Introduction

The US economy is stuck, with only weak growth.  While the 2008 economic collapse was stopped and then partially reversed through a number of bold government programs (including TARP, the Troubled Asset Relief Program launched under Bush, and the Obama stimulus package), the economy is now growing at too slow a rate to see a significant and sustained reduction in the still high rate of unemployment anytime soon.  The economy is operating far below potential, with a consequent huge loss in what living standards could be.  And the personal human cost of high unemployment is severe in itself.

A primary reason for this continued slow growth is the badly functioning housing market.  Housing prices (see this post) built up in a bubble in the middle of the last decade, reaching a peak in early 2006, and then collapsed.  With the collapse of that bubble, losses built up in US banks and in the US financial system more broadly, leading most spectacularly to the bankruptcy of Lehman Brothers.  The TARP program as well as very aggressive actions by the US Federal Reserve Board succeeded in stabilizing the banks.  But homeowners also lost when the housing price bubble burst, with many now owing more on mortgages than the current value of the mortgaged house itself.

These mortgage holders cannot refinance at the lower interest rates now available on the market, unless they can come up with cash at the time of the refinancing to pay off the balance of the old mortgage in excess of what their new mortgage could be (now normally only 80% of the current home value).  If they do not have such cash, they must struggle to pay the mortgage at the old, higher, interest rates that were obtained when they bought their house during the bubble years (or when they may have refinanced at that time to a higher mortgage amount, or taken out a home equity line of credit on the then higher home value).  Similarly, they cannot sell their house and move to a new location (perhaps in pursuit of a new job opportunity) without bringing cash to the table at the time of closing.

Hence such homeowners remain stuck.  As a consequence, the housing market is not performing as it normally would.  To be blunt, the housing markets, and as a consequence the economy more generally, are constipated.  Economists refer to this as a balance sheet recession, as households (in this case) face financial obligations (their mortgages) in excess of the value of the assets they hold (their homes).  Households hunker down, and try to service their expensive mortgages while trying to save enough to get out of their negative net worth position.  But this can take a long time, and meanwhile the overall economy stagnates.  Japan suffered such a balance sheet recession following the bursting of its asset bubble in 1989 (although for Japan the problem was centered in the corporate sector).  It took more than a decade to recover from this, and to a degree the problem in Japan continues.

One can take a fatalistic approach and say there is not much that can be done.  The Treasury Secretary Timothy Geithner, in an interview with the Wall Street Journal published in its November 21, 2011, edition, appears to take this view.   Asked by the interviewer: “Which happens first?  The economy picks up and housing recovers, or a bottoming and slight recovery in housing helps the economy?”  Geithner responded:  “You can’t engineer a recovery in housing that can lift the broader economy.  It has to be the other way around.”

If true, this would be unfortunate, as the economy will not recover as long as housing is in difficulty.  The purpose of this note is to set out a program which, while ambitious, would be feasible, and which would help unlock those households now facing mortgages that are greater than their homes are worth, and with this unlock the housing markets and the economy more generally.  The scale of the program, as will be detailed below, would be similar in scale to TARP and related programs, which succeeded in stabilizing the banks.  There is a need now to stabilize the households who have similarly suffered from the bursting of the housing bubble, with a similar commitment.

I have labeled the proposal the Comprehensive Mortgage Refinancing Program (CMRP).  The first section below will present the basics of the program, through a simple numerical example.  The section that follows will then elaborate on some of the specifics in how it would work.  I will then present the numbers on how many mortgages would be eligible and the savings these homeowners would enjoy, and aggregate figures on the total costs.  Finally a concluding section will discuss the impact on each of the various entities that would be affected (the households, the lenders, and government), and how each would benefit from the program.  There is a shared interest by each in participating, but leadership by government will be necessary to make it happen.

CMRP in Summary

The program would be built around a government loan (not a grant) to the home owners to allow the mortgage balance to be brought down to 80% of the current estimated home value.  Specifically, all household borrowers with a mortgage balance in excess of 80% of their current home value could participate, if they choose.  It would not be compulsory.  If they do, the house would be appraised, and their existing mortgage balance would be refinanced at a 4% interest rate (approximately the current market rate for 20 or 30 year fixed rate mortgages), for 80% of the home value by the existing mortgage holders and for the remaining amount as a loan on the same terms from the government.  Should the home owner decide to sell his property, perhaps some years hence, the mortgage holders would be repaid (as long as the home is sold for more than the mortgage, which was set at 80% of the value of the home when the program was launched).  The government would be repaid half of any gain above the 80% (half in order to preserve an incentive for the home owner to try to get a good price), while the remaining amount would be treated as a personal loan on the same terms, to be repaid over time.

There are many details still to be covered, but it would be helpful first to present this with a simple numerical example.  Assume that the current value of the home is $200,000, but that the mortgage on it is $250,000.  In common usage, the homeowner is “underwater” by $50,000.  Eighty percent (80%) of the home value is $160,000.  Under CRMP, the mortgage would be refinanced with the existing mortgage holder (or holders, if there is a second lien or a home equity line) providing a new 30 year mortgage at 4% on the $160,000, while the government would provide a loan on the same terms (4%, 30 years) of $90,000.

If the house is then sold for $200,000, the $160,000 mortgage would be paid off, while the government would receive $20,000 (half the difference between the sale price and the $160,000 mortgage), with the remaining $70,000 balance on the government loan to be repaid on the same terms (30 years, 4%) as if it were now a personal loan.  The homeowners could take out the $20,000 and use it as a downpayment on a new home, or could prepay the government if they wish.

Elaboration on the Program

Some of the specifics:

1. The lender with the first lien on the home (and normally the largest single lender) would cover all the closing costs involved (including the cost of the appraisal by an independent professional firm, chosen by the government) as well as all the administrative costs involved both initially and over time.  No points would be charged on the new mortgage either.  The lenders will benefit greatly by this program, and can absorb such costs.
2. The program would only be for households where the mortgage is for their principal residence.  The program is not designed to rescue businessmen or others who speculated on a continual rise in home prices during the bubble, nor for the lenders to such speculators.
3. The program is also not designed for borrowers who cannot cover the debt service on these loans.  It is designed for those households who are servicing their debt, perhaps with difficulty but servicing it nevertheless.  They will gain as the new mortgage terms will be at 4%, versus the higher rates that they currently pay (probably normally in the 6 to 7% range, as these rates were typical during the bubble, or possibly even higher if they took out loans at low initial rates which then stepped up after a few years to higher rates).  There are, unfortunately, also households who cannot afford the homes they moved to even at a 4% rate.  Such cases need to be addressed on an individual basis, where there will be foreclosures as well as major losses to the mortgage holders who made such irresponsible loans.  Other programs exist to help in such cases, but this is not the objective of the proposed CRMP.
4. The new loans from the government ($90,000 in the example) would be for 30 years at a fixed 4% rate, with the same level payments as for a 30 year fixed rate mortgage.  But one might include an incentive to pre-pay such loans, so that they do not last for decades unless truly needed.  One might include an automatic increase in the rate by say 1% point in year 10, 1% point again in year 15, and so on.  Even with a modest 2% annual inflation in home prices on average from their current level, prices would be 22% higher in 10 years and 35% higher in 15 years.  Homeowners could refinance at that point with a regular commercial mortgage, if beneficial to them, and repay the government obligation.
5. The seniority of the creditors (i.e. the holders of the first lien, the second lien, any home equity credit lines, etc.) would be kept as they are now.  In the initial refinancing to 80% of the current home value (i.e. to the $160,000 in the example, from the $250,000 initial exposure), each lender will have a proportional reduction in their exposure.  But then if the house is sold for less than $160,000 (or whatever the current mortgage balance would be at some future date, after some period of repayment), there would be losses taken by these mortgage holders, in the order of their seniority as now.  That is, the mortgage holder with a first lien would be paid first, then those with a second lien, and so on.  The holders of these second liens and home equity lines will still benefit a great deal under this program, as the government has in effect already paid them the difference between the initial total mortgage exposure and the 80% home value ($90,000 in the example).  Plus there will not be further losses unless home prices fall by a further 20% from where they are now (as the new mortgages will be 80% of the current value).  But to the extent there are such further major losses, they will bear this.

The Overall Magnitude

An important question to address is what might be the scale of such a program, in terms of the amounts to be refinanced and what the government share of this would be.  The best data from which one can compute this is provided by CoreLogic, a private firm that provides analytical and consulting services on real estate.  They maintain a comprehensive state-by-state data base with estimates of the numbers of mortgages that are underwater, and by how much.  The figures can be worked out from numbers quoted in their most recent press release, available here.

Specifically, CoreLogic estimates that as of the third quarter of 2011, 22 million mortgage borrowers in the US have loans which are greater than 80% of their current home values.  This would define the pool of potential participants under CRMP.  Of the 22 million, CoreLogic estimates that 10.7 million face a mortgage loan greater than 100% of their current home value (i.e. are underwater), with 6.3 million of these having only a first lien on the home, while the remaining 4.4 million have a first lien as well as a second lien (or more).

For the 6.3 million underwater with only a first lien, the average mortgage balance was $222,000, and they were underwater by an average of $52,000, thus implying that their average estimated home value was $170,000.  For the 4.4 million with also a second or other liens, the average mortgage balance was $309,000, and they were underwater by an average of $84,000, implying an average estimated home value of $225,000.  I assumed that the average home value of those 11.3 million with loans between 80 and 100% of their home value, was the same as the weighted average of the homes underwater (equal to about $192,600), and that on average the mortgage balance outstanding on these homes was halfway between the 80 and 100% bounds.

From these numbers, one can calculate that the total mortgage balance outstanding in the US in excess of 100% of the underlying home value, is $699 billion.  In addition, a further $630 billion is outstanding on the mortgage amounts between 80 and 100% of the home values (including all of the 22 million homes with mortgages in excess of 80% of the home values).  Hence the total amount that the government might possibly need to lend, if there is 100% participation by all such eligible mortgage borrowers, would be $1,329 billion.  And the amounts that the lenders would need to provide (for the uniform 80% mortgages) would be $3,390 billion, down from their current exposure of $4,719 billion (where the government share makes up the difference).

These would be the maximum exposures.  However, it is doubtful that 100% of home mortgage borrowers would participate.  The reasons would be various, but would include the requirement that only mortgages on principal personal residences would be eligible.  In addition, CoreLogic noted that in its data, only 69% of the 22 million home mortgage borrowers with outstanding loans greater than 80% of their current home value, have mortgages at interest rates of 5% or more.  It would be these home owners, with high interest rate mortgages, who would gain the most from participation in the proposed program.

While it is impossible to say with any certainty how many mortgage borrowers would choose to participate (a reasonable guess might be somewhere in the 50 to 75% range), for the purposes here, I will assume that 69% do.  Therefore, the outstanding loans to be made by the government to the households would total $917 billion (69% of $1,329 billion), while the new 80% mortgages from the private lenders would total $2,339 billion (69% of $3,390 billion).

A $917 billion program from the government to benefit homeowners and unlock the housing market is of course huge.  But it is similar in scale to the potential exposure the government took on under TARP and related programs to stabilize the banking system.  TARP itself was approved for up to $700 billion, although substantially less was in the end used.  Similar US Federal Reserve Board support to AIG and to JP Morgan for the Bear Stearns purchase totaled $140 billion.  There has also been approved purchases by the US Treasury of equity in Fannie Mae and Freddie Mac of up to $400 billion.   These programs have thus totaled $1,240 billion, plus there were a number of smaller programs.

But it should also be noted that while the potential government losses totaled this $1,240 billion, the actual losses so far have been small.  The US Fed has not lost anything on its programs, including programs that provided massive liquidity support to the banks.  The current estimate of the net cost of TARP to the government is only $19 billion, mostly on programs to support housing where recovery of the funds was never anticipated.  The Government in fact made a significant profit on TARP funds lent to the banks.

Indeed, the main anticipated cost to government of these programs to stabilize the financial system is expected to come from losses in the support provided to Fannie Mae and Freddie Mac.  The Congressional Budget Office expects that these losses will total $389 billion over the next ten years.  To the extent the CRMP proposal being made here is implemented, these losses to Fannie Mae and Freddie Mac would likely be reduced.

One also needs to note that while the government would make loans to the home owners of an estimated $917 billion, these loans would be made at an interest rate of 4% initially (and then possibly bumped up by a percentage point in years 10, 15, and so on, until the loans are paid off).  But the current cost of a 10-year US Treasury bond is less than 2.0%  (indeed only 1.90% as of this writing).  Thus the US Treasury will be earning a positive spread on these loans, where one should note that all administrative expenses under this program would be covered by the primary mortgage lender.  But there will still be defaults, and it is not possible to predict with any certainty how large these will be.

Overall, however, the positive spread the government will earn on the loans that are repaid, plus the savings in terms of reduced losses by Fannie Mae and Freddie Mac, make it possible that the final net cost to government will be small, as it was on TARP.  Plus there will be the broader benefits to the economy from a program to unlock the housing markets, which will in turn lead to more tax revenue to the government.

Finally, the individual home owners will benefit from the lower interest rates on the refinanced mortgages.  While no portion of the loan is being forgiven, they will now pay at a uniform 4% rate rather than the higher rates they are paying currently.  The savings to them will depend on what their current mortgage rates are, and these will vary.  The rates will also be higher on second liens and on home equity lines than on mortgages holding a first lien, and will vary based on whether they have fixed or floating rate loans, step-up payments due, and so on.

But to illustrate, for an average mortgage outstanding of $214,400 (the weighted average in the CoreLogic data cited above), and assuming their current interest rate is 6 1/2% on a 20 year fixed rate loan, the savings would be $6,900 per year in moving to a 30 year fixed rate loan at 4%.  This is a savings of 36%, and would total $152 billion (about 1% of GDP) for all the households.  This in itself would provide a substantial boost to the economy, as much of this will likely be spent.  And for the households that are underwater, and who have second liens and/or home equity lines in addition to a first mortgage, where the average mortgage is $309,000, the savings would be $9,950 per year.

Conclusion:  The Impact on Each Party

It is important to recognize that each of the major groups involved in CRMP would benefit from its implementation:

1. The home owners who cannot now refinance their mortgage because the mortgage is greater than 80% of the current value of their home, will be able to refinance at 4%, the current market rate.  They will not only realize regular monthly savings compared to what they currently often have to struggle to pay, but they will also be able to sell their house, should they now wish, perhaps to move to a different part of the country to pursue a job opportunity.  This will also help unlock the housing market, with attendant broader benefits to all the home owners in the country.
2. Mortgage lenders would with CRMP face fewer mortgage defaults and losses from foreclosures.  And losses from foreclosures are normally much more than simply the excess of the mortgage amount over the estimated current home value, as foreclosed homes typically sell at a significant further discount, plus there are substantial legal and other costs in going through the foreclosure process.  Hence they will welcome a government program where the government provides a personal loan to cover the amount of the mortgage in excess of 80% of the current home value.  It is true that such lenders would prefer the home owners to continue to pay at the above market interest rates of perhaps 6 1/2% or so that they are locked into, but they also recognize that many such borrowers will soon choose to walk away from these mortgage commitments.
3. And while the Federal Government will take on substantial new debt to fund the loans it will make, the net cost in the end is likely to be small.  It will lend the funds at a positive spread, and while there will be costs from defaults, government will also gain from lower losses incurred by Fannie Mae and Freddie Mac.  There will also be higher tax revenues from a better functioning economy, due to a better functioning housing market and as consumer spending rises in a sustainable way.

But while a program such as CRMP makes sense, it is difficult to see in the current political environment that something of this nature will be implemented.   The country’s vision has become too narrow, with no willingness to take bold actions.  As a result, it is much more likely that one will see the slow and unsteady recovery typical of balance sheet recessions where little is done to cure the underlying structural problems.