It is a pleasure to join you today for the annual meeting of the New Jersey Bankers Association.
In a recent speech1 I talked about a number of the challenges facing the economy as we seek to secure the recovery and build for the future. Today I will focus on one of these: the problems in the U.S. housing market. In focusing today on housing, I would emphasize that this is only one factor behind the frustratingly slow economic recovery. Nevertheless, it is an important one that deserves our attention. As always, what I have to say today reflects my own views and not necessarily those of the Federal Open Market Committee (FOMC) or the Federal Reserve System.
The New York Fed is deeply committed to contributing to efforts to resolve the housing crisis that still afflicts our nation. Our economists monitor the housing market and analyze its impact on the national economy. My outreach staff work with community groups and housing practitioners to support local programs that aid distressed homeowners. Our lawyers perform pro-bono work for homeowners facing foreclosure and advise on legal reforms, while our researchers and market analysts have developed proposals to mitigate current problems and improve the future structure of housing finance.
In my presentation today, I will explain why a more robust housing market matters for the wider economy. I will also highlight those factors specific to the housing market that create a risk that it will linger in an extremely weak state for longer than is necessary and remain vulnerable to further weakness.
I will make the case for further housing-related policy interventions that would help stabilize home prices, improve the housing outlook and generate an earlier recovery in housing activity. I believe such actions would support growth and make monetary policy more effective in promoting the dual mandate given to the Federal Reserve by Congress—maximum sustainable employment in the context of price stability.
The Housing Bust
Given the excesses of the housing boom that took place in the early to middle part of the last decade, a substantial downward adjustment in housing was unavoidable. Since peaking in the second quarter of 2006, home prices have fallen by about 30 percent,2 returning to levels that last prevailed in mid-2003. In the states of Arizona, California, Florida and Nevada, where home prices rose much more than the national average, prices have declined by roughly 40 percent. Over the past year, house prices have declined by another 4 percent, though they remained roughly flat, when one excludes distress sales.
Home construction collapsed with new single family starts falling from a peak rate of 1.75 million homes per year to a low of 360,000 in early 2009. After a very slight pick-up, they have been bumping along at about a 425,000 annual rate for over a year.
Although the number of mortgages becoming delinquent is off its peak, it remains very elevated and increased in the third quarter of 2011.3 According to the Mortgage Bankers Association of America, at the end of that quarter there were roughly 1.5 million first lien mortgages 90 or more days past due, and 2 million first lien mortgages in some stage of foreclosure.
Over time, without an improvement in housing and the labor market, more loans will become seriously delinquent. Moreover, absent more aggressive efforts to find economically efficient alternatives to foreclosure — loans already seriously delinquent and in foreclosure will become real estate owned by the lender, or REO. My staff estimate that the flow of properties into lender REO in 2012 and 2013 could be as high as 1.8 million per year, up from around 1.1 million in 2011 and around 600,000 in 2010.
The overhang of properties in the foreclosure/REO pipeline continues to exert downward pressure on house prices and housing activity. Recent surveys by MacroMarkets and by Fannie Mae indicate that households expect negligible gains in home prices over the next few years, with many expecting that prices will continue to decline.
The ongoing weakness in housing has made it more difficult to achieve a vigorous economic recovery. Housing has inhibited economic activity through a number of channels.
First, the strong rebound in residential investment that typically powers economic recoveries following deep recessions has been absent. For example, nine quarters following the ends of the mid-1970s and early 1980s recessions, residential investment had climbed by 65 percent and 55 percent, respectively, above the level at the corresponding troughs. In contrast, during this cycle, there has been no rebound. In the third quarter of 2011, residential investment was still somewhat below the level in the second quarter of 2009.
Second, by eroding household wealth, the decline in housing prices has contributed to greater weakness in consumption. For most American households the family home is by far the largest tangible asset, a buffer against misfortune, and often a significant savings vehicle to fund retirement.4
Since home values peaked in 2006, homeowners have lost more than half their home equity—about $7.3 trillion—and expectations of future gains have also declined. At present roughly 11 million households are in negative equity with the aggregate amount of negative equity estimated to be roughly $700 billion.5
Third, the weakness in home prices has reduced credit availability. Because the home is the primary source of collateral for most households, financial institutions are less willing to lend when house prices have declined substantially and could fall further. In addition, problems in the foreclosure process and obstacles to efficient modification of loans in securitizations, which I'll describe in a moment, make a house of a given value less attractive as collateral to secure a loan. As a result, as home prices have fallen, financial institutions have tightened lending standards. This has reduced the amount of credit available for households, including for small business formation.
Fourth, the big drop in house prices has made it more difficult for borrowers to refinance. This has undercut to a degree the ability of monetary policy to support demand.
Traditionally, refinancing activity has been an important channel through which lower interest rates support spending and employment. Refinancing frees up borrower income. Because those who refinance tend to have a higher marginal propensity to consume from income than the investors receiving the mortgage payments, such refinancing activity increases overall demand. However, the fall in home equity combined with a pro-cyclical tightening of refinancing standards and fees means that many borrowers who would like to refinance cannot do so.
Problems in Housing
By many measures—including affordability ratios and price-to-rent ratios—home prices have fallen back to or below their pre-boom levels. Indeed, housing no longer appears overvalued. In many former boom localities, the price of existing homes has fallen below the cost of building a new home.6 Also, after several years of depressed construction the national stock of housing no longer looks excessive relative to the nation's population.
Nevertheless, housing currently remains in excess supply, in part, because household formation rates remain depressed.
Normally, when there is excess supply, markets self-correct fairly efficiently. Prices fall until they reach a level at which supply and demand are balanced. Unfortunately, the housing market has a number of specific features that complicate this adjustment process and mean it could undershoot — and remain weaker for longer than is necessary.
Housing is a very long-lived asset. Once new building ceases in stressed markets, further reductions in demand cannot be matched by further supply reductions.7 This means that the market clearing price for housing is highly sensitive to those factors that influence demand.
Moreover, there is a feedback loop from prices to demand, with weaker prices leading to weaker demand. If house prices are anticipated to increase, then the expected capital gain helps to offset the financing cost of owning a home. However, if households believe that house prices could decline further, the expected capital loss adds to the financing cost of owning a home and prospective buyers will be reluctant to buy even if mortgage rates are low.
That said, it should also be recognized that the factors that affect demand and home prices are not immutable. They depend, in part, on policy choices. Consider, for example, the fact that most home purchases are financed with long-term mortgage credit. This means that the demand for housing is highly sensitive to the availability and cost of that credit.
Outcomes in housing are also affected by the legal and operational infrastructure of the mortgage market. This apparatus has been wholly inadequate for dealing with a systemic shock to housing and employment, and is amplifying bad outcomes. A flawed resolution regime—very different from the more efficient regime for resolving excess debt in private commercial property—exacerbates the risks of under-shooting.
Conflicts of interest are embedded in mortgage securitization structures and between first and second lien-holders. There are challenges to how the industry records and transfers interests in these mortgages. For mortgages that are part of securitizations, the legal obstacles to loan modification that preserve value are substantial. Meanwhile, mortgage servicers lack the systems, processes, governance, resources and incentives to deal effectively with a large volume of delinquent borrowers—who often suffer poor customer treatment at a time of personal distress.8
The result is inefficiently low levels of mortgage modification, too few "short sales" 9 by mutual agreement between distressed borrowers and lenders, and too much reliance on the foreclosure process. This process also has many legal impediments and has become even more inefficient as the volume of loans in the pipeline increases. One consequence has been lengthening delays in completing foreclosures—including unavoidable foreclosures. These delays have increased the economic damage caused by foreclosure and reduced recovery rates.
The damage caused by inefficient foreclosures is worth emphasizing. It is not just the direct loss of value in terms of deterioration of homes that are ultimately foreclosed. In addition, as I have personally witnessed in my outreach visits across the New York Fed's District, foreclosures impose significant costs on families, surrounding neighborhoods and communities. Unmaintained homes facing foreclosure drag down the value of nearby homes. Also, property taxes on non-foreclosed homes often go up due to a declining tax base and this puts additional downward pressure on housing prices.
Foreclosures also undercut housing demand. Households that go through foreclosure need to repair their credit records and to accumulate new downpayments. For several years they are no longer available as potential home buyers and this also contributes to softer demand and weaker housing prices.
Further declines in home prices also could be reinforcing because they would likely make the already daunting foreclosure/REO pipeline even larger. Currently, households in negative equity or "underwater" on their loans owe $65,000 on average more on their mortgages than the value of their homes. Since the recession, most defaults have been by necessity — that is, they have been prompted by a job loss, health problems or divorce coupled with the household being in negative equity. No doubt, however, there is some degree or duration of negative equity that may cause households to conclude that a default is inevitable and should be taken sooner rather than later. If this point were reached, defaults by choice could significantly swell the flow of homes into the foreclosure pipeline.
The Case for Housing Intervention
All these factors imply that there are many potential equilibrium outcomes in terms of housing demand and home prices—some considerably more desirable than others. Negative price expectations and flawed financing and administrative mechanisms, if left unaddressed, can contribute to ongoing weakness in housing demand and make it harder to generate a robust economic recovery.
Persistent weakness in housing is particularly problematic because it acts as a drag on spending and job creation in an environment in which such weakness can not be easily offset by other policy adjustments. Housing policy should seek to break adverse feedback loops, promote more economically efficient outcomes in housing and support growth.
In recent speeches,10 I have called for a "comprehensive approach" to stabilize the national real estate market and lay the foundations for recovery. I believe this should include measures to improve access to mortgage credit, reduce obstacles to refinancing, lessen the flow of homes into foreclosure through bridge financing and accelerated principal reduction, and to facilitate the absorption of REO back into use as owner- or renter- housing.
Many of these policies simply seek to solve or bypass the legal and incentive problems in the market today and mimic efficient private actions that might be taken if mortgage loans were all held in portfolio and accounting rules did not discourage net present value maximization.
Others seek to solve coordination problems and take account of the externalities or spillovers to generate outcomes that are superior from the broadest perspective of the nation as a whole.
Earlier this week Chairman Bernanke sent members of Congress a thoughtful analysis of housing policy11 by staff at the Board of Governors in Washington that evaluates a number of these same issues. What I would like to do today is to elaborate on some specific proposals we have developed in New York, in order to try to persuade you that there are workable solutions to our housing problems at costs that I deem very reasonable relative to the economically inefficient path we are on. Of course, I am open to other promising alternatives that would address the same issues.
I note also that a truly comprehensive approach would also include long-term reform —including reform of Fannie Mae and Freddie Mac—to put housing finance on a more stable footing and to equip the market to deal more effectively with any future systemic shocks.
But I will not discuss these issues in-depth today.
Access to Mortgage Credit
Restoring demand for housing requires new mortgage debt to be offered to creditworthy borrowers on reasonable terms. This is important because residential mortgage lending terms tightened significantly during the crisis, and have not eased subsequently.12
Indeed, over 50 percent of all new loans—and 70 percent of new prime conforming loans—go to borrowers with FICO scores of 760 or above, compared with about 30 percent before the crisis. Borrowers with credit scores of 720 or less now make up less than 30 percent of new originations, although they represent 52 percent of the population with credit scores.
Steps should be taken to improve households' access to credit. Currently, lenders are discouraged from making new prime conforming loans due to excessively stringent put-back rules on these mortgages. Such rules give Fannie Mae and Freddie Mac broad scope to force the originator to repurchase at par loans that go delinquent if there are any errors in the mortgage loan paperwork—even if the error has no direct link to the delinquency.
It is essential that lenders are accountable for the representations they make and are appropriately penalized for any misrepresentations. However, doing this via a blanket repurchase regime leads to bad outcomes at a time when there is an unusually high risk of job loss leading to delinquency. It would be better to establish alternative ways to penalize lenders for misrepresentation that were not in practice so closely tied to the risk of job loss.13
The current repurchase regime exposes lenders originating prime conforming loans to significant credit risk and encourages them to focus only on the lowest-risk customers.
Lenders are applying tougher restrictions than the Fannie Mae and Freddie Mac formally require in order to limit the risk that they will have to take back such loans when they become delinquent.
For mortgages guaranteed by Fannie Mae and Freddie Mac, a better balance between incentives for sound underwriting and credit availability could be achieved by inserting a materiality test in the rep and warranty agreement.
In the future, it would be worth exploring whether a better regime in general would be to have representations and warranties with a finite duration, but supplemented by a rigorous quality control check of the underwiting based on a random sample of the mortgages as they are being securitized.14
Today's tough underwriting standards and high risk-based premiums for conforming loan borrowers should also be reviewed. I do not want to see a return to the lax standards and underpricing of credit risk of the boom period. But the guarantee fees for new purchase mortgages should be based on the expected losses on these mortgages—not the realized losses on loans of earlier vintages.
There is also a strong case for tackling the downward bias on appraisals, which are voiding many transactions between willing buyers and willing sellers. Incentives for individual appraisers favor conservatism today—just as they favored over-optimism during the boom.
This problem is aggravated by the difficulty in establishing benchmark prices in markets with few transactions and a high share of distress sales. The industry could establish appraisal standards that avoid such procylical outcomes.
Increasing refinancing would support the housing market by promoting aggregate demand and employment. Refinancing creates additional cash flow for borrowers to absorb any adverse income shocks and this reduces the likelihood of default, distress sales and foreclosures.
However, there are significant obstacles to refinancing in current circumstances. Declines in home equity have been aggravated by tighter standards, high refinancing fees, burdensome administrative processes, and legal risks to the lenders refinancing the loan.
Because the taxpayer, via Fannie and Freddie, is already exposed to the risk of conforming loans defaulting, it makes no sense to make it expensive or difficult for borrowers with these loans to refinance. After all, refinancing reduces the credit risk to which taxpayers are exposed. It is also inefficient to fully re-underwrite applications for such refinancings. For example, appraisals could be greatly simplified and made less expensive by using automated valuation models.
The Home Affordable Refinance Program, or HARP, was set up to facilitate refinancing of high loan-to-value prime conforming mortgages. The announced goal of the HARP was to refinance 3 million to 4 million mortgages. However, shortcomings with the original program mean that less than a million HARP refinancings have taken place.
Legal risks discouraged lenders from participating more fully in HARP to date. Notably, if a lender refinanced a loan made by another company, it was liable for the underwriting associated with the original loan. The profit from refinancing was unattractive relative to the risk of being forced to repurchase the loan. Given this, lenders had little incentive to compete to refinance loans and this led to higher mortgage rates and less refinancing.
Many of these issues have been addressed by recent revisions to HARP and related decisions by Fannie Mae, but more could and should be done. Remaining legal issues could be alleviated by harmonizing standards between Fannie and Freddie.
I would like to see refinancing made broadly available on streamlined terms and with moderate fees to all prime conforming borrowers who are current on their payments. This could substantially increase the number of refinancings.
To stabilize the housing market, we also need to weaken the link between unemployment and new foreclosures. The primary reason for distress sales and mortgage defaults today is the loss of a job. Even if it were possible, it would be inefficient for homeowners with mortgages to self-insure against a job loss.
Some support for unemployed homeowners has been provided in some states under the Hardest Hit program. I favor a broad program to provide bridge financing for all qualified borrowers with demonstrated ability to service their debts who become unemployed involuntarily—though categorically not those who quit or are terminated with cause. In economic terms, this is a form of collective insurance.15
My staff estimates that there are 4 million to 5 million "at risk" homeowners who can afford their mortgage, but who would struggle to make their payments in the event that the primary earner in the household became unemployed.
With the current weak labor markets, they estimate around 600,000 of these at-risk households will experience an involuntary job loss that lasts longer than a month over the next year. The average annual amount that would be required to keep the mortgage current for these households while unemployed and receiving unemployment insurance is around $21,000. This would be a loan not a grant.16
This implies an annual bridge lending program of $15 billion per year during the current stress period. This seems expensive, but this is not the cost of such a program, just the cash outlay in the early years. In the future, loan repayments would help to offset the program cost.
Moreover, when one factors in the reduced negative externalities achieved by limiting distress sales, the expected program cost would be even lower.
Of course, any bridge financing loans would need to be properly secured in order to protect taxpayers interests. In situations in which the value of the loan already exceeded the value of the home, it might be appropriate to require lenders to write down the excess debt as a condition for the borrower obtaining a bridge loan. This would ensure that the bridging funds did not "bail out" the lender. Writing down excess debt would likely be in the lender's rational self-interest given the borrower's situation, since a distress sale or foreclosure at current market prices would generally yield larger losses.
Earned Principal Reduction
Investment firms that purchase delinquent mortgages routinely reduce principal in order to maximize value on these loans. It would make sense for Fannie and Freddie to do this as well in order to minimize loss of value on the delinquent loans they guarantee.
However, I am uncomfortable with the notion that "underwater" borrowers who owe more on their mortgages than their homes are worth should have to go delinquent before they have a chance of securing a reduction in their mortgage debt. I believe we should also develop a program for earned principal reduction for borrowers who are underwater but keep on making their mortgage payments. Such a program would strengthen the incentives for mortgage holders who are underwater to continue to stay current on their loans, and reduce the likely number of defaults and REO sales.
Negative equity is one of the most significant predictors of mortgage default across all types of mortgage products, including when you control for credit scores and level of income documentation. While recent underwater borrowers have not been defaulting at a rapid pace, they are also not prepaying, implying that underwater mortgages remain "at risk" for a prolonged period of time to any adverse shock to the borrower.
It is admittedly challenging to predict how underwater borrowers will ultimately perform and this performance will be sensitive to the path of home prices. Nonetheless, analysis by my staff that looks at likely borrower behavior over an extended time horizon suggests that without a significant turnaround in home prices and employment, a substantial proportion of those loans that are deeply underwater will ultimately default — absent an earned principal reduction program.
One option developed by my staff is for Fannie Mae and Freddie Mac to give underwater borrowers on loans that they have guaranteed the right to pay off the loan at below par in the future under certain circumstances, including that the borrowers have continued to make timely payments. For instance, the borrower could be given an open-ended option to pay off the loan at an LTV of 125 percent, and the right to pay off the loan at an LTV of 95 percent after three years of timely payments.
The borrower would be protected from further declines in home prices, but in return would give up a portion of any upside from future capital gains on the home via a shared appreciation agreement. Note that under this arrangement some of the reduction in the loan amount would be paid by the borrower as the outstanding balance was amortized by continued monthly payments.
Even with aggressive policies to minimize the flow of loans into foreclosure, my staff estimates that large volumes of properties would still flow into lenders' REO over the next few years. This growing overhang could continue to depress prices.
Thus, along with more incentives to promote short sales, steps should be taken to facilitate the timely and orderly absorption of these properties back into the market including as renter-occupied housing. An interagency group including the Federal Reserve is working on issues relating to REO-to-rental conversions. Appropriate policies would simultaneously lessen downward pressure on home prices and upward pressure on rents.
Among other steps, investors could be encouraged to purchase REO to be made available as rental housing. Fannie Mae and Freddie Mac could increase the number of loans offered to individual investors, provided the investor puts up an adequate amount of equity for each mortgage. And REO properties in a given locality could be bundled for sale.
The government might consider a package of tax incentives for purchases of REO that are used as rental properties. Such incentives might include a reduction of the current 27½ year depreciation period and/or a reduction of capital gains tax liability if the property is held for a minimum period, such as five years.
One idea developed by my staff—let's call it "homes for heros"—would be to create a new tax credit or other home purchase subsidy specifically for veterans of our foreign wars that would enable these veterans to purchase such properties at a discount. There are over 2½ million Gulf War II veterans alone, many of whom served multiple tours of duty overseas, and a significant proportion of them might otherwise not be able to purchase homes today.
Our nation owes them a great debt, and such a policy would provide a boost to housing demand quickly.
Some argue that any interventions to improve the real estate market would lead to moral hazard—that is these intervention would reward bad behaviors. I think these concerns are overstated. First, the programs that have been proposed can be designed with the proper incentives to limit moral hazard and to encourage desireable behavior.
Second, in contrast to the earlier phase of housing crisis when mostly subprime borrowers came under stress, many of the borrowers running into trouble with their mortgages today took out conforming loans subject to standard downpayment requirements, had respectable FICO scores and borrowed a moderate amount to finance an average house. The problem was that these purchases occurred near the peak in the market and now many of the buyers have suffered an adverse life shock such as unemployment or illness. This isn't a moral hazard issue, this is just the bad luck associated with the timing of the purchase and an exceptionally weak jobs market. Punishing such misfortune accomplishes little.
The Taxpayer Interest
I believe that the proposals I put forward are strongly in the public interest. I also believe that they would be good for taxpayers. Consider, for instance, the proposal for earned principal reduction. While this would certainly involve outlays, these would be offset by reduced defaults. Whether there is a net cost or not depends on the reduction in the default rate that would come with principal reduction and how this reduction would affect the performance of the overall housing market and the economy.
Based on recent data on borrower behavior,17 my staff calculates that the taxpayer (through the effect on Fannie Mae and Freddie Mac) would be better off with earned principal reduction under a base case of roughly flat house prices and persistent weakness in the jobs market. Under a scenario of modest house price increases, the combination of fewer defaults and shared appreciation also produces a net benefit.
This result occurs before taking into account the positive externalities of nudging house prices onto a stronger path, which would reduce the magnitude of losses on loans that do default. On an expected value basis, such a program appears still more compelling, since Fannie Mae and Freddie Mac are currently exposed to the downside risk of further declines in home prices.
Of course, just focusing on the impact on Fannie Mae and Freddie Mac is much too narrow even from a taxpayer perspective. If you believe as I do that improving the outlook for housing would materially strengthen the over economy, one must also include the fiscal benefits generated by stronger economic growth.
Housing Policy and Monetary Policy
In closing, the housing policy agenda I describe would address one factor that has impeded the economic recovery. Implementing such policies would improve the economic outlook and make monetary accommodation more effective.
However, because the outlook for unemployment is unacceptably high relative to our dual mandate and the outlook for inflation is moderate, I believe it is also appropriate to continue to evaluate whether we could provide additional accommodation in a manner that produces more benefits than costs, regardless of whether action in housing is undertaken or not.
Monetary policy and housing policy are much more complements than substitutes.
As I hope I have convinced you today, while the Fed has will do all it can to achieve our dual mandate of maximum sustainable employment in the context of price stability, we have to recognize that there is more to economic policy than just monetary policy. Low interest rates help housing, but cannot resolve the problems in that sector that are pressing on wider economic activity. With additional housing policy interventions, we could achieve a better set of economic outcomes than with just monetary policy alone.
Thank you for your kind attention. I would be happy to take a few questions.
1 William C. Dudley. 2011. "Securing the Recovery and Building for the Future." Remarks at United States Military Academy at West Point, New York, November 17.
2 Based on the CoreLogic index.
3 FRBNY Quarterly Report on Household Debt and Credit 2011 Q3.
4 Because housing is such an important asset for many households, household expectations of future income may also be influenced by what they observe happening to house prices
5 These households have particularly strong incentives to increase saving to recreate a buffer against adverse events and to enable them to move house. Source of data is CoreLogic.
6 In Las Vegas, for instance, the ratio of house prices to replacement costs is now 0.7, having peaked at about 2.0 during the boom.
7 The supply curve is fixed and vertical at the post-bubble stock of housing so the entire adjustment to changes in demand takes place via prices rather than quantities.
8 A number of these issues were documented by an interagency task force that included the Federal Reserve and are being addressed through consent orders requiring supervised institutions to remedy deficiencies.
9 When the amount of debt secured by a property exceeds its likely net sales price, a "short sale" can be a less costly alternative to foreclosure. The short sale agreement between the homeowner and lender often grants the lender final approval of the sales price and may release the homeowner from any deficiency.
10 "The National and Regional Economic Outlook," October 24, 2011, and "Securing the Recovery" November 17, 2011.
11 Board of Governors of the Federal Reserve System. 2012. White Paper, "The U.S. Housing Market: Current Conditions and Policy Considerations," January 4.
12 According to the Senior Loan Officer Opinion Survey.
13 We should want the cost of misrepresenting loans to be invariant to the state of the jobs market.
14 Risk retention will also apply to some forms of mortgages in future.
15 In the early 1980s, Pennsylvania created such a loan program to assist homeowners dealing with unemployment due to the closing down of steel mills. The program was called the Homeowners's Emergency Mortgage Assistance Program, or HEMAP. Since its inception, over 80 percent of the loans made under the program have been fully repaid and the borrowers able to remain in their homes. A new program could be funded by the federal government or individual states.
16 In addition, the repayment terms for the bridge loan once the owner becomes reemployed should be contingent on the owner's reemployment earnings.
17 Lender Processing Services data.