I. Macroeconomic Implications of Housing Trends
II. Length and Duration of Housing Downturns, and Other Recent Research from the Boston Fed
III. Issues Complicating the Resolution of Housing Problems
IV. Implications for Policymakers
As someone who is an ardent consumer of good economic analysis, I have often been the beneficiary of work done by the New England Economic Partnership. So it is certainly a pleasure to be with you today to participate in your discussion of current trends in the economy, and credit and housing matters. [Footnote 1]
Over the past year, the economy has been buffeted by a series of shocks including financial turmoil; an emergency acquisition of a major investment bank to avoid a sudden, disorderly failure [Footnote 2]; falling national housing prices; and oil prices that have exceeded $130 a barrel. Despite all these challenges, the U.S. unemployment rate is at 5 percent; the rise in prices of “core” consumer goods and services (Personal Consumption Expenditures) to a little above 2 percent, while unwelcome, has not been large compared with past episodes; and the economy has been growing, albeit at a much slower pace then we would prefer. These relatively benign outcomes to date are at least partly the result of recent monetary and fiscal policy actions taken to mitigate some of the problems facing the economy.
In keeping with the title of your conference – “Crunched: Credit, Housing and the Consequences for New England” – I am going to focus on the continued downside risk posed by falling asset prices, particularly in the housing sector. Residential investment (the housing component of GDP [Footnote 3]) has declined for nine straight quarters, and in all likelihood will decline this quarter as well. This has been an unusually long period of continuous decline in the housing sector, and coupled with falling housing prices nationally, has complicated predictions of what will happen in the economy going forward.
I would like to begin by first looking at some of the macroeconomic issues related to the housing sector, and considering why – despite being a relatively small component of GDP – housing continues to be a source of such significant uncertainty in the economy.
Then secondly, I will discuss some of the varied and interesting research done recently at the Federal Reserve Bank of Boston by economists and analysts (including Prabal Chakrabarti, Chris Foote, Kristopher Gerardi, Lorenz Goette, Kai-Yan Lee, Adam Shapiro, and Paul Willen). This work highlights some striking differences, and some similarities, between the current housing problems and those experienced in New England in the early 1990s. In particular I will note the observation that foreclosures remained elevated well after their peak in the last downturn, and the finding that multi-family properties are playing a much more important role in the current housing downturn.
Thirdly, I am going to discuss why developing solutions to the current housing problems has not been easy, and why the housing sector, despite being sensitive to interest rates, remains very weak notwithstanding a significant amount of monetary stimulus in the form of Federal Funds rate reductions. In particular, I will highlight the important role of so-called “piggyback loans,” or second mortgages; the implications of homeowners having “negative home equity” as a result of falling house prices; and the complications created by the securitization of many mortgages.
Finally, I am going to briefly discuss some implications for policymakers in seeking to develop and implement responses to the problems posed by falling house prices and rising delinquencies and foreclosures.
Figure 1 highlights how precipitously single-family housing starts have declined since their peak. This is actually the largest peak-to-trough decline in housing starts to occur in nearly 50 years. The decline is occurring in what would seem, in fact, to be a relatively benign economic environment, with U.S. unemployment at 5 percent and interest rates low by historical standards.
However, the reduction in credit availability generated by recent, significant disruptions in the securitization of mortgages has made it particularly difficult for borrowers seeking subprime [Footnote 4] and “jumbo” loans. What was likely to be a weak residential market has become a market characterized by major declines as builders seek to reduce their swelling housing inventories and as continuing price declines raise the costs of holding on to land or unsold homes.
These problems have begun to have an impact on some financial institutions that had taken outsized positions in construction loans, or whose business model was focused on the types of residential lending most impacted by the downturn. Of course, the dramatic decline in housing starts should expedite the process of reducing the inventories of unsold homes, and over time provide a basis for stabilizing the housing market – although with sales continuing to decline, the most recent readings show little or no progress in cutting inventories relative to sales.
A potential problem for financial institutions is the interaction between falling housing prices and homeowners’ home equity lines of credit (HELOCs [Footnote 5]) and junior liens [Footnote 6] (home equity loans). As housing prices have risen over the course of this decade, products have proliferated that allow consumers to tap into the wealth created by home-price appreciation. As can be seen in Figure 2, home equity loans and HELOCs have expanded rapidly – together they now account for over $700 billion of bank assets (which accounts for about one-third of the $2.1 trillion in one-to-four-family mortgages held by the nation's commercial and savings banks as of March 31, 2008).
Using home equity can be particularly attractive to consumers because of the tax advantages of borrowing with the home as collateral, the lower interest rate on mortgage loans relative to other forms of consumer debt, and the ability to expand the credit line during a period of rapid appreciation in home prices.
Despite the advantages to consumers, these products are potentially risky to lenders during periods of falling home prices. Most financial institutions will not take possession of homes when borrowers default on their home-equity loans (junior liens), because this would require them to assume the first mortgage. With falling housing prices, assuming the first mortgage is not an attractive option from the lender’s perspective. An additional complication is that default will frequently not occur until the home equity line of credit has been fully utilized, as borrowers who are having difficulty making payments can tap their credit line again and again. Thus, credit problems may first appear as rising balances rather than a failure to make payments.
As Figure 3 illustrates, the delinquency rates (specifically, the share of loans 90 or more days past due or nonaccruing) on junior liens and home equity lines are rising quite rapidly. One should note that most banks were not lending to so-called subprime borrowers, but nonetheless, the delinquency rate on one-to-four family mortgages just surpassed the levels seen in the early 1990s [Footnote 7].
As noted earlier, the rapid rise in delinquencies for home equity lines and junior liens held at banks is occurring despite an unemployment rate of about 5 percent – so, should the unemployment rate rise and housing prices continue to fall, financial stresses caused by the housing correction could well spread beyond the large banks involved in complex securitizations, and the smaller banks with sizeable portfolios of construction loans, to a larger set of financial institutions. Not surprisingly, in regions where housing prices are falling and delinquencies are rising, we are seeing that banks are beginning to reduce the lines of credit available to borrowers.
A second potential problem for banks is that other types of consumer debt can suffer “collateral damage” from the difficulties in the housing sector. Regions that have experienced falling house prices over the past year are also areas that are seeing some of the larger increases in delinquencies on consumer loans. Banks with significant consumer loan exposures are observing that consumer delinquencies are becoming elevated in regions most affected by housing problems.
To date, credit problems have been focused on large financial institutions, many of whom have shown the capacity to raise new capital. However, should the challenges in housing continue, problems could expand beyond securitized assets to have an impact on the nonsecuritized assets held by smaller banking institutions. It is possible that these institutions may not be able to tap additional capital quite as easily as larger institutions, and if so they may be forced to constrain other lending to address any losses. So I leave you with the cautionary note that while small and medium sized businesses have not generally experienced significant problems with credit availability to date, further deterioration in housing markets, if it occurs, could carry over and begin to impact lenders who serve such borrowers.
New England is no stranger to falling asset values. As you no doubt know, during the early 1990s, New England experienced a steep decline in housing prices. We at the Boston Fed think it may be useful to compare that experience to what we have experienced to date with falling housing prices, and we are pursuing a number of research avenues to do that.
A new research paper by Foote, Gerardi, Goette and Willen is one example [Footnote 8]. The authors have, among other things, done extensive data analysis that allows us to compare the current situation to that earlier episode, using publicly available data from the Massachusetts Registry of Deeds compiled into a database by the Warren Group. I will say more about our researchers’ findings in a moment.
Separately, we are launching today an interactive mapping feature for our website that tracks the rise in foreclosures and subprime activity, then and now. The next few slides will give you a rough sense of how it displays data -- although the site itself will show the year-by-year progressions, dynamically, and I am going to just show you static shots of given years for our purposes here today.
As you can see in Figures 4 and 5, Massachusetts moved quite rapidly from a situation of relatively limited foreclosures in 1990 to a period of very high foreclosures in 1992. The timing is interesting. By late 1989, Massachusetts house prices had begun to fall, but delinquencies and foreclosures did not really accelerate until there was also a significant weakening of the economy. In fact, the unemployment rate in Massachusetts, which had declined to 3.1 percent in late 1987, peaked at 9.1 percent in the second half of 1991. Declining housing prices alone did not cause very elevated foreclosures; it was significantly compounded by an economic shock such as the loss of a job which disrupted the ability of many borrowers to make payments. As house prices fell, many homeowners who lost their jobs in the early-1990s recession could not sell their homes to pay off their mortgages because they owed more than their homes were worth. For unemployed homeowners with “negative equity,” foreclosure was often the only option.
The more recent period also points to the importance of falling house prices and negative equity in foreclosures. In the more recent period (shown in Figures 6 and 7), the foreclosure rate – which was not particularly elevated in 2005 – had become quite elevated by 2007 as house prices softened. This increase in foreclosures occurred even though the Massachusetts unemployment rate averaged 4.5 percent for the year.
Why are foreclosures so high today, given that the economic situation is so much better than it was during the early 1990s? Even in expansions, many homeowners undergo adverse life events – like a job loss, a divorce, a spike in medical expenses, or the like – that disrupt their ability to make mortgage payments. Of course, with regard to unemployment, such household-level disruptions are not as prevalent in expansions as they are in recessions. But when such a life event does occur, it can still precipitate a foreclosure if the homeowner has negative equity because of a fall in house prices.
Another reason for elevated foreclosures today concerns changes in the susceptibility of mortgages to economic shocks. In the late 1980s, many borrowers had made significant down payments and had good credit histories. But the recent ability of borrowers with weak credit histories and little or no down payments to purchase homes, often with subprime loans (and sometimes with minimal income verification), means that a greater share of today’s mortgages are a good bit more susceptible to the types of disruptive life events that precipitate foreclosure. These borrowers were fine when housing prices were rising because if needed, they had the ability to refinance or sell their homes and pay off (or more than pay off) their mortgages. In contrast, in the current environment of falling housing prices, borrowers who made small down-payments or have otherwise risky mortgages are now more likely to end up in foreclosure if they experience an adverse event that interrupts their ability to make mortgage payments.
So, in short, we have seen similar foreclosure numbers this time around without a technical recession, and with a more modest fall in home prices. Boston Fed researchers attribute that to the prevalence of riskier loans and higher combined loan-to-value ratios in general.
Figures 8 and 9 provide some background on the foreclosure experience by house type. Several interesting patterns are apparent. In the early 1990s, foreclosures rose rapidly for all types of homes as the economy deteriorated. While foreclosures peaked in 1992, they remained quite elevated through much of the decade, despite the eventual rebounding of the economy. In addition, that period’s speculative fervor in New England real estate markets had a particularly large impact on the condo market, where a disproportionate share of the foreclosures occurred.
Today, problems appear most pronounced in the multifamily market. The two- and three-family houses that are prevalent in low- and moderate-income neighborhoods in Massachusetts have experienced very elevated foreclosure rates. Specifically, multi-family units accounted for 28.4 percent of the foreclosures in 2006-2007, but have constituted only 10.8 percent of residential purchases since 1990 [Footnote 9] (Figure 9).
The impact on particular communities can be seen in Figure 10, [Footnote 10] which shows foreclosure petitions by property type in given cities and towns in Massachusetts. In Boston, 47 percent of foreclosure petitions between January and August 2007 were multi-family properties. Many of the cities with high concentrations of low- and moderate-income borrowers have also seen particularly high rates of foreclosure petitions [Footnote 11] on multi-family properties. Our researchers suggest that one explanation for the higher foreclosure rate on multi-family properties is found in higher combined loan to value (CLTV) ratios, which make them more vulnerable when prices are falling. [Footnote 12]
States like Massachusetts and Rhode Island have a much larger concentration of multi-family homes than do many other regions of the country [Footnote 13]. Streets with a large concentration of foreclosures in major New England cities frequently are dominated by multi-family units. Recently both Boston and Chelsea have begun programs to purchase foreclosed multi-family units, to try to prevent these properties from becoming vacant and slipping into disrepair, which would bring broader collateral problems for the affected neighborhoods.
Because many of these properties are investor owned – or the current owner occupies one of the units and has limited financial means to deploy should the property require major capital improvements – programs focused on owner-occupied single-family properties will not address these problems. One area that will need focus going forward involves strategies to prevent streets of foreclosed multi-unit properties from leading to broader sections of urban decay.
Several lessons from the historical comparison can be highlighted. First, should the economy worsen and suffer a period of significant job losses, the housing problem could become much more severe. Second, past episodes of elevated foreclosures lingered well after the peak in foreclosures had passed, indicating that the duration of today’s situation may be longer than some are anticipating. Third, because the current problems have been concentrated in vulnerable low- and moderate-income areas with a prevalence of multi-unit properties, such multi-unit properties will be a more important public policy issue in the current situation.
As I have noted, a major change in the U.S. housing sector in recent years has been the emergence of the ability to buy homes with little or no down payment and the increased prevalence of such purchases. Many of these purchases are structured with a first loan that covers 80 percent of the value of the house and a second loan, sometimes called a “piggyback” loan, which is taken out at the same time as the first. The piggyback loan tends to be at a higher interest rate reflecting the greater risk of the loan, but the interest is tax deductible and allows the borrower to avoid purchasing private mortgage insurance (which lenders require if the borrower is paying less than 20 percent in down payment).
As can be seen in Figure 11, until recently, an increasing percentage of homes were being financed with these piggyback loans. In Massachusetts, piggyback loans were used with 26 percent of subprime mortgages in the second quarter of 2003 but that percentage increased to 65 percent of subprime mortgages in the third quarter of 2005. Piggyback loans were about twice as prevalent for subprime borrowers as prime borrowers, and were particularly prevalent in multi-family purchases. In fact, 75 percent of subprime multi-family purchases in the third quarter of 2005 involved a piggyback loan.
At the time of purchase, subprime borrowers frequently used the same lender to acquire both the first loan and the piggyback loan. The complicating factor is that these loans were frequently then securitized, with the first mortgage and the piggyback loan often sold into separate portfolios of loans to be securitized. Thus, even while piggyback loans were becoming more abundant, many mortgage-backed securities contained no piggyback loans, implying that the borrower’s loans were frequently packaged in separate securitizations.
The combination of piggyback loans, securitization, and legal complexities makes it much more difficult to modify loans, particularly if the first and second mortgages have been packaged into separate securitizations. Dealing with two different lenders or servicers, as well as different securitizations, can significantly complicate efforts to restructure loans. Also, loan servicing agreements are not identical and indeed include a variety of forms of authorization (of the servicer) to modify loans. Frequently the only leverage the owner of the piggyback loan can and does exercise with a distressed borrower is demanding compensation should the borrower seek to modify both loans.
In short, while a variety of programs and policies are being considered in order to improve the loan modification process for distressed borrowers and their lenders, a key consideration involves figuring out how to modify loans when the interests of the primary loan holder may be different than those of the holder of the piggyback loan.
A second complication involves targeting the modifications to the homeowners who really need them. We have seen that falling house prices tend to raise foreclosures by causing a negative equity situation which limits homeowners’ options (e.g., to refinance or sell and pay off the loan) if their financial situation deteriorates.
The rising share of homeowners with negative equity in their house – meaning their home is now worth less than their loan(s) – might seem to suggest a large group for whom loan modifications would be a “slam dunk.” Principal write-downs to bring “underwater” mortgages above water may well be the appropriate solution for some troubled borrowers [Footnote 14], but lenders and loan servicers continue to be reluctant to provide modifications in a wholesale manner. Negative equity does not imply foreclosure with certainty – in fact, most people with negative equity situations continue to make their mortgage payments. Lenders and loan servicers who are aware of this fact may be reluctant to offer a modification if they believe that the homeowner has the capacity to continue making payments, no matter the hardship.
To “unpack” this a bit, I want to spend a moment looking at the history of negative-equity borrowers over the past two decades, drawing on a new working paper just produced by Boston Fed researchers [Footnote 15].
In the late 1980s, house-price declines in New England meant a large number of borrowers had negative equity in their house. The Registry of Deeds data allow us to study the historical relationship between negative equity and foreclosures for a limited set of borrowers – specifically, people who purchased homes in Massachusetts on or after January 1987, when the Warren Group data begin. Using these data to calculate a rough proxy for housing equity, Foote, Gerardi and Willen calculate that 100,288 of these post-1987 purchasers, or 36 percent, were in a position of negative equity as of the fourth quarter of 1991 (a high fraction because January 1987 came just before the peak of the housing cycle). [Footnote 16] Yet of this negative-equity group, only 6,453 (or just 6.4 percent) defaulted on their mortgages within the following three years.
What about today? Using the same rough method to approximate home equity for the same subset of homeowners, the data imply that about 10 percent of post-1987 purchasers were in a position of negative equity as of the fourth quarter of 2007. Assuming that the unemployment rate and benchmark interest rate (the 6-month London Interbank Offered Rate or LIBOR) stay at their fourth-quarter 2007 levels, the statistical default model of Gerardi, Shapiro and Willen predicts that less than 10 percent of these homeowners with negative equity will default.
Continued declines in house prices, higher unemployment, and possibly a greater willingness to default on home mortgages might raise this estimate of future defaults. Even so, many lenders will not be inclined to make concessions unless borrowers clearly lack the ability to pay. This is one reason that workouts or distressed borrowers’ mortgages continue to proceed loan by loan, rather than through wholesale changes in mortgage terms. Lenders want borrowers with the capacity to pay to continue to uphold their legal obligation to do so – and any wholesale change in mortgage terms could benefit borrowers that are still able to fulfill the original agreement. To avoid this problem, some policy proposals [Footnote 17] maintain the legal obligation for the full loan for the borrower, but involve significantly reduced payments in return for the lender sharing in any future appreciation of the house.
Another impediment to restructuring loans that have been packaged into securitizations involves the legal restrictions imposed on many securitization agreements [Footnote 18]. These agreements allow the servicer to pass through payments by borrowers without incurring any tax liability, as long as certain conditions are met. One of the conditions is usually that modifications cannot be made to the loans unless the loan is either in default or default is reasonably foreseeable. This determination is generally made loan by loan, with wholesale modifications being interpreted by some servicers as possibly violating the service agreement and the legal framework used in many securitizations and thus potentially imposing tax liability.
Despite some promising proposals, these factors – the presence of piggyback loans, the desire of lenders to continue to get paid unless the borrower does not have the capacity to make payments, and the legal restrictions and tax-liability concerns surrounding securitization agreements – make any across-the-board modification of loan terms challenging.
As long as loan modifications are taking place on a loan-by-loan basis (a time-consuming and labor-intensive process), to significantly increase the number of loan modifications and avoid significant numbers of foreclosures, many servicers will need to quickly expand the pool of staff capable of making these evaluations.
To reiterate, falling housing prices continue to be a significant source of down-side risk to the economy. Previous periods of real estate problems have taken significant time to be worked out, with foreclosures remaining elevated well after their peak. The current foreclosure problem has been exacerbated by the difficulties related to many of the problem loans being held in securities.
To help mitigate the current situation it is important for servicers to have the capacity and the incentive to make loan modifications where appropriate. A challenge continues to arise around getting borrowers and servicers together to determine if loan modifications are appropriate.
In addition, it is clear that servicers were not set up for the spate of elevated foreclosures we have already experienced, and further increases in foreclosures would certainly test the capacity of servicers to have sufficient staff with the expertise to make loan modification determinations. Nonetheless, I am hopeful that the many mortgage forums that are being organized around the country to try to get servicers and borrowers together to work with problem loans will help some borrowers.
The large number of multi-family dwellings entering foreclosure presents a serious challenge to cities. Already there are streets in some cities with a high concentration of foreclosed multi-family properties, putting at risk many of the socioeconomic advances made in the last decade in some of these neighborhoods. Cities, working together and with community organizations and state officials, [Footnote 19] will need to find ways to make sure these low- and moderate-income neighborhoods do not regress as a result of the current wave of foreclosures. One promising tactic is rigorous municipal-code enforcement that ensures foreclosed properties are maintained. Furthermore, servicers must have a process for maintaining foreclosed properties and they must do more to ensure that these properties do not remain vacant for long periods of time.
Finally, the legal structure for securitizations and mortgage-servicing agreements clearly did not foresee the widespread emergence of distressed borrowers, delinquencies, and foreclosures that are being experienced. Changes to servicing agreements and securitization structures that allow more flexibility during times of duress and also address servicers’ liability concerns are necessary. Since different investors in a mortgage-backed security have different incentives, more clearly defining the governance of these securities would make it easier during difficult times to make modifications that are in the borrower’s, servicer’s, and investor’s interests.
The extent of eventual housing problems is highly dependent on the outlook for the economy and the future path of housing prices. Fortunately, aggressive monetary and fiscal policy actions have been taken to help mitigate some of the downside risk. These policies will likely result in some pick up in economic activity in the second half of this year, which should help to stabilize the housing market.
According to interagency guidance issued, in 2001, “The term ‘subprime’ refers to the credit characteristics of individual borrowers. Subprime borrowers typically have weakened credit histories that include payment delinquencies and possibly more severe problems such as charge-offs, judgments, and bankruptcies. They may also display reduced repayment capacity as measured by credit scores, debt-to-income ratios, or other criteria that may encompass borrowers with incomplete credit histories. Subprime loans are loans to borrowers displaying one or more of these characteristics at the time of origination or purchase. Such loans have a higher risk of default than loans to prime borrowers. Generally, subprime borrowers will display a range of credit risk characteristics that may include one or more of the following: Two or more 30-day delinquencies in the last 12 months, or one or more 60-day delinquencies in the last 24 months; Judgment, foreclosure, repossession, or charge-off in the prior 24 months; Bankruptcy in the last 5 years; Relatively high default probability as evidenced by, for example, a credit bureau risk score (FICO) of 660 or below (depending on the product/collateral), or other bureau or proprietary scores with an equivalent default probability likelihood; and/or Debt service-to-income ratio of 50 percent or greater, or otherwise limited ability to cover family living expenses after deducting total monthly debt-service requirements from monthly income.”
An entity that initially qualifies as a REMIC may cease to qualify if a sufficiently large portion of its qualified mortgages are “significantly modified” and the modified obligations are not qualified mortgages. A qualified mortgage is usually treated as having been significantly modified if the change in terms would be treated as an exchange of obligations under section 1001 and its regulations. However, certain changes to the terms of an obligation are not considered signification modifications for this purpose, notwithstanding the fact that they may be significant modifications under section 1001 and § 1.1001-3. The four permitted modifications are (1) Changes in the terms of the obligation occasioned by default or a reasonably foreseeable default; (2) Assumption of the obligation; (3) Waiver of a due-on-sale clause or a due on encumbrance clause; and (4) Conversion of an interest rate by a mortgagor pursuant to the terms of a convertible mortgage.
As echoed in many Servicing Agreements, the most pertinent modification is one induced by either default or impending default. The problem lies in ascertaining with some certainty whether default is “reasonably foreseeable” and thus servicers have been reticent to go forward with modifications based on this shaky standard when the potential consequences are so dire.
1991:Q1 - 2008:Q1
Source: Commercial and Savings Bank Call Reports