Housing and Economic Recovery
Economic Outlook Seminar Stockholm, Sweden
It is a great pleasure to be in Sweden to participate in today's session on the current economic outlook and tomorrow's conference on capital markets in the post-crisis environment. It is vital that policymakers and practitioners study and apply the recent lessons of the global economy and financial interdependence, so I am glad to be able to join you in these discussions.
I should note, of course, that the views I express today are my own, and not necessarily those of my colleagues on the Federal Reserve's Board of Governors or the Federal Open Market Committee (the FOMC).
On a personal note, I should mention that my grandfather John Johanson left Sweden at the age of six, changing his name to Rosengren and joining his father - who was deploying what he had learned as a gardener in Sweden to open greenhouses in America. My relatives were able to take what they had learned in Sweden and apply it - with perhaps some variation - in the United States. In a similar way, there is much to consider about the economic experience in Europe and the U.S. over the past four years. What is similar? What is different? And what can be learned and applied?
The window into these broader questions that I would like to use today is housing. One of the hallmarks of the recent financial crisis, and subsequent economic disruption, has been the very significant role played by unusually large movements in residential real estate values, within and across countries.
Given our setting today, let me start by comparing U.S. and Swedish home price movements since 1995, shown in Figure 1. In the U.S., house prices increased steadily and reached a peak in 2006, before beginning an unprecedented decline - not just in selected regions of the country, but nationally - resulting in aggregate declines of more than 30 percent. In contrast, Swedish real estate prices increased even more substantially, and have not experienced any substantial decline. Indeed, they continued to rise after the global financial crisis.
While the U.S. housing price decline was a nationwide phenomenon, there were relatively more severe declines in some parts of the country, such as the so-called "sand states" of Florida, Arizona, Nevada, and California. Figure 2 shows the home price declines in various metropolitan areas (a subset of those included in the S,P/Case-Shiller 20-city composite index) and the lines drop furthest and fastest for the metro areas in Nevada, Arizona, Florida, and California.
As I talk with economists from countries whose housing values have risen markedly but not experienced sharp declines, I have been struck by two things. First, they are often confident that national (versus regional) house-price reductions are unlikely. And secondly, most assume that a decline in house prices would have a measured impact on the economy should that in fact occur. But the experience of Japan in recent decades and the U.S. more recently[Footnote 1] should provide some caution - given that the economic retrenchment that followed these significant declines in home values exceeded most people's expectations.
While sustained rapid increases in house prices are frequently followed by significant declines, I am certainly not predicting price declines in any country, given my limited knowledge of market characteristics. But I do recommend a thoughtful consideration of the recent U.S. experience, given its severity and the fact that it was not foreseen. Indeed, while residential investment [Footnote 2] now comprises only 2.2 percent of Gross Domestic Product (GDP) in the U.S., our housing problems have had a tremendously disproportionate impact on economic performance and on the recovery of growth and employment.
I would like to discuss some of the reasons for that disproportionate impact today. By way of preview, let me mention that I plan to highlight the following points.
First, while housing is a small component of U.S. GDP, it is also quite volatile. Historically, an outsized proportion of U.S. economic growth in the first two years of a recovery from a recession is generated by residential investment. The sector is very responsive to monetary policy[Footnote 3] since the ability to purchase a home can be quite sensitive to the interest rate and the resulting cost of financing.
Second, the characteristics of a country's mortgage-finance market determine the impact that will come from a change in the rates directly influenced by monetary policy. In the U.S., most homes are financed by 30-year fixed-rate mortgages, so a fall in long-term interest rates really only affects existing home owners to the extent they refinance. As a result, the U.S. gets less effect from the movement of short-term, monetary policy interest rates compared to countries where the primary mortgage financing instruments are floating-rate loans. In countries with this institutional feature, both new and existing home buyers get more immediate boosts from lower short-term rates. Figure 3 helps to illustrate this by showing the greater responsiveness of U.S. adjustable-rate mortgage rates to the target federal funds rate, versus 30-year fixed-rate loans.
Third, many U.S. financial institutions have significant exposure to real estate, either through direct lending or through the purchase of mortgage-backed securities. As a result, declines in real estate prices can have a substantial impact on the capital of financial institutions, which impacts their ability to finance not only the housing sector, but also other sectors of the economy.[Footnote 4]
Fourth, falling home prices have not only impaired the capital of many financial institutions, but in a very real sense have disrupted the transmission of monetary policy. This is because falling home prices have brought about a situation where the availability of credit is more important in many cases than the cost of credit. Borrowers may not qualify for lower rates - in other words, the credit might not be available to them - because their household balance sheet is weaker than prior to the recession, because the collateral values of homes and commercial real estate have fallen, and because lender perception of risks inherent in the economy and in collateral valuations has risen - all at a time when lenders' capital has been depleted.
Fifth, I will argue that changes in some housing policies in the U.S. could make it less difficult for monetary policy to affect the economy when housing prices have fallen.
I will also discuss the actions taken last week at the FOMC to promote a faster recovery in the housing sector and the economy more broadly. I am very supportive of these actions.
Let me stress that I know housing markets have critical institutional features that differ greatly across countries - including financing arrangements, mortgage insurance, the role of government, tax treatment, mortgage servicing, and borrower liability in foreclosure. This makes it difficult to generalize from the U.S. experience.
What may be transferable, however, is the observation that sharp declines in housing prices can have additional negative effects, with broad implications for macroeconomic outcomes and monetary policy - broader, perhaps, than may be assumed and incorporated into most statistical models of the economy. I don't intend to talk about whether housing prices will fall, but rather want to emphasize that when they do fall the effect on the economy is often much greater than many people expect.
Let me pause to mention at the outset that a number of my colleagues at the Federal Reserve are studying and speaking about these important issues. Most recently Elizabeth Duke, a member of the Fed's Board of Governors and herself a former banker, spoke at the Board's Policy Forum on the Housing Market Going Forward and I recommend her address to you.[Footnote 5] I share Governor Duke's view that in the U.S. we "currently have a housing market that is so severely out of balance that it is hampering our economic recovery." I may not be as convinced as she is about one possible remedy - that banks may be suitable landlords upon converting foreclosed properties to rentals - but I recommend absorbing her perspectives on things that can be done in the near term to help the housing market stabilize and rebalance.[Footnote 6]
Also, colleagues at the Boston Fed including Paul Willen and Chris Foote have done a great deal of empirical work on these issues, and are exploring the key institutional characteristics of the housing market that play important roles in the current problems. Among other things, they are working on estimates of the macroeconomic effects of possible refinancing plans.
The Role of Housing in the Slow U.S. Recovery
The decline in housing prices has created a multitude of difficulties for the U.S. economic recovery. I would like to provide some evidence of the problem, and discuss its origins.
The housing sector in the U.S. played a major role in the recent financial crisis, and in making the recovery from the resulting recession so anemic. In the U.S. we had seen regional declines in home prices during previous recessions, but had not seen a national decline in house prices in the post-war era. So for most lenders, investors, and homeowners, the sustained and substantial decline in home prices was not anticipated. Even sophisticated investors had placed only a very small probability on such a correlated decline. A key lesson from the past several years is that homes are not riskless investments - and that highly rated, regionally-diversified mortgage securities do not necessarily carry low risk.
As mentioned earlier, even though residential investment is a small share of GDP (today only 2.2 percent), it is quite interest-sensitive - it can decline quite dramatically as interest rates rise, and expand quickly when interest rates are relatively low. So it has been a disproportionally important part of the monetary policy transmission mechanism.
In the current situation, however, U.S. mortgage rates are quite low but residential investment has not been the engine of growth that it normally is in economic recoveries. As shown in Figure 4, exports have been a source of strength in the first two years of the U.S. recovery, and business fixed investment has grown at approximately the same rate in this recovery as in the previous three. Yet the household sector has been particularly weak. Consumption, which accounts for approximately 70 percent of U.S. GDP, has grown only about half as much in the first two years of the recovery as it did in the previous three recoveries. And the shortfall for residential investment is even more striking. In the previous three recoveries, residential investment grew over 30 percent on average in the first years of the recovery - but has actually decreased in the first two years of this recovery.
Figure 5 explores this pattern further by showing the contributions to total GDP growth that residential investment has provided in the first and second years of recent recoveries. In the past three recoveries, residential investment on average contributed more than 20 percent of GDP growth in the first year and 10 percent of growth in the second. So in the typical recovery, residential investment contributes much more to growth in the first two years of the recovery than would be expected given its share of GDP. This reflects the fact that interest rates generally fall during a recession and the initial stages of a recovery - and residential investment is usually quite responsive to interest rates.
In this recovery, residential investment was only a small share of the growth in the first year of the recovery and declined in the second year. So the unusually slow recovery is, in part, due to the uncharacteristically muted response of residential investment to very low interest rates.
More than one observer has commented that we are seeing a different pattern this time that equates almost to a "negative feedback loop." High unemployment leads to risk aversion, which decreases demand for new housing. But without construction activity we are not seeing the typical uptick in housing-related jobs.
Having said a bit about residential investment and the slow recovery, I would like to now touch on other problems that stem from the decline in housing prices. Because residential investment is not the only facet of the economy that is affected by the drop in home values.
Indeed, the decline in housing prices has contributed to a sharp decline in household wealth. In fact, Figure 6 shows (in both dollar and percent terms) that there was a significant decline in household net worth from 2005 to 2009, and that it particularly impacted racial and ethnic minorities in the U.S.[Footnote 7] These insights are drawn from a Pew Research Center study that used data from the Census Bureau's Survey of Income and Program Participation.[Footnote 8]
To explore this a bit more, Figure 7 shows that while homes are an important component of net worth for Americans regardless of racial and ethnic group, African American and Hispanic households are particularly reliant on home values as a share of their wealth (between 50 and 60 percent, versus between 30 and 40 percent for White and Asian households in the U.S.). While stock prices have increased substantially from their crisis lows, those households whose primary source of wealth is their home have been disproportionately affected by the decline in house prices. And intuitively, households whose net worth has been significantly impacted are likely to reduce their consumption levels by more than households whose net worth has recovered.
Figure 8 shows the erosion in home equity for U.S. households, again in both dollar and percent terms. Home equity is not only a source of funds to finance consumer purchases, but also frequently serves as a source of funds for individuals starting new businesses. With the decline in home equity, many entrepreneurs do not have that route available to finance start-ups or to expand existing small businesses.
The weak housing sector also has an impact on employment. Figure 9 shows that far fewer jobs have been created in the first two years of this recovery (the left bar in each pair) than in previous recoveries (the right bar in the pair). In fact, construction jobs have continued to decline during the first two years of this recovery - we have lost over a half a million construction jobs since the recovery began. While construction employment is typically volatile during a recovery, on average the sector adds roughly 150,000 jobs.
Indeed, Figure 10 shows that employment in construction has declined by 9 percent in the first two years of this recovery compared to growth over 4 percent during the previous three recoveries. And weak construction employment and activity also reduces the demand for labor in sectors that support construction.
Figure 11 highlights that the rate of mortgage delinquency in the U.S. has been much greater during this recession than in previous ones. Despite two years of recovery (albeit a very slow one), the mortgage delinquency rate remains well above the peak level seen over the previous 30 years. And Figure 12 shows that with significant foreclosures and a weak economy, both the housing vacancy rate and the rental vacancy rate remain quite high by historical standards.
So the problems in housing have greatly complicated the recovery in the U.S. Not only has residential investment been unusually weak, but "to add insult to injury" consumption, business formation, and employment have also been affected by problems in the housing sector.
Housing's Impacts on the Policy Response
Monetary policy in the U.S., entrusted to the Federal Reserve by Congress, has responded aggressively to both the severe recession and the weak recovery. Short-term interest rates were cut essentially to zero.
Hitting the so-called "zero lower bound" has certainly brought challenges to the ability of monetary policy to stimulate the economy. However, once the zero lower bound was reached, the Federal Reserve used a variety of tools to try to push down longer-term interest rates, as merited by economic conditions. Our efforts have included asset purchases both traditional (Treasury securities) and nontraditional (mortgage-backed securities). We have also altered our communications strategy. At the August FOMC meeting we announced that the Fed expects to maintain short-term interest rates (the target range for the federal funds rate) near the zero bound at least through mid-2013 unless there is a significant improvement in economic conditions relative to our forecast, and as long as the inflation outlook remains subdued.[Footnote 9] At the September FOMC meeting we announced a plan to extend the maturity of our holdings of securities and reinvest principal payments from our holdings of agency debt and agency mortgage backed securities in agency mortgage-backed securities. The recent actions have helped lower long-term interest rates significantly, which is intended to promote growth in interest-sensitive sectors of the economy such as housing.
Figure 13 shows that in the U.S. the 10-year Treasury rate, and 30-year mortgage rates, have reached historically low levels. Figure 14 illustrates that the combination of low rates and falling house prices significantly improved the affordability of homes.[Footnote 10] In addition, Figure 15 shows that both population and household formation in the U.S. have continued to rise over the past five years.[Footnote 11]
The problem is, while many households have been able to refinance and some new home buyers have taken advantage of the improved affordability of houses, falling home prices have made many unable to refinance, and many unable or unwilling to purchase a home. While the housing market would have been more severely impacted (and housing prices fallen more substantially) had the Federal Reserve not eased monetary policy aggressively, the impact of our policies on the economy has clearly been impaired by falling housing prices.
Monetary policy would have had a greater impact if households were able to respond to the lower interest rates as if house prices had not fallen. The data can help one gauge the scope of the potential. The widely followed CoreLogic estimate is that as of the second quarter, 10.9 million, or 22.5 percent, of all mortgaged residential properties had negative equity, where the debt exceeds the value of the house.[Footnote 12] These so-called upside-down homeowners typically have a much harder time refinancing than owners with positive equity.
Along these lines, CoreLogic estimates that about three-quarters of owners with negative equity are paying "above-market" interest rates, which the company defines as rates above 5.1 percent.[Footnote 13] By contrast, slightly more than half of positive-equity owners are paying above-market rates. All in all, CoreLogic estimates that there are about 28 million borrowers who are paying above-market rates and thus might potentially benefit from a refinance.[Footnote 14]
There are several proposals that attempt to facilitate refinancing for homeowners who have been negatively impacted by the drop in housing prices. These proposals do face hurdles, including how to address private mortgage insurance and second liens. However, a program that made it possible for many homeowners to refinance, even if they were upside down, would likely provide significant reductions in mortgage payments to individuals who are likely to have a relatively high propensity to consume. Clearly getting more money into the hands of homeowners who would spend it could help to fuel GDP growth. This would reduce one of the impediments to a more significant effect from the monetary policy actions taken to date.
I hasten to add that there is already a government program to allow underwater borrowers to refinance, the Home Affordable Refinance Program (HARP). This program allows underwater borrowers with Fannie Mae or Freddie Mac loans to refinance at lower rates. Unfortunately, the program has helped fewer borrowers than was originally hoped. Fed Governor Betsy Duke outlined some of the potential reasons why, in the talk I mentioned earlier. They include loan-level price adjustments (LLPAs) that raise interest rates for many borrowers and thereby reduce the benefit of refinancing; originator worries about "buybacks" forced on them by Fannie Mae and Freddie Mac; junior lien-holder resistance to re-subordinating their loans; and mortgage insurance policies.
The Federal Housing Finance Agency (FHFA) is now investigating whether there are ways to enhance the program to benefit more borrowers.[Footnote 15] As this work proceeds, I hope the FHFA considers dropping or reducing LLPAs in cases when a GSE loan is refinanced into another GSE loan. Such a refinance actually reduces the GSE's credit risk (they already guarantee the existing mortgage and the homeowner will be able to take advantage of lower rates, freeing up cash flow).
A second potential way to help resuscitate the housing sector would be to facilitate the reduction in the number of vacant homes.[Footnote 16] One way to do this would be to help responsible investors purchase vacant homes and then convert them into rental properties. Today, a majority of investor deals are financed with cash, which suggests that there is room for improvement in the financing of investor properties.
A step in the right direction is a new loan product from Fannie Mae that allows a cash-out refinance of properties purchased with cash within the past six months, helping to "recycle" investor dollars through the market. But more could be done. Two possibilities include allowing investors to access FHA 203(k) purchase-rehab loans and increasing the cap on the number of Freddie Mac loans that are available to a single investor. In short, improving access to credit for responsible investors to purchase and improve properties would not only help provide affordable rental housing. It would also help to boost housing demand and limit some of the collateral problems from the large supply of vacant homes.[Footnote 17]
A third way to help today's housing market would be to provide as much certainty as possible regarding future government policy in this area. The GSEs are likely to play a different role in the new system.[Footnote 18] The uncertainty has real effects on how well our current system works. A case in point concerns GSE repurchase requests on delinquent loans.[Footnote 19]
Finally, I would like to be able to guarantee that the United States will never again suffer a significant house-price decline. But instead I can suggest some improvements that could make the housing market of the future more robust in the face of disruptive shocks. One suggestion concerns speeding up[Footnote 20] the resolution of delinquent mortgages that have no hope of becoming current again. The government's Home Affordable Foreclosure Alternative program (HAFA) is intended to speed the resolution of troubled homeownerships by offering monetary incentives to borrowers and servicers who agree to short sales or deeds-in-lieu of foreclosure. But like the larger HAMP program, take-up rates for HAFA have been disappointing.[Footnote 21]
The ultimate solution may be to rethink the foreclosure process from the start, and economic theory could play a role here.[Footnote 22] By influencing the costs and benefits of non-foreclosure transfers, a new system of ownership transfer might better align the incentives of borrowers and lenders, while still protecting the legitimate property rights of homeowners. If so, society as a whole would benefit.
In conclusion, I would reiterate that problems in the housing sector of the U.S. economy have been a major impediment to recovery. As you know, we took significant actions last week at the FOMC to promote a faster recovery in the housing sector and the economy more broadly.[Footnote 23]
I am very supportive of the actions, which were designed to promote stronger growth and reduce unemployment rates, consistent with our dual mandate, by putting downward pressure on longer-term interest rates and helping make financial conditions more accommodative. The actions announced last week involve extending the average maturity of our holdings of securities, by selling $400 billion of Treasury securities with maturities of less than three years, and purchasing $400 billion in securities with maturities of from six to thirty years. I would highlight that 29 percent of the securities being purchased are in the maturity range over 20 years.
The FOMC also announced that to support conditions in mortgage markets we will now reinvest principal payments from our holdings of agency debt and agency mortgage-backed securities in agency mortgage-backed securities, rather than purchasing Treasury securities.
There was a substantial market reaction to the announcement. Thirty-year Treasury yields declined by 18 basis points and thirty-year mortgage rates declined by 22 basis points that day - and fell further the following day.
Pictured in Figure 16 is the spread between the 30-year mortgage yield (a wholesale yield) and the 10-year Treasury yield. The spread illustrates in part what the Fed's actions are addressing - the upward movement in the spread since mid 2011 - and the market's reaction, a 17-basis-point drop in the spread, following the Fed's announcement. Since some of this action had been anticipated and already reflected in market prices, I view this as a very significant decline.
While there are a variety of impacts that lower long-term Treasury and mortgage yields can have, certainly one is encouraging more refinancing and home purchases than would happen in the absence of action. In my own view, the Federal Reserve should continue to closely monitor spreads between mortgage and Treasury yields and consider taking action if that spread widens significantly, as it had recently.
While the housing sector and the economy would have been weaker in the absence of the Fed's actions to lower rates, I believe it is equally important to examine housing policies that might be changed to reduce the impediments to monetary policy and more generally help facilitate recovery. There should be strong encouragement for the GSEs to focus on the housing recovery so home buyers and those that already have loans can fully benefit from the lower interest rates generated by our monetary policy action. Given that Fannie Mae and Freddie Mac are currently under conservatorship by the U.S. government, I believe they should play a larger role in achieving the public policy goals inherent in addressing these housing-market problems.
While monetary policy cannot fully offset the many problems stemming from the recession, it can certainly mitigate some of the effects. A policy action that, for instance, reduces the unemployment rate by half a percent over time will not return the economy to full employment, of course, but will still mean 750,000 jobs that would not have been created in the absence of the action. Of course, carefully weighing the potential costs of additional action with the expected potential benefits involves a high degree of uncertainty, given we have not operated at the zero bound with falling housing prices in the post-war period. However, with unemployment at 9.1 percent and with inflation in the medium term expected to remain below 2 percent, monetary and fiscal policies should be focused on returning the economy to full employment.
Given what I have laid out today, let me close with this. We need to recognize the urgency of the situation and the broader economic implications of housing's continued struggles. All parties in these markets should redouble their efforts, given the broader economic effects and their painful toll.