Last year was surely one of the most memorable years ever experienced by the home mortgage market. Owing largely to the lowest mortgage interest rates in more than three decades and rising home prices, close to 10 million of regular home mortgages were refinanced. I use the term regular mortgage to exclude both home equity and construction loans. The outsized dollar volume of these refinancings--by our estimates, $1-3/4 trillion net of cash-outs--was an all-time record and represented almost one-third of the value of all regular home mortgages outstanding at the beginning of last year. Total regular mortgage originations, at $2-1/2 trillion, also proceeded at a record pace.

As part of 2002's process of refinancing, households "cashed out" almost $200 billion of accumulated home equity, net of fees, taxes, points, and commissions. That represented almost 3 percent of estimated total home equity at the beginning of the year, up slightly from the 2001 share. In no year prior to 2001, as best we can judge, did cash-outs exceed 1-3/4 percent of total home equity. Of last year's cash-outs, approximately $70 billion was apparently applied to repayment of home-equity loans, and a significant part was employed to reduce higher-cost credit card debt, judging from the slowed pace of growth in installment debt outstanding.

Previous Federal Reserve surveys of the disposition of cash-outs indicate that a substantial amount--perhaps half--was used to finance home modernization and personal consumption expenditures, outlays that directly affect GDP and jobs, and that likely was the case again last year. Low mortgage rates doubtless motivated much of this spending, but the ready availability of home equity for extraction appears to have also played a substantial and independent role in prompting additional household expenditures. Even as recently as the late 1980s, a family that wanted to use housing wealth to finance consumption would have faced an expensive and time-consuming process. Although substantial home equity wealth has existed for many years, only in the last decade or so has secured borrowing against home equity become a cost-effective source of credit in a wide variety of circumstances.

An even greater support to the economy than cash-outs last year was the extraction of home equity associated with a record 6.4 million existing home sales, including condos, at record prices. This pace of ownership turnover of the existing housing stock also reflected the near-record-low mortgage rates.

We estimate that mortgage originations for existing home purchase last year topped $600 billion. Subtracting the home sellers' repayments of the remaining debt of their outstanding mortgages, we infer a net increase of approximately $350 billion in debt on the homes that turned over last year. That debt increase exactly matches the extraction of previously built-up equity on those homes plus fees and taxes folded into the loans. Not surprisingly, the change in mortgage debt is highly correlated with the realized capital gains on the turnover of those homes.

We do not have a direct measure of how the equity extracted from last year's home turnover was expended. It is likely, however, that home sellers, after setting aside a down payment for the family's next home, expended a considerable part of their home equity extraction on goods and services.

In addition to the extraction of equity financed by regular mortgages last year, approximately $130 billion was drawn through a net increase of home equity loans, also a record, and also presumed to be the consequence of low mortgage rates as well as accelerated appreciation of homes. Federal Reserve studies of the disposition of home equity loans suggest a slightly higher rate of repayment of credit card and other consumer nonmortgage debt than our cash-out disposition surveys concluded.

All in all, the amount of previously built-up equity extracted from owner-occupied homes last year, net of fees and taxes, totaled $700 billion by our calculations, or more than 10 percent of estimated equity at the beginning of the year. Home equity extraction for the economy as a whole is, of necessity, financed by debt. In fact, the $700 billion of equity extraction is similar to the increase in mortgage debt last year.

Despite the exceptionally large extraction of equity, the total remaining equity at the end of 2002 was higher in dollar terms than at the beginning of the year, owing to a 7 percent increase in existing home prices over the four quarters of last year and $300 billion in new home construction, net of mortgage extensions on those homes. Mortgage debt as a percent of the market value of homes accordingly rose somewhat more than 1 percentage point during the year.

Mortgage debt service costs as a percent of the disposable income of homeowners last year were little changed from 2001. The estimated 10 percent of homeowners' disposable income allocated to mortgage debt service in the third quarter of last year was well below the highs in 1991, though the ratio of homeowners' mortgage debt to their disposable income rose to a record high.

Refinance and home purchase originations, seasonally adjusted, peaked in the fourth quarter of last year. It is difficult to imagine that pace being maintained in the current quarter. Seasonally adjusted applications for refinancings, as reported by the Mortgage Bankers Association, are off their peaks but still impressively high. So are home purchase mortgage applications. This suggests a somewhat less robust market for mortgage originations this quarter.

With home price increases now subsiding, and mortgage interest rates no longer declining at last year's impressive pace, some slowdown in the rate of mortgage debt expansion is to be expected. That is likely to be the case for equity extraction from home turnover as well. As I noted earlier, that extraction appears to parallel the realized capital gains on home sales, which mainly reflect the number of existing home sales. Home price change, of course, is also a factor, but it is the average change over the length of occupancy--nearly ten years for the typical homeowner--that matters, not recent price trends.

Similarly, refinancing and the cash-outs associated with them accelerated last year as the decline in mortgage interest rates on new loans far exceeded the modest decline in the average rate on all outstanding regular home mortgages. This wider spread between the current mortgage rate and the rate on outstanding loans, as you all know, markedly increased the incentive to refinance. So any stabilizing of rates on new mortgages would narrow the spread as portfolio rates continue to decline as a consequence of continued refinancing at interest rates still below the portfolio average. A narrowed spread would likely reduce cash-outs, lessening the level of equity extractions from this source. We do not have estimates of total home equity loans beyond the end of last year, but lines of credit at commercial banks--approximately a fourth of total home equity loans--rose during the first 7 weeks of 2003, considerably more than the typical seasonal change.

In summary, the frenetic pace of home equity extraction last year is likely to appreciably simmer down in 2003, possibly notably lessening support to household purchases of goods and services.

* * *

The very large flows of mortgage funds over the past two years have been described by some analysts as possibly symptomatic of an emerging housing bubble, not unlike the stock market bubble whose bursting wreaked considerable distress in recent years. Existing home prices (as measured by the repeat-sales index) rose by 7 percent during 2002, and by a third during the past four years. Such a pace cannot reasonably be expected to be maintained. And recently, price increases have clearly slowed.

It is, of course, possible for home prices to fall as they did in a couple of quarters in 1990. But any analogy to stock market pricing behavior and bubbles is a rather large stretch. First, to sell a home, one almost invariably must move out and in the process confront substantial transaction costs in the form of brokerage fees and taxes. These transaction costs greatly discourage the type of buying and selling frenzy that often characterizes bubbles in financial markets. Second, there is no national housing market in the United States. Local conditions dominate, even though mortgage interest rates are similar throughout the country. Home prices in Portland, Maine, do not arbitrage those in Portland, Oregon. Thus, any bubbles that might emerge would tend to be local, not national, in scope.

Third, there is little indication of a supply overhang in newly constructed homes. The level of overall new home construction, including manufactured homes, appears to be well supported by steady household formation and not dependent on high and variable replacement needs or second-home demand. Census Bureau data suggest that one-third to one-half of new household formations in recent years result directly from immigration.

In evaluating the possible prevalence of housing price bubbles, it is important to keep in mind that home prices tend to consistently rise relative to the general price level in this country. In fact, over the past half century, the annual pace of home price increases has been approximately 1 percentage point faster on average than the rise in the GDP deflator. This higher home-price inflation rate results from persistently slower productivity growth in new home construction than in the economy overall. This lag in productivity growth drives up new home prices relative to the general price level and, by arbitrage, it drives up the prices of existing homes as well. In addition, local building and land use restrictions continue to constrain the supply of buildable land in many areas, whose price increases also tend to outstrip the rate of inflation.

Clearly, after their very substantial run-up in recent years, home prices could recede. A sharp decline, the consequences of a bursting bubble, however, seems most unlikely. Nonetheless, even modestly declining home prices would reduce the level of unrealized capital gains and presumably dampen the pace of home equity extraction. Home mortgage cash-outs and home equity loan expansion would likely decline in the face of declining home prices. However, the five-year old home building and mortgage finance boom is less likely to be defused by declining home prices than by rising mortgage interest rates.

Should rates rise, it is entirely possible that new and existing home sales would decline, leading to a lower level of realized capital gains on homes, a further narrowed refinance spread and, as a consequence, less overall home equity extraction. It is worth bearing in mind, however, that any sustained increase in rates presumably would occur only in the context of a more vigorous upturn in the pace of business activity, suggesting that the net effect on housing activity might be relatively limited.

* * *

Most of the aforementioned data are new and derive from a much broader home mortgage data system in the early stages of development by the Federal Reserve Board. As a consequence, all specific numbers I have cited this morning are preliminary and subject to revision.

The evident increasing importance of tracking these very large flows of housing-related credits that have become so prominent in the past decade or two has led the Federal Reserve Board into a major statistical program to fill a notable gap in currently available data.

The Board has for years estimated outstanding home mortgage debt and published these estimates in our Flow of Funds accounts. We publish quarterly estimates of the stock of one- to four-family home mortgage debt as compiled from the vast majority of institutions that originate or hold home mortgage loans. These quarterly estimates of the levels of mortgage debt, however, do not allow us to monitor important flows such as refinancings, cash outs, and repayments. That shortfall needed to be remedied.

The availability of considerable detail on the vast majority of originations required by the Home Mortgage Disclosure Act (HMDA) was a catalyst of this new project. Our latest tabulation of HMDA data, for the year 2001, provided detailed characteristics on almost 13 million mortgages with dates of application and origination. This unique body of data enables us to create weekly, monthly, and quarterly data on applications and originations of regular mortgages. These data are available for home purchase and refinance mortgages as well as for originators' purchase and sale of regular mortgages. Incidentally, the proportion of refinance to home purchase mortgages appears consistently higher in the data reported to us than in the data published elsewhere.

Independently, we have constructed a set of quarterly gross mortgage flows consistent with our series on mortgage debt outstanding. These calculations suggest significantly higher estimates of mortgage originations, but otherwise closely parallel the HMDA published originations. We are working on a system to adjust the HMDA internal detail to the wider coverage of originations consistent with our published outstanding mortgage debt on one- to four-family homes.¹

* * *

Home equity extraction directly finances household purchases of goods and services by liquefying previously illiquid assets. It also indirectly finances such purchases by facilitating outlays financed by credit card and other nonmortgage consumer debt. Equity extraction has been a major source of repayment of such debt.

Borrowing against home equity has been a staple of household finance for decades, but as I noted earlier, it has been only in the past decade or so that such practices have been encouraged by lenders. We need to far better understand the economics of this major addition to household finance and its impact on the economy. One hopes, new data will form the basis of better insights.

There can be little doubt that the availability of a ready source of home equity has reduced the costs and uncertainties associated with income volatility, retirement, unexpected medical bills and a host of other life events that can unexpectedly draw down savings. Home equity extraction may be the household sector's realization of the benefit of a rapidly evolving financial intermediation system.

1. Estimates of construction loans are excluded. Return to text