for and 2) afford to guarantee.

The assumption has been, in many quarters, that if the GSEs are to provide substantial liquidity to the subprime or near-prime (say, refis of subprime loans that don't quite meet standard guidelines, often because the LTV is so high) markets, they would do so by portfolio purchases. These loans are not uniform and prime-quality like the usual stuff in the MBS program.

And, after all, the problem appears to be a lack of investor appetite for the stuff. If the GSEs bought it for their MBS programs, they would be able to sell the resulting securities only because they were offering that guarantee on those MBS that you don't get in the private issue market. At some level this means the GSEs would have to "price" the risk on loans that the private market has essentially said it cannot or will not price. Most of us are assuming that the GSEs would have to put a pretty steep G-fee on this stuff in order to pull that off. That, in turn, increases the interest rate on the loan, and you do get into that problem of how the new loan can be more affordable to the borrower than the old loan was at a certain point.

The options for the GSEs, then, are to buy for portfolio or buy for MBS. In terms of portfolio purchases, they have some room left to increase holdings up to their caps (they are both under their caps right now), but that's not, most people think, enough in dollar terms to really make headway with the liquidity problem. So one solution is to raise the caps so they can buy more. Another is for them to sell off some of what they have in portfolio to "make room" for these subprime or near-prime purchases.

We just looked yesterday at Freddie's Q02 report, which showed that its portfolio holdings are about 80% prime MBS and 20% "other" (subprime, Alt-A, and other higher-risk loans). As the idea is not to add more subprime or Alt-A paper to the market, you couldn't have them sell off the hinky stuff, so you'd be asking them to sell off some of the 80%.

I'm not sure I follow the logic, necessarily, of leaving the portfolio caps in place to control risk, and then forcing the agencies to rebalance these portfolios so that a higher percentage of their holdings is in the highest-risk classes. Personally, I think the only way this makes much sense is to limit such purchases to that "bulk" or "inventory" part of the business, precisely because it can be more accurately priced. And, of course, because that means originators can share some of the pain here: by "more accurately priced" I mean the GSEs can insist on a reasonable discount. They can free up dollars on lender balance sheets by taking the loans, but they don't have to do so at a profit to the lenders. We're talking liquidity injections here, not transferring bad pricing decisions from private lenders to the GSEs. You will, however, note that this means targeting the big banks and mortgage companies for "relief," not the little community banks and credit unions and so on who don't do bulk deals.

That leaves the flow or current production problem to be solved by the MBS programs, not the portfolios. This is what the GSEs mean when they talk about putting together new mortgage products for these distressed refis, for instance. These new products put a set of standard guidelines or underwriting practices on the table that lenders can originate to in current production. The struggle for everyone will be to put a price on it that makes investors, the GSEs, the lenders, and the borrowers come out ahead.

I will suggest that anyone who thinks the GSEs can do this for free is deluded. In order to fix the mess we're in, someone is going to be subsidizing something somewhere. That does not necessarily mean a direct taxpayer cash subsidy, at least not in immediate terms. But it may mean forcing the GSEs to underprice their risk, and if they do too much of that, the taxpayers will own the problem.

I don't have an answer for that, but I do suggest that those who are really freaked out over the raising of the portfolio cap issue think it through: that isn't necessarily worse than "rebalancing" the portfolios or having the GSEs issue securities at "below market" G-fees. Remember that "rebalancing" the portfolios would require the GSEs to sell off their good stuff into a market that isn't offering top dollar for anything, good, bad, or indifferent.

by Anonymous on 8/31/2007 06:57:00 AM

There's nothing like reading the New York Times' version of an anonymous White House official's version of what Bush is going to say but has not yet said about subprime lending at 6:00 a.m. to really start your day off. Oh, look, there's a white rabbit!

Seriously. Get this:

Administration officials, who asked not to be identified, briefed a handful of news organizations on the proposals to be announced by Mr. Bush at an appearance in the White House Rose Garden on Friday morning.

The main objective of the package, one senior official said, is not to affect the stock markets but to help low-income homeowners, many of them concentrated in certain neighborhoods in several distressed areas of the country, such as Ohio and Michigan.

“The primary focus is to help individuals who have an opportunity to stay in their homes to stay in their homes,” this official said. “The subprime mortgage situation is having a crushing effect on a lot of communities right now.”

Despite the assertion that affecting the markets is not the goal, one administration official said Thursday evening that concern about Wall Street’s reaction did affect the timing of the briefing. He said there was a fear that if the White House announced in the morning that Mr. Bush would be making an announcement on housing, there could be confusion as buyers and sellers of mortgage securities guessed what the announcement would be.

But secondarily, this official said, helping homeowners keep their homes and refinance or renegotiate the terms of the mortgages could have a stabilizing effect on the financial institutions that have these mortgages in their portfolios, and help them write down the value of the mortgages or sell them off at a loss.

“You can’t solve the problems in the financial markets unless you can make some progress on the retail end of it,” said this official. “This is also a step to get banks to start loaning again.”

White House flunkies require anonymity to discuss Bush's "announcement on housing"? There could be, like, reprisals if they used their names? They're like, what, bloggers?

It's better for the markets if a muddled version goes out Thursday night than if we just get the actual version during trading hours? Is this really what these people think the "efficient markets" theory means? Are we all really confident that the MBS market won't still be forced to "guess" what this means after the announcement?

And what if this version isn't, actually, muddled? What if Bush really is going to propose a mechanism for lenders to refinance loans that can be sold at a loss? And what if that "stabilizes" things because heretofore banks with lending portfolios have been unable to make refinance loans at a loss, but once the President tells them they can, they'll lend more? This could, like, totally revolutionize Econ 101.

Fortunately, I've already got my diploma, and they can't make me take Econ 101 again. I think I will just wait for the, uh, official announcement before attempting to post anything more on the subject. I wouldn't want to move the MBS market in the wrong direction or anything.

by Calculated Risk on 8/30/2007 08:45:00 PM

From Reuters: Moody's president sees unprecedented illiquidity (hat tip Viv)

The credit market is experiencing an unprecedented loss of confidence due to the lack of transparency over where exposures lie rather than underlying credit quality problems, Moody's Investors Service President Brian Clarkson said on Thursday.

"I've been in the marketplace for 20 years ... what we're experiencing is an extreme lack of confidence and lack of liquidity. I have never seen this before," Clarkson told Reuters in an interview. "A lot of it has to do with transparency: it's not clear who owns what."

As an aside, watch for layoffs on Wall Street next week:

"Everyone in New York is expecting to hear about more job cuts in September. There'll be a wave of them."

by Calculated Risk on 8/30/2007 05:30:00 PM

UPDATE: Changed 2nd graph to make the price change clearer (hopefully). Original 2nd graph is at the bottom. Nominal prices in San Diego fell a cumulative 18% in terms in the early '90s.

In an earlier post, I presented some graphs based on the new Goldman Sachs housing forecast. One of the graphs, reproduced below, showed a possible bottom for residential investment (RI) in Q4 2008.

Click on graph for larger image.

To this graph, I added a caveat:

NOTE: Please don't confuse a bottom in RI, with a bottom in housing prices. During previous housing busts, existing home prices continued to fall long after residential investment bottomed.

To clarify this statement, here is a graph showing prices vs. residential investment for the early '90s housing bust. The prices are based on the S&P/Case-Shiller price indices for the U.S. and San Diego. These are nominal prices; for real prices the time lag between the bottom for RI and existing prices would be even longer.

In the '90s housing bust, residential investment (as a percent of GDP) bottomed in Q1 1991.

The bottom for nominal U.S. house prices was in February 1994, about 3 years after the bottom for RI. In real terms (not shown) the bottom was in 1996.

The bottom for nominal San Diego prices was in March 1996, five years after the bottom in RI.

Note: Original graph.

This is the typical pattern for housing busts, and it is because prices tend to be sticky and don't adjust immediately to the market clearing price. It's important to remember that a bottom in RI will probably precede a bottom in existing home prices by a few years.

by Anonymous on 8/30/2007 03:50:00 PM

This is the slide presentation from Freddie's Q02 conference call this morning; there's some interesting data on the credit quality of the loan portfolio. I have not yet seen a link to a transcript of the call. If you want to listen to a recording, it's available here.

by Calculated Risk on 8/30/2007 01:28:00 PM

From Rex Nutting at MarketWatch: Commercial paper market still shrinking

Outstanding commercial paper in the U.S. financial system dropped sharply for a third straight week, indicating that a severe credit crunch has not eased in the market that supplies most large companies with operating funds.

Outstanding paper fell by $62.8 billion, or 3.1%, in the week ending Wednesday to $1.98 trillion, bringing the total decline in the past three weeks to $244 billion, or 11%, the Federal Reserve reported Wednesday.

Commercial paper consists of short-term promissory notes issued by corporations to raise cash for their operational needs. Most of the paper has maturities between 30 days and 270 days; anything longer than that requires a registration statement with the Securities and Exchange Commission.

An estimated $1 trillion in commercial paper will mature in the next few months. ...
...
The declines in outstanding paper have been felt strongest in the asset-backed portion of the market, which represents about half of all commercial paper. These securities are backed by assets such as credit-card receivables or mortgages. In the latest week, asset-backed paper fell by $59.4 billion, or 5.6%. In the past three weeks, this kind of paper has fallen by $184.9 billion, or 15.6%.

by Calculated Risk on 8/30/2007 12:12:00 PM

Last night, I posted the current Goldman Sachs housing forecast.

A frequently asked question is why starts are forecast to be higher than new home sales. New Home sales come from a subset of housing starts. Housing starts also include owner built units, rental apartments, and other units that would still not be included, if sold, in the New Home sales report.

The new home sales estimate reported by the Census Bureau includes only new single-family residential structures that include both the structure and the land. The Census Bureau defines single-family homes as either fully detached structures or certain attached homes with an unbroken ground-to-roof separating wall. This definition includes some condominiums (side by side units), but does not include condominium units with another unit above or below. Starts for large multi-story condominium projects are included in the housing starts report, but sales and inventory are not included in the New Home sales report.

Just remember that new home sales come from a subset of housing starts. Working through the numbers, the Goldman Sachs forecast for starts to fall to 1.1 million units SAAR, and new home sales to fall to 0.650 million units SAAR is logically consistent, and consistent with historical data.

The following two graphs show the Goldman Sachs forecast for residential investment (RI) and new home sales.

Click on graph for larger image.

The GS forecast is given as a percent change in RI. This graph converts that forecast into RI as a percent of GDP (with some added assumptions about GDP).

The GS forecast is for another significant downturn in residential investment, and their forecast takes RI as a percent of GDP close to or below 4%, similar to previous housing downturns. GS is currently forecasting RI to bottom in Q4 '08. This is similar to my current view.

NOTE: Please don't confuse a bottom in RI, with a bottom in housing prices. During previous housing busts, existing home prices continued to fall long after residential investment bottomed.

The second graph shows Goldman Sachs' new home sales forecast through 2008. GS shows sales bottoming at 650K in Q1 and Q2 2008, with a very modest increase towards the end of '08.

No wonder the home builders will be meeting with Bernanke next week. See: Hovnanian CEO says risk of recession heightened. If these forecasts are accurate, these two graphs spell doom for some home builders.

To see the Goldman Sachs forecasts for prices and existing home sales, see table at bottom of Goldman Sachs Housing Forecast.

by Anonymous on 8/30/2007 11:48:00 AM

Credit Suisse's report on the August remittance data for the ABX basket of mortgage securities (no publically available link) includes information, for the first time as far as I know, about loan modifications.

Of the 80 deals surveyed, only 42 include modification information in the remittance reports. That does not necessarily mean there were no mods in the 38 deals; it just means their deal documents do not require that this information be included in the remittance reports. That may change, as more and more servicers are providing this information voluntarily.

For the 42 deals reporting modification activity, 47 loans from 10 different deals were modified. One deal, GSA06HE3, accounted for 21 of the total 47. Of those 21, 20 involved a rate reduction and 11 involved capitalization of principal and interest (past due payments). The average rate reduction was 2.20%. (I note that 2/28 ARMs increase at least 2.00% at their first adjustment.) CS notes that 5 of these loans actually ended up with an increase in the monthly payment (because capitalization offset the rate reduction).

The remaining 9 deals reporting mods involved 4 reporting 1 mod, 1 reporting 2, 2 reporting 3, 1 reporting 5, and 1 reporting 9.

CS does not speculate on why the Goldman deal reports so many more mods than the others. I, on the other hand, speculate that it has to do with the clarity on the subject provided in the prospectus for that deal:

Each servicer will be required to act with respect to mortgage loans serviced by it that are in default, or as to which default is reasonably foreseeable, in accordance with procedures set forth in the applicable servicing agreement. These procedures may, among other things, result in (i) foreclosing on the mortgage loan, (ii) accepting the deed to the related mortgaged property in lieu of foreclosure, (iii) granting the mortgagor under the mortgage loan a modification or forbearance, which may consist of waiving, modifying or varying any term of such mortgage loan (including modifications that would change the mortgage interest rate, forgive the payment of principal or interest, or extend the final maturity date of such mortgage loan) or (iv)accepting payment from the borrower of an amount less than the principal balance of the mortgage loan in final satisfaction of the mortgage loan. These procedures are intended to maximize recoveries on a net present value basis on these mortgage loans.

The claim has been bandied about lately that the servicing agreements for these deals just don't permit mods. That is clearly false as a generalization. Some obviously do. I see nothing in this prospectus that limits workouts (any of them listed here) to some stated percent of the deal balance.

However, it is clear that some deals do have language that restricts modifications, or at least is sufficiently ambiguous that servicers are unable to process with them without investor approval of the workout. Therefore, I for one proceed with great caution in looking at reported numbers on modifications for any sort of "trend" or reflection of the broader market of RMBS outside of the ABX buckets. It is likely to be a matter of "those who can, do, and those who can't, don't," and no one to my knowledge has attempted yet to quantify precisely how many can and how many can't. Given that just over half of the ABX deals report on modifications, I'd think it reasonable to speculate that at least half do allow mods, but that is speculation in the absence of digging through all those prospectuses, and I don't actually care about all of this that much.

by Calculated Risk on 8/30/2007 10:01:00 AM

OFHEO House Price Index Shows Smallest Quarterly Increase Since 1994

U.S. home prices increased only slightly in the second quarter of 2007 according to the OFHEO House Price Index (HPI). The HPI, which is based on data from sales and refinance transactions, was 0.1 percent higher in the second quarter than in the first quarter of 2007. This is below the revised growth rate of 0.6 percent for the previous quarter and the lowest since the fourth quarter of 1994. Prices in the second quarter of 2007 were 3.2 percent higher than they were in the same quarter of 2006, the lowest annual price change since the 1996-97 period.

OFHEO’s purchase-only index, based solely on purchase price data, indicates less appreciation for U.S. houses over the past year than does the all transactions HPI. The purchase-only index increased 2.6 percent between the second quarter of 2006 and the second quarter of 2007, compared with 3.2 percent for the HPI. However, for the second quarter, the purchase-only index increase was slightly higher at 0.5 percent (seasonally-adjusted).

The figures were released today by OFHEO Director James B. Lockhart as part of OFHEO’s quarterly report analyzing housing price appreciation trends. “House prices were basically flat in the second quarter despite tightening credit policies, rising foreclosure rates, and weakening buyer sentiment,” said Lockhart. “Significant price declines appear localized in areas with weak economies or where price increases were particularly dramatic during the housing boom.”

The data in this release only include price information through June. To the extent that recent mortgage market instability may have affected housing demand and prices, those effects would be evident in OFHEO’s next HPI release.

More later.

by Anonymous on 8/30/2007 09:25:00 AM

The LAT reports: "Sub-prime borrowers not alone" (why they are still hyphenated in the LAT style manual after all these years is not clear):

It's not just sub-prime borrowers who are having trouble getting affordable home loans.

Because mortgage investors stung by growing defaults in the sub-prime sector are shunning all but the most traditional loans, creditworthy borrowers are getting hammered if they want mortgages with payment options or the "jumbo" loans used routinely in Southern California and other high-priced home markets.

If you get such a loan, you'll pay a higher rate than before. And to add insult to injury, it's taking more time for all mortgages to get approved and funded, market experts say.

Rancho Palos Verdes marketing consultant Steve Ammons discovered the new jumbo reality after he began shopping for a mortgage on a Manhattan Beach home that he and his daughter own and rent out.

Ammons and his daughter have credit scores of 750 to 760, he said Wednesday, making them prime borrowers, the most creditworthy. What's more, the house is worth an estimated $1.6 million and has a current $650,000 mortgage, so there is plenty of equity to serve as a cushion for the lender.

Ammons is looking for a 30-year, fixed-rate mortgage, but with a twist -- the option to pay only the interest on the loan during the first 10 years. Such loans have been widely used by landlords seeking to maximize their cash flows. Ammons said a jumbo loan with an interest-only option that he took out just last year on another rental property had an interest rate of 5.5%.

But the story was different this time: A loan officer at Washington Mutual Inc. quoted him a rate of 9.75%, saying that the lender "had to charge such high rates so that they could sell off the loans," said Ammons, who has since put off refinancing.

"It's rough out there," said Doug Duncan, chief economist for the Mortgage Bankers Assn. . . .

San Diego mortgage banker John Robbins, chairman of the national mortgage bankers group, said he saw hope for the jumbo loan market's recovery. But he said borrowers such as Ammons were going to "have to be patient for a month or two."

"This market is going to come back," Robbins said. "For a gentleman like this with a lot of equity in his house, just trying to stabilize his mortgage payments, I don't think help is too far away."

I do not wish to be insulting or injurious, but I think it may be a while before non-owner-occupied jumbo IO refis in California get back to that 5.50% thing.