PIIE Blog | RealTime Economic Issues Watch
The Peterson Institute for International Economics is a private, nonprofit, nonpartisan
research institution devoted to the study of international economic policy. More › ›
Subscribe to RealTime Economic Issues Watch Search
RealTime Economic Issues Watch

The Last Bullet

by | October 24th, 2011 | 02:59 pm

US policymakers are running out of options to solve our massive unemployment problem and get the economy growing again. The Administration’s jobs bill faces resistance in Congress. The best option that can be implemented without a vote of Congress is to work through the market that started this mess in the first place—housing. The Administration has made a good start by announcing that it plans to make it easier for underwater mortgagors to refinance and repair their balance sheets, but some of the details to be filled in will be important in determining the ultimate effectiveness of the program. In addition, to boost the overall economy and to maximize the benefits of mortgage refinance, the Federal Reserve should announce new large-scale purchases of agency guaranteed mortgage backed securities (MBS) with the goal of keeping the 30-year mortgage rate between 3 and 3.5 percent through the end of 2012. These purchases would have beneficial spillovers into almost all other financial markets.

The Cartridge Is Almost Empty

The Fed’s decision in September to sell short-term Treasuries and buy long-term Treasuries (known as Operation Twist) has put downward pressure on long-term interest rates. But, with the 10-year yield already down to about 2 percent, the scope for further reductions is somewhat limited. One percent is probably the effective lower bound on the 10-year Treasury yield.

The Fed currently pays banks interest at 0.25 percent on reserves held at the Fed. This rate can and should be lowered to 0.0, but such a move is small beer as the interest payments total only $4 billion per year.

Financial market participants currently expect that the Fed will keep short-term rates near zero for at least the next two years, consistent with the Fed’s announcement in August. The short-term interest rate implied in the Treasury yield curve for one to two years ahead is 0.4 percent and that for two to three years ahead is only 0.8 percent. There is little scope to push these rates lower. Any announcement about future Fed policy more than three years ahead is not likely to have any effect on financial markets because no monetary policy commitment can be credible that far ahead. In particular, it is generally agreed that sound monetary policy should aim to return the economy to trend within three years and no one believes that the interest rate will remain near zero after we have returned to full employment.

Some have proposed that the Fed announce a desired path for the future price level (or future nominal GDP) that is higher than that currently expected by the markets. It is argued that such an announcement would raise inflation expectations, lower real interest rates, and stimulate economic activity. However, it is not clear that such an announcement, by itself, would have much effect. Indeed, during the past two years, there has been little tendency for market forecasts to move toward Fed forecasts. To increase its effectiveness, any such announcement should be accompanied by concrete actions to push market conditions in a supportive direction. The best option available is a massive program of MBS purchases.

Why Housing?

There are several reasons for the Fed to focus on the market for housing finance:

  1. The market for agency MBS is one of the largest markets in which the Fed is allowed to operate. The Fed is not allowed to buy equity, real estate, or corporate debt.

  2. MBS yields are the most important factor behind mortgage rates. As investors have flocked to the perceived safety and liquidity of Treasuries, the spread between mortgage rates and yields on 10-year Treasury notes has risen considerably. Fed purchases are especially effective at reducing those interest rates whose spreads to Treasuries are wider than normal.
  3. Lower mortgage rates help to repair the balance sheets of households hurt by the collapse of the housing bubble, both by lowering the cost of debt service through refinance and by supporting purchases of houses and thus the price of housing.

To have maximum effect, the Fed should indicate a target range for mortgage rates and commit to maintaining rates in this range through December 2012. Given the weakness of the economy, there is no significant chance that the Fed would wish to reverse its easy policy stance before the end of 2012. By making such a commitment, the Fed would encourage mortgage originators to create extra capacity to handle a large volume of activity. A relatively long commitment would provide assurance to those who are considering whether to begin shopping for a new home.

The target ranges should be 3 to 3.5 percent for conforming 30-year mortgages, and 2.5 to 3 percent for conforming 15-year mortgages and 10/1 adjustable rate mortgages (ARMs). To achieve these targets for mortgage rates, the Fed should aim for current coupon 30-year MBS yields of 2.5 percent, and lower yields on MBS with shorter maturities. Because the spread of mortgage rates over MBS yields tends to jump up when yields decline, the primary mortgage rate would start close to 3.5 percent and gradually decline over the first few months of the program as the backlog of applications is run down.

How the Administration Can Help

The Federal Housing Finance Agency (FHFA) announced on October 24 that it plans to strengthen and extend the Home Affordable Refinance Program (HARP) through December 2013.1 HARP is a temporary program by the federal housing agencies, Fannie Mae and Freddie Mac, to allow borrowers whose loans are currently guaranteed by the agencies to refinance despite having loan-to-value ratios above the normal 80 percent limit. Although HARP has existed for roughly two years, only a small fraction of potentially eligible borrowers have been able to refinance (DeMarco 2011). There are several reasons for the failure of HARP to deliver on its early promise:

  1. Low public awareness combined with homeowner fears of being rejected.

  2. Fears by mortgage originators that the housing agencies may “put back” these new loans if they subsequently move into default, which leads them to apply unduly high underwriting standards.
  3. Refusal by some holders of second liens to subordinate their claims to the refinanced first mortgage.
  4. Restrictions on applying the original mortgage insurance policy to the refinanced loan, particularly if the refinanced loan has a different servicer.

The Administration plans to improve HARP in several dimensions, most notably by removing the 125 percent loan-to-value ceiling, by eliminating the need for a new appraisal in many cases, by lowering fees for borrowers, and by waiving some representations and warranties that lenders are required to make to the housing agencies.

Details on the latter point will be released by November 15, and these details will be critical for the program’s success. It is widely agreed that banks have rejected or discouraged all but the most creditworthy borrowers from taking advantage of HARP out of reluctance to take on the risk of these representations and warranties. However, as FHFA noted in the Q&A’s to its October 24 announcement, nearly all loans eligible under HARP have been seasoned for more than three years; defects in loans usually show up in the first few years of the loan, thus the remaining value of representations and warranties associated with these loans is low. A recent CBO study (Remy, Lucas, and Moore 2011) found that waiving all representation and warranty requirements in HARP would have a negligible cost.

To address the other main reason for low take-up under HARP—the lack of awareness of HARP among borrowers—the FHFA should direct the housing agencies to write to all mortgagors whose loans are eligible under HARP to inform them of their eligibility to refinance at current market rates.

According to Goodman et al (2011, 5) second liens are no longer a major barrier to refinancing. As many observers have noted, refinancing of first mortgages is beneficial for second-lien holders because it frees up cash flow that can be used to service the second lien. Should second lien holders become a significant obstacle, the Administration may need to launch a campaign of moral suasion, or jawboning, against recalcitrant financial institutions that hold second liens.

Finally, borrowers with mortgage insurance may have fewer options than others to refinance, but the proposed waiver of representations and warranties should make existing servicers more willing to refinance even these loans.2

How Big?

Remy, Lucas, and Moore (2011) estimate that a program similar to the one described above would result in $428 billion additional refinancings with annual savings to household borrowers of $7.4 billion. They estimate that the program would have a small positive effect on the net worth of the housing agencies, but that it would have a small net cost to the federal government (in present value, not annual, terms) of less than $1 billion. This cost is entirely accounted for by losses in the Fed’s portfolio of existing MBS that would be prepaid. However, they do not include any positive effect on federal tax revenues from the additional stimulus implied by the fact that the program beneficiaries (indebted households) have far higher marginal propensities to consume than the existing MBS investors, who are mainly financial institutions and foreign governments. If the program increased US GDP by as little as $1 billion per year for two or three years, the additional tax revenues would exceed the costs.

The Remy, Lucas, and Moore study assumed that the 10-year Treasury yield would be nearly 4 percent under the program, which implies a 30-year conforming mortgage rate of roughly 5.25 to 5.5 percent. The current 10-year yield of around 2 percent is far lower than that assumed by Remy, Lucas, and Moore, and a new Fed program of large-scale asset purchases would lower the 10-year yield even further, based on the effects of the Fed’s 2008-09 MBS program as documented in Gagnon et al  (2011). A Fed commitment to keep mortgage rates near 3 percent for 12 months, combined with a strengthened HARP, would greatly increase both the incentive and the opportunity for households to refinance relative to the assumptions of Remy, Lucas, and Moore (2011).3

Judging from the pattern of previous refinancing waves, a sustained decline in mortgage rates of 2 to 2.5 percentage points in combination with reform of HARP likely would cause a surge of mortgage originations equal to more than half of the existing stock of agency-backed home mortgages. That would total at least $3 trillion. The median decline in the mortgage interest rate would be more than 2 percentage points, implying an overall reduction in household interest expense of $60 billion to $80 billion per year.

Macroeconomic Effects

The annual savings to borrowers would be about 0.5 percent of GDP. Because of the long-lasting nature of these savings, the total effect on household spending would be greater than that of an equivalent but temporary tax cut.4 In addition, the availability of record-low mortgage rates for a fixed period of time likely would spur potential new home buyers into the market and boost home building and sales.

Even more important, if the Federal Reserve supported the refinancing boom by purchasing $2 trillion of new MBS, for example, the existing MBS holders would have to find another market in which to invest $2 trillion. This avalanche of money would surely push up stock prices, push down bond yields, support real estate prices, and push up the value of foreign currencies. All of these financial developments would stimulate US economic activity. Based on a recent Fed study (Chung et al 2011) Fed purchases of this magnitude would increase US GDP by more than 2 percent after about two years, creating nearly 3 million additional jobs. This estimate includes only a small part of the effects operating through the mortgage refinance channel discussed above, so that the total effects on the economy would be even larger, perhaps creating 4 million extra jobs or more.


Chung, Hess, Jean-Philippe Laforte, David Reifschneider, and John Williams. 2011. Have We Underestimated the Likelihood and Severity of Zero Lower Bound Events? Working Paper 2011-01. Federal Reserve Bank of San Francisco.

DeMarco, Edward. 2011. The Conservatorships of Fannie Mae and Freddie Mac: Current and Future Operations. Speech at the American Mortgage Conference, September 19.

Federal Housing Finance Agency. 2011. FHFA, Fannie Mae and Freddie Mac Announce HARP Changes to Reach More Borrowers. News Release, October 24.

Gagnon, Joseph, Matthew Raskin, Julie Remache, and Brian Sack. 2011. The Financial Market Effects of the Federal Reserve’s Large-Scale Asset Purchases. International Journal of Central Banking 7, no. 1: 3-44.

Goodman, Laurie, Roger Ashworth, Brian Landy, and Lidan Yang. 2011. HARP: What Is Likely to Be Done to Make It More Effective. Amherst Mortgage Insight, New York: Amherst Securities Group LP, August 31.

Remy, Mitchell, Deborah Lucas, and Damien Moore. 2011. An Evaluation of Large-Scale Mortgage Refinancing Programs. CBO Working Paper Series 2011-4, Washington: Congressional Budget Office.


1. Another program, the Home Affordable Modification Program (HAMP), seeks to reduce the principal owed by borrowers who are struggling to meet their monthly payments. The legal and policy issues associated with HAMP are considerably more complex than those associated with HARP. However, HARP has the potential to benefit many more borrowers than HAMP because most borrowers are not behind in their monthly payments.

2. See Goodman et al (2011) for further discussion of issues related to mortgage insurance and refinancing.

3. Remy, Lucas, and Moore also assumed the program would apply only to 30-year and 15-year fixed-rate loans in agency MBS and not to ARMs or to loans held directly by the agencies. The excluded categories have outstanding principal balances of roughly $1.5 trillion and there is no reason to exclude them from HARP.

4. There would be a small offset from reduced income of wealthy households who currently hold the MBS being refinanced. However, the marginal propensity to consume of such households is lower than that of the borrowers. Furthermore, the vast majority of MBS are held by foreign governments, the Fed, and financial institutions. Any knock-on effect on spending from these institutions is likely to be negligible.

See also related post How Much Refinancing?

Comments (0)

Leave a Comment