The municipal bond insurance is a wonderful business. Even though municipalities "run on empty" financially, there is always more money to be taken by simply raising taxes. Thus, even though municipalities hardly qualify as AAA based on their financial statements, historically they have paid as AAA because of their "off balance sheet" resources.
The business of insuring their bonds as AAA has been lucrative for MBIA and Ambac as well as others. Typically, property and casualty companies underwrite insurance premiums with the recognition that losses will occur. The bond insurers, on the other hand, underwrite risk that is "zero loss" expected. But the bond insurers expanded their mandate beyond municipal underwriting into securitization underwriting and are experiencing great difficulty.
Rather than take the approach of Davis Selected Advisors (DSA) which is purchasing 5% of MBIA, Berkshire Hathaway (BRK) is creating its own municipal insurance company. In this way, Buffett does not inherit the problems of someone else's underwriting.
There seem to be two powerful lessons here. First, BRK likes to buy existing companies through their Casey Stengel "getting paid for home runs other fellows hit." So, the challenges at MBIA and Ambac do have not enough clarity for Mr. Buffett to make a purchase. Second, the municipal bond insurance business is as good as it looks.
Friday, December 28, 2007
Wednesday, December 26, 2007
That's Not Champagne Popping
The most recent information on national home prices was released today. This chart describes information through the end of October. As I have pointed out in recent posts, every percentage point drop has significant, leveraged impact on the financial system. The bursting of this bubble does not appear to be running out of air!CDO Squareds: A Cliffhanger
The stock of MBIA (MBI) dropped over 25% last week on disclosure of exposure to $8 billion of CDO squareds. Analysts contended that this was information that management had withheld, while MBI announced that the information had been disclosed on August 2, 2007. But the real question in everyone's mind was, "what's a CDO squared?" and "should I be afraid of these things?"
A CDO-squared is a type of collateralized debt obligation (CDO) where the underlying portfolio includes tranches of other CDOs. The following chart describes these:
Given the double-layer structure, CDOs-squared were perceived as having an added protection against losses, because tranches in underlying CDOs and the CDO-squared are protected by subordination at each level. This week's reaction showed that is no longer the belief.
The CDO squared structure amplifies the characteristics of a CDO. In an earlier post, I detailed how a CDO allows for a reduction of risk while enhancing return as long as there is low correlation of bad events. The CDO squared causes a further reduction of risk while enhancing returns as long as the bad events do not occur in one CDO, but are spread among the CDOs.
Because of this characteristic, CDO squareds are often referred to as being subject to the so-called "cliff" risk, a phenomenon where the tranche gets wiped out quickly once losses reach it. Given today's deteriorating environment, the marketplace clearly regards that risk as possible. Stay tuned.
A CDO-squared is a type of collateralized debt obligation (CDO) where the underlying portfolio includes tranches of other CDOs. The following chart describes these:
Given the double-layer structure, CDOs-squared were perceived as having an added protection against losses, because tranches in underlying CDOs and the CDO-squared are protected by subordination at each level. This week's reaction showed that is no longer the belief.The CDO squared structure amplifies the characteristics of a CDO. In an earlier post, I detailed how a CDO allows for a reduction of risk while enhancing return as long as there is low correlation of bad events. The CDO squared causes a further reduction of risk while enhancing returns as long as the bad events do not occur in one CDO, but are spread among the CDOs.
Because of this characteristic, CDO squareds are often referred to as being subject to the so-called "cliff" risk, a phenomenon where the tranche gets wiped out quickly once losses reach it. Given today's deteriorating environment, the marketplace clearly regards that risk as possible. Stay tuned.
Poppin' the Hood
In order to understand the recent headlines about the housing crisis, some technical knowledge is necessary. The basic storyline is the same as in the past: the public looks in the rear view mirror for the best course of action, while the financial community supplies both the rationale and the means for pursuit of that course of action. This works until it doesn't work.
Housing is a basic human cost. As interest rates declined, property values rose for a variety of reasons, transforming a typical cost into a theoretical profit. People were suddenly able to purchase themselves bigger houses (or at least a house) because it was thought to be a wise investment. Critical to facilitating this reasoning is capital-generating machinery, as in "where is everyone getting this money?" Below is a graphic depiction of the machinery:
At the top (in purple) is the originator. This is our local bank or mortgage broker. In varying ways, the originator sells the loan to a pool of mortgages and receives cash.
The diversified pool of mortgages (in yellow) becomes a series of residential mortgage-backed securities (RMBS) through the alchemy of Wall Street magic where two halves can make two wholes and a whole lotta fees. The challenge of capital is the supply-demand constraint - the supply of low risk investments are limited by the demand for low returns. But for RMBS, CDOs (Collateralized Debt Obligations) neatly evaded these constraints, creating "low risk" investments with "high returns." Here's how:
By slicing the RMBS into graded tranches based on cash flow characteristics, securities were created with elevated returns for lower apparent risk by virtue of "diversification" - the theoretical free lunch of academic business. By pooling securities together with lower rated characteristics, say 80% likelihood of payments, which are estimated to pay at different times, securities can be created which model a 90% likelihood of payment.
Through this mechanism, higher risk capital returns are theoretically available without the higher risk. An example may give clarity. If I am a "B" student, it is because I average "B" work over the course of several years. Yet to attain that "B" average, there are many times that I do "A" work, but occasionally do "C" or "D" or even "F" work. Of course, the "F" work is very infrequent, but nevertheless drags me down into the "B" range. By something like CDO machinery, I could erase my "D" and "F" grades as long as they do not happen in the same semester. By doing this, I become an "A" student without changing my study habits.
You can imagine the demand for such a mechanism; it would become school-wide. Unfortunately, such latitude affects study habits. Instead of having the fear of a low grade spoiling my "B" average, I now see that I simply do not have to get all my bad grades at the same time. So as the entire student body moves to higher grade averages, their study habits are going to hell. In this way, the machinery of tranching which lulled the markets into a lower sense of risk by assuming non-correlating payment patterns (either by region or by FICO score or by type of mortgage) also elevated the risk by loosening mortgage underwriting standards.
While the CDOs are now beset with problems because of the high correlation of problems (all "F"s in the same semester), a major challenge is at the bottom of the picture (in red) in the zone of "first loss piece." As the graph indicates, there is normally some relationship between the originator and the first loss piece.
In order to facilitate the transaction, the originator would often serve as a "liquidity provider." If the originator wished to generate a higher volume of mortgages and gain a competitive edge, the originator could serve in this capacity. The upside was enhanced prospective returns (as this is the highest return piece) and, even if lost, at least make the deal happen - with all its lucrative gains and fees.
Some of the more "conservative" originators discovered that it was possible to retain this piece and then hedge it through a "counterparty swap" as a form of insurance against loss. However, as losses increase, the problems here worsen because the counterparty swaps fail (just as insurors do at times), bringing the losses, en masse, back to the originators.
Housing is a basic human cost. As interest rates declined, property values rose for a variety of reasons, transforming a typical cost into a theoretical profit. People were suddenly able to purchase themselves bigger houses (or at least a house) because it was thought to be a wise investment. Critical to facilitating this reasoning is capital-generating machinery, as in "where is everyone getting this money?" Below is a graphic depiction of the machinery:
At the top (in purple) is the originator. This is our local bank or mortgage broker. In varying ways, the originator sells the loan to a pool of mortgages and receives cash.The diversified pool of mortgages (in yellow) becomes a series of residential mortgage-backed securities (RMBS) through the alchemy of Wall Street magic where two halves can make two wholes and a whole lotta fees. The challenge of capital is the supply-demand constraint - the supply of low risk investments are limited by the demand for low returns. But for RMBS, CDOs (Collateralized Debt Obligations) neatly evaded these constraints, creating "low risk" investments with "high returns." Here's how:
By slicing the RMBS into graded tranches based on cash flow characteristics, securities were created with elevated returns for lower apparent risk by virtue of "diversification" - the theoretical free lunch of academic business. By pooling securities together with lower rated characteristics, say 80% likelihood of payments, which are estimated to pay at different times, securities can be created which model a 90% likelihood of payment.
Through this mechanism, higher risk capital returns are theoretically available without the higher risk. An example may give clarity. If I am a "B" student, it is because I average "B" work over the course of several years. Yet to attain that "B" average, there are many times that I do "A" work, but occasionally do "C" or "D" or even "F" work. Of course, the "F" work is very infrequent, but nevertheless drags me down into the "B" range. By something like CDO machinery, I could erase my "D" and "F" grades as long as they do not happen in the same semester. By doing this, I become an "A" student without changing my study habits.
You can imagine the demand for such a mechanism; it would become school-wide. Unfortunately, such latitude affects study habits. Instead of having the fear of a low grade spoiling my "B" average, I now see that I simply do not have to get all my bad grades at the same time. So as the entire student body moves to higher grade averages, their study habits are going to hell. In this way, the machinery of tranching which lulled the markets into a lower sense of risk by assuming non-correlating payment patterns (either by region or by FICO score or by type of mortgage) also elevated the risk by loosening mortgage underwriting standards.
While the CDOs are now beset with problems because of the high correlation of problems (all "F"s in the same semester), a major challenge is at the bottom of the picture (in red) in the zone of "first loss piece." As the graph indicates, there is normally some relationship between the originator and the first loss piece.
In order to facilitate the transaction, the originator would often serve as a "liquidity provider." If the originator wished to generate a higher volume of mortgages and gain a competitive edge, the originator could serve in this capacity. The upside was enhanced prospective returns (as this is the highest return piece) and, even if lost, at least make the deal happen - with all its lucrative gains and fees.
Some of the more "conservative" originators discovered that it was possible to retain this piece and then hedge it through a "counterparty swap" as a form of insurance against loss. However, as losses increase, the problems here worsen because the counterparty swaps fail (just as insurors do at times), bringing the losses, en masse, back to the originators.
Tuesday, December 25, 2007
Vicious Cycle
An excessive demand for housing has resulted in creative financing to meet the needs of the "non-conforming" borrower. As this financing became available, housing prices appreciated. Then as housing appreciated, financing became more widely available. But this feedback loop has now moved in the other direction. The following graph depicts the popping of the housing bubble through the third quarter of 2007:
This self-reinforcing loop has become more powerful because of the structure of the mortgage market. Instead of holding mortgages on their balance sheets, banks have sold mortgages to pools which securitized the mortgages, creating bonds with the mortgages serving as collateral. These bonds are called mortgage-backed securities (MBS).
MBS are structured so that "tranches" are set up which receive varying cash flows. The tranches with the first cash flows have the highest rating, but the lowest stated returns. Many investors purchased lower rated tranches in search of higher yields. Their purchase was based on the belief that real estate values might flatten out, but would not decline. The following graph depicts the relationship of housing prices to the prices of different tranches. Any decline and investment results for lower rated tranches get ugly quickly.
These graphs also demonstrate why local banks are not crying loudly about the current environment, but the large sophisticated investment banks are. The banks have moved these housing loans off their balance sheet into pools managed by major Wall Street houses. If these declines continue, the erosion of lower tranches will worsen.
This self-reinforcing loop has become more powerful because of the structure of the mortgage market. Instead of holding mortgages on their balance sheets, banks have sold mortgages to pools which securitized the mortgages, creating bonds with the mortgages serving as collateral. These bonds are called mortgage-backed securities (MBS).
MBS are structured so that "tranches" are set up which receive varying cash flows. The tranches with the first cash flows have the highest rating, but the lowest stated returns. Many investors purchased lower rated tranches in search of higher yields. Their purchase was based on the belief that real estate values might flatten out, but would not decline. The following graph depicts the relationship of housing prices to the prices of different tranches. Any decline and investment results for lower rated tranches get ugly quickly.
These graphs also demonstrate why local banks are not crying loudly about the current environment, but the large sophisticated investment banks are. The banks have moved these housing loans off their balance sheet into pools managed by major Wall Street houses. If these declines continue, the erosion of lower tranches will worsen.
The Mortgaging Class of 2006
As Wall Street shudders about the mortgaging class of 2006, that is those who obtained mortgages during 2006, I thought it useful to examine this class to asertain where problems might lie.
This analysis does not necessarily describe those who purchased a home in 2006. They may have purchased homes in earlier years and simply refinanced. They have borrowed against their homes either for purchase or for better payment terms or for cash.
There were two types of loans created: conforming and nonconforming. A "conforming" loan is a mortgage that conforms to certain guidelines (such as debt-to-income ratio limits and documentation requirements) set by Fannie Mae and Freddie Mac. For the class of 2006 these limits were $ 417,000 for single homes and $ 208,500 for second mortgages.
For "nonconforming" loans, three groups of borrowers were distinguished by their FICO scores. FICO is the acronym for Fair Isaac Corporation, a publicly-traded corporation (an interesting investment idea) that created the most widely used credit score model. The credit score model is based on information gathered from a credit reporting agency (such as Equifax). A FICO score is between 300 and 850, with 60% of scores between 650 and 799. According to Fair Isaac the median score is 723 (half of scores above and below).
Non-conforming Group ------------FICO Score
"subprime," ..........................................630
"alt-A," .................................................709
"prime-Jumbo"..................................... 741
Subprime borrowers have below average scores. Alt-A borrowers have nearly average scores, but have trouble documenting their incomes. For example, business owners are generally Alt-A because their earning power is understated due to the tax structure of their businesses. Prime-Jumbo borrowers have better than average scores, but their mortgages are too large to be conforming.
In 2006, $3.0 trillion in mortgages were originated. Of this, only about half or $1.5 trillion was conforming. In past years, conforming loans have comprised a much higher percentage of loans (over 70%). The breakdown for 2006:
Non-conforming Group ------------Loans (Blns.)
"subprime" ..........................................600
"alt-A" .................................................400
"prime-Jumbo"....................................480
Thinking that home prices were rising, these purchasers were willing to pay the increased costs of being "non-conforming" to make their desired home purchases. As an earlier post indicated, the rapid increase in these three categories did not rapidly increase home ownership, but was a means of "movin' on up." Now that home prices are declining, are they going to be "movin' on out"?
This analysis does not necessarily describe those who purchased a home in 2006. They may have purchased homes in earlier years and simply refinanced. They have borrowed against their homes either for purchase or for better payment terms or for cash.
There were two types of loans created: conforming and nonconforming. A "conforming" loan is a mortgage that conforms to certain guidelines (such as debt-to-income ratio limits and documentation requirements) set by Fannie Mae and Freddie Mac. For the class of 2006 these limits were $ 417,000 for single homes and $ 208,500 for second mortgages.
For "nonconforming" loans, three groups of borrowers were distinguished by their FICO scores. FICO is the acronym for Fair Isaac Corporation, a publicly-traded corporation (an interesting investment idea) that created the most widely used credit score model. The credit score model is based on information gathered from a credit reporting agency (such as Equifax). A FICO score is between 300 and 850, with 60% of scores between 650 and 799. According to Fair Isaac the median score is 723 (half of scores above and below).
Non-conforming Group ------------FICO Score
"subprime," ..........................................630
"alt-A," .................................................709
"prime-Jumbo"..................................... 741
Subprime borrowers have below average scores. Alt-A borrowers have nearly average scores, but have trouble documenting their incomes. For example, business owners are generally Alt-A because their earning power is understated due to the tax structure of their businesses. Prime-Jumbo borrowers have better than average scores, but their mortgages are too large to be conforming.
In 2006, $3.0 trillion in mortgages were originated. Of this, only about half or $1.5 trillion was conforming. In past years, conforming loans have comprised a much higher percentage of loans (over 70%). The breakdown for 2006:
Non-conforming Group ------------Loans (Blns.)
"subprime" ..........................................600
"alt-A" .................................................400
"prime-Jumbo"....................................480
Thinking that home prices were rising, these purchasers were willing to pay the increased costs of being "non-conforming" to make their desired home purchases. As an earlier post indicated, the rapid increase in these three categories did not rapidly increase home ownership, but was a means of "movin' on up." Now that home prices are declining, are they going to be "movin' on out"?
(Grand) Housing Fundamentals
Clearly pricing has changed with regard to stocks which have exposure to the housing market. I'm sure Warren Buffett spoke truly when he said that "we have usually made our best purchases when apprehensions about some macro event were at a peak." As I approach the much anticipated audits of the fourth quarter, I am reviewing the issue that's scaring everyone today: the housing market. The following graph describes home ownership percentages over the last thirty years.
With regard to ownership changes, it would seem intuitively that the "subprime" opportunity would have created a dramatic increase in home ownership rates. While the chart does show an upward movement, the increase seems slight - moving from 64% to 68%. What's the big deal? To me the most striking part is what it does not state: the change in the quality and size of the average house.
During the past five years, I have been amazed to see grand structures (sometimes called "McMansions" to describe their ubiquitousness) go up everywhere. From coast to coast, I noticed lots getting cleared of 50s and 60s ranch style homes and replaced with structures that could barely be contained by the lots. My question was always the same, "where are these people getting all that money?" This graph does not capture the real action that I saw - that it wasn't about the drama of increased ownership, but rather about the "movin' on up" nature of the housing boom.
If this is the case, then there needs to be a way to enable financial structures to support the change in quality. For example, in Europe 100-year loans are available. Domestically, a similar change occurred when the quality of cars improved. Suddenly, auto loans could be structured for six years, rather than three. If housing loans were extended beyond the traditional loans and interest rates were assisted, then perhaps these grand houses could be retained by the current owners. It may be that instead of housing finance being too creative, it hasn't been creative enough.
With regard to ownership changes, it would seem intuitively that the "subprime" opportunity would have created a dramatic increase in home ownership rates. While the chart does show an upward movement, the increase seems slight - moving from 64% to 68%. What's the big deal? To me the most striking part is what it does not state: the change in the quality and size of the average house.
During the past five years, I have been amazed to see grand structures (sometimes called "McMansions" to describe their ubiquitousness) go up everywhere. From coast to coast, I noticed lots getting cleared of 50s and 60s ranch style homes and replaced with structures that could barely be contained by the lots. My question was always the same, "where are these people getting all that money?" This graph does not capture the real action that I saw - that it wasn't about the drama of increased ownership, but rather about the "movin' on up" nature of the housing boom.
If this is the case, then there needs to be a way to enable financial structures to support the change in quality. For example, in Europe 100-year loans are available. Domestically, a similar change occurred when the quality of cars improved. Suddenly, auto loans could be structured for six years, rather than three. If housing loans were extended beyond the traditional loans and interest rates were assisted, then perhaps these grand houses could be retained by the current owners. It may be that instead of housing finance being too creative, it hasn't been creative enough.
Monday, December 24, 2007
Making it Thainfully clear
Today's announcement that Merrill Lynch (MER) is raising capital is not news. In an attempt to bolster capital levels as a result of an $8.4 billion write down of assets in the third quarter and an anticipated $8 billion write down in the fourth quarter, MER has been seeking capital infusions and the sale of "non-strategic" assets.
MER announced a $6.2 billion capital infusion: $5 billion from Temasek, an investment company owned by the government of Singapore and $1.2 billion from Davis Selected Advisors (DSA). I do not pretend to estimate the merits of foreign government investments in U.S. stocks; however, DSA's decisions are assessable. DSA has generated good results for their clients by focusing on fundamentals, not "fads," reminding me of Warren Buffett's comment that "Fear is the foe of the faddist, but the friend of the fundamentalist."
DSA committed $1.2 billion to the purchase of MER stock at a price of $48 - roughly 15% off the present trading price. This commitment represents about 1.3% of the $92 billion of AUM. This is neither a large position for DSA as DSA does make larger, more concentrated investment positions, nor is it a negligible position. 1% is, as a friend of mine says, "real money."
Fear has driven prices down significantly. MER is almost off 50% from its 2007 highs. Intrinsic value has eroded, as $16 billion of write-downs represents a significant percentage of the $40 billion book value of MER. However, the problems are known, while the exact size is not. At a price of 15% off today's values, fair values are available. Another 20% more than that and fear will be my good buddy.
MER announced a $6.2 billion capital infusion: $5 billion from Temasek, an investment company owned by the government of Singapore and $1.2 billion from Davis Selected Advisors (DSA). I do not pretend to estimate the merits of foreign government investments in U.S. stocks; however, DSA's decisions are assessable. DSA has generated good results for their clients by focusing on fundamentals, not "fads," reminding me of Warren Buffett's comment that "Fear is the foe of the faddist, but the friend of the fundamentalist."
DSA committed $1.2 billion to the purchase of MER stock at a price of $48 - roughly 15% off the present trading price. This commitment represents about 1.3% of the $92 billion of AUM. This is neither a large position for DSA as DSA does make larger, more concentrated investment positions, nor is it a negligible position. 1% is, as a friend of mine says, "real money."
Fear has driven prices down significantly. MER is almost off 50% from its 2007 highs. Intrinsic value has eroded, as $16 billion of write-downs represents a significant percentage of the $40 billion book value of MER. However, the problems are known, while the exact size is not. At a price of 15% off today's values, fair values are available. Another 20% more than that and fear will be my good buddy.
Tuesday, December 18, 2007
Realtor Business
Realtors have managed to retain the traditional 6% compensation structure, despite the typical pressure capitalism exerts on margins. An article in this week's Dallas Business Journal highlights some interesting facts. During the last eight years, licensees in Dallas (as measured by MLS subscribers) have increased by over 100%, while the volume of property sold has increased by just over 75%. These figures might indicate a saturation point that would exert pressure on commission rates in Texas. However, California and Florida have 11.2 real estate broker licensees per 1,000 people, while Texas has 4.3 (using 2004 numbers). These figures indicate that Texas is far from saturation.
My guess is that the commission structure is not challenged by increased realtor competition because the best realtors are going to set the rates. Rather, the real challenge seems to be that the commission structure will be affected by increased technology. For example, Google has just gotten street level pictures of every home and street in the Dallas - Fort Worth metroplex. It is amazing to be able to drive around my neighborhood with a 360 degree view. The implications for real estate brokerage as Google moves into more businesses could be significant.
My guess is that the commission structure is not challenged by increased realtor competition because the best realtors are going to set the rates. Rather, the real challenge seems to be that the commission structure will be affected by increased technology. For example, Google has just gotten street level pictures of every home and street in the Dallas - Fort Worth metroplex. It is amazing to be able to drive around my neighborhood with a 360 degree view. The implications for real estate brokerage as Google moves into more businesses could be significant.
Monday, December 17, 2007
Auto Manufacturing
In the 1994 annual report of Berkshire Hathaway, Warren Buffett commented, "we believe that our formula - the purchase at sensible prices of businesses that have good underlying economics and are run by honest and able people - is certain to produce reasonable success." When it comes to the auto manufacturing industry, the phrase "have good underlying economics" is the test.
Two years ago (1.25.2006), I pointed out Kerk Kerkorian's investment in GM as a good indicator for potential profits. (It did.) But I also commented on my unwillingness to follow him in, writing, "As a one to three year hold, GM stock has significant potential to appreciate. But, as a long term hold, the business is not attractive, even though there is clearly money to be made."
The news in today's WSJ confirmed my business apprehension. In an article discussing the possibility of Ford selling Jaguar and Range Rover to Tata Industries (surprising in itself), the pricing was discussed, "It isn't clear how much Tata or the other bidders would be willing to pay for the brands. Merrill Lynch & Co. estimated this year the combined sale of the Jaguar and Land Rover brands would raise $1.3 billion to $1.5 billion. Ford acquired Jaguar for $2.5 billion in 1989 and Land Rover for $2.75 billion in 2000" - an estimated $4 billion or 75% loss!
If Ford is unable to value these auto manufacturing businesses any more clearly than that, what ability could I (or most outsiders) possibly have in arriving at "sensible prices?"
Two years ago (1.25.2006), I pointed out Kerk Kerkorian's investment in GM as a good indicator for potential profits. (It did.) But I also commented on my unwillingness to follow him in, writing, "As a one to three year hold, GM stock has significant potential to appreciate. But, as a long term hold, the business is not attractive, even though there is clearly money to be made."
The news in today's WSJ confirmed my business apprehension. In an article discussing the possibility of Ford selling Jaguar and Range Rover to Tata Industries (surprising in itself), the pricing was discussed, "It isn't clear how much Tata or the other bidders would be willing to pay for the brands. Merrill Lynch & Co. estimated this year the combined sale of the Jaguar and Land Rover brands would raise $1.3 billion to $1.5 billion. Ford acquired Jaguar for $2.5 billion in 1989 and Land Rover for $2.75 billion in 2000" - an estimated $4 billion or 75% loss!
If Ford is unable to value these auto manufacturing businesses any more clearly than that, what ability could I (or most outsiders) possibly have in arriving at "sensible prices?"
Subscribe to:
Posts (Atom)