Breaking the Back of the Financial Crisis

December 12, 2008

A Plan to Break the Back of the Financial Crisis, And Strengthen the Housing Finance System

Jack Guttentag and Igor Roitburg

The main objective of our proposal is to break the back of the financial crisis by sharply reducing mortgage foreclosures, while liquefying a major part of the existing mortgage stock. A second purpose is to provide the foundation for a more stable housing finance system in the future.


Requirements of a Successful Program

The root source of the financial crisis is the vicious cycle of declining home prices and foreclosures. The way to break that cycle is for Government to encourage the modification of mortgage contracts in ways that enable borrowers in distress to return to good standing and stay there. The private sector by itself can’t do enough modifications, of types that will not lead to re-defaults, to make a material difference.

Our plan does the following:

Eliminates negative equity on all modified loans. Among other things, negative equity is implicated in high re-default rates on modified loans.

Provides Incentives For Servicers/Investors to Write-down Loan Balances. The Government shares part of the write-down with the investor, and back-stops payment insurance on modified loans.

Provides a Mechanism for Government to Be Repaid: Because the initial cost of the program is high, it is important that Government recover its outlays in the future when the economy has turned around and the Federal deficit has to be reduced.

Re-underwrites Modified Loans: To keep re-default rates down and provide a basis for establishing insurance premiums, modified loans should be re-underwritten.

Provides Improved Program Administration: To eliminate logjams in processing modifications, a major part of the workload should be shifted from understaffed mortgage servicers to mortgage insurers with excess capacity.

Eligibility Requirements For Loan Modifications

To be eligible for the program, borrowers must occupy their home as their principal residence, and their mortgage must be in default. The default requirement is very unfair to borrowers who exercised prudence and foresight in selecting a mortgage they could manage, and it is unfair to borrowers who have managed to stay current on their payments at substantial pain and cost. It also places the Government in the uncomfortable position of encouraging borrowers to default so that they can take advantage of the program. And, of course, innocent taxpayers get stuck with the bill. 

These are unfortunate but unavoidable by-products of an effective program to curb foreclosures, which is necessary to stop the decline in home prices, which is necessary to terminate the financial crisis. Inequities stemming from the crisis and associated recession, which include widespread unemployment and enormous losses of wealth, swamp the inequities stemming from programs designed to break the crisis. 

Need to Eliminate Negative Equity on Modified Loans

Negative equity is when the loan amount exceeds the value of the property; it is what borrowers term being "upside down." It needs to be eliminated on modified loans to avoid a high incidence of re-defaults, to make the loans insurable, and to restore mobility to upside-down borrowers.

There is a lot of wishful thinking that so long as borrowers can afford the monthly mortgage payment, they will continue to pay, disregarding their negative equity. That may have been true historically when negative equity was unusual and when it did arise, it was small. There is mounting evidence, however, that substantial negative equity – say loan balances 20% or more above property value -- causes some borrowers to default who can otherwise afford their payments.

Further, negative equity has enormous social costs above and beyond the impact on foreclosures. Borrowers with negative equity have no mobility; they can’t move to where the jobs are without defaulting on their mortgage. Members of the armed forces are a particular source of concern -- if they have negative equity and are ordered to a new base, they are forced to default.

Borrowers with negative equity also are likely to cut their discretionary spending. In an economy entering a recession, this is bad news.

While negative equity is a millstone around the neck of the economy, getting servicing agents and the investors they represent to implement them, is a major challenge.

Reluctance of Loan Servicers to Write-Down Balances

With a few exceptions, servicing agents view balance reductions as a last resort. One reason could be that a reduction in the balance reduces servicing fees, whereas a reduction in the interest rate or extension of the term, do not. (Servicing fees are expressed as a percent of the loan balance). Our impression, however, is that this is probably less important than the burden of having to justify balance reductions to investors. Servicers are contractually bound to select the modification that is least costly to investors, and in the short-term, balance reductions are the most costly.

On loans paid off before term, which almost all are, a modification aimed at reducing the payment to some target amount costs an investor less if done through a term extension or a rate reduction rather than a balance reduction. (See note). In addition, rate reductions can be temporary, which is not possible with balance reductions.

Note: Assume the loan has a balance of $100,000 at 6% with 360 months remaining and a payment of $599.56. If an affordable payment is 20% lower at $479.65, the balance must be reduced by 20% to $80,000, or the rate must be reduced to 4.0385%. If the loan runs for 5 years, the balance collected by the investor will be $74,443 in the first case, $90,503 in the second. The cost of the rate reduction will be $16,061 less than the cost of the balance reduction. The cost difference disappears only if the loan runs to term.

Term extensions are also less costly to the investor, but on 30-year loans that are only a few years old, the payment can’t be reduced very much. Term extensions are an effective way of reducing the payment on 10 and 15-year mortgages, but these comprise a very small .share of loans in distress.

Balance reductions do have one major advantage for investors: they reduce negative equity, which reduces the re-default rate on loans that are modified. A balance reduction can be justified to the investor, even though the required rate reduction would have a lower initial cost, if the balance reduction had a materially lower probability of re-default.

Until recently, data were not available to support that claim, but that is gradually changing. Some of this evidence is reported by Kate Berry in National Mortgage News, December 1, 2008. For example, she cites data provided by Rod Dubitsky of Credit Suisse, who found that the re-default rate on modifications that reduced the loan balance was about half that of other modifications.

 This, along with the special inducements to reduce balances described below, should be more than enough to make balance reductions a no-brainer for investors.

Restoring Confidence and Marketability

A major feature of the financial crisis is the general loss of confidence in the quality of financial assets other than those guaranteed by the Federal Government. With the loss of confidence has come the loss of ascertainable values and marketability. This process began in the home mortgage market, and its reversal ought to begin there as well.

Under our program, modified loans will acquire an ascertainable value and become marketable by eliminating the risk of re-default to the investor. To accomplish this, modified loans: a) Will be written down to 90% of the current market value of the property; b) Will be re-underwritten by a private mortgage insurer (PMI); and c) will carry payment insurance issued by the PMI.

Government will support the process in two ways. First, it will make a contribution to the balance write-downs equal to 10% of the current market value of the properties (See note). A significant part (if not all) of the write-down cost will be recoverable from borrowers in the future, as discussed below.

Note: For example, if the loan balance is $100,000 and current property value is $80,000, the balance must be written down to 90% of $80,000, or $72,000, of which government will contribute 10% of $80,000 or $8,000.  

If a loan has a second lien, the Government advance will cover 12%, with the investor responsible for removing the second lien. If a loan has an existing mortgage insurance policy, the investor and the PMI will negotiate the share of the write-down assumed by each.

Second, Government will provide backup reinsurance on the modified mortgages insured by PMIs, as described below.

Mortgage Payment Insurance on Modified Loans

The mortgage payment insurance proposed here (MPI) differs from traditional mortgage insurance (TMI) in protecting against cash flow interruptions as well as losses from foreclosure. Under MPI, a private mortgage insurer (PMI) guarantees that investors will continue to receive scheduled payments of interest and principal on time following a borrower default. If the default is not corrected, payments from the insurer continue until the foreclosure process is completed. At that point the investor is reimbursed for the unpaid balance plus foreclosure cost.

The MPI proposed here has 100% coverage, and will carry full faith and credit reinsurance by the Government National Mortgage Association (GNMA), which will assume responsibility for payments in the event of a PMI failure. To encourage PMIs not to be overly restrictive in their underwriting, for three years the Government will absorb the first 10% of loss on all claims. The PMI’s liability would be 40% of the loss, after the 10% paid by the Government. Government will also take the last 50% of loss. GNMA will receive a reinsurance premium of 20 basis points, most of which should turn out to be profit.

Repayment of Government Contributions to Balance Write-Downs

We have not estimated the Government outlays that will be required to support balance write-downs, but they will be large. In a contracting economy, there is no danger of inflation from these and the other Government bail-out programs. When the economy recovers, however, the enormous amount of liquidity injected into the financial system will become an inflationary danger, which is why it is imperative that as much as possible of these outlays have a return address.

The Government’s contributions to balance write-downs will be secured by a second lien owned entirely by the Government. The lien will be non-interest-bearing for the first 5 years, after which it will accrue interest at a rate set by the Government. The lien must be repaid on sale of the property, or after 10 years if the property has not been sold. The payment will be for the amount owed, or 50% of the appreciation of the property from the value stipulated in the MPI policy, whichever is less.

Interest Rates on Modified Loans

The interest rate on modified loans will be the lower of a) the existing rate; b) the rate that will make the new payment including the insurance premium equal to the existing payment; c) the rate that will make the new payment plus taxes, insurance and other debt service 38% of the borrower’s income; and d) the 10-year Treasury constant maturity rate plus 1%.

Role of Private Mortgage Insurers

MPI offered by PMIs at market terms is a logical and extremely economical extension of traditional mortgage insurance (TMI). While MPI provides much more valuable protection to investors than TMI, in most cases it will cost insurers less, and therefore should not result in higher insurance premiums. See Appendix Note A on The Cost of MPI.

PMIs play a central role in the program, since they will be responsible for:

      *Determining the new interest rate, following the rules stipulated above.

      *Establishing actuarially-based insurance premiums to be paid by the borrower.

      *Assuring that the borrower has provided informed consent to the transaction.

      *Obtaining the appraisal upon which the new loan balance is based.

Parameter Values

The parameter values cited above are subject to modification. They are as follows:


Maximum New Balance as Percent of Current Market Value: 90%

Government Write-Down as Percent of Current Market Value: 10%; With Second Lien: 12%

Borrower Repayment:

Interest Rate on Lien (1st 5 Years): 0%

Interest Rate on Lien (After 5 Years): TBD

Percent of Appreciation: 50%

Period Until Repayment Is Due: 10 Yrs

Government’s Reinsurance Liability

First Loss Position: 10%

Period 3 Yrs

Last Loss Position 50%

Modified Interest Rate

Maximum Expense to Income Ratio: 38%

Spread Over Treasury 10-Year Rate 1%

Summary of Obligations and Benefits For Major Participants






Write down balance of loan to 90% of current market value of home

Government pays for part of write-down (10% of current market value of home)


Reduce interest rate on loan to 10-yr Treasury plus 1%

Loan returned to good standing


Guarantee of timely receipt of all mortgage payments plus repayment of principal, backed by Government


Resumption of servicing fees



Pay MPI insurance premium

Loan balance reduced to 90% of current market value of home


Give Government second lien equal to Government write-down contribution; agree to repay in future

Loan rate reduced, payments made affordable


Demonstrate ability to make reduced payments


Mortgage Insurer

Review and underwrite each loan

Receive MPI premiums


Administer MPI program

New and profitable line of business



Pay investors 10% of current market value

Receive second lien equal to write-down contribution


Reinsure payment insurance

Receive reinsurance premium



MPI on New Loans

Coincident with the program for dealing with distressed mortgages, PMIs with support from Fannie Mae and Freddie Mac may offer MPI on new purchase loans and refinances as part of a program to stimulate economic recovery and strengthen the housing finance system. This MPI would have a coverage cap acceptable to the agencies, on the order of 30-35% depending on the loan type and other factors.

Since this program will not carry any Government reinsurance, we expect that Fannie Mae and Freddie Mac will be the major purchasers in the short-term. To make the program go, the agencies must commit to purchase these loans at their lowest posted rates. This is consistent with the long-term interest of the agencies, whether they become Government agencies, are fully privatized, or revert back to their previous mixed status.

A major reason that the GSEs should support MPI is that MPI will sharply reduce the systemic vulnerability of the housing finance system (see below). Since it is now completely clear that the GSEs cannot separate their own fortunes from those of the system, they have a vital stake in how the system evolves in the future.

Further, MPI would eliminate the need for risk-based pricing by the GSEs, which creates needless controversy. All risk-based pricing would be done by PMIs. In addition, by increasing the premium income of PMIs, MPI would help them ride out the crisis, reducing the need for the agencies to cope with further downgrades and possible failures.

Strengthening the System

MPI will provide the foundation for a more stable housing finance system in the future. Insurers will allocate about half of their insurance premiums to contingency reserves, as they do now on loans carrying TMI. The higher insurance premiums charged on riskier loans automatically generate higher contingency reserves, which cannot be touched for 10 years unless there is an emergency.

When PMIs insure riskier loans at higher premiums, allocations to reserves increase automatically. This is in contrast to the current system of pricing greater risk with higher interest rates, which does not involve any reserving at all.

In addition, with MPI the risk remains with the firm that priced it. This is also in sharp contrast to the existing system, where the ultimate risk tends to be shifted to parties far removed from the point where risk decisions are made. These points are discussed further in Appendix Note B on Creating a More Stable Housing Finance System.

Appendix A: The Cost of Mortgage Payment Insurance

Investors in mortgages face two kinds of default risk. Collateral risk is the risk that the investor who forecloses on a loan and sells the property will fail to recover the unpaid balance of the loan plus the foreclosure costs. On loans with small down payments on which the collateral risk is the highest, private mortgage insurance is available to protect investors.

Investors also face cash flow risk. While they ultimately may be made whole from their collateral and mortgage insurance, until that happens a loan in default is a non-performing asset which is not generating any income and is not saleable except at substantial loss. There is no insurance now available against cash flow risk on individual mortgages.

It is natural to assume that since MPI covers both the cash flow risk and the collateral risk, the required mortgage insurance premiums would be substantially larger than those on TMI. In fact, more often than not they are smaller, and when they are larger, they are not much larger!

This astounding fact stems from two sources. The first is that insuring against cash flow risk and collateral risk in combination is incredibly efficient. All of the payments the insurer advances in its role as cash flow insurer are simply prepayments – dollar for dollar – of the ultimate amount they must pay at foreclosure in their role of collateral risk insurer. The only net loss to the insurer is the interest opportunity cost on the funds advanced, which turns out to be small.

The second reason that MPI premiums are so small is that, by assuming all the default risk instead of just a piece of it, MPI eliminates interest rate risk premiums, and lower rates reduce losses on loans that default. A lower rate means more rapid amortization and therefore a lower balance, and it also means smaller accruals of unpaid interest.

We will illustrate with an example based on wholesale price quotes covering two loans as of November 27, 2007, when the market was less unsettled than it is today. The loans were the same except for a few critical differences that made one of them prime and the other Alt-A. The features of the loans are shown in Table 1.

Table 1

Characteristics of Prime and Alt-A Loans

Loan Characteristic

Prime Loan

Alt-A Loan

Price or Value:



Loan ($) / (LTV):

$400,000 / 90%

$400,000 / 90%

TMI Coverage:



Borrower FICO:



Property Type:

Single Family

Single Family


Primary Residence


Loan Purpose:


Cash Out Refi




Loan Rate:



TMI Premium:



Note: Loan is 30-year fixed-rate, property is in California, and lock period is 30 days. Rates are wholesale at zero points as of November 21, 2007, insurance premiums are from the MGIC Rate Finder at that time. Note that rate risk and mortgage insurance premiums have both risen since the table was prepared in November, 2007. In attempting to update the table, we found that the Alt-A loan was no longer being priced, by the market or by MGIC.

The prime loan was to purchase the home as a primary residence with full documentation, whereas the Alt-A loan was to take cash-out by refinancing an investment property with no documentation. The Alt-A loan carried a rate 3.875% higher, and a mortgage insurance premium .62% higher.

We assumed the Alt-A loan went into default followed by foreclosure and calculated losses with a TMI policy. We then used the same default/foreclosure scenario to calculate the losses on an MPI policy with the interest rate reduced to 6%.

We found that the total losses were $21,111 lower with MPI, with the detail shown in Table 2 below. Both insurer and investor share in this loss reduction. The insurer’s cap (maximum loss exposure) is reduced by $4,473 while the investor’s losses are reduced by $16,638.

Table 2

Breakdown of Cost Savings on MPI at 6% Relative to TMI at 9.875%

Incremental Costs of MPI  
Payment Advances, Default to Foreclosure


Interest Cost of Payment Advances to Insurer




Cost Savings of MPI  
Interest Charges Due at Foreclosure


Larger Borrower Equity at Default


Larger Borrower Equity at Foreclosure


Interest Gained on Payment Advances by Investor




Net Saving on MPI


Note: It is assumed that the loan defaults after 24 months, it takes 12 months after default to foreclose and another 9 months after foreclosure to sell the property to a third-party purchaser, house value at disposition is 20% lower than at origination, loss on cash flow advances is calculated at 6%, and foreclosure expenses are based on those developed by HUD in Providing Alternatives to Mortgage Foreclosure: A Report to Congress, March 1996.

Our example involved a relatively large rate reduction. Smaller rate reductions generate smaller savings, as shown in Table 3 below. The costs there are calculated in the same way as in Table 2, except that the interest rate on the Alt-A loan, and therefore the rate reduction associated with MPI, takes different values.

Table 3

Loss Reductions on MPI as a Function of Interest Rate Reduction

Loss Reduction From Using MPI Rather Than TMI

Interest Rate Reduction

Total Loss Reduction

Loss Reduction to Insurer

Loss Reduction to Investor


$21,111 (12.2%)




$16,648 (9.9%)




$11,368 (6.9%)



1. 000%

$5,876 (3.7%)




$3,043 (2.0%)



0. 000%

$151 (0.0%)



If there is no interest rate reduction, MPI will cost the insurer more than TMI, but not much more, and the investor will always incur a smaller loss with MPI.

Appendix B: Creating a More Stable Housing Finance System

Extending the Reserving Principle: Borrowers today pay for default risk in two ways. The first and larger charge is to impose a risk premium in the interest rate. The risk premium is a rate increment above that charged on a "prime" transaction, which carries the lowest risk. The greater the perceived risk, the larger is the premium.

A weakness of the interest rate risk premium system is that, with few exceptions, risk premium dollars not needed to cover current losses are realized as income by investors. They are not reserved and available to meet future losses. This is a serious limitation because losses tend to bunch.

For example, interest rate risk premiums collected on loans originated in 2000 had very low losses because of the marked appreciation in house prices in subsequent years. Most of the risk premiums collected on these loans became investor income. Loans originated in 2006, in contrast, had large losses but none of the excess premiums from the 2000 vintage were available to help meet those losses.

Another weakness of the interest rate risk premium system is that premiums are based not on long-run actuarial loss experience but on the return investors require to compensate for the risk of "going broke." These are substantially higher than premiums based on actuarial experience. Yet in the absence of reserving, interest rate risk premiums are never high enough to meet the losses that occur in a crunch, such as the one we are in now.

In addition, interest rate risk-based premiums along with underwriting requirements can change markedly over short-periods with changes in market sentiment. They ease during periods of euphoria such as 2000-2005, and then sharply reverse course when sentiment changes, as in 2006-2008.

The second method of charging borrowers for default risk is to charge a mortgage insurance premium. Today, borrowers may be required to purchase mortgage insurance if their down payment on a home purchase, or their equity in a refinance, is less than 20%.

In contrast to interest rate risk premiums, more than half of the mortgage insurance premiums collected from borrowers are placed in contingency reserves that cannot be touched for 10 years except in an emergency. The reserves that accumulate during long periods when losses are small are available when a foreclosure crunch comes – as right now.

The reserving process requires mortgage insurance companies to view expected losses over a long time horizon. While premium structures change over time, such changes are based on revised estimates of losses over long periods, rather than on short-term swings in market sentiment.

The upshot is that a mortgage system in which borrower payments for risk are reserved is more stable, and the average premium paid by borrowers is much lower, than one in which borrower payments are divided between current losses and income. Unfortunately, for every risk-based dollar paid by borrowers today that is subject to reserving, they pay ten or more risk-based dollars that are not subject to reserving. Adoption of MPI would tend to move the system in the direction of greater reserving.

Eliminating the Agency Problem: In discussions of the current crisis, one feature of the existing housing finance system that has been much commented upon is that the parties making risk decisions are not the parties that end up assuming the risk. This is what economists term an "agency" problem, where one party (the agent) is supposed to act in the interest of another (the principal), even though their interests are not the same.

Various techniques have been developed to assure that the actions of the agent are consistent with the interests of the principal. For example, when loan originators sell loans, the purchaser often has the right to sell them back if they don’t meet the principal’s requirements. The problem is that these mechanisms don’t always work the way they are supposed to, and during a period of euphoria in the market, such as we experienced during 2000-20005, they may not work at all.

MPI eliminates the agency problem in connection with default risk. The PMI underwrites the loan, and the PMI assumes all or virtually all of the risk.

Want to shop for a mortgage on a level playing field?

Why Shop for a Mortgage with the Professor?

  1. Receive His Help in Finding the Type of Mortgage That Best Meets Your Needs
  2. Shop Prices Posted Directly by His Certified Lenders
  3. Shop Prices Fully Adjusted to Your Deal
  4. Shop Prices That Are Always Current
  5. Get Him as Your Ombudsman Just in Case

Read More About the Support and Protections Listed Above

Sign up with your email address to receive new article notifications