Why the System is Vulnerable to Crisis
February 18, 2008, Revised August 27, 2009
An enormous amount of ink has been spilled on the mortgage market crisis, and I have contributed my share. Yet I am now convinced that the most important factor underlying the crisis, which has been in plain view all along, has been overlooked. It is the way in which the mortgage industry manages default risk.
There are, in fact, two systems for managing default risk. In both, the borrower pays a premium scaled to estimates of the risk of the transaction. But, while one system has worked well, the other has been a disaster.
The system that has worked well is mortgage insurance. Borrowers are required to purchase mortgage insurance if their down payment on a home purchase, or their equity in a refinance, is less than 20%. Mortgage insurance covers lender losses up to an agreed-upon coverage amount.
The 7 companies that now sell mortgage insurance place more than half of every premium dollar they collect from borrowers in reserve accounts. This is mandated by law. The well-founded premise is that mortgage losses tend to bunch during major periods of default, which occur about every 12 to 15 years. The reserves that accumulate during long periods when losses are small are available when a crunch finally comes – as it did in 2007-08.
The stocks of these companies took a hammering during this period, but their capital remained intact. All losses were paid out of reserve accounts accumulated for that very purpose. This is in sharp contrast to the rest of the system, where losses depleted enormous amounts of capital. The mortgage insurers did the job for which they were chartered.
The second system, and unfortunately the larger of the two, is to charge borrowers a risk premium in the interest rate. The risk premium can be viewed as a rate increment above that charged on a "prime" transaction, which is one that carries the lowest risk.
As borrower, property and transaction characteristics diverge from those of a prime transaction, rate increments increase. In the mainstream segment, risk premiums can run to 1.5-2%; in the Alt-A segment, characterized generally by weak documentation, they can get to 3%, sometime more; in the sub-prime segment, characterized generally by poor credit, they can reach 5% or more.
The weakness of the risk premium system is that, with a few exceptions, and in sharp contrast to the way in which the mortgage insurance system works, risk premium dollars not needed to cover current losses are realized as income by investors. They are not available to meet future losses. This makes the system extraordinarily vulnerable to a major default episode, such as the one we are in right now.
Portfolio lenders, who hold the mortgages they originate, do carry loan loss reserves, but the tax laws discourage significant contributions to these accounts. In any case, most loans are sold in the secondary market and end up as the collateral underlying mortgage-backed securities. Each individual security carries reserves, but there is no carryover from one security to another.
Every mortgage security carries "credit enhancement", which are special protections for investors. One common form of credit enhancement, called "excess spread", channels part of the risk premiums into a special reserve account which is available for meeting losses. However, at some point the funds in the account that are not needed to meet losses are paid out to investors who have purchased the right to them.
A cardinal principle of securitization is that each security must stand on its own bottom. For legal and operational reasons, reserves cannot be shifted between securities. Thus, even though the losses on securities issued during 2000-2004 were generally small, none of the funds in those reserve accounts were available to meet losses on securities issued in 2006-2007, which have been high.
A paradox of this system for pricing default risk is that interest rate risk premiums are both too large and too small. If properly reserved, the risk premiums prevailing before the crisis would have been many times larger than those now required to meet the current default crunch. Because they were not properly reserved, they are completely inadequate. Today, risk premiums are much higher than before the crisis, but without proper reserving, they will be too small to cover losses from the next bulge in defaults.
A solution exists and it does not require the dismantling of the existing system. The key is to expand the role of mortgage insurance and extend the reserving principle to the entire system. If this could be done, it would result in a sharp drop in risk premiums paid by borrowers, and a sharp drop in the vulnerability of the system to systemic crises. It could even help get us out of the current mess.
The writer wishes to acknowledge Igor Roitburg for his contributions to this article.
An enormous amount of ink has been spilled on the mortgage market crisis, and I have contributed my share. Yet I am now convinced that the most important factor underlying the crisis, which has been in plain view all along, has been overlooked. It is the way in which the mortgage industry manages default risk.
There are, in fact, two systems for managing default risk. In both, the borrower pays a premium scaled to estimates of the risk of the transaction. But, while one system has worked well, the other has been a disaster.
Mortgage Insurance
The system that has worked well is mortgage insurance. Borrowers are required to purchase mortgage insurance if their down payment on a home purchase, or their equity in a refinance, is less than 20%. Mortgage insurance covers lender losses up to an agreed-upon coverage amount.
The 7 companies that now sell mortgage insurance place more than half of every premium dollar they collect from borrowers in reserve accounts. This is mandated by law. The well-founded premise is that mortgage losses tend to bunch during major periods of default, which occur about every 12 to 15 years. The reserves that accumulate during long periods when losses are small are available when a crunch finally comes – as it did in 2007-08.
The stocks of these companies took a hammering during this period, but their capital remained intact. All losses were paid out of reserve accounts accumulated for that very purpose. This is in sharp contrast to the rest of the system, where losses depleted enormous amounts of capital. The mortgage insurers did the job for which they were chartered.
Interest Rate Risk Premiums
The second system, and unfortunately the larger of the two, is to charge borrowers a risk premium in the interest rate. The risk premium can be viewed as a rate increment above that charged on a "prime" transaction, which is one that carries the lowest risk.
As borrower, property and transaction characteristics diverge from those of a prime transaction, rate increments increase. In the mainstream segment, risk premiums can run to 1.5-2%; in the Alt-A segment, characterized generally by weak documentation, they can get to 3%, sometime more; in the sub-prime segment, characterized generally by poor credit, they can reach 5% or more.
Risk Premium System Does Not Generate Reserves
The weakness of the risk premium system is that, with a few exceptions, and in sharp contrast to the way in which the mortgage insurance system works, risk premium dollars not needed to cover current losses are realized as income by investors. They are not available to meet future losses. This makes the system extraordinarily vulnerable to a major default episode, such as the one we are in right now.
Portfolio lenders, who hold the mortgages they originate, do carry loan loss reserves, but the tax laws discourage significant contributions to these accounts. In any case, most loans are sold in the secondary market and end up as the collateral underlying mortgage-backed securities. Each individual security carries reserves, but there is no carryover from one security to another.
Every mortgage security carries "credit enhancement", which are special protections for investors. One common form of credit enhancement, called "excess spread", channels part of the risk premiums into a special reserve account which is available for meeting losses. However, at some point the funds in the account that are not needed to meet losses are paid out to investors who have purchased the right to them.
A cardinal principle of securitization is that each security must stand on its own bottom. For legal and operational reasons, reserves cannot be shifted between securities. Thus, even though the losses on securities issued during 2000-2004 were generally small, none of the funds in those reserve accounts were available to meet losses on securities issued in 2006-2007, which have been high.
Risk Premiums Are Too Large, and Also Too Small
A paradox of this system for pricing default risk is that interest rate risk premiums are both too large and too small. If properly reserved, the risk premiums prevailing before the crisis would have been many times larger than those now required to meet the current default crunch. Because they were not properly reserved, they are completely inadequate. Today, risk premiums are much higher than before the crisis, but without proper reserving, they will be too small to cover losses from the next bulge in defaults.
A solution exists and it does not require the dismantling of the existing system. The key is to expand the role of mortgage insurance and extend the reserving principle to the entire system. If this could be done, it would result in a sharp drop in risk premiums paid by borrowers, and a sharp drop in the vulnerability of the system to systemic crises. It could even help get us out of the current mess.
The writer wishes to acknowledge Igor Roitburg for his contributions to this article.