Journal Issue

Are U.S. Households Living Beyond Their Means?

International Monetary Fund. External Relations Dept.
Published Date:
March 2004
  • ShareShare
Show Summary Details

Consumer spending, household wealth, and real estate prices in the United States

THE BOOM in consumer spending, now in its 12th year, has weathered many adverse shocks that have hit the U.S. economy in recent years, including terrorist attacks and heightened security concerns, a sharp decline in equity prices and the 2001 recession, and a series of corporate and financial scandals. Throughout, resilient household demand has not only sustained domestic growth but has also played a key role in supporting the global economy.

Some observers, however, have viewed the surge in consumer spending with apprehension. The personal saving rate has fallen to historic lows, consumer debt levels have risen, and the household home equity ratio has dropped to an all-time low. This has led to concern that the rise in private consumption may not be sustainable and that a subsequent weakening could throw the recovery off track. Fears have especially been expressed that consumers could be exposed to a collapse of what many view as a housing “bubble” in the United States, given the spectacular increase in real estate prices in some markets.

Photo shows shoppers crowding a big department store in New York during the holiday rush.

So are U.S. consumers living beyond their means? Examining the evidence based on background work for the IMF’s 2003 Article IV consultation with the United States—the annual round of discussions on macroeconomic and structural policies—this article finds that U.S. households are in a better financial position than they may appear.

Why household spending is strong

The strength of U.S. consumer spending is closely related to increases in personal incomes and wealth. Households have benefited from the rise in housing and stock markets over the past decade, with net housing and equity wealth rising by $3¼ trillion and $2¾ trillion, respectively, between the end of 1995 and the second quarter of 2003. Other components of financial wealth—primarily cash and bond holdings—also rose strongly, by $6½ trillion, bringing the total increase in personal wealth to about 120 percent of GDP during this period. As shown in Chart 1, there has been a close inverse relationship between household net worth and the personal saving rate, consistent with the permanent income hypothesis, under which household expenditures respond principally to changes in long-term wealth rather than short-term fluctuations in income.

Chart 1Saving less

As U.S. household net worth rises, personal saving tends to fall.

Citation: 41, 1; 10.5089/9781451953411.022.A013

Sources: Bureau of Economic Analysis. Federal Reserve Bank: and IMF staff calculations.

1Prior to the 2003 comprehensive revisions.

In addition, consumer spending has been supported by expansionary macroeconomic policies over the past three years. A series of tax cuts has contributed to a $230 billion decline in personal income tax payments to the federal government, and government spending has increased. The economic forecasting firm Macroeconomic Advisers estimates that federal, state, and local fiscal stimulus added 1–2½ percentage points to growth every year during 2001–3.

The effect of monetary policy has also been significant. Action by the Federal Reserve to lower the federal funds target rate from 6½ percent in 2000 to 1 percent in 2003—a 50-year low—has spurred successive waves of mortgage refinancing, releasing substantial financial resources to fund consumer spending. This has been facilitated by innovation and improved risk management in U.S. financial markets, broadening household access to credit and providing consumers with greater flexibility in managing balance sheets.

In addition to mortgage refinancing, households have borrowed from other sources to finance consumption purchases, including higher credit card debt and short-term personal loans. Debt ratios have steadily increased since the mid-1990s, reaching a record high of 115 percent of disposable income in mid-2003 (see Chart 2). Although the increase in private indebtedness could eventually give reason for concern, other indicators suggest that households may have been careful to limit their financial exposure:

Chart 2Growing household debt

Although U.S. household debt has trended upward, payments on debt service and other financial obligations relative to disposable income have barely risen.

Citation: 41, 1; 10.5089/9781451953411.022.A013

Sources: Federal Reserve Bank; and IMF staff calculations.

• As a result of low interest rates, the ratio of debt service to income has risen by a relatively small margin.

• The Federal Reserve’s broad measure of debt service (including rent and other recurring monthly payments) has remained at about 18 percent of disposable income, suggesting that many households have simply substituted mortgage debt payments for rent (the U.S. home ownership rate climbed to a record 68½ percent as of mid-2003).

• Growth in consumer debt relative to household assets has been relatively moderate, and the 2003 equity market recovery has further bolstered the financial position of U.S. households.

Given the wealth losses caused by the collapse in the stock market bubble, many observers would have expected household savings to rise more sharply because of the close inverse relationship between the two variables. As detailed in Faulkner-MacDonagh (2003), however, the impact on consumption has been mitigated by a portfolio shift in household wealth. He found that the empirical relationship between more liquid forms of wealth and consumption is considerably stronger than for equity wealth. Therefore, while the wealth decline has been concentrated in equities, the rise in other forms of wealth—such as bonds and cash—in the past several years has helped to mute the effects of the stock market decline on consumer spending.

Faulkner-MacDonagh’s results also suggest that households’ saving behavior has so far remained broadly consistent with developments in income and wealth levels. On the basis of household data for mid-2003, his model predicts a saving rate of around 4½ percent, which is somewhat higher than the actual saving rate, estimated to be 3¼ percent at the time. However, this difference is not so large as to suggest that any adjustment could not be accommodated in a gradual fashion. That said, the comprehensive revision of the U.S. national accounts has since led to a downward adjustment in the household saving rate to 2¼ percent, highlighting concern about the scope for an eventual uptick. (At the time this article went to press in early 2004, the model prediction could not be updated because revisions of some key variables used in the estimation were unavailable.)

Some analysts have argued that U.S. household saving would need to increase to reduce global current account imbalances and to prepare for the retirement of the baby-boom generation. Other things being equal, an increase in the domestic saving rate would reduce the U.S. current account deficit and the need for foreign capital inflows. However, while the U.S. personal saving rate appears relatively low from an international perspective, it may be consistent with the high efficiency of U.S. capital markets and use of capital by domestic businesses. For example, a 1996 study by the McKinsey Global Institute suggests that the low personal saving rate is consistent with the relatively high returns to U.S. investments.

A similar finding is reached by a 2003 Congressional Budget Office (CBO) report on baby boomers’ retirement prospects. The CBO concluded that, on average, future retirees tend to have higher current income and wealth than earlier generations and are accumulating wealth at roughly the same rate. This would suggest that most baby boomers would also have higher retirement incomes. However, the report also warns that many low-income households may have failed to acquire sufficient retirement assets and thus remain dependent on Social Security benefits.

Housing: the next asset price bubble?

During the decline in equity prices over the past several years, the strong U.S. housing market has been a key factor in sustaining consumer spending. At the same time, the booming U.S. real estate market has often been compared with that in the United Kingdom, Spain, and Australia, which were also thought to have soared well beyond levels justified by economic fundamentals. Given the potential implications of a sharp correction in real estate prices for household wealth and consumption—as well as for overall growth and financial stability—it is important to examine more closely the prospects for the U.S. housing market.

Underlying concerns about real estate markets is the observation that the annual rate of increase in median house prices in the United States accelerated to 6 percent in 2002 from around 4½ percent in 1995, with even faster increases in many metropolitan markets. As a result, real house prices have risen above their long-term trend after a protracted recovery from the last downturn in the late 1980s.

“The balance sheet of the U.S. household sector still looks robust.”

These price developments partly reflect a growing demand for higher-quality housing. A quality-adjusted price index for new homes (which corrects for changes in size, features, and appliances) has risen at a much slower pace, barely exceeding its long-term average in 2002 (see Chart 3), and the growing divergence between average and median price indices also suggests that price increases in recent years have tended to be concentrated at the higher end of the real estate market.

Chart 3Switch to quality

Price increases have been broadly in line with their long-term trend, with the quality-adjusted price index close to its long-term average.

Citation: 41, 1; 10.5089/9781451953411.022.A013

Sources: U.S. Census Bureau; Department of Housing and Urban Development; and National Association or Realtors.

Regionally, the housing market has exhibited more divergent patterns. At one extreme, recent increases in the real value of existing homes in the South and the Midwest represent the first sustained improvement in market conditions in more than two decades. At the other, prices in the West and Northeast—especially in the major cities—have been considerably more volatile in recent decades, reflecting boom and bust cycles in the information technology and energy sectors.

The upward trend in U.S. real estate prices has been supported by a number of factors, many of which are related to the positive economic environment of the late 1990s:

• House prices do not appear particularly out of line with disposable income (see Chart 4). Moreover, declining mortgage interest rates have allowed homeowners to reduce mortgage payments through refinancing or to seek more expensive homes at the same monthly payment.

• Gains in household wealth since the early 1990s may have finally filtered through to the housing market, and diminished expectations of continued rapid stock market gains may have persuaded investors to rebalance portfolios in favor of real estate.

• With home ownership rates increasing sharply for individuals in their 30s, the coming of age of the last cohorts of the baby-boom generation may have had a large impact on the housing market. In addition, a decline in average household size and a pickup in immigration may also have helped boost demand.

• The growing use of mortgage-backed securities has made the U.S. mortgage market significantly more efficient, reducing costs for mortgage applicants and improving access to mortgage loans for lower-income households.

Chart 4Keeping pace

House price-income ratios have been relatively steady and have even fallen when viewed in quality-adjusted terms.

Citation: 41, 1; 10.5089/9781451953411.022.A013

Sources: U.S. Census Bureau; Department of Housing and Urban Development; National Association of Realtors; and IMF staff calculations.

Lingering concerns about house prices

Notwithstanding strong fundamentals, a number of concerns about current house prices remain. First, if productivity and labor income growth falls short of expectations, the burden of mortgage debt would decline more slowly than households may anticipate, which could lead to a shift in market conditions.

Second, an increase in interest rates could affect households’ debt-service capacity, as well as dampen housing demand. Such concerns are mitigated, however, by a decline in the share of adjustable-rate mortgages to below 20 percent of newly closed mortgage loans (compared with 30 percent in the 1980s) and by the refinancing wave that has helped increase the average maturity of outstanding mortgage debt.

Finally, a growing divergence between house prices and rents is seen by some as an indication that house prices are set to decline, analogous to stock prices responding to elevated price-earnings ratios. For example, a 2003 study by John Krainer for the San Francisco Federal Reserve Bank finds that the ratio of house prices to rents currently exceeds its long-term average by 10–15 percent. He demonstrates, however, that the ratio would return to its average if rents continued to grow in line with their long-term trend and house prices were flat for two to three years—not an unusually long period of sluggish real estate markets.

What does this imply for overall house prices? A number of recent studies have discounted the possibility of a nationwide housing bubble (see Kaufman and Mühleisen, 2003, for a survey; and Case and Shiller, 2003). These studies have all concluded that prices in most areas are broadly consistent with increases in personal income although each study identified a (different) group of cities where price levels were excessive relative to fundamentals. In total, only about 20 major metropolitan areas (out of 250) were identified as being excessively priced in more than one study. However, these 20 markets include the largest metropolitan regions in the United States and may therefore account for a more substantial share of the overall housing market than their number suggests.

Empirical results presented by Kaufman and Mühleisen also indicate that house prices are largely consistent with fundamentals, especially in the Midwest and the South. In the Northeast and the West, house prices are about 15–20 percent above the level predicted on the basis of current economic conditions. Even so, weakening labor market conditions in some areas (around California’s Silicon Valley, for example) have so far had a limited effect on housing price inflation, suggesting that it would take a large drop in employment or incomes to see a significant across-the-board decline in house prices.

Not far out of line

In summary, the low U.S. household saving rate is, to a large extent, explained by the significant gains in income and wealth achieved over the past decade. Nonetheless, household spending may eventually have to slow somewhat to close the apparent gap between the personal saving rate and underlying fundamentals, and there is also the question of whether the strong trend increase in private indebtedness can be sustained in a less supportive interest rate environment. However, the balance sheet of the U.S. household sector still looks robust, partly as a result of the equity market upswing since mid-2003, and households appear to have used mortgage refinancing in part to insulate themselves from possible interest rate increases. Finally, although household exposure to the real estate market has grown, and there are signs of possible overheating in major urban markets, aggregate housing prices do not appear so far out of line with macroeconomic fundamentals that an orderly return to equilibrium cannot be achieved.

Chris Faulkner-MacDonagh is an Economist and Martin Mühleisen is a Deputy Division Chief in the IMF’s Western Hemisphere Department.


Other Resources Citing This Publication